MALDI MS: A Practical Guide to Instrumentation, Methods, and Applications, Second Edition [2 ed.] 9783527333318, 9783527335961, 3527333312

This authoritative book on MALDI MS, now finally available in its second edition and edited by one of its inventors, giv

274 19 10MB

English Pages 480 [477] Year 2013

Report DMCA / Copyright

DOWNLOAD PDF FILE

Recommend Papers

MALDI MS: A Practical Guide to Instrumentation, Methods, and Applications, Second Edition [2 ed.]
 9783527333318, 9783527335961, 3527333312

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Edited by Franz Hillenkamp and Jasna Peter-Katalinic MALDI MS

Related Titles Lovric, J.J.

Li, L. (ed.)

Introducing Proteomics – From Concepts to Sample Separation, Mass Spectrometry and Data Analysis

MALDI Mass Spectrometry for Synthetic Polymer Analysis

2011 ISBN: 978-0-470-03524-5, also available in digital formats

García, Á., Senis, Y. (eds.)

Platelet Proteomics Principles, Analysis, and Applications 2011 ISBN: 978-0-470-46337-6, also available in digital formats

Shah, H.H. (ed.)

Mass Spectrometry for Microbial Proteomics 2010 ISBN: 978-0-470-68199-2, also available in digital formats

Cole, R.B. (ed.)

Electrospray and MALDI Mass Spectrometry Fundamentals, Instrumentation, Practicalities, and Biological Applications, Second Edition 2010 Print ISBN: 978-0-471-74107-7, also available in digital formats

2010 Print ISBN: 978-0-471-77579-9, also available in digital formats

Kool, J., Niessen, W.M. (eds.)

Analyzing Biomolecular Interactions by Mass Spectrometry 2013 Print ISBN: 978-3-527-33464-3, also available in digital formats

Edited by Franz Hillenkamp and Jasna Peter-Katalinic

MALDI MS A Practical Guide to Instrumentation, Methods, and Applications Second Edition

Editors Prof. Dr. Franz Hillenkamp Institute for Medical Physics University of Münster Robert-Koch-Str. 31 48149 Münster Germany Prof. Dr. Jasna Peter-Katalinic Department of Biotechnology University of Rijeka Radmile Matejčić 2 51000 Rijeka Croatia Cover High speed time lapse photograph of IR-MALDI plumes generated with an optical parametric oscillator (OPO) laser (for more details see Fig. 1.2)

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and speciically disclaim any implied warranties of merchantability or itness for a particular purpose. No warranty can be created or extended by sales representatives or written sales materials. The Advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor authors shall be liable for any loss of proit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. Library of Congress Card No.: applied for British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliograie; detailed bibliographic data are available on the Internet at . © 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Boschstr. 12, 69469 Weinheim, Germany Wiley-Blackwell is an imprint of John Wiley & Sons, formed by the merger of Wiley’s global Scientiic, Technical, and Medical business with Blackwell Publishing. All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not speciically marked as such, are not to be considered unprotected by law. Print ISBN: 978-3-527-33331-8 ePDF ISBN: 978-3-527-67373-5 ePub ISBN: 978-3-527-67374-2 Mobi ISBN: 978-3-527-67372-8 oBook ISBN: 978-3-527-33596-1 Typesetting Toppan Best-set Premedia Limited, Hong Kong Printing and Binding Markono Print Media Pte Ltd, Singapore Cover Design Adam-Design, Weinheim, Germany Printed on acid-free paper

V

Contents

Preface to the Second Edition XI List of Contributors XIII

1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.8.1 1.8.2 1.8.3

2

2.1 2.2 2.3 2.3.1 2.3.2 2.4 2.4.1 2.4.2 2.4.3

The MALDI Process and Method 1 Franz Hillenkamp, Thorsten W. Jaskolla, and Michael Karas Introduction 1 Analyte Incorporation 4 Absorption of the Laser Radiation 6 The Ablation/Desorption Process 9 Ionization 13 Fragmentation of MALDI Ions 17 MALDI of Noncovalent Complexes 21 The Optimal Choice of Matrix: Sample Preparation 24 Surface Preparation 27 Anchor Sample Plates 27 Matrix Additives and Inluence of the Sample Plate Surface 29 Abbreviations 30 References 31 MALDI Mass Spectrometry Instrumentation 41 Peter B. O’Connor, Klaus Dreisewerd, Kerstin Strupat, and Franz Hillenkamp Introduction 41 Lasers for MALDI-MS 42 Fragmentation of MALDI Ions 48 MALDI at Elevated Pressure 48 Tandem Mass Spectrometry of MALDI Ions 49 Mass Analyzers 52 Axial TOF Mass Spectrometers 53 Relectron TOF Mass Spectrometers 55 Tandem TOF Mass Spectrometers 56

VI

Contents

2.4.4 2.4.5 2.4.6 2.5 2.5.1 2.6 2.6.1 2.6.2 2.7 2.7.1 2.7.2 2.7.3 2.7.4 2.7.5 2.8

Orthogonal TOF Mass Analyzers 60 Tandem Mass Spectrometry in oTOF Mass Analyzers 61 Ion Detectors and Data Processing in MALDI-TOF Analyzers 62 Fourier Transform Ion Cyclotron Resonance Mass Spectrometers 64 Tandem Mass Spectrometry on FTICR Mass Spectrometers 70 Quadrupole Ion Trap Mass Spectrometers 72 RF-Only Ion Guides and LIT Mass Spectrometers 77 Tandem Mass Spectrometry on QIT Mass Spectrometers 77 Hybrid Mass Spectrometers 79 Quadrupole TOF Mass Spectrometers 79 Quadrupole FT Mass Spectrometers 80 QIT-TOF Mass Spectrometers 82 Ion Mobility oTOF Mass Spectrometers 83 Orbitrap 87 Future Directions 92 Deinitions and Acronyms 93 References 96

3

MALDI-MS in Protein Chemistry and Proteomics 105 Karin Hjernø and Ole N. Jensen Introduction 105 Sample Preparation for Protein and Peptide Analysis by MALDI-MS 108 Strategies for Using MALDI-MS in Protein Biochemistry 111 Peptide Mass Mapping of Puriied Proteins 113 Peptide Sequencing by MALDI-MS/MS 114 Analysis of Post-Translational Modiications 116 Applications of MALDI-MS in Proteomics 119 Protein Identiication by MALDI-MS Peptide Mass Mapping 119 Quantitation of Proteins by MALDI-MS 122 Computational Tools for Protein Analysis by MALDI-MS 123 Clinical Applications of MALDI-MS 124 Conclusions 125 Acknowledgments 125 References 126

3.1 3.2 3.3 3.3.1 3.3.2 3.3.3 3.4 3.4.1 3.4.2 3.5 3.6 3.7

4 4.1 4.2 4.3 4.4 4.5 4.6

MALDI-Mass Spectrometry Imaging 133 Bernhard Spengler Introduction 133 History of Mass Spectrometry Imaging (MSI) and Microprobing Techniques 136 MALDI in Micro Dimensions: Instruments and Mechanistic Differences 137 Visualization of Mass Spectrometric Information 140 Data Processing and Data Exchange 143 Matrix Deposition for High-Resolution Imaging 144

Contents

4.7 4.7.1 4.7.2 4.7.3 4.8 4.8.1 4.8.2 4.8.3 4.9 4.10 4.11 4.12

5

5.1 5.2 5.2.1 5.2.2 5.2.3 5.2.4 5.2.5 5.3 5.4 5.4.1 5.4.1.1 5.4.1.2 5.4.1.3 5.4.1.4 5.4.1.5 5.4.1.6 5.4.1.7 5.4.2 5.5 5.6 5.6.1 5.6.2 5.7

Organisms, Organs, and Tissues: MALDI Imaging at Various Lateral Resolutions 148 Phospholipid Analysis 148 Peptide Analysis 150 Drug Monitoring 154 Whole-Cell and Single-Cell Analysis 154 Cellular Analysis 156 Individually Isolated Cells 156 Direct Cellular and Subcellular Imaging 157 Cell Sorting and Capturing 158 Direct Protein Identiication and Localization 160 Identiication and Characterization: Requirements for Mass Resolution and Accuracy 162 Conclusions 163 Acknowledgments 163 References 164 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics 169 Stefan Berkenkamp, Dirk van den Boom, and Daniele Fabris Challenges in Nucleic Acid Analysis by MALDI-MS 169 Genetic Markers 175 Restriction Fragment Length Polymorphisms (RFLPs) 177 Microsatellites/Short Tandem Repeats (STRs) 177 Single Nucleotide Polymorphisms (SNPs) 178 Characterization of Base Modiications and Covalent Adducts 180 Detection of Noncovalent Complexes of Nucleic Acids 186 Assay Formats for Nucleic Acid Analysis by MALDI-MS 190 Applications in Genotyping 192 MALDI-TOF-MS SNP and Mutation Analysis 192 The PinPoint Assay 193 The PROBE Assay 195 The MassEXTEND Assay 196 The GOOD Assay 198 The Invader Assay 200 The Incorporation and Complete Chemical Cleavage Assay 202 The Restriction Fragment Mass Polymorphism Assay 203 MALDI-TOF MS for Haplotyping 203 Applications in Comparative Sequence Analysis 205 Applications in Quantitation of Nucleic Acids for Analysis of Gene Expression and Gene Ampliication 215 Analysis of DNA Mixtures and Allele Frequency Determinations in DNA Pools 215 Analysis of Gene Expression 219 Future Perspectives for the MALDI-MS Analysis of Nucleic Acids 222

VII

VIII

Contents

Acknowledgments 223 References 223 6 6.1 6.1.1 6.1.2 6.2 6.2.1 6.2.2 6.2.3 6.2.4 6.2.5 6.2.6 6.3 6.3.1 6.3.2 6.3.3 6.3.4 6.4 6.4.1 6.4.2 6.5

7 7.1 7.1.1 7.1.2 7.1.3 7.1.4 7.1.5 7.1.6 7.2 7.2.1 7.2.1.1 7.2.2

MALDI-MS of Glycans and Glycoconjugates 239 Hélène Perreault, Erika Lattová, Dijana Šagi, and Jasna Peter-Katalinic Introduction 239 Glycans in Glycoproteins: Types and Importance 239 Glycosphingolipids 241 Proiling of Glycans and Glycosphingolipids 242 Importance of Glycan Proiling and Techniques Used for This Purpose 242 Importance of Glycosphingolipid Proiling and Characterization; Techniques Used 243 MALDI-MS of Glycans and Glycoprotein Components 243 N- and O-Glycan Release 248 Preparation of Glycans for MALDI-MS Analysis 250 Preparation of Glycosphingolipids for MALDI-MS Analysis 253 Structural Determination 255 MS and MS/MS of N-Glycans 255 O-Glycosylation by MS and MS/MS 260 Exoglycosidase Arrays 262 Characterization of Glycopeptides 263 Quantitative Analysis 265 Quantitative Analysis of Glycans 265 Quantitative Analysis of Glycopeptides (e.g., i-Tag, i-Traq) 266 Conclusions 267 References 267 Lipids 273 Jürgen Schiller and Beate Fuchs Introduction 273 Why Are Lipids of Such Great Interest? 273 Problems in Lipid Analysis: A Short Comparison of the Different Methods 276 Analysis of Lipids by Mass Spectrometry 277 Capabilities and Limitations of MALDI-TOF-MS in the Field of Lipid Analysis 278 Choosing an Appropriate Matrix 278 Sample Preparation, Extraction, and Puriication 279 Analysis of Individual Lipid Classes and Their Characteristics 281 The Apolar Lipids: Diacylglycerols, Triacylglycerols, Cholesterol, and Cholesteryl Esters 281 Triacylglycerol Mixtures and Vegetable Oil Analyses 285 Zwitterionic Phospholipids: Sphingomyelin, Phosphatidylcholine, and Phosphatidylethanolamine 285

Contents

7.2.3

7.2.4 7.3 7.3.1 7.4 7.5 7.6 7.7

8 8.1 8.2 8.2.1 8.2.1.1 8.2.1.2 8.2.1.3 8.2.1.4 8.2.2 8.2.2.1 8.2.2.2 8.2.2.3 8.2.2.4 8.2.3 8.3 8.4

9 9.1 9.2 9.2.1 9.2.2 9.2.3 9.2.4 9.3 9.3.1 9.3.2 9.3.3

Acidic Phospholipids: Phosphatidic Acid, Cardiolipin, Phosphatidylglycerol, Phosphatidylserine, Phosphatidylinositol, and Phosphorylated Phosphoinositides 289 Free Fatty Acids 291 MALDI-TOF-MS of Typical Lipid Mixtures 292 Brain Lipids 296 Characterization of Typical Oxidation Products of Lipids 297 MALDI-MS Imaging 299 Combining TLC and MALDI for Lipid Analysis 301 Summary and Outlook 303 Acknowledgments 304 Abbreviations 305 References 306 MALDI-MS for Polymer Characterization 313 Liang Li Introduction 313 Technical Aspects of MALDI-MS 314 Sample Preparation Issues 314 Matrix 315 Cationization Reagent 317 Solvent 320 Solvent-Free Sample Preparation 325 Instrumental and Measurement Issues 326 Mass Resolution and Accuracy 326 Sensitivity and Dynamic Range 331 Mass Range 335 MS/MS Capability 338 Data Processing Issues 341 Attributes and Limitations of MALDI-MS 344 Conclusions and Perspectives 352 References 354 Small-Molecule Desorption/Ionization Mass Analysis 367 Lucinda H. Cohen, Fangbiao Li, Eden P. Go, and Gary Siuzdak Introduction 367 Matrix Choices for Small-Molecule MALDI 368 Organic Matrices 368 Inorganic Matrices 370 Liquid Matrices 374 Matrix-Free Approaches 374 Sample Preparation 377 Electrospray Sample Deposition 379 Analyte Derivatization 379 Analyte Pre-Concentration 380

IX

X

Contents

9.3.3.1 9.3.3.2 9.3.4 9.4 9.5 9.5.1 9.5.2 9.5.3 9.5.4 9.5.5 9.5.6 9.5.7 9.5.8 9.5.9 9.6

Prestructured Sample Supports 380 DIOS with Solid Liquid Extraction 381 Matrix Suppression 383 Qualitative Characterization of LMM Molecules 383 Analyte Quantitation by MALDI 388 Selection of IS 388 Methods for Improving Quantitative Performance 389 Quantitation of Pharmaceutical Compounds 389 Enzyme Activity and Inhibition Studies 391 Quantitative Analysis of Samples from Complex Biological Matrices 393 Environmental Applications of Quantitative MALDI 394 Separation Methods Coupled with MALDI and DIOS 397 TLC-MALDI 397 Capillary and Frontal Afinity Liquid Chromatography 399 Conclusions 402 Acknowledgments 402 Abbreviations/Acronyms 402 References 404

10

Computational Analysis of High-Throughput MALDI-TOF-MS-Based Peptide Proiling 411 Thang V. Pham and Connie R. Jimenez Introduction 411 MALDI-MS Data Preprocessing 413 A Worklow for Data Acquired on a 4800 MALDI-TOF/TOF Mass Spectrometer 416 Identiication of Peptide Ion Peaks 417 Statistical Analysis of Preprocessed Data 418 Unsupervised Methods 421 Supervised Methods 422 Concluding Remarks 426 References 426

10.1 10.2 10.2.1 10.2.2 10.3 10.3.1 10.3.2 10.4

11 11.1 11.2 11.3 11.4 11.5 11.6 11.7

Biotyping of Microorganisms 431 Markus Kostrzewa The Technique 431 Standard Identiication of Bacteria and Other Microorganisms 432 Applicability and Performance in Routine Laboratories 433 Direct Specimen Analysis 434 Subtyping 435 Resistance Testing 435 Outlook 436 References 437

Index 445

XI

Preface to the Second Edition

This book on matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS), irst published in 2006, has obviously fulilled a long felt demand among the community of bioorganic mass spectrometrists. It was sold out after only a few years. To prepare a second edition has been a considerable task. MALDIMS is still a very active and developing ield, requiring essential changes and additions to the irst edition while keeping it to a handy size and particularly staying truthful to the concept of it being a “practical guide” more than an in-depth treatment of the basics. Chapter 1 has been amended by the results of Karas and Jaskolla on new design matrices and the mechanisms of ion formation in MALDI, which has substantially added to our understanding of the processes and how to optimize them for practical applications. They have also inluenced our view of the method as such and have led to revision of other parts of the chapter. While most of the instruments described in the irst edition are still in use in many laboratories, two newcomers have literally revolutionized particularly the routine applications: the orbitrap and the ion mobility instruments. Both are now covered in much detail in Chapter 2. Proteomics had already been at a rather mature state at the time the irst edition was published. Most additions and improvements in this ield are somewhat special to a given problem and are not covered in detail in this book. Here, as in most of the other applications, the reader is referred to the extensive list of original literature at the end of each chapter. Chapter 4 has been essentially rewritten. It now concentrates on MALDI imaging, a ield which has seen a dramatic development in recent years and promises to continue on this path. The discussion of biomarkers has been referred to the new Chapter 10 on bioinformatics. MALDI imaging requires the treatment of the raw data with rather sophisticated software tools, speciic for this application. It is, therefore, contained in Chapter 4, rather than Chapter 10. MALDI-MS of nucleic acids, covered in Chapter 5, has still not found as widespread an application as, for example, the analysis of proteins, mostly because of competing techniques such as second generation sequencers which are now in routine use. One interesting MALDI application is the analysis of RNA and other

XII

Preface to the Second Edition

modiied nucleic acids, where straight sequencing leads to a loss of important information. Application of MALDI-MS to analysis of protein-linked N- and O-glycans in Chapter 6 has been generally revised and updated. In consideration of growing attention to glycomics in biology and medicine, eficient protocols for glycans and glycopeptides have been described to encompass the carbohydrate complexity both for rapid mapping as well as for quantiication, and those for glycosphingolipids added. Revisions in Chapter 7 are related to the increased interest in lipid analysis, notably boosted by introduction of the new “omics” ield – lipidomics – and by developments of MS imaging as a robust new application of MALDI-MS. In addition, novel potentials of lipid analysis in applications of the direct desorption from solid surfaces and MALDI-MS imaging to diagnostics using lipids as disease markers are described. Chapter 8, on the analysis of synthetic polymers, remained essentially unchanged in the second edition. Chapter 9, on small molecule desorption/ionization mass analysis, relects the use of MALDI-MS in the development of new pharmaceutical agents, again a ield of important applications and developments. Bioinformatics, now covered in the new Chapter 10, was obviously missing in the irst edition. Thang V. Pham and Connie R. Jimenez of the Free University of Amsterdam describe how they use bioinformatics software in their search for tumor markers in human samples. While this is a rather special application, they have taken great care to refer the reader to the original literature which describes the principles in other ields of application. Biotyping of microorganisms, covered in Chapter 11, has been added as a last minute topic. This chapter is not as comprehensive as the others, but we considered it important, because it is the irst and so far only large scale routine clinical MALDI application. It has boomed over the last two years and is still in a phase of intense development. The most important aspects of this new application are discussed in the chapter; for details, readers are referred to the many listed references. Münster, Rijeka, April 2013

Franz Hillenkamp Jasna Peter-Katalinic

XIII

List of Contributors

Stefan Berkenkamp SEQUENOM Inc. 3595 John Hopkins Court San Diego, CA 92121 USA sberkenkamp@sequenom Dirk van den Boom SEQUENOM Inc. 3595 John Hopkins Court San Diego, CA 92121 USA [email protected] Lucinda H. Cohen Merck Research Laboratories DMPK Bioanalytical Group Mail Stop RY800B201 Rahway, NJ 07065 USA [email protected] Klaus Dreisewerd University of Münster Institute for Medical Physics and Biophysics Robert-Koch-Str. 31 48149 Münster Germany [email protected]

Daniele Fabris University at Albany The RNA Institute 1400 Washington Avenue Albany, NY 1222 USA [email protected] Beate Fuchs University of Leipzig Institute of Medical Physics and Biophysics, Faculty of Medicine Härtelstr. 16–18 04109 Leipzig Germany [email protected] Eden P. Go Department of Chemistry University of Kansas Lawrence, KS 66047 USA [email protected] Franz Hillenkamp University of Münster Institute for Medical Physics and Biophysics Robert-Koch-Str. 31 48149 Münster Germany [email protected]

XIV

List of Contributors

Karin Hjernø University of Southern Denmark Department of Biochemistry and Molecular Biology Campusvej 55 5230 Odense Denmark [email protected]

Markus Kostrzewa Vice President – Clinical Mass Spectrometry Bruker Daltonik GmbH Fahrenheitstr. 4 28359 Bremen Germany [email protected]

Thorsten W. Jaskolla University of Münster Institute for Medical Physics and Biophysics Robert-Koch-Str. 31 D-48149 Münster Germany [email protected]

Erika Lattová Department of Chemistry University of Manitoba 144 Dysart Road Winnipeg, MB R3T 2N2 Canada [email protected]

Ole N. Jensen University of Southern Denmark Department of Biochemistry and Molecular Biology Campusvej 55 5230 Odense Denmark [email protected] Connie R. Jimenez VU University Medical Center VUmc-Cancer Center Amsterdam, Department of Medical Oncology CCA 1-46, OncoProteomics Laboratory De Boelelaan 1117 1081 HV Amsterdam The Netherlands [email protected] Michael Karas Johann Wolfgang Goethe University of Frankfurt Institute of Pharmaceutical Chemistry, Biocenter Max-von-Laue-Str. 9 60438 Frankfurt am Main Germany [email protected]

Fangbiao Li Merck Research Laboratories DMPK Bioanalytical Group Mail Stop RY800B201 Rahway, NJ 07065 USA [email protected] Liang Li University of Alberta Department of Chemistry Chemistry Centre W3-39 Edmonton, AB T6G 2G2 Canada [email protected] Peter B. O’Connor University of Warwick Department of Chemistry Gibbet Hill Road Coventry CV4 7AL UK [email protected] Hélène Perreault Department of Chemistry University of Manitoba 144 Dysart Road Winnipeg, MB R3T 2N2 Canada [email protected]

List of Contributors

Jasna Peter-Katalinic Department of Biotechnology University of Rijeka Radmile Matejčić 2 51000 Rijeka Croatia [email protected] Thang V. Pham VU University Medical Center VUmc-Cancer Center Amsterdam, Department of Medical Oncology CCA 1-46, OncoProteomics Laboratory De Boelelaan 1117 1081 HV Amsterdam The Netherlands [email protected] Dijana Šagi Sanoi-Aventis Deutchland GmbH Industriepark Höchst, Geb. H773 65926 Frankfurt am Main Germany [email protected] Jürgen Schiller University of Leipzig Institute of Medical Physics and Biophysics, Faculty of Medicine Härtelstr. 16–18 04109 Leipzig Germany [email protected] -leipzig.de

Gary Siuzdak The Scripps Research Institute Center for Metabolomics and Mass Spectrometry Departments of Chemistry, Molecular and Computational Biology BCC007 10550 North Torrey Pines Road La Jolla, CA 92037 USA [email protected] Bernhard Spengler Justus Liebig University Giessen Institute of Inorganic and Analytical Chemistry Schubertstr. 60, Bldg 16 35392 Giessen Germany [email protected]. uni-giessen.de Kerstin Strupat Thermo Fisher Scientiic Life Science Mass Spectrometry Hanna-Kunath-Str. 11 28199 Bremen Germany [email protected]

XV

1

1 The MALDI Process and Method Franz Hillenkamp, Thorsten W. Jaskolla, and Michael Karas

1.1 Introduction

Matrix-assisted laser desorption/ionization (MALDI) is one of the two “soft” ionization techniques besides electrospray ionization (ESI) which allow for the sensitive detection of large, nonvolatile and labile molecules by mass spectrometry. Over the past 27 years, MALDI has developed into an indispensable tool in analytical chemistry, and in analytical biochemistry in particular. In this chapter, the reader will be introduced to the technology as it stands now, and some of the underlying physical and chemical mechanisms as far as they have been investigated and clariied to date will be discussed. Attention will also be focused on the central issues of MALDI, that are necessary for the user to understand for the eficient application of this technique. As an in-depth discussion of these topics is beyond the scope of this chapter, the reader is referred to recent reviews [1–4]. Details of the current state of instrumentation, including lasers and their coupling to mass spectrometers, will be presented in Chapter 2. As with most new technologies, MALDI came as rather a surprise even to the experts in the ield on the one hand, but also evolved from a diversity of prior art and knowledge on the other hand. The original notion had been that (bio)molecules with masses in excess of about 500–1000 Da could not be isolated out of their natural (e.g., aqueous) environment, and even less be charged for an analysis in the vacuum of a mass spectrometer without excessive and unspeciic fragmentation. During the late 1960s, however, Beckey introduced ield desorption (FD), the irst technique to open a small road into the territory of mass spectrometry (MS) of bioorganic molecules [5]. Next came secondary ion mass spectrometry (SIMS), and in particular static SIMS, as introduced by A. Benninghoven in 1975 [6]. This development was taken a step further by M. Barber in 1981, with the bombardment of organic compounds dissolved in glycerol with high-energy atoms, which Barber coined fast atom bombardment (FAB). It was in this context, and in conjunction with the irst attempts to desorb organic molecules with laser irradiation, that the concept of a “matrix” as a means of facilitating desorption and enhancing ion MALDI MS: A Practical Guide to Instrumentation, Methods, and Applications, Second Edition. Edited by Franz Hillenkamp and Jasna Peter-Katalinic. © 2014 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2014 by Wiley-VCH Verlag GmbH & Co. KGaA.

2

1 The MALDI Process and Method

yield was born [7]. The principle of desorption by the bombardment of organic samples with the ission products of the 252Cf nuclear decay, later termed plasma desorption (PD), was irst described by R. Macfarlane in 1974 [8]. Subsequently, the groups of Sundqvist and Roepstoff greatly improved the analytical potential of this technique by the addition of nitrocellulose, which not only cleaned up the sample but was also suspected of functioning as a signal-enhancing matrix [9]. The irst attempts at using laser radiation to generate ions for a mass spectrometric analysis were reported only a few years after the invention of the laser [10, 11]. Vastola and Pirone had already demonstrated the possibility of recording the spectra of organic compounds with a time-of-light (TOF) mass spectrometer. Subsequently, several groups continued to pursue this line of research, mainly R. Cotter at Johns Hopkins University in the USA and P. Kistemaker at the FOM Institute in Amsterdam, the Netherlands. Indeed, for a number of years the Amsterdam group held the high-mass record for a bioorganic analyte with a spectrum of underivatized digitonin at mass 1251 Da ([M + Na]+), desorbed with a CO2-laser at a wavelength of 10.6 μm in the far infrared (IR) [12]. Independently of, and parallel to, these groups, Hillenkamp and Kaufmann developed the laser microprobe mass analyzer (LAMMA) [13], the commercial version of which was marketed by Leybold Heraeus in Cologne, Germany and which is now on exhibition in the section on New Technologies of the Deutsches Museum in Munich, Germany. The instrument originally comprised a frequencydoubled ruby laser at a wavelength of 347 nm in the near ultraviolet (UV), and later a frequency-quadrupled Nd:YAG-laser at a wavelength of 266 nm in the far UV. The laser beam was focused to a spot of ≤1 μm in diameter to probe thin tissue sections for inorganic ions and trace elements such as Na, K, and Fe. The mass analyzer of the LAMMA instruments was also a TOF mass spectrometer, and was the irst commercial instrument with an ion relector, which had been invented a few years earlier by B.A. Mamyrin in Leningrad. The sensitivity-limiting “noise” of the LAMMA spectra were signals that were soon identiied as coming from the organic polymer used to embed the tissue sections, as well as other organic tissue constituents. It was this background noise which triggered the search for a systematic analysis of organic samples and which eventually led to the discovery of the MALDI principle in 1984. The principle and its acronym were irst described in 1985 [14], and the irst spectrum of the nonvolatile bee venom mellitin, an oligopeptide at mass 2845 Da, in 1986 [15]. Spectra of proteins with masses exceeding 10 kDa and 100 kDa were reported in 1988 [16], and details presented at the International Mass Spectrometry Conference in Bordeaux in 1988, respectively. Both, ESI and MALDI were developed independently but concurrently, and when their potential for the desorption of nonvolatile, fragile (bio)molecules was discovered, the scientiic community was mostly impressed by the ability of these techniques to access the high mass range, particularly of proteins. However, FAB- and PD-MS had at that time already generated spectra of trypsin at mass 23 kDa and other high-mass proteins. What really made the difference in particular for the biologists was the stunning sensitivity which, for the irst time, made

1.1 Introduction

MS compatible with sample preparation techniques used in these ields. For MALDI, the minimum amount of protein needed for a spectrum of high quality was reduced from 1 pmol in 1988 to a few femtomoles only about a year later. Today, in favorable cases, the level is now down in the low attomole range. Many other developments – both instrumental (see Chapter 2) as well as speciic sample preparation recipes and assays (see other chapters of the book) – took place during the following decade, and the joint impact of all of these together has today made MALDI-MS an indispensable tool not only in the life sciences but also in polymer analysis, food sciences, pharmaceutical drug discovery, or forensic jurisprudence. The use of a chemical matrix in the form of small, laser-absorbing organic molecules in large excess over the analyte is at the core of the MALDI principle. Several developments for laser desorption schemes took place in parallel to and following publications of the MALDI principle. These all attempted to replace the chemical matrix by a more easy-to-handle physical matrix, or a more simple combination of the two. The best known of these was the system of Tanaka and coworkers, which was irst presented at a Sino-Japanese conference in 1987; the details were subsequently published in 1988 [17]. The matrix comprised cobaltnanoparticles suspended in glycerol as the basic system into which the analyte was dissolved, similar to the sample preparation of FAB. Several other nano- and micro-particles were tested later, and results obtained that were comparable to those of Tanaka [18]. For his technique of surface-assisted laser desorption/ionization (SALDI), Sunner and coworkers used dry carbon and graphite substrates [19]. Another technique which has attracted much interest for the analysis of smaller molecules (and which is described in more detail in Chapter 9) was reported by Siuzdak [20]. This method, termed desorption/ionization on silicon (DIOS), uses preparations of neat organic samples on porous silicon. Several other methods and acronyms use similar systems, such as nanowires or sol–gel systems. All of these methods use the substrate on which the analyte is prepared for the absorption of the laser energy, and are characterized by a sensitivity which is lower than that of MALDI by several orders of magnitude, as well as a strongly increased ion fragmentation which limits the accessible mass range to somewhere between 2000 and 30 000 Da, depending on the method. There is reason to believe that all of these methods are based on a thermal desorption at the substrate/analyte interface, with the high internal excitation of the ions and low ion yield typical for thermal desorption processes. The very high heating and cooling rates, together with high peak temperatures of the substrates as well as the suspension of the absorbers in glycerol, apparently soften the desorption somewhat, the latter most probably through adiabatic cooling in the expanding plume; derivatization of the surfaces can up-concentrate the analyte of interest at the surfaces to increase the sensitivity. Indeed, a yoctomole (10−21 mole) sensitivity has been achieved in this way with a perluorophenyl-derivatized DIOS system for a small hydrophobic peptide [21]. Recently, another technique termed matrix-assisted inlet ionization (MAII) was described by Trimpin et al., which enables the generation of multiply charged

3

4

1 The MALDI Process and Method

analyte ions similarly to those observed with ESI [22]. Although the analytes to ionize are also cocrystallized with typical MALDI matrices, the energy required for material ablation can be supplied by laser irradiation (laserspray ionization with laser pulse energies of about 10-fold that typically used for MALDI MS [22]), but is not limited to it [23]. It is assumed that analyte ion generation occurs independently of, and subsequent to, the ablation process in a heated inlet tube connecting the atmospheric pressure source to the vacuum of the mass spectrometer. Until now, the process of charge generation has not been completely understood; however, matrix evaporation of ablated highly charged clusters/droplets within the heated tube seems to explain the generation of multiply charged analytes. Due to the aforementioned differences to common MALDI MS, this technique is not discussed in this book.

1.2 Analyte Incorporation

What, then, is so special about the chemical matrix in MALDI? Some of its important features, such as the absorption of the laser energy, are easily understood, but rather surprisingly the overall process of the desorption and ionization has not yet been fully described, almost 30 years after its invention. Considerable progress regarding the mechanism of analyte desorption and protonation was recently achieved [24, 25]. Meanwhile, the search for better (i.e., more sensitive) matrices does not remain completely empirical, as some of the critical parameters for eficient analyte protonation (see Section 1.5) are uncovered, although other aspects such as prediction and targeted manipulation of the matrix morphology remain [26]. One important feature is the way in which the matrix and analyte interact in the MALDI sample. In a typical UV-MALDI “dried droplet” sample preparation, small volumes of an about 10−6–10−9 M solution of the analyte and a near-saturated (ca. 0.01–0.1 M) solution of the matrix are mixed; the solvent is then evaporated before the sample is introduced into the vacuum of the mass spectrometer. Upon solvent evaporation, the matrix crystallizes in a speciic morphology to form a bed of small crystals that range in size from a few to a few hundred micrometers, depending on the matrix, the solvent, the substrate surface characteristics, and further details of the preparation. The typical molar analyte-to-matrix ratio ranges from about 10−2 for less-sensitive compounds such as poorly protonable drugs without basic functionalities, to approximately 10−7 for highly sensitive analytes, for example, quaternary ammonium-derivatives such as phosphatidylcholines or many basic peptides. The sample preparation is discussed in more detail in Section 1.8. One of the early surprises in MALDI development was that all of the well-functioning matrices known at that time incorporated the analytes in the crystals quantitatively (up to a maximum concentration), and in a homogeneous (on the light microscopic resolution level of 0.5 μm) distribution. This was shown for the UV-MALDI matrices 2,5-dihydroxybenzoic acid (2,5-DHB) [27], sinapic acid [28], α-cyano-4-

1.2 Analyte Incorporation

hydroxycinnamic acid (HCCA) [29], and 3-hydroxypicolinic acid [29] as well as the IR-MALDI matrix succinic acid [30]. This homogeneous incorporation, in conjunction with the also homogeneous energy deposition and material ablation (for a discussion, see Section 1.3) resulted in the codesorption of intact nonvolatile and labile molecules with the matrix and, in addition, in a cooling of their internal energy in the expanding plume of material. Although the mechanisms and driving force for analyte incorporation are still largely unknown, attractive ion–ion interactions between dissolved protonated analytes and matrix acid anions during sample preparation seem to alleviate analyte inclusion/incorporation within the growing matrix crystals [26]. At typically slightly acidic sample preparation conditions, many analytes (such as most peptides and proteins) carry net positive charges due to protonated basic and predominantly neutralized acidic functionalities. Although matrix acid functionalities are also mostly neutralized under such pH conditions, the large molar matrix-to-analyte excess effects matrix anion amounts suficiently high for at least stoichiometric analyte interactions in solution. These ion–ion interactions might provide the explanation of why almost all compounds used as matrices for the analysis of basic group containing analytes – which is the case for many natural drug classes – exhibit acid functionalities. Neutral α-cyanocinnamic acid (CCA) derivatives, for example, matrix amides or esters with weaker ion–dipole interactions between protonated analytes and neutral matrices, indeed resulted in strongly diminished analyte signal intensities. Nevertheless, further prerequisites for successful analyte inclusion/incorporation presumably must be fulilled. Krueger et al. found clear proof for homogeneous analyte incorporation by using colored pH indicators [31], whereas Horneffer et al. have shown in a systematic study of different position isomers of dihydroxybenzoic acids that only 2,5-DHB incorporates homogeneously and quantitatively, whereas other isomers such as 2,6-DHB do not incorporate at all, while some others incorporate only randomly [32]. Confocal laser scanning images of the protein avidin, labeled with the luorochrome Texas Red for single crystals of 2,5-DHB and 2,6-DHB, are shown in Figure 1.1. No obvious correlation between analyte incorporation and the crystal structure of these isomers was found. The state of the incorporated analyte molecules in the matrix crystals is another interesting question. Based on results obtained for the incorporation of pH-indicator dye molecules, Krueger et al. have concluded that molecules retain their solution charge state in the crystal, which implies that they also retain their solvation shell [31]. Horneffer et al. have found a high density of cavities of 10–2000 nm size in crystals of both 2,5-DHB and 2,6-DHB by electron microscopy [33]. At irst sight, these cavities could be assumed to contain analyte molecules with residual solvent. However, if this is the case it is dificult to understand why 2,5-DHB – but not 2,6-DHB – incorporates analytes into these cavities; attempts to localize goldlabeled proteins in the cavities of 2,5-DHB were also inconclusive [33]. A solventless method for sample preparation was developed originally for the MALDI-MS of synthetic polymers, which often are not soluble in standard solvents [34]. In this method, matrix and analyte powders are mixed and ground in a mortar

5

6

1 The MALDI Process and Method (a)

(b)

(c)

(d)

Figure 1.1 Confocal laser scan luorescence

images of single crystals of (a, b) 2,5-dihydroxybenzoic acid and (c, d) 2,6-dihydroxybenzoic acid. Both matrices were doped with the protein avidin, labeled with a Texas Red luorescent dye. The images were recorded at a x,y-plane 12 μm into the

crystals. Panels (a) and (c) show dark shadowgraphs of the shape of the crystals against the bright green BODIPY 493/503 luorescence of the immersion liquid (false color photography). Panels (b) and (d) show the red Texas Red luorescence of the labeled proteins.

or ball-mill and then applied to a MALDI target support. It was shown that analyte spectra can be obtained from such preparations, even though the analyte is only chemisorbed at the matrix crystal surfaces [35]. However, the desorption is much less “soft” than MALDI-MS from samples with incorporated analytes, leading to a strongly increased metastable fragmentation of the ions and an upper mass limit for proteins of 30–55 kDa.

1.3 Absorption of the Laser Radiation

The role of the optical absorption of the matrix in the transfer of energy from the laser beam to the sample is governed by Beer’s law [14] H = H 0 ∗ e − αz

(1.1)

where H is the laser luence at depth z into the sample, H0 is the laser luence at the sample surface, and α is the absorption coeficient (see Chapter 2, Section 2.2 for a deinition of the luence). The absorption coeficient α equals the product of the wavelength-dependent molar absorption coeficient αn which is a property of the matrix compound and the concentration cn of the absorbing molecules in the sample. The molar absorption coeficient αn has a maximum value for UV-MALDI irradiation of typical matrices between 5 × 103 and 5 × 104 l mol−1 cm−1 at the peak absorption wavelength. Molar absorption coeficients of this order of magnitude and at low wavelengths in the range of 300–400 nm are only pro-

1.3 Absorption of the Laser Radiation

vided by molecules with aromatic systems (typical matrix structures for instance contain phenyl or styryl derivatives) supported by electron-donating groups such as hydroxy residues. The exact wavelength of maximum absorption and its magnitude are determined by the position and nature of the ligands of the core ring, and are tabulated in a variety of reference sources. Some care should be exercised in using the tabulated values for αn, because they all refer to dilute solutions of the compounds. Compared to the absorption proiles of dissolved compounds, the absorption bands of MALDI samples in the solid state are typically broadened and slightly red- or blue-shifted in dependence on the strength of the chromophore–solvent interactions of the dissolved compounds [25]. The concentration cn of absorbers (chromophores) is unusually high in solid-state MALDI samples (about 10 mol l−1), taking into account the typical solid-state density of crystals of roughly 2 g cm−3 (e.g., 2,3-dihydroxybenzoic acid exhibits a density of 1.54 g cm−3), because all of the solvent is evaporated before the sample is introduced into the vacuum. As a result, the typical UV absorption coeficient α ranges from about 5 × 104 to 5 × 105 cm−1 at a laser wavelength of 337 or 355 nm. The inverse of α is called the penetration depth δ, and this has values of only 20 to 200 nm. It is the depth into the sample, at which the luence has decreased to about 30% of the value at the surface. It is also an order of magnitude estimate of the depth of material ablated (desorbed) per single laser pulse in MALDI. Because of this very shallow ablation depth, a given location of the sample can usually be irradiated many times before the material is exhausted. For the MALDI process, the energy absorbed per unit volume Ea/V of the sample (loosely called “energy density”) is the process-determining quantity. This can be derived from Eq. (1.1) by simple differentiation to: E a /V = α ∗ H

(1.2)

Equation (1.2) is at the core of the MALDI process. If a matrix is chosen with a suficiently high absorption coeficient α, a relatively low luence H0 sufices for achieving the critical “energy density” necessary to initialize ablation and ionization of a top layer of the sample. Values for H0 of 50–500 J m−2 are representative for most UV-MALDI applications. As discussed in Chapter 2, Section 2.2, pulsed lasers with a pulse width of a few nanoseconds are employed in UV-MALDI. At a luence of about 100 J m−2 and a pulse width of 2 ns, the “intensity” (irradiance) of the laser beam at the sample surface is only 1011 W m−2 or 107 W cm−2, which is not enough to induce any nonlinear absorption such as nonresonant two-photon absorption. For the linear absorption, the absorbed energy per unit volume can be controlled meticulously with a suitable variable attenuator in the laser beam, a feature which has emerged as being crucial for the successful MALDI of large molecules, because the desorption of nonvolatile, labile molecules can only be achieved in a narrow range of energy “density” between low-energy conditions insuficient for ablation and ionization and high-energy conditions leading to extensive analyte fragmentation (see Section 1.6). The other essential feature of this laser absorption is that the energy is transferred more or less uniformly to a macroscopic sample volume (except for

7

8

1 The MALDI Process and Method

the attenuation of the luence into the sample and the luence proile, as discussed in Chapter 2, Section 2.2). This is very different from the situation in SIMS or PD, where incident particles create minute tracks of atomic dimensions of very high energy density in the sample, with a strong radial decline of energy density. This strongly heterogeneous energy distribution is the main reason for the limitation of these methods for the intact desorption of larger molecules. The luence can also be converted into a value for the photon lux – that is, the number of photons impinging on the sample per single laser pulse. A typical luence of 100 J m−2 [36] corresponds to a photon lux of 1.7 × 1016 photons per cm2; each carrying an energy of 3.7 eV at the wavelength of 337 nm of the N2 laser. A molar absorption coeficient of 104 l mol−1 cm−1 represents a physical absorption cross-section of the chromophore of 1.6 × 10−17 cm2, resulting in an average of 0.3 photons absorbed per surface matrix molecule (about 110 kJ mol–1 matrix for 337 nm photons) for any given laser exposure at this luence. For these considerations, it is assumed that the vast majority of electronic excitation energy is converted into lattice energy by internal conversion (as compared to processes such as luorescence and chemical reactions). This is a very high density of excitation energy, close to the sum of all noncovalent bond energies of the ablated matrix volume. It is, therefore, not surprising that such a large amount of deposited energy leads to an explosive ablation of the excited sample volume. On the other hand, it renders even resonant twophoton absorption by the matrix rather unlikely. The high density of excited molecules does, however, result in a rather high rate of energy pooling in the sample, in which two neighboring electronically excited molecules pool their energy, with one of them acquiring twice the energy of the irst excited singlet state (S1,v = 0) and the other falling back to its ground state [37]. This energy pooling is an important feature in some models for the ionization of the matrix molecules, which requires at least the energy of two photons for an initial photoionization of the matrix molecules [3, 38]. It is elucidated in more detail in Section 1.5. The situation is similar, but not equal, for IR-MALDI. Optical absorption in the IR region of the spectrum represents a transition between vibrational and/ or rotational molecular states. The probabilities for these transitions are typically lower than the electronic transitions in the UV by one to two orders of magnitude. The strongest such transitions are those of the O–H and N–H stretch vibrations near a 3 μm wavelength. The absorption coeficient of water or vacuumstable ice, but also of the common IR-MALDI matrix glycerol, reaches peak values of 104 cm−1 in this wavelength region, corresponding to a penetration depth of about 1 μm, which is more than 20-fold that of typical penetration depths in the UV. As a result, the ablated mass per laser exposure in IR-MALDI exceeds that of UV-MALDI by at least a factor of 10, and the sample consumption rate is accordingly higher. Typical laser luences for IR-MALDI range from 103 to 5 × 103 J m−2. Nonlinear absorption processes are even less likely for such luences in the IR- as compared to UV-MALDI, and for the photon energy of only 0.4 eV or less even the absorption of several photons by a given chromophore or energy pooling cannot possibly excite single molecules to anywhere near their ionization energy.

1.4 The Ablation/Desorption Process

1.4 The Ablation/Desorption Process

As discussed above, every laser exposure of a sample leads to the removal of a bulk volume – that is, many monolayers of matrix molecules of the sample. The term “desorption” is, therefore, somewhat ill-chosen for this process, and was so even for the ield desorption for which it was originally coined. Ablation (removal of bulk material from surfaces) is the more speciic term, and is used interchangeably with desorption throughout this chapter. The processes of material ablation and the ionization of a minor fraction of the matrix and analyte molecules are, no doubt, intimately intertwined, and both take place on a micrometer geometric and a nanosecond time scale. It is experimentally very dificult – if not impossible – to sort out the complex contributions of the physical processes induced by the laser irradiation in all detail. Despite this complexity, it is of considerable merit to treat the ablation and ionization mechanisms separately. From such a discussion, some basic understanding can be derived, particularly, because the vast majority of the ablated material comes off neutral. Regarding the energy loss out of the excited sample volume during the laser pulse, at least two situations need to be considered which are known as “thermal” and “stress coninement” regimes: Energy dissipation by heat conduction during the laser pulse can be neglected in all cases of UV- as well as IR-MALDI. For a penetration depth of UV-laser radiation of 100 nm, the time constant for heat conduction of typical UV-MALDI matrices is about 10 ns [39] – still a factor of two to three longer than the typical laser pulse width (typically 3 ns for nitrogen lasers and 4–7 ns for Nd:YAG lasers). In the IR, the 10- to 1000-fold smaller absorption coeficients compared to UV matrices [40] causes penetration depths of about 1 μm with corresponding heat conduction time constants of about 1 μs, a factor of about 10 longer than the longest pulse width of lasers (Er:YAG) used in that case. Heat conduction is therefore not important as an energy loss process in MALDI. The very rapid heating of the sample by the laser radiation resulting from a comparably slow heat conduction is known as a “thermal coninement” regime. This will generate a thermoelastic pressure pulse in the absorbing sample volume which travels out of the excited volume at the speed of sound, carrying away part of the deposited energy. With the speed of sound in typical crystalline matrices of 2000–3000 ms−1 and depths of roughly 100 nm in the UV, the acoustic time constant is less than 100 ps, much shorter than the laser pulse width of a few nanoseconds. Even though energy is constantly carried away by the pressure wave, this amounts only to a very small fraction of the total deposited energy. Due to the comparably short transport time for propagation of the pressure pulses out of the irradiated volume by using nanosecond-lasers, the pressure within the excited volume never reaches values high enough to substantially inluence the ablation process [39]. For IR-MALDI, the situation can be very different because of the larger penetration depth, resulting in a larger acoustic time constant of about 1 ns. For the desorption with an Er:YAG laser, the pulse width of 100 ns is long compared to

9

10

1 The MALDI Process and Method

the acoustic time constant, with only a negligible pressure build-up in the excited volume. The pulse width of the optical parametric oscillator (OPO) laser of only 6 ns, however, is on the same order of magnitude as the acoustic time constant, and the system stays close to what is called the “acoustic coninement” or “stress coninement.” In this case a very high pressure of several tens of megapascals can build up in the excited volume, which can in turn cause spallation with the generation of microcracks and the ejection of larger bulks of material [40]. Rohling et al. [41] have investigated the ablation processes by measuring the recoil pressure of the ablated material with a fast acoustic transducer onto which the sample was prepared, while Leisner et al. [42] have studied the expanding plume of ablated material with high-speed time-lapse photography, both at a wavelength of 2.94 μm. Both measurements were much easier for IR-MALDI and glycerol as a matrix, because of the greater amount of material ablated. For the short pulse width and near-acoustic coninement, these authors saw pressure pulses of very high amplitude as expected, and time durations comparable to the laser pulse. For the 100-ns pulses of the Er:YAG laser, the pressure amplitude was low, but lasted for several microseconds. The plume photographs revealed that material is removed from the sample for times of up to over 100 μs in both cases. This is certainly somewhat of a surprise, because the TOF analysis had revealed that the ions are only generated during an initial phase of not longer than about 300 ns using glycerol [43]. Similar experiments were conducted by Rohling under UV-MALDI conditions [44], using the liquid matrix nitrobenzylalcohol for better sample homogeneity and a desorption wavelength of 266 nm. Expectedly, the recoil pressure was very low – lower even than that of the long-pulse IR-laser – because of the smaller amount of removed material. The recoil pressure pulse lasted for only less than 25 ns, the time resolution of the detection. The plume photographs revealed a material ejection for up to at least several microseconds, again much longer than the ion generation time of at most a few nanoseconds. Some typical plume photographs are shown in Figures 1.2 (IR-irradiation) [42] and 1.3 (UV-irradiation) [44]. The results of these experiments can tentatively be explained by the following models. In IR-MALDI with 100 ns-long Er:YAG-laser pulses, the absorbing volume is superheated to a temperature that is substantially above the boiling temperature due to a slower gas-phase bubble formation and heat diffusion, followed by an explosive volume ejection of material through boiling by heterogeneous nucleation [40]. The situation is similar for UV-MALDI. The longer time course of material ejection in the IR as compared to the UV is caused by a deeper penetration of the radiation into the sample, and a correspondingly higher inertia and residual heat of the excited volume. For the 6-ns pulses in IR-MALDI of the OPO-laser, the ablation process is substantially different. The strong thermoelastic wave is relected at the sample vacuum interface, thereby reversing its phase. It then travels back into the sample as a tensile wave, transferring the material beyond the liquid spinodal, as described by Vogel and Venugopalan for soft-tissue ablation [45]. Upon this transition, the material goes through a phase explosion by homogeneous nucleation. Even though all of these experiments were conducted on liquid samples to keep reproducibility high, they relect, most probably, also the

1.4 The Ablation/Desorption Process (a)

(b)

Figure 1.2 High-speed time-lapse photo-

graphs of IR-MALDI plumes generated with an optical parametric oscillator laser with 6-ns pulse width (left panels) and an Er:YAG-laser with 100-ns pulse width (right panels). Both lasers were operated at 2.94 μm wavelength. Matrix, glycerol; time resolution, 8 ns; spatial resolution, 4 μm. The top three panels represent gradients of gaseous material density creating gradients

of the index of refraction in the plume, recorded in a dark-ield illumination mode. The lowest panel represents particle emission in the plume after some μs recorded with light scattered at 90° to the illumination beam. The thin lowest lines indicate the top surface of the glycerol drop; the other striations in the dark-ield images are artifacts of optical interference.

situation for crystalline solid samples. A contribution by the gaseous components such as CO2 through thermal decomposition of matrix molecules as well as trans– cis-photoisomerizations of matrices with, for example, cinnamic acid core structures are also discussed as a source of pressure build-up in the excited volume. Theoretical models for the ionization as well as molecular modeling suggest that the ablation process generates large amounts of (molten) clusters and material particles, besides gaseous components [46]. During the early expansion phase of the plume, dark-ield images reveal homogeneous gradients of the index of refraction. However, the spatial resolution of these plume photographs is only a few micrometers; a distribution of clusters of small size, expected during the early phase of the plume expansion, cannot therefore be revealed by this experimental set-up. Nevertheless, nanometer-size particles have indeed been detected by light scattering upon growing the ablated particles by butanol condensation in a cooled tube, although these experiments have been carried out under atmospheric

11

12

1 The MALDI Process and Method

(a)

Figure 1.3 High-speed photographic images

of UV-MALDI plumes generated with a frequency-quadrupled Nd:YAG laser of 266 nm wavelength and 8 ns pulse width. Matrix, nitrobenzylalcohol; time resolution, 8 ns; spatial resolution, 4 μm. (a) Dark-ield

(b)

image, 45 ns after laser exposure; (b) 90° scattered light image, 311 ns after exposure. The thin lowest line indicates the top surface of the nitrobenzylalcohol drop; the other striations in the dark-ield image are artifacts of optical interference.

pressure conditions [47]. In cases when clusters carry net charges they might release analyte ions upon quantitative dissociation (see Section 1.5), provided that suficient internal thermal cluster energy is available. This energy can derive from chemical intracluster reactions which might be, for example, proton-transfer reactions, fragmentation, dissociation of [2+2]-photodimerized matrices [48], or thermal relaxation of cis- to trans-isomers after trans–cis-photoisomerization. On the other hand, clusters can lose energy by adiabatic expansion in the plume, and the majority of clusters most probably never dissociate completely [49]. In cases of insuficient internal energy, incomplete cluster dissociation of clusters with net charges contributes to the chemical noise/background of mass spectra [49]. Most probably, large clusters (particles) do not contain suficient internal energy for quantitative dissociation and are stable at least on a nanosecond time scale. This assumption of stable large clusters is further supported by the generation of highly charged analyte ions inside a heated inlet tube attached to the ion source, which is assumed to provide the lacking energy for the complete dissociation of otherwise stable large clusters containing analytes [23], although the parameters used do not correspond with commonly used MALDI conditions. In addition, the origin of the cluster charges generated by this approach remains unclear, since analyte ion generation is not limited to lasers as an energy supply [23]. Consequently, large cluster/particle emission does not seem to be relevant for typical MALDI ion generation. Garrison, Zhigilei [40, 50] and coworkers, as well as Knochenmuss [38], have modeled the ablation process using molecular dynamics simulations. Qualitatively, these simulations correctly predict many of the features observed experimentally. For example, corresponding to model calculations of Zhigilei and coworkers, mainly single molecules and a small number of dimers and trimers will leave the surface below a threshold luence. Starting at the threshold luence

1.5 Ionization

(deined as the minimum energy per irradiated area required for the abovediscussed collective ejection process of larger molecular clusters), the intense detection of MALDI ions becomes possible [51]. It must also be observed that these simulations contain signiicant simpliications and, most probably more restrictive, must be scaled to very small volumes and short time regimes because of limited computation capacity. These models have become signiicantly reined over the past few years and will, no doubt, continue to do so. In this respect they will clearly contribute to an understanding of MALDI processes in the future.

1.5 Ionization

The mechanisms which lead to the formation of charged matrix and analyte molecules in the MALDI process are even more poorly understood than the physics of the material ablation/desorption. For a better understanding, it is important to distinguish between the ionization of matrix molecules and that of the analytes. For the standard UV-MALDI laser wavelengths of 337 or 355 nm (i.e., photon energies of 3.7 and 3.5 eV, respectively), more than two photons are needed for the photoionization of a free matrix molecule. However, with increasing matrix aggregate size the ionization potential is lowered somewhat [3, 52] and may come within reach of the sum energy of two photons for the condensed phase. In case a small energy gap has still to be overcome, thermal energy may make up for the difference. The typical MALDI laser photon luxes are too low for any signiicant nonresonant and resonant two-photon absorption to take place, but the excited energy of molecules (called excitons) is known to be highly mobile in the crystals with stacked aromatic π-electron systems, such as those of matrices [53]. This allows for the combination of the energy of two excitons with interacting electronic states by which one of the two molecules becomes de-excited to the ground state and the other molecule rises to a higher excited (Sn) state. Very eficient energy pooling between excited matrix molecules in the crystals has indeed been demonstrated at laser luences typical for MALDI, even though the majority of the absorbed energy becomes converted into lattice energy on a picosecond time scale by internal conversion [37]. Subsequent to photoionization, capture of the released photoelectron by neutral matrix molecules will give rise to complementary negative radical matrix ions beside the positive ions [54]. Due to large cross-sections of dissociated slow electrons, the escape of photoelectrons will be restricted to the very top matrix layers [38]. The frequent observation of radical matrix ions such as matrix+• (Eq. 1.3 [54]) and/or [matrix + 2H]+• and matrix−• (Eq. 1.4 [54]), as well as [matrix-2H]−• ions, besides the expected even-electron ions, among them [matrix + H]+ – or [matrixH]− – (for assumed formation reactions, see Eqs (1.5) and (1.6) [54]) as well as a prominent [matrix dimer + H-2H2O]+ for 2,5-DHB and [matrix dimer + H]+ for HCCA is at least in agreement with the photoionization model, if not a strong indication for the correctness of this model (Figure 1.4).

13

14

1 The MALDI Process and Method

Figure 1.4 UV-MALDI mass spectrum of myoglobin. Matrix: DHBs (2,5-dihydroxybenzoic acid

[2,5-DHB] plus 2-hydroxy-5-methoxysalicylic acid [MSA]; 9 : 1 [w/w]). Wavelength, 337 nm; mass analyzer, relectron TOF. ν matrix h → matrix ** → matrix + i + e −

(1.3)

matrix + e − → matrix −i

(1.4) +

i

matrix + i + matrix → [matrix + H ] + [matrix − H ]

(1.5)

matrix − i + matrix → [matrix − H ]− + [matrix + H ]i

(1.6)

Although no precise numbers have been determined experimentally, it is, most probably, safe to assume that the ion yield for the matrix (i.e., the ratio of ions to neutrals) which depends on the matrix as well as the laser luence is somewhere in the range of 10−6 to 10−4 [55, 56]. The ion yield of the analytes can be much higher, on the order of 0.1–1% for typical compounds, and above 10% in exceptional cases. The intensities of the ion signals, as determined from the spectra, are not independent of each other, because charge-transfer processes from matrix to analyte species as well as among each species are taking place in the expanding plume and possibly already in the solid state upon laser irradiation. In favorable cases, the spectra even show intense analyte ion signals with negligible matrix ions, despite a typical 104 excess of the matrix [57]. As the MALDI ionization process starts with the aforementioned matrix ionization, the typically much higher analyte ion yield compared to that of the matrix already indicates a charge transfer and accumulation process in favor of the analyte. Two models for analyte ion formation have been proposed. The older model – which had not had a well-deined name before 2013 and is now proclaimed as Coupled Physical and Chemical Dynamics (CPCD) model – assumes neutral analyte molecules in the expanding plume – regardless of whether the analytes were incorporated in the matrix crystals as neutral species or were quantitatively neutralized by their counterions upon cluster dissociation in the case of precharged incorporated analyte molecules. Subsequent to photoionization of the matrix (Eqs 1.3 and 1.4) and secondary intermolecular matrix reactions leading to the generation of protonated as well as deprotonated matrix ions (Eqs 1.5 and 1.6)

1.5 Ionization

[54], charge transfer to and from the neutral analyte molecules in the plume leads to the generation of [analyte + H]+ Eq. (1.7) and [analyte-H]− Eq. (1.8).

[matrix + H ]+ + analyte → matrix + [analyte + H ]+

(1.7)

[matrix − H ]− (carboxylate ) + analyte ↔ matrix + [analyte − H ]− (carboxylate ) (1.8)

Knochenmuss has developed a quantitative mathematical model for matrix and analyte ion generation assuming matrix photoionization as the initial step of ion generation. Initially, this was limited to a description of the desorption of individual molecules but, in a later approach, this model was extended to the ejection of a mixture between single molecules and clusters [58]. One argument against photoionization as initial starting point for ion generation is the observation that MALDI spectra, obtained with IR lasers at 1.94 μm wavelength, closely resemble the UV-MALDI spectra, including radical matrix ions observed (e.g., for the succinic acid matrix). The photon energy of only 0.6 eV at this wavelength certainly excludes photoionization. However, similarity of the spectra alone does not sufice as proof of identical ionization processes at the two wavelengths, and indeed the ion yield for analytes in the IR is about an order of magnitude lower than in the UV. The more recent “lucky survivor” model in its inal form assumes that basic analytes, such as most proteins, are incorporated into the matrix as precharged species corresponding to their solution charge state at the point in time they are cocrystallized. This assumption is based on the observation that pH-indicator molecules retain their color and charge state upon crystal incorporation for acidic, neutral, or basic matrices [31]. Furthermore, the addition of extremely strong acids to matrix–analyte preparations led to the gas-phase detection of precharged analyte–acid anion adducts originating from solution [59, 60]. For most common acidic matrices, almost all peptides will carry a (multiple) positive excess charge; counterions will then typically be either matrix or acid (if added to the preparation) anions. For preventing neutralization, the charged peptides must be incorporated in an at least partially solvated form for separation from their counterions by solvent. In a second step, the model assumes a break-up of the crystal lattice into larger and smaller clusters upon desorption, some of them with analyte ions and counterions which will neutralize each other upon cluster dissociation by loss of neutral matrix and possibly remaining solvent molecules. Karas and Krüger assumed that some of these analyte-containing clusters statistically carry one or more excess protonated or deprotonated matrix ions. Upon cluster dissociation, these excess charges are thought to prevent quantitative analyte ion neutralization, leading to the generation of protonated and deprotonated analyte ions [2]. These singly charged analyte ions are the so-called “lucky survivors” of the neutralization process. This model elegantly explains the observation of mostly singly charged ions in MALDI spectra, and works equally well for UV- as well as IR-MALDI. The formation of matrix–analyte clusters in the desorption process is in agreement with experiments in which the ablation of large(r) aggregates was detected under

15

16

1 The MALDI Process and Method

MALDI-conditions [47] and also predicted by molecular modeling [40, 50]. However, these model calculations also predicted that large clusters would have insuficient internal energy for a complete dissociation with evaporation of the neutrals. This result may also originate from insuficient simulation times and/or to too-limited sizes of the simulation box [58], and does not necessarily relect the situation for smaller to medium-sized clusters. While the initially charged matrix – analyte cluster ions never appear in the MALDI spectra with the typical solutes used, such as triluoroacetic acid (TFA) or ammonium salts, they have been detected for extremely strong acids [60]; their conjugate anions are extremely weak bases and thus stabilize the formed ion pairs with protonated analyte sites. One of the strengths of the lucky-survivor model is that it can be equally well applied to account for the formation of negative ions (if deprotonable groups are present) and positive ions from basic analytes such as peptides and proteins (in solution precharged by protonation; Eqs 1.9 and 1.10), as well as for positive (if protonable groups are present) and negative ions from acidic analytes (in solution precharged by deprotonation; Eqs 1.11 and 1.12) such as nucleic acids:

{[analyte + nH ]

n+

+ n [counterion − H ]− + [matrix + H ]+

}

+

+

→ [analyte + H ] + n counterion + matrix

(1.9)

{[analyte + nH ] + n [counterion − H ] + [matrix − H ] (carboxylate )} → {analyte + n counterion + [matrix − H ] (carboxylate )} n+













↔ [analyte − H ] (carboxylate) + n counterion + matrix

{[analyte − nH ]

n−

(1.10)

+ +

+

+ n [counterion + H ] + [matrix + H ]

}

+

→ [analyte + H ] + n counterion + matrix

{[analyte − nH ]

n−

(1.11)

}

+ n [counterion + H ]+ + [matrix − H ]− (carboxylate ) −





→ {analyte + n counterion + [matrix − H ] − (carboxylate)} − −

↔ [analytte − H ] (carboxylate) + n counterion + matrix

(1.12)

where {} is a symbol for the intermediate cluster. The commonly observed lower sensitivity of basic analytes (e.g., peptides and proteins) as anions in the negativeion mode, in contrast to the typically higher detection sensitivities of their corresponding protonated species, is relected by the CPCD model as well as the lucky survivor. Typical matrix molecules with proton afinities roughly between 800 and 900 kJ mol−1 [61], and especially the even more acidic halogenated α-cyanocinnamic acid derivatives recently introduced by Jaskolla et al. [26, 62, 63], are in their protonated states strong gas-phase acids in comparison to basic amino acids with much higher proton afinities (up to 1025 kJ mol−1 for arginine) [64]. This results in an eficient proton transfer and formation of positively charged analyte ions, as postulated by the CPCD model of Eq. (1.7). The “lucky survivor” model also postulates an eficient generation of protonated basic analytes, as randomly incorporated protonated matrix molecules (resulting from photoionization) in clusters with precharged basic analytes and corresponding counterions lead to a very exo-

1.6 Fragmentation of MALDI Ions

thermic counterion neutralization by [matrix+H]+ upon cluster dissociation, which in turn results in the survival of a precharged analyte (Eq. 1.9). In contrast, no such difference in basicity between the negatively charged ions of matrix molecules and peptides exists, because both are typically carboxylate anions with very close proton afinities. Peptide anion generation competes with simple dissociation of the neutral analyte–matrix anion adduct and limits the yield of analyte anions (see Eq. 1.8 for the CPCD model and Eq. 1.10 for the “lucky survivor”). Similar arguments hold for the polyanionic nucleic acids. While the addition of ammonium ions results in a quantitative detection of the free acids as singly charged cations or anions, substitution of the ammonium ions by tetraalkyl ammonium ions leads to the observation of adduct ions, increasing in intensity with increasing length of the alkyl chains. Both models – protonation of neutral analytes in the gas phase according to the CPCD model as well as the revised lucky survivor – are discussed in detail in Refs [3, 24, 65]. With the help of deuterated matrices Jaskolla and Karas have recently proven that both models – gas-phase (de)protonation (CPCD model) as well as “lucky survivor” – exist as parts of a greater overall mechanism [24]. These authors concluded that neutral peptides are protonated according to the CPCD model, whereas for slightly basic analytes such as nicotinamide both models contribute to charging. Basic analytes (e.g., peptides) are predominantly protonated by the “lucky survivor,” although several factors such as laser luence, preparation solution pH, analyte size and the difference between matrix and analyte proton afinity may affect or even swap the dominating charging mechanism [24]. The previously detected inluence of the matrix proton afinity or, being more precise, the achievable sensitivity increase by use of low-proton afinity matrices [26], is now in excellent agreement with the predictions made by this uniied analyte protonation mechanism. For analytes of very low proton afinity, such as neutral carbohydrates and many synthetic polymers, cationization by Na+, K+ or other metal cations is usually observed in MALDI (see Chapters 6–8). The cationization in all likelihood takes place in the expanding plume, and requires a codesorption of the analyte and the cations. Hence, the best results are obtained from sample locations where both species exist in close neighborhood, such as in the center of DHB-dried droplet preparations. Speciic protocols have been developed for the MALDI of such analytes [66, 67].

1.6 Fragmentation of MALDI Ions

The fragmentation of MALDI ions is a mixed blessing, as in all of MS. It can, on the one hand, lead to a substantial loss of spectra quality such as loss of mass resolution or even complete loss of the signal of the intact parent ion, as has been

17

18

1 The MALDI Process and Method

shown for the loss of sialic acids in the analysis of glycoconjugates in relectronTOF analyzers [68]. On the other hand, intrinsic or induced fragmentation is an indispensable tool for the acquisition of structural information in MSn experiments. The nomenclature for the fragmentation – particularly the differentiation in post-source-decay (PSD) and in-source-decay (ISD) ions – is closely related to TOF analyzers, because they still constitute the majority of spectrometers used for the analysis of MALDI ions. The details of how to analyze fragment ions with different types of instrument are discussed in Chapter 2 (see Section 2.4.3 in particular). Even though some limited ion stability – and thus fragmentation – was obvious in the early days of MALDI by peak tailing on the low mass side, and was attributed to small neutral losses of peptide and protein ions, the fact that MALDI can generate substantial prompt and metastable fragmentation of analyte ions was obscured at an early stage. This was due to two facts: (i) the laser microprobe instrument (LAMMA 1000) used for the initial experiments indeed minimized fragmentation because of a very weak acceleration ield strength and a low total (3 keV) ion energy; and (ii) the next-generation of MALDI-TOF instruments were linear instruments in which only less-intense prompt but no metastable fragment ions can be observed directly. It was only when Kaufmann and Spengler began careful investigations of ion stability using a deceleration stage in a linear TOF instrument that the potentially high degree of metastable fragmentation was detected [69]. This was the starting point for the development of the so-called PSD analysis of metastable ions which has led to today’s MALDI-TOF/TOF instruments (the irst “TOF” refers to generation and isolation of the precursor ion, and the second “TOF” to separation and identiication of the generated fragment ions) [70, 71]. Modern instruments enable faster and better-resolved PSD measurements by ion fragmentation in a collision cell, followed by an increase in the potential energy of the (precursor and fragment) ions and their subsequent separation by means of a post-acceleration cell, for example, by the “LIFT” technology [72] or the TOF/TOF-version by Vestal et al. [73]. The character of the PSD fragmentation – that is, the classes of fragment ions observed – is in full agreement with a collisional activation process. Despite this general agreement, PSD mass spectra are often more complex than collisionally induced dissociation (CID) mass spectra, showing internal fragments and products of consecutive fragmentations, and pointing to more complex excitation mechanisms of the intramolecular degrees of freedom. An overview between different MALDI fragmentation processes is available in Refs [74, 75]. Besides collisional excitation through collisions of the analyte ions with neutrals in the plume upon applying an acceleration voltage, absorption of the laser energy by the matrix with subsequent thermal energy transfer to analytes already in the solid state, as well as subsequently in the early phase of the dense plume, are possible causes for fragmentation. In addition to exothermic plume reactions such as electron attachment, the exothermicity of gas-phase proton-transfer reactions is a further widely suggested driving force for analyte fragmentation [76, 77]. The thermal energy (temperature) of analyte ions depends in complex fashion on

1.6 Fragmentation of MALDI Ions

instrumental parameters such as laser wavelength, luence and focus, as well as on the matrix-speciic absorption properties [25]. These laser parameters further inluence the plume properties with increasing axial velocities at higher laser luences and laser pulse length-dependent sizes of ablated clusters [58] which, in combination with strength and delay of the ion extraction ield, affects the collisional activation. Information about the different contributions to ion excitation has been listed in a report which compares vacuum MALDI to atmospheric-pressure (AP)-MALDI. These results were obtained by well-understood correlations between the extent of unimolecular fragmentation of substituted benzylpyridinium ions and their internal ion energy [78]. A comprehensive summary of the current knowledge of MALDI fragmentation is available in Refs [78–80]. The effect of collisional excitation upon prompt acceleration was demonstrated for matrix ions. These AP-MALDI investigations were later extended to determine initial MALDI plume excitation processes [81], including the application to more representative test samples than thermometer molecules, such as model peptides and deoxynucleosides. Schulz et al. detected a rough correlation between the extent of fragmentation and the exothermicity of gas-phase analyte protonations by protonated matrices: the lower the matrix proton afinity (PA) and higher the analyte gas-phase basicity, the more internal analyte energy was available for fragmentation [81]. Similar conclusions were drawn by Stevenson et al., when investigating analyte deprotonation reactions. With increasing gas-phase basicities of the investigated deprotonated matrix anions, more internal energy was released upon analyte deprotonation, leading to a more intense analyte fragmentation [76]. In contrast, a recent analysis concluded that the energy released upon exothermic gas-phase reactions would be insuficient for analyte fragmentation [82]. In agreement with this, 4-chloroα-cyanocinnamic acid (ClCCA) is much softer than HCCA [83], even though it has a clearly lower PA [26]. This points to the existence of additional parameters inluencing the matrix “hard-/softness” (causing intense or low analyte fragmentation, respectively). In additional experiments, it was found that the matrix “softness” roughly correlated with the initial ion velocities determined in a linear TOF. In some – but not all – cases, “hard” matrices correlated with low initial velocities [76, 81, 84]. Indeed, ClCCA was seen to exhibit an initial velocity that somewhat exceeded that of HCCA [84], although it was unclear if this increase would fully explain the detected differences in fragmentation. In addition to the assumption that the exothermicity of gas-phase reactions is the most important parameter, additional parameters such as expansion cooling in the MALDI desorption process might further inluence the matrix “hard-/softness.” In a practical approach, matrices are classiied and ranked from “hard” to “soft.” For example, HCCA – the matrix which is preferentially applied for “peptide-mass ingerprint” analyses in proteomics because of its rather homogeneous sample morphology and relatively high sensitivity – is one of the “hardest” matrices in that ranking. Because of its degree of fragmentation induction, it is also the matrix of choice for PSD- or TOF/TOF-experiments. “Super DHB” (DHBs; 2,5-DHB with a 5–10% addition of 2-hydroxy-5-methoxybenzoic acid), 6-aza-2-thiothymine

19

20

1 The MALDI Process and Method

(ATT), and especially 3-hydroxypicolinic acid (3-HPA) [81], are on the “soft” end of the list and are therefore preferentially applied for the analysis of larger proteins. By using such a “soft” matrix and optimizing all instrumental parameters (low extraction ield strength and long extraction delay times for improved resolution and lower plume densities resulting in less collisions and collisional activation upon ion acceleration), small-neutral loss can be minimized and a good mass resolution (close to the theoretical limit) is obtainable for a linear TOF coniguration, even for medium-sized proteins up to bovine serum albumin [85]. Alternative matrices for the MALDI-MS analysis of high-molecular-weight compounds comprise ionic liquids [86] (see Section 1.8.3). Common to PSD and the lowenergy CID mechanisms is the randomization of the internal energy among all internal degrees of freedom with (metastable) fragmentation preferentially of the most labile bonds [87]. A very different fragmentation mechanism was irst reported by R.S. Brown et al. [88, 89], whereby fragment ions are formed “promptly” upon ion generation/ excitation with a time delay of ≤100 ns, which is substantially less than the typical delay times in delayed-extraction TOF instruments. Therefore, these are referred to as in-source-decay (ISD) ions. The ISD spectra of proteins contain signals of c- and sometimes z-type fragment ions, in addition to some a-, b-, and y-ions [90, 91] (for fragment ion nomenclature, see Chapter 3), but is not limited to peptides and proteins and can also be used for the sequencing of, for example, oligonucleotides such as DNA or RNA [92] or carbohydrates [93, 94]. This type of fragmentation is observed for both positive and negative ions [90], indicating that ISD is independent of proton transfer. Initially, 2,5-DHB was the matrix of choice for high ISD yields [90, 95]; nevertheless, upon the introduction of a method for quantifying matrix ISD-eficiencies [96] it was mainly replaced by more reactive compounds such as 5-aminosalicylic acid (5-ASA) [97] or 1,5-diaminonaphthalene (1,5-DAN), which exhibit very eficient radical transfer abilities to analytes [96, 98]. However, the latter matrix is carcinogenic and generates strong, potentially interfering matrix cluster ion peaks up to about 1000 Da [77]. More recent developments have included nontoxic and highly eficient ISD matrices, for example, 2aminobenzamide and 2-aminobenzoic acid [91]. The similarity of ISD spectra to those of electron capture dissociation prompted an initial discussion on the possible role of electrons in ISD, but it is clear today that ISD is mediated by hydrogen radicals [95, 99], as was demonstrated by the in-source reduction of disulide bonds [100]. In the positive-ion mode this includes hydrogen abstraction from matrix molecules to peptides, followed by a unimolecular fragmentation of the peptide radical or a cleavage upon collision with a second matrix molecule [101]. Unfortunately, the nomenclature of ISD and PSD relates to the speciic source geometry of linear TOF instruments rather than to the two different fragmentation processes. This has led to some confusion as to which process is causing which fragment [80, 82]. Although ISD yields substantially more complete fragment-ion series as compared to PSD, previous intentions to use ISD for practical applications (e.g., in proteomics [102, 103]) have not yet fully materialized. This is due mostly to the

1.7 MALDI of Noncovalent Complexes

low intensity of the fragment ion signals, which only recently has been considerably improved by the introduction of the aforementioned higher reactive matrices [104, 105]. An additional problem is that true MS/MS experiments cannot be carried out in linear TOF instruments, as a precursor selection for peptides from a mixture is not possible. This leads to complex fragment spectra with overlapping ion signals deriving from multiple fragmented precursor species if no previous precursor separation (e.g., LC-MS/MS) was performed.

1.7 MALDI of Noncovalent Complexes

After about 25 years’ use of MALDI-MS, it is clear that the successful analysis of noncovalent interactions and complexes is the exception rather than the rule. Yet, when considering most typical MALDI protocols, this cannot be a surprise. Most “classical” matrices are organic acids and typically are used in water/organic solvent mixtures, often acidiied by formic acid (FA) or TFA – that is, under conditions which should result in the dissociation of most (if not all) noncovalent complexes. However, signals of noncovalent complexes have indeed been obtained from such preparations [106, 107]. At least for some such systems, the dissociation seems to be suficiently slow that a rapid evaporation of the solvent conserves at least a certain fraction of the complexes. Unfortunately, adjustment of the pH and the omission of organic solvent is often not a viable alternative, as acidic matrices will be deprotonated and totally change their crystallization and incorporation properties as salts. Many of the salts and buffers that are used to adjust ionic strength/pH are, therefore, not MALDI-compatible in the desired concentrations, or at least they compromise performance with respect to sensitivity and mass resolution. Nevertheless, the use of sinapic acid with a higher water solubility allows for the preparation of purely aqueous matrix solutions. In combination with a physiological pH, at which sinapic acid still crystallizes as a neutral species, noncovalent protein complexes were retained during sample preparation and could consistently be detected using MALDI [108]. The incorporation of analytes into the matrix crystals, which as such has been shown to make the desorption process softer or altogether possible [35], is another step that might lead to dissociation of the complexes. This question has been addressed in order to understand the so-called “irst-shot phenomenon,” which had been observed much earlier but was resolved only recently [109, 110]. For a number of selected matrices, the signals of protein–protein complexes are observed only in the spectra of irst exposures of a given sample spot. Subsequent exposures yield only signals of the monomer units. However, by employing a combination of MALDI-MS and confocal laser scanning microscopy (CLSM) of noncovalent complexes with luorochromes, which exhibit a luorescence resonance energy transfer (FRET), it could be shown that in these systems the complexes would dissociate upon incorporation, whereas intact complexes would precipitate at the surface of the matrix crystals. The next crucial step is the intact desorption and

21

22

1 The MALDI Process and Method

ionization of the complexes, and their survival in the gas phase of the expanding plume. Such dissociation upon desorption is even more likely if the complexes are localized at the crystal surface rather than being incorporated. The type of interaction within the complex is also a decisive parameter. From energetic considerations it is clear that the stability of noncovalent complexes is highest for ionic interactions, followed by ion–dipole forces and hydrogen bonding. Interestingly, the formation of strong ionic complexes has even been used to facilitate MALDI measurements of highly acidic (and thus negatively charged) biocompounds, such as DNA, oligonucleotides and heparin-derived oligosaccharides. This was achieved by admixing highly basic peptides (e.g., histones), followed by a mass determination of the intact stable complex in the positive-ion mode [111]. Hydrophobic interactions were shown to be particularly labile and to result in a predominant gas-phase dissociation of complexes bound largely by such forces [112]. This effect can be explained by a loss of the hydrophilic environment of polar solvents in the liquid state upon transfer into the vacuum. The facile detection of noncovalent complexes by ESI shows, that the transfer of noncovalent complexes into the vacuum does not necessarily result in dissociation, if the internal energy does not sufice for intermolecular bond-breaking and transition to the very different conformational monomeric states. It would also appear that large contact areas between the constituents of the complex and the corresponding contributions of salt- and hydrogen bridges, as well as hydrophobic interactions typical of many protein–protein complexes, help to stabilize the complex. Consequently, complexes including small molecules, as are typical for ligand–receptor and antigen–antibody systems or speciic channel-forming antibiotics, are much more dificult to analyze when using MALDI-MS. Their afinities depend on the exact conservation of the conformation in small epitopes of the protein, which is more easily lost in the MALDI process than is the complete quaternary structure. Interestingly, spectra of the intact biotin–streptavidin complex, which has a dissociation constant of 10−15 M and is one of the strongest complexes known to date, have never been obtained by MALDI. Another issue which complicates the use of MALDI for the analysis of noncovalent complexes is the formation of nonspeciic multimeric and adduct ions. This effect is even more pronounced for the analysis of noncovalent protein complexes, as high concentrations (10 pmol μl−1 or higher) of analyte are typically used to overcome the reduced sensitivity of TOF instruments in the high mass range. Furthermore, an elevated laser luence, as well as deviation from optimal preparation protocols, are aggravating effects. It is, therefore, very important to clearly differentiate speciic from nonspeciic interactions, and this is typically achieved by using a known nonbinding/noninteracting control compound. Because of these limitations, the analyses of noncovalent interactions by MALDI are typically qualitative rather than quantitative. Within these boundary conditions, a number of reports have described the successful detection of several types of noncovalent complexes [113–118]. During the early days of MALDI, high-intensity signals of noncovalent protein complexes

1.7 MALDI of Noncovalent Complexes

using a nicotinic acid matrix and a laser wavelength of 266 nm were reported, for example of the tetrameric glucose isomerase [119] and a trimeric porin [120]. Comprehensive reviews have been provided by Hillenkamp [121] and Farmer and Caprioli [122] on this subject. In addition, a report by Zehl and Allmaier on the inluence of instrumental parameters for the detection of quaternary protein structures starts from a careful review of the state of the art [123], and provides a critical discussion on the above-described problems. These authors used 2,6-dihydroxyacetophenone as (non-acidic) matrix with the addition of ammonium acetate or diammonium citrate (DAC) to adjust the solution conditions to stabilize the protein complexes. By employing these milder preparation conditions, it was possible to directly measure intact hemoglobin complexes from diluted crude blood samples [124]. Another only slightly acidic matrix which tolerates even highly concentrated additive amounts such as DAC is ATT; this was also used to investigate nucleic acids and their noncovalent complexes and adducts [125]. IR-MALDI enables comparably soft desorption processes and is, therefore, more appropriate for noncovalent complex analysis than UV-MALDI. This is relected by the successful detection of intact double-stranded DNA as well as enzyme– oligosaccharide complexes in combination with glycerol as matrix which were not detectable or underwent signiicant fragmentation when UV-MALDI was employed [126, 127]. In more recent approaches, ionic liquid matrices have been used for the facilitated detection of intact noncovalent complexes, most probably due to the omitted requirement of intact complex incorporation into the matrix crystals [86]. Some aggravating circumstances such as the irst-shot phenomenon can be circumvented by the accumulation of only irst shots ired at fresh sample surfaces. As a consequence of these hindered conditions, numerous approaches regarding the investigation of noncovalent complexes use stabilizing chemical crosslinking between the complex partners in combination with high-mass MALDI [122, 128, 129]. It appears, however, that the potential of MALDI in this area is far from being fully explored. Recently, a new approach was presented to investigate the formation and identiication of noncovalent complexes, based on the detection of the “intensity fading” of one complex partner upon the addition of increasing amounts of the other [130], rather than a more ambitious veriication of the intact complex. This approach avoids problems related to the detection of the intact complex in the high mass range, and can be carried out at analytically relevant micromolar and submicromolar concentration levels. Up-to-date overviews debating the challenges of mass spectrometric detection of noncovalent interactions, including new developments and chemical crosslinking, have been presented by Bich and Zenobi [131] and Mädler et al. [132]. In summary, the use of MALDI to investigate noncovalent interaction is far less straightforward than typical applications for peptides and proteins under denaturing conditions. Success is not predictable, and careful control experiments must be implemented to differentiate speciic from nonspeciic interactions.

23

24

1 The MALDI Process and Method

1.8 The Optimal Choice of Matrix: Sample Preparation

Unfortunately – but expectedly – there is no single MALDI matrix or sample preparation protocol which is suited to all analytical problems and analytes in MALDI-MS. A few of the more general considerations are discussed in the following section; more speciic information is provided in the applications chapters of this book. A representative list of commonly used matrices and their main applications is provided in Table 1.1. There are different matrices of irst choice for different classes of analytes and analytical problems. For example, HCCA is used in the majority of proteomics applications for the analysis of peptide-mass-ingerprints generated by protein enzymatic digests, as well as for MS/MS fragmentation experiments (as discussed in Chapter 3). The recently discovered ClCCA even improves this performance especially in combination with 337 nm instruments [26, 62]. On the other hand, 2,5-DHB – and especially DHBs (i.e., DHB with an admixture of 5–10% 2-hydroxy5-methoxybenzoic acid) – with its pronounced crystallization into large crystals of approximately 100 μm size is particularly suited to protein analysis. The reasons for this are: (i) that its softness prevents strong small-neutral losses and peak tailing; and (ii) that the crystals incorporate the proteins but exclude the majority of common contaminants. In addition, there is a large variety of matrix compounds speciically useful for the sensitive detection of speciic analyte classes, for example, 2-[(2E)-3-(4-tert-butylphenyl)-2-methylprop-2-enylidene]malononitrile (DCTB) for polymers and fullerenes [133], 9-aminoacridine [134] and paranitroaniline [135] for lipids, 2-(4′-hydroxybenzeneazo)benzoic acid (HABA)/1,1,3,3tetramethylguanidine (TMG) ionic liquid matrices (ILMs) for the intact analysis of heparin sulfate oligosaccharides [136], α-cyano-2,4-diluorocinnamic acid (DiFCCA) [63] and lumazine [137] for phospholipids, an ionic HCCA copper complex for copper-binding peptides [67], or proton sponge for interference-free negative-ion mode analysis of low-mass metabolites [138]. A practical overview of the various matrices and preparation techniques can be found, for example, on the Internet at http://msr.dom.wustl.edu/Research/ MALDI_TOF_Mass_Spec_and_Proteomics/MALDI_Sample_Preparation_ Methods.htm, as well as on the Internet pages of commercial suppliers of MALDI matrices, for example, http://www.sigmaaldrich.com/analytical-chromatography/ analytix-newsletter/analytix-2001.html#analytix6. The “dried droplet” standard MALDI sample preparation is very simple. Here, the sample and matrix are dissolved in solvents or solvent systems that are miscible with each other, and mixed either before deposition onto or directly on the MALDI sample support. The matrix–analyte droplet of typically 1 μl volume is then slowly dried in air, or under a forced low of cold air. This results in a deposit of crystals which, depending on the matrix and preparation conditions (e.g., solvent evaporation rate), vary between submicrometer and several hundred micrometers in size. In cases of solvent systems with a high polarity, surface tension leads to a nonhomogeneous distribution of the individual crystals near the rim of the

1.8 The Optimal Choice of Matrix: Sample Preparation Table 1.1 A selection of commonly used MALDI matrices.

Matrix

Structure

Wavelength

Major applications

6-Aza-2-thiothymine (ATT)

HS

UV 337 nm, 355 nm

Proteins, peptides, noncovalent complexes; near-neutral pH

UV 337 nm, (355 nm)

(Glyco/Phospho) Peptides, (halogenated) lipids, fragmentation

UV 337 nm, 355 nm

Peptides, fragmentation

UV 337 nm, 355 nm

Proteins, peptides, carbohydrates, synthetic polymers

UV 337 nm, 355 nm

Proteins, peptides, noncovalent complexes; near-neutral pH

IR 2.94 μm, 2.79 μm

Proteins, peptides; liquid matrix

UV 337 nm, 355 nm

Nucleic acids

UV 266 nm

Proteins, peptides, adduct formation

UV 337 nm, 355 nm

Proteins, peptides

IR 2.94 μm, 2.79 μm

Proteins, peptides

N N

OH N CH3

4-Chloro-αcyanocinnamic acid (ClCCA)

Cl COOH NC

α-Cyano-4-hydroxycinnamic acid (HCCA)

HO COOH NC

2,5-Dihydroxybenzoic acid (plus 10% 2-hydroxy-5-methoxybenzoic acid) (2,5-DHB(s)) k,m,n-Di/Trihydroxyacetophenone (D/THAP)

OH COOH HO X

X

X

COCH3 X

X

Glycerol

X = OH or H

OH HO

3-Hydroxypicolinic acid (3-HPA)

OH

OH N COOH

Nicotinic acid (NA)

COOH N

Sinapic acid (SA)

H 3CO HO COOH H 3 CO

Succinic acid

HOOC

IR, infrared; UV, ultraviolet.

COOH

25

26

1 The MALDI Process and Method

preparation; this led to the development of crystal positioning/detection systems for automated MS measurements in modern MALDI mass spectrometers. The best MALDI performance for classic crystalline matrix preparations is usually achieved only at certain locations of the crystals, which often requires manual interference and active control by the experimenter; this is why most MALDI instruments are equipped with a microscopic sample observation system. The cause of these “sweet spots” has been the subject of much speculation, the commonly held notion being that of an inhomogeneous distribution of analyte within the crystals. However, this has been disproved by Horneffer et al., who found a homogeneous distribution of luorescently labeled analytes in the crystals of a representative number of different matrix crystals by CLSM studies (see Section 1.2), as well as by Qiao et al., who detected uniformly incorporated proteins in DHB and sinapic acid by means of highly spatially resolved MALDI imaging and luorescence microscopy [139]. A different (ionization) state of the analyte molecules in different locations, local differences in matrix–analyte interactions [140], as well as different matrix crystallizations due to variations of the local analyte concentrations [141] or heterogeneous orientation of the matrix crystal surfaces relative to the spectrometer axis and perpendicular to which the ions are ejected in conjunction with the limited angular acceptance of the mass spectrometer, might also cause the observed sweet spots. As a rule of thumb, addition of the analyte solution should not noticeably change the crystallization behavior of the neat matrix; an unchanged morphology already indicates that the solution excess of the matrix with respect to the analyte is maintained in the crystals, and that the contaminant level is low. Any solute component which already at low concentrations dramatically changes the morphology of the subsequently crystallized matrix or prevents crystallization altogether – for example, low-volatility solvents such as dimethyl sulfoxide or common detergents – will prohibit a successful MALDI analysis. Over the years, a large number of modiications of, or alternatives to, the drieddroplet technique have been developed. These many variations are often the personal preferences of MALDI users for sample preparation, and the subject may appear to be an art, or even a “black art.” Nonetheless, dried-droplet sample preparations are still most widely used with high success due to their simple and rapid handling. It also appears that instrumental developments using lasers with higher frequencies (e.g., modern Nd:YAG solid-state lasers with repetition rates of up to 1 kHz), together with the automation of entire MALDI measurements, have eased the problems of heterogeneity to some extent. Additionally, modiications of the dried-droplet preparation such as, for example, a rapid evaporation of the solvent of standard dried-droplet preparations by reduced pressure [142] or the rapidevaporation crystallization by using highly volatile solvents such as acetone [143], both result in smaller and more homogeneously distributed crystals with improved spot-to-spot reproducibilities and minimized sweet spots. Among the many modiications and variations of the simple dried-droplet preparation, two alternatives stand out as particularly useful and widespread, namely “surface preparation” and “anchor sample plates.”

1.8 The Optimal Choice of Matrix: Sample Preparation

1.8.1 Surface Preparation

Surface preparations or predeposited matrix crystal layers are also often called thin-layer or two-layer preparations, and were introduced to enhance sensitivity, homogeneity of the preparation, automation, and liquid chromatography (LC)spotting. Surface preparation was a true innovation [144]. As discussed above for the rapid-evaporation crystallization, the HCCA matrix generates a relatively homogeneous microcrystalline layer upon rapid evaporation of the organic solvent. Subsequent addition of the aqueous peptide analyte solution – which must contain a low amount of organic solvent – redissolves only the very top layers of the nearwater-insoluble matrix and incorporates and concentrates the analyte upon recrystallization using the nondissolved matrix bed as seed crystals [145]. It has been shown that the addition of a low matrix amount to the analyte solution (containing a somewhat higher organic solvent amount) to be applied onto the microcrystalline matrix layer increases the achievable sensitivity for proteins. The structurally relatively homogeneous sample surface further increases the mass resolution, particularly in linear TOF mass spectrometers. Unfortunately, this approach is restricted to matrices which do not fully dissolve in the usually aqueous analyte solvent. However, despite its limited solubility, the matrix is partially dissolved by the analyte solution, and true incorporation of the analyte into the matrix takes place. It was shown, moreover, that contaminants such as salts could be rinsed from the surface with a splash of ice water, without causing any major analyte loss which results in clearly higher sensitivities. Hence, surface preparation became the starting point for the development of disposable MALDI targets with predeposited matrix spots [146]. The generation of more homogeneous microcrystalline sample layers by rapid evaporation of the solvent (e.g., in vacuo) has also been used for a variety of other matrices (Figure 1.5). 1.8.2 Anchor Sample Plates

Anchor plates for the preparation of multiple samples have small hydrophilic islands, typically of 100–500 μm diameter, placed on a hydrophobic surface [147]. The hydrophobic surface prevents spreading of the sample solution over a larger area, as otherwise observed for dried-droplet preparations. Instead, the hydrophilic solution contracts onto these islands, thereby concentrating the matrix and analyte onto a small deined area upon solvent evaporation. This coninement to a smaller volume is particularly useful for analytes of low concentration in combination with proportionally lowered matrix concentrations, and also facilitates automated analyses of the ixed-location samples. Anchor sample plates are also commercially available as disposable targets prespotted with matrix and calibration spots.

27

28

1 The MALDI Process and Method (a)

20 µm

1 µm

(b)

20 µm

1 µm

(c)

20 µm

1 µm

(d)

20 µm

1 µm

(e)

20 µm

1 µm

Figure 1.5 Matrix crystals resulting from

standard dried-droplet preparations on stainless steel plate (left) and from sublimed preparations on hydrophobically coated aluminum oxide surfaces (right). From upper

to lower: (a) HCCA; (b) sinapic acid, (c) 4-methyl-CCA; (d) 4-hydroxy-3-methyl-CCA; (e) 4-hydroxy-3-methoxy-CCA. Image reproduced from Ref. [145] with permission from Elsevier.

1.8 The Optimal Choice of Matrix: Sample Preparation

1.8.3 Matrix Additives and Inluence of the Sample Plate Surface

A few other modiications of the sample preparation have also proven useful in speciic cases: •

Mixtures of several different matrices have been reported for an improved performance, for example, 1 : 1 (w/w) mixtures between HCCA and DHB resulted in a lowered chemical noise and slightly improved sequence coverages [148]. However, so far only DHBs (a mixture of 90–95% 2,5-DHB and 5–10% 2-hydroxy-5-methoxybenzoic acid) has found relatively widespread application for proteins [149]. The additive softens the desorption and limits the small neutral loss, thereby improving mass resolution. A mixture of different trihydroxyacetophenones is sometimes used for the analysis of nucleic acids [150].



Additives to matrix preparations are mostly used for intrinsic sample clean-up. These additives do not absorb the laser radiation but may inluence the crystallization behavior of the matrix to some extent as a side effect. The most frequently used method is to add diammonium citrate (DAC) as a cation scavenger to preparations of highly anionic samples such as nucleic acids containing high amounts of alkali metal counterions [151]. The addition of ammonium phosphate to improve peptide-ingerprint mass spectra by the suppression of matrix cation clusters, and the use of up to 1% phosphoric acid to improve DHB analysis of phosphopeptides, may be due to a more eficient incorporation of positively precharged phosphopeptides with neutral phosphate groups into the growing matrix crystals, are two recent successful examples of this approach [152, 153].

Modiied surfaces of sample plates can be used for the afinity capture of analytes from crude mixtures directly on target. Such surfaces can signiicantly enhance detection sensitivity and can be used for simple sample clean-ups. Titania (TiO2)coated surfaces, particles or sol–gel systems, for example, have been shown to concentrate phosphopeptides very selectively from peptide ingerprint samples [154, 155]. Surface-enhanced laser desorption ionization (SELDI) uses so-called “protein chips” for the detection of peptides and proteins from complex biological luids such as blood or urine, often for the identiication of diagnostic biomarkers for speciic carcinomas [156, 157]. These protein chips can contain various media for positive or negative ion exchange or reverse-phase chromatography, as well as speciic antibodies or DNA. The functionalized surface is immobilized on a MALDI sample plate for the selective enrichment of constituents of the complex mixture applied, whereas the not bound supernatant is removed by washing. Unfortunately, a large number of unsubstantiated claims for the detection of disease-related biomarkers has discredited this approach, mostly as a result of poor mass spectrometric performance. Liquid matrices can avoid the undesirable heterogeneity of crystalline MALDI samples. Liquid matrices (e.g., nitrobenzylalcohol and nitrophenyloctylether)

29

30

1 The MALDI Process and Method

were introduced in the early UV-MALDI reports, but never found widespread application because their performance proved to be signiicantly inferior to that of the traditional solid matrices, particularly because of extensive adduct formation. A more recent development has been the synthesis of ILMs, which are synthesized by preparing a 1 : 1 solution of a classical organic acid matrix with an organic base, for example, pyridine or 3-aminoquinoline [158]. Such preparations allow for higher shot-to-shot- as well as spot-to-spot-reproducibility compared to crystalline matrices, although their sensitivities are typically lowered. Unfortunately, the most sensitive ILMs are either solid or very highly viscous liquids. A possible solution to this problem is the addition of a polar liquid of low volatility, such as glycerol, which would stabilize the generated ion-pairs and alleviate liquid preparations [159]. Such liquid-support matrices exhibit high analyte homogeneities with nearly constant ion signal intensities across the whole preparation [160]. To date, these types of matrices seem to be mostly restricted to the MALDIMS analysis and the quantiication of small, stable analytes. Nevertheless, recent optimization approaches using sterically hindered bases such as N,Ndiisopropylethylamine or N-isopropyl-N-methyl-N-tert-butylamine also allow for the detection of high-molecular-weight analytes such as large proteins [86]. Today, ILMs are becoming increasingly important for imaging MALDI-MS, due to their homogeneous sample surface coverage going along with the elimination of the hot-spot phenomenon [161, 162]. Further information regarding sample preparation and the effects of speciic bases, as well as additional application areas, is provided in Refs [140, 163]. A solvent-free preparation is of particular interest for the analysis of synthetic polymers for which a common solvent with a suitable matrix is not available, and for which sizeable amounts of material are usually on hand [164]. In this protocol, the mixed analyte, matrix, and often a cationizing compound (e.g., AgCl or NaCl) are ground thoroughly in a mortar or ball-mill and loaded onto a MALDI target as powders, or after having been pressed into pellets. Good mass spectra can be obtained for analytes up to a mass of ca. 30 kDa, even though the analytes are not incorporated into the matrix crystals [35]. Abbreviations

5-ASA ATT CCA CID ClCCA CLSM DAC 1,5-DAN DCTB DHB/2,5-DHB 2,6-DHB

5-Aminosalicylic acid 6-Aza-2-thiothymine α-Cyanocinnamic acid Collisional induced dissociation 4-Chloro-α-cyanocinnamic acid Confocal laser scan microscopy Diammonium citrate 1,5-Diaminonaphthalene 2-[(2E)-3-(4-tert-butylphenyl)-2-methylprop-2-enylidene] malononitrile (2,5-)Dihydroxybenzoic acid 2,6-Dihydroxybenzoic acid

References

DHBs DiFCCA DIOS ESI FA FAB FD FRET HABA HCCA 3-HPA ILM IR ISD LAMMA LC MALDI AP-MALDI IR-MALDI UV-MALDI MS MSA OPO PA PD PSD SALDI SELDI SIMS TFA TMG TOF TOF-MS UV

“super” DHB (mixture of 90–95% DHB and 5–10% 2-hydroxy5-methoxybenzoic acid) α-Cyano-2,4-diluorocinnamic acid Desorption/ionization on silicon Electrospray ionization Formic acid Fast atom bombardment Field desorption Fluorescence resonant energy transfer 2-(4-Hydroxyphenylazo)benzoic acid α-Cyano-4-hydroxycinnamic acid 3-Hydroxypicolinic acid Ionic liquid matrix Infrared In-source decay Laser microprobe mass analyzer Liquid chromatography Matrix-assisted laser desorption/ionization MALDI at atmospheric pressure MALDI with infrared laser wavelengths MALDI with ultraviolet laser wavelengths Mass spectrometry 2-Hydroxy-5-methoxysalicylic acid Optical parametric oscillator Proton afinity Plasma desorption Post-source decay Surface-assisted laser desorption/ionization Surface-enhanced laser desorption/ionization Secondary ion mass spectrometry Triluoroacetic acid 1,1,3,3-Tetramethylguanidine Time-of-light Time-of-light mass spectrometer Ultraviolet

References 1 Deisewerd, K. (2003) The desorption

4 El-Aneed, A., Cohen, A., and Banoub, J.

process in MALDI. Chem. Rev., 103, 395–425. 2 Karas, M. and Krüger, R. (2003) Ion formation in MALDI. Chem. Rev., 103, 427–439. 3 Knochenmuss, R. and Zenobi, R. (2003) MALDI ionization: the role of in-plume processes. Chem. Rev., 103, 441–452.

(2009) Mass spectrometry, review of the basics: electrospray, MALDI, and commonly used mass analyzers. Appl. Spectrosc. Rev., 44, 210–230. 5 Beckey, H.D. (1977) Principles of Field Ionization and Field Desorption Mass Spectrometry, International Series in Analytical Chemistry, Pergamon Press.

31

32

1 The MALDI Process and Method 6 Benninghoven, A. and Sichtermann, W.

7

8

9

10

11

12

13

14

15

(1978) Detection, identiication and structural investigation of biologically important compounds by secondary ion mass spectrometry. Anal. Chem., 50, 1180–1184. Zakett, D., Schoen, A.E., Cooks, R.G., and Hemberger, P.H. (1981) Laserdesorption mass spectrometry/mass spectrometry and the mechanism of desorption ionization. J. Am. Chem. Soc., 103, 1295–1297. Torgerson, D.F., Skrowonski, R.P., and Macfarlane, R.D. (1974) New approach to the mass spectroscopy of nonvolatile compounds. Biochem. Biophys. Res. Commun., 60, 616–619. Jonsson, G.P., Hedin, A.B., Håkansson, P.L., Sundqvist, B.U.R., Save, B.G., Nielsen, P.F., Roepstorff, P., Johansson, K.E., Kamensky, I., and Lindberg, M.S. (1986) Plasma desorption mass spectrometry of peptides and proteins adsorbed on nitrocellulose. Anal. Chem., 58, 1084–1087. Fenner, N.C. and Daly, N.R. (1966) Laser used for mass analysis. Rev. Sci. Instrum., 37, 1068–1072. Vastola, F.J., Mumma, R.O., and Pirone, A.J. (1970) Analysis of organic salts by laser ionization. Org. Mass Spectrom., 3, 101–104. Posthumus, N.A., Kistemaker, P.G., Meuzelaar, H.L.C., and Ten Noever de Brauw, M.C. (1978) Laser desorptionmass spectrometry of polar nonvolatile bioorganic molecules. Anal. Chem., 50, 985–991. Hillenkamp, F., Kaufmann, R., Nitsche, R., and Unsöld, E. (1975) Laser microprobe mass analysis of organic materials. Nature, 256, 119–120. Karas, M., Bachmann, D., and Hillenkamp, F. (1985) Inluence of the wavelength in high-irradiance ultraviolet laser desorption mass spectrometry of organic molecules. Anal. Chem., 57, 2935–2939. Hillenkamp, F. (1986) Laser desorption mass spectrometry. A review, in Secondary Ion Mass Spectrometry SIMS V. Springer Series in Chemical Physics, vol. 44 (eds A. Benninghoven, R.J. Colton, D.S. Simons, and H.W. Werner), SpringerVerlag, Berlin, pp. 471–475.

16 Karas, M. and Hillenkamp, F. (1988)

Laser desorption ionization of proteins with molecular masses exceeding 10 000 daltons. Anal. Chem., 60, 2299–2301. 17 Tanaka, K., Waki, H., Ido, Y., Akita, S., Yoshida, Y., and Yoshida, T. (1988) Protein and polymer analyses up to m/z 100 000 by laser ionization time-of-light mass spectrometry. Rapid Commun. Mass Spectrom., 2, 151–153. 18 Schürenberg, M., Dreisewerd, K., and Hillenkamp, F. (1999) Laser desorption/ ionization mass spectrometry of peptides and proteins with particle suspension matrixes. Anal. Chem., 71, 221–229. 19 Sunner, J., Dratz, E., and Chen, Y.-C. (1995) Graphite surface-assisted laser desorption/ionization time-of-light mass spectrometry of peptides and proteins from liquid solutions. Anal. Chem., 67, 4335–4342. 20 Wei, J., Buriak, J.M., and Siuzdak, G. (1999) Desorption–ionization mass spectrometry on porous silicon. Nature, 399, 243–246. 21 Trauger, S.A., Go, E.P., Shen, Z., Apon, J.V., Compton, B.J., Bouvier, E.S.P., Finn, M.G., and Siuzdak, G. (2004) High sensitivity and analyte capture with desorption/ ionization mass spectrometry on silylated porous silicon. Anal. Chem., 76, 4484–4489. 22 Trimpin, S., Inutan, E.D., Herath, T.N., and McEwen, C.N. (2010) Laserspray ionization, a new atmospheric pressure MALDI method for producing highly charged gas-phase ions of peptides and proteins directly from solid solutions. Mol. Cell. Proteomics, 9, 362–367. 23 Trimpin, S., Wang, B., Inutan, E.D., Li, J., Lietz, C.B., Harron, A., Pagnotti, V.S., Sardelis, D., and McEwen, C.N. (2012) A mechanism for ionization of nonvolatile compounds in mass spectrometry: considerations from MALDI and inlet ionization. J. Am. Soc. Mass Spectrom., 23, 1644–1660. 24 Jaskolla, T.W. and Karas, M. (2011) Compelling evidence for lucky survivor and gas phase protonation: the uniied MALDI analyte protonation mechanism. J. Am. Soc. Mass Spectrom., 22, 976–988. 25 Soltwisch, J., Jaskolla, T.W., Hillenkamp, F., Karas, M., and Dreisewerd, K. (2012)

References

26

27

28

29

30

31

32

33

34

Ion yields in UV-MALDI mass spectrometry as a function of excitation laser wavelength and optical and physico-chemical properties of classical and novel halogen-substituted MALDI matrices. Anal. Chem., 84, 6567–6576. Jaskolla, T.W., Lehmann, W.-D., and Karas, M. (2008) 4-Chloro-αcyanocinnamic acid is an advanced, rationally designed MALDI matrix. Proc. Natl. Acad. Sci. USA, 105, 12200–12205. Strupat, K., Karas, M., and Hillenkamp, F. (1991) 2,5-Dihydroxybenzoic acid: a new matrix for laser desorptionionization mass spectrometry. Int. J. Mass Spectrom. Ion Processes, 111, 89–102. Beavis, R.C. and Bridson, J.N. (1993) Epitaxial protein inclusion in sinapic acid crystals. Appl. Phys., 26, 442–447. Horneffer, V. (2002) Matrix-AnalytWechselwirkungen bei der MatrixUnterstützten Laserdesorptions/ ionisations Massenspektrometrie (MALDI-MS). Dissertation, University of Münster. Kampmeier, J., Dreisewerd, K., Schürenberg, M., and Hillenkamp, F. (1997) Investigations of 2,5-DHB and succinic acid as matrices for IR and UV MALDI: part 1: UV and IR laser ablation in the MALDI process. Int. J. Mass Spectrom. Ion Processes, 169/170, 31–41. Krueger, R., Pfenninger, A., Fournier, I., Glückmann, M., and Karas, M. (2001) Analyte incorporation and ionization in matrix-assisted laser desorption/ ionization visualized by pH indicator molecular probes. Anal. Chem., 73, 5812–5821. Horneffer, V., Forsmann, A., Strupat, K., Hillenkamp, F., and Kubitschek, U. (2001) Localization of analyte molecules in MALDI preparations by confocal laser scanning microscopy. Anal. Chem., 73, 1016–1022. Horneffer, V., Reichelt, R., and Strupat, K. (2003) Protein incorporation into MALDI-matrix crystals investigated by high resolution ield emission scanning electron microscopy. Int. J. Mass Spectrom., 226, 117–131. Skelton, R., Dubois, F., and Zenobi, R. (2000) A MALDI sample preparation

35

36

37

38

39

40

41

42

43

method suitable for insoluble polymers. Anal. Chem., 72, 1707. Horneffer, V., Glückmann, M., Krüger, R., Karas, M., Strupat, K., and Hillenkamp, F. (2006) Matrix-analyteinteraction in MALDI-MS: pellet and nano-electrospray preparations. Int. J. Mass Spectrom., 249/250, 426–432. Dreisewerd, K., Schürenberg, M., Karas, M., and Hillenkamp, F. (1995) Inluence of the laser intensity and spot size on the desorption of molecules and ions in matrix-assisted laser desorption/ ionization with a uniform beam proile. Int. J. Mass Spectrom. Ion Processes, 141, 127–148. Lüdemann, H.-C., Redmond, R.W., and Hillenkamp, F. (2002) Singlet-singlet annihilation in UV-MALDI studied by luorescence spectroscopy. Rapid Commun. Mass Spectrom., 16, 1287–1294. Knochenmuss, R. (2004) Photoionization pathways and free electrons in UV-MALDI. Anal. Chem., 76, 3179–3184. Dreisewerd, K. (2003) The desorption process in MALDI. Chem. Rev., 103, 395–425. Zhigilei, L.V. and Garrison, B. (2000) Microscopic mechanisms of laser ablation of organic solids in the thermal and stress coninement irradiation regimes. J. Appl. Phys., 88, 1281–1298. Rohling, A., Menzel, C., Kukreja, L.M., Hillenkamp, F., and Dreisewerd, K. (2003) Photoacoustic analysis of matrix-assisted laser desorption/ ionization processes with pulsed infrared lasers. J. Phys. Chem. B, 107, 12275–12286. Leisner, A., Rohling, A., Röhling, U., Dreisewerd, K., and Hillenkamp, F. (2005) Time-resolved imaging of the plume dynamics in infrared matrixassisted laser desorption/ionization with a glycerol matrix. J. Phys. Chem. B, 109, 11661–11666. Menzel, C., Dreisewerd, K., Berkenkamp, S., and Hillenkamp, F. (2002) The role of the laser pulse duration in infrared matrix-assisted laser desorption/ionization mass spectrometry. J. Am. Soc. Mass Spectrom., 13, 975–984.

33

34

1 The MALDI Process and Method 44 Rohling, A. (2006) Untersuchungen

45

46

47

48

49

50

51

52

53

54

zum Desorptionsprozess in der UV-Matrix-unterstützten Laserdesorption/-ionisation mit einer Flüssigmatrix. Dissertation, University of Münster. Vogel, A. and Venugopalan, V. (2003) Mechanisms of pulsed laser ablation of biological tissues. Chem. Rev., 103, 577–644. Kelly, R. and Miotello, A. (1999) Contribution of vaporization and boiling to thermal-spike sputtering by ions or laser pulses. Phys. Rev. E, 60, 2616–2625. Alves, S., Kalberer, M., and Zenobi, R. (2003) Direct detection of particles formed by laser ablation of matrices during matrix-assisted laser desorption/ ionization. Rapid Commun. Mass Spectrom., 17, 2034–2038. Hoyer, T., Tuszynski, W., and Lienau, C. (2007) Ultrafast photodimerization dynamics in a-cyano-4-hydroxycinnamic acid and sinapinic acid crystals. Chem. Phys. Lett., 443, 107–112. Sachon, E., Clodic, G., Blasco, T., and Bolbach, G. (2007) Protein desolvation in UV matrix-assisted laser desorption/ ionization (MALDI). J. Am. Soc. Mass Spectrom., 18, 1880–1890. Zhigilei, L.V. and Garrison, B.J. (1999) Mechanisms of laser ablation from molecular dynamics simulation: dependence on the initial temperature and pulse duration. Appl. Phys. A, 69, 75–80. Zhigilei, L.V., Kodali, P.B.S., and Garrison, B.J. (1998) A microscopic view of laser ablation. J. Phys. Chem. B, 102, 2845–2853. Lin, Q. and Knochenmuss, R. (2001) Two-photon ionization thresholds of matrix-assisted laser desorption/ ionization matrix clusters. Rapid Commun. Mass Spectrom., 15, 1422–1426. Setz, P.D. and Knochenmuss, R. (2005) Exciton mobility and trapping in a MALDI matrix. J. Phys. Chem. A, 109, 4030–4037. Ehring, H., Karas, M., and Hillenkamp, F. (1992) Role of photoionization and photochemistry in ionization processes of organic molecules and relevance for matrix-assisted laser desorption

55

56

57

58

59

60

61

62

63

ionization mass spectrometry. Org. Mass Spectrom., 27, 472–480. Mowry, C.D. and Johnston, M.V. (1993) Simultaneous detection of ions and neutrals produced by matrix-assisted laser desorption. Rapid Commun. Mass Spectrom., 7, 569–575. Quist, A.P., Huth-Fehre, T., and Sundqvist, B.U.R. (1994) Total yield measurements in matrix-assisted laser desorption using a quartz crystal microbalance. Rapid Commun. Mass Spectrom., 8, 149–154. Knochenmuss, R., Dubois, F., Dale, M.J., and Zenobi, R. (1996) The matrix suppression effect and ionization mechanisms in matrix-assisted laser desorption/ionization. Rapid Commun. Mass Spectrom., 10, 871–877. Knochenmuss, R. and Zhigilei, L.V. (2010) Molecular dynamics simulations of MALDI: laser luence and pulse width dependence of plume characteristics and consequences for matrix and analyte ionization. J. Mass Spectrom., 45, 333–346. Friess, S.D. and Zenobi, R. (2001) Protein structure information from mass spectrometry? selective titration of arginine residues by sulfonates. J. Am. Soc. Mass Spectrom., 12, 810–818. Krüger, R. and Karas, M. (2002) Formation and fate of ion pairs during MALDI analysis: anion adduct generation as an indicative tool to determine ionization processes. J. Am. Soc. Mass Spectrom., 13, 1218–1226. Jørgensen, T.J.D., Bojesen, G., and Rahbek-Nielsen, H. (1998) The proton afinities of seven matrix-assisted laser desorption/ionization matrices correlated with the formation of multiply charged ions. Eur. J. Mass Spectrom., 4, 39–45. Jaskolla, T.W., Papasotiriou, D.G., and Karas, M. (2009) Comparison between the matrices α-cyano-4-hydroxycinnamic acid and 4-chloro-α-cyanocinnamic acid for trypsin, chymotrypsin, and pepsin digestions by MALDI-TOF mass spectrometry. J. Proteome Res., 8, 3588–3597. Teuber, K., Schiller, J., Fuchs, B., Karas, M., and Jaskolla, T.W. (2010) Signiicant

References

64

65

66

67

68

69

70

71

72

sensitivity improvements by matrix optimization: a MALDI-TOF mass spectrometric study of lipids from hen egg yolk. Chem. Phys. Lipids, 163, 552–560. Harrison, A.G. (1997) The gas-phase basicities and proton afinities of amino acids and peptides. Mass Spectrom. Rev., 17, 201–217. Knochenmuss, R. (2006) Ion formation mechanisms in UV-MALDI. Analyst, 131, 966–986. Pfenninger, A., Karas, M., Finke, B., Stahl, B., and Sawatzki, G. (1999) Matrix optimization for matrix-assisted laser desorption/ionization mass spectrometry of oligosaccharides from human milk. J. Mass Spectrom., 34, 98–104. Wu, Z., Fernandez-Lima, F.A., Perez, L.M., and Russell, D.H. (2009) A new copper containing MALDI matrix that yields high abundances of [peptide + Cu]+ ions. J. Am. Soc. Mass Spectrom., 20, 1263–1271. Karas, M., Bahr, U., Hillenkamp, F., Tsarbopoulos, A., and Pramanik, B.N. (1995) Matrix dependence of metastable fragmentation of glycoproteins in MALDI TOF mass spectrometry. Anal. Chem., 67, 675–679. Spengler, B., Kirsch, D., and Kaufmann, R. (1992) Fundamental aspects of post-source decay in MALDI mass spectrometry. J. Phys. Chem., 96, 9678–9684. Spengler, B. (1997) Post-source decay analysis in matrix-assisted laser desorption/ionization mass spectrometry of biomolecules. J. Mass Spectrom., 32, 1019–1036. Chaurand, P., Luetzenkirchen, F., and Spengler, B. (1999) Peptide and protein identiication by matrix-assisted laser desorption ionization (MALDI) and MALDI-post-source decay time-of-light mass spectrometry. J. Am. Soc. Mass Spectrom., 10, 91–103. Suckau, D., Resemann, A., Schuerenberg, M., Hufnagel, P., Franzen, J., and Holle, A. (2003) A novel MALDI LIFT-TOF/TOF mass spectrometer for proteomics. Anal. Bioanal. Chem., 376, 952–965.

73 Vestal, M. and Hayden, K. (2012) A new

74

75

76

77

78

79

80

81

82

approach to TOF mass spectrometry for high performance MS and MS-MS. Proceedings, 60th Annual ASMS Conference on Mass Spectrometry and Allied Topics, Vancouver, BC, Canada, May 20–24, 2012. Harvey, D.J. (1999) Matrix-assisted laser desorption/ionization mass spectrometry of carbohydrates. Mass Spectrom. Rev., 18, 349–451. Zhang, L. and Reilly, J.P. (2009) Extracting both peptide sequence and glycan structural information by 157 nm photodissociation of N-linked glycopeptides. J. Proteome Res., 8, 734–742. Stevenson, E., Breuker, K., and Zenobi, R. (2000) Internal energies of analyte ions generated from different matrixassisted laser desorption/ionization matrices. J. Mass Spectrom., 35, 1035–1041. Demeure, K., Gabelica, V., and De Pauw, E.A. (2010) New advances in the understanding of the in-source decay fragmentation of peptides in MALDITOF-MS. J. Am. Soc. Mass Spectrom., 21, 1906–1917. Gabelica, V., Schulz, E., and Karas, M. (2004) Internal energy build-up in matrix-assisted laser desorption/ ionization. J. Mass Spectrom., 39, 579–593. Luo, G., Marginean, I., Ye, L., and Vertes, A. (2008) Competing ion decomposition channels in matrixassisted laser desorption ionization. J. Phys. Chem. B, 11, 6952–6956. Yoon, S.H., Moon, J.H., and Kim, M.S. (2010) A comparative study of in- and post-source decays of peptide and preformed ions in matrix-assisted laser desorption ionization time-of-light mass spectrometry: effective temperature and matrix effect. J. Am. Soc. Mass Spectrom., 21, 1876–1883. Schulz, E., Karas, M., Rosu, F., and Gabelica, V. (2006) Inluence of the matrix on analyte fragmentation in atmospheric pressure MALDI. J. Am. Soc. Mass Spectrom., 17, 1005–1013. Bae, Y.J., Moon, J.H., and Kim, M.S. (2011) Expansion cooling in the matrix

35

36

1 The MALDI Process and Method plume is under-recognized in MALDI mass spectrometry. J. Am. Soc. Mass Spectrom., 22, 1070–1078. 83 Leszyk, J.D. (2010) Evaluation of the new MALDI matrix 4-chloro-αcyanocinnamic acid. J. Biomol. Tech., 21, 81–91. 84 Glückmann, M. and Karas, M. (1999) The initial ion velocity and its dependence on matrix, analyte and preparation method in ultraviolet matrix-assisted laser desorption/ionization. J. Mass Spectrom., 34, 467–477. 85 Bahr, U., Stahl-Zeng, J., Gleitsmann, E., and Karas, M. (1997) Delayed extraction time-of-light MALDI mass spectrometry of proteins above 25 000 Da. J. Mass Spectrom., 32, 1111–1116. 86 Crank, J.A. and Armstrong, D.W. (2009) Towards a second generation of ionic liquid matrices (ILMs) for MALDI-MS of peptides, proteins, and carbohydrates. J. Am. Soc. Mass Spectrom., 20, 1790–1800. 87 Kaufmann, R., Kirsch, D., and Spengler, B. (1994) Sequencing of peptides in a time-of-light mass spectrometer: evaluation of postsource decay following matrix-assisted laser desorption ionization (MALDI). Int. J. Mass Spectrom. Ion Processes, 131, 355–385. 88 Brown, R.S., Carr, B.L., and Lennon, J.L. (1996) Factors that inluence the observed fast fragmentation of peptides in matrix-assisted laser desorption. J. Am. Soc. Mass Spectrom., 7, 225–232. 89 Reiber, D.C., Brown, R.S., Weinberger, S., Kenny, J., and Bailey, J. (1998) Unknown peptide sequencing using matrix-assisted laser desorption/ ionization and in-source decay. Anal. Chem., 70, 1214–1222. 90 Brown, R.S., Feng, J., and Reiber, D.C. (1997) Further studies of in-source fragmentation of peptides in MALDI. Int. J. Mass Spectrom. Ion Processes, 169/170, 1–18. 91 Smargiasso, N., Quinton, L., and De Pauw, E. (2012) 2-Aminobenzamide and 2-aminobenzoic acid as new MALDI matrices inducing radical mediated in-source decay of peptides and proteins. J. Am. Soc. Mass Spectrom., 23, 469–474. 92 Hagan, N.A., Smith, C.A., Antoine, M.D., Lin, J.S., Feldman, A.B., and

93

94

95

96

97

98

99

100

Demirev, P.A. (2012) Enhanced in-source fragmentation in MALDITOF-MS of oligonucleotides using 1,5-diaminonaphthalene. J. Am. Soc. Mass Spectrom., 23, 773–777. Wang, B.H., Hopkins, C.E., Belenky, A.B., and Cohen, A.S. (1997) Sequencing of modiied oligonucleotides using in-source fragmentation and delayed pulsed ion extraction matrixassisted laser desorption ionization time-of-light mass spectrometry. Int. J. Mass Spectrom., 169, 331–350. Yamagaki, T., Suzuki, H., and Tachibana, K. (2005) In-source and post-source decay in negative-ion matrix-assisted laser desorption/ ionization time-of-light mass spectrometry of neutral oligosaccharides. Anal. Chem., 77, 1701–1707. Takayama, M.N. (2001) N-Cα bond cleavage of the peptide backbone via hydrogen abstraction. J. Am. Soc. Mass Spectrom., 12, 1044–1049. Demeure, K., Quinton, L., Gabelica, V., and De Pauw, E. (2007) Rational selection of the optimum MALDI matrix for top-down proteomics by in-source decay. Anal. Chem., 79, 8678–8685. Sakakura, M. and Takayama, M. (2010) In-source decay and fragmentation characteristics of peptides using 5-aminosalicylic acid as a matrix in matrix-assisted laser desorption/ ionization mass spectrometry. J. Am. Soc. Mass Spectrom., 21, 979–988. Fukuyama, Y., Iwamoto, S., and Tanaka, K. (2006) Rapid sequencing and disulide mapping of peptides containing disulide bonds by using 1,5-diaminonaphthalene as a reductive matrix. J. Mass Spectrom., 41, 191–201. Köcher, T., Engström, A., and Zubarev, R.A. (2005) Fragmentation of peptides in MALDI in-source decay mediated by hydrogen radicals. Anal. Chem., 77, 172–177. Schnaible, V., Weing, S., Resemann, A., Suckau, D., Bücker, A., Wolf-Kümmeth, S., and Hoffmann, D. (2002) Screening for disulide bonds in proteins by MALDI in-source decay and LIFT-TOF/ TOF-MS. Anal. Chem., 74, 4980–4988.

References 101 Asakawa, D. and Takayama, M. (2012)

102

103

104

105

106

107

108

109

Fragmentation processes of hydrogendeicient peptide radicals in matrixassisted laser desorption/ionization in-source decay mass spectrometry. J. Phys. Chem. B, 116, 4016–4023. Reiber, D.C., Grover, T.A., and Brown, R.S. (1998) Identifying proteins using matrix-assisted laser desorption/ ionization in-source fragmentation data combined with database searching. Anal. Chem., 70, 673–683. Takayama, M. and Tsugita, A. (2000) Sequence information of peptides and proteins with in-source decay in matrix assisted laser desorption/ionization-time of lightmass spectrometry. Electrophoresis, 21, 1670–1677. Calligaris, D., Villard, C., Terras, L., Braguer, D., Verdier-Pinard, P., and Laitte, D. (2010) MALDI in-source decay of high mass protein isoforms: applications to α- and β-tubulin variants. Anal. Chem., 82, 6176–6184. Zimmerman, T.A., Debois, D., Mazzucchelli, G., Bertrand, V., De Pauw-Gillet, M.-C., and De Pauw, E. (2011) An analytical pipeline for MALDI in-source decay mass spectrometry imaging. Anal. Chem., 83, 6090–6097. Rosinke, B., Strupat, K., Hillenkamp, F., Rosenbusch, J., Dencher, N., Krüger, U., and Galla, H.-J. (1995) Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) of membrane proteins and non-covalent complexes. J. Mass Spectrom., 30, 1462–14. Cohen, L.H.R., Strupat, K., and Hillenkamp, F. (1997) Analysis of quaternary protein ensembles by matrix-assisted laser desorption/ ionization mass spectrometry. J. Am. Soc. Mass Spectrom., 8, 1046–1052. Song, F. (2007) A study of noncovalent protein complexes by matrix-assisted laser desorption/ionization. J. Am. Soc. Mass Spectrom., 18, 1286–1290. Horneffer, V., Strupat, K., and Hillenkamp, F. (2006) Localization of noncovalent complexes in MALDIpreparations by CLSM. J. Am. Soc. Mass Spectrom., 17, 1599–1604.

110 Wortmann, A., Pimenova, T., Alves, S.,

111

112

113

114

115

116

117

118

and Zenobi, R. (2007) Investigation of the irst shot phenomenon in MALDI mass spectrometry of protein complexes. Analyst, 132, 199–207. Terrier, P., Tortajada, J., Zin, G., and Buchmann, W. (2007) Noncovalent complexes between DNA and basic polypeptides or polyamines by MALDITOF. J. Am. Soc. Mass Spectrom., 18, 1977–1989. Bich, C., Baer, S., Jecklin, M.C., and Zenobi, R. (2010) Probing the hydrophobic effect of noncovalent complexes by mass spectrometry. J. Am. Soc. Mass Spectrom., 21, 286–289. Gruic-Sovulj, I., Lüdemann, H.C., Hillenkamp, F., Weygand-Duraseciv, I., Kucan, Z., and Peter-Katalinić, J. (1997) Detection of noncovalent tRNA aminoacyl-tRNA synthetase complexes by matrix-assisted laser desorption/ ionization mass spectrometry. J. Biol. Chem., 272, 32084–32091. Vogl, T., Roth, J., Hillenkamp, F., and Strupat, K. (1999) Calcium-induced non-covalently linked tetramers of MRP8 and MRP14 detected by ultraviolet matrix-assisted laser desorption/ionization mass spectrometry. J. Am. Soc. Mass Spectrom., 10, 1124–1130. Kiselar, J.G. and Downard, K.M. (2000) Preservation and detection of speciic antibody-peptide complexes by matrix-assisted laser desorption ionization mass spectrometry. J. Am. Soc. Mass Spectrom., 11, 746–750. Woods, A.S. and Huestis, M.A. (2001) A study of peptide-peptide interaction by matrix-assisted laser desorption/ ionization. J. Am. Soc. Mass Spectrom., 12, 88–96. Woods, A.S., Koomen, J.M., Ruotolo, B.T., Gillig, K.J., Russel, D.H., Fuhrer, K., Gonin, M., Egan, T.F., and Schultz, J.A. (2002) A study of peptide-peptide interactions using MALDI ion mobility o-TOF and ESI mass spectrometry. J. Am. Soc. Mass Spectrom., 13, 166–169. Luo, S.H., Li, Y.M., Qiang, W., Zhao, Y.F., Abe, H., Nemoto, T., Qin, X.R., and Nakanishi, H. (2004) Detection of

37

38

1 The MALDI Process and Method

119

120

121

122

123

124

125

126

127

speciic noncovalent interaction of peptide with DNA by MALDI-TOF. J. Am. Soc. Mass Spectrom., 15, 28–31. Karas, M., Bahr, U., Ingendoh, A., and Hillenkamp, F. (1989) Laser desorption/ ionization mass spectrometry of proteins of mass 100 000 to 250 000 Dalton. Angew. Chem. Int. Ed. Engl., 28, 760–761. Hillenkamp, F., Karas, M., Beavis, R.C., and Chait, B.T. (1991) Matrix-assisted laser desorption ionization mass spectrometry of biopolymers. Anal. Chem., 63, 1193A–1203A. Hillenkamp, F. (1998) New Methods for the Study of Biomolecular Complexes; NATO ASI Series C: Mathematical and Physical Science 510, Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 181–191. Farmer, T.B. and Caprioli, R.M. (1998) Determination of protein-protein interactions by matrix-assisted laser desorption/ionization mass spectrometry. J. Mass Spectrom., 33, 697–704. Zehl, M. and Allmaier, G. (2005) Instrumental parameters in MALDITOF mass spectrometric analysis of quaternary protein structures. Anal. Chem., 77, 103–110. Zehl, M. and Allmaier, G. (2004) Ultraviolet matrix-assisted laser desorption/ionization time-of-light mass spectrometry of intact hemoglobin complex from whole human blood. Rapid Commun. Mass Spectrom., 18, 1932–1938. Lecchi, P. and Pannelle, L.K. (1995) The detection of intact double-stranded DNA by MALDI. J. Am. Soc. Mass Spectrom., 6, 972–975. Kirpekar, F., Berkenkamp, S., and Hillenkamp, F. (1999) Detection of double-stranded DNA by IR- and UV-MALDI mass spectrometry. Anal. Chem., 71, 2334–2339. Dybvik, A.I., Norberg, A.L., Schute, V., Soltwisch, J., Peter-Katalinić, J., Vårum, K.M., Eijsink, V.G.H., Dreisewerd, K., Mormann, M., and Sørlie, M. (2011) Analysis of noncovalent chitinase-chitooligosaccharide complexes by infraredmatrix assisted laser desorption

128

129

130

131

132

133

134

135

ionization and nanoelectrospray ionization mass spectrometry. Anal. Chem., 83, 4030–4036. Bennett, K.L., Kussmann, M., Björk, P., Godzwon, M., Mikkelsen, M., Sørensen, P., and Roepstorff, P. (2000) Chemical cross-linking with thiol-cleavable reagents combined with differential mass spectrometric peptide mapping – a novel approach to assess intermolecular protein contacts. Protein Sci., 9, 1503–1518. Gasilova, N. and Nazabal, A. (2012) Monitoring ligand modulation of protein-protein interactions by chemical cross-linking and high-mass MALDI mass spectrometry. Methods Mol. Biol., 803, 219–229. Yanes, O., Villanueva, J., Querol, E., and Aviles, F.X. (2005) Functional screening of serine protease inhibitors in the medical leech Hirudo medicinalis monitored by intensity fading MALDITOF MS. Mol. Cell. Proteomics, 4, 1602–1613. Bich, C. and Zenobi, R. (2009) Mass spectrometry of large complexes. Curr. Opin. Struct. Biol., 19, 632–639. Mädler, S., Erba, E.B., and Zenobi, R. (2013) MALDI-TOF mass spectrometry for studying noncovalent complexes of biomolecules. Top. Curr. Chem., 331, 1–36. Ulmer, L., Mattay, J., Torres-Garcia, H.G., and Luftmann, H. (2000) Letter: the use of 2-[(2E)-3-(4-tert-butylphenyl)-2methylprop-2-enylidene]malononitrile as a matrix for matrix-assisted laser desorption/ionization mass spectrometry. Eur. J. Mass Spectrom., 6, 49–52. Sun, G., Yang, K., Zhao, Z., Guan, S., Han, X., and Gross, R.W. (2008) Matrix-assisted laser desorption/ ionization time-of-light mass spectrometric analysis of cellular glycerophospholipids enabled by multiplexed solvent dependent analyte matrix interactions. Anal. Chem., 80, 7576–7585. Estrada, R. and Yappert, M.C. (2004) Alternative approaches for the detection of various phospholipid classes by matrix-assisted laser desorption/

References

136

137

138

139

140

141

142

143

144

145

ionization time-of-light mass spectrometry. J. Mass Spectrom., 39, 412–422. Przybylski, C., Gonnet, F., Bonnaffé, D., Hersant, Y., Lortat-Jacob, H., and Daniel, R. (2010) HABA-based ionic liquid matrices for UV-MALDI-MS analysis of heparin and heparan sulfate oligosaccharides. Glycobiology, 20, 224–234. Calvano, C.D., Carulli, S., and Palmisano, F. (2010) 1H-Pteridine-2,4dione (lumazine): a new MALDI matrix for complex (phospho)lipid mixtures analysis. Anal. Bioanal. Chem., 398, 499–507. Shroff, R. and Svatoš, A. (2009) Proton sponge: a novel and versatile MALDI matrix for the analysis of metabolites using mass spectrometry. Anal. Chem., 81, 7954–7959. Qiao, H., Piyadasa, G., Spicer, V., and Ens, W. (2009) Analyte distribution in MALDI samples using MALDI imaging mass spectrometry. Int. J. Mass Spectrom., 281, 41–51. Tholey, A. and Heinze, E. (2006) Ionic (liquid) matrices for matrix-assisted laser desorption/ionization mass spectrometry – applications and perspectives. Anal. Bioanal. Chem., 386, 24–37. Szăjli, E., Fehér, T., and Medzihradsky, K.F. (2008) Investigating the quantitative nature of MALDI-TOF MS. Mol. Cell. Proteomics, 7, 2410–2418. Weinberger, S.R., Boernsen, K.O., Finchy, J.W., Robertson, V., and Musselman, B.D. (1993) Proceedings, 41st Annual ASMS Conference on Mass Spectrometry and Allied Topics; San Francisco, CA, May 31–June 5, 1993, pp. 775a–775b. Vorm, O., Roepstorff, P., and Mann, M. (1994) Improved resolution and very high sensitivity in MALDI TOF of matrix surfaces made by fast evaporation. Anal. Chem., 66, 3281–3287. Dai, Y., Whittal, R.M., and Li, L. (1999) Two-layer sample preparation: a method for MALDI-MS analysis of complex peptide and protein mixtures. Anal. Chem., 71, 1087–1091. Jaskolla, T.W., Karas, M., Roth, U., Steinert, K., Menzel, C., and Reihs, K.

146

147

148

149

150

151

152

153

(2009) Comparison between vacuum sublimed matrices and conventional dried droplet preparation in MALDITOF mass spectrometry. J. Am. Soc. Mass Spectrom., 20, 1104–1114. Röhl, H., Goethel, S., and Reihs, K. (2005) MPep MALDI chips for high-sensitivity and high-throughput peptide analysis by MALDI-TOF MS. Nat. Methods, 2, 6. Application Note. Schuerenberg, M., Luebbert, C., Eickhoff, H., Kalkum, M., Lehrach, H., and Nordhoff, E. (2000) Prestructured MALDI-MS sample supports. Anal. Chem., 72, 3436–3442. Laugesen, S. and Roepstorff, P. (2003) Combination of two matrices results in improved performance of MALDI MS for peptide mass mapping and protein analysis. J. Am. Soc. Mass Spectrom., 14, 992–1002. Karas, M., Ehring, H., Nordhoff, E., Stahl, B., Strupat, K., Hillenkamp, F., Grehl, M., and Krebs, B. (1993) Matrix-assisted laser desorption/ ionization mass spectrometry with additives to 2,5-dihydroxybenzoic acid. Org. Mass Spectrom., 28, 1476–1481. Zhu, Y.F., Chung, C.N., Taranenko, N.I., Allman, S.L., Martin, S.A., Haff, L., and Chen, C.H. (1996) The study of 2,3,4-trihydroxyacetophenone and 2,4,6-trihydroxyacetophenone as matrices for DNA detection in matrixassisted laser desorption/ionization time-of-light mass spectrometry. Rapid Commun. Mass Spectrom., 10, 383–388. Pieles, U., Zürcher, W., Schär, M., and Moser, H.E. (1993) Matrix-assisted laser desorption ionization time-of-light mass spectrometry: a powerful tool for the mass and sequence analysis of natural and modiied oligonucleotides. Nucleic Acids Res., 21, 3191–3196. Smirnov, I., Zhu, X., Taylor, T., Huang, Y., Ross, P., Papayanopoulos, I.A., Martin, S.A., and Pappin, D.J. (2004) Suppression of alpha-cyano-4hydroxycinnamic acid matrix clusters and reduction of chemical noise in MALDI-TOF mass spectrometry. Anal. Chem., 76, 2958–2965. Kjellström, S. and Jensen, O.N. (2004) Phosphoric acid as a matrix additive for

39

40

1 The MALDI Process and Method

154

155

156

157

158

MALDI MS analysis of phosphopeptides and phosphoproteins. Anal. Chem., 76, 5109–5117. Larsen, M.R., Thingholm, T.E., Jensen, O.N., Roepstorff, P., and Jorgensen, J.D. (2005) Highly selective enrichment of phosphorylated peptides from peptide mixtures using titanium dioxide microcolumns. Mol. Cell. Proteomics, 4, 873–886. Chen, C.T. and Chen, Y.-C. (2005) Fe3O4/TiO2 core/shell nanoparticles as afinity probes for the analysis of phosphopeptides using TiO2 surfaceassisted laser desorption/ionization mass spectrometry. Anal. Chem., 77, 5912–5919. Li, P., Zhang, D., and Guo, C. (2012) Serum biomarker screening for the diagnosis of early gastric cancer using SELDI-TOF-MS. Mol. Med. Report., 5, 1531–1535. Zhang, K., Yuan, K., Wu, H., Li, Q., Wang, Y., Chen, S., Zhang, L., Gu, H., and Fu, R. (2012) Identiication of potential markers related to neoadjuvant chemotherapy sensitivity of breast cancer by SELDI-TOF MS. Appl. Biochem. Biotechnol., 166, 753–763. Armstrong, D.W., Zhang, L.K., He, L., and Gross, M.L. (2001) Ionic liquids as matrixes for matrix-assisted laser desorption/ionization mass spectrometry. Anal. Chem., 73, 3679–3686.

159 Sze, E.T.P., Chan, T.-W.D., and Wang,

160

161

162

163

164

G. (1998) Formulation of matrix solutions for use in matrix-assisted laser desorption/ionization of biomolecules. J. Am. Soc. Mass Spectrom., 9, 166–174. Towers, M.W., Mckendrick, J.E., and Cramer, R. (2010) Introduction of 4-chloro-α-cyanocinnamic acid liquid matrices for high sensitivity UV-MALDI MS. J. Proteome Res., 9, 1931–1940. Meriaux, C., Franck, J., Wisztorski, M., Salzet, M., and Fournier, I. (2010) Liquid ionic matrixes for MALDI mass spectrometric imaging of lipids. J. Proteomics, 73, 1204–1218. Chan, K., Lanthier, P., Liu, X., Sandhu, J.K., Stanimirovic, D., and Li, J. (2009) MALDI mass spectrometry imaging of gangliosides in mouse brain using ionic liquid matrix. Anal. Chim. Acta, 639, 57–61. Soukop-Hein, R.J., Warnke, M.M., and Armstrong, D.W. (2009) Ionic liquids in analytical chemistry. Annu. Rev. Anal. Chem., 2, 145–168. Trimpin, S., Keune, S., Rader, H.J., and Mullen, K. (2006) Solvent-free MALDI-MS: developmental improvements in the reliability and the potential of MALDI in the analysis of synthetic polymers and giant organic molecules. J. Am. Soc. Mass Spectrom., 17, 661–671.

41

2 MALDI Mass Spectrometry Instrumentation Peter B. O’Connor, Klaus Dreisewerd, Kerstin Strupat, and Franz Hillenkamp

2.1 Introduction

Matrix-assisted laser desorption/ionization (MALDI) ions can be mass-analyzed using a number of different types of mass spectrometers, ranging from the inexpensive quadrupole ion trap mass spectrometer (QIT-MS), which generates fairly low-quality spectra, to the Fourier transform mass spectrometer (FT-MS), which generates extremely high-quality spectra, but at signiicantly greater cost. However, by far the most common type of analyzer for MALDI ions is the time-of-light mass spectrometer (TOF-MS). The choice of mass analyzer is governed by the speciic properties of the MALDI ion source on the one hand, and the analytical task of the operator on the other hand. The typical MALDI source generates a pulse of ions of at most a few nanoseconds in duration at a rate of 1 to 100 pulses per second; thus, they are most compatible with mass spectrometers which are effectively pulsed-ion detectors, including instruments which trap the ions for later analysis. Mass spectrometers which operate on a continuous beam of ions, such as magnetic sector and quadrupole instruments, tend to suffer substantial problems in terms of sensitivity and are generally not suitable for a pulsed-ion source. For this reason, while several experimental instruments have been designed and constructed, there is very little utility in a MALDI-magnetic sector or MALDI quadrupole mass spectrometer, and these instrument geometries have not generally been followed up with commercial development. More recently, frequency-tripled Nd:YAG lasers with pulse repetition rates of 1 kHz or more have found increasing application in MALDI ion sources and, together with suitable trapping devices, these can be operated in a quasi-continuous mode. Nonetheless, because MALDI usually generates a pulse of ions several nanoseconds wide, it is particularly suited to TOF analyzers, making TOF the dominant mass analysis technology for MALDI ions, though several variants exist. These and other instruments used for mass analysis of MALDI ions will be discussed in this chapter. A second important feature of MALDI is that it can generate ions with very large masses of greater than 1 MDa. In contrast to electrospray, MALDI ions carry only MALDI MS: A Practical Guide to Instrumentation, Methods, and Applications, Second Edition. Edited by Franz Hillenkamp and Jasna Peter-Katalinic. © 2014 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2014 by Wiley-VCH Verlag GmbH & Co. KGaA.

42

2 MALDI Mass Spectrometry Instrumentation

a few charges, with the singly charged ions generally predominant. Again, the TOF is by far the best-suited analyzer for such large ions because it has, in principle, an unlimited mass range. However, the high initial axial velocity of MALDI ions and their broad kinetic energy distribution causes problems for all mass analyzers, which in turn limits the mass resolution of TOF instruments and makes eficient capture in storage devices such as ion traps or Fourier transform-ion cyclotron resonance (FTICR) instruments dificult. The desorption/ionization also transfers considerable energy into the internal vibrational energy of ions, which is only partly removed by the adiabatic expansion of the plume. This additional vibrational energy activates the ions and results in metastable ions that then undergo unimolecular decay over a period of microseconds to seconds, depending on the relative stability of the cleaved bond (i.e., the rate of that particular fragmentation reaction and on the amount of excess internal energy deposited in the molecule on desorption). Whilst this metastable decay can be useful for obtaining structural information on the ion, such as the sequence of peptides, it can also be problematic if the part of the molecule to be analyzed is that part which fragments (and is lost) on desorption. The latter situation occurs frequently with labile species such as sulfated peptides, phosphopeptides, and sialylated glycolipids. The metastable fragmentation of ions can also prevent the detection of parent ions altogether in mass analyzers, which require milliseconds to seconds for ion detection. This problem can be partially solved by increasing the pressure in the ion source during desorption, as discussed in Section 2.3.1. In some newer MALDI mass spectrometers, this concept has been incorporated into the design of the MALDI sources, and this approach has been shown to be very successful in stabilizing labile molecules during desorption and ionization. The aim of this chapter is to provide an overview of the instrumentation currently available for the mass analysis of MALDI ions. It is not intended as an exhaustive treatise on each instrument, and consequently if more detail is required the reader should examine the listed reference material in detail. In the irst sections of the chapter, the use of lasers for MALDI is discussed, and how they are coupled to the mass spectrometer, together with details of vibrational cooling and tandem mass spectrometry of MALDI ions. This is followed by a description of the different mass analyzer designs, with emphasis placed on commercially available instrument conigurations. Deinitions of all acronyms, as well as technical descriptions of terms (such as peak centroid and resolving power) are included at the end of the chapter.

2.2 Lasers for MALDI-MS

In the past, a variety of gas and solid-state lasers have been used successfully for MALDI-MS. With MALDI, only pulsed lasers are useful, because the energy necessary for desorption and ionization must be transferred to the sample in a time which is short compared to the thermal diffusion time. Axial time-of-light (TOF)

2.2 Lasers for MALDI-MS

Figure 2.1 Wavelength and photon energy of frequently used desorption lasers.

analyzers also require a short laser pulse for a good time and mass resolution. The lasers are characterized by their emission (wavelength, pulse width, and pulse energy) and beam (beam diameter and divergence) parameters. The wavelength, λ, and single photon energy, hc/λ, of the most commonly used lasers are shown in Figure 2.1. For ultraviolet-MALDI (UV-MALDI), nitrogen (N2) gas lasers are by far the most commonly used because of their simplicity, small size, and relatively low cost. They emit at a wavelength of 337 nm, close to the absorption maximum of many commonly used matrices such as 2,5-DHB, HCCA, and 3-HPA (for a discussion of the different matrices, see Chapter 1). The main limitation of N2-lasers is their limited pulse repetition frequency of typically 100 eV, high-energy CAD will generate additional types of fragment that can carry speciic structural information such as the side-chain structure in peptides (d-, w-, and z-ions). In the past, UV photodissociation has been used with MALDI ions by several research groups, and has generated some rather unusual fragmentation patterns which are currently being investigated. Single-photon UV photodissociation has the advantage over CAD in that the timing of the fragmentation event can be controlled to the picosecond. If fragmentation is prompt, then the fragment ion resolution in a TOF instrument will not suffer; however, if fragmentation is not prompt, then metastable lifetimes can be accurately studied because the activation time is known exactly. Unfortunately, the energy needed for fragmentation is extremely high, and the most commonly used laser for this type of work is the luorine excimer laser which emits at 157 nm (∼7.9 eV per photon). Even then, a suitable chromophore must be present, which is generally the case for peptides, but may not be the case for other molecules such as oligosaccharides. More generally, IR photodissociation is used, commonly called infrared multiphoton photodissociation (IRMPD). This mode normally uses CO2 lasers which emit near 10.6 μm (∼0.1 eV per photon), and overlaps with many stretching and bending modes in organic molecules (particularly as CO2 lasers emit in a fairly wide wavelength range from 9.1 to 10.6 μm). This is effectively a slow-heating method which requires the absorption of hundreds or thousands of photons prior to cleavage. Thus, it is not suitable for TOF instruments where the ion residence time inside the laser beam is very short. It is, however, a very good fragmentation method for QIT-MS and FTICR-MS (and also hybrids; see Section 2.7), where the ions can be easily irradiated for milliseconds. In IRMPD, fragment ion spectra similar to those of metastable decay are generated, but with a high coverage because the primary fragment ions (daughter ions) remain conined within the laser beam and continue to absorb energy, thus generating secondary (grand-daughter) and higher-order (great-grand-daughter, etc.) fragments. While this is a very simple method and increases the number of fragments and overall coverage of fragmentation within a molecule, it also can result in extensive secondary fragmentation which yields many internal fragments that are often dificult to interpret. Thus, the laser power and duration of the laser pulse must be carefully adjusted to achieve the extent of fragmentation necessary to solve the analytical problem at hand. Usually, ECD and ETD are not applicable to MALDI ions as they involve the capture of a low-energy electron by a positively charged precursor ion; this reduces the charge by one. As MALDI ions are usually singly charged, ECD would result in primarily undetectable neutral species. However, there is usually a low abundance of multiply charged ions in MALDI which increase in abundance with the mass of the precursor. Thus, MALDI-ECD or ETD has been demonstrated [28], but is usually very dificult. However, while ECD usually uses low-energy electrons in the 0–1 eV range, if higher energy electrons in the 8–12 eV range are used, then fragmentation can be induced on +1 and −1 charge state ions, respectively called

51

52

2 MALDI Mass Spectrometry Instrumentation

electron-induced dissociation (EID) [28] and electron detachment dissociation (EDD) [29]. In this energy range, electrons appear to scatter from peptide ions, causing fragmentation without partially neutralizing them. Although neither method is currently used to any great degree, as they are relatively dificult to adjust, they may become useful in the future.

2.4 Mass Analyzers

Mass spectra and mass analyzers are typically characterized by a number of parameters, the most important of which are mass accuracy, mass resolution, sensitivity/ limit of detection, and signal-to-noise ratio (SNR). Mass accuracy is usually the most important parameter as it deines the most important data within an experiment. Mass accuracy is usually presented as parts per million (ppm); for example, 20 ppm means that the real mass of an ion measured to 1000 Da has a predicted error margin of ±20 mDa. While it is not usually speciied, the best way to deine this error is at ±2 standard deviations in the errors on the calibration, which means that a measured mass is ∼95% likely to be within that range. It is not critically important how this expected error is deined, provided that it is clearly stated exactly how it is calculated. Mass accuracy is limited by the quality of the calibration (how accurate it was to start with, how much it drifts with time, how much space charge has shifted the current spectrum relative to the instrument calibration, etc.), and by how accurately the center of the peak can be determined (which is determined by the number of points on a peak and how accurately the peak follows a theoretical peak shape). The mass resolution of a given ion signal is deined as the mass spacing at which spectral features (peaks) can be clearly separated, and is reported in daltons. More often, the term “resolution” actually refers to “resolving power,” which is measured as m/δm, with m the mass of the peak center and δm the resolution, often deined as the width of a peak at half height or full-width at half-maximum (FWHM). The parameter of resolving power has no units. A poor mass resolution will degrade mass accuracy because the exact center (the centroid) of broad peaks cannot be found with great accuracy. Even more importantly, a poor mass resolution may fail to separate overlapping peaks or isotopic distributions, which will distort peak shapes and thus distort mass accuracy [30]. Mass resolution is determined by the instrument used and its particular quirks, and will be discussed below. Sensitivity is deined as the slope of the intensity/moles of sample plot. It is often incorrectly used to refer to the limits of detection – which is the smallest amount of sample that can be used to achieve a detectable signal (often deined as a peak with SNR >3). Sensitivity and limits of detection are universally determined by the sample, not by the instrument. Every commercially available instrument can easily achieve low-femtomole or better limits of detection for clean, synthetic samples of peptides. However, samples from biological sources are

2.4 Mass Analyzers

extremely dificult to clean, and consequently all other biological molecules and contaminants may well be above the concentration of the analyte. If the sample contains 1 fmol of peptide, but 0.1 pmol of surfactant or salt, then the mass spectrometer (no matter which type) will detect surfactant and salts, even if many research groups have reported data claiming that their mass spectrometer can detect attomoles of that particular peptide! Finally, the SNR is just that – the signal intensity divided by the nearby average noise value. The only trick here is to deine exactly how the average noise value is calculated. Although there are many methods available, the most widely accepted calculation is to use the root-mean-square of the noise, neglecting noise spikes – provided that they do not interfere with the real sample peaks. Normally, it is very easy to identify noise spikes because almost all real ions have an isotopic peak distribution. 2.4.1 Axial TOF Mass Spectrometers

By far, the most common type of mass analyzer for MALDI applications is the axial TOF spectrometer, the principles of which are outlined in Figure 2.4. Axial TOF instruments are ideally suited to MALDI sources because they need only very short pulses of ions which then ly down the light tube and hit a detector, such as a microchannel plate (MCP). The ion current signal is then simply plotted versus arrival time at the detector, using a digital storage oscilloscope. The oscilloscope must be triggered carefully so that its scan is started when the ions are generated; this is usually achieved by splitting a tiny fraction of the laser light off onto a photo-diode, which generates the trigger. For signal averaging and calibration, it is critical that this trigger signal has a time jitter which is very small (usually 20 000 and a mass accuracy of 5–10 ppm, although selected and very highly tuned experiments have achieved 2–5 ppm accuracy. One important caveat for achieving these results is that the peaks must be isotopically resolved and a high SNR present. For example, with large proteins the mass resolution of relectron TOF mass spectrometers is limited by small neutral losses to about 120 at mass 150 kDa of a monoclonal antibody for UV-MALDI. Infrared MALDI generates cooler ions, which increases the mass resolution in this case to 230 [45]. 2.4.3 Tandem TOF Mass Spectrometers

Tandem mass spectrometry is a critical methodology used for the structural analysis of all types of molecule. The procedure involves selecting a precursor ion, fragment-

2.4 Mass Analyzers

ing it, and generating a mass spectrum of the resultant fragment ions. Structural information on the ions can be obtained in axial TOF analyzers using either metastable decay (in which ion selection is avoided, but this can complicate the analysis) or externally induced fragmentation (see Figure 2.3). The classic example is that a peptide of interest (from a biological source or from an enzymatic digest of a protein) can be selected and fragmented; the fragments will then reveal the sequence of the peptide and often will allow determination of sites of deamidation, phosphorylation, or sequence variation from mutation of the peptide’s parent DNA. For the metastable decay fragments, groups conducting research into MALDITOF have made a rather arbitrary distinction between in-source decay (ISD) [24, 25] and post-source decay (PSD) [46] ions. The ISD ions are generated by the desorption/ionization event on a time scale which is short compared to the transit time through the acceleration region or, in the DE mode, on a time scale which is short compared to the delay time of the ion extraction. In both cases, the fragment ions receive the full acceleration energy and will be detected in the spectrum at their correct mass. Strictly speaking, ions which fragment in the source during the rest of the acceleration time period would also generate “in-source” fragments, but they will experience a very wide range of different acceleration energies depending on where in this region the precursor ion decays. They will, therefore, appear in the spectrum distributed over a large mass range and will mostly be lost in the noise (though they are occasionally observed as weird baseline drifts and humps). ISD fragments are often those that result from a chemical reaction involving hydrogen radicals [23, 47] generating c- and z- type fragment ions. While these fragments are in principle very useful (similar to ECD, they can contain nearcomplete sequence information of not too-large proteins) [48], their intensity is usually too low for practical purposes. The metastable decay of ions due to an excess internal energy can usually be described by irst-order kinetics, with time constants of typically micro- to milliseconds. Early fragments during the ion acceleration are mostly lost in the noise, for the reason explained above. Precursor ions, decaying in the ield-free region of the light tube are called PSD fragment ions. Delayed extraction, in general, reduces the abundance of these PSD fragment ions because the high- pressure plume disperses somewhat before application of the ion extraction potential, thus reducing the collisional activation of the ions when the electric ield drags them through the dense cloud of neutrals. Conversely, the abundance of fragment ions can be increased by increasing the pressure in the source when the acceleration voltage is applied. PSD fragment ions will have the same velocity as the precursor. In a linear instrument, both the precursor and the fragments will strike the detector at the same time and thus go undetected. Whereas, the PSD precursor and its fragments have equal velocities, the total kinetic energy of the precursor ion will, in good approximation, be split between the fragments in the ratio of their masses. In the relectron (an energy-dispersive element, as explained above) the fragment ions will be dispersed and poorly focused onto the second detector. In general, if they are detected at all, they generate very poorly resolved peaks that are barely above

57

58

2 MALDI Mass Spectrometry Instrumentation

the baseline noise of the instrument. However, PSD ions can be refocused onto the detector by reducing the relectron voltage [46, 49]. For each reduced relectron voltage a small range of fragment masses is obtained that are well resolved and which, at the reduced relector potential, retrace the same path as the precursor ion at full voltage. The valid m/z range over which these fragments will be focused on the detector is determined by the size of the detector (lower m/z ions will shift to the “inward” side of the detector) and by the ion-focusing optics (if any are used). Thus, by sequentially stepping down the relectron voltages, thereby acquiring spectra at a series of Vr values, and stitching together the valid m/z ranges of the resulting spectra, a complete mass scan of the metastable fragment ions can be generated. This experiment is known as a “PSD scan.” Calibration of the mass scale of PSD spectra is much more complex than for simple TOF spectra, and requires special attention; consequently, for this the reader is referred either to the literature [37–40] or to the user’s manual of their instrument. The dispersion of fragment light times in the simple constant electric ield relector can be partially compensated over a large mass range if the electric relector ield is chosen as nonlinear (“curved”) in a speciic manner [43, 44]. This greatly simpliies the PSD scans, albeit at some expense of the overall mass resolution. An alternative solution to this problem is the “lift” TOF mass spectrometer [50]; in this instrumental coniguration the fragment ions are “lifted up” in their potential whilst inside a special section of the light tube. This adds kinetic energy so that the range of fragment energies relative to their total energy is reduced, thereby improving the fragment mass resolution over a mass range as large as 800 to 3000. Mass spectra generated from metastable fragments would become very crowded and dificult to interpret if the fragments of all ions generated in the source were to be displayed in a single spectrum. It is, therefore, necessary to select a single mass – or at most a mass window of a few ions – for this analysis. Generally, a pair of electrodes is used to delect unwanted ions, but a set of thin wires biased alternatively positive and negative can also achieve delection [51]. There is usually a trade-off between the width of the selected mass window and the transmission of the gate. For example, narrower windows, set to transmit only a monoisotopic mass signal, will have a compromised transmission of often below 50%. Often, the metastable ion fragmentation will not sufice to extract all of the structural information of interest, and consequently the internal vibrational energy of the molecules must be increased to generate more fragments. The most common method to accomplish this is by collisions with neutral gas molecules, typically in a specially designed collision cell. A typical tandem TOF instrument (often referred to as a TOF/TOF instrument) is shown in Figure 2.6 [52–55]. The irst linear drift tube separates the ions into packets of different m/z. An ion gate in front of the collision cell is then switched open at a correctly chosen delay time for a short period, such that only the precursor ions of interest are passed. The kinetic energy of the ions entering the collision cell is controlled by adjusting the offset potential. For example, if the source generates ions at 10 keV per charge, then the collision cell could be held at 9500 V to allow ions to collide at 500 eV per

2.4 Mass Analyzers

Figure 2.6 MALDI tandem time-of-light mass spectrometry.

charge. This collision energy can be varied, but useful values are in the range of 25 to a few hundred eV per charge [56]. These higher voltages are in the so-called “high-energy CAD” range where carbon–carbon bonds can be cleaved, allowing the structural analysis of branched hydrocarbons and isomeric species such as leucine/ isoleucine. The collision cell with its fragment ions then acts as the ion source for the second high-resolution relectron TOF mass spectrometer. Once the ion is dissociated, the fragment ions are accelerated to a high velocity once again, and then allowed to time-separate in the second high-resolution relectron TOF. The mass calibration of the MS/MS spectrum is more complicated than a normal MS spectrum because the fragment ions have a residual velocity from the precursor; thus, an additional constant velocity offset is added to the calibration equation to correct for this. In general, the fragment ion spectra are lower in resolving power (∼[1–2] × 103) and mass accuracy (∼50 ppm) than the precursor ions. The “lift” TOF mass spectrometers [50, 57] can be used for “top-down” mass spectrometry of whole proteins. Although this has been achieved in practice, it either yields huge numbers of very small peptide fragments (e.g., immonium ions) or it sequences the termini of the protein nicely, but does not yield the large complementary ions needed for good top-down analyses. However, this “top- down” mass spectrometry parameter space has not yet been fully explored for TOF/TOF instruments. TOF/TOF instruments also have their limitations: •

In order to obtain suficient initial signal for TOF/TOF experiments, the laser luence for the MALDI source is usually increased by a factor of two. This generates many more ions, but at the expense of a wide kinetic energy distribution in the plume and, consequently, the resolution suffers, both in isolation and in detection.

59

60

2 MALDI Mass Spectrometry Instrumentation



Ions in the collision cell experience a statistically distributed number of collisions, and thus acquire a range of velocities on exiting the chamber. This results in wider than normal kinetic energy distributions in the second TOF, and thus a greatly reduced resolution of the fragment ion signals.



Ion positions (both parent and fragment) are generally scattered along the ion path prior to extraction, which causes them to have a distribution of ion kinetic energies in the second stage. For this reason, the resolving power for the fragment ions is often an order of magnitude or more lower than the instrument’s “best case” precursor ion scan. It is also frequently dificult to distinguish fragments formed in the collision cell from metastable fragments that are formed between the source and the timed-ion-selector.

2.4.4 Orthogonal TOF Mass Analyzers

A more recently developed instrumental system for MALDI mass spectrometry is the orthogonal time-of-light (oTOF) mass analyzer [3, 4, 58, 59] (Figure 2.7). This instrument generates MALDI ions in a plume as usual, but instead of extracting them along the axis of the plume, a voltage pulse is used to delect them sideways so that the ion beam path is delected perpendicular to the original direction of motion. This eliminates the high initial axial velocity distribution of the plume from the arrival times of the ions, thus simplifying the calibration equation and

Figure 2.7 Orthogonal MALDI relectron time-of-light mass spectrometry.

2.4 Mass Analyzers

improving mass accuracy. In current, commercial oTOF instruments, the ions are well collimated by the ion-transfer optics and have a minimal radial velocity, so that they can usually achieve a higher resolving power and mass accuracy compared to axial TOF instruments. Orthogonal TOF instruments can routinely achieve (1–2) × 104 resolving power and 104 resolving power (particularly TOFs using time-to-digital converters for detection), these instruments must operate at threshold luence where only a few ions are generated in each spectrum, and then hundreds to thousands of spectra are usually generated and signal-averaged. FTICR-MS instruments normally use several laser shots to generate a good cloud of ions, but then use only a single excite/detect pulse. So, the speed advantage of the TOF instruments is somewhat lessened for higher-resolution or SNR level experiments. MALDI-FTICR-MS instruments are severely limited in the high m/z regime, and generally cannot detect above ∼5–10 kDa e−1 without major losses in resolution [84, 85]. However, this does mean that MALDI-FTICR-MS instruments are ideal for generating very high-accuracy MS and MS/MS data on smaller molecules – particularly in peptide mass ingerprinting and proteomics experiments. 2.5.1 Tandem Mass Spectrometry on FTICR Mass Spectrometers

FTICR mass spectrometers have the highest lexibility of all mass spectrometers in terms of methods of fragmentation. In quadrupole-FTICR-MS instruments, the ions can be isolated, accumulated, and accelerated into collision cells for collision-

2.5 Fourier Transform Ion Cyclotron Resonance Mass Spectrometers

ally activated fragmentation. For ion-trap-FTICR-MS hybrid instruments, ions can be resonantly fragmented in the ion trap itself. Moreover, once the ions are in the ICR trap, they can be manipulated extensively. FTICR-MS instruments operate on the principle of ion cyclotron resonance. As ions have resonant frequencies, these frequencies can be used to isolate the ions prior to further fragmentation or manipulation. For example, a resonant frequency pulse on the excite plates (E+/− in Figure 2.8b) will eject the ions at, or near, that frequency. Furthermore, frequency sweeps – carefully deined to not excite the ion of interest – can be used to eject unwanted ions. However, the most elegant method for ion isolation is that of Stored Waveform Inverse Fourier Transform (SWIFT) [86] in which an ion-excitation pattern of interest is chosen, inverse Fourier-transformed, and the resulting time domain signal stored in memory. This stored signal is then clocked-out, ampliied, and sent to the excite plates when needed. The typical isolation waveform in SWIFT uses a simple excitation box with a notch at the frequencies of the ion of interest, ± a few kHz. As with all other mass spectrometers, CAD is the predominant method of generating fragments for MS/MS experiments. The easiest way to perform CAD is to resonantly excite an ion of interest and then simultaneously to trigger a pulse of gas into the cell. This works well, but has the disadvantage that fragment ions are generated off-axis (i.e., with a large magnetron moment), which means that they are often dificult to detect or that they rapidly fall out of the cell. To counter this problem, a number of different collisional activation methods were created including Multiple Excitation Collisional Activation (MECA), which involves using many, low-level collisional activation events [87], Very Low-Energy CAD (VLE- CAD), which involves using a resonant excite, a 180% phase shift of the excitation pulse, and resonant de-excitation [88], and Sustained Off-Resonant CAD (SORI- CAD), which excites the ions a few kHz off-resonant so that they experience a “beat” pattern which alternately excites and de-excites them [89, 90]. These methods use low-level excitation and phase shifting to ensure that the ions of interest are always near the axis of the cell, such that the fragments are also generated near the center of the cell. Of these methods, SORI-CAD is generally the most widely used because of its simplicity and high eficiency. These methods result in many low-energy collisions with the background gas; this causes the molecule slowly to heat up, resulting in cleavage of the most labile bonds irst. Moreover, because the procedure is a resonant activation method, once the parent ion dissociates, the daughter ions are not resonant and cannot be further activated, and hence primary fragmentation predominates. For peptides, this results in generally incomplete sequence coverage, with a few labile peptide cleavages dominating the spectra. For MALDI ions, photodissociation is the next most common method for generating fragments. Here, a chromophore is required to absorb the laser light, because of which infrared multiphoton dissociation (IRMPD) [91] using a CO2 laser is preferred because such lasers emit in a wavelength where bending and rotational modes of biomolecules absorb. This is particularly true because the CO2 emission wavelengths are fairly broad, typically from 9.1–10.6 μm. Once the ions

71

72

2 MALDI Mass Spectrometry Instrumentation

are in the cell and isolated, IRMPD simply requires the laser to be triggered for a few hundreds of milliseconds; the ions then absorb hundreds to thousands of photons and dissociate through the lowest energy channels. For peptides, typical values are ∼200–500 ms irradiation time at 10 W laser power, but for DNA or oligosaccharides the typical values are 25–50 ms irradiation time at 10 W. UV photodissociation has been attempted on FTICR instruments with limited success, generating predominantly immonium ions for peptides [92], but the full characterization of this method has not yet been attempted. The new fragmentation techniques of ECD/EDD [29, 93, 94] and electron transfer dissociation (ETD) [95, 96] have not been applied widely to MALDI because they require multiply charged ions, which are generally weak with MALDI. However, the methods are occasionally applicable, and a newer method termed EID [28] has been developed which may prove to be very useful indeed (see also Section 2.3.2).

2.6 Quadrupole Ion Trap Mass Spectrometers

MALDI ion sources have also been used with quadrupole ion trap (QIT) mass spectrometers, which are also referred to as Paul-traps as they were invented by Wolfgang Paul who won the 1989 Nobel prize in physics for these studies. Figure 2.12a illustrates the approximate geometry of a classic, hyperbolic QIT mass spectrometer. These ion traps are similar to the RF-ion guides used above, in that the radial ring (r-direction in Figure 2.12a) and the end-caps (± z-direction electrodes) are driven 180° out of phase with an RF electric ield (typically ∼1 MHz and a few hundred volts peak-to-peak). As the ield oscillates, ions that are in the center of the ield (at the saddle point in Figure 2.12b) undergo characteristic motion which

(a)

(b)

Figure 2.12 The shape of Paul ion trap mass spectrometers.

2.6 Quadrupole Ion Trap Mass Spectrometers

is deined by the RF amplitude and frequency and by the mass of the ions. Some ions will be stably conined (provided that they have a low enough initial kinetic energy), those ions that are too heavy will slowly wander out of the trap and hit the electrodes, while ions that are too light will be quickly ejected. Normally, QIT mass spectrometers are operated with a few 10–3 mbar of helium as a buffer gas to collisionally cool the ions’ translational kinetic energy. Newer linear ion traps (LITs) operate on the same principles, but only in the X–Y plane. The Z-direction is bounded by placing end-plate electrodes at a high potential, similar to the approach used to trap ions in FTICR instruments. With QITs, a relatively simple way of analyzing the mass range available is through the use of Mathieu stability diagrams. These stability diagrams are derived from the differential equations of motion of ions in oscillating electric ields: 8eU dc d 2u + (aU − 2qu cos 2ξ ) u = 0 a z = −2au = 2 dξ mr02Ω2

qz = −2qu =

4eVrf mr02Ω2

where u represents the coordinate axes (x or y), ξ = Ωt/2, m = mass, e = electric charge, Udc is the DC potential difference between the ring and the endcaps, Vrf is the RF amplitude (zero-to-peak voltage), Ω is the applied RF frequency, and r0 is the radius of the trap. Note that the az values and qz values are inversely related to the mass/charge ratio of an ion. The Mathieu stability diagram plots a-values versus q-values and marks the boundaries where ions will be conined to the trap. Figure 2.13 is a plot of the Mathieu stability diagram with four points A–D marked. The trajectory of ions at those four points is shown in plots A–D on the right. Point A is within the r and z stability regions of the diagram, as shown in plot A, while point B is stable in the z-direction but unstable in the r-direction and falls out of the trap radially. Point C is stable in the r-direction but unstable in the z-direction and falls out qz = 0.908

B

A

az 0.2 Z stable B

A

0.0 r and z stable

D –0.2

az = 0.02, qz = 0.7

az = 0.05, qz = 0.1

C C

D

r stable

–0.4

–0.6 0.0

0.5

1.0

qz

az = –0.2, qz = 0.2

az = –0.04, qz = 0.2

Figure 2.13 Mathieu stability diagram with four stability points marked. Typical corresponding

ion trajectories are shown at the right. Please observe that diagrams A–C are rotated 90 degrees relative to Figure 2.12.

73

74

2 MALDI Mass Spectrometry Instrumentation

axially. Point D is just barely unstable, and this ion persists for a long time before eventually falling out of the trap. These four trajectories are representative, but the true trajectories are highly dependent on the ions’ initial velocities, positions, and initial momentum vectors. It must be remembered that each of the points in Figure 2.13 is deined by the az and qz equations above and, thus, variance in their values can represent variance in m/z, Vrf, or Udc (potentially also Ω or r0, but these are usually ixed). Therefore, by slowly ramping up Vrf or Udc (or both), ions can be shifted from stable to unstable trajectories and ejected from the trap. If done correctly, these ions can be made to impact on a conversion dynode (which ejects an electron when hit with a positive ion with suficient kinetic energy) and a secondary electron multiplier detector. When the electron current signal on the detector is plotted versus the az or qz values calculated from the Vrf or Udc ramp, a mass spectrum is produced. Normally, the conversion dynode/electron multiplier detector is placed on the z-axis, so that the qz = 0.908 value as marked in Figure 2.13 is the transition point where an ion becomes unstable in the z-direction (at az = 0, or Udc = 0). As always in physics, the Mathieu equations and stability diagrams make certain assumptions – that is, they neglect certain effects that are deemed to be negligible or uncommon. In particular, they neglect the effect of background gas pressure and the effect of the initial trajectory of the ions, assuming that the ions start in the center of the trap at rest. They also neglect ion–ion coulombic, “space- charge” interactions. All ion modes of motion (x, y, and z) in a quadrupole ion trap are stable, which means that the ion motion in that dimension is bounded and, in the absence of other stimuli, will eventually allow the ions to cool to a resting position in the center of the cell. The addition of a background gas dampens the amplitude of the ions’ excursions, slowly cooling them to the center of the trap. Therefore, it is particularly easy to trap ions for indeinite periods of time in QITs. Any ion with any initially trapped trajectory will eventually end up at the exact center of the trap. However, if there are many ions, this effect increases local space charge, so that there is a balancing act in which damping pushes the ions inward while coulombic repulsion pushes them outward. Thus, a deined number of ions can be trapped, and this is known as the “space-charge” limit. Furthermore, as collisions increase, the ions are translationally cooled to the center of the trap, but are vibrationally excited at the same time. For this reason, helium is usually used as a collision gas because it decreases the center-of-mass collision energy of the ions and decreases vibrational activation. However, at high ion loadings in the trap, ions will be pushed far enough outward that the vibrational activation will warm some ions enough to start fragmentation. Hofstadler et al. have termed this effect “multipole storage-assisted dissociation” (MSAD) [19]. The trapping dimensions in QITs and LITs are essentially simple harmonic oscillators, at least near the center of the traps. Because of this, in addition to trapping ields, ions have their own resonant frequencies inside the trap which can be used to excite them resonantly to high excursion trajectories. At high trajectories, their vibrational activation will cause them to fragment, allowing MS/ MS experiments to be conducted. Resonant excitation is also used to selectively

2.6 Quadrupole Ion Trap Mass Spectrometers

Figure 2.14 MALDI ion trap mass spectrometry.

eject ions or to eject them using a SWIFT function [86], similar to that used in FTICR-MS. Collisional activation experiments in QIT instruments have the interesting effect that, while the precursor ion is fragmenting, the fragment ions are not resonant and quickly cool back to the center of the trap, which means that the fragmentation is very selective. This effect is the same as has been observed for SORI-CAD on an FTICR mass spectrometer. A typical MALDI-QIT instrument diagram is shown in Figure 2.14, except that a microchannel plate is drawn instead of a conversion dynode/electron multiplier detector. This instrument is similar to the design of Krutchinsky et al. [11]. In this instrument, MALDI ions are generated in a plume that is directed into a transfer RF-only multipole (usually a hexapole or octopole). Ions can be either transferred directly into the ion trap or trapped and cooled in the RF-only multipole before being pulsed into the trap. Either way, in order to achieve the best sensitivity and detection limits in the instrument, it is critical that the ion pulse is synchronized to the RF of the trap. If the RF voltage is high when the ion pulse reaches the trap, the ions will be repelled; if it is a large negative voltage, the ions will be accelerated and thrown across to the other side of the trap (note: this assumes positive ions; negative ions are the reverse). For highest transmission and trapping eficiencies, ions must enter the trap when the RF voltage is close to zero. Since the cycle period of the RF trapping ield is about 1 μs, the window for entrance is ∼10 ns, which is comparable to the pulse width of the UV lasers typically used (∼3–5 ns). If the ion cloud is allowed to disperse in time, then the trapping eficiency will be low. MALDI ions trapped in a QIT will experience a competition between activation and cooling, as discussed above. The MALDI laser generates a plume with ions of considerable internal energy, generating metastable ions. Plume expansion will result in adiabatic cooling of the ions, while collisions with neutrals upon extraction will result in activation. Transfer and trapping of the ions, which usually occurs at ∼1 × 10−4 mbar of helium, will generally cool the ions. However, the long time frame of ion trap experiments (typically tens of milliseconds) means that metastable ions have an opportunity to fragment so that some dissociation is inevitable. Clearly, MALDI-QIT mass spectrometry experiments involve some balancing between activation and cooling, similar to many other MALDI

75

76

2 MALDI Mass Spectrometry Instrumentation

instruments. MALDI sources, as noted above, generate a large number of matrix cluster ions which contribute to a “chemical noise” background. In general, chemical noise is basically matrix clusters with sodium, protons, and any other contaminants (e.g., detergents, phosphate, ingerprint oils) present in the sample. This generates a hump of peaks at every m/z value throughout the spectrum – a “peak at every mass” with an intensity that correlates to hundreds of ions. Because QIT systems are capable of detecting single ions hitting the detector, Krutchinsky and Chait recognized that, even if an expected analyte peak is not observed at a particular m/z-value, MS/MS scans which eject all other peaks prior to fragmentation should be able to observe the MS/MS spectrum of the analyte because the chemical noise is essentially eliminated in these scans. Thus, these authors devised a “hypothesis-driven” scan mode [97] where expected m/z ranges are isolated and fragmented, even if there is no peak observed. For example, if a peptide is observed that may be partially phosphorylated at low stoichiometry, an m/z region 80 Da higher is quickly isolated and fragmented. If the phosphopeptide is there, its tandem mass spectrum should be seen. The same concept can be applied to any predicted (but not observed) peak. Overall, QITs and the more recent LITs are robust, cheap mass spectrometers with high scan speed and high sensitivity. The MS/MS capability of these instruments is very advanced, due to the fact that all ion modes are stable – which means a very high percentage of fragment ions can be observed. The resolving power of these instruments is about (1–2) × 10–3 and the mass accuracy at m/z 1000 is usually about 0.1%, or ∼1 Da. While the m/z range of these instruments is variable, on commercial instruments it is usually limited to between 100 and 2000 Da. Thus, provided that mass accuracy or resolution is not important, and that the m/z range to be observed is below ∼2000 Da, these instruments offer a cheap, eficient, and robust means of analyzing samples. However, these caveats are also their downfall. Frequently, QITs are used in proteomics experiments which require high mass accuracy (and resolution) in the mass range from ∼700 to 4000 Da for tryptic peptides. The resulting data are then used to search databases in an attempt to assign identities to the proteins from which the peptide MS/MS data are generated. Because of the low mass accuracy and the depth of normal sequence databases, the false-positive rate on these database searches is often >50% (i.e., >50% of the assignments are wrong, but which 50% is not clear) [23, 24]. Consequently, extensive data curation and manual data interpretation is required to correct for this false-positive rate, though the improved data quality as obtained with FT-MS instruments eliminates this problem. Furthermore, using a MALDI source for proteomic samples often results in masses that are beyond the m/z range of a QIT. For example, tryptic peptides from glycoproteins routinely generate peptide and glycopeptide masses beyond 3 kDa. Overall, QITs are extremely cheap, robust mass spectrometers that generate high-eficiency MS/MS data, but they are limited to the low-molecular mass range 5 in QIT and LIT instruments [99]. There is a caveat to this, however, namely that QIT and LIT instruments, being subject to the Mathieu stability diagram, can only trap a certain range of ions, with a very sharp cut-off in the trapping range at low m/z. If the fragment ions that are generated fall below this cut-off, they are not stable and will fall out of the trap. Furthermore, the addition of a resonant excitation pulse at a particular m/z further destablilizes low-m/z ions. The rule of thumb is that excitation at a particular m/z will eject all ions below one-third of that m/z value (e.g., excitation at m/z 1000 will eject all ions below m/z ∼300), though this rule is being greatly weakened by some new high-voltage pulsed-excitation methods [95]. ECD (see Section 2.5.1), which was irst introduced in FTICR instruments, has also been reproduced in QIT instruments [100, 101], albeit with great dificulty due to the problem of achieving simultaneous stable trapping of electrons and ions. However, a derivative method, ETD [95, 96], has been shown to be extremely effective. ETD involves the reaction of even-electron, multiply charged ions ([M + nH]n+ with n ≥ 2) with low-molecular-mass radical anions (M−•). Provided that the correct radical anion molecule is chosen, it will donate an electron to the even-electron ion, making it into a radical cation ([M + nH](n−1)+•), which then undergoes similar free-radical chemistry that results in ECD-like fragments. Due to the high eficiency of trapping of fragment ions in QIT and LIT instruments, and to the higher rate of reaction of multiply charged ions over singly charged ions, the reaction can be driven to generate extremely high-quality MS/MS spectra. Today, ETD is being widely implemented with many varied designs on most hybrid instruments which have multipole ion traps. Although photodissociation has also been performed extensively on QIT instruments [102–110], the results have not been so widely useful that the method has been implemented on commercial systems. There are two reasons for this: (i) the excursion paths of ions in QIT instruments is often wider than the laser beams, so that some ions may not be irradiated; and (ii) with IRMPD in particular, where hundreds to thousands of photons must be absorbed in order to heat the molecules up for dissociation, ions experience a competition between heating (from the laser) and cooling (from collisions and blackbody radiation). At 10−3 mbar, the collisional

2.7 Hybrid Mass Spectrometers

cooling is substantial, so that higher laser powers or controlling the pressure via pulsed valves is necessary. UV photodissociation has not been extensively explored in QIT and LIT instruments.

2.7 Hybrid Mass Spectrometers

Each of the individual types of mass spectrometer – TOF, FTICR, quadrupole/ multipole ion guides, and QITs – has both advantages and disadvantages, as discussed above. However, several research groups have developed “hybrid” MALDI mass spectrometers which combine the various components to achieve interesting combined advantages. 2.7.1 Quadrupole TOF Mass Spectrometers

MALDI-QTOF instruments basically follow an ion optical arrangement similar to that shown in Figure 2.16. The irst quadrupole, Q0, is simply used to cool the ions’ kinetic energy and focus them onto the quadrupole axis so that they enter Q1 with a well-deined initial trajectory, and is always used in RF-only mode. Q1 is a mass-iltering quadrupole which can either be used in RF-only mode to pass a wide range of mass/charge values, or in mass-iltering mode to select particular

Figure 2.16 Quadrupole time-of-light hybrid.

79

80

2 MALDI Mass Spectrometry Instrumentation

ions. Ions leaving Q1, with either broad m/z ranges or selected ions, are then directed into Q2. By adjusting the offset voltages of the quadrupoles, the ions can be accelerated into Q2, where they can collide and fragment. Fragment ions are then directed into the orthogonal TOF where they can be detected. QTOF hybrids are simple to use and reasonably lexible. The quadrupoles allow beam focusing, mass selection and dissociation, while the TOF provides high sensitivity and resolving power. The orthogonal injection geometry into the TOF decouples the surface desorption from the detection so that irregular, bumpy MALDI crystals or other surfaces (such as thin-layer chromatography plates) can be used. The plume velocity does not affect mass accuracy, so a wider range of laser luence can be used, which improves sensitivity. Moreover, the focusing effect of the quadrupoles allows ions to be injected into the oTOF in a thin beam; this improves the resolution and as a result the QTOF instruments can easily achieve >104 resolving power. Furthermore, this geometry allows fragmentation in Q2 with a rapid and sensitive detection of the products. In addition, ions can be accumulated in Q2 prior to dumping them into the TOF. This “selected ion accumulation” can improve detection sensitivity by as much as 100-fold, depending on the duty cycle of the instrument. The LINAC arrangement from Sciex [111] assists with these experiments by allowing an improved ejection of these stored ions from Q2. 2.7.2 Quadrupole FT Mass Spectrometers

Placing a quadrupole mass ilter in front of a FT mass spectrometer is also useful. The irst MALDI-Q-FTICR-MS was built by Brock et al. [8], and is shown schematically in Figure 2.17. This instrument can generate ions, focus them through the irst quadrupole and octopole into the mass-resolving quadrupole, selectively accumulate or fragment ions in the accumulation hexapole, and then transfer the accumulated ions to the ICR cell using the inal octopole and quadrupole ion optics. Selected ion accumulation allows the operator to ill the ICR cell to its space charge limit with only the ion of interest, which allows higher-quality MS/MS spectra in the cell. It also, for the same reason, eliminates most chemical noise in the cell. Furthermore, mass selection in a quadrupole is much faster and simpler than selection in the ICR cell. Bruker and Thermo Fisher have each adopted similar ion optical arrangements for their commercial MALDI-FTICR-MS instruments. Bruker’s instrument (Figure 2.18) uses a combined MALDI/electrospray ion (ESI) source which focuses the ions through two ion funnels in succession into a quadrupole for ion mass selection and ion accumulation, followed by a hexapole for CAD fragmentation. A divided octopole before the quadrupole and a hollow electron emitter behind the ICR cell respectively permit ETD and ECD for structural analyses. A hexapole ion guide transfers the ions into the ICR cell. Compared to the original electrostatic ion transfer optics, this has the disadvantage of a rather restrictive upper mass limit for MALDI-generated ions.

2.7 Hybrid Mass Spectrometers

81

Figure 2.17 Quadrupole Fourier transform mass spectrometer (FTMS). Reprinted with

permission from Ref. [8]. © 2003 American Chemical Society.

ICR cell Hexapole ion transfer optics

lon funnels Unified ESI / MALDI source

CI source for ETD

ESI sprayer

Hollow electron emitter for ECD

Divided octopole

Beam valve

Quadrupole Collision chamber

ESI capillary

Turbo Pump 1

Turbo Pump 2

Hexapole

Turbo Pump 3

Turbo Pump 4

er

Las

MALDI target plate

Figure 2.18 MALDI-ESI/FTMS hybrid for Solarix instruments from Bruker. Illustration

courtesy of Bruker Daltonik.

Refrigerated superconducting Magnet

82

2 MALDI Mass Spectrometry Instrumentation FTMS Data Linear Ion Trap Data

Actively Shielded 7 Tesla Magnet

Figure 2.19 Thermo LTQ-FT mass spectrometer. Reprinted from Ref. [112] with permission

from Elsevier.

The standard Thermo LTQ-FT instrument [112] (Figure 2.19) does not have a mass selection quadrupole prior to its linear ion trap, but it is able to use the linear ion trap as a rapid-scanning, low-resolution mass spectrometer in addition to ejecting ions into a long multipole which transports them to the ICR cell. Typically, three scans can be performed in the ion trap at the same time as performing one highresolution scan in the FTICR. The Thermo LTQ ion trap adds substantial scan lexibility, automatic gain control, and single-ion detection limits, but without any iltering quadruple prior to the trap there is no selected ion accumulation capability. Additionally, the Thermo LTQ software substantially limits the ability of the user to adjust pulse sequences or experimental parameters; this has the advantage of preventing inexperienced users from detuning the instrument, but limits experienced users from exploring and testing different pulse sequences. In 2011, Thermo Fisher replaced its standard FT-MS instrument with the Orbitrap (see Section 2.7.5). 2.7.3 QIT-TOF Mass Spectrometers

A more recent hybrid involves the coupling of QITs to TOF mass spectrometers. This instrument coniguration originated in the laboratory of Lubman, but has recently been commercialized by Shimadzu (Figure 2.20). These instruments allow MSn experiments in the ion trap, followed by ∼104 resolving power in the TOF. While these performance characteristics are reasonable, no reliable data have yet been published by independent research groups to conirm them. The dificulty is that ions, when ejected into a TOF, must be conined to a small spatial region with little or no initial kinetic energy variance in order to achieve a higher resolving power. Ions oscillating in an ion trap generally have a wide range of initial velocity vectors, so the RF trapping ield must be shut off just as the ions are ejected into the TOF. Although this can be done, the results obtained to date have been less than encouraging.

2.7 Hybrid Mass Spectrometers

Figure 2.20 MALDI ion trap time-of-light mass spectrometry.

Nonetheless, this type of hybrid is interesting for studying ion–molecule reactions. Ion traps in general are very good for studying these reactions because ions are contained under high pressure for long periods of time. In particular, the newer ETD experiments should be possible on such an instrument, as this would allow the generation of both odd-electron and even-electron fragmentation, which tend to be complementary. 2.7.4 Ion Mobility oTOF Mass Spectrometers

Ion mobility spectrometry (IMS) adds another dimension to MS analysis by separating ions in the gas phase according to their collisional cross-section – that is, their size and shape. The physical principle behind IMS is the collision of the ions with neutral gas molecules or atoms, typically N2, He, or Ar, varying in pressure from 1 mbar to ambient pressure. Because the collisional cross-section of an ion is dictated by its exact conformation in addition to its mass, IMS can allow the differentiation of structural isomers or of molecules with similar or even equal mass that would not be discernible in a pure MS measurement [113]. Furthermore, if the ion mobility gas is spiked with a suitable chiral modiier, even stereoisomers (i.e., molecules differing only by their R/S-coniguration) may be differentiated [114]. If crude samples are analyzed, IMS can be utilized beneicially as a “ilter” to simplify the complex mass spectrum. This feature is particularly useful in the MALDI-MS analysis of small molecules, because it can be used to remove interference from abundant matrix cluster ions. In MALDI imaging applications, IMS can serve to separate whole classes of biomolecules that are ionized simultaneously from the tissue. “Trend lines” can be established in the mobility chromatogram by plotting the ion m/z values on the x-axis and the IMS arrival time on the y-axis.

83

84

2 MALDI Mass Spectrometry Instrumentation

Figure 2.21 MALDI-IMS mass spectrometry

conformation space obtained for a mixture of model species representing the four major biomolecular classes (peptides, lipids, carbohydrates, and oligonucleotides). For each class, 7 to 17 model species with

different masses have been analyzed. Dashed lines indicate trend lines for the four classes, around which members of the groups are typically detected. Reproduced with permission from Ref. [116].

This allows the differentiation of peptides, carbohydrates, lipids and oligonucleotides (Figure 2.21) [115, 116]. IMS is most widely used in combination with hybrid quadrupole TOF mass spectrometers. The most prominent example is the commercially available Synapt G2-S mass spectrometer, the coniguration of which is depicted in Figure 2.22. The Synapt uses the so-called traveling-wave or T-Wave™ principle, whereby ions are not dragged through the collision gas by a constant electrical ield (as in classical drift tubes) but rather by a train of voltage pulses [117, 118]. The T-wave cells consist of stacked-ring electrodes. When an opposite-phase RF is applied to consecutive electrodes a potential well is formed in the central region of the rings, and consequently T-wave cells can eficiently serve as ion guides, similar to quadrupole or hexapole ion guides. In the traveling-wave mode, a DC voltage is superimposed temporarily on the RF of an electrode. This DC voltage is subsequently swept over successive electrodes so as to generate a moving electrical ield – the traveling wave. Ions falling within a certain m/z and shape region can “surf” on the traveling wave, while all others may “roll over” and thus spend a more extended time in the drift cell (Figure 2.23). In effect, a good resolving power R = Ω/dΩ, where Ω is the collisional cross-section, can be achieved in this way, and the length of the IMS cell can be kept reasonably short. For the Synapt G2, a possible IMS resolution of about 40 is typically denoted. In applications where overlapping ion

2.7 Hybrid Mass Spectrometers

Figure 2.22 Schematic diagram of the MALDI Synapt G2-S HDMS instrument Illustration courtesy of Waters Corp.

Figure 2.23 Schematic diagram illustrating mobility separation in the T-Wave cell. Electrodes to which a DC voltage pulse is temporarily applied are shown as gray boxes.

The high-mobility ions have kept up with the wave and the low mobility ions have rolled over the top. Reproduced with permission from Ref. [117].

85

86

2 MALDI Mass Spectrometry Instrumentation

packages may produce a broadened but unresolved IMS peak, it may be suficient to select the rising versus the falling edge of the chromatographic peaks for the differentiation of compounds. However, this approach will often require the comparison of drift times with those of the pure compounds. Unfortunately, to achieve optimal resolution for selected ion packages may come at a cost of the overall transmission window, because this is restricted to 200 time bins. The time window for the drift time measurement ranges from 33 μs to 90 ms. IMS lends itself to working in combination with MALDI ion sources because of the pulsed nature of the ionization event. To further improve the “starting conditions” for the gas-phase separation, the MALDI Synapt uses a “trap” cell to store the incoming ions before they are released in packets into the ion mobility cell (Figure 2.22). This feature also allows the instrument to be used without further modiications when a continuous ESI ion source is mounted instead of a MALDI source. In this case, a short helium cell is incorporated in between the trap and the IMS cell, and this interface reduces unwanted scattering and fragmentation processes during ion transfer from the trap to the IMS cell. It also allows the IMS cell to be operated at higher N2 pressures of up to 3 mbar for improved resolution. The role of the “transfer” cell following the IMS cell is to transfer the mobility-separated ion packets to the TOF analyzer. If no gas-phase separation of ions is needed, the IMS cell is simply operated as a transmitting element in the so-called “RF-only” mode. An oTOF (as described in Section 2.4.2) is used for the high-resolution m/z analysis in the inal stage of the Synapt instruments, except that analysis in a double-pass relectron is possible for improved mass resolution. The instrument can be operated in three different modes that provide different mass resolutions and sensitivities. In the sensitivity and resolution modes, the relector is operated as a single-pass ion mirror, providing a mass resolution (FWHM) of >10 000 and >20 000 for the two modes. However, in resolution mode the sensitivity is lower by a factor of about ten. In high-resolution mode, the ion packages are relected three times by the ion mirror for improved light time compensation of their kinetic energy distribution. This increases the mass resolution to >40 000, albeit at the cost of some further loss in sensitivity. The typical time bin of an ion mobility separation is in the millisecond range. Because this time window is substantially larger than the duration of the TOF measurement, many single mass spectra can be recorded during one IMS separation. Under optimal conditions the overall transmission eficiency is close to that achieved without IMS. Depending on the complexity of the sample, limits of detection in the subfemtomolar range are achieved in the MALDI-IMMS analysis of peptides. In the Synapt, the MALDI ions are generated at a ine vacuum of about 0.1 mbar and transferred via a hexapole ion guide into a “stepwave ion guide.” The “stepwave” serves to reduce the background of neutral particles, because only ions are lifted onto the ion axis of the mass spectrometer. An initial selection of ions of interests is achieved with a quadrupole preceding the trap cell. For most standard applications a m/z range of up to 4000 or 8000 Da can be selected, depending on

2.7 Hybrid Mass Spectrometers

the RF generator used. For the RF-only mode (no ion selection) the transmission range is about fourfold higher. Both, the trap and the transfer cells can be utilized for the low-energy CAD of selected ions. If the trap cell is used, the generated fragments are separated in the IMS cell according to their ion mobility constants. Dissociation of these species in the transfer cell provides MS3 data and, hence, additional structural information. The second-generation fragments are grouped according to the drift time of their respective precursors, and may thus be differentiated even if fragments with the same mass were produced from different precursor ions. If only the transfer cell is used for CAD, the IMS separation allows fragment ions stemming from different precursor ions of similar m/z value to be distinguished. This feature can be very useful in the MALDI-MS analysis of complex samples, for example, that of tryptic fragments eluted from SDS–PAGE gel-separated proteins, as well as in MALDI-MS imaging. The Synapt G2, as well as other IMS/MS instruments, offers a wide choice of experimental parameters for adaptation to speciic analytical problems; however, the ability to obtain optimal results requires substantial understanding and skill by the operator. In non-commercial research instruments, MALDI has also been coupled with classical drift tube cells [119]. In these cells, a constant near-uniform electric ield is created along the axis of the cell. Drift tube cells have the advantage of providing a closer linearity between drift time through the cell and collisional cross-section and, therefore, allow a more direct determination of this parameter – and hence the coniguration of the biomolecules. Lately, calibration algorithms have been described that allow more accurate cross-section values also from T-Wave cells [120]. While both T-Wave and classical drift tubes separate the injected ion packages in space, a third widely used IMS principle builds upon a differential mobility analysis and establishes a temporal separation. This principle is often also referred to as high-ield asymmetric ion mobility spectrometry (FAIMS [121]). Due to its functional principle, FAIMS cells are better suited in combination with a continuous ion beam (e.g., provided by ESI) and ambient pressure conditions. Therefore, FAIMS cells are typically not used in combination with MALDI. 2.7.5 Orbitrap

The Orbitrap is the latest addition to the collection of different mass analyzers. Even though the basic principle was irst described in 1923, the technique has only recently been reduced to practice by Alexander Makarov [122]. As shown in Figures 2.24 and 2.25, the core of the trap consists of a central spindle-shaped electrode, surrounded coaxially by a specially shaped cylinder as the counter-electrode. The exact shape of the electrodes is critical and requires very careful machining. A DC-electric ield is provided by applying a voltage of typically a few kV between the two electrodes. Ions of suitable energy, when injected tangentially into the space between the two electrodes, will circle around the center electrode in nearcircular trajectories overlayed by an oscillatory motion along the symmetry

87

88

2 MALDI Mass Spectrometry Instrumentation

LTD XL Linear Ion Trap

(a)

C-Trap

HCD Collision Cell

MALDI Source

(b) laser mirror CCD camera Orbitrap Mass Analyzer ND filters lens

dichromatic mirror

aperture / lens, # 2

p = 0.1 mbar

skimmer, #4 to q0, q1, LTQ XL, Orbitrap analyzer

sample plate, # 1

q00 to pump

RF DC aux rods rods #3 collision gas

Figure 2.24 (a) Schematics of the linear ion trap–Orbitrap hybrid mass analyzer equipped with a MALDI source from Thermo Scientiic; (b) Thermo Scientiic MALDI source design

and schematics of the collisional cooling interface. Adapted with permission from Thermo Scientiic.

axis of the trap. If the two electrodes are of the correct shape, the oscillatory axial motion of the ions will be simple harmonic with a frequency proportional to their (m/z)1/2-value. It is this axial frequency ωa which is recorded and taken as measure of m/z. The frequency of the circular motion depends in a complicated fashion on the ion energy initial position, and so on, and cannot therefore be used for mass analysis. For the mass measurement, a tightly focused ion package with ions of different m/z ratios is injected into the electrostatic ion trap (i.e., Orbitrap device). Ion packages are tangentially moving into the Orbitrap through a small hole in the outer electrode, after having been prepared for injection by a specially shaped so-called curved linear ion trap (C-trap). The pressure in the Orbitrap device must be 10−9 mbar or less in order to prevent collisions with gas molecules that would result in a rapid axial dephasing of the ion packet and a loss of signal. The image current across an equatorial gap between the two halves of the outer cylinder is detected in the measurement. This time-domain signal must then be Fourier-transformed

2.7 Hybrid Mass Spectrometers

static, variable aperture attenuator N2 Laser

lens L1

mirror

mirror Thermo Scientific Q ExactiveTM

objective lens - centrally bored movable sample stage - x, y, z

mirror - centrally bored

observation mirror - off axis

CCD Camera

laser beam path ion trajectory observation path base plate

= = = =

Figure 2.25 TransMIT’s AP SMALDI source design for mass spectral imaging attached to Thermo Scientiic Q Exactive instrumentation. Adapted with permission from Thermo Scientiic.

to obtain the axial frequency distributions, and with a suitable calibration factor the m/z-values of the injected ions are revealed, very much like the signal in FTICR-MS (see Section 2.5). In contrast to typical FTICR and Quadrupole ion trap mass spectrometers, however, the Orbitrap device provides a large space charge limit and can be illed with up to 106 charges. This results in a high tolerance towards an increased number of charges and reduced space charge effects – that is, deviations of mass measurement – in the Orbitrap device. Finally, the mass spectral performance provides an improved SNR in Orbitrap detection. The Orbitrap technology is the property of Thermo Fisher Scientiic, which has now marketed two different product families with several variants. The Q Exactive™ instrument, as shown in Figure 2.25, is a bench-top Orbitrap analyzer with a 90° bend of the ion path to limit the footprint of the instrument. Its mass-resolving power is 140 000 at m/z 200, and the accessible mass range is up to 6000 Da.

89

90

2 MALDI Mass Spectrometry Instrumentation

A schematic of the MALDI LTQ Orbitrap XL is shown in Figure 2.24. This is a hybrid ion trap–Orbitrap instrument in certain respects similar to the ion trap FTICR mass spectrometers described in Section 2.7.2. The instrument can be operated as either a stand-alone linear quadrupole ion trap or in similar fashion to a bench-top Orbitrap device, or as a hybrid of the two. In the former case ions are injected into the LTQ XL linear ion trap and ejected sidewards for detection. In the Orbitrap-only-mode, all multipoles in front of the C-trap are set to massindependent transmission, after which the ions are injected into the Orbitrap device. In hybrid mode, the ions can be preselected in the LTQ XL before being injected into the Orbitrap analyzer for high mass resolution. The LTQ XL and Orbitrap devices can also be operated quasi-independently: whilst the Orbitrap device analyzes ions of a former injection cycle at high resolution in the background, the LTQ XL can perform several independent MS/MS scans of the next injection cycles at low resolution. In its various scan modes the LTQ XL has a mass resolution of 1 Da up to 2500 Da, and a scan range of up to 4000 Da. The resolving power of the Orbitrap mass analyzer is up to 130 000 @ m/z 400 (LTQ Orbitrap XL), with the same upper mass limit as the LTQ XL. The dynamic range in Orbitrap mode has been shown to reach 5000 : 1 in single-scan operation [123]. The hybrid Orbitrap instrument offers the option of standard CID with a precursor selection and dissociation in the LTQ XL with a collision energy of some 10 eV. The Exactive instrument does not have the CID option, but if an AP MALDI source is attached it can generate desorption-induced ISD spectra. Both instruments can be equipped with yet another quadrupole device positioned after the C-trap. In an initial step, ions are transmitted through the C-trap into this additional HCD multipole where they may be subjected to a “higher-energy collision dissociation (HCD)”. The collision energy of some 10 to a few 100 eV performed in the HCD quadrupole increases the fragmentation yield; peptide fragment ion spectra are seen all the way down to the immonium ions. After fragmentation the ions are back-transferred into the C-trap, and subsequently into the Orbitrap device for mass analysis. The back end of the HCD collision cell can be conigured for the attachment of an ETD ion source in the ESI mode of operation. Although, originally, the Orbitrap device was designed for ESI, with an AP ion source, different MALDI sources have since been developed and are now commercially available. Contrary to classical MALDI-sources, these operate at atmospheric or intermediate pressure of typically 10−2 mbar. The collisional cooling of such sources limits the PSD of ions to prepare them for longer detection times in comparison with axial TOF analyzers, for example. A schematic of the Thermo Scientiic MALDI source [124] is shown in Figure 2.24b. The MALDI source region includes: a movable sample plate (#1) with a 1 μm positioning accuracy and approximately 12 × 8 cm2 free travel to access sample plate sizes of the standardized microtiter plate format, a RF-only quadrupole with auxiliary DC resistive rods (DC aux) mounted parallel to and in the gaps between the RF rods (#3). The skimmer (#4) separates the source region held at about 0.1 mbar from the next pressure region. On the front side, a 4 mm aperture (#2) is placed equidistant from both the RF rods assembly and the sample plate,

2.7 Hybrid Mass Spectrometers

about 1 mm away. The purpose of this is to serve as an ion optics element, providing a physical barrier between the high RF voltage of the quadrupole rods and the sample plate, and it also carries a DC voltage to shape the ion cloud and focus ions into the q00 region. An upper limit of the aperture diameter of 4 mm was chosen so as not to obstruct the laser beam. The ion optics downstream of the MALDI source is identical to that in the electrospray instrument starting from the ∼10−3 mbar region behind the skimmer (#4), 10–6 mbar in the linear trap and down to the Orbitrap mass analyzer. In the standard coniguration, ions are generated using a commercial N2 (λ = 337 nm) laser (model MNL-100, 106-LD; LTB Lasertechnik Berlin GmbH) with a repetition rate of up to 60 Hz, a maximum energy per pulse of 75 μJ, and 3 ns pulse duration. The incidence angle of the laser beam relative to the sample plate normal is 32°, and the desorption spot size on the sample ranges from ∼60–80 μm2 to 100–120 μm2, depending on the laser luence applied. At irst approximation, this spot size deines the spatial resolution in the imaging mode (for more recent developments on higher spatial resolution, see below). The potential of the DCaux rods is pulsed in such a way as to secure maximal ion transmission into the analyzer. The typical timing is such that the ions of a preselected number of laser exposures are accumulated. Further information on the source design of the intermediate pressure MALDI source and the implementation of Automatic Gain Control in MALDI mode of operation can be found elsewhere [124]. The ion source of the Exactive instrument, as shown in Figure 2.25 (red, green and blue boxes), has been speciically developed for MALDI imaging applications and, accordingly, is of a very different design. The general set-up is similar to that described in Ref. [125] (for more information on MALDI imaging, see Chapter 4 of this book). This source was developed in the laboratory of Bernhard Spengler at the University of Giessen, is marketed by TransMIT GmbH ([email protected]), and combines a maximum spatial resolution with the high mass-resolving power of the Orbitrap analyzer. The optical beam path resembles that of a laboratory microscope. A lens (L1) focuses the laser beam to an intermediate focus, which is then imaged onto the sample surface by a multi-lens quartz objective. The objective has a numerical aperture of 0.429; that is, the beam uses a 50° full plane angle of the space in front of the sample. At luences typically used for the mass analysis of standard MALDI samples or histological sections in the imaging mode, these optics desorb ions from a spot of minimally 3 μm diameter. The objective has a central bore which holds an extended heated capillary for ion transport into the orbitrap, and which separates atmospheric pressure from the irst stage instrument pressure. The objective has a free working distance of 40 mm, allowing for ion optical elements to collect a maximum of the desorbed ions into the capillary. Fixed and variable attenuators and two alignment mirrors complement the optical beam path. A nitrogen laser (MNL-100, LD-106; LTB Lasertechnik Berlin GmbH) is also used for this set-up. The x, y, z-movable sample stage has a positioning accuracy of ≤1 μm and a free travel of 50 mm in all three dimensions. One of the main applications of both Orbitrap systems and their MALDI ion sources is for the imaging of biological specimens. The Thermo Fisher Scientiic

91

92

2 MALDI Mass Spectrometry Instrumentation

source is limited in its spatial resolution to somewhat above the size of a typical biological cell, whereas the TransMIT source can resolve even subcellular structures. However, such high spatial resolution comes at the expense of detectable ions, due to the constant number of analyte molecules per sample area. The laser luence H needed to desorb a detectable amount of ions is increased with decreasing spot size, while the total laser pulse energy per spot remains almost constant [126]. This increases the optical penetration depth into the sample (see Chapter 1, Section 1.3) and, as a result, fewer spectra can be collected from a given sample location for a given tissue section thickness. More importantly, however, the increased laser luence may lead to an increased ion excitation and fragmentation of labile biological macromolecules, thus limiting their non-ambiguous detection. In contrast, Figure 4.10 (see Chapter 4.7.1) shows a convincing example that takes full advantage of the high resolving power of the Orbitrap mass analyzer. Here, two phospholipids of only 56 mDa difference in mass showed a completely separated spatial distribution in a specimen of mouse urinary bladder wall. Such small mass differences would be expected to be all but rare in the analysis of biological specimens, with their thousands of different molecular species in any given cell, and to remain unidentiied in almost all other (less highly mass-resolving) mass spectrometers used for MALDI imaging. The Thermo Scientiic ion source can be retroitted to the LTQ Orbitrap XL line of instruments, while the TransMIT ion source can be retroitted to the Exactive series instrumentation. The MassTech AP MALDI PDF+ ion source has also been adapted to both hybrid Orbitrap devices, as well as to bench-top Orbitrap systems. The AP MALDI PDF+ ion source can be langed to the atmospheric pressure interface (API) of ion trap-, ion trap-Orbitrap-, or bench-top Orbitrap instruments. The ion source is equipped with a frequency-tripled Nd:YAG laser operating at 355 nm with 3 ns pulse duration, while the laser beam is typically coupled into a 400 μm core diameter iber forming a spot size of 500 × 600 μm2 on the plate (a smaller-diameter iber can also be applied). The ease-of-use of this ion source for regular MALDI applications, and the possibility of a rapid source exchange between MALDI and ESI, have led to this source becoming an excellent tool for today’s analytical studies. Details on the coupling of this source with the Exactive series instrumentation are available from the manufacturer.

2.8 Future Directions

Currently, mass spectrometry instrumentation for MALDI ion sources has considerable room for growth. Although the above-described hybrids are highly effective, very few are presently commercially available; nonetheless, new hybrids continue to be developed and revised fragmentation techniques are being pursued. Within the ield, however, the greatest need for improvement is in sample preparation. Currently, several companies can provide sample preparation robots to perform a range of tasks, from the detection and slicing of gel electrophoresis

Deinitions and Acronyms

spots, to enzymatic digestion, to spotting the resulting peptides onto the MALDI sample plate along with the matrix. These robots not only greatly improve the reliability and robustness of the analysis, but are also clearly very helpful for achieving maximum data throughput. Indeed, some robots are also capable of depositing spots or trails of chromatographic efluent, allowing these separation methods to be coupled with MALDI. One interesting method in this respect is the coupling of thin-layer chromatography (TLC) or other planar separation techniques with MALDI-qTOF [6, 127–129] and MALDI- FTICR-MS [14] instruments. Although the coupling of TLC to MALDI-TOF has been previously explored [130–132], the use of such an approach has been limited due to the low resolving power achievable with axial TOF instruments due to the irregular, bumpy surfaces involved. Yet, both MALDI-qTOF and MALDI-FTICR-MS instruments decouple the source from the analyzer, so that irregular surfaces, surface static charging and variable kinetic energy distributions are no longer problematic. Consequently, this method is likely to be extended to a wide variety of surface techniques, including two-dimensional gels and/or microluidic channels. Since the advent of MALDI, explorations of the different types of lasers and matrices that can be used have been continuing. One particularly interesting approach here is in the use of IR-MALDI [132–137] (in the 2 to 10 μm range) and liquid matrices (including water). Indeed, some use has been made of these systems, and today IR-MALDI lasers appear to generate ions with less internal vibrational energy than do traditional UV lasers. Furthermore, IR-MALDI appears to generate higher charge state ions. Whilst there is substantial scope for further investigation in this realm, however, the ield is currently limited by a lack of reliable IR lasers. Deinitions and Acronyms

2,5-DHB 3-HPA ADC AGC AP-MALDI CAD Centroid

2,5-Dihydroxybenzoic acid, a common MALDI matrix 3-Hydroxypicolinic acid Analog-to-digital converter Automatic gain control Atmospheric pressure MALDI Collisionally activated dissociation The center-of-mass of a peak. The best method for determining peak centroids is a nonlinear leastsquares itting of the raw peak data with a function which has the correct peak shape for the instrument. Other methods include: (i) using a center-of-mass calculation (weighted average) of the raw peak data (pretty good); (ii) itting the upper half of the peak with a parabola and reporting the zero of the second derivative (OK if there are many points on the peak); and (iii) taking the highest point of the peak and deining this to be the centroid (generally a bad method, but rapid).

93

94

2 MALDI Mass Spectrometry Instrumentation

CID DE Dalton (Da)

ECD EDD EID Er:YAG

ETD FFT Fragmentation eficiency FT FTICR FWHM

HCCA (also CHCA) ICR IR IRMPD Limit of detection

LINAC LIT MALDI

Collision-induced dissociation Delayed extraction A Dalton is deined as 1/12th of the mass of a 12C atom. It differs from an atomic mass unit (amu), which is deined as 1/16th of the mass of a 16O atom. Electron capture dissociation Electron detachment dissociation Electron-induced dissociation Erbium-doped yttrium–aluminum–garnet, a crystal used for lasers. It lases at a fundamental wavelength of 2.94 μm. Electron transfer dissociation Fast Fourier transform [137] The ratio of the sum of the fragment ions divided by the reduction in the precursor ion abundance. Fourier transform Fourier transform ion cyclotron resonance Full-width at half-maximum, the standard way of deter- mining peak width. It is used to calculate resolving power (see below). alpha-Cyano-4-hydroxycinnamic acid Ion cyclotron resonance Infrared Infrared multiphoton dissociation The smallest signal that can be detected reliably in the mass spectrometer. Because ion sources and transfer optics are not 100% eficient at converting analyte molecules into a detectable signal, the correct way to determine this value is to measure the number of analyte molecules (in moles) that are deposited onto the MALDI target and then run a dilution series until the signal from the mass spectrometer falls below a chosen SNR (usually ∼3). It is best to report the inal analyte molecule number and the SNR at that number. Dilution series experiments are frequently plagued by systematic errors, so that measuring detection limits can often be problematic. These systematic errors are often due to sample carry-over down the dilution series due to sample sticking to the insides of pipettes or sample target plates. Extreme care must be taken to avoid these problems in order to be able to report accurate limits of detection. Linear accelerator Linear ion trap Matrix-assisted laser desorption/ionization

Deinitions and Acronyms

Mass accuracy

MCP MECA MS MSAD MSn NA

Nd:YAG

Noise

OPO oTOF PPM PSD QIT qTOF Resolution

Mass spectrometers are usually characterized by their internal and external calibration mass accuracy, or how closely the measured mass matches the theoretical mass. It is usually reported in Da or in parts per million, but as mass accuracy varies across the spectrum in many mass spectrometers it must also be reported at a particular m/z value. Every mass spectrometer has a characteristic mass accuracy capability, but achieving this “best-case” limit usually requires careful control of the instrumental parameters and ensuring that the ion source and ion optics are clean. Microchannel plate Multiple excitation collisional activation Mass spectrometry Multipole storage assisted dissociation nth order tandem mass spectrometry Numerical aperture, NA = sin(Θ/2), where Θ is the full-angle of the divergence of the beam; thus, NA = 0.2 equals Θ = 22°. Neodymium-doped yttrium–aluminum–garnet, a crystal that is used for lasers. It lases at a fundamental wavelength of 1064 nm, but can be frequency tripled to 355 nm. Most mass spectrometers suffer from white noise (Johnson noise and shot noise), RF interference (i.e., radio stations, luorescent light ballast supplies, switching power supplies, quadrupoles, etc.), and chemical noise (chemical interference peaks). Generally, the noise level reported for signal/noise calculations should be reported as root-mean-square (RMS) of the white noise, but this value can sometimes be dificult to determine if the experiment has much chemical noise. RF noise is not usually dense enough to be a major problem. Regardless, if SNR values are reported, it is important to report exactly how noise is calculated. Optical parametric oscillator. A device for scanning or choosing wavelengths in the 1.5 to 3 μm range. Orthogonal time-of-light Parts per million Post-source decay Quadrupole ion trap Quadrupole time-of-light The lowest m/z spacing where different peaks can be distinguished. It is usually determined at FWHM, but

95

96

2 MALDI Mass Spectrometry Instrumentation

Resolving power

RF Sensitivity

SEM SORI-CAD SWIFT TDC TEM00

TLC TOF UV VC-MALDI VLE-CAD

10% valley and 90% valley positions are also sometimes used. Resolution has units of m/z. Resolving power is the most common way of discussing resolution in a mass spectrometer. It is calculated as M/ΔM, where ΔM is FWHM. For example, an ion at 1000 m/z with a FWHM peak width of 0.3 m/z will have a resolving power of 3000. Resolving power is a unitless parameter. Radiofrequency, anywhere from a few hundred kHz to a few hundred MHz Deined as the slope of a concentration versus signal intensity plot. This term is often used incorrectly to mean detection limits. Secondary electron multiplier Sustained off-resonance irradiation-CAD Stored waveform inverse Fourier transform Time-to-digital converter Transverse electric ield mode indicates that the electric ield of the laser beam is perpendicular to its direction of propagation. TEM00 implies a Gaussian beam proile. Thin-layer chromatography Time-of-light Ultraviolet Vibrationally cooled MALDI Very low-energy-CAD

References 1 Holle, A., Haase, A., Kayser, M., and

Höhndorf, J. (2006) Optimizing UV-laser focus proiles for improved MALDI performance. J. Mass Spectrom., 41, 705–716. 2 Schurenberg, M., Schulz, T., Dreisewered, K., and Hillenkamp, F. (1996) Matrix-assisted laser desorption/ ionization in transmission geometry – instrumental implementation and mechanistic implications. Rapid Commun. Mass Spectrom., 10, 1873–1880. 3 Krutchinsky, A.N., Loboda, A.V., Spicer, V.L., Dworschak, R., Ens, W., and Standing, K.G. (1998) Orthogonal injection of matrix-assisted laser desorption/ionization ions into a

time-of-light spectrometer through a collisional damping interface. Rapid Commun. Mass Spectrom., 12, 508–518. 4 Loboda, A.V., Krutchinsky, A.N., Bromirski, M., Ens, W., and Standing, K.G. (2000) A tandem quadrupole/ time-of-light mass spectrometer with a matrix-assisted laser desorption/ ionization source: design and performance. Rapid Commun. Mass Spectrom., 14, 1047–1057. 5 Loboda, A.V., Ackloo, S., and Chernushevich, I.V. (2003) A highperformance matrix-assisted laser desorption/ionization orthogonal time-of-light mass spectrometer with collisional cooling. Rapid Commun. Mass Spectrom., 17, 2508–2516.

References 6 Ivleva, V.B., Sapp, L.M., O’Connor, P.B.,

7

8

9

10

11

12

13

and Costello, C.E. (2005) Ganglioside analysis by thin-layer chromatography matrix-assisted laser desorption/ ionization orthogonal time-of-light mass spectrometry. J. Am. Soc. Mass Spectrom., 16, 1552–1560. Moyer, S.C., Budnik, B.A., Pittman, J.L., Costello, C.E., and O’Connor, P.B. (2003) Attomole peptide analysis by high pressure matrix-assisted laser desorption/ionization Fourier transform mass spectrometry. Anal. Chem., 75, 6449–6454. Brock, A., Horn, D.M., Peters, E.C., Shaw, C.M., Ericson, C., Phung, Q.T., and Saloma, A.R. (2003) An automated matrix-assisted laser desorption/ ionization quadrupole Fourier transform ion cyclotron resonance mass spectrometer for “bottom-up” proteomics. Anal. Chem., 75, 3419–3428. Budnik, B.A., Moyer, S.C., Pittman, J.L., Ivleva, V.B., Sommer, U., Costello, C.E., and O’Connor, P.B. (2004) High pressure MALDI-FTMS: implications for proteomics. Int. J. Mass Spectrom. Ion Processes, 234, 203–212. O’Connor, P.B., Budnik, B.A., Ivleva, V.B., Kaur, P., Moyer, S.C., Pittman, J.L., and Costello, C.E. (2004) A high pressure matrix-assisted laser desorption ion source for Fourier transform mass spectrometry designed to accommodate large targets with diverse surfaces. J. Am. Soc. Mass Spectrom., 15, 128–132. Krutchinsky, A.N., Kalkum, M., and Chait, B.T. (2001) Automatic identiication of proteins with a MALDI-quadrupole ion trap mass spectrometer. Anal. Chem., 73, 5066–5077. Makaraov, A., Denisov, E., Lange, O., and Horning, S. (2006) Dynamic range of mass accuracy in LTQ orbitrap hybrid mass spectrometer. J. Am. Soc. Mass Spectrom., 17, 977–982. O’Connor, P.B., Mirgorodskaya, E., and Costello, C.E. (2002) High pressure matrix-assisted laser desorption/ ionization Fourier transform mass spectrometry for minimization of ganglioside fragmentation. J. Am. Soc. Mass Spectrom., 13, 402–407.

14 Ivleva, V.B., Elkin, Y.N., Budnik, B.A.,

15

16

17

18

19

20

21

22

Moyer, S.C., O’Connor, P.B., and Costello, C.E. (2004) Coupling thin layer chromatography with high pressure matrix assisted laser desorption/ ionization Fourier transform mass spectrometry for the analysis of ganglioside mixtures. Anal. Chem., 76, 6484–6491. Laiko, V.V., Moyer, S.C., and Cotter, R.J. (2000) Atmospheric pressure MALDI/ ion trap mass spectrometry. Anal. Chem., 72, 5239–5243. Laiko, V.V., Baldwin, M.A., and Burlingame, A.L. (2000) Atmospheric pressure matrix-assisted laser desorption/ionization mass spectrometry. Anal. Chem., 72, 652–657. Doroshenko, V.M., Laiko, V.V., Taranenko, N.I., Berkout, V.D., and Lee, H.S. (2002) Recent developments in atmospheric pressure MALDI mass spectrometry. Int. J. Mass Spectrom. Ion Processes, 221, 39–58. Kellersberger, K.A., Tan, P.V., Laiko, V.V., Doroshenko, V.M., and Fabris, D. (2004) Atmospheric pressure MALDIFourier transform mass spectrometry. Anal. Chem., 76, 3930–3934. Moyer, S.C., Marzilli, L.A., Woods, A.S., Laiko, V.V., Doroshenko, V.M., and Cotter, R.J. (2003) Atmospheric pressure matrix-assisted laser desorption/ ionization (AP MALDI) on a quadrupole ion trap mass spectrometer. Int. J. Mass Spectrom. Ion Processes, 226, 133–150. Moyer, S.G., and Cotter, R.J. (2002) Atmospheric pressure MALDI. Anal. Chem., 74, 468A–476A. Von Seggern, C.E., Zarek, P.E., and Cotter, R.J. (2003) Fragmentation of sialylated carbohydrates using infrared atmospheric pressure MALDI ion trap mass spectrometry from cation-doped liquid matrixes. Anal. Chem., 75, 6523–6530. Von Seggern, C.E., Moyer, S.C., and Cotter, R.J. (2003) Liquid infrared atmospheric pressure matrix-assisted laser desorption/ionization ion trap mass spectrometry of sialylated carbohydrates. Anal. Chem., 75, 3212–3218.

97

98

2 MALDI Mass Spectrometry Instrumentation 23 Kocher, T., Engstrom, A., and Zubarev,

24

25

26

27

28

29

30

31

32

R.A. (2005) Fragmentation of peptides in MALDI in- source decay mediated by hydrogen radicals. Anal. Chem., 77, 172–177. Brown, R.S., Carr, B.L., and Lennon, J.J. (1996) Factors that inluence the observed fast fragmentation of peptides in matrix- assisted laser desorption. J. Am. Soc. Mass Spectrom., 7, 225–232. Brown, R.S., Feng, J.H., and Reiber, D.C. (1997) Further studies of in-source fragmentation of peptides in matrixassisted laser desorption-ionization. Int. J. Mass Spectrom. Ion Processes, 169, 1–18 Roepstorff, P. and Fohlman, J. (1984) Proposal for a common nomenclature for sequence ions in mass spectra of peptides. Biomed. Mass Spectrom., 11, 601. Biemann, K. (1990) Sequencing of peptides by tandem mass spectrometry and high-energy collision-induced dissociation. Methods Enzymol., 193, 455–479. Budnik, B.A., Jebanathirajah, J., Ivleva, V.B., Costello, C.E., and O’Connor, P.B. (2004) Fragmentation methods of odd and even electron ions in HP MALDI FTMS. Proceedings, 52nd Annual Conference of the American Society for Mass Spectrometry, Nashville, TN. Budnik, B.A., Haselmann, K.F., and Zubarev, R.A. (2001) Electron detachment dissociation of peptide di-anions: an electron-hole recombination phenomenon. Chem. Phys. Lett., 342, 299–302. Sleno, L., Volmer, D.A., and Marshall, A.G. (2005) Assigning product ions from complex MS/MS spectra: the importance of mass uncertainty and resolving power. J. Am. Soc. Mass Spectrom., 16, 183–198. Wiley, W.C. and McLaren, I.H. (1955) Time-of-light mass spectrometer with improved resolution. Rev. Sci. Instrum., 29, 1150–1157. Juhasz, P., Roskey, M.T., Smirnov, I.P., Haff, L.A., Vestal, M.L., and Martin, S.A. (1996) Applications of delayed extraction matrix- assisted laser desorption ionization time- of-light mass

33

34

35

36

37

38

39

40

41

spectrometry to oligonucleotide analysis. Anal. Chem., 68, 941–946. Vestal, M.L., Juhasz, P., and Martin, S.A. (1995) Delayed extraction matrixassisted laser-desorption time-of-light mass-spectrometry. Rapid Commun. Mass Spectrom., 9, 1044–1050. Brown, R.S. and Lennon, J.J. (1995) Mass resolution improvement by incorporation of pulsed ion extraction in a matrix-assisted laser-desorption ionization linear time-of-light massspectrometer. Anal. Chem., 67, 1998–2003. Colby, S.M. and Reilly, J.P. (1996) Space-velocity correlation focusing. Anal. Chem., 68, 1419–1428. King, T.B., Colby, S.M., and Reilly, J.P. (1995) High- resolution MALDI-TOF mass-spectra of 3 proteins obtained using space-velocity correlation focusing. Int. J. Mass Spectrom. Ion Processes, 145, L1–L7. Juhasz, P., Vestal, M.L., and Martin, S.A. (1997) On the initial velocity of ions generated by matrix-assisted laser desorption ionization and its effect on the calibration of delayed extraction time-of-light mass spectra. J. Am. Soc. Mass Spectrom., 8, 209–217. Christian, N.P., Arnold, R.J., and Reilly, J.P. (2000) Improved calibration of time-of-light mass spectra by simplex optimization of electrostatic ion calculations. Anal. Chem., 72, 3327–3337. Gobom, J., Mueller, M., Egelhofer, V., Theiss, D., Lehrach, H., and Nordhoff, E. (2002) A calibration method that simpliies and improves accurate determination of peptide molecular masses by MALDI-TOF ms. Anal. Chem., 74, 3915–3923. Moskovets, E., Chen, H.S., Pashkova, A., Rejtar, T., Andreev, V., and Karger, B.L. (2003) Closely spaced external standard: a universal method of achieving 5 ppm mass accuracy over the entire MALDI plate in axial matrix-assisted laser desorption/ionization time-of-light mass spectrometry. Rapid Commun. Mass Spectrom., 17, 2177–2187. Mamyrin, B.A., Karataev, V.I., Schmikk, D.B., and Zagulin, B.A. (1973) The

References

42

43

44

45

46

47

48

49

50

mass-relectron, a new nonmagnetic time-of-light mass spectrometer with high resolution. Sov. Phys. JETP, 37, 45–48. Mamyrin, A. (1994) Laser-assisted relectron time-of-light massspectrometry. Int. J. Mass Spectrom. Ion Processes, 131, 1–19. Cotter, R.J., Iltchenko, S., and Wang, D.X. (2005) The curved-ield relectron: PSD and CID without scanning, stepping or lifting. Int. J. Mass Spectrom. Ion Processes, 240, 169–182. Cornish, T.J. and Cotter, R.J. (1993) A curved-ield relectron for improved energy focusing of product ions in time-of-light mass- spectrometry. Rapid Commun. Mass Spectrom., 7, 1037–1040. Berkenkamp, S. and Hillenkamp, F. (2005) Infrared MALDI with an orthogonal ion extraction TOF for high performance analysis of large proteins. Proceedings, 53rd American Society for Mass Spectrometry and Allied Topics, San Antonio, Texas, June 5–9, 2005. Kaufmann, R., Chaurand, P., Kirsch, D., and Spengler, B. (1996) Post-source decay and delayed extraction in matrix-assisted laser desorption/ ionization/relectron time-of-light mass spectrometry – are there trade-offs. Rapid Commun. Mass Spectrom., 10, 1199–1208. Karas, M., Gluckmann, M., and Schafer, J. (2000) Ionization in matrix-assisted laser desorption/ionization: singly charged molecular ions are the lucky survivors. J. Mass Spectrom., 35, 1–12. Reiber, D.C., Grover, T.A., and Brown, R.S. (1998) Identifying proteins using matrix- assisted laser desorption/ ionization in-source fragmentation data combined with database searching. Anal. Chem., 70, 673–683. Spengler, B. (1997) Post-source decay analysis in matrix-assisted laser desorption/ionization mass spectrometry of biomolecules. J. Mass Spectrom., 32, 1019–1036. Suckau, D., Resemann, A., Schuerenberg, M., Hufnagel, P., Franzen, J., and Holle, A. (2003) A novel MALDI lift-TOF/TOF mass

51

52

53

54

55 56

57

58

59

60

spectrometer for proteomics. Anal. Bioanal. Chem., 376, 952–965. Brock, A., Rodriguez, N., and Zare, R.N. (1998) Hadamard transform time of light mass spectrometry. Anal. Chem., 70, 3735–3741. Cotter, R.J., Gardner, B.D., Iltchenko, S., and English, R.D. (2004) Tandem time-of-light mass spectrometry with a curved ield relectron. Anal. Chem., 76, 1976–1981. Yergey, A.L., Coorssen, J.R., Backlund, P.S., Blank, P.S., Humphrey, G.A., Zimmerberg, J., Campbell, J.M., and Vestal, M.L. (2002) De novo sequencing of peptides using MALDI/ TOF-TOF. J. Am. Soc. Mass Spectrom., 13, 784–791. Campbell, J.M., Stein, S.E., Blank, P.S., Epstein, J., Vestal, M.L., and Yergey, A.L. (2005) Fundamentals of peptide fragmentation as a function of laser luence in a MALDI TOF-TOF mass spectrometer. Abstracts Papers Am. Chem. Soc., 229, U148–U148. Vestal, M.L. and Campbell, J.M. (2005) Biol. Mass Spectrom., 402, 79–108. Medzihradszky, K.F., Campbell, J.M., Baldwin, M.A., Falick, A.M., Juhasz, P., Vestal, M.L., and Burlingame, A.L. (2000) The characteristics of peptide collision- induced dissociation using a high- performance MALDI-TOF/TOF tandem mass spectrometer. Anal. Chem., 72, 552–558. Suckau, D., and Cornett, D.S. (1998) Protein sequencing by ISD and PSD MALDI-TOF ms. Analusis, 26, M18–M21. Shevchenko, A., Loboda, A., Ens, W., and Standing, K.G. (2000) Maldi quadrupole time-of-light mass spectrometry: a powerful tool for proteomic research. Anal. Chem., 72, 2132–2141. Verhaert, P., Uttenweiler-Joseph, S., de Vries, M., Loboda, A., Ens, W., and Standing, K.G. (2001) Matrix-assisted laser desorption/ionization quadrupole time-of-light mass spectrometry: an elegant tool for peptidomics. Proteomics, 1, 118–131. Kraus, H. (2002) Cryogenic detectors and their application to mass

99

100

2 MALDI Mass Spectrometry Instrumentation

61

62

63

64

65

66

67

68

69

70

spectrometry. Int. J. Mass Spectrom. Ion Processes, 215, 45–58. Uchaikin, S.V., Probst, F., and Seidel, W. (2001) Developing of a fast cryodetector read-out for mass spectrometry. Physica C, 350, 177– 179. Frank, M. (2000) Mass spectrometry with cryogenic detectors. Nucl. Instrum. Methods Phys. Res. A, 444, 375–384. Christ, P., Rutzinger, S., Seidel, W., Uchaikin, S., Probst, F., Koy, C., and Glocker, M.O. (2004) High detection sensitivity achieved with cryogenic detectors in combination with matrixassisted laser desorption/ionisation time-of-light mass spectrometry. Eur. J. Mass Spectrom., 10, 469–476. Frank, M., Mears, C.A., Labov, S.E., Benner, W.H., Horn, D., Jaklevic, J.M., and Barfknecht, A.T. (1996) Higheficiency detection of 66 000 Da protein molecules using a cryogenic detector in a matrix-assisted laser desorption/ ionization time-of-light mass spectrometer. Rapid Commun. Mass Spectrom., 10, 1946–1950. Ferguson, P.L. (2007) Agilent Acqiris High Speed Analog Signal Averager Improves Mass Accuracy, Dynamic Range in oaTOFMS Agilent Inc. Application Note http://www.stanford. edu/∼wonghoi/AP200/5989-7554EN.pdf Comisarow, M.B. and Marshall, A.G. (1974) Fourier transform ion cyclotron resonance spectroscopy. Chem. Phys. Lett., 25, 282–283. Marshall, A.G., Hendrickson, C.L., and Shi, S.D. (2002) Scaling ms plateaus with high-resolution FTICRMS. Anal. Chem., 74, 252A–259A. Amster, I.J. (1996) Fourier transform mass spectrometry. J. Mass Spectrom., 31, 1325–1337. Ledford, E.B., Jr, Rempel, D.L., and Gross, M.L. (1984) Mass calibration law for quadrupolar potential. Anal. Chem., 56, 2744–2748. Francl, T., Sherman, M.G., Hunter, R.L., Locke, M.J., Bowers, W.D., and McIver, R.T. (1983) Experimental determination of the effects of space charge on ion cyclotron resonance frequencies. Int. J. Mass Spectrom. Ion Processes, 54, 169–187.

71 Shi, S.D.H., Drader, J.J., Freitas, M.A.,

72

73

74

75

76 77

78

79

80

Hendrickson, C.L., and Marshall, A.G. (2000) Comparison and interconversion of the two most common frequency-tomass calibration functions for Fourier transform ion cyclotron resonance mass spectrometry. Int. J. Mass Spectrom. Ion Processes, 196, 591–598. Zhang, L.K., Rempel, D., Pramanik, B.N., and Gross, M.L. (2005) Accurate mass measurements by Fourier transform mass spectrometry. Mass Spectrom. Rev., 24, 286–309. O’Connor, P.B., Duursma, M.C., Rooij, G.J., van Heeren, R.M.A., and Boon, J.J. (1997) Correction of time of light shifted polymeric molecular weight distributions in MALDI-FTMS. Anal. Chem., 69, 2751–2755. Gerlich, D. (1992) in State-Selected and State-to-State Ion-Molecule Reaction Dynamics, vol. LXXXII (eds C.-Y. Ng and M. Baer), John Wiley & Sons, Inc., New York, pp. 1–176. Kaiser, N.K., Wu, S., Zhang, K., Prior, D.C., Bushback, M.A., Anderson, G.A., and Bruce, J.E. (2005) Development of a novel ion guide for improved ion transmission in FTICR mass spectrometry. Proceedings, 53rd American Society for Mass Spectrometry and Allied Topics, San Antonio, Texas, June 5–9, 2005. Caravatti, P. (1990) US Patent no. 4,924,089, issued 8 May, 1990. Shi, S.D.H., Drader, J.J., Hendrickson, C.L., and Marshall, A.G. (1999) Fourier transform ion cyclotron resonance mass spectrometry in a high homogeneity 25 Tesla resistive magnet. J. Am. Soc. Mass Spectrom., 10, 265–268. Guan, S., Wahl, M.C., and Marshall, A.G. (1993) Elimination of frequency drift from FTICR mass spectra by digital quadrature heterodyning: ultrahigh mass resolving power for laser-desorbed molecules. Anal. Chem., 65, 3647–3653. Easterling, M.L., Mize, T.H., and Amster, I.J. (1999) Routine part-permillion mass accuracy for high-mass ions: space-charge effects in MALDI FTICR. Anal. Chem., 71, 624–632. Nikolaev, E.V., Miluchihin, N.V., and Inoue, M. (1995) Evolution of an ion

References

81

82

83

84

85

86

87

88

89

cloud in a FTICRMS during signal detection: its inluence on spectral line shape and position. Int. J. Mass Spectrom. Ion Processes, 148, 145–157. Belov, M.E., Zhang, R., Strittmatter, E.F., Prior, D.C., Tang, K., and Smith, R.D. (2003) Automated gain control and internal calibration with external ion accumulation capillary liquid chromatography-electrospray ionizationFourier transform ion cyclotron resonance. Anal. Chem., 75, 4195–4205. Keller, B.O. and Li, L. (2001) Detection of 250 000 molecules of substance p by MALDI-TOF mass spectrometry and investigations into the fundamental limits of detection in MALDI. J. Am. Soc. Mass Spectrom., 12, 1055–1063. Moyer, S.C., Budnik, B.A., Pittman, J.L., Costello, C.E., and O’Connor, P.B. (2003) Attomole peptide analysis by high-pressure matrix-assisted laser desorption/ionization Fourier transform mass spectrometry. Anal. Chem., 75, 6449–6454. O’Connor, P.B., Costello, C.A., and Earle, W.E. (2002) A high voltage rf oscillator for driving multipole ion guides. J. Am. Soc. Mass Spectrom., 13, 1370–1375. Jaber, A.J., Kaufman, J., Liyanage, R., Akhmetova, E., Marney, S., and Wilkins, C.L. (2005) Trapping of wide range mass-to-charge ions and dependence on matrix amount in internal source maldi-ftms. J. Am. Soc. Mass Spectrom., 16, 1772–1780. Marshall, A.G., Wang, T.-C.L., and Ricca, T.L. (1985) Tailored excitation for Fourier transform ion cyclotron resonance mass spectrometry. J. Am. Chem. Soc., 107, 7893–7897. Lee, S.A., Jiao, C.Q., Huang, Y., and Freiser, B.S. (1993) Multiple excitation collisional activation in FTMS. Rapid Commun. Mass Spectrom., 7, 819–821. Boering, K.A., Rolfe, J., and Brauman, J.I. (1992) Control of ion kinetic energy in ion cyclotron resonance spectrometry: very-low-energy collision-induced dissociation. Rapid Commun. Mass Spectrom., 6, 303–305. Gauthier, J.W., Trautman, T.R., and Jacobson, D.B. (1991) Sustained off-resonance irradiation for CAD

90

91

92

93

94

95

96

97

98

involving FTMS. CAD technique that emulates infrared multiphoton dissociation. Anal. Chim. Acta, 246, 211–225. Mirgorodskaya, E., O’Connor, P.B., and Costello, C.E. (2002) A general method for precalculation of parameters for sustained off resonance irradiation/ collision-induced dissociation. J. Am. Soc. Mass Spectrom., 13, 318–324. Little, D.P., Speir, J.P., Senko, M.W., O’Connor, P.B., and McLafferty, F.W. (1994) Infrared multiphoton dissociation of large multiply-charged ions for biomolecule sequencing. Anal. Chem., 66, 2809–2815. Williams, E.R. and McLafferty, F.W. (1990) 193-nm laser photoionization and photodissociation for isomer differentiation in Fourier-transform mass-spectrometry. J. Am. Soc. Mass Spectrom., 1, 361–365. Zubarev, R.A., Kelleher, N.L., and McLafferty, F.W. (1998) Electron capture dissociation of multiply charged protein cations – a nonergodic process. J. Am. Chem. Soc., 120, 3265–3266. Zubarev, R.A., Kruger, N.A., Fridriksson, E.K., Lewis, M.A., Horn, D.M., Carpenter, B.K., and McLafferty, F.W. (1999) Electron capture dissociation of gaseous multiply-charged proteins is favored at disulide bonds and other sites of high hydrogen atom afinity. J. Am. Chem. Soc., 121, 2857–2862. Coon, J.J., Syka, J.E.P., Schwartz, J.C., Shabanowitz, J., and Hunt, D.F. (2004) Anion dependence in the partitioning between proton and electron transfer in ion/ion reactions. Int. J. Mass Spectrom., 236, 33–42. Syka, J.E.P., Coon, J.J., Schroeder, M.J., Shabanowitz, J., and Hunt, D.F. (2004) Peptide and protein sequence analysis by electron transfer dissociation mass spectrometry. Proc. Natl Acad. Sci. USA, 101, 9528–9533. Krutchinsky A.N. and Chait, B.T. (2002) On the nature of the chemical noise in MALDI Mass Spectra. J. Am. Soc. Mass Spectrom., 13, 129–134. Garrett, T.J. and Yost, R.A. (2006) Analysis of intact tissue by

101

102

2 MALDI Mass Spectrometry Instrumentation

99

100

101

102

103

104

105

106

intermediate-pressure MALDI on a linear ion trap mass spectrometer. Anal. Chem., 78, 2465–2469. Schwartz, J.C., Syka, J.E.P., and Quarmby, S.T. (2005) Improving the fundamentals of msn on 2d linear ion traps: New ion activation and isolation techniques. Proceedings, 53rd Annual Conference of the American Society for Mass Spectrometry and Allied Topics, San Antonio, Texas, June 5–9, 2005. Baba, T., Hashimoto, Y., Hasegawa, H., Hirabayashi, A., and Waki, I. (2004) Electron capture dissociation in a radio frequency ion trap. Anal. Chem., 76, 4263–4266. Silivra, O.A., Ivonin, I.A., Kjeldsen, F., and Zubarev, R.A. (2004) Ion-electron reactions in a quadrupole ion trap: Realization, irst results and prospects. Proceedings, 52nd Annual Conference of the American Society for Mass Spectrometry, Nashville, TN. Kim, T.Y., Thompson, M.S., and Reilly, J.P. (2005) Peptide photodissociation at 157 nm in a linear ion trap mass spectrometer. Rapid Commun. Mass Spectrom., 19, 1657–1665. Payne, A.H. and Glish, G.L. (2001) Thermally assisted infrared multiphoton photodissociation in a quadrupole ion trap. Anal. Chem., 73, 3542–3548. Cui, W.D., Hadas, B., Cao, B.P., and Lifshitz, C. (2000) Time-resolved photodissociation (TRPD) of the naphthalene and azulene cations in an ion trap/relectron. J. Phys. Chem. A, 104, 7160–7160. Creaser, C.S. and O’Neill, K.E. (1997) Photodissociation and collisionally activated dissociation tandem mass spectrometric studies of diluoro[triazol1-ylmethyl]benzhydrols and related compounds in a quadruple ion trap. Int. J. Mass Spectrom. Ion Processes, 165, 13–23. Colorado, A., Shen, J.X.X., Vartanian, V.H., and Brodbelt, J. (1996) Use of infrared multiphoton photodissociation with swift for electrospray ionization and laser desorption applications in a quadrupole ion trap mass spectrometer. Anal. Chem., 68, 4033–4043.

107 Welling, M., Thompson, R.I., and

108

109

110

111

112

113

114

115

Walther, H. (1996) Photodissociation of MgC60+ complexes generated and stored in a linear ion trap. Chem. Phys. Lett., 253, 37–42. Stephenson, J.L., Booth, M.M., Boue, S.M., Eyler, J.R., and Yost, R.A. (1996) Analysis of biomolecules using electrospray ionization–ion-trap mass spectrometry and laser photodissociation. Biochem. Biotechnol. Appl. Electrospray Ionization Mass Spectrom., 619, 512–564. Williams, J.D., Cooks, R.G., Syka, J.E.P., Hemberger, P.H., and Nogar, N.S. (1993) Determination of positions, velocities, and kinetic energies of resonantly excited ions in the quadrupole ion-trap mass-spectrometer by laser photodissociation. J. Am. Soc. Mass Spectrom., 4, 792–797. Louris, J.N., Brodbelt, J.S., and Cooks, R.G. (1987) Photodissociation in a quadrupole ion trap mass-spectrometer using a iber optic interface. Int. J. Mass Spectrom. Ion Processes, 75, 345–352. Loboda, A., Krutchinsky, A., Loboda, O., McNabb, J., Spicer, V., Ens, W., and Standing, K. (2000) Novel LINAC II electrode geometry for creating an axial ield in a multipole ion guide. Eur. J. Mass Spectrom., 6, 531–536. Diehnelt, C.W., Peterman, S.M., and Budde, W.L. (2005) Liquid chromatography-tandem mass spectrometry and accurate m/z measurements of cyclic peptide cyanobacteria toxins. Trends Anal. Chem., 24, 622–635. Hongli, L., Giles, K., Bendiak, B., Kaplan, K., Siems, W.F., and Hill, H.H. (2012) Resolving structural of monosaccharide methyl glycosides using drift tube and traveling wave ion mobility mass spectrometry. Anal. Chem., 84, 3231–3239. Dwivedi, P., Ching, W., Matz, L.M., Clowers, B.H., Siems, W.F., and Hill, H. (2006) Gas-phase chiral separations by ion mobility mass spectrometry. Anal. Chem., 78, 8200–8206. Woods, A., Ugarov, M., Egen, T., Koomen, J., Gillig, K.J., Fuhrer, K., Gonin, M., and Schultz, J.A. (2004)

References

116

117

118

119

120

121

122

123

124

125

Lipid/peptide/nucleotide separation with MALDI-ion mobility TOF-MS. Anal. Chem., 76, 2187–2195. Fenn, L.S., Kliman, M., Mahsut, A., Zhao, S.R., and McLean, J.A. (2009) Characterizing ion mobility-mass spectrometry conformation space for the analysis of complex biological samples. Anal. Bioanal. Chem., 294, 235–244. Giles, K., Pringle, S.D., Worthington, K.R., Little, D., Wildgoose, J.L., and Bateman, R.H. (2004) Applications of a travelling wave-based radio-frequencyonly stacked ring ion guide. Rapid Commun. Mass Spectrom., 18, 2401–2414. Kiss, A. and Heeren, R.M.A. (2011) Size, weight and position: ion mobility spectrometry and imaging MS combined. Anal. Bioanal. Chem., 399, 2623–2634. Bohrer, B.C., Merenbloom, S.I., Koeniger, S.L., Hilderbrand, A.E., and Clemmer, D.E. (2008) Biomolecule analysis by ion mobility spectrometry. Annu. Rev. Anal. Chem., 1, 293–327. Ridenour, W.B., Kliman, M., McLean, J.A., and Caprioli, R.M. (2010) Structural characterization of phospholipids and peptides directly from tissue sections by MALDI traveling-wave ion mobilitymass spectrometry. Anal. Chem., 82, 1881–1889. Guevremont, R. (2004) High-ield asymmetric waveform ion mobility spectrometry: a new tool for mass spectrometry. J. Chromatogr. A, 1058, 3–19. Makarov, A. (2000) Electrostatic axially harmonic orbital trapping: a highperformance technique of mass analysis. Anal. Chem., 72, 1156–1162. Strupat, K., Kovtourn, V., Bui, H., Viner, R., Stafford, G., and Horning, S. (2009) MALDI produced ions inspected with a linear ion trap-orbitrap hybrid analyzer. J. Am. Soc. Mass Spectrom., 20, 1451–1463. Koestler, M., Kirsch, D., Hester, A., Leisner, A., Guenther, S., and Spengler, B. (2008) Rapid Commun. Mass Spectrom., 22, 3275–3285. Guenther, S., Koestler, M., Schulz, O., and Spengler, B. (2010) Int. J. Mass Spectrom., 294, 7–15.

126 Dreisewerd, K., Muthing, J., Rohling,

127

128

129

130

131

132

A., Meisen, I., Vukelic, Z., PeterKatalinic, J., Hillenkamp, F., and Berkenkamp, S. (2005) Analysis of gangliosides directly from thin-layer chromatography plates by infrared matrix-assisted laser desorption/ ionization orthogonal time-of-light mass spectrometry with a glycerol matrix. Anal. Chem., 77, 4098–4107. Guittard, J., Hronowski, X.P.L., and Costello, C.E. (1999) Direct matrixassisted laser desorption/ionization mass spectrometric analysis of glycosphingolipids on thin layer chromatographic plates and transfer membranes. Rapid Commun. Mass Spectrom., 13, 1838–1849. Dreisewerd, K., Kolbl, S., Peter-Katalinic, J., Berkenkamp, S., and Pohlentz, G. (2006) Analysis of native milk oligosaccharides directly from thin-layer chromatography plates by matrixassisted laser desorption/ionization orthogonal-time-of- light mass spectrometry with a glycerol matrix. J. Am. Soc. Mass Spectrom., 17, 139–150. Gusev, A.I., Proctor, A., Rabinovich, Y.I., and Hercules, D.M. (1995) Thin-layer chromatography combined with matrix- assisted laser desorption ionization mass spectrometry. Anal. Chem., 67, 1805–1814. Gusev, A.I., Vasseur, O.J., Proctor, A., Sharkey, A.G., and Hercules, D.M. (1995) Imaging of thin-layer chromatograms using matrix/assisted laser desorption/ ionization mass spectrometry. Anal. Chem., 67, 4565–4570. Dreisewerd, K., Berkenkamp, S., Leisner, A., Rohling, A., and Menzel, C. (2003) Fundamentals of matrix-assisted laser desorption/ionization mass spectrometry with pulsed infrared lasers. Int. J. Mass Spectrom. Ion Processes, 226, 189–209. Menzel, C., Dreisewerd, K., Berkenkamp, S., and Hillenkamp, F. (2002) The role of the laser pulse duration in infrared matrix-assisted laser desorption/ionization mass spectrometry. J. Am. Soc. Mass Spectrom., 13, 975–984.

103

104

2 MALDI Mass Spectrometry Instrumentation 133 Menzel, C., Dreisewerd, K.,

135 Berkenkamp, S., Menzel, C., Karas, M.,

Berkenkamp, S., and Hillenkamp, F. (2001) Mechanisms of energy deposition in infrared matrix-assisted laser desorption/ionization mass spectrometry. Int. J. Mass Spectrom. Ion Processes, 207, 73–96. 134 Menzel, C., Berkenkamp, S., and Hillenkamp, F. (1999) Infrared matrix-assisted laser desorption/ ionization mass spectrometry with a transversely excited atmospheric pressure carbon dioxide laser at 10.6 μm wavelength with static and delayed ion extraction. Rapid Commun. Mass Spectrom., 13, 26–32.

and Hillenkamp, F. (1997) Performance of infrared matrix-assisted laser desorption/ionization mass spectrometry with lasers emitting in the 3 μm wavelength range. Rapid Commun. Mass Spectrom., 11, 1399–1406. 136 Berkenkamp, S., Karas, M., and Hillenkamp, F. (1996) Ice as a matrix for IR-matrix-assisted laser desorption/ ionization: mass spectra from a protein single crystal. Proc. Natl. Acad. Sci. USA, 93, 7003–7007. 137 Cooley, J.W. and Tukey, J.W. (1965) Fast Fourier transform algorithm. Math. Comput., 19, 297–301.

105

3 MALDI-MS in Protein Chemistry and Proteomics Karin Hjernø and Ole N. Jensen

3.1 Introduction

Proteins are a major constituent of living cells, and mediate the majority of biological processes in all organisms, from microbes to mammals. During the past six decades, research into protein biochemistry and molecular cell biology has led to the elucidation of the functions and structures of many proteins. However, the overall architecture and dynamics of subcellular structures and the mechanisms of molecular signaling networks remain to be established. Proteins have been – and continue to be – explored for use in biotechnology, pharmacology, and biomedical applications. Consequently, a range of protein-based drugs are now available commercially, and many proteins are targets for small-molecule drugs. Recombinant enzymes are widely used for catalysis in bioprocessing, ranging from the production of food and food ingredients to the manufacture of textiles and the creation of biofuels. In addition, a large part of the protein research today is focused towards biomarker discovery and disease diagnostics (see Chapters 4 and 10). Thus, protein research is at center stage in many cell biology and biotechnology laboratories, both in academia and in industry. Within the cell, proteins are synthesized in a multistep process which includes the transcription of DNA into RNA, the processing of RNA into mature mRNA and, inally, translation of the mRNA into protein. Each of these steps is prone to molecular events that lead to alterations of the protein product (Figure 3.1). For example, the gene may contain base substitutions and mutations that eventually alter the amino acid sequence of the protein. Likewise, the processing and maturation of RNA may lead to the elimination of distinct exons, thereby creating variant gene products, the so-called splice variants. Both, co- and post-translational modiication of the polypeptide backbone lead to further diversity and heterogeneity of the gene products – that is, the proteins. Many proteins are substrates for dynamic reversible modiications (e.g., phosphorylation and acetylation) that regulate their biological activity and interactions, depending on environmental cues and the metabolic status of the cell. It is estimated that, on average, each eukaryotic gene is translated into at least 10 different modiied forms. Thus, the elucidation of MALDI MS: A Practical Guide to Instrumentation, Methods, and Applications, Second Edition. Edited by Franz Hillenkamp and Jasna Peter-Katalinic. © 2014 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2014 by Wiley-VCH Verlag GmbH & Co. KGaA.

106

3 MALDI-MS in Protein Chemistry and Proteomics

Figure 3.1 A range of events may inluence

protein structure and integrity. Molecular events at the DNA and RNA levels generate heterogeneity at the protein level. Co- and post-translational modiications and processing leads to the heterogeneity of proteins; analytical artifacts during protein

extraction, puriication and analysis produces further heterogeneity. Mass spectrometry can detect and locate chemical changes to proteins via determination of the accurate molecular mass of protein and the derived peptides.

protein primary structure and post-translational modiications is crucial to studies of biological processes. Modern analytical technologies are applied to elucidate the intricate protein machinery of the cell and to discover and develop new protein-based biotechnology products. Proteomic studies – that is, the systematic analysis of proteins from cells, tissues or whole organisms – calls for highly eficient and sensitive analytical methods for the identiication, characterization, and quantiication of proteins. Today, mass spectrometry (MS) is the most useful analytical technique for protein and proteome analysis, as it provides a relatively simple platform for determining one of the fundamental properties of biological molecules, namely the molecular mass. In this respect several strategies can be taken, depending on the complexity of the sample and the information sought. The basic steps in such MS-based strategies for protein analysis, which are summarized in Figure 3.2, include: (i) the biochemical characterization of puriied proteins; and (ii) two parallel strategies for the analysis of proteomes or subproteomes – that is, the multitude of proteins within a given cell or cell organelle at any one time. The irst approach (Figure 3.2a) is very useful for the validation of native or recombinant proteins destined for structure analysis by nuclear magnetic resonance (NMR) or X-ray diffraction, as well as in the quality control of proteins for biotechnological applications. The method is based on an ability to obtain pure protein in soluble form, and in a solvent that is suitable for matrix-assisted laser desorption/ionization (MALDI) -MS analysis. In most cases, the identity of the protein is known à priori, and the amino acid sequence is available either from protein databases or by translation from the corresponding gene sequences. The mass determination of a puriied intact protein will reveal any discrepancies from the molecular mass of a “naked” protein – that is, the calculated molecular mass as determined by the amino acid sequence predicted from the cognate gene. Similarly, MS analysis of peptides derived from a protein by enzymatic digestion

3.1 Introduction

Figure 3.2 MALDI-MS is a very powerful tool

in many types of analysis involving proteins, from in-depth analyses of individual proteins to studies of large complex protein samples, as in proteomic studies. Here, the outlines of two different strategies are shown: (a) the analysis of a puriied protein for which the sequence is known; and (b) the analysis of all proteins in, for example, a given cell (i.e.,

the proteome of the cell). Both, the LC- and gel-based strategies are shown, and these can be combined with quantiication strategies and studies of modiied proteins. As can be seen, many of the steps involved, such as protein digestion, peptide mass mapping and peptide fragmentation, are identical and used in all of the strategies shown.

will not only conirm the amino acid composition of the individual peptides but also reveal the presence of any chemical modiications by a mass increment or a mass deicit relative to the expected masses of the unmodiied peptides. This type of analysis is referred to as “peptide mass mapping.” When used for proteomics analysis and the analysis of large protein complexes (Figure 3.2b), MS must be combined with protein or peptide separation strategies, in order to deal with the high complexity of such samples. These are typically one-dimensional (1-D) or two-dimensional (2-D) gel electrophoresis or liquid chromatography (LC) approaches. The individual proteins/peptides are then identiied by searching of the obtained MS-data against in-silico-digested proteins from databases of known or predicted sequences. Correlations between the experimental and theoretical data are used for scoring of the proteins and thereby determining the identity of the proteins in the sample being analyzed. Further details on the different approaches available to achieve such identity are provided later in the chapter. Before being analyzed by MS, soluble biomolecules must be converted into gasphase ions. Today, MALDI and electrospray ionization (ESI) are the two main

107

108

3 MALDI-MS in Protein Chemistry and Proteomics

techniques used to produce peptide and protein ions for MS analysis. ESI mainly produces multiply protonated peptide ions [M + nH]n+, whereas MALDI generates mainly singly protonated peptide ions [M + H+], although multiply charged species are sometimes observed. Proteins may generate both singly and multiply charged ions in MALDI. Tandem mass spectrometry (MS/MS) is very useful for the amino acid sequencing of peptides, and has been used widely in both protein biochemistry and proteomics to identify proteins, to deduce the sequence of a peptide, and to detect and locate post-translational modiications. Until around a decade ago, the concept of amino acid sequencing by MS-technologies was synonymous with ESI-MS/MS, but today MALDI-MS/MS techniques are implemented in high-performance instruments such that the quality of MALDI tandem mass spectra is comparable with that of ESI-MS/MS spectra. Currently, MALDI tandem mass spectrometers exist in a number of geometries, including TOF-TOF, Q-TOF, ion trap and orbitrap analyzers that each provide unique analytical features for the sequencing of peptides and proteins by MS/MS (details of the instrumentation for different types of MS/MS are provided in Chapter 2). Here, we will describe a range of applications of MALDI-MS, from the concepts of in-depth analysis of puriied proteins to applications of MALDI-MS in a broader, proteomics-based research where proteins are identiied, characterized, and quantiied. In addition, issues of sample preparation, protein characterization and identiication strategies and bioinformatic tools for data interpretation will be discussed. The concepts of peptide fragmentation, sequencing and derivatization, analysis of post-translational modiications and the clinical applications of MALDIMS are also briely outlined.

3.2 Sample Preparation for Protein and Peptide Analysis by MALDI-MS

In this section, some of the general sample preparation issues that inluence protein analysis by MALDI-MS experiments are described. Often, the intact proteins need to be puriied to near-homogeneity using biochemical or immunological methods, for example, chromatography, afinity puriication or immunoprecipitation prior to MS analysis. Volatile buffers, such as ammonium bicarbonate and ammonium acetate should be used for the inal stages of puriication of proteins, if possible. The pH should be adjusted to 10 000 Da) the fragmentation degrades the resolving power in relectron TOF-MS, and linear TOF-MS becomes the preferred instrument coniguration. Linear TOFs can achieve a mass resolving power of 500–1000 for oligonucleotides above 10 000 Da, and of 3000 for smaller oligonucleotides.

173

174

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

Figure 5.3 Sections of UV MALDI mass

spectra of 70, 80, and 90mer DNA strands desorbed from 3-3-hydroxypicolinic acid as a matrix and analyzed in an orthogonal TOF. The MALDI ions irst pass a quadrupole before they enter a relectron TOF section. Independent of the ion source used and the mass of the molecules, the O-TOF analyzer provides a constant instrumental resolution throughout the whole mass range rather than a mass-dependent resolution as in axial MALDI-TOF. The indicated full width at

half-maximum (FWHM) values essentially relect the envelope to the isotopic distribution. Note the peak broadening with increasing mass on the right slope of the protonated species due to adduct formation. For larger nucleic acids it becomes increasingly dificult to remove or prevent the attachment of small molecules to the highly negatively charged backbone, and the cleaning steps prior to analysis become crucial.

into a relectron-TOF placed in the orthogonal direction (this instrument coniguration is described in more detail in Chapter 2, Section 2.4.4). As can be seen in Figure 5.3, hybrid instruments with orthogonal extraction can enhance the analytical performance for DNA samples in terms of both mass resolution and upper mass limits. Even with these improvements, however, the practical limit achieved by UV-MALDI remains around 30 kDa (∼100 nt). For larger analytes, performance degrades rapidly due to the mounting challenges of controlling adduct formation and gas-phase fragmentation. The utilization of infrared (IR) lasers provides the means to limit ion fragmentation, due to their considerably softer desorption. Indeed, it has been shown that the fragmentation of DNA ions is greatly reduced when desorption is performed at wavelengths of 2.94 or 10.6 μm, using glycerol or succinic acid as matrices [22, 23]. For example, these conditions have enabled the detection of restriction enzyme products of up to 1400 nt single-stranded DNA (Figure 5.4) [24, 25]. Unfortunately, these combinations of desorption wavelengths and matrix selection also lead to excessive adduct formation. In this example, extensive glycerol adduction limited the mass resolution to about 50 and prevented the accurate determination of the exact mass of the DNA free acid.

5.2 Genetic Markers

Figure 5.4 IR-MALDI mass spectrum of

large DNA analyzed from glycerol as a matrix. The spectrum depicts single-stranded DNA molecules of length up to 1.4 kb generated from restriction-digested plasmid DNA puriied by ethanol precipitation. The

achieved mass resolution is limited to about 50 for DNA molecules of this size, caused by excessive adduct formation of glycerol and its desorption products. The demonstrated mass accuracy of 1% and sensitivity in the low femtomole range is remarkable.

If some of the physico-chemical properties of nucleic acids pose many challenges to their analysis by MALDI-TOF, then some others afford intrinsic advantages as compared to proteins/peptides and other biomolecules. For instance, their primary structure is much more homogeneous than that of proteins, consisting of only four relatively similar building blocks. Because of this structural simplicity and homogeneity, relative and even absolute (with an internal standard) quantiication can be readily accomplished and this constitutes the basis for a number of assays (as discussed below). The remainder of this chapter is mostly devoted to descriptions of the different assays that have been developed for the analysis of NAs, under the somewhat restrictive boundary conditions of UV-MALDI-TOF-MS.

5.2 Genetic Markers

During the past few decades, intensive worldwide efforts have been dedicated to establishing a reference sequence of the human genome, a draft sequence of which was announced in 2000 [26, 27]. By September 2002, many of the remaining gaps had been illed and the sequence was further reined such that over 90% of the sequence could be presented in its inal form with an accuracy of greater than 99.99%. Meanwhile, the genomes of a substantial number of other species have been reported, as a result of which the emphasis has shifted from de novo sequencing to the analysis of differences, for example between individuals, ethnic groups and diseased versus healthy populations. It is in this arena rather than in de novo sequencing that MALDI MS has found multiple applications.

175

176

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

In this sense the completion of the Human Genome Project constitutes a major scientiic milestone, but it must also be seen as a starting point for the “real” exploration of the human genome. Currently, there is still a rather “poor” understanding of the architecture, functionalities and regulation of the human genome, and although a complete DNA sequence is almost available, debate persist regarding the total number of genes involved. Recent estimates have pointed to approximately 30 000 genes, a number that is much smaller than was originally estimated. The “variety” and complexity seen at the mRNA and protein levels is much higher than the number of genes might suggest. One of the processes contributing to the estimated number of well over 100 000 proteins arising from the rather low number of genes is alternative splicing. Among the processes not yet fully explored are epigenetic modiications such as methylation and histone acetylation, as well as RNA editing and certain forms of somatic recombination. Aside from questions focusing on how genomic information translates into the mRNA and protein levels, another important aspect remains of whether there is something like “a reference sequence” for a species, and how variable are human genomes? Currently, the genetic diversity among humans is estimated to be about 0.1%, such that two unrelated persons will share about 99.9% of their DNA sequence. Given a genome size of three billion bases, this means that they differ in millions of bases. The elucidation of this intra-species genetic diversity remains an important and still formidable task. This is important, as the assessment of genetic variability will provide insight into genotype–phenotype interaction and thus will help in an understanding of how genes contribute to diseases, in disease predisposition, and also perhaps how an individual will respond to a drug. This is clearly a formidable task, since cataloging these genotypic variations requires the analysis of large numbers of individuals, either for a large number of known genetic markers or, preferably, for their complete individual genome sequences. Genetic diversity within the human genome manifests itself mostly as single nucleotide polymorphisms (SNPs), restriction fragment length polymorphisms (RFLPs, forming a subset of all SNPs), short tandem repeats (STRs), and “random” base changes. The use of genome-scanning technologies has also uncovered another type of genotypic variation that is mainly characterized by larger genomic deletions, duplications and copy number variations [28, 29]. During recent years, substantial efforts have been focused on the development of suitable technologies, which enable the cost-eficient, large-scale assessment of these markers in multiple populations. Among these technologies, MS has shown great promise for an ability to sustain the high-throughput analysis of genetic markers. Naturally, most of the methods in which MS is applied have focused on the development of assay formats suited to the analysis of these markers. To help understand the focus of most research groups developing tools for DNA analysis by MS, and the strategy upon which they might embark, the nature and importance of the most common genetic markers – and their assays – are described in greater detail in the following sections.

5.2 Genetic Markers

5.2.1 Restriction Fragment Length Polymorphisms (RFLPs)

Restriction endonucleases are enzymes that recognize speciic DNA sequences in either a genome or a given stretch of DNA, and cleave the DNA in a predeined manner. Consequently, changes in a given DNA sequence may create new such cleavage sites or abolish existing ones. Restriction enzymes were increasingly used for genome analysis during the late 1970s, at which time the visualization of restriction cleavage patterns was achieved by Southern blotting of the cleavage products and hybridization with labeled probes. When the restriction patterns of different individuals were compared, it became apparent that the nucleotide variation rate between individuals was higher than expected. DNA-based variations at restriction sites, which alters the restriction fragment length pattern, were accordingly termed RFLPs [30]. With the knowledge available today, RFLPs can be deined as a subset of SNPs located in the recognition sequence for the restriction site or the cleaved sequence. RFLPs are of a biallelic nature, which means that usually only two (of the four possible) nucleotides occur in a given population. The allele-status renders the DNA permissive to cleavage or not, and thus provides a “yes” or “no” answer, which can be used as a simple marker to follow a genetic trait. However, isolated RFLPs alone are not very informative and their abundance is too low to allow in-depth genetic studies. Although RFLPs have demonstrated their merits as a valuable tool in genome analysis and genetics, they have since been superseded by either more informative or more abundant markers. An additional problem is that the process of Southern blotting is cumbersome and dificult to automate. 5.2.2 Microsatellites/Short Tandem Repeats (STRs)

Microsatellites are stretches of DNA that consist of repeating units of two, three, or four nucleotides (also referred to as di-, tri-, or tetranucleotide repeats). The length of such a microsatellite, which is based on the number of repeated units, can vary between the homologs of each chromosome of an individual and also between individuals, thus constituting a polymorphism [31]. The different repeat lengths of the microsatellite constitute the alleles of the polymorphism. Usually, multiple alleles are observed for any particular microsatellite in a given population, and thus the information content of this type of marker is much higher than for biallelic markers; indeed, this is one of the features that has made them useful for genetic linkage studies. The probability of identifying heterozygous individuals for a selected marker is very important in linkage studies and in forensic applications. In the past, tens of thousands of microsatellites were identiied and a portfolio with genome-wide coverage was developed for genetic linkage studies [32–34]. The analysis of microsatellites can be enabled by employing rather simple polymerase chain reaction (PCR) ampliication techniques. Microsatellite marker analysis does

177

178

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

not require Southern blotting, and thus offers a reasonable potential for automation, and the separation and detection of the PCR products is usually performed using gel or capillary electrophoresis [35, 36]. However, the screening of thousands of individuals for tens of thousands of microsatellite alleles has remained challenging and cumbersome. However, within recent years much of the focus has shifted away from microsatellites, and these are now mainly employed in the ield of identiication (forensics), where less informative markers are not easily applicable, or in the analysis of genomes of other organisms, where other types of markers are (as yet) scarce. 5.2.3 Single Nucleotide Polymorphisms (SNPs)

SNPs are usually deined as single base-pair positions in genomic DNA for which two sequence alternatives – termed “alleles” – exist, and the frequency of the least abundant allele is greater than 1%. Usually, SNPs are biallelic; moreover, about two-thirds of all SNPs are C/T sequence variations, while the inal third is shared among the other three types (C/A, C/G, and T/A) with comparable frequencies [37–39]. Sequence changes occurring with a frequency less than 1% are usually referred to as “mutations.” SNPs are the genetic marker with the highest abundance in the human genome, occurring with a frequency of about one in every 1000 bp. Given a human genome size of three billion base pairs, the genome of individuals will differ in millions of bases. Most of these SNPs will not reside in coding regions, and among the coding SNPs a fraction will be non-synonymous. Interestingly, about one-half of all coding SNPs identiied to date are non-synonymous, and thus lead to an amino acid exchange. The combinations of SNPs in regulatory regions as well as amino acid-coding SNPs have a substantial inluence on inter-individual differences in expression, and in the proteome. The high abundance of SNPs and the ease of developing assays for their analysis has recently attracted much attention from the scientiic community. The importance of SNPs to an understanding of genetic diversity and their value in establishing genotype-to-phenotype correlations has been recognized by many academic and commercial research entities [40, 41]. This has led to the foundation and implementation of large-scale SNP Discovery programs, among which The SNP Consortium (TSC) has received the most public attention. The TSC is a collaboration between 13 multinational companies and prominent academic institutions. The initial goal of the TSC was the delivery of 300 000 evenly spaced SNPs to the public domain by the end of 2001. The human diversity program of the Human Genome Project (HGP) is another program providing additional SNP markers. Currently, the details of about four million SNPs have been reported in dbSNP (http://www.ncbi.nlm. nih.gov/SNP/), a public database of the National Center for Biotechnology information (NCBI), and thus the initial goals have been exceeded by far. The tremendous effort associated with the generation of a SNP catalog has furthered technology development on a broad scale. Eficient methods for the

5.2 Genetic Markers

discovery of SNPs were developed, among which indirect methods such as the in-silico mining of existing data for polymorphisms through sequence overlays were employed. Direct methods have also been developed, which include for example reduced representative shotgun sequencing [42] (the method primarily used by the TSC), denaturing gradient high-performance liquid chromatography (HPLC) [43], single-strand conformation polymorphism analysis (SSCP) [44, 45], re-sequencing hybridization chips [46, 47], and direct DNA sequencing [48]. The discovery phase is usually followed by an experimental veriication of these SNPs, and an assessment of their allelic frequencies in various populations. This step employs a different set of technologies, which were developed speciically for SNP analysis. Today, MS plays an important role in the cataloging of SNPs and in assessing allele frequency. The assay formats developed for this purpose are discussed in detail in the next sections. With a catalog of almost four million SNPs available, the scientiic community has now entered into the next large-scale project, namely the HapMap, the focus of which is the identiication and characterization of haplotype blocks in the human genome. (Haplotypes are deined as groups of closely linked genetic markers that tend to be inherited together and are rarely separated by recombination events [49].) Having large amounts of SNPs available is a irst step, but the next step is to elucidate how these SNPs are grouped together and how these groups (blocks) are inherited. So, the main questions are: (i) How are SNPs used in genome research?; and (ii) How would knowledge of the haplotype groups advance genome research? The majority of SNPs do not reside in coding regions, but by virtue of their abundance they still represent the prominent percentage of SNP sets used in genome-wide association and linkage disequilibrium studies. Both are geared towards the elucidation of the polygenic origins of common diseases. The basis of this is a comparison of SNP allele frequencies among groups of affected and unaffected individuals. Statistically signiicant frequency changes are thought to be associated with phenotypic differences between groups, while the associated SNPs pinpoint to the genomic region within which the “causative” genomic polymorphism or mutation of interest resides. SNP-based genome-wide studies have not only been suggested as methods to identify the genetic underpinning of common diseases, but they may also allow for the elucidation of individual responses to drug treatment, which represents another important piece of the puzzle towards individualized medicine. Genome-wide linkage and association studies require the analysis of hundreds of thousands to millions of SNPs in thousands of individuals or DNA samples. This is a daunting task and there are several technical ways to go about this, as will be discussed in subsequent sections of this chapter. Another consideration is the HapMap. If haplotype groups can be described for the human genome, and each haplotype group can be characterized by a subset of “tagSNPs,” then the overall number of SNPs required for genome-wide analysis can be reduced drastically.

179

180

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

5.2.4 Characterization of Base Modiications and Covalent Adducts

The unique feature that sets MS apart from the other platforms for NA analysis is its ability to recognize covalent modiications on the basis of their characteristic mass signatures. This capability is particularly valuable for characterizing the numerous modiications displayed by NA components, which arguably outnumber those observed in proteins. Indeed, a variety of modiied deoxyribonucleotides has been described, which may be originated by normal cellular functions, undesired damage processes, or deliberate human intervention [50]. In the case of ribonucleotides, a total of 109 modiied variants have been discovered thus far in Nature and described in the RNA Modiication Database [51, 52]. The leading analytical platforms can only recognize modiications that cause detectable variations of spectroscopic or electrophoretic properties. Hybridization methods are hampered by the absence of complementary nucleotides capable of pairing with the modiied bases to maintain the idelity of copy during strand replication. In contrast, MS-based approaches can detect the presence of modiied bases, provide information on adduct structure and stoichiometry, and locate their sequence position through a variety of sequencing strategies [53, 54]. These intrinsic capabilities form the basis for numerous applications that take advantage of the sensitivity, speed, and speciicity afforded by MS. In the analysis of natural modiications, MALDI-MS was successfully employed to characterize nucleosides isolated from urine by HPLC or capillary electrophoresis (CE) [55]. The possibility of determining the identity and concentration of such species enabled their evaluation as putative biomarkers for leukemia, breast cancer, and other malignancies. Additionally, fragmentation data obtained by postsource decay (PSD) [56] conirmed the structure assignments and contributed to their positive identiication. In an example involving the analysis of synthetic variants, MALDI-MS was used to support the development of a novel strategy for the detection of carcinogen-modiied nucleotides, which was proposed as a possible alternative to the leading 32P-postlabeling methods [57]. This strategy relied on the introduction of appropriate imidazole dyes to enable detection by CE and laserinduced luorescence. Model biomarkers were generated by derivatizing deoxyribonucleotides with benzo[α]pyrenediol epoxide to produce the corresponding carcinogenic adducts. MALDI-MS was then carried out with THAP as matrix to conirm the structure of the synthetic model and to support the optimization of the labeling procedure. This type of information is accessible not only at the level of individual nucleosides and nucleotides, but also in the context of larger NA samples, which may call also for the additional task of locating the sequence position of the modiication. This capability can be exempliied by the characterization of carcinogen adducts of oligodeoxynucleotides ranging from 4 to 11 nt, which was accomplished on a Fourier transform-ion cyclotron resonance (FT-ICR) [58] analyzer by taking advantage of judicious matrix selection [59, 60]. Speciically, 3HPA was employed to obtain accurate mass information and to assess sample purity, whereas either

5.2 Genetic Markers

2,5-dihydroxybenzoic acid (DHB), or a mixture of anthranilic acid and nicotinic acid were used to prompt the formation of structurally signiicant fragments. A time delay was also inserted between the laser-desorption and ion-detection steps to enable metastable decomposition [60]. Alternatively, informative sequence ions were produced by collision-induced dissociation (CID) [61] in the FT-ICR cell. When accuracy is commensurate with the analyte’s size, the presence of covalent modiications can be directly revealed by accurate mass determination. Owing to the small number of fundamental units available for these types of biopolymers (i.e., A, T, G, and C for DNA; a, u, g, and c for RNA), any series of oligonucleotides of a given length and random base composition displays a much more limited mass degeneracy than that manifested by randomized polypeptides of equivalent length, which may contain up to 20 different building blocks [62]. It has been shown that, when mass accuracy is at least 100 ppm, a unique mass can be unambiguously assigned to each base composition possible for oligonucleotides up to 14 nt [63]. Therefore, any signiicant deviation from such mass could relect the presence of modiication, thus prompting further investigation. Detecting base modiications, however, is generally more straightforward when either base composition or sequence are exactly known, which enables an immediate comparison to be made between the observed experimental mass and that calculated for the target oligonucleotide. When working with known substrates, a direct mass determination can immediately provide the incremental mass of the modiier and the stoichiometry of modiication, thus offering valuable information on the possible structure and mechanism of adduct formation. One common strategy for these types of investigation consists of performing analyses before and after addition of the reagent of interest to readily appreciate the outcome. For example, MALDI-TOF was employed to characterize the products obtained by treating a model 18-mer deoxyribonucleotide with Pt(II) compounds [64]. The results clearly differentiated Pt(NH3)3-, Pt(NH3)2-, and PtCl (NH3)2-adducts, and revealed the addition of up to four modiications per substrate under the selected conditions. In this case, the speciic interactions involved in metal coordination were preserved by a 4 : 1 anthranilic acid/nicotinic acid matrix mixture. In a similar study, the photoactivated reaction of Ru(II) complexes with short DNA duplexes was investigated by a series of techniques, including MALDITOF [65]. The analyses were performed with a matrix consisting of a 2 : 1 mixture of THAP and diammonium hydrogen citrate. The results displayed not only adducts of individual strands, but also those of intact duplex, which helped compare the intra- and inter-strand reactivities of these photoactive compounds. In general, the type of information accessible by these determinations can be very broad, and depends not only on the characteristics of the available mass analyzer but, more importantly, on the actual experimental design. The position of modiied nucleotides can be located according to any of the sequencing approaches described earlier, which rely on either gas-phase fragmentation techniques, or the analysis of appropriate ladders generated in solution by chemical/enzymatic methods. Although the gas-phase sequencing of oligonucleotides up to 100 nt has been reported [66, 67], the practical limit for these

181

182

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

determinations is closer to ∼30 nt. When the analyte of interest exceeds this size, it is generally preferable to follow a bottom-up strategy that may involve digestion in solution phase, mapping of the ensuing hydrolytic products, and then gas-phase fragmentation of the putative species containing the modiication. Although this strategy was initially implemented on an ESI platform to identify post-transcriptional modiications in Escherichia coli ribosomal RNA (rRNA) [68, 69], later it was also employed with MALDI to map the modiications in the 5S rRNA subunits of other microorganisms [70]. In this example, the products obtained by RNase A hydrolysis of Sulfolobus acidocaldarius 5S rRNA are shown in Figure 5.5a. The signal detected at 1662.24 m/z could match only a possible 32–36 hydrolytic product that implied the absence of cleavage between c32 and g33. The fact that the observed mass deviated exactly by 14 Da from the value corresponding to its putative sequence cggac suggested that the methylation of c32 may be responsible not

Figure 5.5 (a) MALDI-TOF spectrum of

RNase digestion products of 5S rRNA of S. acidocaldarius; (b) Post-source decay mass spectrum obtained from the precursor ion

circled in panel (a). In these experiments, 3HPA was employed both as a matrix and as a denaturant. Adapted from Ref. [70].

5.2 Genetic Markers

only for the detected mass shift, but also for the inhibition of RNase A activity. When submitted to PSD analysis, the modiied oligonucleotide provided characteristic fragments that conirmed its sequence and placed the site of methylation on c32 (Figure 5.5b). In the context of this general bottom-up strategy, a wide variety of enzymatic and chemical reagents have been explored over the years to streamline the experimental procedures, minimize interference, and increase overall sensitivity. As part of the strategies involving the analysis of either intact analytes (topdown) or hydrolytic products (bottom-up), MALDI-MS has enabled the investigation of base methylation [71–73], carcinogen adducts [74, 75], oxidative damage and abasic lesions [76–78], as well as intra- and inter-strand crosslinking [79–81]. These approaches have been employed also to support the addition of chromophores and tags to facilitate detection [55, 82–85], the introduction of base-speciic modiiers meant to reveal mass-silent nucleotides [86–88], and the synthesis of modiications that increase the stability of the respective oligonucleotides [89–91]. Nucleotide analogs and stable isotopes, as common staples of MS research, have also been employed to investigate the gas-phase behavior of oligonucleotide ions and to understand the mechanism of fragmentation under typical MALDI conditions. For example, cytidine analogs including 5-methyl-2′-deoxycytidine, 5-bromo2′-deoxycytidine, cytidine arabinoside and 2′-luorocytidine, were introduced into model oligonucleotides to evaluate the stability of the N-glycosidic bond during analysis [89]. The results showed that a greater stabilization was achieved by substituting the 2′ position with electron-withdrawing functions, such as hydroxyl and luorine, than by modifying the 5 position of the nucleobase ring. The cleavage mechanism of the N-glycosidic bond in DNA, which results in base loss and subsequent backbone fragmentation, was investigated in deoxyoligonucleotides by performing exhaustive hydrogen/deuterium exchange directly on the sample stage. In positive-ion mode, the observations suggested that bases may be lost in either neutral or ionic form upon protonation/deuteration during the MALDI process [15]. Further studies showed that proton transfer may lead to charging of the nucleobase with the formation of a zwitterionic form that may then undergo base elimination in both positive and negative-ion mode, which triggers backbone dissociation [14]. Stable isotope labeling has been employed to enable the accurate determination of base composition and the sequence validation of PCR products [85, 92]. According to this strategy, >99% 13C/15N-labeled triphosphodeoxynucleotides (d*NTPs) were employed during the PCR ampliication of target sequences to obtain their respective labeled copies. Only one type of d*NTP at a time was included in each PCR reaction mixture (Figure 5.6), in such a way as to produce unique recognizable deviations from the mass of the unlabeled version. Indeed, knowing the shift corresponding to the 13C/15N labels in each individual nucleotide allowed the number of units of such nucleotides in the corresponding PCR product to be determined [92]. Subsequently, this strategy was employed to successfully determine the identity of uncalled bases in GC-rich regions that had been yielding

183

184

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

Figure 5.6 Determination of base composition by multiple PCR labeling reactions. “*”

indicates stable-isotope labeling. Reproduced with permission from Ref. [92]. ©1999 American Chemical Society.

inconsistent data via electrophoresis-based methods [85]. Unlike luorescent dyes that may hamper polymerase activity due to their different physico-chemical properties, the isotope-labeled nucleotides possess the same characteristics as their unlabeled counterparts, and thus can be incorporated at the same rate into the nascent biopolymer chain. Alternative strategies that do not require labeled building blocks have been developed in which stable isotopes can be incorporated directly into target analytes, for example, through enzymatic reactions. This approach was demonstrated by submitting RNA samples to RNase T1 digestion in buffers containing H218O, which induced the formation of hydrolytic products with labeled 3′-end phosphates [93, 94]. This labeling strategy was employed not only to distinguish endonuclease products from incomplete or nonspeciic cleavage products, but also to perform a relative quantiication of the RNA. In this case, two samples were digested in either 18O-labeled or regular water, and then combined immediately prior to MALDI-MS analysis (Figure 5.7). Monitoring of the “doublet” signals separated by 2 Da (i.e., the isotopic mass shift introduced by 18O) enabled the relative RNA content to be determined in the two samples [94]. It is important to note that, whilst the introduction of 2H can be signiicantly affected by back-exchange with protons, not only in the solution phase before analysis but also in the expanding plume during MALDI, no undesirable effects have been reported either for

Figure 5.7 General overview of the described 18O labeling and MALDI-MS approach for quantitation of RNA samples. Reproduced with permission from Ref. [94]. © 2005 American Chemical Society.

5.2 Genetic Markers 185

186

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics 18

O incorporated during hydrolysis, or for 13C/15N labels included in the NA building blocks. Finally, MALDI-MS has been applied also to the characterization of NA analogs with a modiied backbone structure, such as peptide nucleic acids (PNAs). This class of DNA-mimic consists of regular nucleobases such as adenine, thymine, cytosine and guanine, which are connected by a peptide-like backbone in a way that allows for correct hybridization with complementary DNA/RNA strands [95, 96]. Traditional protein matrices, such as sinapinic acid (SA), DHB and αcyano-4-hydroxy cinnamic acid (CHCA), as well as 6-aza-2-thiothymine (ATT), were shown to be superior to more traditional NA matrices, such as 3HPA, picolinic acid, and THAP [97]. Based on these indings, a strategy was developed for revealing sequence polymorphism within the human leukocyte antigen DQA focus [98]. According to this, the target DNA was immobilized to streptavidin beads through biotinylation and then added with a series of PNA probes corresponding to possible alleles. After removing the nonhybridized probes by stringent washes, the beads were analyzed directly by MALDI, which enabled a selective detection of the hybridized probes released from the DNA strand immobilized onto the surface. 5.2.5 Detection of Noncovalent Complexes of Nucleic Acids

When ESI and MALDI were irst introduced, they were immediately recognized as soft ionization techniques, based on their ability to generate abundant molecular ions with little or no analyte fragmentation. In the case of ESI, this notion was further cemented by numerous reports describing the observation of labile noncovalent complexes of biopolymers, including relatively short oligonucleotide duplexes [99, 100]. In the case of MALDI, intact duplexes were among the irst examples of noncovalent complexes to be successfully detected by this ionization technique [25, 101, 102]. . The utilization of ATT and ammonium citrate was found to be critical for the desorption of an intact 12mer/16mer duplex, which could not be observed with typical NA matrices, such as 3HPA and THAP [101]. The neutral character of this matrix (pKa = 6.31) was largely credited for its ability to prevent denaturation of the Watson–Crick pairing interactions that stabilized the duplex. The utilization of ATT, however, did not appear to alleviate the “irst-shot” phenomenon observed for noncovalent complexes, according to which only dissociated species can be typically detected beyond the very irst laser event (see also Chapter 1, Section 1.7). In this example, signals were detected also for the separate strands, but their presence was attributed to incomplete annealing in solution, rather than to dissociation effects induced by MALDI. The absence of homodimers of either strand served to demonstrate that the observed complex was the result of speciic interactions between complementary sequences, instead of nonspeciic clustering that may result from excessive oligonucleotide concentration [103]. An excellent correlation was later demonstrated between the gas-phase stability of deoxyoligonucleotide duplexes and their melting melting temperature (Tm) in

5.2 Genetic Markers

Figure 5.8 IR-MALDI relectron TOF-mass

spectrum of a mixture of double-stranded DNA fragments recorded from a glycerol/ ammonium acetate matrix. The sample was generated by digestion of the plasmid

pBluescript KS+ with the restriction enzymes BglI and RsaI. The sum of 20 single-shot spectra was recorded. Adapted with permission from Ref. [25].

solution, which is a direct function of their G-C content and the greater number of inter-strand hydrogen bonds afforded by these base pairs [104]. Optimal conditions for duplex detection were investigated in both UV- and IR-MALDI [25]. The study results conirmed that UV-MALDI with an ATT matrix performed well for samples up to ∼70 bp, whereas 3HPA induced strand dissociation under otherwise identical conditions. In contrast, IR-MALDI with a glycerol matrix enabled the observation of duplexes up to 920 bp and demonstrated much lower sample demands (Figure 5.8). Further, the latter was found to be rather sensitive to the presence of ammonium acetate and Tris–HCl buffer, which increased the stability of strand association by increasing the sample’s ionic strength. In subsequent studies, UV-MALDI analysis beneited from the inclusion of additives capable of stabilizing the inter-strand interactions [105]. For example, spermine was shown to enhance the detection of intact DNA duplexes when added to ATT in amounts that were approximately equimolar with the analyte. Further, the additive was also capable of displacing alkali adducts and reducing the incidence of base loss during desorption/ionization. These stabilizing effects were attributed to the ability of its positive charges to reduce the coulombic repulsion between negatively charged strands, thus limiting their dissociation. A mechanism was also suggested by which the presence of a co-matrix aided crystal growth by complexing the NA species during crystallization on the MALDI probe, thus increasing the yield of intact duplex. When ethidium bromide and methylene blue were evaluated under the same conditions, duplex stabilization was modest, whereas relatively stable complexes could be detected between the ligands and the double-stranded substrate [105]. These results were consistent with the intercalating properties of such molecules, which are responsible for their speciic binding to helical structures. In contrast, electrostatic interactions were shown to contribute prominently to the stability of noncovalent complexes between single-stranded deoxyoligonucleotides and

187

188

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

strongly positive ligands, such as guanidinium derivatives [106] and polybasic compounds [107]. In the case of guanidinium analogs, the binding stability was determined not only by the strength of the salt bridges between guanidinium functions and phosphate groups, but also by the hydrophobic nature of the aromatic substituents and their ability to establish π-stacking interactions with the nucleobases [106]. Consistent with the effects of ionic strength on electrostatic interactions, the binding of polyamines, polyimines and basic polypeptides was noticeably weakened by increasing the concentration of ammonium acetate in the samples [107]. This effect was opposite to that observed for hydrogen-bonding interactions, such as those between complementary strands in helical structures, which are instead strengthened by increasing salt content. The study revealed also that ATT had the tendency of enhancing the signal of duplexes over that of ligand complexes, whereas 2,6-dihydroxyacetophenone (DHAP) had the opposite effect. For reasons that are not understood, the former appeared to increase the stability of hydrogen bonds that stabilize double-stranded association, while the latter favored the electrostatic interactions at the basis of the binding of polybasic ligands [107]. In addition to enabling the analysis of intact NA–NA and ligand–NA complexes, MALDI was proven capable of preserving also the association of peptide–NA and protein–NA assemblies. The ability to desorb noncovalent complexes between basic peptides/proteins and single-stranded deoxyoligonucleotides was initially demonstrated by using a DHB matrix in positive-ion mode [108]. Subsequently, this was also demonstrated for double-stranded substrates with ATT [109] and for RNA oligonucleotides with THAP [110], both in negative-ion mode. The parameters determining the success of this type of analysis were evaluated in experiments that used model homo-oligomers and short peptides to generate nonspeciic complexes. These studies highlighted the importance of the pH-controlled formation of ion pairs between negatively charged NA species and strongly basic peptides (i.e., R- and K-rich) [108]. As discussed above, ionic strength and matrix selection were proven critical for the detection of these noncovalent complexes stabilized predominantly by electrostatic effects [109]. As is generally true for all MS-based approaches, the MALDI analysis of noncovalent complexes takes advantage of the fact that electrostatic binding gains strength in the gas-phase from the absence of shielding by solvent and counterions. However, it has been demonstrated that the overall outcome may relect also more subtle, weaker effects that play determinant roles in the speciic recognition between cognate partners. This point was demonstrated not only by the studies on duplex constructs, in which the speciic binding of complementary strands could be correctly discriminated from nonspeciic aggregation between random oligonucleotides, but also by the investigation of speciic protein–NA complexes of biological signiicance [111–113]. For example, assemblies containing two protein units and one or two equivalents of tRNA were readily observed in samples containing yeast aminoacyl-tRNA synthetases and their respective tRNA substrates [111]. These results were consistent with the fact that the active forms of this class of enzymes are known to be dimeric. The concomitant detection of protein mono-

5.2 Genetic Markers

Figure 5.9 MALDI-mass spectrum of a

mixture of yeast Tyr-tRNA-synthetase (TyrRS), tRNATyr, and tRNASer. The amount of material per sample was: 0.86 pmol of TyrRS, 3.2 pmol of tRNATyr, 3.2 pmol of tRNASer, 0.12 μmol of ATT, 38 nmol of ammonium acetate, and

0.4 μmol of glycerol. The sum of 22 single-shot spectra was recorded. Reproduced with permission from Ref. [111]. © the American Society for Biochemistry and Molecular Biology.

mers was attributed to equilibrium effects in solution, or to MALDI-induced dissociation of the initial complexes, whereas low-intensity multimers were explained with possible nonspeciic clustering in the expanding desorption plume. Based on these results, analytical artifacts were believed to be responsible for the observation of complexes comprising noncognate tRNAs, even though these enzymes are known to bind noncognate substrates with much lower afinity. On the other hand, when competitive binding experiments were performed by mixing each enzyme with both cognate and noncognate substrates, only cognate assemblies were observed (Figure 5.9). Therefore, the results demonstrated that MALDI-MS could be employed to take reliable “snapshots” of complexes that were formed in solution according to speciic rules of enzyme–substrate recognition. The results served also as a reminder that, as for any other MS-based approach for noncovalent complexes, designing proper control experiments is critical to enable unambiguous data interpretation and evaluation. The investigation of intact protein–NA assemblies may be accomplished also by following strategies that do not depend on the stability of noncovalent interactions under MALDI conditions. Crosslinking approaches have been developed to stabilize intrinsically reversible binding with permanent covalent bridges, thus enabling the determination of entire assemblies without unwanted dissociation [113–115]. For example, this strategy was successfully followed to investigate complexes of the E. coli single-stranded DNA-binding (SSB) protein with model deoxyoligonucleotides [113]. In the absence of crosslinking, only monomeric and homodimeric SSB were detected in positive-ion mode by using SA as a matrix. The tetrameric assembly of SSB, which represents the active form, was observed

189

190

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

only upon reaction with disuccinimidyl suberate to introduce stable inter-subunit crosslinks between primary amines (K side-chains and N-terminus). When a polydeoxythymidine (poly-dT) substrate was added in solution, full-ledged complexes consisting of tetrameric SSB bound to poly-dT were readily detected either before or after the crosslinking reaction. These results clearly indicated that the interactions established by the deoxyoligonucleotide were capable of stabilizing not only protein–NA but also protein–protein binding, thus enabling an analysis of the entire assembly even at the intrinsically low pH typical of this matrix system. The combination of crosslinking and MALDI analysis has been employed also to characterize the points of contact between bound components, and to identify proteins that may be interacting speciically with a target NA substrate [116–119]. Typical strategies involve either direct UV-photocrosslinking, or the activation of complexes in which appropriate photoafinity labels were incorporated in the NA moiety. Nuclease and protease digestion is typically performed to release relatively small peptide–oligonucleotide heteroconjugates, which can be then isolated/ puriied by chromatography before analysis. For example, the photoactive analog 5-iodo-deoxyuridine was introduced in speciic positions of a 20 bp deoxyoligonucleotide displaying the recognition site of the restriction endonuclease MboI from Moraxella bovis [119]. After UV-activation of the MBoI–deoxyoligonucleotide complex, the products were treated with trypsin, chymotrypsin/trypsin, or proteinase K. The heteroconjugates were separated from the hydrolytic peptides by Fe3+-immobilized metal afinity chromatography (IMAC). The NA moieties were then cleaved by treatment with HF to obtain peptides bearing only the initial photolabeled deoxyuridine (dU). Finally, the modiied peptides were characterized by MALDI-MS and MALDI-MS/MS by using a CHCA matrix under typical conditions employed for peptide/protein analysis. These experiments were able to pinpoint amino acid residues that had been in direct contact with the bound deoxyoligonucleotide, such as K209, which were identiied as critical players in speciic base recognition [119]. Similar approaches have been successfully implemented by crosslinking assemblies that were reconstituted in vitro from separate components [119–121], as well as extracted intact from cells/tissues [116, 118, 122–124]. However, as already demonstrated for ESI-based platforms, these approaches could be readily applied to study targets that are still immersed in their normal cellular environments by performing in situ crosslinking [125].

5.3 Assay Formats for Nucleic Acid Analysis by MALDI-MS

The development of assay formats for mass spectrometric analysis has followed the needs in genetic marker analysis and DNA diagnostics. In this section, a brief overview will be provided of the plethora of assay formats developed for the analysis of NAs from biological samples in general, and the various genetic markers in particular. A more detailed description follows in subsequent sections. Several recent reviews have summarized developments in the ield of mass spectrometric

5.3 Assay Formats for Nucleic Acid Analysis by MALDI-MS

NA analysis [126–129]. Hence, in order to avoid redundancy, the early developments will be only briely summarized at this point, and special attention will be paid to recent developments that make large-scale use of MALDI-TOF MS in genomic science. As outlined earlier, sample preparation is a key aspect for the robust analysis of NAs by MALDI-TOF-MS. The challenge of sample preparation can be subdivided into three categories: • • •

Which puriication format is suitable to provide NA products suficiently conditioned for MALDI-TOF-MS analysis? Which enzymatic reaction generates NA products of a size (and molecular mass) suitable for robust analysis by MALDI-TOF-MS? Which preparation of the matrix/analyte mixture provides suficient performance for the various applications of NA analysis by MALDI-TOF-MS?

Signiicant efforts were initially devoted to the puriication of NA products from enzymatic reactions prior to MALDI-TOF-MS. As will be indicated in later sections that describe the enzymatic assay formats in more detail, several components of enzymatic reactions are detrimental to the MALDI-process. Hence, puriication formats must be able to remove these components, including detergents, surfactants, proteins and unincorporated nucleotides, and must provide the means for the eficient removal of high concentrations of salt commonly used in enzymatic reactions (sodium-, potassium-, magnesium- chlorides and sulfates). Among the puriications formats evaluated for their potential in MALDI-TOFMS analysis, spin-columns with size separation capabilities as well as solid-phase puriication systems such as reversed-phase beads/columns and streptavidincoated beads, have proved very eficient [130–134]. The assay formats described in the following section very often rely on one of these puriication formats. While these methods provided eficient means for the puriication of enzymatic products, they also proved to be a hurdle for high-throughput applications: centrifugation steps, such as those required for spin columns, are dificult to automate. Solid-phase puriication systems usually require several reagent addition and washing steps, and therefore require more elaborate pipetting systems. It must also be considered that solid-supports, and in particular streptavidincoated paramagnetic beads, have a restricted binding capacity, and that the binding capacity will vary with the length of the NA product bound to the surface. Hence, the overall analyte concentration will be limited by the binding capacity of the solid support, and if multiple gene regions are to be analyzed simultaneously (referred to as multiplexing), then differences in the length of ampliication products can lead to unwanted biases in their representation after puriication. This cannot necessarily be offset by increasing the amount of coated beads, mainly because of the cost contribution and because of leakage of the capture protein (usually streptavidin) into the analyte solution (leading to protein signals dominating the mass spectrum). It was noted more recently, however, that the effect of detergents/surfactants and proteins can be minimized in most assay concepts simply by dilution of the

191

192

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

sample. Because of the exponential nature of the PCR commonly used to amplify gene regions prior to the analysis for mutations or SNPs, the NA analytes usually have a suficiently high concentration that supports the concept of sample dilution prior to MALDI analysis. The main challenge therefore resides in the ion-exchange step to condition the NA phosphate backbone for MALDI. Two concepts addressing the conditioning of the phosphate backbone will be described in the following section. The irst method, as introduced by Gut and Beck during the mid-1990s, uses DNA alkylation to generate nonionic NAs and therefore avoids the issue altogether [135]. The second – now commonly used – method employs a simple addition of porous ion-exchange beads to the diluted analyte-solution prior to mixing the analyte with matrix. Lastly, preparation of the matrix and matrix/analyte mixture has received considerable attention. The drivers of this point were the reproducibility and homogeneity of the crystallization process, as well as the potential for automation of the analyte transfer process. With respect to both, two signiicant developments are noteworthy that allowed MALDI-TOF-MS to be used routinely for NA analysis by non-experts and in high-throughput settings. Both developments were aimed at addressing the reproducibility of the inal step of sample preparation by using prefabricated miniaturized arrays of matrix spots, and transferring only submicroliter amounts of sample onto these arrays [136–138]. This development proved to be a cornerstone for enabling the quantitative analysis of NAs, as will be described below.

5.4 Applications in Genotyping

For each of the most common genetic markers described in the previous sections, attempts have been made to develop solutions using MALDI-TOF-MS. Here, investigations involving the analysis of RFLPs [139, 140] and microsatellites [134, 141– 145] using MALDI-TOF-MS will be reviewed, and attention focused on those applications which have found more widespread use in genomics/genetics research. 5.4.1 MALDI-TOF-MS SNP and Mutation Analysis

Assay formats speciically suited to SNP analysis by MALDI-TOF-MS have been a prime focus of recent method development. This is in part the case, because the analysis of large DNA molecules has not progressed suficiently to allow highthroughput use of MS (as discussed in Section 5.1). With major genomics activities focusing on SNP analysis, however, new opportunities have arisen for MS, with research groups having developed multiple new assays all of which had a common feature, namely that the size and nature of the generated products are well suited for mass spectrometric analysis in general, and particularly for MALDITOF-MS.

5.4 Applications in Genotyping

Earlier reports on the use of MALDI-TOF-MS for the detection of mutations and/or sequence changes have employed more conventional techniques such as restriction endonuclease digests [146] or the ligase chain reaction [147]. Initially, allele-speciic hybridization was also combined with mass spectrometric detection. Smith and coworkers, for example, used peptide NA probes hybridized against an immobilized PCR template for genotyping [148]. Finally, the direct measurement of PCR products has been attempted as a means of NA analysis and genotyping [17, 24, 92, 149, 150]. The majority of recently described methods and applications have used a primer extension concept to determine the genotype present in a particular SNP position and sample. One of the drivers for these assay concepts was the basic limitation in mass resolution and mass accuracy, which did not allow for a “simple” mass determination of PCR products to accurately deine genotypes. A common scheme among these methods begins with ampliication of the target region by PCR. A detection primer is then annealed immediately adjacent to the mutation site/polymorphic site and extended by a polymerase; the extension is terminated either on the polymorphic site or within a few bases thereafter, through the incorporation of terminator nucleotides (usually ddNTPs; see also Figure 5.10). In the simplest case, the products generated in these reactions are between 17 and 28 nt in length (ca. 5000–8500 Da) and are thus readily amenable for mass spectrometric analysis. The different concepts introduced deviate in the exact way that the primer extension reaction is performed and in the puriication formats applied. 5.4.1.1 The PinPoint Assay In one approach termed the PinPoint assay, introduced by Haff and coworkers, four dideoxy terminators are used in the post-PCR primer extension reaction [151]. The primer is only extended by one nucleotide and termination occurs directly at the polymorphic site. The genotype is then determined from the molecular mass of the extension/termination product, as this is directly correlated to the type of nucleotide incorporated at the SNP site. Discrimination of alleles is based on the mass difference between the four terminators (9 Da for ddA/ddT, 40 Da for ddC/ ddG, 16 Da for ddA/ddG, 25 Da for ddT/ddG, 15 Da for ddT/ddC, and 24 Da for ddC/ddA). A calculation of the molecular mass of the possible combinations of the four nucleotides for a string length between 17 and 30 nt shows that a single MALDI-TOF read could, in principle, resolve multiples of such primer extension products. Ross and coworkers were able to achieve a 12-plex analysis [152]. Intelligent assay design combined with “mass tuning” of primers, where nontemplated nucleotides were used to allow eficient mass-intercalation of multiple primer and primer extension products, increased the plexing level even further. However, in most cases the molecular biology limits these efforts. The equal ampliication of multiple target regions in one reaction is dificult to achieve. Single base extension methods such as the PinPoint assay have an advantage for multiplexed SNP marker analysis. Notably, they can combine all assays irrespective of the alleles to be genotyped simply by using the four dideoxynucleotides

193

194

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

Figure 5.10 Schematic representation of

analysis of single nucleotide polymorphisms (SNPs) by primer extension and MALDI-TOFMS. The genomic target region is irst ampliied by the polymerase chain reaction (PCR). An oligonucleotide primer is then annealed immediately adjacent to the polymorphic site. This primer is extended allele-speciically using a DNA polymerase and a nucleotide mix that leads to a termination of the primer extension reaction either after one or two nucleotide additions. The length of the extension products is determined by the allele present in the analyzed sample. In the depicted case, the

primer is extended by one nucleotide (incorporation of a ddG) for the G allele, and by two nucleotides (incorporation of dA and ddC) for the C allele. Analysis of the extension products by MALDI-TOF-MS allows the unambiguous determination of the genotype present in the sample. Several variations of this basic concept have been developed where the primer is extended only by a single base, and where mass-modiied nucleotides are employed. All these variations aim to increase the ability to interrogate more than one genomic region at a time and therefore use the available mass window more eficiently.

5.4 Applications in Genotyping

(ddNTPs), although the concurrent use of all four ddNTPs may present a technical challenge. For longer primers, the resolution of certain types of mass spectrometer may be insuficient to resolve the mass difference of 9 Da for a heterozygous A/T mutation, as both peaks appear in the spectrum. Indeed, for primers with molecular weights in excess of 5000 Da the mass accuracy might not even routinely be suficient to distinguish between a ddA or ddT extension in homozygous samples. Furthermore, the mass differences between C-A (24 Da) and C-G (40 Da) fall very close to sodium and potassium adducts. This introduces the risk of false-positive heterozygote genotypes when salts are not completely removed from the analyte. A simple way to avoid ambiguities related to the “inconvenient” mass difference between pinpoint products is the use of mass-modiied terminator nucleotides, as introduced by Smith and coworkers [153]. The rationale design of terminator nucleotides with a “perfect” mass separation can be achieved by selecting appropriate mass-modifying chemical moieties. 5.4.1.2 The PROBE Assay The primer oligo base extension (PROBE) assay, as introduced by Little and coworkers, avoids mass resolution issues with a different terminator strategy [154, 155]. The post-PCR primer extension reaction is performed in the presence of one ddNTP and three dNTPs. The mixes are chosen in such a way that the two allelespeciic extension products always differ in length and mass by at least one nucleotide. The mass difference of about 300 Da decreases the demand on the mass resolution and accuracy of the TOF-MS under high-throughput conditions, and adds a considerable safety margin. Assays can be multiplexed by intercalation in the same way as in Pinpoint. However, assays can now only be multiplexed if their alleles can be differentiated with the same ddNTP/dNTP combination. Termination with a mix containing only one dideoxynucleotide bares another risk. Polymerase pausing artifacts, where the extension reaction is not speciically terminated, can have the same mass as “real” termination products. Pausing on dATP (primer + 313.2 Da) for example leads to the same mass as termination on ddGTP (primer + 313.2 Da), and this could lead to a misinterpretation of results (again false heterozygous genotypes). These situations can, however, be circumvented through the design of tri-terminator mixes exhibiting clear mass differences between all possible products. An interesting approach for the increase of multiplexing levels in MALDI-TOFMS genotyping has recently been introduced by Ju and coworkers [156, 157]. The basis of the genotyping assay is a conventional single-base primer extension reaction in which terminator nucleotides are used that carry a biotin group. The chemical linker between the biotin group and each of the four ddNTPs has been selected such that the mass difference between the terminator nucleotides doubles. The reaction products are incubated with streptavidin-coated magnetic particles. Only extended primers carry a biotin moiety (at their 3′ end) and will thus be immobilized. Unextended primers and reaction components can be eficiently separated with the supernatant. After release from the streptavidin beads, the extension products can be analyzed using MS. Thus, Ju et al. combined the

195

196

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

beneits of solid-phase puriication, which usually yields very pure analytes, with the selection of only extended primers and with mass-modiied terminator nucleotides. The absence of unextended primers removes undesirable additional masses/signals and allows a more eficient use of the available mass window. As mentioned above, the use of terminators optimized for mass difference avoids ambiguities in spectra interpretation at limited mass resolution. Although this approach is technically quite attractive, it suffers from high processing costs and complications in automation. The sample puriication requirements of NAs in MS have always been seen as a limiting factor. Adduct formation (mainly with sodium and potassium as the predominant cations in enzyme reaction buffers) is probably the biggest concern in MALDI-MS, although buffer detergents and large amounts of protein add signiicantly to the list of issues. Various methods for the removal of these ingredients from the analyte solution prior to MS analysis have been suggested. Solid-phase approaches, such as coated magnetic beads, have probably been the most successful in this respect. The automation of solid-phase processes is feasible with reasonable efforts in instrumentation and robotics, and the samples are usually very “clean.” However, the processing of tens to hundreds of 384-well microtiter plates on a daily basis, coupled to the multitude of pipetting steps associated with the addition and removal of washing liquids, present a major bottleneck. In addition, coated particles are usually very costly. The issue of liquid handlers being the ratelimiting step (especially given the high acquisition speed of MS) is very often dealt with through the integration of multiple pipettors/dispensers for parallel processing which, however, is also expensive. Subsequently, several research groups made the development of less-demanding puriication schemes for SNP analysis a focal point of their investigations. Two different approaches have resulted from these activities. First, a concept described by Meyer and coworkers employed the addition of NH+4 -conditioned ion-exchange resin subsequent to completion of the post-PCR primer extension reactions [158]. An overnight incubation with the ion-exchange resin removed sodium and potassium suficiently from the primer extension products (usually these are only 17–25 nt in length) and did not require any further sample puriication. The addition of ion-exchange resin to the solution containing analyte and matrix had actually been reported earlier [10, 159]. Upon matrix crystallization, the resin particles accumulate in the center of the microliter preparations, and can easily be removed manually before the sample is introduced into the mass spectrometer. For nanoliter preparations and automated analysis, as is commonly applied nowadays, the commercially available resin particles of about 100 μm in size interfere with correct matrix crystallization and this results in unacceptably inhomogeneous samples. 5.4.1.3 The MassEXTEND Assay The PROBE assay has been further developed into a homogeneous assay (now termed homogeneous MassEXTEND; hME) using the same puriication principle, whereby all sequential enzymatic steps are performed by a “simple” addition of the reagents to the reaction well. No washing steps are required, and the ion-

5.4 Applications in Genotyping

exchange resin is added to the crude solution in the last step [160]. In order to circumvent long incubation times which are undesirable for high-throughput processes, the use of ion-exchange resin was coupled with a dilution of the sample with deionized water and optimization of resin types (mesh size and conditioning of the resin prior to use), which yields incubation times of only 10 min. Current processing volumes in the low microliter range allow such a step without the necessity of transferring the sample into new reaction vials. As homogeneous assays do not allow for the removal of all components affecting the MALDI process, the dilution plays an important role. If suficient amounts of analyte are produced, dilution can be used as a vehicle to reduce the concentration of buffer and reaction components to the point that they do not interfere signiicantly with the sample crystallization and desorption process. The reduction of analyte volumes (as little as 10 nl is currently used in combination with micro-arrayed MALDI targets/chips) also aided the reduction of sample purity requirements [138, 161]. Originally, homogeneous assays were not an obvious solution for the development in NA analysis by MS. Rather, they represented a strong contrast to the scientiic literature which reported the impact of common ingredients of molecular biological reactions on MS performance. The inal analyte solution contains almost everything (protein, salts, detergents, large NA PCR templates, genomic DNA, nucleotides) previously described as degrading the spectra quality. As described, recent technological improvements in MALDI-TOF-MS, as well as in the miniaturization of MALDI sample preparation, have surely aided in this development. The emergence of homogeneous assays must be attributed to the necessity for simple, cost-eficient processes in high-throughput operations. These represent a compromise between performance (analytical yield and accuracy) and cost for the assay, the process, and the operation, which includes equipment time and resources. Recently, hME assays have gone through several iterations and improvements, such that 36–40plex reactions are now routinely analyzed. Due to the mass range limitation of MALDI-TOF of about 9–10 kDa and other hurdles related to the prior sample creation (i.e., the PCR step), a signiicant challenge will emerge for measuring samples beyond a 60–70plex. Due to the high peak density it is essentially unavoidable that a few signiicant peaks will overlap or suffer from too-low signal intensities and, as a consequence, some of the multiplexed assays may not lead to a result. A representative mass spectrum of a 57-plex MassEXTEND reaction is shown in Figure 5.11. Here, the extension products span the mass range from around 5000 Da to 9000 Da, and are suficiently separated for a clear assignment of genotypes based on the presence of mass signals. The successful use of MALDI-TOF-MS as a high-throughput genotyping method at multiple sites worldwide has conirmed that this compromise can be deemed successful. The issue in high-throughput processing has been addressed in various ways. Ivo Gut, for example, introduced the concept of charge-tagging and subsequent alkylation of the phosphate backbone as a means of increasing ion-yield and avoiding analyte puriication [135]. Early experiments demonstrated improvements of

197

198

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

Figure 5.11 MALDI-TOF mass spectrum of a

57-plexed SNP genotyping reaction. Multiplexed analysis of SNPs builds on the same scheme as depicted in Figure 5.9. However, to make more optimal use of the available mass window, multiple genomic regions are coampliied in the same reaction and the ampliied SNPs are each targeted with a speciic extension primer. These primers are designed so that their molecular mass and the molecular mass of all their extension products do not overlap and still

allow unambiguous identiication of the SNP that they were targeting and the allele that was present in the sample. Each assay occupies three mass signal positions representing the unextended primer, and the irst and second extension products. With careful consideration of the mass resolving power, between six and 70 SNPs can be interrogated in a single reaction and mass spectrometric readout. The vertical dotted lines indicate the position of a relevant peak in the multiplexed spectrum.

the mass spectra when the NA analyte carries a preferred site for a single positive or negative charge [83]. In order to avoid heterogeneous cationization of the phosphate backbone by sodium or potassium, the phosphate backbone is alkylated. However, this approach relies on a quantitative alkylation which becomes increasingly dificult for species larger than 8mers. Reduction of the primer extension products of >18 nt to small fragments can be achieved by the incorporation of phosphorthioate linkages near the 3′ end of the extension primer and the addition of an exonuclease digest step which stops at the phosphorthioate link. The concept was originally evaluated using an oligonucleotide model system. 5.4.1.4 The GOOD Assay A streamlined version of the above-described concept has been implemented as the “GOOD” assay for SNP analysis [162]. The assay was adapted to cope with the peculiarities of oligonucleotide synthesis, enzymatic reactions and alkylation eficiencies. The positive-ion MALDI-MS versions of the GOOD assays will be

5.4 Applications in Genotyping

described here in more detail. The charge-tag concept for positive-ion MALDI-MS lends itself to a more facile assay development. Common to other approaches, the target region is irst ampliied by PCR, after which the remaining dNTPs from the ampliication are degraded by the addition of shrimp alkaline phosphatase (SAP). In a subsequent simple add-on step, an extension primer is annealed adjacent to the polymorphism/mutation. This primer carries a charge-tag near the 3′ end and a phosphorothioate bridge on the 5′ side of the charge tag. A DNA polymerase allele-speciically extends the primer using a suitable mixture of alpha-thio dNTPs/ ddNTPs. A subsequent treatment with a 5′-speciic phosphodiesterase digests the primer sequence down to the irst phosphorthioate, leaving a small oligonucleotide with a charge-tag and the SNP-speciic nucleotide. The phosphorothioate bridges of these oligomers are then quantitatively alkylated using methyl iodide. The alkylated products are signiicantly less susceptible to cationization as the backbone is now neutralized. In summary, the “GOOD” assay combines charge-tagging, backbone neutralization by alkylation and enzymatic treatments to allow for a homogeneous primer extension assay with subsequent detection of the products by MS. The addition of reagents during the subsequent steps “dilutes” the ingredients, which hampers the MALDI process. This is similar to the dilution by addition of water as described above for the MassEXTEND assay. Furthermore, the allele-speciic products are rather small after phosphodiesterase treatment (usually the length of tetramers/ pentamers), and this helps the eficiency of the alkylation reaction. One limitation of the original “GOOD” assay was the limited availability of modiied nucleotides for subsequent coupling of positive charge-tags. However, this issue has been addressed with the introduction of new β-cyanoethyl phosphoramidites, which allows for easier charge-tagging [163]. The need for toxic reagents for alkylation of the phosphorothioate backbone was a further serious issue that limited the more widespread use of the GOOD assay. However, two improvements have been introduced recently to avoid the need for alkylation [164]. Primers carrying methylphosphonate groups towards the 3′-end yield an “alkylated”, charge-neutral 3′-end by synthesis; these provide the same inhibition of the phosphodiesterase digest as primers carrying phosphorothioate bridges. The incorporation of standard ddNTPs or alpha-thio ddNTPs during the primer extension reaction generates a negative charge for MALDI-MS analysis. Methylphosphonate-containing primers are usually not the preferred substrate for DNA polymerases. The introduction of Tma31 FS, an enzyme capable of extending methylphosphonate primers and preferring the incorporation of ddNTPs over dNTPs, aided in the development of an improved “GOOD” assay. Some of the general issues in multiplexed MALDI-TOF-MS genotyping related to mass accuracy, resolution and sensitivity could also be resolved if the products of primer extension reactions were to fall into a more benign mass window. Several approaches were conceived to generate short analytic fragments by introducing site-speciic cleavage points into the primers. The recently introduced GenoSNIP assay, for example, uses photocleavable linkers in the primer to shorten the extension product prior to MALDI-TOF-MS. Primers used in the GenoSNIP

199

200

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

assay contain a photocleavable o-nitrobenzyl moiety, which replaces a nucleotide. This linker produces an abasic site within the primer, which does not inhibit its correct annealing to the target region. Exposure to UV irradiation cleaves the extended primer at the introduced moiety and allows for a separation of the informative extended 3′ end from the noninformative primer. The GenoSNIP assay uses streptavidin-coated particles for puriication of the extended primer from the reaction prior to release of the 3′ end by photocleavage. Sauer and coworkers merged the concepts of photocleavage for primer shortening and the GOOD assay to avoid the necessity of solid-phase puriication and to overcome some shortcomings of the GOOD assay [165]. The combined approach uses primers containing phosphorthioates and a photocleavable linker towards the 3′ end of the extension primer. The allele-speciic extension/termination is performed with α-S-ddNTPs. Photocleavage and subsequent alkylation generate short assay-/allele-speciic DNA fragments. The photocleavage leaves a 5′ phosphate group at the 3′ fragment, and thus conveniently generates singly charged molecules without the introduction of real charge-tags or necessity for phosphodiesterase treatment, as in previous versions of the GOOD assay. Although shorter DNA fragments are easier to analyze, it has to be noted here that this also reduces the compositional space available for multiplex design. A restriction to a length of 4mers with the last position ixed to a speciic ddNTP leaves not more than 20 possible compositions with mass differences large enough to be resolved with MALDI-TOF-MS. This presents a rather serious restriction if the assay panel to choose from is limited. The most eficient way of multiplexing might be to allow for cleavage sites to be distributed anywhere in the extension primer. In this way, the compomer type and length can be used to spread assays over a deined mass range. However, to what extent the alkylation reaction can be eficiently employed on longer cleavage products remains to be explored. 5.4.1.5 The Invader Assay One of the few non-primer extension-based methods for MALDI-based genotyping is the Invader assay [166]. This uses two sequence-speciic oligonucleotides (an “invader” oligonucleotide and a probe oligonucleotide), which hybridize to a target sequence; the principle is depicted in Figure 5.12. The probe oligonucleotide and target region form a duplex which includes the polymorphic site and a noncomplementary 5′ overhang of the probe. The invader is designed such that its 3′ nucleotide invades into this duplex at the polymorphic site, forming a sequence overlap at this position. DNA repair enzymes called lap endonucleases (FENs) speciically recognize and cleave the unpaired region on the 5′ end of the probe oligonucleotide. This generates a reporter molecule, which can be used to signal the cleavage event. Each genotyping assay requires three oligonucleotides: the invader and two allele-speciic probes. The noncomplementary sequence of the 5′-end of these probes can be designed to allow discrimination between the probes that have been cleaved. If thermostable enzymes are used, this process can be run near to the melting temperature of the duplex formed between probe and target sequence, so that cleaved products are cycled off and replaced by noncleaved

5.4 Applications in Genotyping

201

Allele 2 (Y)

Allele 1 (R) primary system 5’ -A

3’

GC AG

cleavage site

Primaryy Allele 1 probe p

R N

3’ 5’

R

5’

GA ARM 1 GG CG CA CT T Invader oligonucleotide

5’

GC A

ARM 2

CA CG TT T

Invader oligonucleotide

3’

Genomic DNA representing Allele 1 (R)

cleavage site

Primary Allele 2 probe

Y N

3’

Y

5’ -A AC

5’

Genomic DNA representing Allele 2 (Y)

5’ -T TT

cleavage site

5’ -T T TTTCATCTCAGAC-3’Secondary probe for ARM 1 ARM 1 5’-AACGAGGCGCACTTR 3’NH2-AAUUGCTCCGCGTGAAAAAGTAGAGTCTG-5’Secondary target for ARM 1

5’-TTT-biotin

biotin

biotin

secondary d system t

Identifier molecule

5’-TTTT-biotin (1538da)

1234

m/z

1234

1538

intensity

intensity

intensity

(1234da)

Sample Homozygote for Allele 1 (R)

cleavage site

TCTCTTACACTGT-3’Secondary probe for ARM 2 ARM 2 5’-AGCAGGCACACGTTTY 3’NH2-AAUCGTCCGTGTGCAAAAGAGAATGTGACA-5’Secondary target for ARM 2

1538

m/z

m/z

Sample Heterozygote

Sample Homozygote for Allele 2 (Y)

Figure 5.12 Schematic representation of the

Invader assay for MALDI-TOF-MS based genotyping. The Invader assay uses two sequence-speciic oligonucleotides (an “Invader” oligonucleotide and a probe oligonucleotide) that hybridize to the target sequence. The oligonucleotides form a duplex which includes the polymorphic site and a noncomplementary 5′ overhang. The invader oligonucleotide creates a sequence overlap at the polymorphic site that is speciically recognized by a DNA repair enzyme (a lap endonuclease [FEN]) and cleaved. The cleavage releases the 5′ overhang of the probe oligonucleotide and

this released sequence can then be detected by MALDI-TOF-MS. For each SNP, two allele-speciic oligonucleotides are used that also differ in their 5′ overhang. MALDI-TOFMS analysis of the reaction products allows identiication of the released overhangs, and therefore identiication of the alleles present in the sample. In most cases a so-called “squared invader” assay is used. This variations uses the irst released 5′ overhang as an invader oligonucleotide for a separate, generic reporter system that releases a second 5′ overhang speciically designed for detection by MALDI-TOF-MS.

202

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

probes. This process is eficient enough that the Invader assay does not require prior ampliication of the target region by PCR. Smith and coworkers combined this genotyping method, which is operates directly on genomic DNA, with MALDI-MS analysis. In a reaction called the Invader Squared assay, the primary cleavage product serves as an Invader molecule for a second Invader reaction. The “squared” cleavage product is then puriied via streptavidin–biotin, which allows separation and conditioning of the cleavage products for MALDI-TOF-MS [167]. 5.4.1.6 The Incorporation and Complete Chemical Cleavage Assay A process called the Incorporation and Complete Chemical Cleavage assay (ICCC) was introduced recently as an alternative genotyping principle to the most commonly used primer extension methods [168]. During PCR, one of the natural deoxyNTPs is completely replaced by a chemically labile nucleotide. Following PCR, the generated amplicon can be cleaved speciically at the incorporation sites, which generates a set of cleavage products that can be analyzed using MS. The genotype can be deduced from speciic cleavage products and their corresponding mass signals. Usually, one of the PCR primers is designed adjacent to the polymorphic site such that the cleavage products containing the primer sequence (which does not contain any modiied nucleotides) up to the irst cleavage site generate the allele-determining mass signals. Assay design is a function of the alleles to be discriminated, the primer position, and the choice of modiied nucleotide. The combination of these three factors determines the signals pairs, which can be used for genotyping. 7-Deaza-7-nitro-dATP/dGTP and 5-hydroxy-dCTP/ dUTP have been proposed as chemically labile nucleotides for base-speciic cleavage. Both nucleotide sets form standard Watson–Crick base-pairing, which is important to maintain the level of artifacts generated by modiied nucleotides at a minimum. Cleavage is mediated through an oxidant and incubation with an organic base. Two important factors determine the utility of this process: (i) the modiied (chemically labile) nucleotide must be incorporated eficiently enough by a DNA polymerase for it to completely replace its natural counterpart in PCR without decreasing PCR yield; and (ii) the cleavage reaction must be complete for each incorporated modiied nucleotide (to avoid potential misinterpretation of mass signals), and must not change the chemical nature of any other nucleotide in the amplicon. There are several process-inherent advantages for the proposed method. The incorporation of modiied nucleotides during PCR allows a combination of the ampliication and genotyping process in one reaction, and eliminates the need for subsequent enzymatic reactions. This simpliies the processing steps, which is important for automation. As the double-stranded amplicon is directly used, the cleavage reaction generates products (i.e., mass signals) from both strands. This provides redundant information for the determination of genotypes and, by the internal conirmation, increases the accuracy of the result. The underlying principle of this method, namely base-speciic cleavage, is now used more widely for the sequence analysis of target regions. Further applications

5.4 Applications in Genotyping

of this principle are discussed in Section 5.5. The main concern surrounding ICCC as a general method for MS-based sequence analysis can be attributed to the slightly lower ampliication yield with modiied nucleotides, which might not prove suficient for an analysis of >500 bp. Furthermore, oxidative agents used for chemical modiication of nucleotides are often hazardous and thus not desirable for high-throughput processing. Both issues can, however, be overcome by identifying not only DNA polymerases with increased incorporation rates for modiied nucleotides but also alternative cleavage reagents. This is very similar to the evolution of the GOOD assay described above. 5.4.1.7 The Restriction Fragment Mass Polymorphism Assay One of the most recent assay formats used for genotyping by MALDI-TOF-MS is termed the restriction fragment mass polymorphism (RFMP) assay [169]. This assay was developed to perform the genotyping of polymorphic regions, where the close proximity of multiple polymorphisms renders the development of post-PCR primer extension readouts extremely dificult. In the RFMP assay, the region of interest is ampliied with PCR primers carrying IIS restriction endonuclease recognition sites. Following PCR ampliication, the PCR product is digested with a IIS restriction endonuclease that cleaves at a speciied position of up to 20 nt distance from the recognition site. If both PCR primers carry a IIS recognition site, this process allows for an enzymatic removal of the primer sequences from the PCR product, and generates a cleavage product of a length suitable for direct analysis by MALDI-TOF-MS. Kim et al. used this approach for genotyping polymorphic regions of the hepatitis C virus (HCV) genome [169]. 5.4.2 MALDI-TOF MS for Haplotyping

The use of SNPs as polymorphic markers for association/linkage disequilibrium (LD) studies by now supersedes conventional studies using microsatellite markers. The huge number of available SNPs and their relatively homogeneous distribution in the human genome compensates for their biallelic nature. In some instances, however, haplotype structures (the collection of genotypes found in a single allele or chromosome) rather than individual SNPs can be the principal determinant of a phenotypic consequence. Typically, they can provide additional statistical power in mapping disease genes, especially when the SNPs contributing to a disease are not directly observed, or when their interaction does not follow a simple additive effect [170–174]. Correspondingly, determination of the allelic phase of SNPs (the haplotype) is an important component of genetic studies. Determining the haplotype for several SNP markers in a diploid cell is challenging. The availability of pedigree genotype information, for example, can be used to determine offspring haplotypes. In addition, computational algorithms have been developed to impute haplotypes based on available genotype information in a population [170, 175]; both methods, however, can fail to reconstruct the correct haplotypes. An alternative to these methods is direct molecular hap-

203

204

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

lotyping. All genotypes determined after the physical separation of the two homologous chromosome sets/genomes are in phase by default, and thus represent the haplotype. Physical separation requires rather elaborate methods such as cloning or somatic cell hybrid construction [176, 177]. Other methods make use of allele-speciic PCR to amplify only one homolog of a gene/gene region; alternatively, single-molecule PCR is used whereby each of the two homologs is represented on a statistical basis [178–181]. To date, reports on true highthroughput haplotyping methods are scarce. The use of MALDI-TOF-MS for high-throughput haplotyping was evaluated in two recent reports. Ivo Gut and coworkers combined an allele-speciic PCR for the selective ampliication of only one of the two homologs of a gene with multiplex genotyping of further SNPs contained in the PCR product by the GOOD assay [181]. Allele-speciic PCR uses primers which hybridize with the 3′-end on the polymorphic site, and by choosing the nucleotide the ampliication can be directed towards the desired allele. DNA polymerases will recognize mismatches in the hybridization of the primer to the undesired allele, and will not extend these primer template structures. If the allele frequency of SNPs in the genomic region of interest is unknown, however, multiple SNPs have irst to be typed to identify the informative heterozygous positions. These can then be used for allele-speciic ampliication, followed by re-genotyping of SNPs contained in the allele-speciic amplicon, which then yields the haplotype information. If a suficient number of SNPs with known heterozygosity is already available from databases, a brute-force approach can be utilized whereby each highly polymorphic SNP can be used as an anchor position to perform allele-speciic PCR on all samples. Samples homozygous for the selective SNP will be easily identiied by producing heterozygous genotypes in the mass spectrometric analysis, and can be excluded from haplotype analysis. The combination of allele-speciic PCR with primer extension and MALDI-TOFMS for haplotyping may help to overcome some of the throughput hurdles of current technologies. MALDI-TOF-MS offers multiplex SNP haplotyping capabilities while maintaining a high analysis speed and accuracy. Although allele-speciic PCR can be established as a robust method, it has often been observed that reactions “leak”; that is, the undesired allele sometimes coampliies through primer mismatch and results in ambivalent genotypes. Hence, a careful adjustment of primer design (GC-tails, mismatch design) and ampliication conditions must be performed to render the reactions more speciic. These optimizations are not feasible for brute-force approaches, however. In order to avoid issues associated with allele-speciic PCR, Ding and Cantor used singlemolecule dilutions of genomic DNA for the separation of two homologous genomic DNAs, and combined this with the MassEXTEND assay [182]. This scheme proposed a dilution of genomic DNA to about one genome copy per PCR aliquot. The estimation of copy numbers of very dilute DNA concentrations follows the Poisson distribution [183], and accordingly, multiple replicates of each individual must be analyzed until the haplotype for multiple SNPs can be constructed. Ding and Cantor estimated from their results that the PCR eficiency from single molecules

5.5 Applications in Comparative Sequence Analysis

was about 90–95% for amplicons of about 100 bp length. This high PCR eficiency, obtained with current ampliication systems, is one of the reasons why this approach – which originally was proposed a decade ago but was soon abandoned due to its impracticability – now seems feasible. When multiplexing PCR and MassEXTEND, Ding and Cantor achieved a haplotyping eficiency of 40–45% per reaction; thus, four replicates should increase the haplotyping eficiency to about 90%. An additional advantage of single-molecule ampliication for haplotyping is the independence of the method from the distance between SNPs. Allele-speciic PCRbased methods require SNPs to be adjacent enough that they can be ampliied on the same PCR product. Single-molecule methods, when combined with the intelligent design of overlapping multiplex assay, allow the haplotyping of more than 20 kb lengths, provided that the genomic DNA does not have any substantial physical breaks. As pointed out earlier, these described methods would, in principle, also be applicable to other detection platforms; however, the accuracy and speed of mass spectrometric analysis, combined with its multiplexing capabilities, makes the use of mass spectrometers especially attractive for analyzing a large number of samples.

5.5 Applications in Comparative Sequence Analysis

MALDI-TOF-MS was proposed as a separator and detector for Sanger sequencing ladders during the early 1990s, when new technologies to sequence the human genome were in great demand. The main appeal of TOF-MS was the extremely high duty cycle, which suggested acquisition and analysis times on the order of seconds, as well as a high potential for automation. For comparison, the separation and detection speed of state-of-the-art sequencing equipment using the Sanger concept is still in the range of hours. In addition to the replacement of laborious electrophoresis steps, MS is interesting for the ield of NA sequencing, as it does not require labeling (either radioactive or luorescent). Moreover, the method measures an inherent physical property – the molecular mass – and this should increase accuracy and minimize any false interpretation of artifacts. The basic concept of sequence determination using mass spectrometric analysis of Sanger sequencing ladders relies on the superposition and alignment of mass signals obtained in the four base-speciic termination reactions originally described by Sanger and coworkers [184]. The molecular mass of termination products, as well as their mass difference, can be used to calculate the sequence. In essence, all mass signals arising from Sanger sequencing should be composed of a combination of the four nucleotides A, C, G and T, plus one terminator nucleotide (ddA, ddC, ddG, or ddT). The discrimination of termination events from polymerase artifacts is feasible, if a suficient mass resolution and mass accuracy is provided by the instrumentation. An early report by Lloyd Smith exempliied the

205

206

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

concept of sequence analysis with MALDI-TOF-MS using a mock DNA sequencing ladder produced by mixing oligonucleotides [185]. Since then, several groups have elaborated on the approach with various schemes and assay formats [130–132, 168, 186–190]. One of the longest sequencing read lengths obtained by MALDI-TOF-MS analysis of Sanger sequencing ladders was reported by Taranenko and coworkers [191], who were able to identify sequencing products of up to 120 nt. With the primer length subtracted, this represents an effective read length of 100 nt. At the same time, a report from Taranenko and coworkers illustrated some of the basic (and often ignored) technological limitations. First, in most reports a model of known sequence was used to evaluate the method, but this bears the risk of supervised data analysis being geared towards the known end result. Second, current mass accuracy and mass resolution present a challenge to accurate sequence determination. Here, the biggest challenge is the discrimination between “real” termination products and polymerase pausing events. A mass difference of 16 Da has to be resolved in a mass window ranging from 5 to 25 kDa. Even more challenging situations can occur in the case of, for example, sporadic false termination signals interfering with “true” termination, in which case mass differences less than 16 Da must be resolved. The two most recent approaches aimed at addressing these issues. First, Kang and coworkers employed a transcriptional synthesis of chain-terminated RNA ladders for sequencing [192]. This approach has two potential advantages: (i) the isothermal transcription process increases the yield of analytes several fold more than a cycle sequencing reaction; and (ii) the analyte is now composed of RNA, which is less prone to ion fragmentation. As expected, the mass separation power reported by Kang and coworkers was less limiting as compared to standard DNA sequencing ladders, although the read length could not be signiicantly extended. One reason for this might be related to the processivity and idelity of the enzyme when noncanonical nucleotides, such as the 3′-deoxyribonucleotide terminators, are employed. Abortive cycling (the unspeciic premature termination of the transcription process within the irst 10 nucleotides) is another factor that surely would inluence spectra interpretation and analyte yield. The most recent report by Ju and coworkers introduced the use of biotinylated dideoxy nucleotides for MALDI-based DNA sequencing [193]. The use of biotinylated chain termination nucleotides addresses the issue of polymerase pausing, a biochemical artifact that is not a speciic issue for MALDI-based sequencing but which became eminent at a very early stage during Sanger sequencing with luorescently labeled sequencing primers. All of the extension products generated from the sequencing primer were carrying the same luorescent label, such that distinction between speciically terminated fragments and unspeciic byproducts was impossible. In an attempt to increase the accuracy and ease of sequence analysis, luorescently labeled chain terminators (ddNTPs) were developed. Biotinylated dideoxynucleotides followed the same principle, allowing the separation of speciically terminated sequencing products from unspeciic/unwanted byproducts through solid-phase puriication systems such as streptavidin-coated magnetic beads. In this case, the mass spectra appeared “clearer” and the risk of

5.5 Applications in Comparative Sequence Analysis

Figure 5.13 MALDI-TOF mass spectrum of a

Sanger sequencing ladder generated by primer extension from PCR products using a mixture of normal elongators (dNTPs) and biotinylated terminator nucleotides (ddNTPs). The termination products are puriied on a streptavidin-coated solid support. The sequence of the target region is

derived by calculating the mass difference between the termination products. Each mass signal in the spectrum is marked with the mass difference to the preceding mass signal and the corresponding nucleotide that this mass difference represents (C = 289.2 Da; A = 313.2 Da; G = 329.2 Da; T = 304.2 Da).

misinterpretation of mass signals was greatly reduced. A representative result of a Sanger sequencing mass spectrum, generated using the biotinylated dideoxynucleotides approach, is shown in Figure 5.13. Even today, the read length is still far lower than what can be achieved with current state-of-the-art sequencing equipment. Yet, despite recent developments, MALDI-based sequencing using the Sanger concept never found its way into the production units of genome centers.

207

208

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

Nonetheless, a new series of methods for MALDI-TOF-MS-based sequencing has recently been implemented which rely on the generation of rather short oligonucleotide fragments by complete endonucleolytic cleavage, such that each cleavage product features at least one deined terminal base. This concept of basespeciic cleavage resembles a sequencing method previously developed by Maxam and Gilbert. From a MS point of view, base-speciic cleavage also relates to peptide mapping used for protein identiication, because mass signals are matched to compositional explanations and sequence strings are reconstructed from the explanations. The degradation of a DNA analyte into short oligonucleotide fragments for NA sequence analysis avoids some of the mass accuracy, resolution and sensitivity issues, which are inherent to primer extension methods. All of the methods described in this section require the a priori knowledge of a validated reference sequence, and cannot be used for de-novo sequencing. Nevertheless, they can be applied to experimental questions for DNA-based identiication or resequencing, because in these applications an experimentally determined sequence is cross-compared to a known reference sequence. Although the race to elucidate the human genome sequence is over, the question must still be asked if there remains a need for high-throughput sequencing methods based on MS. As alluded to earlier, genome sequencing projects have made tremendous progress over the past years with both conventional and novel sequencing technologies [194]. The sequence data accumulated to date has led to the discovery of the most abundant type of genetic marker, namely SNPs, and today the details of over four million SNPs have been deposited in databases. This number of markers may prove suficient for current efforts in genome-wide association studies and the elucidation of linkage disequilibrium in the human genome (i.e., to build a map of the most common haplotypes). However, once particular genomic regions have been identiied as being associated with, or are in-linkage with, a speciic phenotypic trait, there is still a need to identify the exact gene leading to the explored phenotype difference. Usually, higher-density marker panels are required within the particular genomic region. Public databases very often do not contain the details of suficient SNPs for higher-density panels, and thus new SNPs must be discovered in the target region and study population. This is usually accomplished by sequencing the target region in a suficient number of individuals (here, the term “suficient” is dependent on the desired allele-frequency of SNP markers in the panel). Once higher-density panels have conirmed any linkage or association and have identiied the disease gene(s), then disease-causing sequence variants (mutations) must be discovered. This usually involves sequencing the respective genes in multiple affected individuals, and has led to project efforts of considerable size. It seems fair to conclude that there remains a tremendous demand for large-scale sequencing in the human genome. Moreover, given the genetic variability encountered, the results of genetic studies will most likely transcend into multiple diagnostic sequencing applications. The continuing demand for high-throughput sequencing may become even more obvious if other organisms are included, especially those with

5.5 Applications in Comparative Sequence Analysis

much faster generation times (as would be necessary in the ield of infection and healthcare). The core of the new methods for MALDI-TOF-MS-based sequencing is a basespeciic cleavage of the ampliication products. Although Hillenkamp and coworkers originally introduced such a concept in 1997 [195], newer approaches have slightly modiied this approach in order to avoid some of the initial issues. Now, a single-stranded copy of the target sequence is usually generated and cleaved to completion in up to four separate base-speciic reactions. Each reaction reduces the original sequence into a set of oligonucleotides, which is readily separated and analyzed by MALDI-TOF-MS. The analysis employs the concept that a set of compomers (combinations of the nucleotides A, C, G, and T) can be assigned to each mass signal of a base-speciic cleavage reaction. The sequence can then be reconstructed from the set of compomers by combining the information of all four base-speciic cleavage reactions and comparing it to a predicted set of mass signals derived from the in-silico cleavage of the supposedly known reference sequence. Whilst several methods have been developed to achieve the base-speciic cleavage of ampliication products, the main difference between them relates to the type of analyte, with base-speciic cleavage occurring either on DNA or RNA. DNA-based methods achieve base-speciic cleavage directly from a DNA ampliication product (such as a PCR product). As discussed in the section on genotyping (see Section 5.4), this can be achieved for example by the incorporation of chemically modiied nucleotides during PCR, which generate site-speciic cleavage sites in further post-PCR reactions [168]. This concept has certain disadvantages when applied to longer target regions, such as those normally used in SNP Discovery or mutation detection. Chemically modiied (non-natural) nucleotides are often incorporated with a lower eficiency during PCR, and this usually has a more profound impact on ampliication yield when products over 500 bp length are to be generated. Another issue relates to the fact that base-speciic cleavage products are generated from both strands. Although this is desirable for genotyping – because it provides redundancy in the analysis and increases accuracy – it can also lead to issues in the analysis of longer amplicons. Without strand separation prior to cleavage, the mass signals of sense and antisense cleavage products may overlap and may compromise the reconstruction of sequence changes. Both of these issues were addressed in an approach described by Shchepinov and coworkers [196], who introduced the use of acid-labile P–N bond nucleotides to obtain base-speciic cleavage. In this case, 5′-phosphoramidate analogs were chosen because they exist as stable triphosphates, they are incorporated by DNA polymerases, and their cleavage proceeds under acidic conditions such that the addition of MALDI matrix will generate site-speciic cleavage products. Due to the instability of nucleotides under PCR cycling conditions and issues with incorporation eficiencies, the process features immobilization of the PCR product to a solid support, thereby generating single-stranded templates; this is then followed by an isothermal primer extension reaction to incorporate the modiied nucleotides. In this way, a speciic strand of the PCR product can be selected for base-speciic cleavage, while the yield of products is not compromised.

209

210

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

An alternative approach that was introduced by two different groups [197, 198] makes use of the natural enzyme repair mechanisms in prokaryotic and eukaryotic cells. The base uracil is exclusive to RNA and is not a part of the DNA alphabet; however, deoxyuracil can be generated when mutagenic chemicals modify DNA, as for example in the deamination of cytosine. DNA glycosylases, such as uracil DNA glycosylase (UDG), recognize uracil residues and remove the base from the DNA strand; this triggers further enzymatic reactions to repair the affected base pair. As a consequence, the incorporation of dUTP (as an analog of dTTP) during an ampliication reaction (PCR) will tag all T positions in the PCR product for attack by UDG. Subsequently, UDG treatment will generate abasic sites (with an exposed phosphate backbone) that can be cleaved under alkaline conditions. This concept has been used on a wider scale for the prevention of PCR contamination in diagnostic settings. In the context of this section, the method can be used to generate T-speciic cleavage patterns of PCR ampliication products, whereby the target region of interest can be ampliied with a nucleotide mix containing dUTP instead of dTTP (full replacement). In the methods introduced by both von Wintzingerode and coworkers and Elso and coworkers, the dUTPcontaining PCR product was immobilized to a solid-support (using perhaps the streptavidin–biotin system), denatured, and the generated single-strand then subjected to UDG treatment as well as backbone cleavage. The resulting T-speciic cleavage products can be analyzed using MALDI-TOF-MS. In principle, the reaction can also be performed with the double-stranded PCR product [198], but this generates cleavage products from both strands simultaneously. As mentioned earlier, the number of potentially coinciding products could lead to issues in the discovery of sequence changes. For applications such as pathogen identiication, however, the simultaneous generation of sense and antisense cleavage patterns could increase the number of discriminatory signals. More recently developed methods have been focused on RNA transcription and RNase cleavage as a means of generating base-speciic cleavage. In two reports that emerged almost simultaneously, G-speciic cleavage patterns were generated with RNase T1 cleavage and used for the identiication of sequence polymorphisms and for the generation of bacteria-speciic mass signal patterns. This concept employs the ampliication of genomic target regions with primer pairs containing promoter tags. The PCR product is then transcribed in separate reactions from forward and reverse directions, and the transcripts are cleaved at every G upon the addition of RNAse T1 [199, 200]. Although, at irst glance, the concept appears complicated, there are some inherent advantages over DNA-based approaches. As discussed for the analysis of microsatellite by MALDI-TOF-MS, the RNA transcription step ampliies the target molecule at least 50- to 100-fold, which means that about 50- to 100-fold more analyte is available for the mass spectrometric analysis and, correspondingly, the SNRs are signiicantly higher. As RNA is less prone to base loss, the spectra are also virtually devoid of base-loss signals. In addition, the stringency for adjusting the appropriate laser luence is reduced, and this simpliies fully automated data acquisition. The use of RNA

5.5 Applications in Comparative Sequence Analysis

transcription has enabled the development of a homogeneous assay format for base-speciic cleavage, whereby all reagents for the subsequent steps are simply added to the PCR product (provided that the starting volume of the PCR reaction is suficiently small that the total volume does not exceed the maximum volume of the microtiter plate wells). The high concentration of analyte allows dilution of the sample prior to sample transfer to a MALDI target with precrystallized matrix. This eliminates the detrimental effect that some of the buffer components of the enzymatic reactions exhibit for the MALDI process. The use of RNAse T1 provides considerable power in the detection of SNPs. G-speciic cleavage reactions performed from both forward and reverse strands allow on aggregate the detection of 80–90% of all possible SNPs within target regions of up to 500 bp length. The exact numbers vary with the sequence context, and will also largely depend on the allelic nature of the SNP (heterozygous or homozygous). In order to increase the sensitivity of base-speciic cleavage for the detection of sequence changes, a scheme featuring cleavage at all four bases is preferred. Recently, Stanssens and coworkers introduced an RNAse A-based concept to achieve base-speciic cleavage at all four bases [201]. RNase A naturally cleaves RNA at every pyrimidine residue (C and U), and cleavage at all pyrimidines would reduce the target sequence to an uninformative mixture of extremely short oligonucleotides. Stanssens and colleagues avoided this issue by rendering RNAse A base-speciic through the incorporation of noncleavable nucleotides during RNA transcription. A mutant of T7 and SP6 RNA polymerase has the ability to incorporate noncanonical nucleotides, such as dNTPs [202] and, correspondingly, the incorporation of dCTP during transcription can be used to obtain U (T)-speciic cleavage. The incorporation of dUTP (or dTTP) leads to a C-speciic cleavage. Yet, virtually all four bases can be covered if C- and T-speciic cleavages are performed on the forward and reverse transcript. Typical basespeciic cleavage spectra, generated through the above-described process, are shown in Figure 5.14. The ability to generate cleavage patterns of virtually all four bases increases the sensitivity of base-speciic cleavage for the detection of sequence changes. An initial simulation using a 4 Mb sequence region surrounding the ApoE gene as a model system revealed that about 99% of all possible single nucleotide changes (substitutions, insertions, deletions) could be detected at amplicon length of 500 bp. Exact numbers vary with amplicon length (they decrease with increasing length of the amplicon) and are sequence context-dependent. The analysis of base-speciic cleavage patterns is signiicantly more complex than the analysis of primer extension products. For the interpretation of spectral changes, the results of the four complementary cleavage spectra must be integrated. In genotyping, it is possible simply to deine the molecular mass of the potential extension products, and to evaluate the presence of mass signals at either a deined mass (or at both, for a heterozygous individual). In base-speciic cleavage, different and mathematically more challenging concepts must be applied, as the simple comparison of the sample spectrum with in-silico spectra from all possible sequence changes would, computationally, not be time-eficient. In principle, a

211

212

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

Figure 5.14 Mass spectra derived from

comparative sequence analysis by complete base-speciic cleavage of ampliication products and analysis of the cleavage products by MALDI-TOF-MS. In this concept the target region is irst ampliied and then cleaved base-speciically into small oligonucleotides. The oligonucleotides are analyzed by mass spectrometry. Sequence changes are identiied by comparing the experimental mass signal pattern to a simulated mass signal pattern derived from a reference sequence. Sequence changes can introduce or remove cleavage sites from the target region and therefore change the mass signal pattern. The nucleobase composition of new mass signals can be derived from their molecular mass, and this allows identiication of the nature and location of the sequence change. Usually, up to four

cleavage reaction are employed for highly conident identiication of sequence changes. Compared to sequencing by MALDI-TOF-MS using the Sanger method (which usually is restricted to a read length of about 25 bases), this approach provides analysis of target regions up to 500 bp in length, or more. In the depicted case, cleavage has been performed at every cytosine and thymine of the forward and reverse strand (in four separate reactions). The dotted lines represent the expected mass signals (as calculated from the reference sequence). In this example, the T-speciic forward cleavage shows a new, non-expected mass signal at around m/z 4250, and one expected mass signal is not detected (missing) at around m/z 3900. This information is computationally deconvoluted to assign a sequence change.

single base substitution can lead to the following observations: (i) it can remove a cleavage site so as to generate a new, larger fragment; (ii) it can introduce a cleavage site generating two new shorter fragments; and (iii) it does not alter a cleavage site, but rather leads to a mass shift in one of the fragments, leading to a new mass signal with either a lower or a higher mass. On aggregate, the combination of four cleavages can result in a maximum of ive such observations for a heterozygous sequence change, and in up to 10 observations for a homozygous sequence change (as the information of a missing predicted signal is also an observation).

5.5 Applications in Comparative Sequence Analysis

Single base insertions and deletions generate a maximum of nine observations in a homozygous sequence change. A concept for the automated analysis of cleavage spectra has recently been described [203]. Given the availability of a reference sequence and a deined method of cleaving the NA amplicon, it is possible to simulate the expected mass signals for each cleavage reaction in an in-silico experiment. The spectra are then evaluated by comparing the in-silico mass signal pattern with the experimental mass signal pattern. For reasons explained below, additional mass signals (deined as signals present in the sample spectrum, but not in the reference spectrum) are selected as indicators for a deviation of the experimental sequence from the reference sequence. The next step is to determine which NA fragments can account for the additional mass signals in a spectrum. Although it is not possible to assign an exact nucleotide sequence to a mass signal in a spectrum, it is possible to calculate the potential compositions of the four nucleotides A/C/G/T (multiplicity of nucleotides, but with unknown order), which could correspond to the mass signal. The compositional analysis of mass signals is challenging if either the number of building blocks of similar mass is high (as for amino acids, the building blocks of proteins), or if the mass accuracy is limiting. However, the simplicity of DNA/RNA keeps the complexity manageable. DNA (and RNA) is comprised of “only” a four-letter alphabet, and additionally, base-speciic cleavage reduces this four-letter alphabet essentially to three. With complete base-speciic cleavage provided, each cleavage product contains at most three of the four possible bases (the cut base being ixed in terms of compositional analysis). According to an early calculation by Pomerantz and coworkers, there exist only a limited number of base-compositions with a mass difference of up to 2 Da [63]. Thus, even with current linear MALDI-TOF-MS, a composition can be assigned to a mass signal and a list generated of base compositions with a mass that is suficiently close to the observed additional signal. The combined list of additional signals and the corresponding compositions can then be mapped back to the reference sequence and used to search the space of sequence variations matching the observed compositional changes. In addition to the challenge of performing this search in a time-eficient manner for all possible single-base substitutions, insertions and deletions, there are further layers of complication. First, the generation of a useful list of compomers for each additional mass signal is greatly simpliied if the mass accuracy is at least ±1 Da. Despite advances in sample preparation, this mass accuracy can only be obtained with internal recalibration. However, the in-silico mass signal pattern predicted from the reference sequence can be used for eficient internal recalibration over the full mass range, as sequence changes will only alter a minority of the mass signals. Additional challenges related to MALDI-TOFMS are the reduced detectivity in the higher mass range (cleavage products with masses exceeding 8000 Da tend to be not detected) and the varying ionization/ desorption behaviors of cleavage products of the same length but with a different composition or sequence. T-rich fragments, for example, tend to show extremely low SNRs compared to signals of other compositions of comparable mass. Both aspects compromise the ability of algorithms to use missing signals (those

213

214

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

predicted from the reference sequence but not observed in the sample spectrum) as a reliable indicator for sequence changes. According to simulations, however, the majority of sequence changes not detectable are related to mass resolution/ accuracy, which means that the corresponding base-compositions cannot be assigned uniquely with linear MALDI-TOF-MS. Additional challenges are posed by the biochemistry of the process, as any biochemical artifact might contribute to the mass signal pattern in an unpredictable manner, leading to a false interpretation (a false-positive indication of a sequence change). Finally, biology and genetics each contribute their share of complication. For example, the longer the target region scanned the more likely it is that multiple sequence changes co-occurring in an amplicon will be observed. The detailed sequence analysis of several gene regions revealed SNP densities as high as one per 200 bp. Mathematical algorithms must be capable of differentiating multiple events per amplicon and sample. The ability to scan larger target regions for sequence changes with MALDI-TOFMS represents a major milestone. Yet, the question must be asked as to how this method compares with state-of-the art competing technology, and what the incentives are to use base-speciic cleavage and MALDI-TOF-MS compared to any other non-MS technology. At a laser repetition rate of 20 Hz, a sample spot can be measured in about 5 s if the sample is rastered up to ive times at different positions. If it is assumed that the average target length scanned is 500 bp, and that four cleavage reactions are needed to identify a maximum of sequence changes, a single mass spectrometer would be able to read over 2 Megabases in a 24-h period. From a throughput standpoint, this compares favorably with state-of-the art capillary sequencers. If the rate at which the biochemical reactions can be processed is ignored, then the limiting factor in current MALDI-TOF instruments will be the laser repetition rate. The use of 200–1000 Hz lasers would shift the bottleneck to other parts of the instrument, though it can be estimated that the throughput would increase to well over 4 Megabases. The real advantage, however, is not necessarily the throughput. More interesting is the combination of both, the speed of signal acquisition/analysis and the availability of molecular mass information (interpreted here as high accuracy). Fast acquisition/analysis times are very important in diagnostic applications (for example, in pathogen identiication) where the sample-in–result-out timeframe must be short, and where massive parallelization cannot substitute for inherently long process times of a technology. Clearly, with a signal acquisition speed in the microsecond range, MALDI-TOF-MS seems well suited to these applications. Additionally – and perhaps even more importantly – base-speciic cleavage offers collateral security since, in a majority of cases, the detection and identiication of sequence changes is based on multiple “observations” (this aspect of the analysis has been described above). On aggregate, the use of four base-speciic cleavage reactions can provide an inherent redundancy of up to 10 observations (ive additional mass signals supporting the sequence change, and ive “missing” signals predicted from the reference sequence), which is an important aspect in diagnostic applications.

5.6 Applications in Quantitation of Nucleic Acids

The principle of base-speciic cleavage signiicantly expands the application portfolio of MALDI-TOF-MS, because larger target regions can now be analyzed. Essentially, a majority of applications currently dominated by capillary electrophoresis separation/detection now become amenable to mass spectrometric analysis. In addition to SNP discovery/resequencing (as discussed above), the use of base-speciic cleavage has also been described for the genotypic identiication of pathogens [204–206]. Further applications may include the qualitative and quantitative analysis of cytosine methylation in genomic DNA [207], the identiication and characterization of complementary DNAs (cDNAs) and their splice variants, for example to identify potential targets for antibody development against cancer cells, and inally also mutation screening [208]. Recent years have also witnessed approaches to increase the apparent mass resolution [85, 209, 210]. Most genotyping assay formats were designed such that the instrumental mass resolution of approximately ±1 Da, available for linear MALDI-TOF instruments, does not compromise the mass accuracy. However, more challenging assays and higher degrees of multiplexing would beneit from increased mass resolution and accuracy. On the one hand, the use of isotopically depleted nucleotides for MALDI-TOFMS represents an option to increase the apparent mass resolution. In combination with linear TOFs, however, the instrumental limitation in mass accuracy is still dominating. A gain in mass resolution by a factor of 2 does not merit the costs associated with the production of isotopically depleted nucleotides. Newer instruments such as orthogonal TOFs and Orbitraps, on the other hand, will alone also not overcome all of the current limitations, because the isotopic envelope of the analytes will be limiting when natural nucleotides are used. Hence, the success of more challenging applications in NA analysis is linked to both, instrumental as well as biochemical improvements, that is, the large-scale use of isotopically depleted nucleotides.

5.6 Applications in Quantitation of Nucleic Acids for Analysis of Gene Expression and Gene Ampliication 5.6.1 Analysis of DNA Mixtures and Allele Frequency Determinations in DNA Pools

Ross and coworkers were the irst to describe the determination of ratios of primer extension products by MALDI-TOF-MS for the analysis of DNA mixtures [211]. This approach opened the door to new means of applying MALDI-TOF-MS to DNA analysis, notably in cases that included quantitative abilities. The main application for this approach – the relative quantitation of allele frequencies – is achieved by calculating the ratio of the peak area associated with allele-speciic extension products. The combined allele frequency of a biallelic SNP in a sample pool (and in an equimolar mixture of genomic DNA from multiple individuals) is

215

216

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

1 (100%). In order to keep the inluence of the sample heterogeneity low, even relative quantiication requires that multiple spectra are averaged from several locations of a given sample. The allele frequency is then expressed as the ratio of the peak areas of any given allele to the total/combined peak area of all alleles. Several groups have assessed the quantitative capabilities of primer extension assays combined with MALDI-TOF-MS [212–216]. It was found that allele frequencies can routinely be measured down to frequencies of 5%. A comparison with other technologies using luorescence or other detection methods showed that the mass spectrometric results were of comparable accuracy and reproducibility. Depending on assay quality, even frequencies below 5% could be detected, though the accuracy of these values was lower as the peak areas would exceed a 50 : 1 ratio. On a routine basis, and using automated processing, frequencies of between 10% and 90% are detected and analyzed with a standard deviation of 2–3%. A detailed analysis of the major contributors to this standard deviations showed that DNA sample generation and ampliication have a greater impact on accuracy and reproducibility than does the heterogeneity of the crystallization and the MALDI-process. This might come as a surprise to specialists in MALDI-TOF-MS, but it must be noted that most reports used miniaturized sample preparation on silicon chips, which has been shown to minimize these effects. Unstable PCR ampliication can have a dramatic impact on the reproducibility of semi-quantitative analysis. However, such an impact is independent of the detection technology, and usually such assays are identiied after multiple reactions and excluded from further consideration. Amexis and coworkers reported an interesting application of semi-quantitative analysis [217] when they employed the MALDI-TOF-MS analysis of primer extension products in their vaccine quality control process. The ratios of viral quasispecies of the mumps virus were determined between Jeryl Lynn substrains in live, attenuated mumps/measles vaccine. Determination of the ratio of two substrains was performed at ive discriminative nucleotide positions within the viral genome. Methods such as this are important for maintaining vaccine safety. Mueller and coworkers used a semi-quantitative analysis of primer extension products by MALDI-TOF-MS to compare the allelic expression between healthy tissue and tumor tissue [218]. The approach irst screened for informative cases (heterozygous individuals) on genomic DNA, and then compared the expression of the corresponding alleles on the mRNA level. In this way, it could be determined that changes in the allelic expression of their gene of interest correlated with tumor development and progression. In particular, the study evaluated the impact of parental imprinting and so-called “loss-of-imprinting” in tumor development. Interestingly, the study results also revealed that previously used technologies mainly had insuficient sensitivity in assessing the effects accurately. Tumor samples normally represent DNA mixtures, which also contain remainders of healthy cells, and this makes their analysis a challenging application for technologies with insuficient sensitivity. The assessment of allele frequencies in DNA pools has been extensively used as a convenient way to validate SNPs and to characterize their allele frequency in different ethnic groups. Some typical results are shown in Figure 5.15. With the

(a)

Genotype vs. Pooled (Uncorrected) Allele Frequencies Genotype vs. Pooled Allele Frequencies Linear (Genotype vs. Pooled Allele Frequencies)

Pooled Allele Frequecy

1.00 R2 = 0.9532 0.80 0.60 0.40 0.20 0.00 0.00

0.20

0.40

0.60

0.80

1.00

Genotyped Allele Frequecy

(b)

Genotype vs. Pooled (Corrected) Allele Frequencies Genotype vs. Pooled Allele Frequencies Linear (Genotype vs. Pooled Allele Frequencies)

Pooled Allele Frequecy

1.00 R2 = 0.9753 0.80 0.60 0.40 0.20 0.00 0.00

0.20

0.40

0.60

0.80

1.00

Genotyped Allele Frequecy

Figure 5.15 Example of the use of semi-

quantitative analysis of mass signal peak area ratios for the measurement of the relative abundance genotypes in a population. The allele-frequencies derived from the relative peak area ratios of SNP-speciic extension products from a pool of individual samples are compared to those determined by individual genotyping of each sample to estimate the precision of the measurement. (a) A scatter plot of genotyped population allele frequencies (x-axis) versus allele frequencies calculated using pooled population DNAs (y-axis). Results for 48 unique assays are shown. The DNA population pool consisted of 96 individual DNAs at equimolar concentrations (260 ng per individual; = 25 ng DNA μl–1). The calculated allele frequency for each assay represents the average of four replicate reactions. The best-it line and the coeficient of determination (R2) were calculated using Excel 2000 (Microsoft). The allele-frequencies derived from pooled DNAs were also

corrected for ampliication bias of individual alleles, which can contribute to differences between the allele-frequencies derived from pools versus the individual genotypes. For this purpose, the allele ratio of individual heterozygote reaction can be used as a “correction” factor for the allele frequencies determined in the pool reaction. The individual heterozygote should have a 0.50 : 0.50 (1 : 1) allele ratio. Any deviation from this expected ratio represents a “skewing” factor in that reaction. After correction with the heterozygote allele ratios, the allele frequencies from the pool reaction match the genotyped population frequency exactly; (b) Scatter plot comparing the genotyped population allele frequencies (x-axis) versus allele frequencies calculated using pooled population DNAs (y-axis) after correction. The coeficient of determination (R2) improves from 0.9532 to 0.9753, and hence leads to an increase in the precision of allele-frequency estimates.

218

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

rapid assay development capacity of primer extension-based assays, this approach allows for a large-scale initial discrimination between “real” and “false” SNPs derived from databases, and also allows for a discrimination between common and rare SNPs. In order to exploit this approach in large-scale association studies, it was necessary to verify the accuracy of allele frequency estimation in DNA pools by MALDITOF-MS, by comparison with the true allele frequency derived from individual genotyping. A scatter plot with allele frequencies determined for 24 assays in a DNA pool of 96 individuals versus the “true” allele frequency of the same assays observed in individual genotyping, is shown in Figure 5.15a. These data illustrate the uncorrected dataset, and reveal that the correlation between allele frequencies estimated from a DNA pool and the frequencies determined by individual genotyping is not perfect. Numerous factors may contribute to this effect, the majority of which are technology-independent, such as the accuracy with which the DNA pool was prepared or the preferential ampliication of one of the alleles over the other during PCR. Due to the importance of allele frequency estimation in genetic studies, recent reviews have provided a detailed account of strategies and issues for the analysis of DNA pools. It has been frequently found that individual DNAs, heterozygous for a particular SNP, have shown “skewed” distributions of the two alleles. Genetically, there is a 1 : 1 distribution of the two genetic informations (one chromosome carrying one allele, and the second chromosome carrying the other allele), and consequently a 1 : 1 ratio of the two alleles in the SNP assay (or a 50% frequency for both alleles in terms of peak areas) would be expected. This skewing, which is most likely introduced during PCR (by the preferential ampliication of one allele) or during the post-PCR primer extension, leads to a deviation in the estimated allele frequency from the true frequency when the DNA pools are analyzed. A further MALDI-speciic factor might be a slightly lower desorption/ionization eficiency of the higher-mass allele versus the lower-mass allele. However, it is possible to apply a “correction” factor for each individual assay, based on the peak area distributions observed in individual heterozygous DNAs. Deviations from the expected 1 : 1 ratio can be incorporated into a correction factor, which should then be applied to the allele frequency estimations from pools. Today, allele frequency estimations have been corrected by assay-speciic correction factors such that, with the correction applied, the coeficient of correlation is improved (see Figure 5.14b). Although some assay-speciic deviations between true and estimated frequencies remain, the approach has been successfully applied to semi-quantitative SNP allele frequency analysis in sample pools, and also to differential protein binding to mRNA associated with allelic variants of a gene [219]. As noted above, allele frequency information derived from DNA pools not only allows the rapid collection of validated SNP sets; rather, a comparison of the abundance of alleles between different populations represented as DNA pools also offers a suitable means of identifying genotype–phenotype correlations. In this respect, SNPs are used as genetic markers, allowing the identiication of causative

5.6 Applications in Quantitation of Nucleic Acids

genetic loci in complex diseases through the linkage between the SNP marker and the genetic locus (the SNP being in linkage disequilibrium with the causative genetic locus). The use of DNA pools has also been recognized as a potential short-cut to identifying associations between genetic loci and phenotypes [220, 221]. Instead of the costly and cumbersome individual genotyping of hundred thousands of SNPs, the allele frequency of SNPs in various DNA pools stratiied by phenotype can be employed to ind such an association. For this approach, the correctness of the allele frequency estimation compared to the “true” frequency is less of an issue, as long as the relative abundance between the two pools can be assessed accurately and with high reproducibility. Buetow and coworkers were the irst to apply allele-frequency estimations by MALDI-TOF-MS in DNA pools on a genome-wide basis [216]. Recently, several other groups have dissected the major process variables from ampliication to MALDI-TOF-MS analysis and have provided proof-of-principle that the approach allows for the identiication of major genetic contributors to complex diseases [212–215, 222]. Recently, Ding et al. took these applications one step further when they combined the sensitivity and speciicity of the combination of PCR and analysis by MALDI-TOF-MS (which they had demonstrated with the molecular haplotyping approach) with the quantitative features established in the experiments described above. In this way, it was possible to demonstrate an analysis of mutations in circulating NAs which, ultimately, proved to be an important milestone towards noninvasive prenatal diagnostics [223]. 5.6.2 Analysis of Gene Expression

The ield of gene expression analysis has only recently become of broader interest to users of mass spectrometry. The reason for this lag in interest was simply the lack of methods allowing the absolute quantiication of NAs in general, or mRNA in particular. Mass spectrometry only allows an endpoint analysis, and even then a standard for normalization (such as a second allele) is required. Hence, the approaches used have focused on the semi-quantitative analysis of allelic expression which, as described above, is a rather narrow – albeit expanding – ield of research. The irst attempts to enable gene expression analysis by MS were reported only recently, when Smith and coworkers applied the Invader assay (see Section 5.4.1.5 and Figure 5.12) to the detection and relative quantiication of RNA [224]. In order to allow an inter-spectrum (and thus inter-sample) correlation, Invader assays of multiple target genes were multiplexed and a reference gene was included for normalization of the data. In a different approach, Ding and Cantor recently described the use of an internal standard added to cDNA samples to enable the quantiication of NAs with technologies lacking real-time monitoring capabilities (as are employed for realtime PCR) [225]. This approach uses a synthetic oligonucleotide designed to match

219

220

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

the target sequence (i.e., a sequence stretch of the cDNA to be investigated) in all positions, except for a single nucleotide. An internal standard is added to the sample of interest (i.e., a cDNA preparation) at a known concentration prior to ampliication (hence, the process is called “competitive PCR”). The introduction of a single nucleotide change allows a differentiation to be made between the target cDNA and the internal standard by means of a post-PCR primer extension assay. Although the internal standard and cDNA are coampliied during the PCR, the eficiency of ampliication should be equivalent for both as they share very nearly the same sequence and, in particular, the same primer binding sites. The post-PCR primer extension reaction targets the nucleotide difference between the internal standard and the cDNA, and generates two speciic products that resemble the same reaction performed for allele-frequency determination. Analysis of the peak areas allows not only for a relative comparison of cDNA amount versus internal standard, but also for an absolute quantiication, as the concentration of the internal standard is known. The use of an internal standard with the same sequence as the target sequence alleviates common issues related to quantiication. Due to the same PCR ampliication eficiency, the process is PCR-cycle-independent, and real-time monitoring of the ampliication process becomes obsolete, as it will impact on both the internal standard and the cDNA in the same way. The results of a model system used to establish and validate this concept are shown in Figure 5.16. Here, a 90 bp region of the cholesterol ester transfer protein (CETP) gene region was selected to demonstrate the principle. Two sequences were designed: one which exactly matched the gene sequences that served as the unknown, and one which carried a single base change (C→G substitution) in the central region of the oligonucleotide that served as the internal standard. Both molecules were used as templates for competitive PCR in various template ratios. Post-PCR primer extension and MALDI-TOF-MS were used to track the relative ratio of the two standards after PCR ampliication. The observed allele ratios derived from the standards, compared to the expected allele ratios as calculated from the mixture used for PCR, are shown in Figure 5.16. The uncorrected observed allele ratio was shown to track the expected ratio closely, which supports the validity of the approach; also notable was the low standard deviation. All data points were acquired in triplicate (on the PCR level), and the values of standard deviations resembled those obtained for allele-frequency determination in DNA pools (∼3%). Clearly, if these standard deviations were also to be obtained for multiple studies carried out in different laboratories, then the method would allow for the accurate analysis of rather subtle changes in expression analysis. The description by Ding and Cantor is very recent, and further use of the proposed method in various laboratories will demonstrate the utility and accuracy of the approach. The competitor PCR approach described by Ding and Cantor against real-time PCR (rtPCR) was validated in a more recent report from Elvidge et al., who demonstrated a good correlation between real-time PCR data and MALDITOF-MS data [226]. There was, however, a signiicant difference in sensitivity, with low-abundant mRNAs being much more readily detected and quantiied by the

5.6 Applications in Quantitation of Nucleic Acids CETP gene observed ss-template ratios vs expected

Relative Ratio (observed)

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.0

0.2

0.4

0.6

0.8

1.0

Relative Ratio (expected) Figure 5.16 Quantitation of gene expression

by PCR/primer extension and MALDI-TOFMS in a mixture model system. Depicted is a scatter plot of calculated allele ratios (based on mass signal peak area rations of primer extension products) for expression analysis of the CETP gene. Data points represent the average of triplicate reactions. Expected ratios are depicted as a solid line, and observed values as dots. In this model system two artiicial templates (90 bp) were designed based on the sequence of the CETP gene mRNA (Accession# AC023825). One of the templates matched this region of the CETP gene exactly. The second had a 1 bp

mismatch introduced to mimic a mutation and to serve as a second allele in a primer extension reaction. Each template and allele is coampliied at equal rates, as shown in the graph. Deviation from an exact it to expected allele frequencies represent a “skew” (as detailed in Figure 5.9), and can be corrected in the same manner using a heterozygote (in this case artiicial). The concentration of each template added to the reaction is known and therefore the amount of wild-type mRNA (or cDNA) can be determined when the two alleles are at a 1 : 1 ratio (0.5 : 0.5 allele frequency).

combination of rcPCR and MS. It can be envisioned that this approach will ultimately be applied to studies with a focused interest in the expression of a set of genes, because the combination of an internal standard with PCR and primer extension can be easily multiplexed when MALDI-TOF-MS is used for the analysis. Thus, massive amounts of data comparing relative and absolute expression of multiple genes can be generated for large sample sets. One drawback of conventional microarray-based gene expression analysis is that it delivers a large number of data points (or even genome-wide data points) for individual samples. However, once genes or gene sets of interest have been selected, the microarrays do not permit the scanning of a large number of samples in a cost-eficient way. Elvidge et al. also noted in their report that the development of real-time PCR assays for expression analysis may more frequently require time-consuming – and potentially costly – assay optimization rounds. The basic concept described by Ding et al. has recently been extended to exciting new aspects of gene expression, including the quantitation of allele-speciic expression and the quantitative analysis of splice variants [227, 228].

221

222

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

The use of internal standards and competitive PCR is not limited to gene expression analysis. Clearly, there are many more scientiic questions which require the absolute quantiication of NAs, and the concept of competitive PCR will surely help to open these applications to mass spectrometric analysis.

5.7 Future Perspectives for the MALDI-MS Analysis of Nucleic Acids

The applications described here have clearly shown that MALDI analysis can contribute more than just sequence information. It is true that, in the sequencing arena, this analytical platform cannot compete in terms of throughput and viable sequence lengths with approaches that rely on strand ampliication and electrophoretic analysis or, even more so, with other sequencing techniques of the second or third generation. This limitation may, arguably, have prevented MS-based technologies from assuming the same prominent role in genetics and genomics, which they have assumed in proteomics research. During recent years, however, it has become evident that sequence information alone cannot reveal the function of all known NAs. The discovery of small interfering RNA (siRNA) and riboswitches has led to the realization that three-dimensional structure can be at least as important as sequence in determining function. Furthermore, recent estimates that more than 98% of the genome sequence does not code for actual proteins have awakened a keen interest for approaches capable of linking sequence to structure and then function, which will be paramount for annotating these noncoding regions [229]. The versatility displayed by MALDI in the applications reviewed here, in addition to its typical sensitivity, speciicity and speed, ensures that this technology will have important roles to play in the elucidation of noncoding elements and in the identiication of their possible cognate proteins in biological systems. The fact that strand-ampliication approaches are not capable of “copying” modiied nucleotides – owing to the absence of appropriate complementary bases for faithful replication – represents an excellent opportunity for applications aimed at the identiication and quantiication of post-transcriptional modiications. It could be argued that the nearly exclusive reliance on strand-ampliication approaches may thus far have hampered a correct evaluation of the biological signiicance and diagnostic value of such modiications. The sensitivity afforded by MALDI, and its applicability to modiications of virtually any type, will be paramount to support the direct analysis of genuine NA samples that were extracted directly from cells and did not undergo any type of ampliication. Considering the roles of post-translational modiications in proteins, in which they perform a broad range of structural, catalytic, and regulatory functions, the investigation of posttranscriptional modiications in NAs will likely brim with surprises. Finally, the possibility of including MALDI analysis in complex analytical schemes ensures that this platform will remain an excellent research tool for academic, as well as corporate, laboratories. Furthermore, as NA species become

References

more popular not only as possible targets for drug discovery but also as therapeutics in their own right, MALDI will become increasingly involved in early discovery stages, drug metabolism, production support, and quality control procedures.

Acknowledgments

The authors are grateful to Jingue Ju for providing Figure 5.12. They would also like to thank Christian Jurinke, Christiane Honisch, Mathias Ehrich, Paul Oeth and April Kinsler for their help in the preparation of this contribution, and valuable comments.

References 1 Fenn, J.B., Mann, M., Meng, C.K.,

2

3

4

5

6

Wong, S.F., and Whitehouse, C.M. (1989) Electrospray ionization for mass spectrometry of large biomolecules. Science, 246, 246. Karas, M. and Hillenkamp, F. (1988) Laser desorption ionization of proteins with molecular Masses exceeding 10000 daltons. Anal. Chem., 60 (20), 2299–2301. Muddiman, D.C., Null, A.P., and Hannis, J.C. (1999) Precise mass measurement of a double-stranded 500 base-pair (309 kDa) polymerase chain reaction product by negative ion electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. Rapid Commun. Mass Spectrom., 13, 1201–1204. Hannis, J.C. and Muddiman, D.C. (2001) Detection of double-stranded PCR amplicons at the attomol level electrosprayed from low nanomolar solutions using FTICR mass spectrometry. Fresenius J. Anal. Chem., 369, 246–251. Hahner, S., Schneider, A., Ingendoh, A., and Mosner, J. (2000) Analysis of short tandem repeat polymorphisms by electrospray ion trap mass spectrometry. Nucleic Acids Res., 28, 18. Laken, S.J., Jackson, P.E., Kinzler, K.W., Vogelstein, B., Strickland, P.T., Groopman, J.D., and Friesen, M.D. (1998) Genotyping by mass spectrometric analysis of short DNA

7

8

9

10

11

fragments. Nat. Biotechnol., 16 (13), 1352–1356. Jiang, Y. and Hofstadler, S.A. (2003) A highly eficient and automated method of purifying and desalting PCR products for analysis by electrospray ionization mass spectrometry. Anal. Biochem., 316 (1), 50–57. Doktycz, M.J., Hurst, G.B., HabibiGoudarzi, S., McLuckey, S.A., Tang, K., Chen, C.H., Uziel, M., Jacobson, K.B., Woychik, R.P., and Buchanan, M.V. (1995) Analysis of polymerase chain reaction-ampliied DNA products by mass spectrometry using matrix-assisted laser desorption and electrospray: current status. Anal. Biochem., 230 (2), 205–214. McLuckey, S.A., Wu, J., Bundy, J.L., Stephenson, J.L., Jr, and Hurst, G.B. (2002) Oligonucleotide mixture analysis via electrospray and ion/ion reactions in a quadrupole ion trap. Anal. Chem., 74 (5), 976–984. Nordhoff, E., Cramer, R., Karas, M., Hillenkamp, F., Kirpekar, F., Kristiansen, K., and Roepstorff, P. (1993) Ion stability of nucleic acids in infrared matrix-assisted laser desorption/ ionization mass spectrometry. Nucleic Acids Res., 21 (15), 3347–3357. Wu, K.J., Steding, A., and Becker, C.H. (1993) Matrix-assisted laser desorption time-of-light mass spectrometry of oligonucleotides using 3-hydroxypicolinic acid as an ultraviolet-

223

224

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

12

13

14

15

16

17

18

sensitive matrix. Rapid Commun. Mass Spectrom., 7 (2), 142–146. Nordhoff, E., Kirpekar, F., Karas, M., Cramer, R., Hahner, S., Hillenkamp, F., Kristiansen, K., Roepstorff, P., and Lezius, A. (1994) Comparison of IR- and UV-matrix-assisted laser desorption/ ionization mass spectrometry of oligodeoxynucleotides. Nucleic Acids Res., 22 (13), 2460–2465. Zhu, Y.F., Chung, C.N., Taranenko, N.I., Allman, S.L., Martin, S.A., Haff, L., and Chen, C.H. (1996) The study of 2,3,4-trihydroxyacetophenone and 2,4,6-trihydroxyacetophenone as matrices for DNA detection in matrixassisted laser desorption/ionization time-of-light mass spectrometry. Rapid Commun. Mass Spectrom., 10 (3), 383–388. Gross, J., Hillenkamp, F., Wan, K.X., and Gross, M.L. (2001) Metastable decay of negatively charged oligodeoxynucleotides analyzed with ultraviolet matrix-assisted laser desorption/ionization post-source decay and deuterium exchange. J. Am. Soc. Mass Spectrom., 12 (2), 180–192. Gross, J., Leisner, A., Hillenkamp, F., Hahner, S., Karas, M., Schafer, J., Lutzenkirchen, F., and Nordhoff, E. (1998) Investigations of the metastable decay of DNA under ultraviolet matrix-assisted laser desorption/ ionization conditions with post-sourcedecay analysis and hydrogen/deuterium exchange. J. Am. Soc. Mass Spectrom., 9 (9), 866–878. Kirpekar, F., Nordhoff, E., Kristiansen, K., Roepstorff, P., Lezius, A., Hahner, S., Karas, M., and Hillenkamp, F. (1994) Matrix assisted laser desorption/ ionization mass spectrometry of enzymatically synthesized RNA up to 150 kDa. Nucleic Acids Res., 22 (19), 3866–3870. Tang, K., Taranenko, N.I., Allman, S.L., Chang, L.Y., and Chen, C.H. (1994) Detection of 500-nucleotide DNA by laser desorption mass spectrometry. Rapid Commun. Mass Spectrom., 8 (9), 727–730. Lin, H., Hunter, J.M., and Becker, C.H. (1999) Laser desorption of DNA

19

20

21

22

23

24

25

26

oligomers larger than one kilobase from cooled 4-nitrophenol. Rapid Commun. Mass Spectrom., 13 (23), 2335–2340. Schneider, K. and Chait, B.T. (1993) Matrix-assisted laser desorption mass spectrometry of homopolymer oligodeoxynucleotides. Inluence of base composition on the mass spectrometric response. Org. Mass Spectrom., 28 (11), 1353–1361. Schneider, K. and Chait, B.T. (1995) Increased stability of nucleic acids containing 7-deaza-guanosine and 7-deaza-adenosine may enable rapid DNA sequencing by matrix-assisted laser desorption mass spectrometry. Nucleic Acids Res., 23 (9), 1570–1575. Cotter, R.J. (1997) Time-of-Flight Mass Spectrometry. ACS Professional Reference Books, American Chemical Society, Washington, DC, USA. Berkenkamp, S., Menzel, C., Karas, M., and Hillenkamp, F. (1997) Performance of infrared matrix-assisted laser desorption/ ionization mass spectrometry with lasers emitting in the 3 μm wavelength range. Rapid Commun. Mass Spectrom., 11 (13), 1399–1406. Menzel, C., Berkenkamp, S., and Hillenkamp, F. (1999) Infrared matrix-assisted laser desorption/ ionization mass spectrometry with a transversely excited atmospheric pressure carbon dioxide laser at 10.6 um wavelength with static and delayed ion extraction. Rapid Commun. Mass Spectrom., 13 (1), 26–32. Berkenkamp, S., Kirpekar, F., and Hillenkamp, F. (1998) Infrared MALDI mass spectrometry of large nucleic acids. Science, 281 (5374), 260–262. Kirpekar, F., Berkenkamp, S., and Hillenkamp, F. (1999) Detection of double-stranded DNA by IR- and UV-MALDI mass spectrometry. Anal. Chem., 71 (13), 2334–2339. Venter, J.C., Adams, M.D., Myers, E.W., Li, P.W., Mural, R.J., Sutton, G.G., Smith, H.O., Yandell, M., Evans, C.A., Holt, R.A., Gocayne, J.D., Amanatides, P., Ballew, R.M., Huson, D.H., Wortman, J.R., Zhang, Q., Kodira, C.D., Zheng, X.H., Chen, L., Skupski, M., Subramanian, G., Thomas, P.D., Zhang,

References J., Gabor Miklos, G.L., Nelson, C., Broder, S., Clark, A.G., Nadeau, J., McKusick, V.A., Zinder, N., Levine, A.J., Roberts, R.J., Simon, M., Slayman, C., Hunkapiller, M., Bolanos, R., Delcher, A., Dew, I., Fasulo, D., Flanigan, M., Florea, L., Halpern, A., Hannenhalli, S., Kravitz, S., Levy, S., Mobarry, C., Reinert, K., Remington, K., AbuThreideh, J., Beasley, E., Biddick, K., Bonazzi, V., Brandon, R., Cargill, M., Chandramouliswaran, I., Charlab, R., Chaturvedi, K., Deng, Z., Di Francesco, V., Dunn, P., Eilbeck, K., Evangelista, C., Gabrielian, A.E., Gan, W., Ge, W., Gong, F., Gu, Z., Guan, P., Heiman, T.J., Higgins, M.E., Ji, R.R., Ke, Z., Ketchum, K.A., Lai, Z., Lei, Y., Li, Z., Li, J., Liang, Y., Lin, X., Lu, F., Merkulov, G.V., Milshina, N., Moore, H.M., Naik, A.K., Narayan, V.A., Neelam, B., Nusskern, D., Rusch, D.B., Salzberg, S., Shao, W., Shue, B., Sun, J., Wang, Z., Wang, A., Wang, X., Wang, J., Wei, M., Wides, R., Xiao, C., Yan, C., Yao, A., Ye, J., Zhan, M., Zhang, W., Zhang, H., Zhao, Q., Zheng, L., Zhong, F., Zhong, W., Zhu, S., Zhao, S., Gilbert, D., Baumhueter, S., Spier, G., Carter, C., Cravchik, A., Woodage, T., Ali, F., An, H., Awe, A., Baldwin, D., Baden, H., Barnstead, M., Barrow, I., Beeson, K., Busam, D., Carver, A., Center, A., Cheng, M.L., Curry, L., Danaher, S., Davenport, L., Desilets, R., Dietz, S., Dodson, K., Doup, L., Ferriera, S., Garg, N., Gluecksmann, A., Hart, B., Haynes, J., Haynes, C., Heiner, C., Hladun, S., Hostin, D., Houck, J., Howland, T., Ibegwam, C., Johnson, J., Kalush, F., Kline, L., Koduru, S., Love, A., Mann, F., May, D., McCawley, S., McIntosh, T., McMullen, I., Moy, M., Moy, L., Murphy, B., Nelson, K., Pfannkoch, C., Pratts, E., Puri, V., Qureshi, H., Reardon, M., Rodriguez, R., Rogers, Y.H., Romblad, D., Ruhfel, B., Scott, R., Sitter, C., Smallwood, M., Stewart, E., Strong, R., Suh, E., Thomas, R., Tint, N.N., Tse, S., Vech, C., Wang, G., Wetter, J., Williams, S., Williams, M., Windsor, S., Winn-Deen, E., Wolfe, K., Zaveri, J., Zaveri, K., Abril, J.F., Guigo, R., Campbell, M.J., Sjolander, K.V.,

Karlak, B., Kejariwal, A., Mi, H., Lazareva, B., Hatton, T., Narechania, A., Diemer, K., Muruganujan, A., Guo, N., Sato, S., Bafna, V., Istrail, S., Lippert, R., Schwartz, R., Walenz, B., Yooseph, S., Allen, D., Basu, A., Baxendale, J., Blick, L., Caminha, M., Carnes-Stine, J., Caulk, P., Chiang, Y.H., Coyne, M., Dahlke, C., Mays, A., Dombroski, M., Donnelly, M., Ely, D., Esparham, S., Fosler, C., Gire, H., Glanowski, S., Glasser, K., Glodek, A., Gorokhov, M., Graham, K., Gropman, B., Harris, M., Heil, J., Henderson, S., Hoover, J., Jennings, D., Jordan, C., Jordan, J., Kasha, J., Kagan, L., Kraft, C., Levitsky, A., Lewis, M., Liu, X., Lopez, J., Ma, D., Majoros, W., McDaniel, J., Murphy, S., Newman, M., Nguyen, T., Nguyen, N., Nodell, M., Pan, S., Peck, J., Peterson, M., Rowe, W., Sanders, R., Scott, J., Simpson, M., Smith, T., Sprague, A., Stockwell, T., Turner, R., Venter, E., Wang, M., Wen, M., Wu, D., Wu, M., Xia, A., Zandieh, A., and Zhu, X. (2001) The sequence of the human genome. Science, 291 (5507), 1304–1351. 27 Lander, E.S., Linton, L.M., Birren, B., Nusbaum, C., Zody, M.C., Baldwin, J., Devon, K., Dewar, K., Doyle, M., FitzHugh, W., Funke, R., Gage, D., Harris, K., Heaford, A., Howland, J., Kann, L., Lehoczky, J., LeVine, R., McEwan, P., McKernan, K., Meldrim, J., Mesirov, J.P., Miranda, C., Morris, W., Naylor, J., Raymond, C., Rosetti, M., Santos, R., Sheridan, A., Sougnez, C., Stange-Thomann, N., Stojanovic, N., Subramanian, A., Wyman, D., Rogers, J., Sulston, J., Ainscough, R., Beck, S., Bentley, D., Burton, J., Clee, C., Carter, N., Coulson, A., Deadman, R., Deloukas, P., Dunham, A., Dunham, I., Durbin, R., French, L., Grafham, D., Gregory, S., Hubbard, T., Humphray, S., Hunt, A., Jones, M., Lloyd, C., McMurray, A., Matthews, L., Mercer, S., Milne, S., Mullikin, J.C., Mungall, A., Plumb, R., Ross, M., Shownkeen, R., Sims, S., Waterston, R.H., Wilson, R.K., Hillier, L.W., McPherson, J.D., Marra, M.A., Mardis, E.R., Fulton, L.A., Chinwalla, A.T., Pepin, K.H., Gish, W.R., Chissoe, S.L., Wendl, M.C., Delehaunty, K.D.,

225

226

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics Miner, T.L., Delehaunty, A., Kramer, J.B., Cook, L.L., Fulton, R.S., Johnson, D.L., Minx, P.J., Clifton, S.W., Hawkins, T., Branscomb, E., Predki, P., Richardson, P., Wenning, S., Slezak, T., Doggett, N., Cheng, J.F., Olsen, A., Lucas, S., Elkin, C., Uberbacher, E., Frazier, M., Gibbs, R.A., Muzny, D.M., Scherer, S.E., Bouck, J.B., Sodergren, E.J., Worley, K.C., Rives, C.M., Gorrell, J.H., Metzker, M.L., Naylor, S.L., Kucherlapati, R.S., Nelson, D.L., Weinstock, G.M., Sakaki, Y., Fujiyama, A., Hattori, M., Yada, T., Toyoda, A., Itoh, T., Kawagoe, C., Watanabe, H., Totoki, Y., Taylor, T., Weissenbach, J., Heilig, R., Saurin, W., Artiguenave, F., Brottier, P., Bruls, T., Pelletier, E., Robert, C., Wincker, P., Smith, D.R., Doucette-Stamm, L., Rubenield, M., Weinstock, K., Lee, H.M., Dubois, J., Rosenthal, A., Platzer, M., Nyakatura, G., Taudien, S., Rump, A., Yang, H., Yu, J., Wang, J., Huang, G., Gu, J., Hood, L., Rowen, L., Madan, A., Qin, S., Davis, R.W., Federspiel, N.A., Abola, A.P., Proctor, M.J., Myers, R.M., Schmutz, J., Dickson, M., Grimwood, J., Cox, D.R., Olson, M.V., Kaul, R., Shimizu, N., Kawasaki, K., Minoshima, S., Evans, G.A., Athanasiou, M., Schultz, R., Roe, B.A., Chen, F., Pan, H., Ramser, J., Lehrach, H., Reinhardt, R., McCombie, W.R., de la Bastide, M., Dedhia, N., Blocker, H., Hornischer, K., Nordsiek, G., Agarwala, R., Aravind, L., Bailey, J.A., Bateman, A., Batzoglou, S., Birney, E., Bork, P., Brown, D.G., Burge, C.B., Cerutti, L., Chen, H.C., Church, D., Clamp, M., Copley, R.R., Doerks, T., Eddy, S.R., Eichler, E.E., Furey, T.S., Galagan, J., Gilbert, J.G., Harmon, C., Hayashizaki, Y., Haussler, D., Hermjakob, H., Hokamp, K., Jang, W., Johnson, L.S., Jones, T.A., Kasif, S., Kaspryzk, A., Kennedy, S., Kent, W.J., Kitts, P., Koonin, E.V., Korf, I., Kulp, D., Lancet, D., Lowe, T.M., McLysaght, A., Mikkelsen, T., Moran, J.V., Mulder, N., Pollara, V.J., Ponting, C.P., Schuler, G., Schultz, J., Slater, G., Smit, A.F., Stupka, E., Szustakowski, J., ThierryMieg, D., Thierry-Mieg, J., Wagner, L., Wallis, J., Wheeler, R., Williams, A.,

28

29

30

31

32

33

34

35

Wolf, Y.I., Wolfe, K.H., Yang, S.P., Yeh, R.F., Collins, F., Guyer, M.S., Peterson, J., Felsenfeld, A., Wetterstrand, K.A., Patrinos, A., Morgan, M.J., Szustakowki, J., de Jong, P., Catanese, J.J., Osoegawa, K., Shizuya, H., Choi, S., and Chen, Y.J. (2001) Initial sequencing and analysis of the human genome. Nature, 409 (6822), 860–921. Hinds, D.A., Kloek, A.P., Jen, M., Chen, X., and Frazer, K.A. (2006) Common deletions and SNPs are in linkage disequilibrium in the human genome. Nat. Genet., 38 (1), 82–85. Feuk, L., Carson, A.R., and Scherer, S.W. (2006) Structural variation in the human genome. Nat. Rev. Genet., 7 (2), 85–97. de Martinville, B., Wyman, A.R., White, R., and Francke, U. (1982) Assignment of irst random restriction fragment length polymorphism (RFLP) locus (D14S1) to a region of human chromosome 14. Am. J. Hum. Genet., 34 (2), 216–226. Taylor, G.R., Noble, J.S., Hall, J.L., Stewart, A.D., and Mueller, R.F. (1989) Hypervariable microsatellite for genetic diagnosis. Lancet, 2 (8660), 454. Epplen, C., Buitkamp, J., Rumpf, H., D’Souza, M., and Epplen, J.T. (1995) Immunoprinting reveals different genetic bases for (auto)immuno diseases. Electrophoresis, 16 (9), 1693–1697. Sunden, S.L., Businga, T., Beck, J., McClain, A., Gastier, J.M., Pulido, J.C., Yandava, C.N., Brody, T., Ghazizadeh, J., Weber, J.L., Duyk, G.M., Murray, J.C., Buetow, K.H., and Shefield, V.C. (1996) Chromosomal assignment of 2900 triand tetranucleotide repeat markers using NIGMS somatic cell hybrid panel 2. Genomics, 32 (1), 15–20. Dib, C., Faure, S., Fizames, C., Samson, D., Drouot, N., Vignal, A., Millasseau, P., Marc, S., Hazan, J., Seboun, E., Lathrop, M., Gyapay, G., Morissette, J., and Weissenbach, J. (1996) A comprehensive genetic map of the human genome based on 5,264 microsatellites. Nature, 380 (6570), 152–154. Ziegle, J.S., Su, Y., Corcoran, K.P., Nie, L., Mayrand, P.E., Hoff, L.B., McBride,

References

36

37

38

39

40

41

42

43

L.J., Kronick, M.N., and Diehl, S.R. (1992) Application of automated DNA sizing technology for genotyping microsatellite loci. Genomics, 14 (4), 1026–1031. Reed, P.W., Davies, J.L., Copeman, J.B., Bennett, S.T., Palmer, S.M., Pritchard, L.E., Gough, S.C., Kawaguchi, Y., Cordell, H.J., Balfour, K.M., et al. (1994) Chromosome-speciic microsatellite sets for luorescence-based, semi-automated genome mapping. Nat. Genet., 7 (3), 390–395. Kruglyak, L. and Nickerson, D.A. (2001) Variation is the spice of life. Nat. Genet., 27 (3), 234–236. Stephens, J.C., Schneider, J.A., Tanguay, D.A., Choi, J., Acharya, T., Stanley, S.E., Jiang, R., Messer, C.J., Chew, A., Han, J.H., Duan, J., Carr, J.L., Lee, M.S., Koshy, B., Kumar, A.M., Zhang, G., Newell, W.R., Windemuth, A., Xu, C., Kalbleisch, T.S., Shaner, S.L., Arnold, K., Schulz, V., Drysdale, C.M., Nandabalan, K., Judson, R.S., Ruano, G., and Vovis, G.F. (2001) Haplotype variation and linkage disequilibrium in 313 human genes. Science, 293 (5529), 489–493. Reich, D.E., Gabriel, S.B., and Altshuler, D. (2003) Quality and completeness of SNP databases. Nat. Genet., 33 (4), 457–458. Risch, N. and Merikangas, K. (1996) The future of genetic studies of complex human diseases. Science, 273 (5281), 1516–1517. Lander, E.S. (1996) The new genomics: global views of biology. Science, 274 (5287), 536–539. Altshuler, D., Pollara, V.J., Cowles, C.R., Van Etten, W.J., Baldwin, J., Linton, L., and Lander, E.S. (2000) An SNP map of the human genome generated by reduced representation shotgun sequencing. Nature, 407 (6803), 513–516. O’Donovan, M.C., Oefner, P.J., Roberts, S.C., Austin, J., Hoogendoorn, B., Guy, C., Speight, G., Upadhaya, M., Sommer, S.S., and McGufin, P. (1998) Blind analysis of denaturing high-performance liquid chromatography as a tool for mutation detection. Genomics, 52, 44–49.

44 Glavac, D. and Dean, M. (1995)

45

46

47

48

49

50

51

52

53

54

Applications of heteroduplex analysis for mutation detection in disease genes. Hum. Mutat., 6, 281–286. Orita, M., Iwahana, H., Kanazawa, H., Hayashi, K., and Sekiya, T. (1989) Detection of polymorphisms of human DNA by gel electrophoresis as singlestrand conformation polymorphisms. Proc. Natl Acad. Sci. USA, 86, 2766–2770. Hacia, J.G., Brody, L.C., Chee, M.S., Fodor, S.P., and Collins, F.S. (1996) Detection of heterozygous mutations in BRCA1 using high-density oligonucleotide arrays and two-color luorescence analysis. Nat. Genet., 14, 441–447. Hacia, J.G. (1999) Resequencing and mutational analysis using oligonucleotide arrays. Nat. Genet., 21, 42–47. Kwok, P.Y., Carlson, C., Yager, T.D., Ankener, W., and Nickerson, D.A. (1994) Comparative analysis of human DNA variations by luorescence-based sequencing of PCR products. Genomics, 23, 138–144. Couzin, J. (2002) Human genome. HapMap launched with pledges of $100 million. Science, 298 (5595), 941–942. Farmer, P.B. and Sweetman, G.M.A. (1995) Mass spectrometric detection of carcinogen adducts. J. Mass Spectrom., 30, 1369–1379. Limbach, P.A., Crain, P.F., and McCloskey, J.A. (1994) Summary: the modiied nucleosides of RNA. Nucleic Acids Res., 22 (12), 2183–2196. Cantara, W.A., Crain, P.F., Rozenski, J., McCloskey, J.A., Harris, K.A., Zhang, X., Vendeix, F.A.P., Fabris, D., and Agris, P.F. (2011) The RNA modiication database, RNAMDB: 2011 update. Nucleic Acids Res., 39 (Database issue), D195–D201. Nordhoff, E., Kirpekar, F., and Roepstorff, P. (1996) Mass spectrometry of nucleic acids. Mass Spectrom. Rev., 15, 67–138. Limbach, P.A. (1996) Indirect mass spectrometric methods for characterizing and sequencing

227

228

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

55

56

57

58

59

60

61 62

63

oligonucleotides. Mass Spectrom. Rev., 15, 297–336. Liebich, H.M., Muller-Hagedorn, S., Klaus, F., Meziane, K., Kim, K.R., Frickenschmidt, A., and Kammerer, B. (2005) Chromatographic, capillary electrophoretic and matrix-assisted laser desorption ionization time-oflight mass spectrometry analysis of urinary modiied nucleosides as tumor markers. J. Chromatogr., 1071 (1-2), 271–275. Spengler, B., Kirsch, D., and Kaufmann, R. (1991) Metastable decay of peptides and proteins in matrix-assisted laser-desorption mass spectrometry. Rapid Commun. Mass Spectrom., 5, 198–202. Lan, Z.H., Wang, P., and Giese, R.W. (1999) Matrix-assisted laser desorption/ ionization mass spectrometry of deoxynucleotides labeled with an IMI dye. Rapid Commun. Mass Spectrom., 13 (14), 1454–1457. Comisarow, M.B. and Marshall, A.G. (1974) Fourier transform ion cyclotron resonance. Chem. Phys. Lett., 25 (2), 282–283. Hettich, R. and Buchanan, M. (1991) Structural characterization of normal and modiied oligonucleotides by matrix-assisted laser desorption Fourier transform mass spectrometry. J. Am. Soc. Mass Spectrom., 2 (5), 402–412. Stemmler, E.A., Buchanan, M.V., Hurst, G.B., and Hettich, R.L. (1995) Analysis of modiied oligonucleotides by matrix-assisted laser desorption/ ionization Fourier transform mass spectrometry. Anal. Chem., 67 (17), 2924–2930. McLafferty, F.W. (1981) Tandem mass spectrometry. Science, 214, 280–287. Muddiman, D.C., Anderson, G.A., Hofstadler, S.A., and Smith, R.D. (1997) Length and base composition of PCR-ampliied nucleic acids using mass measurements from electrospray ionization mass spectrometry. Anal. Chem., 69 (8), 1543–1549. Pomerantz, S.C., Kowalak, J.A., and McCloskey, J.A. (1993) Determination of oligonucleotide composition from mass spectrometrically measured molecular

64

65

66

67

68

69

70

71

weight. J. Am. Soc. Mass Spectrom., 4, 204–209. Guittard, J., Paciico, C., Blais, J.C., Bolbach, G., Chottard, J.C., and Spassky, A. (1995) Matrix-assisted laser desorption ionization time-of-light mass spectrometry of DNA-Pt(II) complexes. Rapid Commun. Mass Spectrom., 9 (1), 33–36. Ossipov, D., Gohil, S., and Chattopadhyaya, J. (2002) Synthesis of the DNA-[Ru(tpy)(dppz)(CH(3)CN)](2+) conjugates and their photo cross-linking studies with the complementary DNA strand. J. Am. Chem. Soc., 124 (45), 13416–13433. Little, D.P., Thannhauser, T.W., and McLafferty, F.W. (1995) Veriication of 50- to 100-mer DNA and RNA sequences with high-resolution mass spectrometry. Proc. Natl Acad. Sci. USA, 92 (6), 2318–2322. Little, D.P., Aaserud, D.J., Valaskovic, G.A., and McLafferty, F.W. (1996) Sequence information from 42–108-mer DNAs (complete for a 50-mer) by tandem mass spectrometry. J. Am. Chem. Soc., 118, 9352–9359. Kowalak, J.A., Pomerantz, S.C., Crain, P.F., and McCloskey, J.A. (1993) A novel method for the determination of post-transcriptional modiication in RNA by mass spectrometry. Nucleic Acids Res., 21 (19), 4577–4585. Kowalak, J.A., Bruenger, E., Hashizume, T., Peltier, J.M., Ofengand, J., and McCloskey, J.A. (1996) Structural characterization of U*-1915 in domain IV from Escherichia coli 23S ribosomal RNA as 3-methylpseudouridine. Nucleic Acids Res., 24 (4), 688–693. Kirpekar, F., Douthwaite, S., and Roepstorff, P. (2000) Mapping posttranscriptional modiications in 5S ribosomal RNA by MALDI mass spectrometry. RNA, 6 (2), 296–306. Tamura, T., Araki, Y., Yamaoka, S., Inagaki, K., and Tanaka, H. (1997) Determination of methylation speciicity of sequence-speciic DNA methyltransferases using matrix assisted laser desorption/ionization time-of-light mass spectrometry. Nucleic Acids Res., 25 (20), 4162–4164.

References 72 Bonk, T. and Humeny, A. (2001)

73

74

75

76

77

78

79

80

MALDI-TOF-MS analysis of protein and DNA. Neuroscientist, 7 (1), 6–12. Tost, J.R., Schatz, P., Schuster, M., Berlin, K., and Gut, I.G. (2003) Analysis and accurate quantiication of CpG methylation by MALDI mass spectrometry. Nucleic Acids Res., 31 (9). Tretyakova, N., Matter, B., Ogdie, A., Wishnok, J.S., and Tannenbaum, S.R. (2001) Locating nucleobase lesions within DNA sequences by MALDI-TOF mass spectral analysis of exonuclease ladders. Chem. Res. Toxicol., 14 (8), 1058–1070. Brown, K., Harvey, C.A., Turteltaub, K.W., and Shields, S.J. (2003) Structural characterization of carcinogen-modiied oligodeoxynucleotide adducts using matrix-assisted laser desorption/ ionization mass spectrometry. J. Mass Spectrom., 38 (1), 68–79. Bartolini, W.P. and Johnston, M.V. (2000) Characterizing DNA photooxidation reactions by high-resolution mass measurements with matrixassisted laser desorption/ionization time-of-light mass spectrometry. J. Mass Spectrom., 35 (3), 408–416. Luo, H., Lipton, M.S., and Smith, R.D. (2002) Charge effects for differentiation of oligodeoxynucleotide isomers containing 8-oxo-dG residues. J. Am. Soc. Mass Spectrom., 13 (3), 195–199. Zhang, L.-K. and Gross, M.L. (2002) Location of abasic sites in oligodeoxynucleotides by tandem mass spectrometry and by a chemical cleavage initiated by an unusual reaction of the ODN with MALDI matrix. J. Am. Soc. Mass Spectrom., 13 (12), 1418–1426. Wang, Y., Taylor, J.S., and Gross, M.L. (1999) Differentiation of isomeric photomodiied oligodeoxynucleotides by fragmentation of ions produced by matrix-assisted laser desorption ionization and electrospray ionization. J. Am. Soc. Mass Spectrom., 10 (4), 329–338. Gasparutto, D., Da Cruz, S., Bourdat, A.G., Jaquinod, M., and Cadet, J. (1999) Synthesis and biochemical properties of cyanuric acid nucleoside-containing DNA oligomers. Chem. Res. Toxicol., 12 (7), 630–638.

81 Kozekov, I.D., Nechev, L.V., Moseley,

82

83

84

85

86

87

88

M.S., Harris, C.M., Rizzo, C.J., Stone, M.P., and Harris, T.M. (2003) DNA interchain cross-links formed by acrolein and crotonaldehyde. J. Am. Chem. Soc., 125 (1), 50–61. Schuette, J.M., Pieles, U., Maleknia, S.D., Srivatsa, G.S., Cole, D.L., Moser, H.E., and Afeyan, N.B. (1995) Sequence analysis of phosphorothioate oligonucleotides via matrix-assisted laser desorption ionization time-of-light mass spectrometry. J. Pharm. Biomed. Anal., 13 (10), 1195–1203. Berlin, K. and Gut, I.G. (1999) Analysis of negatively “charge tagged” DNA by matrix-assisted laser desorption/ ionization time-of-light mass spectrometry. Rapid Commun. Mass Spectrom., 13 (17), 1739–1743. Sengle, G., Jenne, A., Arora, P.S., Seelig, B., Nowick, J.S., Jäschke, A., and Famulok, M. (2000) Synthesis, incorporation eficiency, and stability of disulide bridged functional groups at RNA 5′-ends. Bioorg. Med. Chem., 8 (6), 1317–1329. Abdi, F.A., Mundt, M., Doggett, N., Bradbury, E.M., and Chen, X. (2002) Validation of DNA sequences using mass spectrometry coupled with nucleoside mass tagging. Genome Res., 12 (7), 1135–1141. Patteson, K.G., Rodicio, L.P., and Limbach, P.A. (2001) Identiication of the mass-silent post-transcriptionally modiied nucleoside pseudouridine in RNA by matrix-assisted laser desorption/ ionization mass spectrometry. Nucleic Acids Res., 29 (10), E49-9. Mengel-Jorgensen, J. and Kirpekar, F. (2002) Detection of pseudouridine and other modiications in tRNA by cyanoethylation and MALDI mass spectrometry. Nucleic Acids Res., 30 (23), e135. Durairaj, A. and Limbach, P.A. (2008) Matrix-assisted laser desorption/ ionization mass spectrometry screening for pseudouridine in mixtures of small RNAs by chemical derivatization, RNase digestion and signature products. Rapid Commun. Mass Spectrom., 22 (23), 3727–3734.

229

230

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics 89 Tang, W., Zhu, L., and Smith, L.M.

90

91

92

93

94

95

96

97

(1997) Controlling DNA fragmentation in MALDI-MS by chemical modiication. Anal. Chem., 69 (3), 302–312. Ono, T., Scalf, M., and Smith, L.M. (1997) 2′-Fluoro modiied nucleic acids: polymerase-directed synthesis, properties and stability to analysis by matrix-assisted laser desorption/ ionization mass spectrometry. Nucleic Acids Res., 25 (22), 4581–4588. Wan, K.X. and Gross, M.L. (2001) Fragmentation mechanisms of oligodeoxynucleotides: effects of replacing phosphates with methylphosphonates and thymines with other bases in T-rich sequences. J. Am. Soc. Mass Spectrom., 12 (5), 580–589. Chen, X., Fei, Z., Smith, L.M., Bradbury, E.M., and Majidi, V. (1999) Stableisotope-assisted MALDI-TOF mass spectrometry for accurate determination of nucleotide compositions of PCR products. Anal. Chem., 71 (15), 3118–3125. Berhane, B.T. and Limbach, P.A. (2003) Stable isotope labeling for matrixassisted laser desorption/ionization mass spectrometry and post-source decay analysis of ribonucleic acids. J. Mass Spectrom., 38 (8), 872–878. Meng, Z. and Limbach, P.A. (2005) Quantitation of ribonucleic acids using 18O labeling and mass spectrometry. Anal. Chem., 77 (6), 1891–1895. Nielsen, P.E., Egholm, M., Berg, R.H., and Buchardt, O. (1991) Sequenceselective recognition of DNA by strand displacement with a thymine-substituted polyamide. Science, 254 (5037), 1497–1500. Egholm, M., Buchardt, O., Christensen, L., Behrens, C., Freier, S.M., Driver, D.A., Berg, R.H., Kim, S.K., Norden, B., and Nielsen, P.E. (1993) PNA hybridizes to complementary oligonucleotides obeying the Watson–Crick hydrogenbonding rules. Nature, 365 (6446), 566–568. Butler, J.M., Jiang-Baucom, P., Huang, M., Belgrader, P., and Girard, J. (1996) Peptide nucleic acid characterization by MALDI-TOF mass spectrometry. Anal. Chem., 68 (18), 3283–3287.

98 Jiang-Baucom, P., Girard, J.E., Butler, J.,

99

100

101

102

103

104

105

106

107

and Belgrader, P. (1997) DNA typing of human leukocyte antigen sequence polymorphisms by peptide nucleic acid probes and MALDI-TOF mass spectrometry. Anal. Chem., 69 (23), 4894–4898. Ganem, B., Li, Y.T., and Henion, J.D. (1993) Detection of oligonucleotide duplex forms by ion-spray mass spectrometry. Tetrahedron Lett., 34, 1445–1448. Light-Wahl, K.J., Springer, D.L., Winger, B.E., Edmonds, C.G., Camp, D.G., Thrall, B.D., and Smith, R.D. (1993) Observation of a small oligonucleotide duplex by electrospray ionization mass spectrometry. J. Am. Chem. Soc., 115, 803–804. Lecchi, P. and Pannell, L.K. (1995) The detection of intact double-stranded DNA by MALDI. J. Am. Soc. Mass Spectrom., 6, 972–975. Little, D.P., Jacob, A., Becker, T., Braun, A., Darnhofer-Demar, B., Jurinke, C., van den Boom, D., and Koster, H. (1997) Direct detection of synthetic and biologically generated double-stranded DNA by MALDI-TOF MS. Int. J. Mass Spectrom. Ion Processes, 169/170, 323–330. Ding, J. and Anderegg, R.J. (1994) Speciic and nonspeciic dimer formation in the electrospray ionization mass spectrometry of oligonucleotides. J. Am. Soc. Mass Spectrom., 6, 159–164. Sudha, R. and Zenobi, R. (2002) The detection and stability of DNA duplexes probed by MALDI mass spectrometry. Helv. Chim. Acta, 85 (10), 3136–3143. Distler, A.M. and Allison, J. (2002) Additives for the stabilization of double-stranded DNA in UV-MALDI MS. J. Am. Soc. Mass Spectrom., 13 (9), 1129–1137. Ohara, K., Smietana, M., and Vasseur, J.-J. (2006) Characterization of speciic noncovalent complexes between guanidinium derivatives and singlestranded DNA by MALDI. J. Am. Soc. Mass Spectrom., 17 (3), 283–291. Terrier, P., Tortajada, J., Zin, G.G., and Buchmann, W. (2007) Noncovalent complexes between DNA and basic

References

108

109

110

111

112

113

114

115

polypeptides or polyamines by MALDITOF. J. Am. Soc. Mass Spectrom., 18 (11), 1977–1989. Tang, X., Callahan, J.H., Zhou, P., and Vertes, A. (1995) Noncovalent protein– oligonucleotide interactions monitored by matrix-assisted laser desorption/ ionization mass spectrometry. Anal. Chem., 67 (24), 4542–4548. Lin, S., Cotter, R.J., and Woods, A.S. (1998) Detection of non-covalent interaction of single and double stranded DNA with peptides by MALDI-TOF. Proteins, (Suppl. 2), 12–21. Thiede, B., and von Janta-Lipinski, M. (1998) Noncovalent RNA-peptide complexes detected by matrix-assisted laser desorption/ionization mass spectrometry. Rapid Commun. Mass Spectrom., 12 (23), 1889–1894. Gruic-Sovulj, I., Ludemann, H.C., Hillenkamp, F., Kucan, I., and Peter-Katalinic, J. (1997) Detection of noncovalent tRNA aminoacyl-tRNA synthetase complexes by matrix-assisted laser desorption/ionization mass spectrometry. J. Biol. Chem., 272 (51), 32084–32091. Luo, S.-Z., Li, Y.-M., Qiang, W., Zhao, Y.-F., Abe, H., Nemoto, T., Qin, X.-R., and Nakanishi, H. (2004) Detection of speciic noncovalent interaction of peptide with DNA by MALDI-TOF. J. Am. Soc. Mass Spectrom., 15 (1), 28–31. Chen, F., Mädler, S., Weidmann, S., and Zenobi, R. (2012) MALDI-MS detection of noncovalent interactions of single stranded DNA with Escherichia coli single-stranded DNA-binding protein. J. Mass Spectrom., 47 (5), 560–566. Jensen, O.N., Barofsky, D.F., Young, M.C., von Hippel, P.H., Swenson, S., and Seifried, S.E. (1993) Direct observation of UV-crosslinked proteinnucleic acid complexes by matrixassisted laser desorption ionization mass spectrometry. Rapid Commun. Mass Spectrom., 7 (6), 496–501. Bich, C., de Bovet, C., Rochel, N., Peluso-Iltis, C., Panagiotidis, A., Nazabal, A., Moras, D., and Zenobi, R. (2010) Detection of nucleic acid-nuclear hormone receptor complexes with mass

116

117

118

119

120

121

122

spectrometry. J. Am. Soc. Mass Spectrom., 21 (4), 635–645. Urlaub, H., Thiede, B., Muller, E.C., and Wittmann-Liebold, B. (1997) Contact sites of peptide-oligoribonucleotide cross-links identiied by a combination of peptide and nucleotide sequencing with MALDI MS. J. Protein Chem., 16 (5), 375–383. Thiede, B., Urlaub, H., Neubauer, H., and Wittmann-Liebold, B. (1999) Identiication and sequence analysis of RNA-protein contact sites by N-terminal sequencing and MALDI-MS. Methods Mol. Biol., 118, 63–72. Thiede, B. and Wittmann-Liebold, B. (2000) Analysis of RNA-protein cross-link sites by matrix-assisted laser desorption/ionization mass spectrometry and N-terminal microsequencing. Methods Enzymol., 318, 438–446. Geyer, H., Geyer, R., and Pingoud, V. (2004) A novel strategy for the identiication of protein-DNA contacts by photocrosslinking and mass spectrometry. Nucleic Acids Res., 32 (16), e132. Naryshkin, N.A., Farrow, M.A., Ivanovskaya, M.G., Oretskaya, T.S., Shabarova, Z.A., and Gait, M.J. (1997) Chemical cross-linking of the human immunodeiciency virus type 1 Tat protein to synthetic models of the RNA recognition sequence TAR containing site-speciic trisubstituted pyrophosphate analogues. Biochemistry, 36 (12), 3496–3505. Connor, D.A., Falick, A.M., Young, M.C., and Shetlar, M.D. (1998) Probing the binding region of the singlestranded DNA-binding domain of rat DNA polymerase beta using nanosecond-pulse laser-induced cross-linking and mass spectrometry. Photochem. Photobiol., 68 (3), 299–308. Urlaub, H., Thiede, B., Muller, E.C., Brimacombe, R., and Wittmann-Liebold, B. (1997) Identiication and sequence analysis of contact sites between ribosomal proteins and rRNA in Escherichia coli 30 S subunits by a new approach using matrix-assisted laser desorption/ionization-mass spectrometry combined with N-terminal

231

232

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

123

124

125

126

127

128

129

130

131

132

microsequencing. J. Biol. Chem., 272 (23), 14547–14555. Thiede, B., Urlaub, H., Neubauer, H., Grelle, G., and Wittmann-Liebold, B. (1998) Precise determination of RNA-protein contact sites in the 50 S ribosomal subunit of Escherichia coli. Biochem. J., 334 (Pt 1), 39–42. Urlaub, H., Hartmuth, K., Kostka, S., Grelle, G., and Luhrmann, R. (2000) A general approach for identiication of RNA-protein cross-linking sites within native human spliceosomal small nuclear ribonucleoproteins (snRNPs). Analysis of RNA-protein contacts in native U1 and U4/U6.U5 snRNPs. J. Biol. Chem., 275 (52), 41458–41468. Steen, H. and Jensen, O.N. (2002) Analysis of protein-nucleic acid interactions by photochemical crosslinking and mass spectrometry. Mass Spectrom. Rev., 21 (3), 163–182. Crain, P.F. and McCloskey, J.A. (1998) Applications of mass spectrometry to the characterization of oligonucleotides and nucleic acids. Curr. Opin. Biotechnol., 9 (1), 25–34. Murray, K.K. (1996) DNA sequencing by mass spectrometry. J. Mass Spectrom., 31 (11), 1203–1215. Tost, J. and Gut, I.G. (2002) Genotyping single nucleotide polymorphisms by mass spectrometry. Mass Spectrom. Rev., 21 (6), 388–418. Guo, B. (1999) Mass spectrometry in DNA analysis. Anal. Chem., 71 (12), 333R–337R. Shaler, T.A., Tan, Y., Wickham, J.N., Wu, K.J., and Becker, C.H. (1995) Analysis of enzymatic DNA sequencing reactions by matrix-assisted laser desorption/ionization time-of-light mass spectrometry. Rapid Commun. Mass Spectrom., 9 (10), 942–947. Shaler, T.A., Wickham, J.N., Sannes, K.A., Wu, K.J., and Becker, C.H. (1996) Effect of impurities on the matrixassisted laser desorption mass spectra of single-stranded oligodeoxynucleotides. Anal. Chem., 68 (3), 576–579. Nordhoff, E., Luebbert, C., Thiele, G., Heiser, V., and Lehrach, H. (2000) Rapid determination of short DNA

133

134

135

136

137

138

139

140

sequences by the use of MALDI-MS. Nucleic Acids Res., 28 (20), E86. Jurinke, C., van den Boom, D., Collazo, V., Luchow, A., Jacob, A., and Koster, H. (1997) Recovery of nucleic acids from immobilized biotin-streptavidin complexes using ammonium hydroxide and applications in MALDI-TOF mass spectrometry. Anal. Chem., 69 (5), 904–910. Ross, P.L. and Belgrader, P. (1997) Analysis of short tandem repeat polymorphisms in human DNA by matrix-assisted laser desorption/ ionization mass spectrometry. Anal. Chem., 69 (19), 3966–3972. Gut, I.G. and Beck, S. (1995) A procedure for selective DNA alkylation and detection by mass spectrometry. Nucleic Acids Res., 23 (8), 1367–1373. Schuerenberg, M., Luebbert, C., Eickhoff, H., Kalkum, M., Lehrach, H., and Nordhoff, E. (2000) Prestructured MALDI-MS sample supports. Anal. Chem., 72 (15), 3436–3442. Little, D.P., Cornish, T.J., O’Donnell, M.J., Braun, A., Cotter, R.J., and Koster, H. (1997) MALDI on a chip: analysis of arrays of low-femtomole to subfemtomole quantities of synthetic oligonucleotides and DNA diagnostic products dispensed by a piezoelectric pipet. Anal. Chem., 69, 4540–4546. Little, D.P., Braun, A., O’Donnell, M.J., and Koster, H. (1997) Mass spectrometry from miniaturized arrays for full comparative DNA analysis. Nat. Med., 3 (12), 1413–1416. Little, D.P., Jacob, A., Becker, T., Braun, A., Darnhofer-Demar, B., Jurinke, C., van den Boom, D., and Koster, H. (1997) Direct Detection of synthetic and biologically generated double-stranded DNA by MALDI-TOF MS. Int. J. Mass Spectrom. Ion Processes, 169/170, 323–330. Wada, Y. (1998) Separate analysis of complementary strands of restriction enzyme-digested DNA. An application of restriction fragment mass mapping by matrix-assisted laser desorption/ ionization mass spectrometry. J. Mass Spectrom., 33 (2), 187–192.

References 141 Butler, J.M., Li, J., Shaler, T.A.,

142

143

144

145

146

147

148

149

Monforte, J.A., and Becker, C.H. (1999) Reliable genotyping of short tandem repeat loci without an allelic ladder using time-of-light mass spectrometry. Int. J. Legal Med., 112 (1), 45–49. van den Boom, D., Jurinke, C., McGinniss, M.J., and Berkenkamp, S. (2001) Microsatellites: perspectives and potentials of mass spectrometric analysis. Expert Rev. Mol. Diagn., 1 (4), 383–393. Krebs, S., Seichter, D., and Forster, M. (2001) Genotyping of dinucleotide tandem repeats by MALDI mass spectrometry of ribozyme-cleaved RNA transcripts. Nat. Biotechnol., 19 (9), 877–880. Braun, A., Little, D.P., Reuter, D., Muller-Mysok, B., and Koster, H. (1997) Improved analysis of microsatellites using mass spectrometry. Genomics, 46 (1), 18–23. Wada, Y., Mitsumori, K., Terachi, T., and Ogawa, O. (1999) Measurement of polymorphic trinucleotide repeats in the androgen receptor gene by matrixassisted laser desorption/ionization time-of-light mass spectrometry. J. Mass Spectrom., 34 (8), 885–888. Liu, Y.H., Bai, J., Zhu, Y., Liang, X., Siemieniak, D., Venta, P.J., and Lubman, D.M. (1995) Rapid screening of genetic polymorphisms using buccal cell DNA with detection by matrixassisted laser desorption/ionization mass spectrometry. Rapid Commun. Mass Spectrom., 9 (9), 735–743. Jurinke, C., van den Boom, D., Jacob, A., Tang, K., Worl, R., and Koster, H. (1996) Analysis of ligase chain reaction products via matrix-assisted laser desorption/ionization time-of-lightmass spectrometry. Anal. Biochem., 237 (2), 174–181. Grifin, T.J., Tang, W., and Smith, L.M. (1997) Genetic analysis by peptide nucleic acid afinity MALDI-TOF mass spectrometry. Nat. Biotechnol., 15 (13), 1368–1372. Jurinke, C., Zollner, B., Feucht, H.H., Jacob, A., Kirchhubel, J., Luchow, A., van den Boom, D., Laufs, R., and Koster, H. (1996) Detection of hepatitis B virus

150

151

152

153

154

155

156

157

158

DNA in serum samples via nested PCR and MALDI-TOF mass spectrometry. Genet. Anal., 13 (3), 67–71. Jurinke, C., Zollner, B., Feucht, H.H., van den Boom, D., Jacob, A., Polywka, S., Laufs, R., and Koster, H. (1998) Application of nested PCR and mass spectrometry for DNA-based virus detection: HBV-DNA detected in the majority of isolated anti-HBc positive sera. Genet. Anal., 14 (3), 97–102. Haff, L.A. and Smirnov, I.P. (1997) Single-nucleotide polymorphism identiication assays using a thermostable DNA polymerase and delayed extraction MALDI-TOF mass spectrometry. Genome Res., 7 (4), 378–388. Haff, L.A. and Smirnov, I.P. (1997) Multiplex genotyping of PCR products with MassTag-labeled primers. Nucleic Acids Res., 25 (18), 3749–3750. Fei, Z., Ono, T., and Smith, L.M. (1998) MALDI-TOF mass spectrometric typing of single nucleotide polymorphisms with mass-tagged ddNTPs. Nucleic Acids Res., 26 (11), 2827–2828. Braun, A., Little, D.P., and Koster, H. (1997) Detecting CFTR gene mutations by using primer oligo base extension and mass spectrometry. Clin. Chem., 43 (7), 1151–1158. Little, D.P., Braun, A., DarnhoferDemar, B., and Koster, H. (1997) Identiication of apolipoprotein E polymorphisms using temperature cycled primer oligo base extension and mass spectrometry. Eur. J. Clin. Chem. Clin. Biochem., 35 (7), 545–548. Kim, S., Edwards, J.R., Deng, L., Chung, W., and Ju, J. (2002) Solid phase capturable dideoxynucleotides for multiplex genotyping using mass spectrometry. Nucleic Acids Res., 30 (16), e85. Kim, S., Shi, S., Bonome, T., Ulz, M.E., Edwards, J.R., Fodstad, H., Russo, J.J., and Ju, J. (2003) Multiplex genotyping of the human beta2-adrenergic receptor gene using solid-phase capturable dideoxynucleotides and mass spectrometry. Anal. Biochem., 316 (2), 251–258. Harksen, A., Ueland, P.M., Refsum, H., and Meyer, K. (1999) Four common

233

234

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

159

160

161

162

163

164

165

166

mutations of the cystathionine beta-synthase gene detected by multiplex PCR and matrix-assisted laser desorption/ionization time-of-light mass spectrometry. Clin. Chem., 45 (8 Pt 1), 1157–1161. Nordhoff, E., Ingendoh, A., Cramer, R., Overberg, A., Stahl, B., Karas, M., Hillenkamp, F., and Crain, P.F. (1992) Matrix-assisted laser desorption/ ionization mass spectrometry of nucleic acids with wavelengths in the ultraviolet and infrared. Rapid Commun. Mass Spectrom., 6 (12), 771–776. Storm, N., Darnhofer-Patel, B., van den Boom, D., and Rodi, C.P. (2003) MALDI-TOF mass spectrometry-based SNP genotyping. Methods Mol. Biol., 212, 241–262. O’Donnell-Maloney, M.J. and Little, D.P. (1996) Microfabrication and array technologies for DNA sequencing and diagnostics. Genet. Anal., 13 (6), 151–157. Sauer, S., Lechner, D., Berlin, K., Lehrach, H., Escary, J.L., Fox, N., and Gut, I.G. (2000) A novel procedure for eficient genotyping of single nucleotide polymorphisms. Nucleic Acids Res., 28 (5), E13. Sauer, S., Lechner, D., Berlin, K., Plancon, C., Heuermann, A., Lehrach, H., and Gut, I.G. (2000) Full lexibility genotyping of single nucleotide polymorphisms by the GOOD assay. Nucleic Acids Res., 28 (23), E100. Sauer, S., Gelfand, D.H., Boussicault, F., Bauer, K., Reichert, F., and Gut, I.G. (2002) Facile method for automated genotyping of single nucleotide polymorphisms by mass spectrometry. Nucleic Acids Res., 30 (5), e22. Sauer, S., Lehrach, H., and Reinhardt, R. (2003) MALDI mass spectrometry analysis of single nucleotide polymorphisms by photocleavage and charge-tagging. Nucleic Acids Res., 31 (11), e63. Kwiatkowski, R.W., Lyamichev, V., de Arruda, M., and Neri, B. (1999) Clinical, genetic, and pharmacogenetic applications of the Invader assay. Mol. Diagn., 4 (4), 353–364.

167 Grifin, T.J., Hall, J.G., Prudent, J.R.,

168

169

170

171

172

173

174

175

and Smith, L.M. (1999) Direct genetic analysis by matrix-assisted laser desorption/ionization mass spectrometry. Proc. Natl Acad. Sci. USA, 96 (11), 6301–6306. Wolfe, J.L., Kawate, T., Sarracino, D.A., Zillmann, M., Olson, J., Stanton, V.P., Jr, and Verdine, G.L. (2002) A genotyping strategy based on incorporation and cleavage of chemically modiied nucleotides. Proc. Natl Acad. Sci. USA, 99 (17), 11073–11078. Kim, Y.J., Kim, S.O., Chung, H.J., Jee, M.S., Kim, B.G., Kim, K.M., Yoon, J.H., Lee, H.S., Kim, C.Y., Kim, S., Yoo, W., and Hong, S.P. (2005) Population genotyping of hepatitis C virus by matrix-assisted laser desorption/ ionization time-of-light mass spectrometry analysis of short DNA fragments. Clin. Chem., 51 (7), 1123–1131. Templeton, A.R., Sing, C.F., Kessling, A., and Humphries, S. (1988) A cladistic analysis of phenotype associations with haplotypes inferred from restriction endonuclease mapping. II. The analysis of natural populations. Genetics, 120 (4), 1145–1154. Judson, R., Stephens, J.C., and Windemuth, A. (2000) The predictive power of haplotypes in clinical response. Pharmacogenomics, 1 (1), 15–26. Martin, E.R., Gilbert, J.R., Lai, E.H., Riley, J., Rogala, A.R., Slotterbeck, B.D., Sipe, C.A., Grubber, J.M., Warren, L.L., Conneally, P.M., Saunders, A.M., Schmechel, D.E., Purvis, I., PericakVance, M.A., Roses, A.D., and Vance, J.M. (2000) Analysis of association at single nucleotide polymorphisms in the APOE region. Genomics, 63 (1), 7–12. Cardon, L.R. and Abecasis, G.R. (2003) Using haplotype blocks to map human complex trait loci. Trends Genet., 19 (3), 135–140. Kruglyak, L. (1999) Prospects for whole-genome linkage disequilibrium mapping of common disease genes. Nat. Genet., 22 (2), 139–144. Stephens, M., Smith, N.J., and Donnelly, P. (2001) A new statistical method for haplotype reconstruction from

References

176

177

178

179

180

181

182

183

184

185

population data. Am. J. Hum. Genet., 68 (4), 978–989. Yan, H., Papadopoulos, N., Marra, G., Perrera, C., Jiricny, J., Boland, C.R., Lynch, H.T., Chadwick, R.B., de la Chapelle, A., Berg, K., Eshleman, J.R., Yuan, W., Markowitz, S., Laken, S.J., Lengauer, C., Kinzler, K.W., and Vogelstein, B. (2000) Conversion of diploidy to haploidy. Nature, 403 (6771), 723–724. Bentley, K.L., Li, W.L., VannBerg, F.O., Choi, J.Y., Yu, J., Kao, F.T., and Ruano, G. (1997) Detailed analysis of a 17q21 microdissection library by sequence bioinformatics and isolation of region-speciic clones. Somat. Cell Mol. Genet., 23 (5), 353–365. Ruano, G., Kidd, K.K., and Stephens, J.C. (1990) Haplotype of multiple polymorphisms resolved by enzymatic ampliication of single DNA molecules. Proc. Natl Acad. Sci. USA, 87 (16), 6296–6300. Ruano, G. and Kidd, K.K. (1989) Direct haplotyping of chromosomal segments from multiple heterozygotes via allele-speciic PCR ampliication. Nucleic Acids Res., 17 (20), 8392. Ruano, G., Fenton, W., and Kidd, K.K. (1989) Biphasic ampliication of very dilute DNA samples via “booster” PCR. Nucleic Acids Res., 17 (13), 5407. Tost, J., Brandt, O., Boussicault, F., Derbala, D., Caloustian, C., Lechner, D., and Gut, I.G. (2002) Molecular haplotyping at high throughput. Nucleic Acids Res., 30 (19), e96. Ding, C. and Cantor, C.R. (2003) Direct molecular haplotyping of long-range genomic DNA with M1-PCR. Proc. Natl Acad. Sci. USA, 100 (13), 7449–7453. Stephens, J.C., Rogers, J., and Ruano, G. (1990) Theoretical underpinning of the single-molecule-dilution (SMD) method of direct haplotype resolution. Am. J. Hum. Genet., 46 (6), 1149–1155. Sanger, F., Nicklen, S., and Coulson, A.R. (1977) DNA sequencing with chain-terminating inhibitors. Proc. Natl Acad. Sci. USA, 74 (12), 5463–5467. Smith, L.M. (1993) The future of DNA sequencing. Science, 262 (5133), 530–532.

186 Koster, H., Tang, K., Fu, D.J., Braun, A.,

187

188

189

190

191

192

193

194

van den Boom, D., Smith, C.L., Cotter, R.J., and Cantor, C.R. (1996) A strategy for rapid and eficient DNA sequencing by mass spectrometry. Nat. Biotechnol., 14 (9), 1123–1128. Roskey, M.T., Juhasz, P., Smirnov, I.P., Takach, E.J., Martin, S.A., and Haff, L.A. (1996) DNA sequencing by delayed extraction-matrix-assisted laser desorption/ionization time of light mass spectrometry. Proc. Natl Acad. Sci. USA, 93 (10), 4724–4729. Kirpekar, F., Nordhoff, E., Larsen, L.K., Kristiansen, K., Roepstorff, P., and Hillenkamp, F. (1998) DNA sequence analysis by MALDI mass spectrometry. Nucleic Acids Res., 26 (11), 2554–2559. Kirpekar, F., Nordhoff, E., Kristiansen, K., Roepstorff, P., Hahner, S., and Hillenkamp, F. (1995) 7-Deaza purine bases offer a higher ion stability in the analysis of DNA by matrix-assisted laser desorption/ionization mass spectrometry. Rapid Commun. Mass Spectrom., 9 (6), 525–531. Fu, D.J., Tang, K., Braun, A., Reuter, D., Darnhofer-Demar, B., Little, D.P., O’Donnell, M.J., Cantor, C.R., and Koster, H. (1998) Sequencing exons 5 to 8 of the p53 gene by MALDI-TOF mass spectrometry. Nat. Biotechnol., 16 (4), 381–384. Taranenko, N.I., Allman, S.L., Golovlev, V.V., Taranenko, N.V., Isola, N.R., and Chen, C.H. (1998) Sequencing DNA using mass spectrometry for ladder detection. Nucleic Acids Res., 26 (10), 2488–2490. Kwon, Y., Tang, K., Cantor, C., Koster, H., and Kang, C. (2001) DNA sequencing and genotyping by transcriptional synthesis of chainterminated RNA ladders and MALDITOF mass spectrometry. Nucleic Acids Res., 29 (3), E11. Edwards, J.R., Itagaki, Y., and Ju, J. (2001) DNA sequencing using biotinylated dideoxynucleotides and mass spectrometry. Nucleic Acids Res., 29 (21), E104-4. Broder, S. and Venter, J.C. (2000) Whole genomes: the foundation of new biology

235

236

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics

195

196

197

198

199

200

201

and medicine. Curr. Opin. Biotechnol., 11 (6), 581–585. Hahner, S., Ludemann, H.C., Kirpekar, F., Nordhoff, E., Roepstorff, P., Galla, H.J., and Hillenkamp, F. (1957) Matrix-assisted laser desorption/ ionization mass spectrometry (MALDI) of endonuclease digests of RNA. Nucleic Acids Res., 25 (10), 64. Shchepinov, M.S., Denissenko, M.F., Smylie, K.J., Worl, R.J., Leppin, A.L., Cantor, C.R., and Rodi, C.P. (2001) Matrix-induced fragmentation of P3′-N5′ phosphoramidate-containing DNA: high-throughput MALDI-TOF analysis of genomic sequence polymorphisms. Nucleic Acids Res., 29 (18), 3864–3872. von Wintzingerode, F., Bocker, S., Schlotelburg, C., Chiu, N.H., Storm, N., Jurinke, C., Cantor, C.R., Gobel, U.B., and van den Boom, D. (2002) Basespeciic fragmentation of ampliied 16S rRNA genes analyzed by mass spectrometry: a tool for rapid bacterial identiication. Proc. Natl Acad. Sci. USA, 99 (10), 7039–7044. Elso, C., Toohey, B., Reid, G.E., Poetter, K., Simpson, R.J., and Foote, S.J. (2002) Mutation detection using mass spectrometric separation of tiny oligonucleotide fragments. Genome Res., 12 (9), 1428–1433. Hartmer, R., Storm, N., Boecker, S., Rodi, C.P., Hillenkamp, F., Jurinke, C., and van den Boom, D. (2003) RNase T1 mediated base-speciic cleavage and MALDI-TOF MS for high-throughput comparative sequence analysis. Nucleic Acids Res., 31 (9), e47. Krebs, S., Medugorac, I., Seichter, D., and Forster, M. (2003) RNaseCut: a MALDI mass spectrometry-based method for SNP discovery. Nucleic Acids Res., 31 (7), e37. Stanssens, P., Zabeau, M., Meersseman, G., Remes, G., Gansemans, Y., Storm, N., Hartmer, R., Honisch, C., Rodi, C.P., Bocker, S., and van den Boom, D. (2004) High-throughput MALDI-TOF discovery of genomic sequence polymorphisms. Genome Res., 14 (1), 126–133.

202 Sousa, R. and Padilla, R. (1995) A

203

204

205

206

207

208

209

mutant T7 RNA polymerase as a DNA polymerase. EMBO J., 14, 4609–4621. Bocker, S. (2003) SNP and mutation discovery using base-speciic cleavage and MALDI-TOF mass spectrometry. Bioinformatics, 19 (Suppl. 1), I44–I53. Lefmann, M., Honisch, C., Bocker, S., Storm, N., von Wintzingerode, F., Schlotelburg, C., Moter, A., van den Boom, D., and Gobel, U.B. (2004) Novel mass spectrometry-based tool for genotypic identiication of mycobacteria. J. Clin. Microbiol., 42 (1), 339–346. Honisch, C., Chen, Y., Mortimer, C., Arnold, C., Schmidt, O., van den Boom, D., Cantor, C.R., Shah, H.N., and Gharbia, S.E. (2007) Automated comparative sequence analysis by base-speciic cleavage and Mass Spectrometry for nucleic acid-based typing. Proc. Natl Acad. Sci. USA, 104, 10649–10654. Herring, C.D., Raghunathan, A., Honisch, C., Patel, T., Applebee, M.K., Joyce, A.R., Albert, T.J., Blattner, F.R., van den Boom, D., Cantor, C.R., and Palsson, B.Ø. (2006) Comparative genome sequencing of Escherichia coli allows observation of bacterial evolution on a laboratory timescale. Nat. Genet., 38, 1406–1412. Ehrich, M., Nelson, M., Stanssens, P., Zabeau, M., Liloglou, T., Xinarianos, G., Cantor, C., Field, J., and van den Boom, D. (2005) Quantitative high-throughput analysis of DNA methylation patterns by base-speciic cleavage and mass spectrometry. Proc. Natl Acad. Sci. USA, 102 (44), 15785–15790. Honisch, C., Raghunathan, A., Cantor, C.R., Palsson, B.O., and van den Boom, D. (2004) High-throughput mutation detection underlying adaptive evolution of Escherichia coli-K12. Genome Res., 14 (12), 2495–2502. Tang, K., Shahgholi, M., Garcia, B.A., Heaney, P.J., Cantor, C.R., Scott, L.G., and Williamson, J.R. (2002) Improvement in the apparent mass resolution of oligonucleotides by using 12C/14N-enriched samples. Anal. Chem., 74 (1), 226–231.

References 210 Abdi, F., Bradbury, E.M., Doggett, N.,

211

212

213

214

215

216

and Chen, X. (2001) Rapid characterization of DNA oligomers and genotyping of single nucleotide polymorphism using nucleotide-speciic mass tags. Nucleic Acids Res., 29 (13), E61-1. Ross, P., Hall, L., and Haff, L.A. (2000) Quantitative approach to singlenucleotide polymorphism analysis using MALDI-TOF mass spectrometry. Biotechniques, 29 (3), 620–626, 628–629. Werner, M., Sych, M., Herbon, N., Illig, T., Konig, I.R., and Wjst, M. (2002) Large-scale determination of SNP allele frequencies in DNA pools using MALDI-TOF mass spectrometry. Hum. Mutat., 20 (1), 57–64. Le Hellard, S., Ballereau, S.J., Visscher, P.M., Torrance, H.S., Pinson, J., Morris, S.W., Thomson, M.L., Semple, C.A., Muir, W.J., Blackwood, D.H., Porteous, D.J., and Evans, K.L. (2002) SNP genotyping on pooled DNAs: comparison of genotyping technologies and a semi-automated method for data storage and analysis. Nucleic Acids Res., 30 (15), e74. Mohlke, K.L., Erdos, M.R., Scott, L.J., Fingerlin, T.E., Jackson, A.U., Silander, K., Hollstein, P., Boehnke, M., and Collins, F.S. (2002) High-throughput screening for evidence of association by using mass spectrometry genotyping on DNA pools. Proc. Natl Acad. Sci. USA, 99 (26), 16928–16933. Bansal, A., van den Boom, D., Kammerer, S., Honisch, C., Adam, G., Cantor, C.R., Kleyn, P., and Braun, A. (2002) Association testing by DNA pooling: an effective initial screen. Proc. Natl Acad. Sci. USA, 99 (26), 16871–16874. Buetow, K.H., Edmonson, M., MacDonald, R., Clifford, R., Yip, P., Kelley, J., Little, D.P., Strausberg, R., Koester, H., Cantor, C.R., and Braun, A. (2001) High-throughput development and characterization of a genomewide collection of gene-based single nucleotide polymorphism markers by chip-based matrix-assisted laser desorption/ionization time-of-light

217

218

219

220

221

222

223

mass spectrometry. Proc. Natl Acad. Sci. USA, 98 (2), 581–584. Amexis, G., Oeth, P., Abel, K., Ivshina, A., Pelloquin, F., Cantor, C.R., Braun, A., Chumakov, K., and Braun, A. (2001) Quantitative mutant analysis of viral quasispecies by chip-based matrixassisted laser desorption/ ionization time-of-light mass spectrometry. Proc. Natl Acad. Sci. USA, 98 (21), 12097–12102. Muller, S., van den Boom, D., Zirkel, D., Koster, H., Berthold, F., Schwab, M., Westphal, M., and Zumkeller, W. (2000) Retention of imprinting of the human apoptosis-related gene TSSC3 in human brain tumors. Hum. Mol. Genet., 9 (5), 757–763. Knight, J.C., Keating, B.J., Rockett, K.A., and Kwiatkowski, D.P. (2003) In vivo characterization of regulatory polymorphisms by allele-speciic quantiication of RNA polymerase loading. Nat. Genet., 33 (4), 469–475. Sham, P., Bader, J.S., Craig, I., O’Donovan, M., and Owen, M. (2002) DNA Pooling: a tool for large-scale association studies. Nat. Rev. Genet., 3 (11), 862–871. Barratt, B.J., Payne, F., Rance, H.E., Nutland, S., Todd, J.A., and Clayton, D.G. (2002) Identiication of the sources of error in allele frequency estimations from pooled DNA indicates an optimal experimental design. Ann. Hum. Genet., 66 (5-6), 393–405. Herbon, N., Werner, M., Braig, C., Gohlke, H., Dutsch, G., Illig, T., Altmuller, J., Hampe, J., Lantermann, A., Schreiber, S., Bonifacio, E., Ziegler, A., Schwab, S., Wildenauer, D., van den Boom, D., Braun, A., Knapp, M., Reitmeir, P., and Wjst, M. (2003) High-resolution snp scan of chromosome 6p21 in pooled samples from patients with complex diseases. Genomics, 81 (5), 510–518. Ding, C., Chiu, R.W., Lau, T.K., Leung, T.N., Chan, L.C., Chan, A.Y., Charoenkwan, P., Ng, I.S., Law, H.Y., Ma, E.S., Xu, X., Wanapirak, C., Sanguansermsri, T., Liao, C., Ai, M.A., Chui, D.H., Cantor, C.R., and Lo, Y.M.

237

238

5 Analysis of Nucleic Acids, and Practical Implementations in Genomics and Genetics (2004) MS analysis of single-nucleotide differences in circulating nucleic acids: application to noninvasive prenatal diagnosis. Proc. Natl Acad. Sci. USA, 101 (29), 10762–10767. 224 Berggren, W.T., Takova, T., Olson, M.C., Eis, P.S., Kwiatkowski, R.W., and Smith, L.M. (2002) Multiplexed gene expression analysis using the invader RNA assay with MALDI-TOF mass spectrometry detection. Anal. Chem., 74 (8), 1745–1750. 225 Ding, C. and Cantor, C.R. (2003) A high-throughput gene expression analysis technique using competitive PCR and matrix-assisted laser desorption ionization time-of-light MS. Proc. Natl Acad. Sci. USA, 100 (6), 3059–3064. 226 Elvidge, G.P., Price, T.S., Glenny, L., and Ragoussis, J. (2005) Development

and evaluation of real competitive PCR for high-throughput quantitative applications. Anal. Biochem., 339 (2), 231–241. 227 Ding, C., Maier, E., Roscher, A.A., Braun, A., and Cantor, C.R. (2004) Simultaneous quantitative and allele-speciic expression analysis with real competitive PCR. BMC Genet., 5 (1), 8. 228 McCullough, R.M., Cantor, C.R., and Ding, C. (2005) High-throughput alternative splicing quantiication by primer extension and matrix-assisted laser desorption/ionization time-of-light mass spectrometry. Nucleic Acids Res., 33 (11), e99. 229 Pheasant, M. and Mattick, J.S. (2007) Raising the estimate of functional human sequences. Genome Res., 17 (9), 1245–1253.

239

6 MALDI-MS of Glycans and Glycoconjugates Hélène Perreault, Erika Lattová, Dijana Šagi, and Jasna Peter-Katalinic

6.1 Introduction 6.1.1 Glycans in Glycoproteins: Types and Importance

Approximately 50% of proteins produced in eukaryotic cells are post-transitionally glycosylated [1]. This type of modiication of proteins distinguishes the biosynthetic capabilities of eukaryotic cells from those of prokaryotes, which lack the intracellular structures and enzymes necessary for glycosylation. These modiied amino acid chains, or glycoproteins, encompass several important classes of macromolecules, including enzymes, hormones, immunoglobulins, transport proteins, cell adhesion molecules, and cytoplasmic proteins [2]. Glycoproteins are produced as pools of different glycoforms with varying glycan structures attached to a mostly invariant peptide backbone. The characteristic glycoform proile is dependent on the glycan structures added during cotranslation and modiied post-translationally. Variations may be found either in the site-occupancy (macroheterogeneity) or in the structures of attached glycans (microheterogeneity). The structural variability of glycans is generally dictated by a panel of highly speciic glycosyltransferases and glycosidases present in the endoplasmic reticulum and Golgi apparatus of the eukaryotic cells [3–5]. Variations in structure and degree of glycosylation site can contribute to both biological function and mass heterogeneity. Glycan structures are important also because they can inluence many of the biological properties of each given glycoprotein, including pharmacokinetics, bioactivity, secretion, in vivo clearance, solubility, receptor recognition, and antigenicity [6–9]. Elucidating the structures of various glycoproteins is therefore an essential component of current biological studies, and constitutes a large part of the ield of glycomics. In conjunction with genomics and proteomics, glycomics completes a triad of disciplines that investigate the structure and function of biological macromolecules. The three most abundant types of glycan found in glycoproteins differ by the nature of their linkage to the protein: N-linked; O-linked; and glycosylphosphatidylinositol (GPI) lipid anchors. MALDI MS: A Practical Guide to Instrumentation, Methods, and Applications, Second Edition. Edited by Franz Hillenkamp and Jasna Peter-Katalinic. © 2014 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2014 by Wiley-VCH Verlag GmbH & Co. KGaA.

240

6 MALDI-MS of Glycans and Glycoconjugates

N-linked glycans have N-acetylglucosamine (GlcNAc) covalently attached to asparagine through an amide bond. These are the most widely studied structural forms of glycosylation, and have the greatest effect on overall protein structure and function [10, 11]. There must be a tripeptide sequon with the consensus sequence Asn-X-Ser/Thr, where X can be any amino acid except proline. N-glycans are modiied from an original consensus 14mer oligosaccharide (Glc3Man9GlcNAc2); the resulting N-glycans are based on the characteristic pentasaccharide core, Man3-GlcNAc2. Further processing in the Golgi apparatus results in three main classes of N-linked glycans: high-mannose, hybrid, and complex [12]. The high-mannose and hybrid types exist mostly as bi- and triantennary forms, while complex glycans can be present as bi-, tri-, and tetra-antennary structures. N-glycans are highly associated with various biological functions of cells, including cell adhesion and cancer metastasis [13–15]. Although several studies have emphasized the N-linked glycosylation of proteins, O-linked glycans are smaller structurally but are equally important in eukaryotic glycoproteins. The most common form of O-glycan is the mucin-type, which involves the attachment of N-acetylgalactosamine (GalNAc) to a serine or threonine residue in a protein [16, 17]. This type of O-glycan is added posttranslationally to the fully folded protein. No consensus sequence has been identiied, although glycosylation often occurs in a region of the protein that contains a high proportion of serine, threonine, and proline [18]. The placement of Pro at either −1 or +3 positions relative to Ser or Thr is often favorable to O-glycosylation. These amino acid residues are thought to enable the region of the protein to assume a conformation accessible to GalNAc transferase. The most common mucin-type structure is the core 1 (Galβ1-3GalNAc) which may be either monoor di-sialylated [19]. O-glycans are normally found in most secretory cells and tissues, and may also be involved in hematopoiesis [20], in inlammation response mechanisms [21], and they can contribute to the formation of ABO blood antigens [22]. Core 2 glycans are the well-studied sialyl-Lewis X glycan structures that mediate cell–cell adhesion. Other O-linked glycans can have Gal, GlcNAc, Fuc, Man or Xyl (xylose) as the initial residue bound to the Ser/Thr residues. O-linked glycoproteins are usually large proteins with masses higher than 200 kDa, and O-glycans generally show less branching than N-glycans [20]. The three typical subgroups of N-linked sugar chains found in mammals, and the different cores found in mucin-type O-linked sugar chains are shown in Figure 6.1 [23]. The GPIs represent an abundant and ubiquitous class of eukaryotic glycolipids [24, 25]. GPI-anchored proteins are characterized by an amide bond formed between the carboxy-terminal residue of the protein and the amino group of the phosphoethanolamine linked to a trimannosyl-nonacetylated glucosamine core (Man3-GlcN). The reducing end of GlcN is attached to phosphatidylinositol (PI) that is then anchored through another phosphodiester linkage to the cell membrane through its hydrophobic region. The carbohydrate Man3-GlcN core may undergo various modiications during secretion from the cell. These types of glycans can also contribute to different functions, ranging from enzymatic to

6.1 Introduction (a)

N-glycans (mammalian)

Asn

High mannose

(b)

Asn

Asn

Hybrid

Complex

Mucin-type O-glycans S/T

Tn antigen

S/T

Core III

Core VI S/T

S/T Core I

S/T

Core IV

Core VII S/T

S/T Core II

S/T

Core V

Core VIII

S/T

N-Acetylglucosamine Galactose

Mannose

Fucose

N-Acetylgalactosamine

N-Acetyl neuraminic acid Figure 6.1 Structures of (a) N-glycans found in mammalian glycoproteins and (b) O-glycan

cores [12, 23].

antigenic activities and adhesion. GPI-anchored proteins also play a critical role in receptor-mediated signal transduction pathways [26]. 6.1.2 Glycosphingolipids

Glycosphingolipids (GSLs) are found in virtually all vertebrate cells and plasma membranes, where they act as binding sites on the cell surface. They are components of the outer lealet of the cell-membrane bilayer and are enriched in microdomains. The hydrophilic part of the GSL is composed of one to several monosaccharide units, whereas a sphingosine and a fatty acid are components of the hydrophobic portion, the ceramide. GSLs have been implicated to have functional relevance in mediating signal transduction, cell–cell recognition and

241

242

6 MALDI-MS of Glycans and Glycoconjugates

adhesion in the lipid microdomains [27]. The diversity of GSLs occurring in any given biological source is invariably relected in the heterogeneity of puriied GSL fractions with respect to both glycan and ceramide structure, and also in the dynamic range of single-component abundance in GSL mixtures. Carbohydrate chains containing different building blocks can be classiied according to their biological origin. In order to characterize all naturally occurring components in such mixtures, the current MS-based glycosphingolipidomic strategies include a preceding separation step to prove the existence of isomeric structures related to the attachment site within the sugar chain, that is, “a-” or “b-” type gangliosides and/or a different linkage site within a monosaccharide building block unit such as Neu5Acα2-3 versus Neu5Acα2-6 linkage, and so on [28, 29]. As generally accepted, the malignant transformation of cells is accompanied by an aberrant cell surface–glycocalyx molecular composition, which is particularly due to alterations in glycoconjugate glycosylation pathways. Various glycosyl epitopes in the GSLs constitute tumor-associated antigens; moreover, when highly expressed some of these can promote invasion and metastases, whereas others may suppress tumor progression [30]. Gangliosides are considered as potential therapeutic targets for cancer treatment, primarily for the production of anticancer ganglioside-based polyvalent vaccines, as shown by a strategy in the prevention of neuroblastoma relapse. Several ganglioside-based vaccines, in particular against melanoma, small-cell lung carcinoma and breast carcinoma, are currently undergoing clinical trials [31].

6.2 Proiling of Glycans and Glycosphingolipids 6.2.1 Importance of Glycan Proiling and Techniques Used for This Purpose

Glycoproiling is simpler than full structural analysis, and for some applications it is the method of choice, owing to the availability of an array of complementary techniques suitable for this type of analysis. Glycoproiling may provide a pattern of peaks or bands that can be unique to a particular glycoprotein or glycoform. Glycoproiles can provide the appropriate information to validate the consistency of production during a natural or recombinant bioprocess. They may be obtained with intact glycoproteins, although more detailed information becomes available through the selective removal of O- or N-linked glycans. Besides MALDI-MS (as will be discussed below), other popular glycoproiling techniques include two-dimensional (2-D) gels, lectin arrays [32], capillary electrophoresis [33], high-performance anion-exchange chromatography coupled with pulsed amperometric detection (HPAEC-PAD) [34], and high-performance liquid chromatography (HPLC) (employing weak-anion exchange [WAX], normal phase [35, 36], or reversed phase [37] systems).

6.2 Proiling of Glycans and Glycosphingolipids

6.2.2 Importance of Glycosphingolipid Proiling and Characterization; Techniques Used

Glycosphingolipid proiling deals with the structural issues of determining both the carbohydrate and lipid portions. Carbohydrate chains can be either linear or branched, and can reach up to 50 monosaccharide units in length. Moreover, they can be substituted by rather labile sialic acids or fucose at different locations. The lipid portion is heterogeneous in the sphingosine base moiety, and can contain different lengths of fatty acids linked by an amide bond. Glycosphingolipid proiling and characterization can be performed via a combination of analytical techniques, including high-performance thin-layer chromatography (HPTLC) and immunostaining, or in separations by HPLC or capillary electrophoresis (CE) prior to MS analysis. Several procedures have been developed and adopted for the eficient characterization of GSLs by MALDI-MS. One novel procedure of automation for the coupling of nano-high-performance liquid chromatography (nano-HPLC) with matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) via an automatic spotting robot device was developed and adapted for the analysis of complex mixtures of GSLs [38]. In one method, when HPTLC-separated GSL species were submitted to a direct MALDI-MS screening, this allowed a highly sensitive and complete characterization in which infrared laser desorption/ionization on an orthogonal-time-of-light (o-TOF) mass analyzer was shown to be instrumental [39]. Subsequently, both protocols were shown to be applicable for the assay of neutral and acidic GSLs in crude mixtures obtained from body luids or cell lines, with low femtomolar levels of individual GSL species detection. Hence, both methods have been proposed as eficient analytical tools for high-throughput glycosphingolipidomic projects. 6.2.3 MALDI-MS of Glycans and Glycoprotein Components

Since the late 1980s, fundamental studies in glycobiology have beneited greatly from the availability of techniques based on mass spectrometry (MS), such as electrospray ionization-MS (ESI-MS) and matrix-assisted laser desorption/ ionization-time-of-light-MS (MALDI-TOF-MS). In addition, the requirements of therapeutic applications have encouraged developments to be made in numerous disciplines, including the separation sciences and instrumentation related to these MS techniques. A combination of basic needs and application has, in recent years, resulted in improved in both MS technologies and high-throughput methods. Despite the availability of high-end instrumentation, however, the quantitative or qualitative determination of oligosaccharides in diverse types of glycoprotein samples remains a major challenge. Moreover, regardless of the sophistication, sensitivity and speciicity of the instruments utilized, the successful qualitative and quantitative determination of oligosaccharides from glycoproteins often depends heavily on sample preparation.

243

244

6 MALDI-MS of Glycans and Glycoconjugates

It is well known that MS provides many advantages over traditional analytical methods, and has been used extensively as the most sensitive technique for the analysis of carbohydrate molecules, especially during the last decade. Indeed, newly available MS instruments allow measurements to be made with complex mixtures, and even permit the identiication of trace amounts of glycans in biological materials [38]. In most cases, the MALDI-MS ingerprints of glycan mixtures will display a singly charged ionic signal per compound, and this tends to simplify the spectral interpretation with respect to ESI. For most experiments conducted on glycans, ubiquitous sodium ions from the glass vessels and other sources become attached to the glycan molecules to form (M + Na)+ ions. For direct measurements on neutral glycans, the (M + Na)+ abundances (peak area) may be compared in a semi-quantitative fashion within the same sample. The MALDI spectra of neutral glycans from different glycoproteins, immunoglobulin G and hen ovalbumin, obtained with two different mass spectrometers, are shown in Figure 6.2 [E. Lattova, unpublished data]. In the case of acidic glycans (e.g., sialylated), however, ionization by the addition of Na+ does not follow the same eficiency pattern, and these glycans must be considered separately (methods for the quantitative analysis of glycans are discussed in Section 6.4). An example of the effect of using different matrices on the ionization of tetrasialylated glycans is shown in Figure 6.3 [42]. It should be noted that, upon irradiation with the laser, not all of the sample will be consumed; consequently, when the initial measurements have been completed the material can be redissolved and treated chemically or enzymatically prior to further mass spectrometric measurements. In the case of glycans, the most common MALDI matrix to be used is 2,5-dihydroxybenzoic acid (DHB). The most recently used matrices have been reviewed [43], with mixtures of DHB and coumarin analogs having been reported to provide enhanced ionization conditions for dextran chains [44]. MALDI-MS measurements on instruments equipped with a linear TOF analyzer are often used to perform a primary characterization of intact glycoproteins. These experiments, when performed on a “clean” sample (i.e., a pure glycoprotein that is free from the presence of any nonvolatile buffers and detergents) can lead to useful information on the total molecular weight, the percentage of glycosylation in terms of weight (if the amino acid sequence is known), and the general levels of homogeneity (or heterogeneity) in protein glycosylation. Although protein puriication is largely beyond the scope of this chapter, a brief discussion will be provided nonetheless. Glycoproteins are available from a variety of biological origins, and may be cleaned and puriied in accordance with the medium from which they have been taken. The most common puriication methods include gel-permeation columns, bioafinity columns, electrophoretic gels, ion-exchange columns, dialysis, and combinations of these techniques. Denaturing agents such as sodium dodecyl sulfate or other detergents should be avoided or removed, because these can yield intense mass spectra themselves and quench the glycoprotein-related signals at any subsequent stage of the analysis.

6.2 Proiling of Glycans and Glycosphingolipids

1591.6

1429.5

1347.5 1372.5

1226.5

Intensity

1575.6

18

1899.7

1737.7

(a)

0 1000

2800

m/z

(b)

1835.7

Figure 6.2 MALDI-MS spectra of neutral

oligosaccharides obtained by PNGase F digestion from 100 μg of glycoprotein. (a) IgG1 (Herceptin), recorded on qQTOF instrument; (b) Ovalbumin, recorded on

m/z

2728.1

2566.0

2403.9 2362.9

2159.8

1997.7

2000.8 2241.8

2038.8

0.0 1000

1956.8

1753.701 1794.7

1347.5

0.4

1185.4

0.8

1509.6 1550.6 1591.6

1388.5 1429.5

1632.6

1226.4

1.2

1023.3

x10 5

2800

UltraleXtreme (Bruker). Oligosaccharides were labeled on-target with PHN without puriication. All ions are [M + Na]+. Not all isomeric structures are shown [E. Lattova, unpublished data].

The ability to obtain a molecular weight measurement for a puriied glycoprotein is useful but not always possible, especially if the amount of sample is limited and many sites are glycosylated with sialic acid residues. Today, linear MALDI-TOF instruments and most commercial ESI-MS systems will allow molecular weight measurements to be made, although ESI is less tolerant than MALDI to the presence of any residual salts, detergents, or other additives. Occasionally, a successful molecular weight determination will clearly show the glycoforms of the protein,

245

246

6 MALDI-MS of Glycans and Glycoconjugates

Figure 6.3 Linear mode (-) MALDI spectra of a sialylated tetra-antennary N-glycan mixture

obtained with (a) ATT/DAC (15 g l–1 20 mM) and (b) THAP/DACA (20 g l–1 20 mM) before and (c) after puriication with graphitized carbon. Reprinted with permission from Ref. [42].

6.2 Proiling of Glycans and Glycosphingolipids

Figure 6.4 (a) MALDI mass spectra

(sinapinic acid) of serum transferrins from a healthy control and two patients with CDG syndrome. Reprinted from Ref. [45] with permission; (b) Glycosylation pattern of an

intact monoclonal antibody obtained on the IR-MALDI-TOF instrument with orthogonal extraction. Reprinted from Ref. [46] with permission.

whereas in other cases a wide, unresolved peak, comprised of all combined glycoforms, will be produced. A typical analysis of intact transferrin, analyzed with linear-mode MALDI-MS, for control samples versus those of patients with a congenital disorder of glycosylation (CDG), is shown in Figure 6.4a [45]. The glycosylation pattern of an intact monoclonal antibody, as obtained using MALDIMS, is shown in Figure 6.4b [46]. Generally, for more detailed glycoprotein MS analyses, two main approaches exist: (i) glycan release followed by characterization of the whole pool; and (ii)

247

248

6 MALDI-MS of Glycans and Glycoconjugates

proteolytic digestion followed by the determination of glycopeptide compositions. The second of these approaches is very useful for obtaining information on sitespeciic variability. Typically, glycoforms of the same tryptic peptide will elute at approximately the same time on a C18 reversed-phase HPLC column, while the direct MALDI analysis of their fraction(s) will show typical glycosylation patterns, as depicted in Figure 6.5a [47]. On-target deglycosylation will result in the bare peptide (M + H)+ signal (Figure 6.5b), and phenylhydrazine tagging (Figure 6.5c) allows observation of the released glycans. 6.2.4 N- and O-Glycan Release

N-glycans are commonly released from glycoproteins by the endoglycosidase enzyme, Peptide N-glycosidase F (PNGase F), an amidase that cleaves between the innermost GlcNAc residue of the glycan and the asparagine residue of the protein (Figure 6.6) [48, 49]. This is a particularly valuable enzyme because it can release all N-glycan structures from mammalian-derived glycoproteins, including high-mannose, hybrid, and complex oligosaccharides. Hydrolysis liberates amino-oligosaccharides, which are slowly hydrolyzed by water and converted into forms with a reducing terminal GlcNAc [50]. As a result of this process, the glycan structures remain intact whilst the asparagine residue from which the sugar was removed is deaminated to aspartic acid, which is the only modiication to the protein. The reducing end of the glycan is then available for labeling – a step that is likely to be useful for subsequent analysis. In order for the glycan cleavage from the protein to occur eficiently, it is important to denature the glycoprotein so as to maximize the accessibility of the enzyme. In this case, reduction with dithiothrietol (DTT), followed by alkylation of exposed –SH groups with iodoacetamide, is suficient to unfold a protein and expose the glycans. Two types of PNGase are commonly used in this procedure, namely PNGase-F and PNGase-A. PNGase-F releases most N-glycans, except those containing a fucose residue linked α1-3 to the reducing GlcNAc terminal [51]. Such glycans are often found in plant and insect glycoproteins, and their release necessitates the use of PNGase-A [50, 52]. EndoH recombinant glycosidase is a more speciic enzyme than PNGase, and cleaves a glycosidic bond within the N-glycan chitobiose core to leave one terminal GlcNAc attached to the protein [53]. This may be a disadvantage, as information related to fucosylation at the reducing termini of glycans is lost. EndoH glycosidase also cleaves only high-mannose and some hybrid N-oligosaccharides, but not highly processed complex N-oligosaccharides. Hydrazinolysis is the most widely used chemical cleavage method, and can be applied to the release of O- and N-linked glycans [54]. This is the only chemical method to release glycans with the reducing terminal GlcNAc intact, as is necessary for glycan labeling. Unfortunately, hydrazinolysis suffers certain disadvantages, including a need for extreme caution during the preparation of reagents and

6.2 Proiling of Glycans and Glycosphingolipids

249

EEQFNSTFR

(a) 2764.4

100

2602.3

2926.7 2967.8

0 m/z

1000

(b)

3500

EEQFDSTFR 1158.5

100

0 1000

(c)

m/z

3500

EEQFDSTFR 1158.5

1899.8 1940.9

1737.7 1778.8

1575.5

100

0 1000

Figure 6.5 MALDI-MS spectra (Bruker,

Bilex-IV) acquired from an HPLC fraction (elution time 28 min) of polyclonal IgG. (a) After trypsin digestion, all ions are [M + H]+;

m/z

(b) After enzymatic deglycosylation with PNGaseF; (c) On-target derivatization with PHN, oligosaccharide ions are [M + Na]+. Adapted from Ref. [47] with permission.

3500

250

6 MALDI-MS of Glycans and Glycoconjugates

Figure 6.6 The hydrolysis target bond of the deglycosylation enzyme PNGase [48, 49].

samples, low overall reaction yields, the toxicity and volatility of hydrazine, and the nonreversible dismantling of the protein chain. At present, there appears to be no universal enzyme for the removal of O-glycans, although hydrazinolysis can be used chemically for the nonselective removal of both O- and N-glycans from glycoproteins. If the conditions are controlled carefully (i.e., mild temperature of 60 °C for 6 h), anhydrous hydrazine may be used to selectively release O-glycans [55]. In this situation, the dangers of “peeling” – that is, the breaking down of glycans by base-catalyzed β-elimination – can be avoided if the reaction is carefully monitored. The purity of the hydrazine used is also very important, with the water content controlled at