Mass Spectrometry Handbook [1 ed.] 047053673X, 9780470536735

Due to its enormous sensitivity and ease of use, mass spectrometry has grown into the analytical tool of choice in most

237 57 40MB

English Pages 1368 [1364] Year 2012

Report DMCA / Copyright

DOWNLOAD PDF FILE

Recommend Papers

Mass Spectrometry Handbook [1 ed.]
 047053673X, 9780470536735

  • 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

MASS SPECTROMETRY HANDBOOK

WILEY SERIES ON PHARMACEUTICAL SCIENCE AND BIOTECHNOLOGY: PRACTICES, APPLICATIONS AND METHODS Series Editor: Mike S. Lee Milestone Development Services Mike S. Lee • Integrated Strategies for Drug Discovery Using Mass Spectrometry Birendra Pramanik, Mike S. Lee, and Guodong Chen • Characterization of Impurities and Degradants Using Mass Spectrometry Mike S. Lee and Mingshe Zhu • Mass Spectrometry in Drug Metabolism and Disposition: Basic Principles and Applications Mike S. Lee (editor) • Mass Spectrometry Handbook

MASS SPECTROMETRY HANDBOOK

EDITED BY MIKE S. LEE Milestone Development Services

A JOHN WILEY & SONS, INC., PUBLICATION

Copyright © 2012 by John Wiley & Sons, Inc. All rights reserved. Published by John Wiley & Sons, Inc., Hoboken, New Jersey. Published simultaneously in Canada. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permissions. 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 may 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 author shall be liable for any loss of proit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com. Library of Congress Cataloging-in-Publication Data: Mass spectrometry handbook / edited by Mike S. Lee. p. cm. Includes index. ISBN 978-0-470-53673-5 (cloth) 1. Mass spectrometry–Handbooks, manuals, etc. I. Lee, Mike S., 1960– QD96.M3M36 2012 543'.65–dc23 2011034171 Printed in the United States of America. ISBN: 9780470536735 10

9

8

7

6

5

4

3

2

1

CONTENTS

FOREWORD

xi

PREFACE

xiii

CONTRIBUTORS

xvii

SECTION I 1

BIOTECHNOLOGY/PROTEINS

1

Targeted Proteomics Using Immunoafinity Puriication

3

Karen R. Jonscher, Lei Jin, John C. Cambier, Shaikh M. Rahman, and Jacob E. Friedman

2

Mass Spectrometry-Based Methods to Investigate Posttranslational Protein Modiications by Lipid Peroxidation Products

23

Navin Rauniyar and Laszlo Prokai

3

Imaging Mass Spectrometry (IMS) for Biological Application

41

Yuki Sugiura, Ikuko Yao, and Mitsutoshi Setou

4

Methodologies for Identifying Microorganisms and Viruses by Mass Cataloging of RNAs

85

George W. Jackson, Rafal Drabek, Mithil Soni, Roger McNichols, Richard C. Willson, and George E. Fox

SECTION II 5

PHARMACEUTICAL

Preclinical Pharmacokinetics: Industrial Perspective

107 109

Ayman El-Kattan and Manthena Varma

6

LC-MS in Drug Metabolism and Pharmacokinetics: A Pharmaceutical Industry Perspective

119

Wenying Jian, Wilson Shou, Richard W. Edom, Naidong Weng, and Mingshe Zhu

7

Quantitative Mass Spectrometry in Support of Pharmacokinetic Studies

171

Xiaoying Xu, Wenkui Li, and Francis L.S. Tse v

vi

8

Contents

Determination of Pharmacokinetic Parameters by HPLC-MS/MS and UPLC-MS/MS

191

Margrét Thorsteinsdóttir, Baldur Bragi Sigurðsson, and Gísli Bragason

9

Methods for Screening Enantioselective Interactions in the Solution Phase Using ESI-MS

209

Kevin A. Schug

10

Hydrogen/Deuterium Exchange Mass Spectrometry (HDX MS) in the Studies of Architecture, Dynamics, and Interactions of Biopharmaceutical Products

227

Igor A. Kaltashov, Cedric E. Bobst, and Rinat R. Abzalimov

11

TOF-SIMS Applications to Bioimaging and Biomolecule Evaluation Methods

243

Satoka Aoyagi

12

Accelerator Mass Spectrometry in Pharmaceutical Development

259

Benjamin J. Stewart, Graham Bench, Bruce A. Buchholz, Kurt W. Haack, Michael A. Malfatti, Ted J. Ognibene, and Kenneth W. Turteltaub

SECTION III 13

CLINICAL ANALYSIS

Mass Spectrometry in Clinical Analysis: Screening for Inborn Errors in Metabolism

271

273

Donald H. Chace

14

Mass Spectrometry for Steroid Analysis

297

William J. Grifiths, Michael Ogundare, Anna Meljon, and Yuqin Wang

SECTION IV 15

FORENSICS

Forensic Applications of Isotope Ratio Mass Spectrometry

339 341

Sarah J. Benson

16 Analysis of Triacetone Triperoxide Explosive by Mass Spectrometry

373

Michael E. Sigman and C. Douglas Clark

SECTION V 17

SPACE EXPLORATION

389

Mass Spectrometry in Solar System Exploration

391

Paul V. Johnson, Luther W. Beegle, and Isik Kanik

18 Application of GC × GC–TOFMS to the Characterization of Extraterrestrial Organic Matter

407

Jonathan S. Watson

SECTION VI 19

HOMELAND SECURITY

Methods of Mass Spectrometry in Homeland Security Applications Ünige A. Laskay, Erin J. Kaleta, and Vicki H. Wysocki

417

419

Contents

20

Homeland Security

441

Christina L. Crawford and Herbert H. Hill, Jr.

21

Mass Spectrometry in Homeland Security

477

Yasuaki Takada

22

Measurements of Surface Contaminants and Sorbed Organics Using an Ion Trap Secondary Ion Mass Spectrometer

491

Gary S. Groenewold, Anthony D. Appelhans, Garold L. Gresham, and John E. Olson

23

Determination of Actinides: Determination of Low-Concentration Urine Uranium 235/238 Isotope Ratios

509

R. Steven Pappas

SECTION VII 24

FOOD ANALYSIS

Mass Spectrometry in Agriculture, Food, and Flavors: Selected Applications

529

531

Maciej Stobiecki, Piotr Kachlicki, and Henryk Jeleń

25 Top-Down Proteomic Identiication of Protein Biomarkers of Food-Borne Pathogens Using MALDI-TOF-TOF-MS/MS

559

Clifton K. Fagerquist and Omar Sultan

SECTION VIII ENVIRONMENTAL 26

Determination of Dithiocarbamate Fungicides in Food by Hydrophilic Interaction Liquid Chromatography/Mass Spectrometry

575

577

Wolfgang Schwack

27

Disinfectant and By-Product Analysis in Water Treatment by Membrane Introduction Mass Spectrometry

593

Chongzheng Na and Terese M. Olson

28

Proton Transfer Reaction Mass Spectrometry (PTR-MS)

605

Yujie Wang, Chengyin Shen, Jianquan Li, Haihe Jiang, and Yannan Chu

29

Determination of Chlorinated Compounds in Dialysis Water and in Biological Fluids/Matrices

631

Diana Poli

SECTION IX 30

GEOLOGICAL

Mass Spectrometry Techniques for Analysis of Oil and Gas Trapped in Fluid Inclusions

645

647

Simon C. George, Herbert Volk, and Adriana Dutkiewicz

31

LA-MC-ICP-MS Applied to U-Pb Zircon Geochronology

675

Alain Cocherie and Michèle Robert

32

Hydrocarbon Processing Maoqi Feng, Thomas Andrews, and Eloy Flores III

707

vii

viii

33

Contents

Hydrocarbon Processing: MALDI-MS of Polydisperse Hydrocarbon Samples

725

Alan A. Herod

34

Renewable Energy: Mass Spectrometry in Biofuel Research

749

Ingvar Eide and Kolbjørn Zahlsen

SECTION X ARCHAEOLOGY

763

35

765

Mass Spectrometry in Archaeology Robert Hedges and James McCullagh

36 Archaeometric Data from Mass Spectrometric Analysis of Organic Materials: Proteins, Lipids, Terpenoid Resins, Lignocellulosic Polymers, and Dyestuff

797

Maria Perla Colombini, Francesca Modugno, and Erika Ribechini

37

Laser Ablation ICP-MS in Archaeology

829

Hector Neff

38

Spatially Resolved MS in the Study of Art and Archaeological Objects

845

Giuseppe Spoto

39

Laser Ablation–Inductively Coupled Plasma Mass Spectrometry for the Investigation of Archaeological Samples

859

Martín Resano, Esperanza García-Ruiz, and Frank Vanhaecke

SECTION XI 40

SURFACE ANALYSIS

885

Mass Spectrometry in Semiconductor Research

887

Stefan Flege and Wolfgang Ensinger

41 Analysis of Thin and Thick Films

943

Philippe Le Coustumer, Patrick Chapon, Agnès Tempez, Yuriy Popov, George Thompson, Igor Molchan, Nicolas Trigoulet, Peter Skeldon, Antonino Licciardello, Nunzio Tuccitto, Ivan Delfanti, Katrin Fuhrer, Marc Gonin, James Whitby, Markus Hohl, Christian Tanner, Nerea Bordel Garcia, Lara Lobo Revilla, Jorge Pisonero, Rosario Pereiro, Cristina Gonzalez Gago, Alfredo Sanz Medel, Mihai Ganciu Petcu, Ani Surmeian, Constantin Diplasu, Andreea Groza, Norbert Jakubowski, Roland Dorka, Stela Canulescu, Johann Michler, Philippe Belenguer, Thomas Nelis, Abdellatif Zahri, Philippe Guillot, Laurent Thérèse, Arnaud Littner, Richard Vaux, Julien Malherbe, Frédéric Huneau, Fred Stevie, and Hugues François-Saint-Cyr

42

SIMS for Organic Film Analysis

961

Taouiq Mouhib and Arnaud Delcorte

43

Ceramics: Contribution of Secondary Ion Mass Spectrometry (SIMS) to the Study of Crystal Chemistry of Mica Minerals

1017

Luisa Ottolini, Emanuela Schingaro, and Fernando Scordari

SECTION XII 44

POLYMERS

1061

ETV-ICPMS for Analysis of Polymers

1063

Maite Aramendía Marzo, Martín Resano, and Frank Vanhaecke

Contents

45

Polymers

1079

Maurizio S. Montaudo and Salvatore Battiato

46

Mass Spectroscopy in Polymer Research

1107

Jale Hacaloglu and Talat Yalcin

47

Laser Mass Spectrometry Applied to the Analysis of Polymers

1135

Jérôme Bour and David Ruch

SECTION XIII ANALYTICAL TECHNIQUES

1143

48

1145

Measuring Thermodynamic Properties of Metals and Alloys Evan H. Copland and Nathan S. Jacobson

49

High-Performance Thin-Layer Chromatography–Mass Spectrometry for Analysis of Small Molecules

1181

Gertrud E. Morlock

50

Laser Ionization Mass Spectrometry of Inorganic Ions

1207

Julius Pavlov and Athula B. Attygalle

51

Mass Spectrometry in the SSITKA Studies

1229

L.G. Pinaeva, E.M. Sadovskaya, A.P. Suknev, V.B. Goncharov, and B.S. Bal’zhinimaev

52

Proton Transfer Reaction Mass Spectrometry: Applications in the Life Sciences

1257

Elena Crespo, Marco M.L. Steeghs, Simona M. Cristescu, and Frans J.M. Harren

INDEX

1283

ix

FOREWORD

It is a pleasure to provide this foreword to the Handbook of Mass Spectrometry, edited by Dr. Mike S. Lee, a PhD graduate of my research group at the University of Florida 25 years ago. Mike is not only an outstanding scientist and a visionary in how mass spectrometry can drive science in a diverse range of disciplines; he is also a master at assembling and leading a team of experts, as he has ably demonstrated with this volume. Mass spectrometry, although barely a hundred years old, has become the dominant force in modern analytical chemistry. It provides unparalleled levels of sensitivity and selectivity for trace analysis, and an impressive range of capabilities and application. Some of these unique capabilities arise from the unique feature of mass spectrometry (compared to other spectrometric methods) that the sample itself (matter) passes through the spectrometer and is separated and detected. Thus mass spectrometry is both a spectrometric method and a separation method! Many of the capabilities of modern mass spectrometry arise from the remarkable advances in instrumentation over the past 30 years, many of which are reviewed in this handbook. Advances in ionization techniques have expanded the applicability of mass spectrometry from small, volatile, and thermally stable molecules to large, nonvolatile, and labile molecules, including intact proteins and polymers. The coupling of mass spectrometry with separation techniques (gas chromatography [GC], liquid chromatography [LC], capillary electrophoresis [CE], and even a second stage of mass spectrometry) has established it as the standard for trace mixture analysis. Innovations in mass analyzers continue to bring improved performance in terms of mass resolution, mass range, and sensitivity. And perhaps most impressively, the pace of advances in mass spectrometry instrumentation and methodologies has not slacked off—we continue to see remarkable advances every year.

I often date the “coming of age” of modern analytical mass spectrometry to a 1982 quote from Chemical & Engineering News: Mass spectrometry has advanced to the point that it’s no longer (as has been said) . . . “the method of choice – if there’s no other way.”

Indeed, mass spectrometry is the method of choice for an amazing range of applications, from structure determination of proteins to forensic toxicology, from fundamental studies of reaction kinetics to imaging tissues. And that breadth of use and dominance of mass spectrometry is well represented in the chapters assembled here. The remarkable growth of mass spectrometry is well represented in the growth of attendance at the Annual Meeting on Mass Spectrometry and Allied Topics of the American Society for Mass Spectrometry, from 700 attendees in the mid-1970s to 7000 today. This relects not only the expanding scope of application of the technique, but also the ease with which modern mass spectrometers can be mastered by users new to the ield, without needing to understand the underlying fundamentals. This handbook provides in its 13 sections and 52 chapters an excellent overview of that wide range of applications. The breadth of coverage makes this an excellent resource for practicing mass spectrometrists as well as to those new to the ield. Welcome to a hopefully stimulating journey through modern mass spectrometry and its breadth of applications! Richard A. Yost University of Florida October 2011 xi

PREFACE

Mass spectrometry is an integral part of modern research in academic, industrial, and clinical laboratories. The Handbook of Mass Spectrometry represents the current state-of-the-art practices in these laboratory settings. The purpose of the handbook is to provide a unique reference that allows for easy access to a variety of applications that involve mass spectrometry. The intent of the handbook is to provide a resource for beginners, practitioners, and experts to obtain vital background, current approaches, and real-world methodology. Further, the handbook can also be viewed as an interactive time capsule to perhaps delineate “where we are,” “where we came from,” and “where we are headed” with regard to these speciic applications—current and emerging. Thus, the handbook is not intended to be comprehensive, but rather to provide unique, in-depth information on speciic techniques and experiences. The evolution of mass spectrometry has been both dramatic and fascinating. Trace analytical measurement, speciically the demand for trace mixture analysis, has created an increased demand for this powerful tool. In many cases, the preference for the trace mixture sample type has transformed the mass spectrometer into a gold standard platform for qualitative and quantitative assays. In its simplest form, a mass spectrometer can be viewed as a molecular weighing machine. Much like we regularly weigh ourselves in the morning to provide an early, facile benchmark for personal health and wellbeing, mass spectrometers are being used for a similar function. Speciically, a mass spectrometer is routinely used to monitor the “well-being” of a speciic analyte. Moreover, the conirmation each analyte (structure or amount), or ensemble of analytes, often provides a sur-

rogate benchmark into a speciic process that relates to a biological or chemical condition. Regardless of the application, mass spectrometrybased methods can be organized into two areas of analytical focus: qualitative (“What is it?”) and quantitative (“How much is there?”) analysis. Similar to the building of a picture puzzle—starting with the edges (the molecular ion!) to deine the size of the puzzle and/or set a deined limit to where all remaining subsequent puzzle pieces (fragment ions!) may it inside the edges—the use of mass spectrometry provides a powerful way to quickly and conidently “deine the edges” by providing molecular weight information. Molecular weight can then become a surrogate for conirmation or even be used for the identiication of a targeted compound, particularly when used in conjunction with an authentic standard or chromatographic technique, for example. Advanced studies that involve two or more dimensions of mass analysis can also be used to obtain speciic structural detail (fragment ions that correspond to speciic pieces of the picture puzzle!) or more selectivity to enable powerful approaches for high throughput quantitation. Moreover, similar to how high-deinition televisions are improving our entertainment experience, the higher resolution mass spectrometry (and chromatography!) technologies are poised to provide a beneit to the scientiic community in perhaps a highly routine manner. Thus, the diverse contributions to the handbook are essentially uniied based on the puzzle analogy. Conident and deinitive “What is it?” and/or “How much is there?” information is obtained via molecular weight measurements provided by the mass spectrometer. The speciic mass spectrometer and, of course, xiii

xiv

PREFACE

speciic chemistries (i.e., sample preparation, chromatography, ionization) help to deine the analytical method. Although the handbook is not necessarily designed to be comprehensive, the contributions represent an impressive array of critical work from diverse areas ranging from biological studies to food analysis to environmental analysis to archaeology. Each chapter in the handbook contains several compulsory elements: (1) essential background and history of the application; (2) detailed analytical methodology; and inally, (3) valuable references for more in-depth study. Each contributor has provided critical updates in their respective ield of expertise. Both current and emerging trends are highlighted. Perhaps a distinguishing feature of the handbook is that nearly all of the chapters provide a detailed description of the actual methodologies used in their respective laboratory— speciically intended so that others may initiate similar work in their respective laboratory. We hope that this unique feature will allow broad base interest and use for all scientists! Certainly, the handbook is quite diverse in scope and application. The handbook is organized into 13 sections—starting with life sciences and culminating with specialized analytical techniques. Section I provides an exciting perspective on the recent applications of mass spectrometry for the identiication of proteins and peptides. These methods represent the emerging role of mass spectrometry in biology-related ields to assist with the determination of both process and function. The section also provides the recent methodology used for imaging studies on biological systems as well as the proiling of microorganisms and viruses. The current state-of-the-art work performed in the pharmaceutical industry is featured in Section II. A continuum of work that begins with drug discovery activities such as pharmacokinetics (surrogate studies to determine dosing regimen in humans) as well as mass spectrometry methods for screening, characterization, and imaging are featured in Section II. The pharmaceutical section concludes with perspectives into drug development with the use of accelerator mass spectrometry. Exciting growth and, perhaps, a renaissance, is currently experienced in the ield of clinical analysis. Section III provides a timely and critical update on the use of mass spectrometry for the screening of inborn errors and steroid analysis in a clinical laboratory setting. The distinct criteria and features necessary for a clinical laboratory—as opposed to a research setting—are powerfully represented and easily understood. Forensics is indeed a challenging area of focus that requires diverse analytical tools as well as a strict protocol of analysis—from sampling to preparation to analysis to reporting. Section

IV contains two important applications of mass spectrometry in this ield. The use of isotope ratio mass spectrometry is highlighted followed by a speciic application that describes the analysis of the explosive triacetone triperoxide. Section V addresses the important role of mass spectrometry in programs involved with space exploration. A fascinating perspective on the use of mass spectrometry for solar system exploration is provided. This chapter is followed by work that features the use of gas chromatography (GC)/gas chromatography–mass spectrometry (GC-MS) for the characterization of extraterrestrial organic matter. Travel and safety has been greatly impacted over the past decade. Section VI contains the recent work that describes the various uses of mass spectrometry for homeland security. Speciic methods are detailed along with the requirements and challenges for this specialized application. The safety of our food and subsequent food supply is of critical worldwide importance. The role of mass spectrometry for food analysis is highlighted in Section VII. A perspective on agriculture, food and lavors is provided to give the reader some historical perspectives and background in food analysis. The recent mass spectrometry application of “top-down” proteomic methods for the identiication of biomarkers of foodborne pathogens highlights future direction and analysis formats. Perhaps a cornerstone of commercial applications of mass spectrometry is in the ield of environmental analysis. Section VIII contains the recent work that details how mass spectrometry is used to monitor targeted analytes such as fungicides, commercial by-products, and targeted carcinogens. Section IX focuses on geology. In this section, the authors provide their unique perspective on mass spectrometry applications that address the analysis of oil and gas, geochronology, and hydrocarbon processing. The section concludes with a chapter on the current status and prospects for renewable energy. Mass spectrometry methods have made signiicant contributions to archaeology. Section X focuses on recent work to give the reader historical and background information as well as speciic studies that require careful ield work (collection of the actual samples!) along with trace analysis using mass spectrometry-based methods. Surface analysis is a challenging area of study with very speciic criteria for analysis. Section XI provides perspective and recent methods in the area of semiconductor research, organic ilm analysis, and characterization of ceramic materials. Section XII provides perspective on the role and uses of mass spectrometry in polymer research. Background and methodology are highlighted from three leading laboratories. Specialized analytical techniques are presented in Section XIII. The section begins with a chapter on the approaches used for the measurement of metals and alloys followed by a variety

PREFACE

of interesting techniques that involve the use of thin layer chromatography, laser ionization, steady-state isotopic transient kinetic analysis, and proton transfer reaction mass spectrometry. It is my sincere hope that the handbook provides the information and details to assist scientists with current work as well as inspire future studies. Also, because of the vast content of work, it is hoped that seemingly unrelated applications provide helpful insight into novel uses of mass spectrometry and promote new areas of research. Finally, I wish to acknowledge the contributions of many—authors, collaborators, editors, and families—

xv

who made this handbook possible. Also, along with the terriic editorial staff at John Wiley & Sons, I would like to give a special acknowledgment to Gladys Mok, Managing Editor at John Wiley & Sons, for her signiicant contributions and premier support during this project. Mike S. Lee Milestone Development Services August 2011

CONTRIBUTORS

Rinat R. Abzalimov, PhD, Department of Chemistry, University of Massachusetts-Amherst, Amherst, MA

Sarah J. Benson, PhD, Australian Federal Police, Forensic & Data Centres, Canberra, ACT, Australia

Thomas Andrews, Division of Chemistry and Chemical Engineering, Southwest Research Institute, San Antonio, TX

Cedric E. Bobst, PhD, Department of Chemistry, University of Massachusetts-Amherst, Amherst, MA

Satoka Aoyagi, PhD, Department of Regional Development, Faculty of Life and Environmental Science, Shimane University, Matsue-shi, Shimane, Japan Anthony D. Appelhans, Idaho National Laboratory, Interfacial Chemistry Department, Idaho Falls, ID Maite Aramendía Marzo, PhD, Centro Universitario de la Defensa, Academia General Militar, Carretera de Huesca, Zaragoza, Spain Athula B. Attygalle, PhD, Center for Mass Spectrometry, Department of Chemistry, Chemical Biology, and Biomedical Engineering, Stevens Institute of Technology, Hoboken, NJ B.S. Bal’zhinimaev, PhD, Boreskov Institute of Catalysis, Siberian Branch of the Russian Academy of Sciences, Novosibirsk, Russia Salvatore Battiato, Institute of Chemistry and Technology of Polymers, National Research Council of Italy, Catania, Italy Luther W. Beegle, PhD, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA Philippe Belenguer, PhD, University of Toulouse, France Graham Bench, PhD, Lawrence Livermore National Laboratory, Center for Accelerator Mass Spectrometry, Livermore, CA

Nerea Bordel Garcia, PhD, University of Oviedo, Spain Jérôme Bour, PhD, Department of Advanced Materials and Structures, Centre de Recherche Public Henri Tudor (CRPHT), Esch sur Alzette, Luxembourg Gísli Bragason, BSc, MBA, ArcticMass, Sturlugata, Reykjavik, Iceland Bruce A. Buchholz, PhD, Lawrence Livermore National Laboratory, Center for Accelerator Mass Spectrometry, Livemore, CA John C. Cambier, PhD, Integrated Department of Immunology, National Jewish Medical and Research Center, Denver, CO Stela Canulescu, PhD, EMPA Materials Science and Technology, Switzerland Donald H. Chace, PhD, MSFS, Pediatrix Analytical, The Center for Research and Education, Pediatrix Medical Group, Sunrise, FL Patrick Chapon, Horiba Jobin Yvon, Longjumeau, France Yannan Chu, PhD, Laboratory of Environmental Spectroscopy, Anhui Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Hefei, China C. Douglas Clark, BS, National Center for Forensic Science, University of Central Florida, Orlando, FL Alain Cocherie, PhD, BRGM, Orleans, France xvii

xviii

CONTRIBUTORS

Maria Perla Colombini, PhD, Dipartimento di Chimica e Chimica Industriale, Università di Pisa, Pisa, Italy

Eloy Flores III, Division of Chemistry and Chemical Engineering, Southwest Research Institute, San Antonio, TX

Evan H. Copland, PhD, CSIRO, Sydney, Australia

George E. Fox, PhD, Department of Biology and Biochemistry, University of Houston, Texas

Christina L. Crawford, Department of Chemistry, Washington State University, Pullman, WA

Hugues François-Saint-Cyr, PhD, CAMECA, France

Elena Crespo, PhD, Life Science Trace Gas Facility, Molecular and Laser Physics, Institute of Molecules and Materials, Radboud University, Nijmegen, The Netherlands

Jacob E. Friedman, PhD, Departments of Pediatrics, Biochemistry and Molecular Genetics, and Reproductive Sciences, University of Colorado School of Medicine, Aurora, CO

Simona M. Cristescu, PhD, Life Science Trace Gas Facility, Molecular and Laser Physics, Institute of Molecules and Materials, Radboud University, Nijmegen, The Netherlands

Katrin Fuhrer, PhD, Tofwerk AG, Switzerland

Arnaud Delcorte, PhD, Institute of Condensed Matter and Nanosciences—Bio and Soft Matter (IMCN/ BSMA), Université catholique de Louvain, Croix de Sud, Louvain-la-Neuve, Belgium Ivan Delfanti, PhD, University of Catania, Italy Constantin Diplasu, PhD, National Institute of Lasers, Plasmas and Radiation Physics, Romania

Mihai Ganciu Petcu, PhD, National Institute of Lasers, Plasmas and Radiation Physics, Romania Esperanza García-Ruiz, PhD, Department of Analytical Chemistry, University of Zaragoza, Zaragoza, Spain Simon C. George, PhD, Department of Earth and Planetary Sciences, Macquarie University, Sydney, NSW, Australia

Roland Dorka, PhD, ISAS Dortmund, Germany

V.B. Goncharov, PhD, Boreskov Institute of Catalysis, Siberian Branch of the Russian Academy of Sciences, Novosibirsk, Russia

Rafal Drabek, PhD, BioTex, Inc, Houston, TX

Marc Gonin, PhD, Tofwerk AG, Switzerland

Adriana Dutkiewicz, PhD, School of Geosciences, University of Sydney, Sydney, NSW, Australia

Cristina Gonzalez Gago, PhD, University of Oviedo Spain

Richard W. Edom, PhD, Janssen Development, Raritan, NJ

Research

&

Ingvar Eide, PhD, Statoil ASA, Research Centre, Department of Energy and Environment, Trondheim, Norway Ayman El-Kattan, PhD, Department of Pharmacokinetics, Dynamics and Metabolism, Pizer Global Research and Development, Groton, CT Wolfgang Ensinger, PhD, Department of Materials Science,Technische Universität Darmstadt, Darmstadt, Germany Clifton K. Fagerquist, PhD, United States Department of Agriculture, Western Regional Research Center, Agricultural Research Service, Albany, CA Maoqi Feng, PhD, Division of Chemistry and Chemical Engineering, Southwest Research Institute, San Antonio, TX Stefan Flege, PhD, Department of Materials Science, Technische Universität Darmstadt, Darmstadt, Germany

Garold L. Gresham, Idaho National Laboratory, Interfacial Chemistry Department, Idaho Falls, ID William J. Grifiths, PhD, Institute of Mass Spectrometry, School of Medicine, Swansea University, Swansea, Wales, UK Gary S. Groenewold, PhD, Idaho National Laboratory, Interfacial Chemistry Department, Idaho Falls, ID Andreea Groza, PhD, National Institute of Lasers, Plasmas and Radiation Physics, Romania Philippe Guillot, PhD, University of Albi, France Kurt W. Haack, PhD, Lawrence Livermore National Laboratory, Biosciences and Biotechnology Division, Livermore, CA Jale Hacaloglu, PhD, Chemistry Department, Middle East Technical University, Ankara, Turkey Frans J.M. Harren, PhD, Life Science Trace Gas Facility, Institute for Molecules and Materials, Radboud University, Nijmegen, The Netherlands Robert Hedges, PhD, Research Laboratory for Archaeology, University of Oxford, Oxford, UK

CONTRIBUTORS

xix

Alan A. Herod, PhD, Chemical Engineering Department, Imperial College London, London, UK

Arnaud Littner, PhD, ALMA Consulting Group, Gennevilliers, France

Herbert H. Hill, Jr., PhD, Department of Chemistry, Washington State University, Pullman, WA

Lara Lobo Revilla, PhD, University of Oviedo, Spain

Markus Hohl, PhD, Tofwerk AG, Switzerland Frédéric Huneau, PhD, University of Bordeaux, France

Michael A. Malfatti, PhD, Lawrence Livermore National Laboratory, Biosciences and Biotechnology Division, Livermore, CA

George W. Jackson, PhD, BioTex, Inc, Houston, TX

Julien Malherbe, PhD, NIST Washington, DC

Nathan S. Jacobson, PhD, National Aeronautics and Space Administration, Glenn Research Center, Cleveland, OH

James McCullagh, PhD, Chemistry Research Laboratory, Department of Chemistry, University of Oxford, Oxford, UK

Norbert Jakubowski, PhD, ISAS Dortmund, Germany

Roger McNichols, PhD, BioTex, Inc, Houston, TX

Henryk Jeleń, PhD, Faculty of Food Science and Nutrition, University of Life Sciences, Poznań, Poland

Anna Meljon, PhD, Institute of Mass Spectrometry, School of Medicine, Swansea University, Swansea, Wales, UK

Wenying Jian, PhD, Janssen Research & Development, Raritan, NJ Haihe Jiang, PhD, Laboratory of Environmental Spectroscopy, Anhui Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Hefei, China Lei Jin, PhD, Integrated Department of Immunology, University of Colorado School of Medicine, National Jewish Medical and Research Center, Denver, CO Paul V. Johnson, PhD, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA Karen R. Jonscher, PhD, Department of Anesthesiology, University of Colorado School of Medicine, Aurora, CO Piotr Kachlicki, PhD, Institute of Plant Genetics, Polish Academy of Sciences, Poznań, Poland Erin J. Kaleta, PhD, Department of Chemistry and Biochemistry, University of Arizona, Tucson, AZ Isik Kanik, PhD, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA Igor A. Kaltashov, PhD, Department of Chemistry, University of Massachusetts-Amherst, Amherst, MA Ünige A. Laskay, PhD, Department of Chemistry and Biochemistry, University of Arizona, Tucson, AZ Philippe Le Coustumer, PhD, University of Bordeaux, France Jianquan Li, PhD, Laboratory of Environmental Spectroscopy, Anhui Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Hefei, PR, China Wenkui Li, PhD, Novartis Pharmaceuticals Corporation, Drug Metabolism & Bioanalytics, East Hanover, NJ Antonino Licciardello, PhD, University of Catania, Italy

Johann Michler, PhD, EMPA Materials Science and Technology, Switzerland Francesca Modugno, PhD, Dipartimento di Chimica e Chimica Industriale, Università di Pisa, Pisa, Italy Igor Molchan, PhD, The University of Manchester, UK Maurizio S. Montaudo, PhD, Institute of Chemistry and Technology of Polymers, National Research Council of Italy, Catania, Italy Gertrud E. Morlock, PhD, Institute of Food Chemistry, University of Hohenheim, Stuttgart, Germany Taouiq Mouhib, PhD, Institute of Condensed Matter and Nanosciences—Bio and Soft Matter (IMCN/ BSMA), Université catholique de Louvain, Croix de Sud, Louvain-la-Neuve, Belgium; Ecole Supérieure de Technologie, Université Hassan, Berrechid, Morocco Chongzheng Na, PhD, Department of Civil Engineering and Geological Sciences, University of Notre Dame, Notre Dame, IN Hector Neff, PhD, Department of Anthropology, Institute for Integrative Research in Materials, Environments, and Society (IIRMES), California State University-Long Beach, Long Beach, CA Thomas Nelis, PhD, University of Toulouse, France Ted J. Ognibene, PhD, Lawrence Livermore National Laboratory, Center for Accelerator Mass Spectrometry, Livermore, CA Michael Ogundare, PhD, Institute of Mass Spectrometry, School of Medicine, Swansea University, Swansea, Wales, UK. John E. Olson, PhD, Idaho National Laboratory, Interfacial Chemistry Department, Idaho Falls, ID

xx

CONTRIBUTORS

Terese M. Olson, PhD, Department of Civil and Environmental Engineering, University of Michigan, Ann Arbor, MI Luisa Ottolini, Consiglio Nazionale delle Ricerche (CNR)-Istituto di Geoscienze e Georisorse, Sezione di Pavia, Pavia, Italy R. Steven Pappas, PhD, U.S. Centers for Disease Control and Prevention, National Center for Environmental Health, Division of Laboratory Sciences, Emergency Response and Air Toxicants Branch, Atlanta, GA Julius Pavlov, Center for Mass Spectrometry, Department of Chemistry, Chemical Biology, and Biomedical Engineering, Stevens Institute of Technology, Hoboken, NJ Rosario Pereiro, PhD, University of Oviedo, Spain L.G. Pinaeva, PhD, Boreskov Institute of Catalysis, Siberian Branch of the Russian Academy of Sciences, Novosibirsk, Russia Jorge Pisonero, PhD, University of Oviedo, Spain Diana Poli, PhD, Department of Clinical Medicine, Nephrology, and Health Sciences, University of Parma, Parma, Italy Yuriy Popov, Horiba Jobin Yvon, France Laszlo Prokai, PhD, DSc, Department of Molecular Biology & Immunology, University of North Texas Health Science Center at Fort Worth, Fort Worth, TX Shaikh M. Rahman, PhD, Department of Pediatrics, University of Colorado School of Medicine, Aurora, CO Navin Rauniyar, PhD, Department of Molecular Biology & Immunology, University of North Texas Health Science Center at Fort Worth, Fort Worth, TX

Emanuela Schingaro, Dipartimento Geomineralogico, Università degli Studi di Bari, Bari, Italy Kevin A. Schug, PhD, Department of Chemistry & Biochemistry, The University of Texas at Arlington, Arlington, TX Wolfgang Schwack, PhD, University of Hohenheim, Institute of Food Chemistry, Stuttgart, Germany Fernando Scordari, Dipartimento Geomineralogico, Università degli Studi di Bari, Bari, Italy Mitsutoshi Setou, PhD, Hamamatsu University School of Medicine, Department of Molecular Anatomy, Hamamatsu, Shizuoka, Japan Chengyin Shen, PhD, Laboratory of Environmental Spectroscopy, Anhui Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Hefei, China Wilson Shou, PhD, Applied Biotechnology, BristolMyers Squibb, Wallingford, CT Michael E. Sigman, PhD, Department of Chemistry and National Center for Forensic Science, University of Central Florida, Orlando, FL Baldur Bragi Sigurðsson, MSc, ArcticMass, Sturlugata, Reykjavik, Iceland Peter Skeldon, PhD, The University of Manchester, UK Mithil Soni, PhD, BioTex, Inc, Houston, TX Giuseppe Spoto, PhD, Dipartimento di Scienze Chimiche, Università di Catania, Catania, Italy Marco M.L. Steeghs, PhD, Life Science Trace Gas Facility, Molecular and Laser Physics, Institute of Molecules and Materials, Radboud University, Nijmegen, The Netherlands

Martín Resano, PhD, Department of Analytical Chemistry, University of Zaragoza, Zaragoza, Spain.

Fred Stevie, PhD, North Carolina State University, Raleigh, NC

Erika Ribechini, PhD, Department of Chemistry and Industrial Chemistry, University of Pisa, Pisa, Italy

Benjamin J. Stewart, PhD, Lawrence Livermore National Laboratory, Center for Accelerator Mass Spectrometry, Livermore, CA

David Ruch, Department of Advanced Materials and Structures, Centre de Recherche Public Henri Tudor (CRPHT), Esch sur Alzette, Luxembourg Michèle Robert, engineer, BRGM, Orléans, France E.M. Sadovskaya, PhD, Boreskov Institute of Catalysis, Siberian Branch of the Russian Academy of Sciences, Novosibirsk, Russia Alfredo Sanz Medel, PhD, University of Oviedo, Spain

Maciej Stobiecki, PhD, Institute of Bioorganic Chemistry, Polish Academy of Sciences, Poznań, Poland Yuki Sugiura, PhD, Hamamatsu University School of Medicine, Department of Molecular Anatomy, Hamamatsu, Shizuoka, Japan A.P. Suknev, PhD, Boreskov Institute of Catalysis, Siberian Branch of the Russian Academy of Sciences, Novosibirsk, Russia

CONTRIBUTORS

Omar Sultan, MS, United States Department of Agriculture, Western Regional Research Center, Agricultural Research Service, Albany, CA Ani Surmeian, PhD, National Institute of Lasers, Plasmas and Radiation Physics, Romania Yasuaki Takada, PhD, Hitachi Ltd, Central Research Laboratory, Kokubunji-shi, Tokyo, Japan Christian Tanner, Tofwerk AG, Switzerland Agnès Tempez, PhD, Horiba Jobin Yvon, France Laurent Thérèse, PhD, University of Albi, France George Thompson, PhD, The University of Manchester, UK Margrét Thorsteinsdóttir, Cand.Pharm., PhD, Faculty of Pharmaceutical Sciences, University of Iceland, Hagi, Reykjavik, Iceland Nicolas Trigoulet, PhD, The University of Manchester, UK Francis L.S. Tse, PhD, Novartis Pharmaceuticals Corporation, Drug Metabolism & Bioanalytics, East Hanover, NJ Nunzio Tuccitto, PhD, University of Catania, Italy Kenneth W. Turteltaub, PhD, Lawrence Livermore National Laboratory, Biosciences and Biotechnology Division, Livermore, CA

xxi

Yujie Wang, PhD, Laboratory of Environmental Spectroscopy, Anhui Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Hefei, China Yuqin Wang, PhD, Institute of Mass Spectrometry, School of Medicine, Swansea University, Swansea, Wales, UK. Jonathan S. Watson, PhD, Planetary and Space Sciences Research Institute, The Open University, Milton Keynes, Buckinghamshire, UK Naidong Weng, PhD, Janssen Research & Development, Raritan, NJ James Whitby, PhD, Tofwerk AG, Switzerland; EMPA Materials Science and Technology, Switzerland Richard C. Willson, PhD, Department of Chemical and Biomolecular Engineering, Department of Biology and Biochemistry, University of Houston; The Methodist Hospital Research Institute, Houston, TX Vicki H. Wysocki, PhD, Department of Chemistry and Biochemistry, University of Arizona, Tuscon, Arizona Xiaoying Xu, PhD, China Novartis Institute for BioMedical Research Co., Ltd., Shanghai Pudong Software Park, Zhangjiang Hi-Tech Park, Pudong, China Talat Yalcin, PhD, Chemistry Department, Izmir Institute of Technology, İzmir, Turkey

Frank Vanhaecke, PhD, Department of Analytical Chemistry, Ghent University, Ghent, Belgium

Ikuko Yao, PhD, Department of Medical Chemistry, Kansai Medical University, Moriguchi, Osaka, Japan

Manthena Varma, PhD, Department of Pharmacokinetics, Dynamics and Metabolism, Pizer Global Research and Development, Groton, CT

Kolbjørn Zahlsen, SINTEF Materials & Chemistry, Department of Biotechnology, Trondheim, Norway

Richard Vaux, Alma Consulting Group, Lyon, France Herbert Volk, CSIRO Earth Science and Resource Engineering, North Ryde, NSW, Australia

Abdellatif Zahri, PhD, University of Toulouse, France Mingshe Zhu, PhD, Department of Biotransformation, Bristol-Myers Squibb Research and Development, Princeton, NJ

SECTION I BIOTECHNOLOGY/PROTEINS

1 TARGETED PROTEOMICS USING IMMUNOAFFINITY PURIFICATION Karen R. Jonscher, Lei Jin, John C. Cambier, Shaikh M. Rahman, and Jacob E. Friedman

1.1

INTRODUCTION

Proteins are multimodular and multifunctional, interacting in complex networks that drive cellular function. Pathological alterations in signaling networks are thought to result in a number of diseases, particularly cancer. Understanding the roles and consequences of protein–protein interactions is therefore a fundamental goal in systems biology. The two-hybrid approach [1] emerged in the early 1990s as the irst method to assay whether two proteins interact in a pair-wise fashion. A number of bait–target strategies were subsequently developed [2], including techniques exploiting afinity puriications coupled to mass spectrometry (MS) to rapidly identify potentially novel protein interactions. Initial studies were performed in yeast [3,4] and were subsequently expanded to mammalian models [5]. Since MS-based proteomics is not necessarily limited to speciic sites or to speciic proteins, it represents an unbiased and direct approach to studying cellular processes [6]. As recently described [6,7], immunoafinity puriication has emerged as the most frequently employed method for multiprotein complex puriication. Its success is based on the principle that multiple members of a complex may be captured when one complex member is enriched, regardless of whether the complexed proteins are directly bound to the target protein. Additionally, puriication of posttranslational modiications has been used extensively to globally proile modiied proteins throughout cellular networks [8,9] and

provides invaluable insights into signal transduction mechanisms. A summary of typical steps employed to generate samples using an immunoafinity-based approach is illustrated in Figure 1.1 and described in detail for two example applications below. Following puriication, peptide mixtures resulting from the digestion of bands or eluates are analyzed using tandem mass spectrometry (MS/MS) and proteins are identiied by database searching and spectral matching. A gel-based approach may be useful when two conditions are being compared—bands exhibiting visual differences can be excised to yield data most likely to contrast biologically signiicant results (note that interesting low abundance proteins may be covered up by more abundant nonspeciic proteins). Another useful method, gel-enhanced liquid chromatography–tandem mass spectrometry (GeLC-MS/MS), has also emerged for the analysis of complex protein mixtures [10] and can be applied to the separation of immunoafinity eluates. In this approach, a protein-containing gel lane is chopped into equivalent sections, digested, and peptide mixtures analyzed. When complexed protein levels are extremely low or sample is limited, elution followed by in-solution digestion may provide a better option, as less protein is lost to sample handling. An important caveat to note is that a protein of interest may be “covered up” by comigrating background or nonspeciic proteins. This is a particular concern for proteins that may comigrate with immunoglobulin (Ig)

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

3

4

TARGETED PROTEOMICS USING IMMUNOAFFINITY PURIFICATION

dŝƐƐƵĞŽƌĐĞůůƐ

/ŶĐƵďĂƚĞǁŝƚŚ ƉƌŝŵĂƌLJĂŶƟďŽĚLJ ;ĨƌĞĞŽƌďĞĂĚďŽƵŶĚͿ

,ŽŵŽŐĞŶŝnjĞͬůLJƐĞ &ƌĂĐƟŽŶĂƚĞ KƌŐĂŶĞůůĞ EƵĐůĞĂƌͬĐLJƚŽƉůĂƐŵŝĐͬ ŵĞŵďƌĂŶĞ tŚŽůĞĐĞůůůLJƐĂƚĞ

ŝŐĞƐƚ

ůƵƚĞ

ŝŶĚƉƌŽƚĞŝŶƐ͕ ŵŽĚŝĮĐĂƟŽŶƐ͕Žƌ ĐŽŵƉůĞdžĞƐ

ŶƌŝĐŚǁŝƚŚ WƌŽƚĞŝŶͬ'ďĞĂĚƐ

FIGURE 1.1 Flowchart of sample preparation steps for immunoafinity puriication worklows. Tissue is homogenized or cells are lysed in an appropriate buffer using protease inhibitors. Whole cell lysates may be used, or subcellular enrichment of organelles, membrane, or cytoplasmic fractions, for example, may be employed for subsequent processing. Primary antibody recognizing the protein or modiication of interest is either bound to protein A- or protein G-coated beads or added to the lysate, allowing either target proteins or modiications to bind. Bead-bound protein A or protein G interacts with the primary antibody, enriching the protein complexes through a series of washes. Finally, proteins are eluted from the beads. The protein mixture, also containing the antibody, is either directly digested with an enzyme or separated by 1D gel electrophoresis, with the protein-containing bands of interest subsequently excised and digested.

heavy and light chains. Using a cross-linker such as DMP (dimethyl pimelimidate) to bind the primary antibody to Protein A will suppress elution of Ig chains when nonreducing conditions are used for elution. However, cross-linking may result in loss of afinity; an optimization worklow should ideally include both cross-linked and noncross-linked trials. The primary challenge of immunoafinity-based worklows lies in the dificulty of separating true low abundance interactors from nonspeciically binding proteins. Use of negative controls, such as preimmune sera or antibodies against other proteins, or, if the model allows, using a knockout or knock down of the protein of interest, can help separate out these background proteins. As described below, cross-linking the antibody, minimizing incubation times and antibody concentrations, optimizing wash buffer stringency, and other approaches may help mitigate the extent of nonspeciic binding. Assessing the utility of at least a few of these parameters should be included in the optimization worklow. A good way to begin optimizing the protocol is to immunoprecipitate the protein of interest and probe for a known interactor using Western blot. Begin

by titrating the primary antibody and beads to ind the minimum amount required to effectively immunodeplete the sample. Then experiment with incubation times. Fewer beads and shorter incubation times will help reduce nonspeciic binding. Ultimately, orthogonal techniques such as co-immunoprecipitation (IP) with Western blot should be used to validate a subset of the potential interactors, whenever possible. In this work we present examples of worklows in which immunoafinity-puriied proteins were either separated using gel electrophoresis and bands exhibiting signiicant change from control were analyzed, or complexed proteins were eluted from the beads, digested in solution, and analyzed. It is important to note that these protocols provide general guidelines and that several optimization steps with multiple iterations of MS will likely be required for puriication of a protein complex of interest. As discussed in more detail below, data analysis and mining are critical for gleaning relevant information from proteomics studies. Careful extraction of peptide signals, determination of properly stringent search engine parameters [11], and reversed or scrambled

EXPERIMENTAL PROTOCOLS

database searching leads to an output data set where signiicance of the identiications may be established with score or probability cutoffs. Although a false discovery rate of ∼1% is often employed in more global approaches [12], establishing criteria for two peptide “hits” to a protein with peptide probabilities of ∼95% is suficient to provide a false discovery rate approaching 0% for immunoafinity puriication applications. Following identiication, data mining is employed to obtain functional information about the proteins to begin to decipher mechanisms that may be triggered by the interaction. The tools used often include those developed for microarray analysis, where gene ontology information is used to cluster proteins with similar cellular compartments, functions, or pathways [13–16]. Downstream assays based on these results, including IP of proteins identiied in the complex, allow investigators to begin to elucidate mechanisms driving cellular function.



• •

• •

EXPERIMENTAL PROTOCOLS

Materials and Solutions •







Cell Lysis Buffer. 0.33% 3-[(3-cholamidopropyl) dimethylammonio]-1-propanesulfonate (CHAPS); 150 mM NaCl; 10 mM sodium pyrophosphate; 10 mM Tris-HCl pH 7.4; 1 mM phenylmethylsulfonyl luoride (PMSF); 0.4 mM ethylenediaminetetraacetic acid (EDTA); 1.8 mg/mL iodoacetamide (IAA); 10 mM NaF; 2 mM Na3VO4; and 1 µg/mL each of aprotinin, leupeptin, and pepstatin. Tissue Lysis Buffer. Buffer A (10 mM HEPES pH 7.9, 1.5 mM KCl, 10 mM MgCl2, 0.5 mM dithiothreitol (DTT), 0.1% IGEPAL CA-630 (SigmaAldrich, St. Louis, MO), and 0.5 mM PMSF) and Buffer B (20 mM HEPES pH 7.9, 25% glycerol, 1.5 mM MgCl2, 420 mM NaCl, 0.5 mM DTT, 0.2 mM EDTA, 0.5 mM PMSF, and 4 µM leupeptin). (Note: These buffers were selected for the analysis of liver proteomes, as described below. Other lysis buffers may be more appropriate for the sample/proteins under investigation. A search of recent literature should provide some direction regarding appropriate buffer selection.) Reagents for preparation of magnetic beads as described below. Protein A or protein G beads. (Note: Either Sepharose [GE Healthcare, Piscataway, NJ] beads or magnetic beads may be used; however, magnetic beads are preferred as they tend to exhibit less nonspeciic binding than Sepharose beads.)

Pipette tips cut 5 mm from the top (will avoid damaging the beads if using Protein A/G Sepharose beads). Laemmli buffer. High performance liquid chromatography (HPLC)grade water (Honeywell Burdick and Jackson [Morristown, NJ] or other high quality liquid chromatography–mass spectrometry [LC/MS]grade water). Desalt spin columns. Reagents for in-gel or in-solution digestion as listed below.

Equipment • • • •



1.2

5

Refrigerated centrifuge Rotary mixer Vacuum centrifuge Gel electrophoresis apparatus and 10% polyacrylamide gel Nanoscale HPLC, tandem mass spectrometer

Lysis Note that the cell numbers and tissue amounts presented here are a guideline. These numbers should be increased if complexed proteins are of low abundance. Cell Lysis 1. Lyse ∼5 × 108 cells using 500–1000 µL cell lysis buffer at 4°C overnight. Note that if low abundance or weakly interacting proteins are to be analyzed, increase the cell numbers to as much as 1010 (as shown by Malovannaya and coworkers [17]). 2. In the morning, centrifuge lysates at 12,000 × g at 4°C for 20 min. Remove supernatants to a clean tube. Tissue Lysis 1. Homogenize ∼100 mg of tissue using a mortar and pestle over liquid nitrogen. For identiication of potentially weakly binding complexes, increase tissue amount to 10–20 g (following Moresco et al. [18]) and increase lysis buffer volume to 5–10 mL. 2. Add homogenized tissue to 0.5 mL of tissue lysis buffer A (ice cold) and ultrasonicate three times, 15 s each. Place sample tubes in an ice bath for at least 1 min between sonications. 3. Incubate samples on ice for 30 min, then centrifuge at 14,000 × g, 4°C for 10 min. Remove the supernatant (cytoplasmic fraction) to clean tube. 4. Resuspend the membrane/organelle fraction pellet in 0.2 mL ice-cold tissue lysis buffer B and

6

TARGETED PROTEOMICS USING IMMUNOAFFINITY PURIFICATION

incubate on ice for 30 min. Following centrifugation at 14,000 × g for 30 min at 4°C, remove the supernatant to clean tube.

bead solution from cap), and place tubes in a magnetic rack. Remove supernatant. Repeat two more times.

Note: For all subsequent steps, be sure to keep samples on ice or at 4°C, however freezing lysates before the immunoafinity puriication should be avoided [18]. If necessary, store lysates at −80°C prior to use.

Prepare 400 µg of primary antibody in 1 mL citrate phosphate buffer and add to beads. Reducing the volume may improve binding and may be included in subsequent optimization steps. Rotate tube end over end for 2–3 h at room temperature.

Total Protein Quantiication Use either the Bradford or bicinchoninic acid (BCA) method to quantify total protein concentrations following the manufacturer’s instructions for a microwell plate assay. Make sure that the lysis buffer components are compatible with the manufacturer’s stated levels. Try several dilutions to ensure the sample concentration is within the linear range of the assay. Immunoafinity Puriication Immunoafinity puriication may be accomplished using either soluble antibodies or antibodies cross-linked to beads. Generally the irst step of the optimization should be done using cross-linked antibodies; cross-linking signiicantly reduces contaminating signals from Ig light and heavy chains. Procedures for both approaches are provided below. Immunoafinity Puriication Using Magnetic Beads Reagents for Magnetic Bead Preparation (Dynabeads Are Typically Used) •





• •

• •

Citrate Phosphate Buffer, pH 5.0. 25 mM citric acid, 50 mM sodium phosphate (Na2HPO4) 0.2 M Triethanolamine (TEA), pH 8.2. 3.71 g triethanolamine-HCl/100 mL water 20 mM DMP. 5.4 mg DMP-2HCl per milliliter of TEA buffer 50 mM Tris pH 7.5 PBS-T. 0.01% Tween-20 (Thermo Fisher Scientiic, Waltham, MA) in phosphate buffered saline 0.1 M glycine pH 2.5–2.7 Storage Solution. PBS-T with 0.02% sodium azide

Dynabeads (Invitrogen Corp., Carlsbad, CA) are packaged as a 5% slurry. Prepare 0.5–1.0 mL of slurry to obtain 25–50 µL of packed beads. A rule of thumb is that 1 mL of slurry binds ∼300 µg of antibody. Incubate with about 400 µg of the primary antibody. Equilibrate Dynabeads Centrifuge beads briely, place tube in a magnetic rack, and remove the supernatant. Add 1 mL citrate phosphate buffer, vortex, spin briely in a minifuge (1 s to remove

Incubate with Primary Antibody

Wash Centrifuge briely, place tubes in a magnetic rack, remove supernatant, and add 1 mL citrate phosphate buffer and wash three times as described in the section “Equilibrate Dynabeads.” Wash two times more with 1 mL 0.2 M triethanolamine-HCl. Cross-Link Remove inal TEA wash from the beads using the magnet and add 1 mL of DMP solution. Incubate 30 min at room temperature, rotating end over end. Clean Up Using the magnet, remove the DMP solution and incubate beads with 50 mM Tris for 15 min to remove free cross-linking reagent. Wash Wash beads three times with PBS-T. Remove Free Antibody Incubate the beads with 0.1 M glycine for 5 min, rotating end over end at 4°C. Wash Wash beads three times with PBS-T. Store Beads Bring beads back to original packaged volume in storage solution for storage at 4°C. For a 1 mL stock, use 950 µL of storage solution. To use, wash beads three times using 1 mL PBS, then three time with 1 mL lysis buffer. Use Western blots and bead titration to determine the minimum amount of beads to use and the minimum amount of time to incubate. A starting point for optimization may be 20 µL of packed beads and 2 h of incubation at 4°C. Elute as described below. Immunoafinity Puriication Using Protein A/G Sepharose Beads Incubation with Primary Antibody Add 3–20 µg of primary antibody to the supernatant and incubate for 2 h to overnight at 4°C. This step

EXPERIMENTAL PROTOCOLS

should be optimized to use the minimal time and generate the least amount of nonspeciic binding. Increasing the incubation time generally increases background and fragmentation due to endogenous protease activity. Increased antibody concentrations are required when using more cells or when the total protein concentration is high. Note: In order to reduce nonspeciic binding of protein aggregates during incubation, ultracentrifuge samples at 100,000 × g for 20 min following incubation and limit incubation time to 2 h. Remove the supernatant to a clean tube. Some groups recommend that the lower 0.1 mL of lysate not be used [17]. Using magnetic beads instead of Sepharose beads may also help. Prepare Protein A/G Sepharose Beads Remove 100 µL of bead slurry for each sample (mix the slurry well prior to removing the beads). Centrifuge at 1500 × g at 4°C for 1–2 min, aspirate the supernatant, and wash with 10× bead volume of cell lysis buffer three times. Bind to Sepharose Beads Remove sample to the tube containing the washed beads and mix end over end at 4°C for 4 h. Centrifuge at 1500 × g at 4°C for 10 min and remove supernatant to clean tube. Wash Wash beads up to ive times with 10× bead volume of cell or tissue lysis buffer by adding buffer, centrifuging at 1500 × g at 4°C for 1 min, and removing supernatant. (Note: To retain weakly interacting proteins, use a low stringency wash buffer, such as PBS or 0.5% NP-40, add 10× bead volume, and briely invert the tube 10 times [17]. Radioimmunoprecipitation assay [RIPA] buffer may also be used to retain more strongly interacting proteins.) Elution Proteins may be eluted from the beads either for subsequent gel electrophoresis and in-gel digestion or for in-solution digestion. Gels For applications involving gel electrophoresis, add one bead volume of 2× Laemmli buffer to the bead pellet, heat at 60°C for 10–15 min, centrifuge, and remove the supernatant to clean tube. To ensure complete elution, add an additional 10 µL of Laemmli buffer to the bead pellet, repeat, and pool supernatants. Solution For in-solution digestion applications, elute the proteins from the antibody-coupled Sepharose beads using one

7

bead volume of a strong acid such as 0.2 M citric acid (pH 2.0), 1% formic acid or 0.5% triluoroacetic acid. If using citric acid, immediately add an equivalent volume of 2 M Tris to neutralize the pH prior to digestion. Alternatively, a strong and volatile base such as ammonium hydroxide can be used, or even 8 M urea. The urea will not effectively elute the protein target but will disrupt interacting proteins. Perform three similar elutions and combine the eluates. Concentrate to dryness in a vacuum centrifuge. If 8 M urea is used for elution, proceed to the protocol for in-solution digestion. Note: The elution should be optimized by checking at least two to three different elution solvents. Eluates may be assessed using a one-dimensional (1D) sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel stained with colloidal Coomassie. Gel Electrophoresis (for Subsequent in Gel Digestion) Separate Proteins Using samples generated in the section “Gels,” load the proteins onto a 1D 10%–14.5% SDS-PAGE gel and separate them using 120–180 V. A 13.3 × 8.7 cm × 1 mm thick gradient gel generally provides suficient separation. If more separation is required, a larger gel can be used. A 15% gel should be used for resolving smaller proteins, while a 7%–10% gel is appropriate for investigating larger proteins. All of these types of gels may be useful for the initial stage of optimization for separating the proteins of interest. A lower separating voltage may provide more resolution. Stain Proteins Following electrophoresis, carefully wash the gels twice with ultrapure water. Be careful to handle the gels only by the edges. Some commercially available ready-made gels have a thick bottom edge which helps prevent the gels from ripping when they are rinsed. Use either colloidal Coomassie blue (such as EZ Blue, Sigma-Aldrich) or a luorescent stain and follow the manufacturer’s directions for staining and ixing. Oriole™ luorescent gel stain (BioRad Corporation, Hercules, CA) is an easy-to-use, sensitive, and highly linear protein gel stain. This stain can be used without prior protein ixation or destaining. For larger protein loads (e.g., 10 mg total protein), use a standard Coomassie blue stain and follow destaining procedure below. Visualize Proteins Image the gel using a laser imager capable of exciting and measuring emission at wavelengths appropriate for the stain employed. For example, Oriole has an excitation maximum at 270 nm and emission at 604 nm, thus

8

TARGETED PROTEOMICS USING IMMUNOAFFINITY PURIFICATION

it is appropriate for UV-based imagers. Coomassie can be imaged using a standard visible light gel scanner. Excise Protein Bands Excise gel bands of interest using a scalpel or a cut pipette tip. Dice the gel slice into 1 mm3 cubes. Place into a 1.5 mL microfuge tube and remove liquid. Gels may be stored at −80°C prior to processing. Sample Preparation for MS In-Gel Digestion Reagents It must be noted, use only high quality HPLC-grade water. •









50 mM Ammonium Bicarbonate (NH4HCO3). 4 mg/mL in water. Destain Reagent. 1:1 acetonitrile (ACN) in 50 mM NH4HCO3. Reducing Reagent. Prepare a 1.5 mg/mL solution of DTT in 50 mM NH4HCO3. DTT is unstable in solution at room temperature so prepare fresh daily. Alkylating Reagent. 10 mg/mL solution of IAA in 50 mM NH4HCO3. This buffer should also be prepared fresh daily (while reducing) and kept in the dark. Trypsin. Use sequencing grade modiied trypsin. Promega (Madison, WI) is widely used.

Destain (This step will remove stain that might interfere with proteolytic digestion. Note that this step is not required if using Oriole stain. Follow manufacturer’s directions for the stain utilized.) 1. Remove any liquid from the microfuge tube. If the gel band was not previously diced, use a sealed pipette tip to cut gel slice into small pieces. 2. Add enough destain reagent to cover gel pieces. Shake or sonicate at room temperature for 10 min. Discard liquid. If gels were stained using a luorescent stain, go to the section “Reduction/Alkylation (for 1D Gel Bands).” 3. Repeat the second step until gel pieces are clear. Usually two to three washes are necessary for Coomassie stained gels. If gel pieces still have blue color, rehydrate by adding 50 mM NH4HCO3 and shake for 10 min at room temperature. Discard liquid and repeat. Reduction/Alkylation (for 1D Gel Bands) (This step allows for denaturing and separation of complexed proteins)

1. If not already dehydrated, dry gel pieces in vacuum centrifuge. Transfer the dried slices to clean 0.6 mL tubes. 2. Add enough of the reducing reagent to fully cover gel pieces, taking into account that the pieces will swell. Usually 20–50 µL is suficient. 3. Incubate at 37°C for 1 h. During this time, prepare the alkylating reagent. 4. After incubating, remove excess reducing reagent and cool to room temperature (this can be done by placing samples at −20°C for 5 min). Add enough alkylating reagent to cover gel pieces. Place tubes in the dark and shake gently for 45 min. 5. Discard supernatant. Wash gel pieces with 50 mM NH4HCO3 for 10 min with shaking at room temperature. Discard supernatant. 6. Wash gel slices twice more using destaining reagent (10 min with shaking at room temperature and each time discarding supernatant). 7. Dehydrate gel pieces by vacuum centrifugation. Transfer the dried slices to clean 0.6 mL tubes. 8. Trypsin digestion: Add enough of a 0.2 µg/µL trypsin solution to swell the gels and incubate on ice or at 4°C for 15 min. Usually 20 µL is suficient. The trypsin solution can be prepared and aliquoted ahead of time (20 microfuge tubes of 10 µL each) and stored in the freezer until ready. 9. Remove excess trypsin solution. Add 50 mM NH4HCO3/10% ACN, enough to cover the gel pieces, but not excess. Place in 37°C incubator. 10. Check pieces in 20 min, adding enough 50 mM NH4HCO3/10% ACN to keep pieces just covered. 11. Digest overnight (∼19 h) at 37°C. Extracting Tryptic Fragments 1. Following digestion, remove microfuge tubes from incubator. Add 1 µL of formic acid to tubes and sonicate or shake at room temperature for 15 min. Centrifuge and remove supernatant to clean tube. 2. Add enough 1.0% triluoroacetic acid/60% ACN to cover gel pieces. Shake or sonicate for 10 min at room temperature, centrifuge, and remove supernatant and combine with supernatant from previous step. 3. Vacuum centrifuge the supernatant extract until concentrated to ∼10 µL. If extract goes to dryness (not good), add 5 µL of 1.0% formic acid and vortex vigorously at room temperature. 4. Store digests at −70°C prior to analysis.

EXPERIMENTAL PROTOCOLS

9

In-Solution Digestion

Mass Spectrometric Analysis

Note that a 100 mM Tris buffer is good to use if samples will be loaded onto a C18 enrichment desalting column prior to separation by nanoscale HPLC. Otherwise, use NH4HCO3.

Note that the analyses of the samples presented here were performed using a high capacity quadrupole ion trap (LC/MSD XCT Ultra [Agilent Technologies, Santa Clara, CA]). Although the quadrupole trap was historically the proteomics workhorse, hybrid quadrupole/ time-of-light (QTOF) and Orbitrap instruments are capable of rapid scanning and high resolution of fragment ions, providing a performance advantage over the quadrupole traps [6]. Parameters presented should be viewed as general guidelines that may be modiied and incorporated into proteomics worklows using alternative instrumentation.

Reagents •









100 mM NH4HCO3. 7.9 mg/mL of HPLC-grade water (or 100 mM Tris, pH 8.5). 8 M Urea. Dilute 480 mg of urea in 1.0 mL of 100 mM NH4HCO3 solution (or 100 mM Tris buffer). Reducing Reagent. Dissolve 3 mg of DTT in 20 µL of 100 mM NH4HCO3 solution to make 1 M DTT (or use 100 mM Tris buffer). Alkylating Reagent. Dissolve 3.6 mg in 100 µL of NH4HCO3 or 100 mM Tris solution to make 200 mM IAA. Trypsin Solution. Make up a 1 mg/mL solution of trypsin in HPLC-grade water or NH4HCO3 (or 100 mM Tris buffer) and add 1–2 µL to each sample.

Digestion Procedure 1. Reconstitute. Reconstitute sample in approximately 20 µL of 8.0 M urea in a 0.5 mL microfuge tube. 2. Reduce. Add 1 µL of reducing reagent and mix the sample by gentle vortex. Reduce the mixture for 1 h at room temperature or in an oven at 37°C. Do not go over 37°C or the urea will react with the sample and generate carbamylated artifacts. Allow the sample to cool to room temperature. 3. Alkylate. Add 20 µL of alkylating reagent and alkylate for 1 h at room temperature in the dark (use aluminum foil to cover the sample). Add 4 µL of reducing reagent to consume any leftover alkylating agent (so the trypsin is not alkylated). 4. Dilute. Add 60 µL of NH4HCO3 or 100 mM Tris solution to dilute the urea before digesting it with trypsin. 5. Digest. Add trypsin solution in appropriate ratio (1:30) to approximate amount of protein by weight. After 1 h, add another microliter of trypsin solution. Digest 4 h to overnight at 37°C. 6. Stop Digestion. In the morning, or following digestion, add 1 µL of 100% formic acid to the sample. Vortex and centrifuge. Freeze sample at −70°C prior to analysis. Use a vacuum centrifuge (Speedvac) for samples with relatively low protein concentrations to maximize signal and minimize loss of sample to the tube.

Loading Samples Load either 5 µL of in-gel digestion sample or 10 µL of in-solution digestion into a polypropylene autosample vial. Establishing Flow Use a 75 µm internal diameter (ID) nanoscale column with C18 packing for separating peptides. Flow rate should be approximately 200–300 nL/min. Adjust following manufacturer’s directions for ultra high performance liquid chromatography (UPLC) pumps. Desalting Samples should be desalted prior to loading on the separating column. This can be accomplished using a pipette tip desalting system (Glygen Scientiic Products, Columbia, MD) or an online enrichment column with a switching valve. Peptide Separation Elute peptides from the separating column using a gradient from 3% to 40% buffer B (90% ACN, 0.1% formic acid). Buffer A is generally 0.1% formic acid. (Note: It is important to purchase high-quality formic acid and other reagents labeled as suitable for LC/MS applications.) For relatively simple mixtures, a gradient length of ∼40 min is generally suficient. Gradient times may be reduced if UPLC is employed. For more complex mixtures, extending the gradient, as well as performing steps of intact protein fraction, enables the identiication of more peptides and proteins. For an in-solution digest, a 2- to 3-h gradient should be suficient. This step will need to be optimized for different sample types. Acquiring Spectra Utilize data-dependent acquisition, where the most abundant peptides are analyzed a limited number of times and then placed in an exclusion list for a given amount of time (e.g., 1 min). The data shown below

10

TARGETED PROTEOMICS USING IMMUNOAFFINITY PURIFICATION

were acquired by excluding the most abundant three to six peptides in a spectrum from subsequent analysis following three spectral acquisitions. Collect spectra over a mass-to-charge ratio (m/z) range of 350–1500 Da. (Note: The desired acquisition mass range will depend on the instrument employed. This mass range is appropriate when using a quadrupole ion trap.) Database Searching Proteomics worklows with MS/MS analysis provide extraordinarily data-rich results. A single digested band can easily result in over 10,000 fragmentation spectra. Therefore, careful analysis of the data must be performed to generate high-quality output protein lists for use in subsequent biological assays. Commonly used [19] database search algorithms include Sequest (Thermo Fisher Scientiic), Spectrum Mill (Agilent Technologies), X!Tandem (The Global Proteome Machine Organization, http://www.thegpm.org/TANDEM), Mascot (Matrix Science, Inc., Boston, MA), ProteinLynx Global Server (Waters Corporation, Milford, MA), Phenyx (Geneva Bioinformatics, SA, Geneva, Switzerland), OMSSA (NCBI, http://pubchem.ncbi.nlm.nih.gov/omssa/), PEAKS (Bioinformatic Solutions, Inc., Waterloo, ON, Canada), ProteinPilot (AB Sciex, Framingham, MA), and Sequest Sorcerer (Sage-N-Research, Inc., Milpitas, CA), some of which are freely available (X!Tandem, OMSSA). Although each implementation is different, these search algorithms operate under the same general principles. These include establishment of the database and search space to be used and a statistical method of comparing experimental and theoretically generated fragmentation mass spectra, ultimately outputting a “ranked” score. Subsequently, users must determine criteria for valid score thresholds and extract potentially interesting results for follow-up.

UniProtKB database was used, updated 04/2010, with 68,507 entries. Selecting Potential Peptides to Search First, peptide experimental mass is compared with the theoretical mass generated from an in silico digest of the protein database, and a subset of peptides with mass within the tolerance window selected in the search space are utilized for subsequent processing. Next, the fragment ion masses are compared with in silico fragmentation of the peptides (with masses within the deined mass tolerance) that passed the irst criteria. Using a variety of different techniques, the algorithms assign a score or probability to the “hit” and generate a protein identiication. Interpreting the Score Data The algorithms generate scores for all of the data within the tolerance window; however, some spectra may be falsely assigned. Generally, users select a score cutoff criterion and only consider data falling within the criterion window. Typically, a minimum of two unique peptides are required for each protein identiication, although some groups have validated identiications from just one unique peptide [18]. Using reversed database searching and receiver operating curves may also help determine an appropriate cutoff. The use of data evaluation methods is not yet standardized; see Kapp and Schütz for an excellent discussion of these issues [19]. Due to the large number of spectra generated in a typical experiment it is impossible to manually inspect all of the data. However, search results from proteins that may be used in downstream biological assays should be manually validated. Mining the Data

Determining the Search Space The search space can be limited by specifying known parameters including sample taxonomy, precursor and fragment ion mass tolerance, enzyme speciicity, numbers of allowed missed enzymatic cleavages, and potential amino acid (AA) modiications. Limiting the search space signiicantly helps reduce false-positive results. Additionally, the selection of a more highly curated database, such as the UniProt Knowledge Base (http:// www.uniprot.org/help/uniprotkb), may also help with reducing false positives. In the examples below, taxonomy was mus musculus, precursor ion tolerance was 1.7, and the fragment ion tolerance was 0.6. Trypsin was selected as the enzyme with two missed cleavages allowed. No AA modiications were included. The

Proteomics data output results in long lists of proteins and scores, and it is often quite dificult to extract signiicant information to use for the development or testing of hypotheses. Initial insights are often obtained by searching the Gene Ontology (GO) annotations [13] for overrepresented terms. A comprehensive list of GObased tools is available at http://www.geneontology.org/ GO.tools.shtml. In the example below, we utilized The Database for Annotation, Visualization and Integrated Discovery (DAVID) v6.7 [20,21] to cluster genes with similar functions, families, or cellular locations to select a group of proteins with characteristics related to our hypothesis that were further investigated. New tools are constantly being developed and a regular review of Web sites and the literature is highly recommended.

APPLICATIONS OF THE PROTOCOLS

1.3 APPLICATIONS OF THE PROTOCOLS 1.3.1 Analysis of the Acetylated Proteome of Mouse Liver in Obesity Acetylation is now recognized as an emerging mechanism for controlling proteins mediating cellular adaptation to metabolic fuels [22] and is governed, in part, by sirtuins (SIRTs), class III NAD+-dependent histone deacetylases (HDACs) that regulate lipid and glucose metabolism in liver during fasting and aging. However, whether acetylation or SIRTs play a pathogenic role in fuel metabolism under conditions of obesity is unknown. In the present study, 5- to 6-week-old male C57BL/6 SVJ mice were fed a high-fat diet (45 kcal% fat) or a standard chow diet for 16 weeks, and then were sacriiced by pentobarbital overdose following treatment. Livers were harvested immediately from anesthetized mice and snap frozen at −70°C in liquid nitrogen before analysis. Utilizing the protocol provided for preparation of tissue homogenates (∼200 mg liver tissue, 10 µg primary antibody, 50 µL packed beads), we prepared nuclear/mitochondrial and cytoplasmic fractions and enriched for lysine acetylated proteins using an antiacetyllysine polyclonal antibody. Complexed proteins

were eluted from the beads by boiling in Laemmli buffer, separated using a 10% polyacrylamide gel, luorescently stained with Lava Purple and imaged. Labeled gel bands in Figure 1.2 that differed between mice fed normal chow (control) and a high-fat diet were excised, digested, and analyzed by MS/MS using a 40min gradient of acetonitrile. Data were analyzed by LABKEY (http://www.labkey.com), an interface utilizing the X!Tandem search algorithm and elements of the Trans-Proteomics Pipeline, including PeptideProphet and Protein Prophet (these are freely available software packages developed by the Seattle Proteome Center within the Institute for Systems Biology, Seattle, WA; http://www.proteomecenter.org/software.php). Receiver operating characteristic (ROC) curves were generated to determine peptide and protein probability cutoffs providing ∼ 1% false discovery rate (typically ∼0.9–1.0). Following the search, results were exported to Microsoft Excel (cytoplasmic sample, 1703 proteins; nuclear/ mitochondrial sample, 307 proteins). Proteins that were observed in multiple samples were assumed to be nonspeciically binding and were excluded from further analysis, as were proteins identiied by one unique peptide observed only once. Potential isoforms and putative proteins were excluded as well, reducing the

3 4 8 5 6 7 9

250

150

150

100 75

100

High Fat

MW 250

Control

Cytoplasmic

MW

Control

High Fat

Nuclear

1 2

11

1

75 50 37

50

2 3 4 5

37

6

25

10 25

7

FIGURE 1.2 Immunoafinity puriication of acetylated proteins reveals increased acetylation in proteins from livers of mice fed a high-fat diet. C57BL/6 SVJ mice were fed either normal chow or a high-fat diet (45% fat) for 16 weeks and fasted overnight before sacriice (n = 3 per group). Samples were extracted from liver as described in the text. Lysine-acetylated liver proteins were immunoafinity puriied from either cytoplasmic or mitochondrial/nuclear extracts, then separated by 1D gel electrophoresis using a 10% SDS-PAGE gel, visualized with Lava Purple and imaged with a Typhoon 9600 imager (GE Healthcare). Labeled bands with differential staining were excised, digested, and identiied by MS/MS and database searching.

12

TARGETED PROTEOMICS USING IMMUNOAFFINITY PURIFICATION

number of identiications to 227 for the cytoplasmic sample and 156 for the nuclear/mitochondrial sample. Proteins were sorted by calculated molecular weight (MW) and MWs differing from the median by more than 20% were assumed to be incorrect identiications. Finally, protein identiications with at least 5% AA coverage of the protein [18] were retained in the analysis. As summarized in Tables 1.1 and 1.2, we identiied 148 proteins from the cytoplasmic sample and 75 proteins from the nuclear/mitochondrial sample with high conidence. Each band typically contained a number of proteins that were identiied. Bands excised from the cytoplasmic samples were larger to incorporate the apparently high number of closely comigrating proteins; therefore, more proteins were typically identiied in those bands than in the nuclear/mitochondrial extract bands. An examination of proteins from the latter sample suggests that a high-fat diet led to hyperacetylation of proteins involved in gluconeogenesis, mitochondrial oxidative metabolism, methionine metabolism, liver injury, and the endoplasmic reticulum (ER) stress response. Not shown here, we observed that in mice lacking SIRT3, a sirtuin localized to the mitochondrion, a high-fat diet further increased the acetylation status of liver proteins compared with high-fat diet-fed wildtype mice and was associated with the disruption of mitochondrial oxidative phosphorylation complexes II, III, and V. Our results suggest that hyperacetylation of mitochondrial proteins may play a pivotal role in mechanisms regulating high-fat diet-induced mitochondrial dysfunction in livers of obese mice. 1.3.2 Identiication of a Major Histocompatibility Complex Class II (MHC-II)-Complexed Death Transducer MHC-II is primarily known to function in the presentation of antigenic peptides to T lymphocytes. However, these molecules have also been observed to transduce signals, leading to either cell activation or apoptotic death. The short, cytoplasmic tails of the two transmembrane proteins comprising MHC-II are not required for induction of apoptosis [23], therefore a protein complexing with MHC-II is likely important in mediating death signaling. In this study, K46 cells (5 × 108) were lysed in cell lysis buffer and immunoafinity puriied, as described above. Shown in Figure 1.3 is a gel separation of proteins from a representative preparation, suggesting the presence of many potentially complexed proteins. Bound proteins were eluted from the beads with citric acid and pH was neutralized with Tris. Following in-solution digestion, ∼30% of the sample was analyzed by nanoscale liquid chromatography–tandem mass spectrometry (nano-LC/

MS/MS) using a 145-min gradient of acetonitrile. Data from ive separate immunoafinity puriications were analyzed by LABKEY and proteins with probabilities yielding less than 3% false discovery rates were included. Initially, a total of 2514 protein identiications were obtained from the combined ive analyses. Since the data iles were obtained from immunoafinity puriication of both MHC-II and a complexed protein (MPYS, subsequently termed TM173), we restricted the list to proteins that were observed in at least three runs, including one run from the MHC-II IP and one run from the MPYS IP. Identiications based on one hit were removed, as were duplicates, and putative uncharacterized proteins, resulting in a list of 237 proteins, of which 83 were isoforms of MHC-II. The list of accession numbers was then clustered using functional annotation tools in DAVID [21] and protein groups present are summarized in Figure 1.4. A recent review of commonly observed protein identiications [24] suggests that overrepresentation of actin, tubulin, and adenosine triphosphate (ATP) synthase isoforms may result from nonspeciic binding or represent common cellular stress responses. Since MHC-II is a transmembrane protein, we concentrated further analysis on the group of transmembrane proteins identiied in this study, summarized in Table 1.3. Of these proteins, only two were potentially multispanning proteins—CD20 and TM173. Fragmentation mass spectra leading to the identiication of these proteins are plotted in Figures 1.5 and 1.6, respectively. CD20 is known to associate with MHC-II but was not thought to be a likely candidate for transducing death signals [23]. As described in the UniProt Knowledge Base, the isotopes of transmembrane emp24 are singlepass transmembrane proteins with cytoplasmic domains not well described, and CDGSH iron sulfur protein, implicated in autophagy, has a short cytoplasmic domain. Ribophorin and solute carrier protein 3 are both single membrane-spanning proteins with longer cytoplasmic domains (150 and 75 AAs, respectively). MHC-II is known to associate with tetraspanins, therefore we selected TM173, a potential tetraspanin with a 204-aminoacid-long cytoplasmic domain, for subsequent study. As we previously described [23], TM173 is a membrane protein with an immunoreceptor tyrosine-based inhibitory motif (ITIM) contained in its cytoplasmic tail. We demonstrated that TM173 becomes phosphorylated upon cross-linking with MHC-II, recruits the inhibitory signaling effectors Src homology region 2 domain-containing phosphatase-1 (SHP-1) and Phosphatidylinositol-3,4,5trisphosphate 5-phosphatase 1 (SHIP), reduces calcium mobilization, and negatively regulates cell growth. Conirmation of the important role of TM173 in MHC-IImediated cell death was obtained in cells with TM173 knocked down.

13

APPLICATIONS OF THE PROTOCOLS

TABLE 1.1

Hyperacetylated Cytoplasmic Proteins from Livers of Mice Fed a High-Fat Diet

Banda

Identiierb

Protein Description

1 1 1

Q8CIE6 Q9Z1Q9 Q8VDN2

1 1 1

Q8VDJ3 Q9JKR6 Q05920

2 2 2 2 3

P17563 O35728 P24549 Q8VCM7 O54734

3 3

P17182 Q8BWF0

3

P53395

3

P70333

3

O35737

3 3

P12790 P40124

3

Q9D2G2

3

Q9JLJ2

3

Q62465

3

P50431

3

P61922

3 4 4 4 4 4 4 4

Q9Z1N5 Q9QZ85 P09411 Q61598 P43883 P15105 Q8VCZ9 P63037

Coatomer subunit alpha Valyl-tRNA synthetase Sodium/potassium-transporting ATPase subunit alpha-1 Vigilin Hypoxia upregulated protein 1 Pyruvate carboxylase, mitochondrial Selenium-binding protein 1 Cytochrome P450 4A14 Retinal dehydrogenase 1 Fibrinogen gamma chain Dolichyldiphosphooligosaccharide– protein glycosyltransferase 48 kDa subunit Alpha-enolase Succinate-semialdehyde dehydrogenase, mitochondrial Lipoamide acyltransferase component of branched-chain alpha-keto acid dehydrogenase complex, mitochondrial Heterogeneous nuclear ribonucleoprotein H2 Heterogeneous nuclear ribonucleoprotein H Cytochrome P450 2B9 Adenylyl cyclase-associated protein 1 Dihydrolipoyllysine-residue succinyltransferase component of 2-oxoglutarate dehydrogenase complex, mitochondrial 4-trimethylaminobutyraldehyde dehydrogenase Synaptic vesicle membrane protein VAT-1 homolog Serine hydroxymethyltransferase, cytosolic 4-aminobutyrate aminotransferase, mitochondrial Spliceosome RNA helicase BAT1 Interferon-inducible GTPase 1 Phosphoglycerate kinase 1 Rab GDP dissociation inhibitor beta Perilipin-2 Glutamine synthetase Probable proline dehydrogenase 2 DnaJ homolog subfamily A member 1

MW (kDa)

AA Coveragec (%)

Protein Prophet Probabilityd

Total Peptidese

Unique Peptidesf

138,446 140,084 112,510

12 9 9

1.0000 1.0000 1.0000

6 9 6

4 4 3

141,612 107,606 127,428

6 6 6

1.0000 1.0000 1.0000

2 8 3

1 3 3

525,14 58,238 54,337 46,671 46,044

45 19 16 14 36

1.0000 1.0000 1.0000 0.9997 1.0000

31 12 13 4 27

11 5 5 1 10

47,010 52,012

26 21

1.0000 1.0000

29 7

7 6

46,188

20

1.0000

5

4

49,280

19

1.0000

3

2

49,199

19

1.0000

3

2

55,760 51,444

17 14

1.0000 1.0000

10 7

6 4

41,470

12

1.0000

4

3

53,384

9

1.0000

4

2

42,965

8

0.9998

2

1

52,585

7

1.0000

2

2

53,220

6

0.9996

2

2

48,904 47,572 44,419 50,537 46,664 41,988 50,723 44,552

6 44 43 33 30 29 28 24

0.9999 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000

2 22 21 14 11 38 5 13

2 7 8 7 7 9 3 5 (Continued)

14

TARGETED PROTEOMICS USING IMMUNOAFFINITY PURIFICATION

TABLE 1.1

(Continued)

Banda

Identiierb

Protein Description

4 4

Q91WN4 Q91YT0

4

P26150

4

O88986

4 4 4

Q8R0F9 Q91VC3 Q922E4

4

O88844

4

A2AKK5

5 5

Q99PG0 Q60759

5

Q9Z2I9

5

Q8QZT1

5

P08249

5

P49429

5 5

Q9QYG0 P51174

5

P45952

5 5 5 5 5

Q62087 P06151 Q9CZU6 P52430 Q9WUZ9

5

Q91YP0

5

Q924Y0

5

Q71RI9

5

Q99JI4

5

P35486

5

P14152

Kynurenine 3-monooxygenase NADH dehydrogenase (ubiquinone) lavoprotein 1, mitochondrial 3 beta-hydroxysteroid dehydrogenase/Delta 5–>4-isomerase type 3 2-amino-3-ketobutyrate coenzyme A ligase, mitochondrial SEC14-like protein 4 Eukaryotic initiation factor 4A-III Ethanolamine-phosphate cytidylyltransferase Isocitrate dehydrogenase (NADP) cytoplasmic Acyl-coenzyme A amino acid N-acyltransferase 1 Arylacetamide deacetylase Glutaryl-CoA dehydrogenase, mitochondrial Succinyl-CoA ligase (ADPforming) subunit beta, mitochondrial Acetyl-CoA acetyltransferase, mitochondrial Malate dehydrogenase, mitochondrial 4-hydroxyphenylpyruvate dioxygenase Protein NDRG2 Long-chain speciic acyl-CoA dehydrogenase, mitochondrial Medium-chain speciic acyl-CoA dehydrogenase, mitochondrial Serum paraoxonase/lactonase 3 L-lactate dehydrogenase A chain Citrate synthase, mitochondrial Serum paraoxonase/arylesterase 1 Ectonucleoside triphosphate diphosphohydrolase 5 L-2-hydroxyglutarate dehydrogenase, mitochondrial Gamma-butyrobetaine dioxygenase Kynurenine–oxoglutarate transaminase 3 26S proteasome non-ATPase regulatory subunit 6 Pyruvate dehydrogenase E1 component subunit alpha, somatic form, mitochondrial Malate dehydrogenase, cytoplasmic

AA Coveragec (%)

Protein Prophet Probabilityd

Total Peptidese

Unique Peptidesf

54,532 48,626

22 20

1.0000 1.0000

9 4

7 3

41,900

20

1.0000

8

4

42,850

14

1.0000

9

3

46,053 46,709 45,235

14 13 12

1.0000 1.0000 0.9956

4 7 2

2 2 1

46,660

11

0.9999

2

2

46,070

8

1.0000

2

2

45,119 43,737

49 47

1.0000 1.0000

23 34

11 9

44,422

36

1.0000

27

10

41,414

31

1.0000

25

7

36,367

27

1.0000

17

6

44,923

27

1.0000

13

6

40,658 44,627

25 21

1.0000 1.0000

5 22

3 7

43,593

21

1.0000

15

5

39,351 43,292 49,014 39,434 45,262

19 19 16 15 13

1.0000 1.0000 1.0000 1.0000 0.9895

10 2 4 5 2

3 2 3 3 1

45,665

12

1.0000

2

2

44,699

12

1.0000

3

2

51,126

10

1.0000

2

2

45,405

10

1.0000

3

3

40,181

10

1.0000

11

3

38,545

8

1.0000

2

2

MW (kDa)

15

APPLICATIONS OF THE PROTOCOLS

TABLE 1.1

(Continued)

Banda

Identiierb

5

P42669

5 6 6

P50247 P70694 Q8VBT2

6 6

P52196 Q9CW42

6

Q8VCX1

6

Q922Q1

6 6

P47962 P57776-2

6

P57776-3

6

Q9JMD3

6

P62880

6

P50172

6

Q9WUM5

6 6

Q9JII6 Q9D051

6

Q6P3A8

6 6

P47911 Q91Z53

6

Q9D5T0

6

Q3UNZ8

6 6 6

Q91Y97 Q91XE0 Q9CXR1

6

Q8VCH0

6

Q921H8

6

Q99MZ7

Protein Description

Transcriptional activator protein Pur-alpha Adenosylhomocysteinase Estradiol 17 beta-dehydrogenase 5 L-serine dehydratase/L-threonine deaminase Thiosulfate sulfurtransferase MOSC domain-containing protein 1, mitochondrial 3-oxo-5-beta-steroid 4-dehydrogenase MOSC domain-containing protein 2, mitochondrial 60S ribosomal protein L5 Isoform 2 of elongation factor 1-delta Isoform 3 of elongation factor 1-delta Phosphatidylcholine transfer protein (PCTP)-like protein Guanine nucleotide-binding protein G(I)/G(S)/G(T) subunit beta-2 Corticosteroid 11-betadehydrogenase isozyme 1 Succinyl-CoA ligase (GDPforming) subunit alpha, mitochondrial Alcohol dehydrogenase (NADP+) Pyruvate dehydrogenase E1 component subunit beta, mitochondrial 2-oxoisovalerate dehydrogenase subunit beta, mitochondrial 60S ribosomal protein L6 Glyoxylate reductase/ hydroxypyruvate reductase ATPase family AAA domaincontaining protein 1 Quinone oxidoreductase-like protein 2 Fructose-bisphosphate aldolase B Glycine N-acyltransferase Dehydrogenase/reductase SDR family member 7 3-ketoacyl-CoA thiolase B, peroxisomal 3-ketoacyl-CoA thiolase A, peroxisomal Peroxisomal trans-2-enoyl-CoA reductase

AA Coveragec (%)

Protein Prophet Probabilityd

Total Peptidese

Unique Peptidesf

34,884

8

0.9795

2

1

47,557 37,048 34,462

6 67 48

0.9999 1.0000 1.0000

2 62 16

1 12 7

32,233 33,991

45 35

1.0000 1.0000

48 18

12 6

37,290

32

1.0000

11

6

34,831

31

1.0000

37

6

37,248 33,335

30 29

1.0000 1.0000

13 7

7 3

31,162

29

1.0000

7

3

32,951

26

1.0000

13

5

37,200

25

1.0000

6

4

34,269

24

1.0000

10

3

32,098

24

1.0000

17

5

36,456 35,768

23 23

1.0000 1.0000

5 17

4 4

37,817

22

1.0000

3

3

34,216 35,329

21 19

1.0000 1.0000

6 3

3 2

40,744

19

1.0000

2

2

37,809

18

1.0000

2

2

39,376 34,098 34,876

17 16 14

1.0000 1.0000 1.0000

7 10 2

4 3 2

41,284

14

1.0000

6

3

41,242

14

1.0000

6

3

32,410

14

1.0000

10

3

MW (kDa)

(Continued)

16

TARGETED PROTEOMICS USING IMMUNOAFFINITY PURIFICATION

TABLE 1.1

(Continued)

Banda

Identiierb

Protein Description

6

Q9DBH5

6

Q9CY50

6 6 6 6 6

Q8CDN6 P48758 Q9D819 Q9DCX8 P62874

6 6 6

P51658 O55125 Q6ZWX6

6

P11725

7

Q9DCX2

7 7 7 7 7 7

P35700 Q9WVL0 P10649 Q9DCM2 P11352 Q9CQQ7

7 7

Q9R257 Q9D6Y7

7

Q9DB20

7 7 7 7 7 7 7 7 7 7 7 7 7

Q9CQC9 Q64471 P35278 P30115 Q9D1G1 P46638 P62492 Q9CZM2 P19157 Q923D2 P51410 P62082 Q9DCS9

7 7 7 7 7

Q80W21 P35980 Q91V41 Q9DBP5 Q9CQA3

7

Q9DCW4

Vesicular integral-membrane protein VIP36 Translocon-associated protein subunit alpha Thioredoxin-like protein 1 Carbonyl reductase (NADPH) 1 Inorganic pyrophosphatase Iodotyrosine dehalogenase 1 Guanine nucleotide-binding protein G(I)/G(S)/G(T) subunit beta-1 Estradiol 17-beta-dehydrogenase 2 Protein NipSnap homolog 1 Eukaryotic translation initiation factor 2 subunit 1 Ornithine carbamoyltransferase, mitochondrial ATP synthase subunit d, mitochondrial Peroxiredoxin-1 Maleylacetoacetate isomerase Glutathione S-transferase Mu 1 Glutathione S-transferase kappa 1 Glutathione peroxidase 1 ATP synthase subunit b, mitochondrial Heme-binding protein 1 Peptide methionine sulfoxide reductase ATP synthase subunit O, mitochondrial GTP-binding protein SAR1b Glutathione S-transferase theta-1 Ras-related protein Rab-5C Glutathione S-transferase A3 Ras-related protein Rab-1B Ras-related protein Rab-11B Ras-related protein Rab-11A 60S ribosomal protein L15 Glutathione S-transferase P 1 Flavin reductase 60S ribosomal protein L9 40S ribosomal protein S7 NADH dehydrogenase (ubiquinone) 1 beta subcomplex subunit 10 Glutathione S-transferase Mu 7 60S ribosomal protein L18 Ras-related protein Rab-14 UMP-CMP kinase Succinate dehydrogenase (ubiquinone) iron-sulfur subunit, mitochondrial Electron transfer lavoprotein subunit beta

AA Coveragec (%)

Protein Prophet Probabilityd

Total Peptidese

Unique Peptidesf

35,455

13

0.9918

3

1

29,689

12

1.0000

7

3

32,106 33,378 32,667 30,257 37,246

11 11 11 9 9

1.0000 1.0000 1.0000 1.0000 1.0000

3 3 3 3 4

2 2 2 2 2

30,510 33,363 35,977

8 8 8

1.0000 1.0000 0.9931

4 2 3

2 1 1

36,122

6

0.9999

2

2

20,893

63

1.0000

13

7

22,176 25,295 25,839 22,165 22,179 24,765

54 53 48 48 46 44

1.0000 1.0000 1.0000 1.0000 1.0000 1.0000

23 41 23 20 6 22

7 9 6 8 3 9

25,494 22,187

41 40

1.0000 1.0000

2 7

2 4

25,988

32

1.0000

17

4

22,382 27,245 23,281 25,229 24,015 24,030 23,907 21,493 23,478 22,066 21,881 22,127 22,687

31 31 31 31 29 25 25 25 23 23 23 22 22

1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000

9 20 6 21 3 7 7 4 6 3 5 8 5

3 6 4 4 2 4 4 2 2 3 2 2 2

25,710 21,513 23,766 21,006 28,770

22 22 20 19 17

1.0000 1.0000 1.0000 1.0000 1.0000

8 7 6 3 3

2 2 3 3 2

25,564

17

0.9998

2

1

MW (kDa)

17

ACKNOWLEDGMENTS

TABLE 1.1

(Continued)

Banda

Identiierb

Protein Description

7 7 7

P62827 Q6ZWN5 Q9WTP7

7 7 7

Q9CQD1 Q99PT1 Q9DCS2

7 7 7 7 7 7

Q00623 P24472 Q61133 Q9CPU0 Q02013 Q91VT4

7

P20108

7 7 7

P84099 P10605 Q99KF1

7

Q922B1

7 7 7

P19253 Q9CR57 Q64105

GTP-binding nuclear protein Ran 40S ribosomal protein S9 GTP:AMP (adenosine monophosphate) phosphotransferase mitochondrial Ras-related protein Rab-5A Rho GDP-dissociation inhibitor 1 UPF0585 protein C16orf13 homolog Apolipoprotein A-I Glutathione S-transferase A4 Glutathione S-transferase theta-2 Lactoylglutathione lyase Aquaporin-1 Carbonyl reductase family member 4 Thioredoxin-dependent peroxide reductase, mitochondrial 60S ribosomal protein L19 Cathepsin B Transmembrane emp24 domaincontaining protein 9 MACRO domain-containing protein 1 60S ribosomal protein L13a 60S ribosomal protein L14 Sepiapterin reductase

a b c d e f

AA Coveragec (%)

Protein Prophet Probabilityd

Total Peptidese

Unique Peptidesf

24,292 22,460 21,053

17 16 16

1.0000 1.0000 1.0000

2 4 3

1 2 2

23,467 27,492 25,573

16 16 16

1.0000 0.9996 1.0000

2 2 2

2 1 1

27,922 24,275 27,549 24,074 28,662 23,333

16 15 12 11 11 10

1.0000 0.9999 1.0000 0.9948 1.0000 0.9952

6 3 5 4 3 3

2 2 2 1 2 1

21,565

10

1.0000

3

2

23,466 23,276 23,432.91

9 9 7

1.0000 0.9957 0.9918

5 3 3

1 1 1

25,415

7

0.9845

3

1

27,578 22,210 20,678

6 6 6

0.9957 0.9957 0.9673

4 3 2

1 1 1

MW (kDa)

Excised band labeled in Figure 1.2. UniProtKB identiier number. Number of AAs identiied from peptides divided by total number of AAs in the identiied protein and multiplied by 100. Probability value calculated by the ProteinProphet module of the LABKEY software algorithm. Total number of fragmentation spectra identiied as belonging to a protein. Number of nonredundant fragmentation spectra identiied as belonging to a protein.

1.4

CONCLUSION

The use of immunoafinity puriication as a sample preparation method for downstream proteomics applications is emerging for applications involving identiication of proteins in functional complexes and global identiication of modiied proteins. This approach allows investigators to use proteomics approaches to address hypothesis-driven research questions. Although a detailed protocol was presented here, it should be noted that each sample type and antibody will require optimization of the protocol—different buffers may work better for lysis, antibodies may bind more or less

strongly—therefore, the information here is presented only as a guideline or starting place.

ACKNOWLEDGMENTS This work was supported by NIH Grants DK59767 and P30-DK48520 (to J.E.F.), a Pilot & Feasibility Award from the University of Colorado, Center for Human Nutrition P30-DK048520-09 (K.J.), and NIH Grant AI020519-22 (to J.C.C.). K.J. thanks W.H. MacDonald at Vanderbilt University for review of the manuscript and extraordinarily helpful discussions.

18

TARGETED PROTEOMICS USING IMMUNOAFFINITY PURIFICATION

TABLE 1.2 Banda

Hyperacetylated Proteins from a Nuclear/Mitochondrial Extract of Livers from Mice Fed a High-Fat Diet

Identiierb

1

sp|Q8C196

2 3 3 3 3 3 4

sp|Q05920 sp|P20029 sp|P38647 tr|Q7TSZ0 sp|P63017 tr|Q504P4 sp|P63038

4 4 4 4 4

tr|Q8C6E3 sp|P32020 sp|P34914 sp|Q63880 tr|B1AXW8

4

sp|Q8VC30

5 5 5

sp|O09173 sp|P05784 sp|P56480

5 5

sp|P10126 sp|Q99JY0

5 5 6 7

sp|Q9D8N0 sp|P97807 sp|P16460 tr|Q3TIT9

8

sp|Q03265

8 8 9 9

sp|Q63836 sp|P17563 sp|Q91Y97 sp|Q8VCH0

9

sp|Q921H8

10 10 10 10

sp|P35700 tr|Q5RJH8 sp|P11352 tr|Q58EV2

a b c d e f

Protein Description

MW (kDa)

AA Coveragec (%)

Protein Prophet Probabilityd

Total Peptidese

Unique Peptidesf

Carbamoyl-phosphate synthase (ammonia), mitochondrial Pyruvate carboxylase, mitochondrial 78 kDa glucose-regulated protein Stress-70 protein, mitochondrial Heat shock protein 9 Heat shock cognate 71 kDa protein Hspa8 protein 60 kDa heat shock protein, mitochondrial Catalase Nonspeciic lipid-transfer protein Epoxide hydrolase 2 Liver carboxylesterase 31 Aldehyde dehydrogenase 4 family, member A1 Bifunctional ATP-dependent dihydroxyacetone kinase/lavin adenine dinucleotide (FAD)– AMP lyase (cyclizing) Homogentisate 1,2-dioxygenase Keratin, type I cytoskeletal 18 ATP synthase subunit beta, mitochondrial Elongation factor 1-alpha 1 Trifunctional enzyme subunit beta, mitochondrial Elongation factor 1-gamma Fumarate hydratase, mitochondrial Argininosuccinate synthase Acetyl-Coenzyme A acyltransferase 2 (Mitochondrial 3-oxoacylCoenzyme A thiolase), isoform CRA_k ATP synthase subunit alpha, mitochondrial Selenium-binding protein 2 Selenium-binding protein 1 Fructose-bisphosphate aldolase B 3-ketoacyl-CoA thiolase B, peroxisomal 3-ketoacyl-CoA thiolase A, peroxisomal Peroxiredoxin-1 Glutathione peroxidase Glutathione peroxidase 1 Apoa1 protein

164,618

38

1.0000

137

37

129,685 72,422 73,528 73,461 70,871 68,779 60,955

23 20 18 18 6 6 17

1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000

43 19 19 19 5 5 11

17 6 7 7 2 2 5

59,835 59,126 62,515 63,318 61,841

15 14 13 7 7

1.0000 1.0000 1.0000 1.0000 1.0000

13 12 18 5 5

5 4 7 2 3

59,691

7

0.9999

3

1

49,990 47,538 56,300

18 16 16

1.0000 1.0000 1.0000

7 7 12

2 4 6

50,114 51,386

11 11

1.0000 1.0000

17 5

4 4

50,061 54,371 46,584 41,830

9 7 19 6

1.0000 1.0000 1.0000 0.9993

3 7 16 3

2 2 6 1

59,753

36

1.0000

62

12

52,610 52,514 39,507 43,995

9 9 19 11

1.0000 1.0000 1.0000 1.0000

5 5 8 3

2 2 4 2

43,953

11

1.0000

3

2

22,176 22,292 22,179 23,022

37 23 23 7

1.0000 1.0000 1.0000 0.9953

6 8 8 2

4 3 3 1

Excised band labeled in Figure 1.2. UniProtKB identiier number. Number of AAs identiied from peptides divided by total number of AAs in the identiied protein and multiplied by 100. Probability value calculated by the ProteinProphet module of the LABKEY software algorithm. Total number of fragmentation spectra identiied as belonging to a protein. Number of nonredundant fragmentation spectra identiied as belonging to a protein.

ACKNOWLEDGMENTS

TABLE 1.3 Transmembrane Proteins Potentially Complexed with MHC-II

MHC-II IP



+

250 kDa 150 kDa

UniProtKB Identiier 4F2_MOUSE

100 kDa 75 kDa 50 kDa

37 kDa

19

CD20_MOUSE MHC-II α MHC-II β

CISD2_MOUSE Q3TVJ8_MOUSE Q6PDC2_MOUSE

25 kDa

FIGURE 1.3 Immunoafinity puriication of MHC-II reveals a number of potentially complexed proteins. MHC-IIassociated proteins were enriched from K46 cells lysates and analyzed using nano-LC/MS/MS. Incubation of protein A beads in the absence of primary antibody was employed as a negative control. The gel shown is a representative result from three similar experiments.

RPN1_MOUSE TM173_MOUSE TMED4_MOUSE

TMEDA_MOUSE

Protein Description Solute carrier family 3 (activators of dibasic and neutral AA transport), member 2 Membrane-spanning 4-domains, subfamily A, member 1 CDGSH iron sulfur domain 2 Signal sequence receptor, delta Transmembrane emp24 protein transport domain containing 9 Ribophorin I Transmembrane protein 173 Transmembrane emp24 protein transport domain containing 4 Transmembrane emp24-like traficking protein 10 (yeast); predicted gene 4024

FIGURE 1.4 Functional classiication of proteins identiied in the MHC-II complex. The DAVID algorithm was used to assess overrepresentation of GO terms for the identiied proteins. The Similarity Term Overlap was set to 3, the Similarity Threshold was 0.2, the Group Membership was 5, and the Multiple Linkage Threshold value was 0.5. Following data reduction, actins, tubulins, and heat shock proteins (25) were clustered into a group with an enrichment score of 13.7, ATPases (9) were clustered with an enrichment score of 5.7, and transmembrane proteins (17) were clustered with an enrichment score of 1.4. GAPDH was left unclustered.

20

TARGETED PROTEOMICS USING IMMUNOAFFINITY PURIFICATION +2, CD 20 Peptide Probability 0.9999, Protein Probability 1.0 MW 31,958

88

b

202

301

400

513

626

714

842

785

1099 1213 1342 1470 1571 1685 1531

971

S N V V L L S A G E K N E Q T

1531 1744 1630 1530 1431 1318 1205 1118 1047 990

861

733

619

489

361

I

260

K

175 y

× 105

y9

y11

Relative Abundance

1.0

y8 0.8

y7 0.6

y10

0.4

y12 y6

b5

0.2

b13

y13

b4

b3 0 200

b10

b14

400

600

800

1000

1200

1400

y14

b15

y15

b16

1600

1800

m/z

FIGURE 1.5 Fragmentation mass spectrum and sequence analysis of peptide identiied from CD20. An almost complete series of singly charged b- and y-type ions were used to unambiguously identify this doubly charged peptide SNVVLLSAGEKNEQTIK from CD20 (MW 31,958 Da). The PeptideProphet probability was 0.9999 and the ProteinProphet probability was 1.000 for this identiication.

+2, Peptide Probabililty 0.9993, Protein Probability 1.0 MW 38,036 Da

b 102

215

344

473

587

700

829

T

L

E

E

I

L

E D V P E

1531 1430 1316 1187 1058 × 105

945

944

832

703

1043 1140 1269 1356 1531 588

489

391

S R

262

175 y

y4

1.2

Relative Abundance

1.0

0.8

b9 0.6

0.4

b5 y5 b6 0.2

0

y2 b2 200

b3

y3 400

y6

b7 y

7

b8 y8

y9

b4 600

800

1000

y10 1200

b11 y11 b12 1400

m/z

FIGURE 1.6 Fragmentation mass spectrum and sequence analysis of peptide identiied from TM173. An almost complete series of singly charged b- and y-type ions were used to unambiguously identify this doubly charged peptide TLEEILEDVPESR from TM173 (MW 38,036 Da). The PeptideProphet probability was 0.9993 and the ProteinProphet probability was 1.000 for this identiication.

REFERENCES

REFERENCES 1. Chien, C.T., Bartel, P.L., Sternglanz, R., Fields, S. (1991) The two-hybrid system: a method to identify and clone genes for proteins that interact with a protein of interest. Proceedings of the National Academy of Sciences of the United States of America, 88, 9578–9582. 2. Suter, B., Kittanakom, S., Stagljar, I. (2008) Interactive proteomics: what lies ahead? BioTechniques, 44, 681–691. 3. Gavin, A.-C., Bosche, M., Krause, R., Grandi, P., et al. (2002) Functional organization of the yeast proteome by systematic analysis of protein complexes. Nature, 415, 141–147. 4. Ho, Y., Gruhler, A., Heilbut, A., Bader, G.D., et al. (2002) Systematic identiication of protein complexes in Saccharomyces cerevisiae by mass spectrometry. Nature, 415, 180–183. 5. Ewing, R.M., Chu, P., Elisma, F., Li, H., et al. (2007) Largescale mapping of human protein-protein interactions by mass spectrometry. Molecular Systems Biology, 3, 2–17. 6. Choudhary, C., Mann, M. (2010) Decoding signaling networks by mass spectrometry-based proteomics. Nature Reviews Molecular Cell Biology, 11, 427–439. 7. Yang, W., Steen, H., Freeman, M.R. (2008) Proteomic approaches to the analysis of multiprotein signaling complexes. Proteomics, 8, 832–851. 8. Ashman, K., López Villar, E. (2009) Phosphoproteomics and cancer research. Clinical & Translational Oncology, 11, 356–362. 9. Choudhary, C., Kumar, C., Gnad, F., Nielsen, M.L., et al. (2009) Lysine acetylation targets protein complexes and co-regulates major cellular functions. Science, 325, 834–840. 10. Schirle, M., Heurtier, M.-A., Kuster, B. (2003) Proiling core proteomes of human cell lines by one-dimensional PAGE and liquid chromatography-tandem mass spectrometry. Molecular & Cellular Proteomics, 2, 1297–1305. 11. Kumar, C., Mann, M. (2009) Bioinformatics analysis of mass spectrometry-based proteomics data sets. FEBS Letters, 583, 1703–1712. 12. Hubbard, S.J., Jones, A.R., Elias, J.E., Gygi, S.P. (2010) Proteome Bioinformatics. Methods in Molecular Biology, Vol. 604. London: Humana Press.

21

13. Ashburner, M., Ball, C.A., Blake, J.A., Botstein, D., et al. (2000) Gene Ontology: tool for the uniication of biology. The Gene Ontology Consortium. Nature Genetics, 25, 25–29. 14. Okuda, S., Yamada, T., Hamajima, M., Itoh, M., et al. (2008) KEGG Atlas mapping for global analysis of metabolic pathways. Nucleic Acids Research, 36, W423–W426. 15. Carvalho, P., Fischer, J., Chen, E., Domont, G., et al. (2009) GO Explorer: a gene-ontology tool to aid in the interpretation of shotgun proteomics data. Proteome Science, 7, 6. 16. Carvalho, P.C., Fischer, J.S.G., Chen, E.I., Yates, J.R., et al. (2008) PatternLab for proteomics: a tool for differential shotgun proteomics. BMC Bioinformatics, 9, 316. 17. Malovannaya, A., Li, Y., Bulynko, Y., Jung, S.Y., et al. (2010) Streamlined analysis schema for high-throughput identiication of endogenous protein complexes. Proceedings of the National Academy of Sciences, 107, 2431–2436. 18. Moresco, J.J., Carvalho, P.C., Yates, J.R. III (2010) Identifying components of protein complexes in C. elegans using co-immunoprecipitation and mass spectrometry. Journal of Proteomics, 73, 2198–2204. 19. Kapp, E., Schütz, F. (2007) Overview of tandem mass spectrometry (MS/MS) database search algorithms. Current Protocols in Protein Science, 49, 25.2.1–25.2.19. 20. Dennis, G., Sherman, B., Hosack, D., Yang, J., et al. (2003) DAVID: database for annotation, visualization, and integrated discovery. Genome Biology, 4, P3. 21. Huang, D.W., Sherman, B.T., Lempicki, R.A. (2008) Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nature Protocols, 4, 44–57. 22. Bao, J., Sack, M. (2010) Protein deacetylation by sirtuins: delineating a post-translational regulatory program responsive to nutrient and redox stressors. Cellular and Molecular Life Sciences, 67, 3073–3087. 23. Jin, L., Waterman, P.M., Jonscher, K.R., Short, C.M., et al. (2008) MPYS, a novel membrane tetraspanner, is associated with major histocompatibility complex class II and mediates transduction of apoptotic signals. Molecular and Cellular Biology, 28, 5014–5026. 24. Petrak, J., Ivanek, R., Toman, O., Cmejla, R., et al. (2008) Déjà vu in proteomics. A hit parade of repeatedly identiied differentially expressed proteins. Proteomics, 8, 1744–1749.

2 MASS SPECTROMETRY-BASED METHODS TO INVESTIGATE POSTTRANSLATIONAL PROTEIN MODIFICATIONS BY LIPID PEROXIDATION PRODUCTS Navin Rauniyar and Laszlo Prokai

Covalent modiications of proteins are essential posttranslational processing steps that involve linkage of various chemical groups (phosphate, methyl, acetyl, and many others) to speciic amino acid residues. These modiications play an important role in rendering the proteins functionally active for numerous cellular functions such as regulating cellular location and dynamic interactions with other proteins, achieving required three-dimensional conformation or catalytic activity. On the other hand, several posttranslational modiications (PTMs) arise from deleterious processes that compromise protein function. Many chemically distinct types of PTMs are known to occur physiologically (A comprehensive list of protein PTMs can be found at http:// www.abrf.org/index.cfm/dm.home?AvgMass=all). PTMs of proteins increase the diversity and complexity of a proteome in an organism by providing covalent variation to protein backbones and side chains. It is imperative to identify covalent protein modiications in order to decode their role in biological processes. Mass spectrometry (MS) has emerged as a powerful technology for protein identiication and characterization of PTMs. However, comprehensive analysis of posttranslationally modiied peptides in complex mixtures is a challenging endeavor because of the necessity to digest proteins to successfully sequence the resulting peptides by MS in bottom-up proteomics approach [1]. The modiied peptides occur in substoichiometric levels compared with corresponding native forms in the analytes, a situation analogous to “inding a needle in a haystack.”

Also, the modiied peptides may ionize less eficiently by electrospray ionization (ESI) and matrix-assisted laser desorption/ionization (MALDI) than those of the regular, unmodiied peptides in a complex mixture, and the low-abundance ions of modiied peptides that coelute with intense ions of unmodiied peptides may be missed by conventional data-dependent acquisition [2]. The recent advent of a new generation of mass spectrometers with ultra-high mass resolution, exceptionally high mass accuracy and sensitivity, and unique ion manipulation capabilities such as electron capture dissociation (ECD) and electron transfer dissociation (ETD), together with development of various excellent database search algorithms, has proven invaluable for global interrogation of protein modiications from the acquired mass spectra. This has resulted in an unprecedented growth in studies oriented toward comprehensively mapping of protein PTMs. Additionally, the implementation of several enrichment techniques, including covalent coupling and afinity capture, for purifying the “subproteome” carrying modiied forms from native species and simultaneously reducing the sample complexity and peptide coelution prior to mass spectrometric analysis, greatly enhances the chance of detecting those modiications that are substoichiometric in nature or are affecting low-abundance proteins. Thus, the combination of high-performance MS with new computational techniques and experimental methods are facilitating large-scale analysis of PTMs.

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

23

24

MASS SPECTROMETRY-BASED METHODS TO INVESTIGATE POSTTRANSLATIONAL PROTEIN MODIFICATIONS

SCHEME 2.1 Reaction of –NH2 groups, imidazole =NH group, and –SH groups of lysine (Lys, K), histidine (His, H), and cysteine (Cys, C) side-chains, respectively, in proteins with 4-hydroxy-2-nonenal (HNE) through MA (a, +156 Da). HNE can also form SB adducts (b, +138 Da) with –NH2 groups.

PTMs usually are site speciic, located at speciic amino-acid residues at a particular conserved sequence motif in proteins. However, modiications by 4-hydroxy2-nonenal (HNE), a reactive carbonyl species (RCS) produced from polyunsaturated fatty acids (PUFAs) during lipid peroxidation, in proteins according to reactions shown in Scheme 2.1, cannot be predicted by computational sequence analysis and, therefore, experimental proteomic techniques have to be developed for their determination. Michael adducts generally represent >99% of HNE protein modiications by HNE, whereas Schiff-base (SB) adduct formation is less prevalent even in the presence of excess HNE [3,4]. A bottom-up protocol for determination of posttranslational protein carbonylation by HNE requires high-eficiency frontend enrichment of modiied peptides from biological samples, for which this chapter describes the method that employs solid-phase hydrazide (SPH) chemistry summarized in Scheme 2.2 [5]. In addition, several MSbased approaches have been developed to identify HNE-modiied peptides through coupling with online liquid chromatographic separation, data-dependent acquisition, and subsequent database searches from tandem mass spectra [6–9]. From this handbook’s thematic viewpoint, speciic highlights of the procedure are data-dependent acquisitions that utilize complementary ion dissociation methods such as collision-induced dissociation (CID) and ECD with tandem mass spectrometry (MS/MS). In case of CID, “third-order” tandem

SCHEME 2.2 Schematic illustration of the procedure used for solid-phase hydrazide (SPH) enrichment of 4-HNEcarbonylated peptides. Adapted from Reference [9], ©2010 Wiley Interscience, with permission.

IDENTIFICATION OF HNE-MODIFIED PEPTIDES IN BIOLOGICAL SAMPLES

mass spectrometry (MS/MS/MS or MS3) triggered by modiication-speciic neutral loss (NL) is also used to interrogate peptides with covalently attached HNE for improved identiication and localization of modiication sites. These methods can be implemented on modern linear ion trap and, especially, ion trap–Fourier transform hybrid mass spectrometers.

2.1

PROCEDURE CONTROL

Although the protocol is aimed at discovery-driven exploration of posttranslational HNE modiications in proteins and, hence, rigorous quality control does not apply, initial demonstration of performance is recommended before application to complex biological samples including cell lysates. Prepare an HNE-modiied protein for such demonstration by incubating a protein purchased from a commercial vendor (see the section “Reagents and Standards”) or, in speciic proteintargeted experiments, isolated for the purpose of the study. Operating procedures pertaining to the SPHbased enrichment procedure, as well as the CID-based NL-driven MS3 and NL-driven ECD acquisitions coupled with online nanoscale liquid chromatography (nano-LC) should be veriied with the tryptic digest of the modiied standard protein(s). Compare the results of performance demonstration to data available from the literature (e.g., [7,9]). For label-free quantitation, consider exploratory validation of spectral counts or extracted ion chromatograms (XICs) versus protein concentrations by spiking the prepared HNE-modiied protein into a sample matrix such as a cell lysate [10].

2.2 IDENTIFICATION OF HNE-MODIFIED PEPTIDES IN BIOLOGICAL SAMPLES BY SOLID-PHASE ENRICHMENT AND NANO-LC–ESI-MS/MS This protocol describes a procedure for the identiication of PTM by HNE to reveal protein targets and speciic sites of covalent attachment. Identiication is performed by combining proteolytic digestion followed by SPH enrichment and nanoscale liquid chromatography– electrospray ionization tandem mass spectrometry (nano-LC–ESI-MS/MS). The modiied proteins are subjected to reduction, alkylation, and subsequent digestion by a proteolytic enzyme. The peptides thus obtained are desalted and the substoichiometric quantities of HNE-modiied peptides are fractionated from unmodiied species using hydrazide-coated glass beadsbased enrichment technique that selectively captures peptide carbonyls (as hydrazones) that are subsequently

25

released from the beads in their original forms by acid-catalyzed hydrolysis. The enriched HNE-modiied peptides are analyzed by several MS-based approaches, including data-dependent acquisition that utilize CID and ECD as a method of peptide dissociation, as well as NL-driven collision-induced dissociation–mass spectrometry (CID-MS3) and NL-driven electron capture dissociation tandem mass spectrometry (ECD-MS/MS) methods for in-depth proiling of carbonylation sites in the peptides. HNE-modiied peptide matches from the database search algorithms are visually inspected to conirm the mass shifts and only peptides with unambiguous sites of modiication within their amino acid sequence should be considered correct. The procedure can be veriied using HNE-modiied synthetic peptide standards such as human angiotensin (AGT) I peptide (DRVYIHPFHL), AGT II octapeptide (DRVYIHPF), and custom synthesized peptides LVLEVAQHLGESTVR and IVYGHLDDPANQEIER (which correspond to tryptic peptides of ATP synthase subunit β and aconitate hydratase, respectively) [7]. HNE-modiied apomyoglobin from equine skeletal muscle may also be used for the purpose of veriication [9]. Although the characteristic elimination of HNE (156 Da) in CID can serve as a signature tag for the modiied peptides, ECD is recommended as a complementary method of ion dissociation to aid the elucidation of primary structural information and assignment of exact carbonylation sites in the protein. A nearly complete fragment ion series is generated with ECD due to eficient peptide backbone cleavage (in most cases over 75%) and is also capable of retaining the labile HNE moiety in the peptides [9]. 2.2.1

Materials

Water, high performance liquid chromatography (HPLC) grade (EMD [Philadelphia, PA] or equivalent) Acetonitrile, HPLC grade (EMD or equivalent) Acetic acid, 99.99% (Sigma [St. Louis, MO] or equivalent) Formic acid, ≥98% (Sigma or equivalent) HNE (5 mg/500 µL in ethanol, Cayman Chemical Company, Ann Arbor, MI) Apomyoglobin from equine skeletal muscle (Sigma) for modiication by HNE (see the section “Reagents and Standards”), or similar protein commercially available or isolated Lysis buffer for mammalian cells: 20 mM TrisHCl/ 50 mM NaCl/6 M urea (pH 8.1)/10 mM sodium pyrophosphate/1 mM sodium luoride/1 mM βglycerophosphate/1 mM sodium orthovanadate/1

26

MASS SPECTROMETRY-BASED METHODS TO INVESTIGATE POSTTRANSLATIONAL PROTEIN MODIFICATIONS

tablet complete Mini protease inhibitor mixture (Roche, Indianapolis, IN) dissolved per 10 mL buffer Laboratory centrifuge (Eppendorf [Hauppauge, NY] or equivalent) Laboratory vortex device (Vortex-Genie [Bohemia, NY] or equivalent) Speedvac concentrator (Eppendorf Vacufuge or equivalent) Kit and equipment for spectrophotometric protein concentration measurement by Lowry, Bradford, or bicinchoninic acid (BCA) assay (Thermo [San Jose, CA] or equivalent) 1.5 and 0.5 mL polypropylene centrifuge tubes (Eppendorf) 50 mM Ammonium bicarbonate buffer, pH 7.4 Phosphate-buffered saline (PBS) Dithiothreitol (DTT, Sigma) Iodoacetamide (IAA, Sigma) Sequencing-grade modiied porcine trypsin (Promega, Madison, WI) Sequencing-grade triluoroacetic acid (TFA) ≥98% (Sigma) Solid-phase extraction (SPE) C18 column (Sep-Pak, Waters, Milford, MA) SPE solvents: Binding buffer (1% acetic acid, v/v, in H2O) and elution buffer (80% acetonitrile v/v, plus 1% acetic acid, v/v, in H2O). SPH reagent (prepared according to Roe et al. [5]) Buffers and solutions for SPH chemistry: Reaction buffer (0.2% acetic acid v/v, 10% acetonitrile v/v, pH 3.6); washing solvent and solutions (distilled water, 1 M aqueous NaCl and 80:20 v/v acetonitrile : H2O); elution solution (10% v/v aqueous formic acid) 250 µL polypropylene autosampler vials with presson caps and telon-lined septa (National Scientiic Company, Rockwood, TN) High-performance nanoscale liquid chromatography system (Eksigent [Dublin, CA] nano-LC™ or equivalent), equipped with autosampler IntegraFrit™ sample trap, 25 mm × 75 µm internal diameter (i.d.) (New Objective [Woburn, MA]) or equivalent preconcentration column PepMap C18 analytical column, 150 mm × 75 µm i.d., 3 µm particle size, 100 Å pore size (LC Packings, San Francisco, CA) or equivalent nano-LC column packed with octadecylsilica particles. Picotip emitter, i.d. 10 ± 1 µm (New Objective)

Loading solvent for reversed-phase liquid chromatography (RPLC): 0.1% (v/v) acetic acid and 5% (v/v) acetonitrile in 94.9% (v/v) water Mobile phase A for reversed-phase (RP) column: 0.1% (v/v) acetic acid and 99.9% (v/v) water Mobile phase B for RP column: 0.1% (v/v) acetic acid and 99.9% (v/v) acetonitrile Hybrid linear ion trap–Fourier transform ion cyclotron resonance (FTICR) mass spectrometer, 7T (LTQ-FTICR, Thermo), or equivalent instrument, equipped with an electrospray/nanospray (ESI/ NSI) source and ECD accessory, operated under the control of Xcalibur software package (Thermo Fisher Scientiic, San Jose, CA; Instrument Coniguration, Instrument Setup, Sequence Setup, and Qual Browser modules appropriately created and used) with Bioworks available to preprocess raw data iles for database search. ECD may be replaced with ETD and the respective manufacturer’s own version of the above software should be used, when alternative instrument is selected. International Protein Index database for the species studied (available online at http://www.ebi.ac. uk/IPI) and its sequence-reversed version with the Mascot (Matrix Science, Boston, MA) search algorithm. Equivalent search software (Proteome Discoverer [Thermo Fisher Scientiic], Protein Prospector [http://prospector.ucsf.edu/prospector/ mshome.htm], Phenyx [GeneBio, Geneva, Switzerland], X!Tandem [http://www.thegpm.org/tandem/], etc.) may also be used. Additional program(s) to validate the MS/MS-based peptide and protein identiications, to summarize results including MS/MS spectral counts (Scaffold, Proteome Software, Portland, OR), and to create aligned extracted-ion chromatograms (XICs, Sieve®, Thermo), and so on. 2.2.2

Preparation of Cell Lysates

1. Harvest the cells by centrifugation at 2000 × g in 1.5 mL centrifuge tubes for 15 min at 4°C. Pour off the supernatant and discard it. Note: During lysis, cells and lysates should be kept at 4°C at all times. Temperature has a profound effect on the catalytic activity of most proteases. 2. Carefully wash the cell pellet twice with ice-cold PBS, and then place the washed cell pellet on ice. Resuspend the pellet in 1 mL of lysis buffer previously chilled to 4°C. Incubate the cells for 30 min on ice with occasional vortexing of the tube. Perform freeze/thaw cycles to complete the rupture of the

IDENTIFICATION OF HNE-MODIFIED PEPTIDES IN BIOLOGICAL SAMPLES

cell membrane and release the contents. Centrifuge the resultant solution at 20,000 × g for 15 min at 4°C to separate the bulk of cell debris from the lysates. Carefully transfer the supernatant to a fresh tube, making sure not to disturb the pellet. 3. Measure protein concentration using Lowry, Bradford, or BCA protein assay and bovine serum albumin (BSA) as a reference. Note: It is recommended to continue with the trypsin digestion step immediately to minimize protein degradation. If lysates need to be stored at this point, then snap-freeze them using a dry ice/ ethanol mixture and store them at −80°C.

2.2.3

Protein Reduction and Alkylation

1. Add DTT, also known as Cleland’s reagent, from a 0.5 M stock solution to reach 10 mM concentration in the cell lysate and incubate for 30 min at 56°C to reduce the disulide bonds. Note: Incubation at higher temperature may cause urea-based carbamylation of lysines and protein N-termini. 2. Allow the reduced protein solution to cool to room temperature and add IAA from a 0.5 M stock solution to reach 20 mM concentration in the reduced cell lysate. Incubate the reaction for 30 min at room temperature in the dark to alkylate the cysteine residues. Modiication of the protein thiol groups by alkylation prevents reformation of disulide bonds due to oxidation upon exposure to air. 3. Quench unreacted IAA by increasing DTT content by 10 mM (see step 1 above) and incubating for 15 min at room temperature in the dark.

2.2.4 Trypsin Digestion 1. Dilute the protein mixture in 50 mM ammonium bicarbonate, pH 7.4, to reduce urea concentration (6 M in the lysis buffer) to about 1.6 M. 2. Add trypsin at a inal protease-to-protein ratio of 1:50 (w/w) and incubate at 37°C overnight. Longer incubations, up to 24 h, may be required depending on the nature of the protein. Note: Trypsin is a serine protease that speciically cleaves at the C-terminal side of arginine (R) and lysine (K). Restrictions to the speciicity of trypsin occur when proline (P) is at the C-terminal side of R or K; these peptide bonds are almost completely resistant to cleavage by trypsin.

27

3. After incubation, allow the digest to cool to room temperature. Remove a small aliquot of the digest and freeze the remaining reaction mixture. Acidify the aliquot with TFA to pH < 2 to stop the digestion. Note: You may determine the extent of digestion by analyzing a portion of the digestion products by routine sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE; no protein bands should be detected). If further proteolysis is required, reincubate the reaction mixture at 37°C until the desired digestion is obtained. After completing the digestion, cool the reaction mixture and lower the pH of the solution below 2 by the addition of TFA. 4. Centrifuge at 2500 × g for 10 min at room temperature and discard the pellet. The peptides must be desalted to remove salts and urea from the digestion buffer. In our laboratory, peptide desalting is achieved by reverse-phase C18 Sep-Pak SPE cartridges (Waters). 2.2.5

Peptide Desalting

The size of the cartridge should be selected based on the amount of protein and the volumes should be adapted accordingly for different sizes. Follow the description enclosed by the manufacturer of the cartridge to perform the procedure. A typical sequence includes the steps below: 1. Wash and condition the cartridge using 1 mL of acetonitrile. Note: Preconditioning of the sorbent with an organic solvent is necessary. Without this step, a highly aqueous solvent cannot penetrate the hydrophobic surface and wet the sorbent. Thus, only a small fraction of the sorbent surface area would be available for interaction with the analyte. 2. Equilibrate with 1 mL of 1% acetic acid (v/v) in H2O. 3. Load the acidiied sample onto the cartridge. 4. Wash/desalt with 1 mL of 1% acetic acid (v/v) in H2O. The washing step removes interferences while retaining the analyte. 5. Elute with 1 mL of 80% acetonitrile (v/v), 19% H2O (v/v), and 1% acetic acid (v/v) to collect the desalted sample. 6. Freeze the eluate obtained after SPE with liquid N2 and lyophilize it over a few hours almost to dryness in a lyophilizer (Labconco, Kansas City, MO). A volume of approximately 10 µL should

28

MASS SPECTROMETRY-BASED METHODS TO INVESTIGATE POSTTRANSLATIONAL PROTEIN MODIFICATIONS

remain. Complete dryness may result in some sample loss. At this point, samples can be stored at –20°C for several weeks or reconstituted in reaction buffer and proceed to enrich carbonylated peptides using SPH strategy. 2.2.6 Enrichment of HNE-Modiied Peptides Using an SPH Reagent In order to enrich and identify suficient number of HNE-modiied peptides, starting from at least 5 mg of total protein is recommended. 1. In a 0.5 mL centrifuge tube, add 6–7 mg of SPH reagent to peptides dissolved in 100–150 µL of reaction buffer, and incubate the resulting mixture overnight at room temperature rotating end over end. The pH of 3.6 favors the formation of hydrazone linkage between the aldehyde group in the carbonylated peptides and the immobilized hydrazide group on the surface of glass beads. 2. Following incubation, pellet the SPH reagent by centrifugation and carefully transfer (so as not to disturb the pellet) the supernatant containing the unbound unmodiied peptides to a tube labeled “SPH low through” and save it for further analysis. The unmodiied peptides do not bind to the SPH beads and, if one is interested, the lowthrough or wash fractions can be analyzed by liquid chromatography–tandem mass spectrometry (LC–MS/MS) to identify the unbound noncarbonylated peptides. 3. To get rid of the unmodiied peptides adsorbed to the glass surface, wash the pelleted SPH beads with 400 µL of reaction buffer. Repeat this step three more times. Subsequent washes should be performed each with 400 µL of 1 M NaCl (four times, 4×), distilled water (4×), 80% acetonitrile (4×), and a inal wash by distilled water (4×). Each washing step involves a mild vortexing (30 s) followed by a brief centrifugation to settle the beads and changing the wash solution. 4. Release the carbonylated peptides from the SPH beads by incubating with 200 µL of 10% formic acid at 60°C for 30 min by rotating end over end. Following incubation, collect and save the released peptides. Repeat this step one more time. Combine eluates from the same sample into the same tube. Lyophilize the eluate containing the HNEmodiied peptides. 5. Resuspend samples in 20 µL of loading solvent. Centrifuge the samples at 14,000 rpm for 10 min

to pellet any solid material. If any pellet is visible, then transfer the supernatant (peptide sample) to a fresh tube. 2.2.7 RPLC–Neutral Loss-Driven Tandem Mass Spectrometry (NL-MS3 and NL-Driven ECD) 2.2.7.1 Nano-LC Separation A protein digest can be loaded onto a nano-LC system, separated by gradient elution and the eluate from the RP column analyzed by MS. Install the 25 mm × 75 µm i.d. IntegraFrit sample trap (New Objective) and the 150 mm × 75 µm i.d. PepMap C18 analytical column (LC Packings) according to the instruction manual by the nano-LC’s manufacturer. 1. For online separation with an Eksigent nano-LC system or equivalent, 250 nL/min should be used as low rate. Mobile phases will be mixed from solvents A (0.1% acetic acid and 99.9% water, v/v) and B (0.1% acetic acid and 99.9% acetonitrile, v/v). 2. Inject 5 µL of the sample onto the sample trap at 1.5 µL/min low rate in a loading solvent for 5-min sample concentration and desalting while the mobile phase is kept isocration at 5% B. 3. Following desalting, elute the peptides from the trap column onto the nano-LC column using a linear 90-min gradient from 5% to 40% mobile phase B at a constant low rate of 250 nL/min through the column. The majority of peptides will elute on a C18 column at ∼40% B. Finally, ramp the gradient up to 90% mobile phase B for high organic wash step. Set a ramp-down period of 5 min, during which the concentration of mobile phase B returns to 5%. 4. Peptides eluted through the Picotip emitter is directly supplied into the nano-ESI source of the mass spectrometer. Maintain the spray voltage and capillary temperature during the gradient run at 2.0–2.4 kV and 250°C, respectively. The camera attached to the nano-ESI source with its output cable plugged into the computer performing data acquisition will enable you to display a real-time zoomed image of the Picotip on the desktop and, thus, to check spray quality during the run. 2.2.7.2 Tandem Mass Spectrometry NL-driven CIDMS3 and NL-driven ECD-MS/MS can be performed, among others, on a hybrid linear ion trap-FTICR mass spectrometer (LTQ-FT, Thermo Finnigan). The use of an instrument capable of high mass accuracy measurements both for precursor ions and product ions

IDENTIFICATION OF HNE-MODIFIED PEPTIDES IN BIOLOGICAL SAMPLES

gives high conidence in protein identiication with greater protein coverage when performing database searches, and it also increases the probability of unambiguous localization of sites for PTMs in peptides. Data-dependent acquisition of MS/MS and MS3 spectra during the online gradient nano-LC separation is programmed through Xcalibur (Instrument Setup). The following settings are used to create a typical method: 1. Select Instrument Setup in the main page of Xcalibur software. 2. Select Data-dependent MS/MS method in the next window. 3. The pop-up window will contain design parameters for the acquisition setup. The top region of the page contains the Run settings group box that allows you to deine the MS acquisition time, divide the acquisition time into segments, and specify the start and end time of each segment. In the top left corner of the Run settings group box, input the total MS detector Acquire time that would be the same as the duration of the HPLC gradient. 4. The Segments indicate the number of scan segments to occur during the period of time speciied in the Acquire Time text box. Start Delay allows us to specify a time delay before the start of acquisition. Set the Segments setting to 1 and the Start Delay setting to 0 for the acquisition. 5. Set the Scan events to 6 for a method where mass spectrometer performs one round of MS scan followed by acquisition of ive MS/MS scans on the ive most intense ions (recorded in the MS scan) in the linear ion trap. So the irst triggered scan event is a full MS scan, and the next ive scan events are MS/MS. During the setting of Scan event 1, select Analyzer as FTMS (Fourier transform mass spectrometry) for mass analysis that provides high resolution and mass accuracy, Mass Range as Normal, Resolution of 50000 at massto-charge ratio (m/z) 400 and Scan Type as Full. Set the Scan Ranges to 350–1500, the Polarity option to Positive for analyzing positive ion, and the Data type to Proile mode of data collection. 6. The Tune method box speciies the path for a ile containing the parameters for the electrostatic lenses and ion trap. Tune methods are deined and saved by using the Tune Plus window. Browse for the appropriate nano-ESI tune ile and click to associate it with the method.

29

7. In the Scan Event 2, select Analyzer as Ion Trap for MS/MS, Mass Range and Scan Rate as Normal, and select Centroid option for Data type. Check the box next to Dependent scan, and then click the Settings button to open the Data Dependent Settings dialog box that allows us to set parameters for performing data-dependent scans. 8. The Global page of the Data Dependent Settings dialog box allows selecting dependent scan settings that apply to all dependent scans. The values of Global and Mass Widths settings are left at the default value. 9. Check the Enabled box in the Dynamic Exclusion page of the Data Dependent Settings dialog box to select parameters to prevent for a subsequent data-dependent scan after it has already triggered a data-dependent scan. This way, dynamic exclusion prevents an abundant peptide ion with a broad elution proile from being continually selected for MS/MS and allows the ion trap to collect MS/MS spectra on other, less intense ions that would otherwise not be examined. The typical settings include Repeat count of 2, Repeat duration of 30 s, Exclusion list size of 50 and Exclusion duration of 180 s. 10. The Segment page allows setting the parameters for the active segment. Check the box for Enable preview mode for FTMS master scans in the Current Segment dialog box. In this mode, the data-dependent decision is made on the basis of the FTMS master scan with lower resolution to increase the duty cycle. Also check boxes corresponding to Enable charge state screening and Enable charge state rejection of singly charged ions for precluding these ions from triggering data-dependent MS/MS scans. 11. The Current Scan Event window of the Data Dependent Settings dialog box allows for selection of the parameters to execute (or not to execute) data-dependent acquisition on currently selected scan event. a. The Minimum signal threshold (counts), which represents the signal intensity needed for a peak recorded in an MS scan to trigger the mass spectrometer to select this peak for MS/ MS, should be set to 1000. Select the ranking of ion abundances per MS scan (Scan event 1) by inserting 1 in the option Mass determined from scan event. b. Set value 1 to Nth most intense ion in order to trigger MS/MS of the most intense ion observed in the MS scan.

30

MASS SPECTROMETRY-BASED METHODS TO INVESTIGATE POSTTRANSLATIONAL PROTEIN MODIFICATIONS

c. Choose Activation type CID as the method of peptide dissociation. d. The Default charge state is set to 2, which makes the assumption that all peptides entering the mass spectrometer are doubly charged, and Isolation width, the window around the ion in the MS scan that is selected for MS/MS, is set as 4.0 (m/z, thomsons). e. The Normalized collision energy is set to 35%. This is the amount of energy delivered to the peptide to cause fragmentation. f. The Activation Q and Activation time are both left at their default setting of 0.250 and 30 ms, respectively. 12. The process is repeated for Scan events 3 through 6 by going back to the MS Detector Setup page and irst pressing the Scan event 3 bar, checking Dependent scan and then pressing Settings button. Once set, the mass spectrometer will continually cycle through these six scans, collecting MS/MS spectra in a data-dependent fashion throughout the duration of the gradient without any user intervention. Proceed to Option 1 below to set up data-dependent NL-driven CID-MS3 acquisitions, or skip to Option 2, if NL-driven ECD-MS/MS is employed. 13. Option 1: Data-Dependent NL-driven CID-MS3 for HNE Modiications a. Choose option Data dependent NL MS3 in Select experiment type group box in Instrument Setup page. b. Input value 3 in the box corresponding to Analyze top N peaks in order to trigger the MS3 scan on top 3 (in intensity) NL ion peaks observed in the MS/MS scan. c. Fill in the Neutral loss masses box. The NL masses for Michael adducts of HNE with respect to parent ion are m/z 78 and 52 for doubly and triply charged peptides, respectively. Press OK to accept changes and close the dialog box. d. Deine the MS acquisition time, number of segments in the acquisition time, and the start and end time of each segment in the Run settings group box (top region of the page), as described previously for conventional datadependent MS/MS acquisition strategy. e. Set the number of events in the Scan events scroll box to 3 and specify the Tune ile in Tune method. f. Select the settings in the Scan Event 1 as described previously.

g. In the Scan Event 2, select Analyzer as Ion Trap, Mass Range and Scan Rate as Normal, and select Centroid option in the Data type text box. • Click the Settings button to open the Data Dependent Settings dialog box that allows you to set parameters for performing a datadependent scans. • The values of Global and Mass Widths settings are left at the default value. • Check the Enabled box in the Dynamic Exclusion page of the Data Dependent Setting dialog box and select values for the parameters mentioned previously. h. The table in Segment Level Neutral Loss page of the Data Dependent Settings dialog box will display the NL masses. • Input value 3 in the text box corresponding to within top N. This prevents random MS3 events and limits the selection of ions for MS3 acquisition only to the NL fragment ions whose intensities are within the top three peaks in the MS/MS spectrum. • Set the value 5 for Analyze top N peaks in Current Scan Event dialog box in order to perform MS/MS on the ive most intense ions detected in Scan Event 1, starting with the most intense ion and proceeding in order of decreasing intensity. • The values for other parameters are set as mentioned previously for Data-dependent MS/MS. The MS will irst perform an accurate m/z survey scan in the ion cyclotron resonance (ICR) cell, followed by parallel CID MS/MS linear in ion trap analysis of the top ive most intense precursor ions selected from interim survey spectra using 35% normalized collision energy with helium as the target gas. i. The execution of Scan Event 3 is dependent on the presence of an NL ion within three most intense ions in the MS/MS of an ion in Scan Event 2. Scheme 2.3 shows the schematic representation of the NL-MS3 method for the analysis of Michael adducts. In this igure, the NL ion showing the designated difference (m/z 78 or 52 for Michael adducts) is selected for an additional round of fragmentation giving MS3 spectra. NL-MS3 strategy can also be successfully implemented for identifying SB modiication site(s) in tryptic peptides of HNE-modiied proteins. However, unlike

IDENTIFICATION OF HNE-MODIFIED PEPTIDES IN BIOLOGICAL SAMPLES

SCHEME 2.3 Flowchart of data-dependent NL-driven MS3 (NL-MS/MS/MS) that employs CID only.

Michael adducts, the chemical structure of SB adducts precludes the use of SPH enrichment strategy for isolating HNE modiication to proteins by this mechanism. Note: For NL-driven CID-MS3 data-dependent acquisitions of SB adducts of HNE, adjust the acquisition parameter so as to isolate and subsequently fragment the ions exhibiting an m/z 69 or 46 difference (representing NL of HNE from doubly or triply charged precursor ions of SB adducts, respectively) from the precursor ion and also if the NL fragment ions passed speciied selection criteria (i.e., they were among the three most intense ions in the MS/MS spectra). The details of the method have been described elsewhere [8]. Option 2: NL-driven data-dependent ECD-MS/MS a. In the NL-driven ECD-MS/MS method employed to HNE-modiied peptides, set the Scan Event 1 and 2 parameters similar to NL-driven CID-MS3 method. Acquire full-scan mass spectra (from m/z 350 to 1500) in the ICR cell with a resolving power of 50,000 at m/z 400 during the irst scan event in the LTQ-FT. Using data-dependent acquisition, isolate pre-

31

cursor ions of the most abundant peptides eluted at a particular chromatographic time point from the nano-LC column with an isolation width of 4 Th and subject to CID fragmentation in the linear ion trap. Induce the dissociation of the precursor ion using an activation time of 30 ms and at normalized collision energy of 35%. b. In Scan Event 3, select Analyzer FTMS for mass analysis. • Select the Normal mass range with a resolution of 25,000 at m/z 400 and specify Proile mode of data collection. • In the Current Scan Event in Data Dependent Settings dialog box check the option Repeat previous scan event with ECD to trigger ECD-MS/MS of precursor ion showing NL in Scan Event 2. Thus, ECD will be triggered on the same precursor ion that shows an NL of 156 Da (loss of 78 or 52 Th from a doubly or triply charged precursor ion) and is one of the three most abundant fragment ions in the previous CID-MS/MS scan. • Specify the default charge state as 2 and isolation width as 4 and check the box corresponding to ECD active. c. Specify the parameters of the ECD in the FT ECD/IRMPD page of the Data Dependent Settings dialog box. • The default electron energy value, 5, is set in the Energy column. • 100 (milliseconds) is set for the Time delay between trapping the ions in the FT-ICR analyzer and the start of ECD. • Input 120 ms for Duration of ECD. 14. Save the methods and run the samples through the data acquisition module of Xcalibur. Scheme 2.4 summarizes the steps incorporated in the NL-ECD-MS/MS experiment.

2.2.8

Data Analysis

Perform data analysis to identify HNE-modiied peptides as follows (linear ion trap–FTMS hybrid instruments), or through a procedure appropriately modiied to conirm your instrument and software used: 1. Raw data iles generated by data-dependent NLdriven CID-MS3 and NL-driven ECD-MS/MS acquisitions are extracted by BioWorks version 3.3 for the subsequent database search using the

32

MASS SPECTROMETRY-BASED METHODS TO INVESTIGATE POSTTRANSLATIONAL PROTEIN MODIFICATIONS

SCHEME 2.4 Flowchart of data-dependent NL-driven ECD-MS/MS that employs both CID and ECD.

Mascot (Matrix Science) search algorithm and the International Protein Index (available online at http://www.ebi.ac.uk/IPI), or similar database. The search against the normal database combined with the corresponding reversed or randomized database is an excellent validation method for MS/MS searches of large data sets. 2. MS/MS spectra is searched using the Mascot search algorithm with parent ion and fragment ion mass tolerances of 10.0 ppm and 0.80 Da, respectively, and specifying trypsin as digesting enzyme with two missed cleavages allowed. For NL-driven CID-MS3 experiments, a parent ion tolerance of 1.5 Da is utilized since the actual NL occurs in the linear ion trap. Carbamidomethylation of Cys, oxidation of methionine, carbamylation of Lys and the N-terminal amino acid residue, and HNE–Michael adduct formation on Cys, His, and Lys are speciied as dynamic modiications. Dynamic mass modiication means both possibilities of modiied or unmodiied amino acid residues are used in database search. In general, probability-based molecular-weight search (MOWSE) scores corresponding to a signiicance threshold of P < 0.05 are considered for peptide identiication.

3. Scaffold (Proteome Software) is used to validate the MS/MS-based peptide and protein identiications. Peptide identiications are accepted if they could be established at greater than 95% probability as speciied by the PeptideProphet algorithm [11]. Protein identiications, where protein probabilities are assigned by the ProteinProphet algorithm [12], are accepted if they can be established at greater than 99.0% probability and contain at least two identiied peptides. 4. Lower peptide probability scores are considered upon manually verifying the peptide assignments made by the database search programs. Manual inspection of MS/MS and MS3 spectra of each HNE-modiied peptide is performed to discard false positives and accept false negatives, if any. A computer-assisted inspection of MS/MS spectra that searches for HNE-histidine immonium ions (m/z 266), NL product ions, and proline-driven fragmentation, can help in discriminating modiied peptides from unmodiied forms. Note: The modiied peptides identiied in the eluate fraction can be considered correct matches when peptides are derived from proteins in which additional, unmodiied peptides are also matched in the supernatant fraction. The presence of a protein based on identiication of a single modiied peptide in the eluate is supported by the detection of other peptides of the protein in the low-through fraction. Also, analysis of noncarbonylated peptides in low-through fractions should be performed to proile proteins in the sample in order to conirm that changes occur in carbonylation stoichiometry rather than the protein abundance. Label-free approaches such as spectral counting or integration of XICs can be used for quantitation, as they have gained recent interest for abundancebased proiling and have beneited signiicantly from current advances in MS instrumentation regarding rapid data-dependent acquisition, ultrahigh sensitivity, and dynamic range [10]. 5. Prior to follow-up biological experiments, further validation of subsets of the identiied HNEmodiied peptides can be done by preparing synthetic peptides [7] and performing their MS/MS and, if applicable, MS3 analyses. The spectra of the HNE-modiied synthetic peptides should be identical to those acquired from the biological sample. As an example, Figure 2.1A shows an MS spectrum of [M + 2H]2+ ion at m/z 714.4 of HNE-modiied apomyoglobin tryptic peptide, LFTGH*PETLEK, where the asterisk indicates His modiication as a Michael adduct. Figure 2.1B shows the MS/MS spectrum of the

IDENTIFICATION OF HNE-MODIFIED PEPTIDES IN BIOLOGICAL SAMPLES 714.4

Relative Abundance

100

ESI-MS of LFTGH*PETLEK 714.9

50

715.4 0

710

[M+2H-HNE]2+

B Relative Abundance

100

[M+H-HNE]+

600

800 m/z

b2

0 200

y92+ y102+ b5 y4

400

600

y8

y9 b10

b 7 y7 800 m/z

Relative Abundance

y72+

1000

1200

1000

1400

ECD-MS/MS [M+2H]2+

D 100

y6 – H2O

Relative Abundance

400

CID-MS3 y6

50

CID-MS/MS

50

0 200 C 100

720

m/z

×5

×5

50

c8 z c7 7 z9

[M+H]+•

A

33

c9 c10

c6 1200 0 200

400

600

800 m/z

1000

1200

1400

FIGURE 2.1 (A) [M + 2H]2+ molecular ions of the HNE-modiied apomyoglobin tryptic peptide, LFTGH*PETLEK, where the asterisk indicates His modiication as a Michael adduct, at FTMS resolution. (B) CID-MS/MS spectrum from this doubly charged HNE-modiied peptide. (C) CID-MS3 of the doubly charged NL ion marked [M + 2H–HNE]2+ in the MS/MS spectrum in (B). (D) ECD-MS/MS spectrum of LFTGH*PETLEK.

same HNE-modiied peptide. The CID fragmentation of doubly charged HNE-modiied peptide produced a prominent signal of HNE NL at m/z 636.3 in the MS/ MS spectrum. No appreciable backbone fragmentation is observed, which prevents the direct localization of HNE adduct in the peptide. Implementation of the NL-driven CID-MS3 technique has allowed for an additional round of CID-based fragmentation on the NL ion (Figure 2.1C) and yielded enough fragment ions for identiication of the peptide

by database search. Although there is no direct evidence for the site of modiication in this MS3 spectra, the His residue may be regarded as the only plausible target site for HNE attachment in this apomyoglobin tryptic peptide according to the chemistry of HNE modiications in proteins [13] (Scheme 2.1) and also considering the trypsin-based bottom-up proteomics. Speciically, the peptide harbors no other candidate nucleophiles except His (H) and Lys (K). The observed cleavage by trypsin at Lys during digestion, however, precludes Lys

34

MASS SPECTROMETRY-BASED METHODS TO INVESTIGATE POSTTRANSLATIONAL PROTEIN MODIFICATIONS

(K) as the site of modiication, because its HNEmodiication in a protein is refractory to attack by this proteolytic enzyme. ECD-MS/MS of the peptide triggered during datadependent NL-driven ECD-MS/MS (Figure 2.1D) provides complementary fragmentation information and retains labile HNE modiications during fragmentation, thereby providing unambiguous identiication of sites of modiication. Because the fragmentation did not involve loss of HNE, a direct assignment of an HNE modiication site to His (H) has been possible from the collected ECD-MS/MS spectrum [9]. In order to proile proteins in the sample to conirm that changes occur in carbonylation stoichiometry rather than in the protein abundance, Scaffold allows for simultaneous comparison of multiple proteomic data sets in which the list of identiied proteins can be sorted by various parameters. The spectral counting technique for relative protein quantitation utilizes the total number of MS/MS spectra identiied for a particular protein as a measure of protein abundance, and consequently this parameter can be used to classify the abundance. Published protocols [14,15] can be used as methods for relative quantitation in which the change in abundance can be determined by the ratios as follows:

distribution with one degree of freedom, allowing Pvalue calculations for each ratio value. Ratios should be considered signiicant at P < 0.05. Also, consider conirmation of spectral counting by integration of corresponding peptide XICs. If antibodies are available for proteins that yielded differences in carbonylation by the bottom-up approach detailed in this chapter, then Western blotting is a truly orthogonal approach to conirm that changes occurred in carbonylation stoichiometry rather than in the protein abundance [17]. Relative quantitation based on XICs can be used to validate spectral counting results for selected peptides and proteins, and employs a similar routine as the spectral counting method. However, rather than total number of identiied spectra as a measure of protein abundance, the mass spectrometric response value (based on ion current) for individual peptides from a given protein is utilized. In our laboratory, XICs of peptides is generated automatically via the Sieve® program (Thermo) using its default parameters for chromatographic alignment and data framing typical for data acquired on an LTQ-FT.

Ratio = [( ηtreated )/( ηcontrol )],

For testing and quality control of the SPH-based enrichment and verifying performance of data-dependent liquid chromatography–electrospray ionization tandem mass spectrometry (LC–ESI-MS/MS), prepare HNEmodiied apomyoglobin by incubating the protein (1 mg/mL) with 2 mM HNE in PBS at 37°C for 2 h. Then, precipitate the modiied protein by adding four volumes of ice-cold acetone and keeping the mixture at −20°C for 2 h. Centrifuge the sample at 13,000 rpm for 10 min at 4°C. The protein pellet should be resuspended in 50 mM NH4HCO3 buffer and subjected to proteolytic digestion according to the procedure described above in the subsection entitled “Trypsin Digestion.”

where ηtreated and ηcontol are the total number of identiied MS/MS spectra (normalized spectral counts from Scaffold) for a particular protein in the treated and control group, respectively. A 50% peptide probability can be used instead of the initial list of high-conidence identiications (99% protein conidence, 95% peptide conidence, and containing 2 unique peptides) in order to include peptides with lower Mascot scores that represent true positive identiications and would improve the overall spectral counting sensitivity. Ratios for proteins showing zero spectral counts in either control or treated groups should be tabulated as UC (unique in control) or UT (unique in treated). Utilize a G statistic test (likelihood ratio test for independence) to determine the statistical signiicance for each protein ratio [14,16]: G = 2[C control ln(C control /t ct ) + C treated ln(C treated /t ct )], where G is the G-test statistic, Ccontrol is (ηcontrol + 1), Ctreated is (ηtreated + 1), and tct is (Ccontrol + Ctreated)/2. The G-statistic value is approximately characterized by a χ2

2.3

2.4 2.4.1

REAGENTS AND STANDARDS

COMMENTARY Background Information

Oxidative stress occurs when the body’s antioxidant defense mechanism succumbs to the damaging effects of the reactive oxygen species (ROS). ROS are produced by a variety of pathways during normal metabolism in an aerobic environment and are also generated during mitochondrial oxidative phosphorylation pathway [18]. Superoxide anion (O2•−), produced as a result of the one-electron reduction system by electron transport

COMMENTARY

chain (ETC) in the normal process of mitochondrial respiration, is the precursor of mitochondrial hydrogen peroxide (H2O2). O2•− is converted to H2O2 by the action of mitochondrial manganese superoxide dismutase (MnSOD, SOD2). Eventually, the conversion of H2O2 into highly reactive hydroxyl radicals (•OH) in the presence of reduced transition metal ions, such as ferrous ion (Fe2+), by the Fenton reaction can cause lipid peroxidation [18]. The membrane lipids—mainly phospholipids— containing PUFAs are predominantly susceptible to free radical-initiated peroxidation [19]. The breakdown of polyunsaturated acyl chains through nonenzymatic Hock cleavage results in the formation of a great diversity of reactive aldehyde products collectively termed RCS and includes malondialdehyde (MDA), HNE, 4-hydroxy-2hexenal (HHE), acrolein, and others [19]. The RCS can cause carbonylation of proteins, resulting in severe damage to their biological functions, which can eventually lead to the cell demise. Of the lipoxidation-derived RCS, HNE is regarded as the main and highly toxic aldehyde formed during peroxidation of n-6 PUFAs, such as linoleic acid (C18:2, n-6) and arachidonic acid (C20:4, n-6) [13,19]. HNE, a bifunctional aldehyde, is susceptible to nucleophilic addition at both the double bond (C3) and the carbonyl moiety (C1) in the molecule. The presence of electron withdrawing carbonyl group at the C1 position in HNE makes the double bond at the C3 position highly electrophilic and, hence, can covalently bind to proteins via Michael addition (MA) to the sulfhydryl group of Cys, imidazole group of His, and ε-amino group of Lys residues, resulting in an increase of molecular mass by 156 Da, as shown in Scheme 2.1 [13]. HNE also forms SB with its carbonyl moiety (C1) and ε-NH2 group of Lys in proteins with a concomitant loss of water, resulting in a mass increase of 138 Da. However, corresponding kinetics are inherently slow, and the SB product is reversible. In the studies involving oxidative stress, HNE-induced protein carbonylation has received the most attention. As opposed to the free radicals, HNE can diffuse across membranes easily and can attack targets distant from the sites of its production. Thus, this aldehyde acts as a toxic second messenger of oxidative stress by disseminating and augmenting initial free radical events. The primary mechanism of HNE-induced toxicity is its covalent modiication of proteins that can diminish the protein function by alterations in its conformations and also endanger the integrity of enzyme complexes [20,21]. In addition, the adducted proteins act as neoantigens and induce a speciic immune response promoting tissue inlammation. HNE has been implicated in a number of pathologies such as atherosclerosis, cancer, neurodegenerative disorders (Alzheimer’s disease, Parkinson’s disease, and amyo-

35

trophic lateral sclerosis), ethanol-associated liver injury, and injury from ischemia/reperfusion [21–23]. Despite the fact that protein carbonylation has deleterious effects on the cell, carbonylated proteins remain at very low physiological concentrations and are present in low substoichiometric quantities in a tissue—often described as a “needle in a haystack” situation pertaining to analytical measurements. Mapping of HNE modiication sites within complex mixtures of proteins is made more challenging in bottom-up proteomics approach where samples are made signiicantly more complex by digesting proteins into multiple peptides and may test the limits of even the most sensitive MS systems. For modiication mapping, the modiied peptides must be present in the sample and observable in the mass spectrum to enable the localization of HNE modiication, whereas any set of peptides can be suficient for identiication purposes as long as criteria for given statistical signiicance are fulilled. The presence of high-abundance unmodiied peptides in samples of biological origin makes it necessary to use an eficient enrichment strategy to selectively prefractionate carbonylated peptides by substantially reducing the bulk of unmodiied species that can suppress ionization. The identiication of oxidized proteins with carbonyl groups provides very important information regarding the class of proteins that are susceptible to damage during oxidative stress. Even more important is identiication of amino acid residues within a protein sequence that undergo modiication, because localizing the site of modiication helps in understanding how and why oxidative stress generally causes loss of protein function in living tissues. MS-based identiication of HNE modiication sites in different proteins in a highly complex biological sample requires a proper and eficient front-end technique for selective enrichment of modiied peptides. Several strategies for the enrichment of HNE-modiied proteins and peptides have been developed in order to circumvent the technical challenge presented by the low stoichiometry of HNE modiication and purify the “subproteome” carrying the HNE-adducts from native unmodiied forms. The carbonylated proteins can be enriched by employing an avidin afinity capture strategy upon labeling with biotin hydrazide [24]. The enriched proteins are fractionated by reversed-phase chromatography followed by tryptic digestion and analysis by MS. Tryptic digestion of enriched carbonylated proteins affords modiied as well as unmodiied peptides, thereby complicating analyses. Alternatively, the derivatization of the carbonyl group found in these modiied proteins with a biotinylated hydroxylamine derivative, N′-amino oxymethylcarbonylhydrazino D-biotin, to form biotinmodiied oxylipid peptide conjugates that are amenable

36

MASS SPECTROMETRY-BASED METHODS TO INVESTIGATE POSTTRANSLATIONAL PROTEIN MODIFICATIONS

to enrichment using avidin afinity capture, has been developed [25]. However, this method features the use of a covalently bound tag to HNE that makes the conjugate bulkier, and hence, the MS/MS spectrum is complicated by side-chain fragment ions from the biotin-labeled aldehyde reactive agent. The presence of nonpeptide ions in the MS/MS spectrum can adversely affect the identity scores of the database search algorithms by providing scores that are often lower than those considered to be acceptable for conclusive identiication of a peptide. Another strategy reported to enrich HNE-modiied peptides is to derivatize the carbonylated proteins with Girard P reagent (GRP; 1-(2-hydrazino-2-oxoethyl)pyridinium chloride), followed by proteolysis and enrichment of the derivatized peptides using strong cation exchange (SCX) chromatography [26]. The hydrazide group in the GRP enables its reaction with carbonyls, whereas the quaternary amine adds positive charge to the carbonylated proteins and peptides, permitting the enrichment of labeled peptides by SCX chromatography at neutral pH. The limitation of the GRP-based enrichment techniques is that it enriches not only GRP-tagged HNE-modiied peptides but also basic peptides. An elegant and eficient label-free method of enrichment of HNE-modiied peptides by SPH chemistry utilizes reversible hydrazide chemistry to selectively capture and then release HNE-modiied peptides from within complex mixtures [5], as shown in Scheme 2.2. In this strategy, HNE-modiied peptides are captured on hydrazide-coated glass beads via the formation of covalent hydrazone linkage between the hydrazide group immobilized on the solid support and the aldehyde group in the modiied peptides facilitating their selective enrichment. The hydrazone bond is acid labile and, hence, the modiied peptides can be easily recovered from the SPH beads under acidic conditions. The use of SPH strategy to enrich carbonylated species prior to mass spectrometric analysis is advantageous compared with direct immunoafinity approaches. While the SPH method has the potential to enrich carbonylated peptides modiied by a wide array of lipid peroxidation end products such as MDA, HNE, HHE, and acrolein, for example, an immunoafinity approach requires separate antibodies against each kind of carbonyl adducts. Derivatization of peptide carbonyls with, for example, 2,4-dinitrophenylhydrazine (DNPH), and then, performing immunoafinity-based enrichment with immobilized antidinitrophenyl antibodies, may preclude the need for separate antibodies [27]. However, these derivatives may form at low yields and, in addition, according to our experience, they tend to detach the DNPH tags with each processing step because of their instability.

HNE-modiied peptides can be characterized by MS/ MS using CID. However, the carbonylated peptides may lose the HNE group readily upon CID, leaving the peptide intact or generating very few product ions rather than fragmenting along the peptide backbone to create y- and b-ion series that can be used to identify the peptide. This loss creates dificulty in identiication of peptide sequence and localization of site(s) of HNE modiication. MS3 analysis is then required to create ions that can be used for the peptide identiication. In NL-driven CID-MS3, if a loss of 78 or 52 Th (from a doubly or triply charged HNE-modiied peptide) is observed from the precursor ion, then the NL fragment ion is selected and subjected to CID fragmentation (i.e., MS3). The application of NL-driven CID-MS3 approach (Scheme 2.3) enhances the number of HNEpeptide adducts identiied after protein sequence database searching [6]. However, MS3 of the NL ion provides no diagnostic mass tag and, hence, may introduce ambiguity associated with the precise position of the modiication site when more than one potential acceptor site exists in a peptide. This dificulty of correctly assigning the site of HNE modiication when more than one possible candidate site is present within the peptide is due to the lack of known consensus sequence representing the chemical selectivity of HNE modiication. In such cases, the use of the NL-driven ECD-MS/MS (Scheme 2.4) method is a beneicial approach to characterize the HNE modiications. ECD produces extensive fragmentation of the peptide backbone, generating nearly complete product ions series of c- and z-ions and provides good amino acid sequence coverage of the fragmented peptide [9]. The peptide fragmentation is less inluenced by the peptide sequence. Moreover, ECD is a very gentle dissociation method that allows labile modiications to remain attached during backbone fragmentation, enabling their detailed analysis. ECD-MS/MS of HNE-modiied peptides gives conclusive data concerning speciic amino acids to which HNE is adducted. The implementation of ECD in unambiguous characterization of HNE-modiied peptides/proteins has been shown [7,9]. Also, since ECD will be triggered only on precursor ions that show an HNE loss of 78 or 52 Th in the CID spectra, depending on the charge of the peptide (2+ or 3+), the duty cycle is reduced compared with full data-dependent ECD analysis. Label-free quantitation is a rapid and fairly accurate approach for the survey of change in protein expression levels between two biological samples and is a promising alternative to stable isotope labeling approaches. It is a highly reproducible method that enables relative quantitation of the proteins across samples [14,15]. In a label-free quantitative approach, each sample has to be analyzed individually and sequentially by MS (unlike

COMMENTARY

labeling techniques where samples are pooled together and analyzed). Quantitative comparisons between proteins in samples are performed by two different ways [10]. The irst is based on peptide ions belonging to a speciic protein: measuring and comparing changes in areas or heights of chromatographic peaks reconstructed (“extracted”) from successively recorded fullscan mass spectra during the analysis. The second is based on the spectral counting: recording and comparing the number of MS/MS spectra identifying peptides of a given protein to assess relative protein abundance. The spectral counting-based quantitation approach is more reproducible and has wider dynamic range compared with chromatographic peak intensity measurements-based quantitation [28–30]. Low quality MS/MS spectra and correctly identiied peptides that map to more than one protein might, however, create problems for peptide quantitation. Nevertheless, several statistically rigorous bioinformatics tools (open-source, commercial, and custom software packages) have been developed for label-free differential quantitation of proteins. These tools perform data normalization, time alignment, peak detection, peak quantiication, peak matching, identiication, and statistical analysis. We anticipate that the procedure reported in this chapter could be widely used to enrich and detect HNEmodiied peptides from protein samples from diverse biological origins. This technology possibly will contribute to further understanding of cellular proteins targets that are susceptible to HNE modiication during oxidative stress.

2.4.2

Deinitions

Cell Lysate An extract containing cellular contents obtained after rupturing cells by various means such as freeze–thaw, sonication, enzymatic, or osmotic mechanisms. Extracted-Ion Chromatogram A chromatographic trace generated by considering the measured intensity of the signal for an ion with a chosen m/z value. The area of a chromatographic peak of an ion increases with its increased concentration in a sample and can be used for relative quantitation. HNE-Modiied Protein This arises by the covalent binding of HNE, an endogenously generated major lipid peroxidation product derived from breakdown of lipids containing ω-6 fatty acids (linoleic and arachidonic acid), to nucleophilic sites in the given protein. HNE reacts with the side chains of cysteine (Cys, C), histidine (His, H), and lysine (Lys, K) in an order of reactivity: Cys >> His > Lys.

37

The covalent addition of HNE to proteins is frequently used as a marker of oxidative stress. Neutral Loss-Driven MS3 This triggers additional fragmentation of the dominant and speciied NLassociated peak observed in the MS/MS scan. It results in the generation of more fragment ions for accurate peptide sequence determination. MS3 spectra are helpful when peptide backbone cleavage by MS/MS is ineficient. Neutral Loss-Driven ECD Represents a method when ECD in the FTICR unit of the instrument is initiated on precursor ions of peptides showing the speciied NL in the MS/MS scan acquired via collision-induced dissociation–tandem mass spectrometry (CID-MS/MS) in the linear ion trap. ECD helps in correct assignment of sites of modiication when more than one candidate site is present, because this method of peptide dissociation tends to retain the modifying group during MS/MS of the precursor ion. Solid-Phase Hydrazide (SPH) Reagent Constituted by hydrazide-derivatized, spherical glass beads useful for the enrichment of peptides carbonylated, e.g., by HNE from complex mixtures. The peptide carbonyls are captured on the SPH reagent under mildly acidic conditions (pH 3.6) via the formation of a covalent but reversible hydrazone linkage. Such coupling method (“chemoprecipitation”) is a powerful way to immobilize modiied peptides. After washing to remove unmodiied species, bound carbonylated peptides are subsequently released from the beads in their original forms under strong acidic conditions. Spectral Count The total number of MS/MS spectra (usually corresponding to redundant and nonredundant peptides) used for identiication of proteins. Spectral count increases with protein abundance, hence spectral counting is a useful approach in assessing relative protein abundance by a label-free quantitative strategy. 2.4.3

Safety

It must be noted that one must review all relevant Material Safety Data Sheets (MSDSs) prior to the analysis and take appropriate safety precautions. MSDSs are available from reagent manufacturers and on the World Wide Web. HNE is a colorless liquid soluble in most organic solvents (e.g., alcohols, hexane, and chloroform). HNE is a cytotoxic and mutagenic lipid oxidation product, and, therefore, this aldehyde should be used with caution and handled appropriately to minimize exposure.

38

2.4.4

MASS SPECTROMETRY-BASED METHODS TO INVESTIGATE POSTTRANSLATIONAL PROTEIN MODIFICATIONS

Critical Parameters and Troubleshooting

2.4.4.1 Sample Preparation The amount of protein needed for analysis of PTM is much higher than that required for the protein identiication, because of the necessity of maximum coverage of the sequence for comprehensive localization of modiications. In the case where very few total peptides were detected, starting with higher number of cells is usually recommended. Large amounts of protein are usually obtained from cell cultures, but it is also feasible to obtain protein extracts from primary tissues. Failure to detect any peptides upon MS analysis may be due to the degradation of proteins prior to tryptic digestion or may be due to problems in the digestion step. To avoid these troubles, make sure protease inhibitors are added before lysis, and also check the integrity/reactivity of trypsin solutions. The pH of solution during digestion should be from 7.5 to 8.5, and urea concentration should be kept below 2 M. The isolation of organelles and subcellular fractionations is a useful method of sample preparation to reduce sample complexity and increase the probability of identifying posttranslational protein modiications by a bottom-up approach. 2.4.4.2 LC–ESI-MS/MS With nano-LC column, critical parameters affecting separations are sample preparation, solvent purity, low rate, void volume associated with tubing (fused-silica or PEEKsil [SGE Analytical Science, Austin, TX]) i.d. and connections, column aging, and column contamination. Follow the manufacturer’s guide to troubleshoot the equipment should you experience loss of performance. An excellent article by Noga et al. [31] is also recommended as a guide for troubleshooting nano-LC. There are several possible ways by which the eficiency of a nano-LC column can be compromised. Most common is the presence of submicroscopic solid material in the biological sample that is capable of clogging the nano-LC column and rendering it unusable as it cannot be back-lushed and decontaminated. Operating the column at high low rate and high pressure can also ruin column packing. Nano-LC columns can easily be “overloaded” with samples, and, as a consequence, carryover can be seen in subsequent runs. Unmodiied peptides as possible carryover may compromise the analysis of HNEmodiied peptides, and may also prompt questions about the eficiency of the enrichment protocol employed. When carryover is suspected, a few “blank” runs are recommended to wash away contaminants before injecting the next sample. The frequently encountered problem in operating nano-LC system is leakages that need careful inspection

of all connections and diagnosis of the reason for the leakage before concluding the occurrence of clogging of the sample trap, nano-LC column, or Picotip emitter, prompting their replacement. Check system pressure with and without column because many pressure-related problems may be due to blockages elsewhere in the system. Leaks at connectors preceding the nano-LC column cause signiicant drop in column pressure and no signal will be observed in the mass spectrometer. Leaks between the column outlet and Picotip emitter are often caused by clogging of the Picotip. Also check for partial leakages for the loss of sensitivity in the system. For the complete or partial loss of signal in the mass spectrometer, check the position and integrity of the Picotip as well. Make sure the tip is not broken or its coating is not damaged; replace it if required. The presence of air bubbles in the mobile phases might also compromise the spray from the Picotip emitter and degassing of mobile phases prior to use is recommended. Note: Use care while handling the tip as the sharp end can pierce the skin. Be aware of the high temperatures of certain parts after the mass spectrometer is shut down or turned off. Also, whenever the instrument is not in standby, hazardous voltages are present on one or more interior parts. 2.4.4.3 Data Analysis Trypsin and/or endoproteinase Lys-C are typically used for protein digestion in order to achieve peptides in the mass range 0.5–4 kDa that are suitable for MS and MS/MS analysis. Studies have shown that combinations of other sequence-speciic or nonspeciic proteases can be used to obtain further amino acid sequence information from proteins. For database searches, more than one differential mass modiication can be included in the search algorithms. However, this increases the search space, resulting in increase in the number of false assignments that may occur by random chance. Incorporation of enrichment step to selectively isolate one class of posttranslationally modiied peptides/proteins enables focusing on one or few types of modiication on all proteins present in a sample. Moreover, high mass accuracy measurements increase the conidence in peptide identiication and identiication of modiication sites. Also, the use of more than one database search programs increases the conidence in analysis of PTMs and is highly recommended. SEQUEST (http://ields.scripps.edu/sequest/) evaluates similarity between the observed MS/MS spectrum and a theoretical spectrum, while Mascot uses a statistical approach to evaluate the probability of observing MS/MS fragment ions. Other search algorithms include OMSAA (http://pubchem.ncbi.nlm. nih.gov/omssa/) and X!Tandem.

REFERENCES

It should be noted that protein carbonylations occur via many different pathways. The hydrazide-coated glass beads are capable of enriching peptides harboring carbonyl modiications other than those by HNE (e.g., 4-hydroxy-2-hexenal, 4-oxo-2-nonenal, metal-catalyzed oxidation, and others). These modiications can be included in the database search algorithms for in-depth mining of posttranslational protein carbonylations.

2.4.5

Method Performance

The performance of this method has been veriied in our laboratory and has been reported [7,9]. The approach for analysis of HNE modiication in peptides/proteins consisted of three steps: (1) selective isolation of HNEmodiied peptides from a complex peptide mixture via chemical method of capture into hydrazidefunctionalized glass bead supports; (2) analysis of the enriched modiied peptides by various LC–MS/MS approaches, including data-dependent acquisition and multistage MS; and (3) identiication of the modiied protein and localization of the modiied residue(s) by correlation of MS/MS and MS3 data with sequence databases. We have observed that the selective enrichment of HNE-modiied peptides prior to mass spectrometric analysis has the added advantage of producing enhanced data quality of MS3 spectrum by increasing the number of ion counts for fragmentation. Furthermore, the enrichment technique has also reduced the acquisition of false MS3 spectra triggered due to isobaric fragment ions produced during CID. We have concluded that SPH enrichment combined with both CID and ECD method of peptide dissociation are advantageous for an in-depth interrogation and unequivocal localization of HNE-induced carbonylation to proteins that occurs via MA (Scheme 2.1).

2.4.6 Time Considerations The time required for preparation of cultured cell lysate is about 60 min and measurement of protein concentration using BCA assay is 40 min. Protein reduction and alkylation take about 75 min; protein digestion is conducted overnight (12–18 h). Peptide desalting and lyophilization need 30 min and 2–4 h (depending on the volume after elution), respectively. Enrichment of HNE-modiied peptides using the SPH chemistry requires overnight incubation of the peptide mixtures with surface-modiied glass beads, about 60 min each (i.e., 2 h combined) for washing and subsequent elution of carbonylated peptides, followed by 3–4 h for lyophilization of the eluate. LC–MS/MS analysis requires about 120 min per run.

39

ACKNOWLEDGMENTS This project was supported by a grant (AG025384) from the National Institutes of Health. Laszlo Prokai is the Robert A. Welch Chair in Biochemistry at the University of North Texas Health Science Center at Fort Worth (endowment number BK-0031).

REFERENCES 1. Farley, A.R., Link, A.J. (2009) Identiication and quantiication of protein posttranslational modiications. Methods in Enzymology, 463, 725–763. 2. Stensballe, A., Jensen, O.N., Olsen, J.V., Haselmann, K.F., Zubarev, R.A. (2000) Electron capture dissociation of singly and multiply phosphorylated peptides. Rapid Communications in Mass Spectrometry, 14, 1793–1800. 3. Bruenner, B.A., Jones, A.D., German, J.B. (1995) Direct characterization of protein adducts of the lipid peroxidation product 4-hydroxy-2-nonenal using electrospray mass spectrometry. Chemical Research in Toxicology, 8, 552–559. 4. Uchida, K., Stadtman, E.R. (1992) Modiication of histidine residues in proteins by reaction with 4hydroxynonenal. Proceedings of the National Academy of Sciences of the United States of America, 89, 4544–4548. 5. Roe, M.R., Xie, H., Bandhakavi, S., Grifin, T.J. (2007) Proteomic mapping of 4-hydroxynonenal protein modiication sites by solid-phase hydrazide chemistry and mass spectrometry. Analytical Chemistry, 79, 3747– 3756. 6. Stevens, S.M., Rauniyar, N., Prokai, L. (2007) Characterization of 4-hydroxy-2-nonenal-modiied peptides by liquid chromatography-tandem mass spectrometry using data-dependent acquisition: neutral loss-driven MS3 versus neutral loss-driven electron capture dissociation. Journal of Mass Spectrometry, 42, 1599–1605. 7. Rauniyar, N., Stevens, S.M., Prokai-Tatrai, K., Prokai, L. (2009) Characterization of 4-hydroxy-2-nonenal-modiied peptides by liquid chromatography-tandem mass spectrometry using data-dependent acquisition: neutral lossdriven MS3 versus neutral loss-driven electron capture dissociation. Analytical Chemistry, 81, 782–789. 8. Rauniyar, N., Prokai, L. (2009) Detection and identiication of 4-hydroxy-2-nonenal Schiff-base adducts along with products of Michael addition using data-dependent neutral loss-driven MS3 acquisition: method evaluation through an in vitro study on cytochrome c oxidase modiications. Proteomics, 9, 5188–5193. 9. Rauniyar, N., Prokai-Tatrai, K., Prokai, L. (2010) Identiication of carbonylation sites in apomyoglobin after exposure to 4-hydroxy-2-nonenal by solid-phase enrichment and liquid chromatography-electrospray ionization tandem mass spectrometry. Journal of Mass Spectrometry, 45, 398–410.

40

MASS SPECTROMETRY-BASED METHODS TO INVESTIGATE POSTTRANSLATIONAL PROTEIN MODIFICATIONS

10. Prokai, L., Stevens, S.M. Jr., Rauniyar, N., Nguyen, V. (2009) Rapid label-free identiication of estrogen-induced differential protein expression in vivo from mouse brain and uterine tissue. Journal of Proteome Research, 8, 3862–3871. 11. Keller, A., Nesvizhskii, A.I., Kolker, E., Aebersold, R. (2002) Empirical statistical model to estimate the accuracy of peptide identiications made by MS/MS and database search. Analytical Chemistry, 74, 5383–5392. 12. Nesvizhskii, A.I., Keller, A., Kolker, E., Aebersold, R. (2003) A statistical model for identifying proteins by tandem mass spectrometry. Analytical Chemistry, 75, 4646–4658. 13. Esterbauer, H., Schaur, R.J., Zollner, H. (1991) Chemistry and biochemistry of 4-hydroxynonenal, malonaldehyde and related aldehydes. Free Radical Biology and Medicine, 11, 81–128. 14. Old, W.M., Meyer-Arendt, K., Aveline-Wolf, L., Pierce, K.G., Mendoza, A., Sevinsky, J.R., Resing, K.A., Ahn, N.G. (2005) Comparison of label-free methods for quantifying human proteins by shotgun proteomics. Molecular & Cellular Proteomics, 4, 1487–1502. 15. Higgs, R.E., Knierman, M.D., Gelfanova, V., Butler, J.P., Hale, J.E. (2005) Comprehensive label-free method for the relative quantiication of proteins from biological samples. Journal of Proteome Research, 4, 1442–1450. 16. Sokal, R.R., Rohlf, F.J. (1995) Biometry: The Principles and Practice of Statistics in Biological Research. New York: W. H. Freeman and Company, pp. 729–731. 17. Alegria-Schaffera, A., Lodgea, A., Vattem, K. (2009) Performing and optimizing western blots with an emphasis on chemiluminescent detection. In Methods in Enzymology, Vol. 463, edited by Burgess, R.R., Deutscher, M.P. Amsterdam, The Netherlands: Elsevier, pp. 573–599. 18. Kokoszka, J.E., Coskun, P., Esposito, L.A., Wallace, D.C. (2001) Increased mitochondrial oxidative stress in the Sod2 (+/–) mouse results in the age-related decline of mitochondrial function culminating in increased apoptosis. Proceedings of the National Academy of Sciences of the United States of America, 98, 2278–2283. 19. Catala, A. (2009) Lipid peroxidation of membrane phospholipids generates hydroxy-alkenals and oxidized phospholipids active in physiological and/or pathological conditions. Chemistry and Physics of Lipids, 157, 1–11. 20. Grimsrud, P.A., Xie, H., Grifin, T.J., Bernlohr, D.A. (2008) Oxidative stress and covalent modiication of protein with

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

bioactive aldehydes. Journal of Biological Chemistry, 283, 21837–21841. Carini, M., Aldini, G., Facino, R.M. (2004) Mass spectrometry for detection of 4-hydroxy-trans-2-nonenal (HNE) adducts with peptides and proteins. Mass Spectrometry Reviews, 23, 281–305. Uchida, K. (2003) 4-Hydroxy-2-nonenal: a product and mediator of oxidative stress. Progress in Lipid Research, 42, 318–343. Aldini, G., Dalle-Donne, I., Facino, R.M., Milzani, A., Carini, M. (2007) Intervention strategies to inhibit protein carbonylation by lipoxidation-derived reactive carbonyls. Medicinal Research Reviews, 27, 817–868. Mirzaei, H., Regnier, F. (2005) Afinity chromatographic selection of carbonylated proteins followed by identiication of oxidation sites using tandem mass spectrometry. Analytical Chemistry, 77, 2386–2392. Chavez, J., Wu, J., Han, B., Chung, W.G., Maier, C.S. (2006) New role for an old probe: afinity labeling of oxylipid protein conjugates by N’-aminooxymethylcarbonylhydraz ino d-biotin. Analytical Chemistry, 78, 6847–6854. Mirzaei, H., Regnier, F. (2006) Enrichment of carbonylated peptides using Girard P reagent and strong cation exchange chromatography. Analytical Chemistry, 78, 770–778. Fenaille, F., Tabet, J.C., Guy, P.A. (2002) Immunoafinity puriication and characterization of 4-hydroxy-2-nonenaland malondialdehyde-modiied peptides by electrospray ionization tandem mass spectrometry. Analytical Chemistry, 74, 6298–6304. Zybailov, B., Coleman, M.K., Florens, L., Washburn, M.P. (2005) Correlation of relative abundance ratios derived from peptide ion chromatograms and spectrum counting for quantitative proteomic analysis using stable isotope labeling. Analytical Chemistry, 77, 6218–6224. Wang, M., You, J., Bemis, K.G., Tegeler, T.J., Brown, D.P. (2008) Label-free mass spectrometry-based protein quantiication technologies in proteomic analysis. Brieings in Functional Genomics and Proteomics, 7, 329–339. Zhu, W., Smith, J.W., Huang, C.M. (2010) Mass spectrometry-based label-free quantitative proteomics. Journal of Biomedicine and Biotechnology, 2010, 840518. Noga, M., Sucharski, F., Suder, P., Silberring, J. (2007) A practical guide to nano-LC troubleshooting. Journal of Separation Science, 30, 2179–2189.

3 IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION Yuki Sugiura, Ikuko Yao, and Mitsutoshi Setou

3.1

INTRODUCTION

3.1.1 Principle and the History of Imaging Mass Spectrometry The development of mass spectrometry (MS) has recently entered into new paradigm. Classical matrixassisted laser desorption/ionization (MALDI) MS has been established as an analytical method for a wide range of molecules with varied physical and chemical characteristics, and now this continuous advancing technology has been utilized into molecular imaging technique (Figure 3.1). The innovation in MS has enabled the provision of additional two-dimensional (2D) axes of recognition on tissue sections as a new approach for the life science ield. Its unique advantages, which are summarized as follows, facilitate imaging mass spectrometry (IMS) as a versatile molecular imaging technique: (1) IMS does not require any speciic chemical labels or probes; (2) IMS is a “nontargeted” imaging method; and (3) the simultaneous imaging of many types of molecular species is possible. With the unique and powerful detection principle facilitated by MS, the matrix-assisted laser desorption/ionization–imaging mass spectrometry (MALDI-IMS) can be used for the visualization of the distribution of large number of biomolecules in the cells and tissues, ranging from small metabolite molecules [1,2] to much larger proteins [3,4]. Figure 3.2 illustrates the general worklow of MALDI-IMS. Basically, re-

searchers take thin tissue slices mounted on conductive glass slides and apply a suitable MALDI matrix to the tissue section, and then the slide is inserted into a mass spectrometer. A focused laser beam is directed at predetermined positions in the tissue slice and the mass spectrometer records the spatial distribution of molecular species (typically with 10–200 µm scan pitch). Automated data collection takes 2–6 h, depending on the number of points assayed. Suitable image processing software can be used to import data from the mass spectrometer to allow visualization of ion distribution images and comparison with the histological images of the sample. In the following sections, we will describe two major application areas; when using IMS as a technique for detection/visualization of various analytes, the application of IMS can fall roughly into two categories: 1. measurement of proteins and peptides, and 2. measurement of small organic compounds (massto-charge ratio [m/z] < 1000). To date, most of the reports concerning MALDI-IMS are with regards to the detection and imaging of proteins or peptides. On the other hand, the amount of research regarding the detection/imaging of small organic molecules has been growing recently. Figure 3.3 shows the result of a PubMed search using “Imaging Mass Spectrometry” as key words (except reviews). Reports were subdivided into the two groups, according

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

41

42

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION

2D

1D J. Plucker Discovery of CR Slit

3D

W.C. Rontgon Discovery of X-ray

A.M. Cormack, G. Hounsfield Development of CT Heart

Cathocle Ray

X-ray

J. I. Thomson Discovery of electron (particle aspect)

Electron Microscope

NMR

L. de Broglie Discovery of matter wave H. Busch Lens function by electro-magnetic field

Magnetic moment experiment by SternGeriach Zeeman splitting

E.M. Purcell, F. Bloch Discovery of NMR

Imaging Mass Spectrometry

Coronary Cross-section artery X-ray CT picture of human (gift from Shimadzu Co. Ltd) coronary artery (Hypertension vol. 49) Y. Fujiyoshi J. Heuser E. Ruska, M. Knoll Development of Quick-freezer deep-etch Electron microscope electron cryomicroscopy

J.J.Thomson First mass spectrometry

Mouse brain section:TEM (I.Yao, MITILS)

Freeze deep-etch of cytoskelton (JCB vol, 86)

R.R. Ernst Fourier transform NMR

K. Wuthrich 3D structure with NMR

TROSY spectrum (Nature vol. 418)

Insulin (PDB: 1his)

(Nature vol.423) P. Lauterbur P. Mansfield MRI

MRI picture of human upper body (BMC Medical imaging vol.6)

Development of imaging mass spectrometry (IMS)

3D?

K. Tanaka, J.B. Fenn MS of proteins with MALDI/ESI

Tryptic digested β-tubulin MA spectrum

*Won Nobel Prize in shaded area

Visulalization of ganglio side distribution in mouse brain (PloS One, vol.3. Setou lab)

FIGURE 3.1 Historical view of the development of analytical instruments. The progress of imaging technology was categorized by its technology (longitudinal axis) and its analytical dimensions (horizontal axis). Shaded researches have received Nobel Prizes. CR, cathode ray; NMR, nuclear magnetic resonance; CT, computed tomography; TEM, transmission electron microscopy.

to the materials analyzed in the each study, and the number of reports in each group was indicated. Notably, the number of reports regarding the IMS of small compounds gradually increased and occupied half of the published work in 2007. In the following section, we

shall introduce representative sample preparation strategies for analyses of distinctly structured analytes: proteins and small organic compounds, interweaving some research applications. Further, statistical analysis methods of MALDI-IMS are also described.

APPLICATIONS OF MALDI-IMS

43

Laser scanning

Tissue section

Ion images Histological image

Number of articles on IMS

FIGURE 3.2 Schematic representations of MALDI-IMS procedures. Usually, the tissue section mounted on an indium thin oxide (ITO)-coated glass slide is covered with a speciic MALDI matrix. Next, the ITO slide is inserted into a mass spectrometer. The MALDI laser scans through a set of preselected locations on the tissue (10–200 µm scan pitch) and the mass spectrometer records the spatial distribution of molecular species. Suitable image processing software can be used to import data from the mass spectrometer to allow visualization and comparison with the histological image of the sample.

30

Small metabolites

Large proteins m/z

20

10

0

2003

2004

2005

2006

2007 December 2008

Year Proteins/peptides

Small molecules

Others

FIGURE 3.3 The result of the PubMed search using “Imaging Mass Spectrometry” as keywords.

3.2 APPLICATIONS OF MALDI-IMS 3.2.1

IMS for Proteins and Peptides

3.2.1.1 Sample Preparation Procedure for Protein Imaging In traditional MS research, analyte molecules are basically extracted and separated from crude biological samples by gas chromatography (GC) or high

performance liquid chromatography (HPLC), mainly to avoid “ion suppression effect” [5–7]; in cases where such crude samples are subjected to MS, numerous molecular species compete for ionization and eventually, molecules that are easily ionized preferentially can be detected while suppressing the ionization of other molecules.

44

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION

Contrastingly, in MALDI-IMS, in which the tissue sample is directly measured, researchers should pay special attention to the fact that samples are extremely complex mixtures of biomolecules. Because tissues and cells are subjected to MALDI-IMS, the sample clean-up procedure is limited. Figure 3.4 shows IMS results of spotted peptide solution (0.5 µL of 100 nM adrenocorticotropic hormone [ACTH]) on both indium tin oxide (ITO) glass slide surface and brain section, and it clearly demonstrates the severe ion suppression effect on the tissue surface; the spotted ACTH peptide was detected only from the ITO surface, but not from tissue surface. As this example indicates, an important point to consider when executing an IMS experiment is the optimization of the sample’s condition, so that analyte objects can be eficiently ionized from crude mixtures. As a practical example, when a mouse brain section to which 2,5-dihydroxybenzoic acid (DHB) has been applied as a matrix is subjected directly to MS in positive ion-detection conditions, strong peaks which are mainly derived from phospholipids were observed in mass region of 700 < m/z < 900 [8]; on the other hand, signals derived from proteins, meanwhile, are scarcely detected at m/z > 3000 [9]. This is because phospholipids ionize much more eficiently than proteins, and they, on the other hand, suppress protein/peptide ionization. Therefore, for detecting/imaging proteins and peptides, removal of such lipids improves the sensitivity for proteins analysis. To this end, tissue sections should be rinsed with organic solvent, to remove lipids from tissue samples [9–11].

3.2.1.2 Rinsing of Tissue Sections: Removal of Unwanted Components As explained, the procedures of rinsing tissue sections or cells are key steps of the sample pretreatment (Figure 3.5). Such rinsing process enhances the sensitivity required for protein/peptide detection by eliminating small molecules, particularly phospholipids, from the tissues [9]. Lemaire et al. have Optical image Before matrix application

After matrix application

Ion image 100%

ITO surface

On tissue

0%

On tissue ITO surface 2440

2450

2460

2480

FIGURE 3.4 An example of the severe ion suppression effect on the tissue surface. A peptide solution (0.5 µL of 100 nM ACTH) was spotted on both ITO glass slide surface and brain-homogenate section. After the spraying of matrix solution (10 mg/mL α-cyano4-hydroxycinnamic acid [α-CHCA] 50% acetonitrile, 0.1% TFA), the spotted peptide was visualized by IMS. Small molecular analysis

Protein analysis

2470

Tissue extraction from animal

Tissue sectioning

Washing with organic solvent

Drying in vacume chamber

Matrix application

IMS measurement

FIGURE 3.5 A scheme for protein and small molecular measurement.

45

APPLICATIONS OF MALDI-IMS

examined the use of variable organic solvents in the rinsing step to enhance the eficiency of protein detection in tissue [9] (Table 3.1). They showed mass spectra obtained from mouse brain sections with or without rinsing using chloroform, acetone, or hexane, and found that the predominant signals in the vicinity of m/z 700 are abolished with rinsing procedure, and conversely, the number of peaks derived from proteins at a high molecular weight (i.e., 5000 < m/z) increased more than 40% after rinsing. We consider that this treatment also helps remove salts, a detrimental factor that otherwise interferes with the matrix–analyte cocrystallization process and thus degrades sensitivity, particularly for large proteins, while also complicating the spectrum assignment by producing both protonated and cationated molecular ions [10]. Figure 3.6 shows an example of protein imaging in the mouse brain section. With a proper sample preparation

TABLE 3.1 The Average Number of Detected Compounds from Rat Brain Treated with Various Organic Solvents [9] Treatment

n

Number of Detected Compounds ± Standard Deviation (%)

Increase in Detection (%)

Chloroform Hexane Toluene Xylene Acetone Untreated

10 5 5 5 5 10

81 ± 22 75 ± 28 68 ± 22 86 ± 13 64 ± 29 60 ± 34

34 25 13 44 7 0

The average number of detected compounds and calculated increase detection for peptides/proteins of m/z > 5000 is determined from the mass spectra recorded on untreated rat brain sections versus organic solvent-treated ones. n, number of experiments.

(A)

HE stained (B)

(C)

m/z 9919

Merged: A and B

(D)

(E)

m/z 14675

Merged: A and D m/z 9919 Choroid plexus

(F)

Overall

3000

4000

5000

6000

7000

8000

9000

10000

11000

12000

13000

m/z 14675 m/z

FIGURE 3.6 With proper sample preparation procedure, IMS facilitates simultaneous imaging of multiple proteins in a single tissue section. The igure shows the ion distribution images for mass peaks at m/z 9919 (B) and 14675 (D), and a microscopic observation of the same mouse brain section stained with HE after IMS measurement (C and E). The ion at m/z 14675 was localized in the white matter region, while m/z 9919 was found in choroid plexus (B and C). Averaged mass spectra from the choroid plexus and entire brain region (overall) were also presented (F).

46

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION

(A)

(B)

(C)

Aβ-(1−42) (E)

(D)

Aβ-(1−39)

Aβ-(1−40) (F)

Aβ-(1−38)

Aβ-(1−37)

FIGURE 3.7 Distribution of six amyloid β variants in the tissue prepared from a mouse model of Alzheimer’s disease. A. Optical image of the sagittal AD brain section. B–F. The molecular images of m/z 4515, 4330.9, 4231.7, 4132.6, and 4075.5 shows the distribution of Aβ-(1-42), Aβ-(1-40), Aβ-(1-39), Aβ-(1-38), and Aβ-(1-37), respectively [14].

procedure shown in Figure 3.5, MALDI-IMS provides simultaneous imaging of multiple proteins in a single tissue section at a time. We additionally note that researchers can histochemically stain the tissue sections which are measured by IMS after removal of the matrix and ixation procedure (in this case, immersed in methanol for 5 min). As can be seen in Figure 3.6, in conjunction with traditional light-microscopic histochemical observation, ion distribution images obtained by IMS provides valuable information regarding molecular distribution; this example demonstrates that ion at m/z 9919 is a speciic protein expressed by cells of choroid plexus, while m/z 14675 is localized in the white matter region of the brain. 3.2.1.3 Application of Protein Imaging to Disease Model Mice One of the advantages of traditional MS, including both MALDI- and ESI-MS, is that they can distinguish even slight structural variations of analyte molecules by their masses. Due to this unique advantage, MS has frequently been applied to the identiication/ characterization of posttranslational modiications in modern proteome researches [12]. In this context, IMS also enables the distinct detection of the protein molecular species as well as the visualization of the distribution of these species on the tissue sections. As an excellent example, in the Stoeckli et al. study, amyloid

β molecular species were visualized, which is generated by cleaving the amyloid precursor protein at a different cleavage site [13]. They revealed the distinct distribution of ive amyloid β variants in the brain sections prepared from a mouse model of Alzheimer’s disease [3,14] (Figure 3.7). From the analytical aspect, while traditional and wellestablished immunohistochemistry technique requires speciic antibodies which recognize each protein variant, and generation of such antibodies is a quite timeconsuming procedure, on the other hand, IMS could determine the distribution of protein molecular variants at once, even via a simple protocol. 3.2.1.4 On-Tissue Digestion Method Traditional MALDI is used to generate intact molecular ions of proteins up to m/z 100,000. In the current IMS technology, however, the upper mass limit of protein detection is approximately 40 kDa because the detection sensitivity severely falls at higher mass [15]. This is a considerable limitation which narrows the application capability of this technology. As another important problem, insoluble proteins to the matrix solution, such as membrane proteins, which is a protein molecule that is attached to the membranes, are dificult to be extracted into the applied matrix solutions and thus hardly crystallize with matrix, and they were in turn dificult to detect. On this

APPLICATIONS OF MALDI-IMS

regard, on-tissue digestion method in which proteins are denatured and enzymatically digested has been developed as an effective solution for these problems [16–18]. In this method, by cleaving large proteins, such proteins can be measured as digested proteins which are observed mainly in the mass range of 900 < m/z < 3000. The protein digestion process also makes it easier to perform tandem mass spectrometry (MS/MS), thus to identify the molecular species directly on the tissue section [16,18] (Figure 3.8). For all these reasons, on-tissue digestion is an attractive alternative method for detecting proteins that cannot be ionized by standard methods. We previously studied this method and found that this process was enhanced by a heat-denaturation process (80°C, for 10 h) and the use of a detergentsupplemented trypsin solution (200 mg/mL trypsin in 25 mM NH4HCO3 and 20 mM n-octylglucoside) (Figure 3.9; see detail protocol for Setou et al. [19]). 3.2.2

IMS for Small Organic Compounds

3.2.2.1 Employment of Optimized Experimental Protocols Is an Essential Issue In the ield of the molecular biology, localization of transcripts is visualized with oligonucleotide probe in situ hybridization, and localization of proteins is visualized using immunohistochemistry based on antibodies. The emergence and continuous development of IMS could add another standard imaging technique for metabolites, whereas we do not have an established visualization technology for them. In fact, until today, the small metabolites (i.e., m/z < 1000) have been intensively investigated by IMS and this research application can be further subdivided into two distinct areas: (1) measurement of endogenous small organic compounds and (2) measurement of exogenous compounds (such as administrated drugs). For both, the nature of MS-based detection principle facilitates the IMS as an effective imaging tool for these metabolites in which any chemical labels and probes are not required. Such uniqueness of IMS provides a capability for simultaneous visualization of multiple metabolites, enabling to follow molecular conversion of these small organic compounds (i.e., metabolism itself) between times or conditions (Table 3.2). However, each of such diverse molecular species could have quite different chemical/physical properties, and therefore, in practical, an optimization process of experimental protocol for each analyte is still an essential issue. As a representative example, Figure 3.10 demonstrates that concentration of the organic solvent (in this case, methanol) in the matrix solution affects the detection sensitivity of lipids and peptides. The results showed that a high composition of methanol (80%–100%) was favorable for lipid detection, while a low concentration solu-

47

tion (20%–40%) was favorable for the detection of peptides, indicating that lipids and peptides could be eficiently extracted from tissue sections into organic and nonorganic solvents, respectively. Figure 3.11 summarizes such key experimental points. As a irst point, we have to choose the appropriate ionization method; for the detection of small metabolites, we have alternative choices other than MALDI, such as secondary ion mass spectrometry (SIMS) [15], nanostructure-initiator mass spectrometry (NIMS) [20,21], desorption/ionization on silicon (DIOS) [22], nanoparticle-assisted laser desorptiopn/ ionization (nano-PALDI) [23], and even laser desorption/ionization (LDI) [24,25]. We consider that MALDI is still the most versatile method, particularly due to the soft ionization capability of intact analyte. However, other methods each have unique advantages; for example, SIMS and nano-PALDI have achieved higher spatial resolution than conventional MALDIIMS, and above all, these mentioned alternative methods are all matrix-free methods, and thus can exclude the interruption of the matrix cluster ion. Next, if MALDI is chosen, experimenters should choose a suitable matrix compound, solvent composition, and further matrix application method for their target analyte. All these factors are critical to obtain suficient sensitivity because they affect eficiency of analyte extraction, condition of cocrystallization, and, above all, analyte–ionization eficiency. In addition, based on the charge state of the analyte molecule, suitable MS polarity (i.e., positive/ negative ion detection mode) should be used in MS measurement. Below, we shall describe the key experimental points for MALDI-IMS applications of representative metabolites. 3.2.2.2 IMS of Endogenous Metabolites: Lipids Among the endogenous metabolites, MALDI-IMS for proiling [8,26–28] and visualizing distribution [1,29] of lipids is the best established application area. In the body, numerous lipid species play speciic functional roles, for example, energy storage, structural components of cell membranes, and important signaling molecule. Such lipid species may be roughly divided into three large categories: complex lipids, which contains phosphate and sugars in their structures (e.g., glycerophospholipids, GPLs); simple lipids, which are alcohol fatty acid esters (e.g., acylglycerols); and derived lipids, produced via hydrolysis of the simple/complex lipids (e.g., fatty acids). Table 3.3 shows the representative application studies of MALDI-IMS for lipids, and as can be seen, explorations of complex lipids have been the most intensively performed. This may be due to their easily charged structures, for example, phosphate group in phospholipids, sialic acids in gangliosides, and sulfate

(A) Apply the matrix solution by spraying... Microdispense the trypsin solution

Incubate at 37°C

or by microdispensing

Overview of on-tissue digestion method

b

80 60 40 20

1000

1200

1600

d m/z 1743.9

1800

0

40 20 0

400

600

800

1000

1200

1400 m/z

400

600

1039.6 y8 1126.6 y9 1240.7 y10

60

892.5 y7

AMGIMNSFVNDIFER

80

679.2 y5

20

x10

100 m/z 1743.9 Histone H2BF

564.0 y4

40

677.1 y5

578.0 y4

60

1460.8

TQDENPVVHFFK

80

873.5 y7 883.4 b8 987.6 y8 1020.5 b9 1116.6 y9 1167.5 b10 1231.6 y10 1314.6 b11 1332.7 b11+H2O

g 100 m/z 1460.8 MBP

2000 m/z

merged

e

f Relative intensity (%)

1400

1743.9

0

1541.9 y13

m/z 1460.8

c

100

* 1743.9

Relative intensity (%)

Optical Image

* 1460.8

a

Relative intensity (%)

(B)

800 1000 1200 1400 1600

m/z



FIGURE 3.8 Tryptic-digested protein imaging and precursor ion mass spectra with positive ion detection mode. A. A scheme showing overview of on-tissue digestion method. B. An example of IMS of tryptic-digested proteins. (a) Optical image of imaging region and (b) accumulated mass spectrum from imaging region. (c, d) Imaging results of m/z1460.8 and 1743.9, which are labeled by asterisks in (b). The merged image (m/z 1460.8 and m/z 1743.9) is shown in (e). These peaks were identiied by direct multistage tandem mass spectrometry (MSn) and were identiied as the fragment ions of (f) myelin basic protein (MBP) and (g) histone H2B [18].

(E) signal-to noise ratio

(A)

18 16 14 12 10 8 6 4 2 0

m/z 1198.7

(a) (b)

(c)

(d)

Relative Intensity

Denaturation process + Detergent-supplemented solution

(B) 1200 1400 1600 1800 2000 2200 2400 2600

Denaturation process only

(C) 1200 1400 1600 1800 2000 2200 2400 2600

Detergent-supplemented solution only

(D) 1200 1400 1600 1800 2000 2200 2400 2600

None m/z 1000 1200 1400 1600 1800 2000 2200 2400 2600

FIGURE 3.10 Concentration of organic solvent in the matrix solution inluences signal detection in the analysis of lipids and peptides. The brain homogenate slices were dispensed with 0.1 µL of DHB solutions containing different concentrations of the organic solvent (0%, 20%, 40%, 60%, 80% and 100% methanol containing 0.1% TFA), and the sensitivity of the signals corresponding to endogenous lipids and peptides was measured. The results showed that a high composition of methanol (80%–100%) was favorable for lipid detection, while a low concentration solution (20%–40%) was favorable for the detection of peptides. The crystal form of the analytes also changed with an increase in the methanol concentration in the matrix solution; needle-like crystals from which peptides were detected changed into aggregates of smaller crystals from which lipids were detected.

FIGURE 3.9 On-tissue digestion process was enhanced by the heat-denaturation process and the use of detergentsupplemented trypsin solution. A. Mass spectra obtained from 10-µm mouse cerebellum sections of trypsin-digested position (see inset). B–D. Mass spectrum obtained from the tissue sections prepared with indicated treatment of denaturation and detergent. Asterisks represent the mass peaks at m/z 1198.7. E. Signal-to-noise ratio of peak at m/z 1198.7 obtained from each tissue section. Bar, 1 µm [19].

TABLE 3.2

Potential Contribution of IMS for Imaging of Metabolites in Tissues or Cells

molecular diversity

Representative Molecular Imaging Method in Tissue or Cells

Probes

Selectivity

Allow Simultaneous Imaging

DNA

FISH (luorescence in situ hybridization)

Oligo nucleotide probe

Targeted

+

RNA

In situ hybridization

Targeted

+

Protein

Immunohistochemistry, green luorescent protein-fused protein Imaging mass spectrometry

Oligo nucleotide probe Antibody

Targeted

+

Metabolite (especially lipids)



Targeted/ nontargeted

+++

50

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION 1 Ionization technique

TABLE 3.3 Summary of Application Studies of MALDIIMS for Lipids Complex lipid

2 If MALDI, choice of matrix

3 Solvent composition of matrix solution

4

Choice of method for matrix application

5 MS polarity

FIGURE 3.11 Representative experimental points to consider in MALDI-IMS of small metabolites.

group in sulfatides. In particular, GPLs is the most frequent studied topic because their rich amounts in the tissues make the analysis easier [30]. However, for analysis of the noncharged and less-abundant molecular components, development of suitable sample preparation protocol to enhance their sensitivity is necessary. This is especially the case for lipid imaging, in which tissue-washing procedure with organic solvents is omitted because lipids are easily lost or migrate from their original location in tissues. When such a crude sample is subjected to MS analysis, numerous molecular species compete for ionization, and this can gives rise to severe ion suppression effects, degrading the sensitivity of such “dificult” lipids [5–7]. Hence, systematical experimental strategies ranging a wide variety of lipid species must be developed for successful lipid IMS. Figure 3.12 shows a proposed experimental strategy for MALDI-IMS of biological lipids. It is reported that proper choice of MS polarity (positive or negative) and use of suitable matrix solution is critical [31,32]. Addition of Salt to the Matrix Solution Enhanced the Sensitivity of Lyso-PC (Polar Lipid) Detection But Decreased the Sensitivity of TG (Nonpolar Lipid) Detection As demonstrated in Figure 3.10, composition of matrix solution is an important and useful “adjustable” parameter to increase the detection sensitivity of target analyte. In addition, if properly used, additive compounds to the matrix solution are another effective factor to achieve selective increase of target analyte signal. In this regard, it is reported that the presence/absence of alkali metal salts in the matrix solution affects the detection sensitivity for polar and nonpolar lipids [32]. Figure 3.13 shows that addition of the potassium acetate to the matrix solution enhanced the sensitivity of lyso-PC (polar lipid) detection, but on the

Simple lipid

Derived lipid

GPLs PCs (Astigarraga et al., 2008 [92]; Garrett et al., 2006 [1]; Jackson et al., 2005a,b [26,54]), PEs (Astigarraga et al., 2008 [92]; Jackson et al., 2005a, 2007b [26,58]), PIs (Astigarraga et al., 2008 [92]; Jackson et al., 2005a, 2007b [26,58]), PSs (Astigarraga et al., 2008 [92]; Jackson et al., 2005a, 2007b [26,58]), PGs (Jackson et al., 2005a, 2007b [26,58]), and Cardiolipins (Wang et al., 2007 [31]) Glycosphingolipids Gandliosides (Chen et al., 2008 [56]; Sugiura et al., 2008 [93]), Sulfatides (Ageta et al., 2008 [94]; Chen et al., 2008 [56]; Jackson et al., 2007b [58]), and Galactocyl-ceramide (Cha and Yeung, 2007 [95]; Taira et al., 2008 [23]) Neutral lipids Triacylglycerols (Astigarraga et al., 2008 [92]) and diacylglycerols (Astigarraga et al., 2008 [101]) Fatty acids (Zhang et al., 2007 [96]) Cholesterol (Altelaar et al., 2006 [72]; Jackson et al., 2005a [26])

PC, phosphatidylcholine; PE, phosphatidylethanolamine; PI, phosphatidylinositol; PS, phosphatidylserine; PG, prostaglandin.

other hand, decreased the sensitivity of triacylglycerol (TG) (nonpolar lipid) detection. The left panel of Figure 3.13 shows that the alkali metal salt in the matrix solution enhanced the sensitivity of phosphatidylcholine (PC) detection, presumably by the merging of [M+H]+ and [M+K]+ ion adduct form into [M+K]+, in the presence of potassium salt additive. Contrastingly, under the presence of a potassium salt, the intensity of TG as the [M+K]+ decreased to approximately half even at 4 mM (right panel). This can be accounted in that the enhancement of ionization for endogenous PCs signals the underexistence of the salt; the signal degradation of neutral lipids was attributed to the ion suppression effect of the enhanced PC ionization. Generation of Multiple Molecular Ion Adducts from a Single PC Molecular Species Was Suppressed by Adding an Alkali Metal Salt to the Matrix Solution The addition of alkali metal salts has another advantage: The formation of multiple molecular species with the same nominal mass can be avoided during the analysis of endogenous phospholipids. With the addition of alkali metal salts, the mass spectra of endogenous lipids (in the m/z range of 400–900) are simpliied to a

Tends to negatively

Tends to positively Charge state

Nonpolar lipids

Polar lipids Polar/nonpolar

Yes Multiple charge?

No Positive ion detection mode 10 mM K-acetate Phosphatidylcholine Sphingomyelin

Positive ion detection mode alkali-metal free Cholesterol Cholesterol ester Diacylclycerol Triacylglycerol

Negative ion detection mode alkali metal free

Negative ion detection mode mM Cs Cardiolipin

Phosphatidylinositol Phosphatidylethanolamine Phosphatidylserine Sulfatide Ganglioside

FIGURE 3.12 Suggested experimental procedure for various lipid molecules.

51

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION

Nonpolar lipids

Polar lipids

Triacylglycerol (54:2) [M+K]+

Relative Intensity

Lyso-PC (17:0)

Potassium acetete

No salt

+

+

[M

10 mM Potassium acetate

No salt

]

]

+K

+H

[M

m/z 548 [M+K]–

+

+

]

]

+K

[M

+H

[M

No salt

m/z 510 [M+H]+

Potassium acetete

0%

No salt

100% relative Intensity

52

M m tate 10 ace K

lt M e m at sa 4 cet No a K

m/z 917

Potassium acetete

100%

10 mM Potassium acetete

No salt

FIGURE 3.13 Addition of the potassium acetate to the matrix solution enhanced the sensitivity of lyso-PC (polar lipid) detection, but on the other hand, decreased the sensitivity of TG (nonpolar lipid) detection. We dispensed the solution of the reference compounds (0.1 µL) on thin sections of the tissue homogenate. We then sprayed the sections with different matrix solutions and measured the sensitivity of detection of each dispensed compound. We tested polar lipids such as lyso-PC and nonpolar lipids such as cholesterol and TG which tend to be charged positively. The results obtained for PC showed that the alkali metal salt in the matrix solution enhanced the sensitivity of PC detection; the most prominent signal derived from lyso-PC(17:0) was attributed to the [M+K]+ ion when using potassium acetate as the additive, while the generation of the [M+H]+ ion was reduced under the condition. These results can be accounted for the merging of ion adduct form into [M+K]+ in the presence of potassium salt additive. On the other hand, alkali metals salts added to the matrix solution during the detection of nonpolar lipids caused a decrease in the signal sensitivity.

considerable extent since multiple adduct ions formed are merged into a single alkali adduct ion. Figure 3.14 shows the representative mass spectra obtained from the rat kidney sections both in the presence/absence of 10 mM potassium acetate. Since PCs preferentially form cations in the form of alkali metal adducts [26,33,34] and tissues are rich in sodium and potassium salts, peaks due to [M+H]+, [M+Na]+, and [M+K]+ ions are detected in the mass spectra of endogenous PCs (Figure 3.14A). This is a critical problem that needs to be addressed because many molecular species are generated from PCs, and hence, a single peak in the spectrum may correspond to multiple ions. Table 3.4 summarizes the m/z values of abundant brain PCs in various ion forms, and it demonstrates that many molecular species share the same nominal mass (as indicated by the same fonts or styles). For example, the mass of a protonated PC(diacyl-16:0/20:4) molecule is identical to that of a sodiated PC(diacyl-16:0/18:1) ion (m/z 782), as can be seen in Figure 3.14A. In fact, MS/MS analysis of the

peak appearing at m/z 782 in the absence of the potassium salt shows that this peak actually corresponds to three different PC ions: [PC(diacyl-16:0/20:4)+H]+, [PC(diacyl-16:0/18:1)+Na]+, and [PC(1-alkyl-16:0/18:2) +K]+. This can be conirmed from the MS/MS spectra shown in Figure 3.14B(e–f); the peak at m/z 782 corresponds to three type of PCs which are a protonated PC ion (by a diagnostic peak at m/z 184 [a]), a sodiated PC ion (by m/z 147 [b]), and a potassiated PC ion (by m/z 163 [c]). On the other hand, upon addition of the potassium salt, peaks due to the aforementioned forms of the PC ions were separated into two distinct peaks at m/z 798 and 820, corresponding to [PC(diacyl-16:0/18:1)+K]+ and [PC(diacyl-16:0/20:4)+K]+, respectively (Figure 3.14A). In addition, MS/MS analysis of the spectra clearly shows that only potassiated PC molecules are present in the mass peaks (Figure 3.14B[g–h]). As can be seen from Table 3.4, by merging into [M+K]+ ion, most of such mass sharing of abundant PC species can be avoided except that of isobaric species.

(A)

[ +Na]+ [ +H]+

Without salt additives

[ +H]+ [ +K]+

[ +H]+ [ +K]+ [ +K]+

Relative Intensity

[ +H]+ [ +K]+ [ +Na]+

741.7 [ +Na]+

[ +Na]+

With salt additives

[ +K]+ [ +K]+[ +K]+

[ +K]+

741.7

740

760 c/t c/t

780

800

PC(diacyl-16:0/16:0) PC(diacyl-16:0/18:1)

820

840

m/z

PC(diacyl-16:0/20:4) PC(diacyl-16:0/18:2)

(B) (a) m/z 184

O

O P

O

(b) m/z 147

O

HO O

(c) m/z 163

+ O Na P

O

O

(d) m/z 86

+ O K

HO P

H2C

O

N

N+

+

184

(e)

m/z 760.5 [PC(diacyl-34:1)+H]+

760 59

86 184

(f)

147 163 86

m/z 782.5 [PC(diacyl-36:4)+H]+ [PC(diacyl-34:1)+Na]+ [PC(1-alyl-34:2)+K]+

59 124

723 782 599

(g)

739 59

163

798

m/z 798.5 [PC(diacyl-34:1)+K]+

86 761

(h) 163

59

637

86

100

124

m/z 820.5 [PC(diacyl-36:4)+K]+

200

300

400

500

600

820

700

800 m/z

FIGURE 3.14 Generation of multiple molecular ion adducts from a single PC molecular species was suppressed by adding an alkali metal salt to the matrix solution. A. Generation of multiple molecular ions from a single PC molecule was suppressed by adding an alkali metal salt to the matrix solution. Spectra obtained from sections of rat brain homogenate in the presence/absence of potassium acetate in the matrix solution are shown. With the addition of potassium acetate to the matrix solution, multiple molecular ion forms of PCs merged into a single potassiated PC ion. B. Structure of the diagnostic ions for (a) protonated PC at m/z 184; (b) sodiated PC at m/z 147; (c) potassiated PC at m/z 163. MS/MS spectra of ion peaks at (e) m/z 760 and (f) 782 in the absence of potassium acetate and spectra of ions peaks at (g) m/z 798 and (h) 820 in the presence of potassium acetate are also shown. c/t, corresponding to.

54

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION

TABLE 3.4 The m/z Values of Abundant Brain PCs as Various Ion Forms Molecular Species PC(diacyl 16:0–18:1) PC(diacyl 16:0–16:0) PC(diacyl 18:0–18:1) PC(diacyl 16:0–20:4) PC(diacyl 16:0–18:0) PC(diacyl 16:0–22:6) PC(diacyl 18:0–20:4) PC(diacyl 18:1–20:4) PC(diacyl 18:0–22:6) PC(diacyl 18:1–18:1) PC(diacyl 16:0–16:1) PC(diacyl 18:1–22:6) PC(diacyl 16:0–20:3) PC(diacyl 18:0–18:2) PC(diacyl 18:1–18:2)

PC

[M+H]+

[M+Na]+

[M+K]+

C34:1 C32:0 C36:1 C36:4 C34:0 C38:6 C38:4 C38:5 C40:6 C36:2 C32:1 C40:7 C36:3 C36:2 C36:3

760 734 788 782 762 806 810 808 834 786 732 832 784 786 784

782 756 810 804 784 828 832 830 856 808 754 854 806 808 806

798 772 826 820 800 844 848 846 872 824 770 870 822 824 822

For Negatively Charged Lipids Containing Multiple Negative Charted Structures, Addition of Alkali Metal Salt Enhanced Generation of Singly Charged Molecular Ions Alkali-metal salts dissolved in the matrix solution enhanced generation of singly charged ion molecules which contain the multiple negatively charged structures; during ionization, residual charged groups in their structures were neutralized by adduct formation with alkali metal cations, and as a result, singly charged ions could be eficiently generated. Since in MALDI process, much less multiple charged ions could be generated than that of ESI, the salt addition eventually can improve their sensitivity by promoting the eficiency of single-charged ion generation. In the study of Hay-Yan J. Wang et al., they added 100 mM of cesium iodide to a matrix solution (docosahexaenoic acid [DHA] 30 mg/mL in 50% ethanol) and successfully proiled multiple species of cardiolipin as single-charged molecular ions by neutralizing additional phosphate group (Figure 3.15). This treatment also beneits to integrate the ion adduct form to [M+Cs–2H]− [31]. That may be the case for gangliosides, which have multiple sialic acids in their sugar chain. Taken together, for the analysis of lipids which tend to have multiple charges, the addition of appropriate concentration of alkali metal salt is advantageous (Figure 3.12). MALDI-IMS of Phospholipids Revealed Cell-Selective Production of PC Molecular Species GPLs comprise a large molecular family in which phosphoric acid is esteriied to a glycerolipid. They are subdivided into distinct classes (e.g., PCs, phosphatidylethanolamines, and phosphatidylinositols) based on the structure of the head group linked to the phosphate, attached at the sn-3 posi-

tion of the glycerol backbone. They are further subdivided into numerous molecular species on the basis of the composition of the fatty acids linked to the sn-1 and sn-2 positions of the glycerol backbone [30]. Using IMS, we can image not only these multiple classes but also related molecular species simultaneously. In particular, the capability to determine the distinct localization of each molecular species, that is, to elucidate the distinct fatty acid composition of biological membranes in different tissue locations, is an important advantage of IMS [30] (Figure 3.16). In the brain, among the classes of GPLs, PCs are the most abundant structural component of neural and glial cell membranes, and the fatty acid constituents of PCs (i.e., molecular species) inluence the membrane’s physical properties, including luidity and curvature [35–38]. Since several types of fatty acids, especially polyunsaturated fatty acids (PUFA), in the PCs are released and converted in response to extracellular stimuli into bioactive lipids that mediate important biological processes [39], information on the distinct distributions of PUFAcontaining molecular species is quite valuable [40,41]. By applying high-magniication IMS to the cerebella cortex, a docosahexaenoic acid-containing phosphatidylcholine (DHA-PC), namely PC(diacyl-18:0/22:6), was found to be enriched in the Purkinje cell layer (Figure 3.17). Optical observation of successive hematoxylin and eosin (HE)-stained brain sections also suggested that PC(diacyl-18:0/22:6) was selectively detected in Purkinje cells (Figure 3.17, arrowheads) and in molecular layers (MLs) in which dendrites of Purkinje cells exist. In contrast, granule cells were impoverished in another DHA-PC, namely PC(diacyl-18:0/22:6). Interestingly, a complementary distribution of two other

40 1077.109

[M−H]–

20

0 1000

1240

1480

1720

1488.77

% Intensity

60

1471.80 1472.79

1449.87 1451.87 1450.85

1469.78

80

1487.83

1469.78

1485.78

100

1485.78

1447.89

1447.89

(A)

[M+K+−2H]–

[M+Na+−2H]–

1960

2200

1579.77

m/z (B) 100

[M+Cs+−2H]–

[M−H]– = (18:2)4 CL

1581.78 1450.89 1449.89

50

1420

1460

[M+K+−2H]–

1583.79

1485.83 1489.86

1472.87

1470.86

[M+Na+−2H]–

1452.92

% Intensity

1447.89

[M−H]–

1500

1540

1580

1620

m/z

FIGURE 3.15 Detection of cardiolipins as cesium (Cs)-adducted negative ions. A. MALDI-MS spectrum of (18:2)4 cardiolipin (CL) from rat heart section. The inset magniies the m/z region between 1420 and 1520. DHA matrix was used at 30 mg/mL. B. MALDI-MS spectrum of (18:2)4 CL from rat heart section in the presence of Cs ions. The relative abundance of sodiated and potassiated CL was signiicantly decreased by Cs addition [31].

56

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION PC(diacyl-16:0/16:0)

PC(diacyl-16:0/20:4) N+

N+

O

O P

O

O

O

O

C

O

CH3 O

H

C

H

CH3

CH3 O

P O

O CH3

O

O

O

O

O

[PC(diacyl-16:0/16:0) + K]+

[PC(diacyl-16:0/20:4) + K]+ HIP CTX TH

Cp

Mad/P m/z 772.4

m/z 820.5 100%

0% Ion intensity

FIGURE 3.16 Distinct localization of phospholipid molecular species revealed by IMS. Different distribution pattern of phospholipids arise from the distinct fatty acid composition of GPLs.

(B)

(A)

(C)

CBX

HE stained (E)

(F)

HPF TH

PC(diacy-18:0-22:6) PC(diacy-16:0-22:6) (D) (G) (H)

CTX

Cp

Merged

FIGURE 3.17 Purkinje cells selectively contained a DHA-PC. High-magniication IMS at raster size of 15 µm revealed the Purkinje cell-selective distribution of PC(diacyl-18:0/22:6) in the cerebellum. Both optical observation of HE-stained successive brain sections (A, E) and ion images of DHA-PCs (B, F) clearly suggest that the PC was enriched in the Purkinje cell layer (arrowheads). Interestingly, a complementary distribution of another abundant DHA-PC, PC(diacyl-16:0/22:6), was enriched in the granule layer of the cerebellum (C, G). D. Merged image. ML, molecular layer; GL, granule layer; W, white matter; CBX, cerebellar cortex; CP, corpus striatum; CTX, cerebral cortex; HPF, hippocampal formation; TH, thalamus [41].

APPLICATIONS OF MALDI-IMS

abundant DHA-PCs, PC(diacyl-16:0/22:6) and PC(diacyl18:1/22:6), was observed in the granule cells of the cerebellum (Figure 3.17c,d). Such cell-type heterogeneity of the fatty-acid constituent in part relects the cells’ heterogeneous membrane properties. Because of its high degree of unsaturation, DHA-GPLs increase membrane luidity and even regulate the functions of membrane-associated proteins [42–44]. Purkinje cells are the largest neurons in the brain, with intricately elaborate dendritic arbors. Thus, higher membrane luidity may be required for effective transport of membrane-associated proteins via the plasma membrane. Thus, the high-level expression of a DHA-PC may contribute to the transportation of membrane proteins in the cells.

57

3.2.2.3 IMS of Endogenous Metabolites: Gangliosides Gangliosides are glycosphingolipids consisting of monoto polysialylated oligosaccharide chains of variable lengths attached to a ceramide unit. They are inserted in the outer layer of the plasma membrane, with the hydrophobic ceramide moiety acting as an anchor while their oligosaccharide moiety is exposed to the external medium [45]. Gangliosides also comprise a large family; their constituent oligosaccharides differ in glycosidic linkage position, sugar coniguration, and the contents of neutral sugars and sialic acid. Along with the oligosaccharide unit, the ceramide moiety of gangliosides also varies with respect to the type of long-chain base (LCB) (sphingosine base) and the fatty acid to which it is coupled (Figure 3.18).

(A)

HO

OI

I IO

O O

O HO

O

O −O

OH O

NHAc O

OH

O

O OH−IO

O−

Ceramide

OH O

O

H

HO I− HO

LCB (d18:1)

OH

O

N C O

Fatty Acid (18:0)

OH HO

−N O

Oligosaccharide

(B) C18-species

Oligosaccharides

OH

O

N

H

LCB d18:1

C O

Fatty Acid (C18:0)

C20-species Oligosaccharides

O

OH

LCB d20:1

H N C O

Fatty Acid (C18:0)

FIGURE 3.18 Structure of GM1a. Gangliosides comprise a large family; their oligosaccharides structures differ in the glycosidic linkage position, sugar coniguration, and the contents of neutral sugars and sialic acid content. The ceramide moiety of gangliosides also has some variation with respect to the type of long-chain base (LCB) (sphingosine-base) and fatty acid moiety (A). Structures of ganglioside molecular species containing C18- and C20-long chain base (LCB) are shown in (B).

58

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION

Previous biochemical studies have revealed that the LCB of the brain ganglioside species has either 18 or 20 carbon atoms (i.e., C18- or C20-sphingosine), and C20sphingosine (C20-LCB species) is present in signiicant amounts only in the central nervous system [46–49]. The LCB content increases signiicantly in rodents and humans throughout life [50–52]. The C20-LCB gangliosides are of great interest because of their characteristic brain speciicity and their dramatic increase during the organism’s life span. However, the ield lacks the capability to provide a precise tissue distribution of C18 and C20 gangliosides. Antibodies to some oligosaccharide moieties are available for visualizing the molecular species with different constituent oligosaccharides [53], but such immunological methods cannot detect the differences in the ceramide structure hidden in the lipid bilayer. Due to the negative charge on the sialic acids and their rich abundance in the brain, gangliosides are easily detected in the m/z 1500–2500 range with IMS in the negative ion detection mode [54–56]. In addition, IMS discriminates not only structural differences in oligosaccharides but also in the lipid moiety, and therefore, the speciic distribution of the C20-LCB species is successfully revealed in the mouse brain. While the C18 species is widely distributed throughout the frontal brain, the C20 species is selectively localized in the ML of the dentate gyrus (DG) (Figure 3.19). Furthermore, the developmental- and aging-related accumulation of the C20 species in the ML-DG can be visualized [55], therefore the tissue location of C20 gangliosides accumulation can be identiied (Figure 3.20). These observations indicate that this brain regionspeciic regulation of LCB chain length is, in particular, important for its distinct function in the brain. As this study clearly demonstrates, the novel capabilities of IMS could shed light on long-standing questions in the biological/clinical ield. 3.2.2.4 IMS of Endogenous Metabolites: Primary Metabolites In the body, primary metabolites are directly involved in normal growth, development, and reproduction processes. Most of these molecules are smaller than typical lipids (15 µm, the sensitivity deteriorates, particularly when high molecular weight proteins are analyzed [66] (Figure 3.30). This difference can be attributed to a phenomenon referred to as the “charging effect” [67]. Generally, biological tissue sections have low intrinsic electric conductivity, and this tendency is considered more apparent with thicker tissue sections. In this state, a surplus electric charge generated by laser irradiation is not lost through the sample stage. Thus, multiple charged ions are generated—and ultimately leads to a signiicant loss of sample ions that would otherwise reach the detector [67]. However, a high technical proiciency is required in order to prepare slices with thicknesses of several micrometers each. Currently,

most samples are prepared with a slice thickness of 10–20 µm [68,69]. These medium-thick sections appear to provide a good compromise between optimal IMS performance and experimental eficiency [70], particularly when a large number of samples need to be analyzed [71]. The “metal sputtering” technique enhances the signal intensity and thus image quality [63,72], presumably by avoiding the “charging effect.” 3. Dehydration of tissue sections for long times can lead to altered signals [73]. Goodwin and colleagues demonstrated that, even within 1 min, signals were altered, both increasing and decreasing. Therefore, tissue slices should be moved to the next step (matrix application) as quickly as possible. Considerable care is required at these stages in order to facilitate a comparison between the biomarkers in independent IMS experiments. 3.3.2.3 Spray-Coating of the Matrix Solution with an Artistic Airbrush Among the several matrix application methods, the spray-coating method is one of the frequently used methods. In this process, an entire tissue section can be coated with relatively small crystals homogenously. For this operation, several instruments, including TLC sprayers and artistic airbrushes, are available. To achieve optimal spraying: (1) minimize the droplet size, to accelerate the dry rate, (2) keep the airbrush in a position at same distance from tissue, and (3) gradually move the airbrush horizontally. Further technical tip to bear in mind when executing this method is to maintain equilibrium between the two rates—one rate, at which a ine aerosol of airbrushed matrix solution slightly moisturizes the tissue section which facilitates analyte extraction, and the other rate, at which the quick solvent evaporates which prevent analyte migration (Figure 3.31 inset). To execute this, important parameters to care for practical spraying include (1) the size of the droplet, (2) the amount of the mist, (3) the angles and distances between the spray nozzle and the tissue section, and (4) laboratory temperature and humidity. These parameters are described below in detail. Preparation of Matrix Solution 1. Weigh appropriate amount of matrix compound and put it into an organic solvent-tolerant microtube. 2. Add 1.0 mL of solvent with organic solventtolerant pipette tips. 3. Dissolve the matrix compound thoroughly by vortexing or performing brief sonication. 4. Store at room temperature until use.

A

Intensity

(a)

OCT

Tissue slice

OCT

Intensity

(b)

4500

7500 m/z

6000

9000

10500

Relative intensity

B

600

800

1000

1200

1400

1600

1800

m/z

2000

6716

FIGURE 3.29 Residual OCT polymer on the tissue slices degrades the mass spectra. A. Decrease in detection sensitivity of the ions originating from proteins due to contamination with OCT. OCT adhering to the tissue section diminishes the detectable peaks. (a) A case in which OCT was used only for supporting the tissue block. (b) A case in which the tissue block was completely embedded with OCT. B. Contamination with OCT leads to the presence of extremely high polymer peaks.

(A) o

17870 18412

14124

11783 12133 12369

10871

9977 10187

8772 8945

7410 7410 7534 8024 8448

4964 5442 5696

4279

3351

Rerative Intensity

8564

laser

(B) (C) (D) (E) (F) (G) 3000

5000

7000

9000

11000

13000

15000

17000

19000

21000 m/z

FIGURE 3.30 Mass spectra obtained from the cerebral cortex region [66]. The mass spectra obtained from the cerebral cortex region in mouse brain slices (a white circle in (A) with thicknesses of 2 µm (B), 5 µm (C), 10 µm (D), 15 µm (E), 30 µm (F), and 40 µm (G). A larger number of mass peaks with high signal-to-noise (S/N) ratios were observed in the spectra obtained from sections with 2, 5, and 10 µm thicknesses than in those obtained from sections with 15, 30 and 40 µm thicknesses.

68

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION

Analyte extraction & interfusion of matrix solution

Temperature & humidity

Number of cycles

Flow rate

Moving rate

Ai

rp

Angle

re s

su

re

Distance Mist size

Tissue

FIGURE 3.31 Representative parameters of spraying operation with an airbrush, required to be controlled among the trials for the reproducible IMS experiment.

Notes: •





Microtubes with organic solvent-tolerant properties can be purchased from Eppendorf Co., Ltd. Use HPLC or LC/MS-grade solvent for preparing matrix solution. Preparing the matrix solution on the day of experiment is recommended.

Matrix Application Using an Airbrush 1. Pour approximately 1.0 mL of the solvent for the matrix solution into the airbrush. 2. Optimize the size of the droplet, the amount of mist, and the angles and distances between the nozzle and the sample. 3. Remove the solvent from the airbrush. 4. Mask the areas outside of the tissue section, which is mounted on the ITO slide glass, with a masking tape. 5. Fix the masked ITO slide glass onto a perpendicular board. 6. Pour the prepared matrix solution into the airbrush and spray onto the tissue section on the ITO slide glass. 7. As a rough guide for optimal coating, spray approximately 0.5–1.0 mL of matrix solution for one ITO slide glass. 8. Remove the masking tape from the ITO slide glass after spraying, and place the sample in an airtight container with dry silica gel.

9. Perform the IMS measurement as soon as possible to minimize the progress of sample damage. 10. Figure 3.32 shows an example of good and unfavorable spraying results by spraying DHB matrix solution (50 mg/mL, 70% MeOH containing 10 mM potassium acetate). 3.3.2.4 Spectrum Normalization In the data analysis of MALDI-IMS, experimenters should not interpret peak intensities of each metabolite simply as metabolite concentrations. Dr. R. Murphy explained this issue using a following equation. Observed ion intensities were result of not only metabolite concentration, but also functions of ionization eficiency of each metabolite compound and local environmental factors [74]: Intensity m / z = f ([metabolite conc.] × [ionization cross-section] × [local environment] × …). In this equation, local environmental factors include analyte extraction eficiency from the distinct tissue structures, and as the most important factor, the matrix crystal condition. Figure 3.33 shows scanning electron microscope (SEM) images of DHB crystals on the tissue section as a result of manual spraying of the matrix solution. Although the presented manual spraying result could be classiied into “good” example (like that shown in Figure 3.31A), as can been seen, the SEM observation revealed rather heterogeneous distribution of DHB crystals on the tissue surface. If tissue surface is scanned with a typical MALDI laser, there could be “hot spot” in which more ions were detected than other location,

EXPERIMENTAL PROCEDURES A. O

B. X

69

C. X SEM images of DHB crystals (×100)

FIGURE 3.32 Result examples of DHB spray coating with the airbrush [41]. A. A properly handled spray-coating step created a uniform matrix crystal layer, and awareness of certain technical points leads to a successful coating step. B. Too small a distance between the airbrush and the tissues (80%, at room temperature). The upper panel shows stereoscopic microscope images, and the lower panel shows phase-contrast microscopic images of the matrix layer formed on the glass slides.

300 µm (×500)

Typical MALDI Laser spot ( 60 µm)

and it in turn results in spot-to-spot variance of signal intensities [75,76]. This problem was at least partially solved using a spectrum-normalization procedure with total ion current (TIC) (Figure 3.33). We previously studied the effectiveness of spectrum normalization with TIC; the obtained spectra were multiplied with arbitrary variables such that all spectra had equal TIC values (i.e., equal integral values of the measured m/z region [m/z 400–900]). Such TIC normalization is available with the “Normalize Spectra” function of FlexImaging 2.0 software (Bruker Daltinics) with ilter function to exclude a number of noise spectra from the normalization process (see details in the software manual). To evaluate the effect of the normalization procedure, we prepared a section of mouse brain homogenate that had a uniform distribution of biomolecules. Figure 3.34A shows the ion images for m/z 772.6 corresponding to PC(diacyl-16:0/16:0), with and without spectrum normalization. After the normalization procedure, the image was corrected such that the ion distribution was uniform throughout the section. The signal intensity was then plotted and found to have a Gaussian distribution. Spectrum normalization with TIC improved the results

60 µm

FIGURE 3.33 Size comparison of DHB crystal and typical MALDI-laser spot. DHB solution was sprayed onto the mouse brain section and the crystals were observed by SEM. The white ellipse represents the typical size of MALDI laser spot. Each matrix crystals typically have 20–60 µm length, and they did not distribute heterogeneously on the tissue surface.

of the IMS of mouse brain sections. Figure 3.34B shows the ion images of a mouse brain section for PC(diacyl-16:0/16:0), with and without spectrum normalization. In the ion image without normalization, the ion distribution was heterogeneous, even between adjacent pixels. Furthermore, the signal intensity was found to decrease with time (arrowhead). In contrast, when the normalization procedure was used, a clear iondistribution pattern that correlated well with the anatomical features of the brain section was obtained [41].

70

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION

A

TIC normalized 25

No normalization No normalization

TIC normalized

Number of pixels

20 15 10 5

~10,500

~9500

~8500

~7500

~6500

~5500

~4500

~3500

~2500

~10,500

~9500

~8500

~7500

~6500

~5500

~4500

~3500

~2500

0

Peak interisity at m/z 772.6

B

Mesurement start

Mesurement end

TIC normalized

No normalization

FIGURE 3.34 Spectrum normalization using TIC improves both the quantitative ability and visualization quality of IMS. A. IMS results for PC(diacyl-16:0/16:0) on a section of mouse brain homogenate, processed with or without TIC normalization (upper panel), and plot of ion intensity distribution for PC(diacyl-16:0/16:0) obtained from a brain homogenate section, with or without TIC normalization (lower panel). B. Ion images of PC(diacyl-16:0/16:0) on an adult mouse brain section, in which spectra were processed with or without TIC-normalization.

Another way of the spectrum normalization is using an external standard (ES) compound spiked in the matrix solution. We have also studied this normalization method for phospholipid imaging; the methylcarbamyl platelet-activating factor (C-PAF) (C-16) was used as the ES compound by considering the following two criteria: (1) no other mass peaks overlap the peak of the ES compound and (2) the ES compound has suficient ionization capability on the tissue section, in which numerous biological compounds compete to ionize. All obtained spectra were multiplied to equalize the intensity of the ES and such normalization produces improved ion images of biomolecules by eliminating the variations in ionization eficiency. First, we determined the optimal concentration of C-PAF for spectrum normalization. The spectra were

normalized so that the C-PAF peaks at m/z 561.4 have equal intensities during the normalizing process. Figure 3.35A shows the ion images of C-PAF (m/z 561) with and without spectrum normalization. Regarding the brain section sprayed with the matrix solution containing 0.5, 5, and 25 mg/mL of C-PAF, spectrum normalization could not be performed successfully at some data points because of insuficient intensity of the ES peak, since mass peaks with insuficient intensity of C-PAF could not be recognized by the software as ES peaks. On the other hand, successful normalization was achieved for the section sprayed with 50 mg/mL of C-PAF, since ion signals of C-PAF equalized among the data points. Figure 3.35B shows the ion images of PC(diacyl16:0/18:1) (m/z 782) with and without spectrum normal-

STATISTICAL PROCEDURES FOR IMS DATA ANALYSIS

71

A

(External standard)

Unnormalized 0.5 mg/mL

5 mg/mL

25 mg/mL

m/z 561 50 mg/mL

5 mg/mL

25 mg/mL

50 mg/mL

Normalized

B

(Endogenous lipids)

Unnormalized 0.5 mg/mL

m/z 782 Optical image

1mm

Corpus callosum Normalized

Hippocampus

FIGURE 3.35 Comparison of unnormalized and normalized ion images. These images are successive mouse coronal brain sections. The ion at m/z 561 is derived from ES and m/z 782 is derived from PC(diacyl-16:0/18:1).

ization. While the inner structure of the brain such as the hippocampus region could not be distinguished in the unnormalized ion images, normalized images (50 mg/mL of C-PAF) showed a clear borderline of the hippocampus, indicated by the absence of PC (diacyl-16:0/18:1) in the corpus callosum (arrowheads). The normalization process with C-PAF clearly improved the ion distribution images, providing sharply deined tissue edges and increased dynamic range. 3.4 STATISTICAL PROCEDURES FOR IMS DATA ANALYSIS 3.4.1 MALDI-IMS with Statistical Analyses Revealed Abnormal Distribution of Metabolites in Colon Cancer Liver Metastasis In the following chapter, a medical application of MALDI-IMS to colon cancer liver metastasis with use

of the presented procedures is described. In addition to the shown procedures, here we employed statistical procedure for eficient analysis of complex IMS data sets. The goal of this study is to discover potential biomarkers which are speciically found in normal or diseased cells. The IMS capability to simultaneously detect multiple metabolites at a time, even with spatial information, facilitates this emerging technique as an effective tool for biomarker discovery, within surgically resected tissues. In fact, a previous study has shown that IMS can discriminate cancer types (such as primary or nonprimary cancer) based on their molecular signature, and can even predict survival rate among human patients [77]. For this kind of purpose, it is necessary to utilize statistical analyses to extract useful information from enormous IMS data sets. The MS of tissues gives an extremely complex spectrum with hundreds of to a thousand peaks obtained from a single data point, and

72

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION

furthermore, several thousands of spectra with spatial data are obtained at one IMS measurement. Because of the complexity and enormousness of the IMS data set, for discovery of biomarkers, manual processing of the data set in order to obtain signiicant information is not a realistic procedure. In this regard, today, multivariate analysis becomes a powerful tool in IMS data analysis. Here, we applied the statistical procedure to the IMS results of the pathological specimen, colon cancer liver metastasis. Colon cancer is a challenging worldwide clinical problem and the incidence rate of colon cancer has been rising rapidly in Japan [78]. Genealogy is known to be a risk factor [79] and as environmental factors, aging [80] and diet, particularly a high intake of animal protein and fat along with a low intake of iber [81], increase the incidence of colorectal cancer. Until today, a number of approaches including a cDNA microarray have revealed characteristics of cancer cells with some success, such as the discovery of speciic gene expressions for drug resistance [82]. In addition to this, IMS approaches presented here which enable comprehensive analysis of metabolite expression patterns in tissues might improve our ability to understand the molecular complexities of tumor cells. In this chapter, we will show altered composition of metabolites in the cancerous tissue revealed by IMS, with both manual data processing and statistical data management. In particular, as a statistical strategy, an unsupervised multivariate data analysis technique that enables us to sort the data sets without any reference information is described. A major method that is related to IMS, namely principal component analysis (PCA), will be described in detail. 3.4.2

Materials and Methods

3.4.2.1 Chemicals TFA was purchased from Merck (Darmstadt, Germany). Methanol was purchased from Wako Pure Chemical Industries (Osaka, Japan). 2,6dihydroxy acetophenone (2,6-DHA) was purchased from Bruker Daltonics. A calibration standard for the low m/z region was prepared by mixing angiotensin III ([M+H]+: 899.47) and Leu-Euk ([M+H]+: 556.28). All the chemicals used in this study were of the highest purity available. 3.4.2.2 Conductive Sheet The conductive sheet was purchased from Tobi Co., Ltd. (Osaka, Japan). This sheet has a thin ITO layer on a polyethylene terephthalate. The sheet was 125 µm thick and its conductivity was 100 Ω. The transparency was 80% (λ = 550 nm), so that we could observe stained tissues with transmitted light.

This lexible sheet made sample handling easy, because the sheet could be cut to an arbitrary size with a paper cutter and samples did not crack easily, which was sometimes problematic with glass slides. 3.4.2.3 Tissue Block Preparation A tissue block with colon cancer liver metastasis was removed from a Japanese patient during an operation, and rinsed with PBS buffer. The tissue was then immediately frozen in liquid nitrogen to minimize degradation, and was kept at −80°C. Informed consent was obtained before the operation. 3.4.2.4 Sample Preparation Before sectioning, the liver block was left for 30 min at −20°C. The tissue sections were sliced to a thickness of 3 µm using a cryostat (CM 3050; Leica) and mounted onto the ITO sheet. A thin matrix layer was applied to the surface by an airbrush. A 2-min spraying of 2,6-DHA solution (30 mg/ mL in 70% methanol/0.1% TFA) was iterated twice. During spraying, the distance between the nozzle and the tissue surface was kept at 15 cm. After drying, the ITO sheet was attached to a metal-coated glass slide by conductive tape to facilitate electrical conduction. 3.4.2.5 Conditions of MS and MALDI-IMS The tissue section was analyzed using a matrix-assisted laser desorption/ionization time-of-light mass spectrometry (MALDI-TOF)/time of light (TOF)-type instrument, Ultralex II TOF/TOF (Bruker Daltonics), which was equipped with a Nd:YAG laser with a 200 Hz repetition rate. External calibration solution was deposited on the surface of the ITO sheet to minimize mass shift. In this experiment, an acceleration voltage was set to 25 kV. 3.4.2.6 IMS A raster scan on the tissue surface was performed automatically. Laser irradiation consisted of 100 shots in each spot. The interval of data points was 100 µm, giving a total of 445 data points in the tissue section. The spectra shown in the results section were accumulated in square sections (300 × 300 µm) of normal and cancerous areas. Here, we did not apply data processing such as smoothing or baseline subtraction. The reconstructions from the spectra were performed by FlexImaging (Bruker Daltonics). 3.4.2.7 Statistical Analysis Statistical analyses were carried out using the ClinProTools 2.2 Software (Bruker Daltonics). For the statistical analyses, the mass spectra were internally recalibrated on common peaks (also known as spectral alignment) and normalized on the TIC. An average spectrum created from all single spectra was used for a peak picking and to deine integration ranges. These integration ranges were used to

STATISTICAL PROCEDURES FOR IMS DATA ANALYSIS

obtain the intensities or areas on the single spectra. The signal intensities were used for all calculations. 3.4.3

Figure 3.37 shows the averaged mass spectra from the cancerous (Figure 3.37a) and normal area (Figure 3.37b). Numerous differences on the mass signals were observed between the normal and cancerous cells. In particular, we found that the signal at m/z 725 showed a massive increase in the cancerous region while the ion at m/z 616 almost disappeared in the cancer cells.

Results and Discussion

3.4.3.1 Comparison of Averaged Mass Spectra in Normal and Cancerous Areas At irst, a tissue section with colon cancer liver metastasis was stained with HE for histological observation (Figure 3.36a). The histochemical staining enables us to distinguish the normal, stroma, and cancer cells which were localized on the left, middle, and right locations of the tissue section, respectively. A successive tissue section was used for MALDIIMS and after the measurement, according to the histological observation, two quadrate areas—one from the normal area and the other from the cancerous area—were selected to collect and average the obtained mass spectra (Figure 3.36b).

3.4.3.2 Visualization of Molecules Speciically Localized in Normal and Cancerous Region Having demonstrated the cancerous/normal tissue speciic localization of ions at m/z 616 and 725, respectively, we proceed to the visualization of their distribution pattern. As expected, the ion distribution images shown in Figure 3.38 demonstrate that they are expressed in the normal/diseased region speciic manner; ion at m/z 725 was clearly localized in the cancerous region while ion at m/z 616 was found only in the normal cell region. The merged image demonstrates that these two ions were complementarily distributed in the specimen. 3.4.3.3 Molecular Identiication with MS/MS Next question is the origin of these two ions. As shown in previous chapters, MS/MS provides the structural information of interest ions and therefore, it enables the molecular identiication. The result of MS/MS with regard to m/z 725 is shown in Figure 3.39A. In the product ion mass spectrum, peaks at m/z 666.5 and 542.5, which correspond to neutral loss (NL) of trimethylamine (59 u, C3H9N) and NL of trimethylamine and cyclophosphate (124 u, C2H5O4P), respectively, were detected. This result indicates that m/z 725 contained an alkali metal adduct phosphocholine, therefore ion at m/z 725 is suggested to be PC or sphingomyelin (SM) [41]. According to the nitrogen rule, ion at m/z 725 having odd nominal mass should contain additional nitrogen in its structure, thus indicating presence of a sphingosine. We concluded that m/z 725 was attributed to be a sodiated molecule of SM (16:0).

(B)

(A) Normal region Cancer region

Stroma region

Accumulated area

FIGURE 3.36 Histological observation of HE-stained liver tissue section with colon cancer metastasis. A. The HE-stained section allows us to distinguish the normal, stroma, and cancer cells which were localized on the left, middle, and right locations of the section, respectively. B. Photograph of the tissue section prepared for MALDI-IMS. From the data points in white squares represented in B, mass spectra were collected and averaged. Bars: 1 mm.

(A) Cancer Relative intensity

73

m/z 725.5

(B) Normal m/z 616.1

610 620 630 640 650 660 670 680 690 700 710 720 730 740 750 760 770 780 790 800 810 820 830 840 850 860 870 880 890 m/z

m/z

FIGURE 3.37 Comparison of averaged mass spectra from the (A) cancerous and (B) normal areas.

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION

Relative intensities

725.458

(B)

Relative intensities

(A)

616.074

74

Ion image at m/z 725.5

Ion image at m/z 616.1

616 618 m/z

620

725 727 m/z

729

FIGURE 3.38 Visualization of molecules speciically localized in normal and cancerous region. Ion distribution images and corresponding mass spectra demonstrate the strong distribution of ion at m/z 616 in the normal area (A), while the ion at m/z 725 showed higher expression in the cancerous area than in the normal area (B).

Regarding the ion at m/z 616, the irst generation of product ion mass spectrum from m/z 616 showed consecutive NLs of 73, 59, and 45 Da (Figure 3.39B[a]). From the previous literature [83,84], it is suggested that m/z 616 corresponded to heme B and that these NLs were derived from loss of the CH2CH2COOH (73 Da), CH2COOH (59 Da), or the COOH (45 Da) group, respectively. The molecular structure of heme B is displayed as an inset. Figure 3.39B(b) shows the second product ion mass spectrum generated from m/z 557, and an additional NL of 59 Da was observed. This fragment was considered to be derived from another CH2COOH in heme B. Here, we demonstrate that SM(16:0) was strongly expressed in the cancerous area. Previous studies reported that in colon cancer, the cancerous cells contain elevated amounts of total phospholipids [85], and in addition, the phospholipid composition of the cellular membrane is altered [85,86] even between cancer cell types, that is, metastases and nonmetastatic cancer [86]. Brasitus et al. studied a relationship between the malignancy and altered lipid composition of the colon cancer, and reported signiicant accumulation of SM, consistent with the presented result [87]. On the other hand, heme B consists of an iron atom and porphyrin, and is known as a prosthetic group in hemoglobin, which is a protein in erythrocytes. Presented results indicate the difference between the blood-rich organ liver and the ischemic metastatic colon cancer [88,89]. 3.4.3.4 IMS Linked to Multivariate Analysis Up to this point, we showed that two small metabolites were speciically expressed between the cancerous and normal tissue areas. In the described data analysis pro-

cedures without statistical methods, we usually averaged the spectra of each region and visually compared the mass peaks between the spectra one by one. As seen in Figure 3.37, with such visual comparisons of spectra, we were certainly able to ind differences among the peak expressions. However, such methodology is ineficient especially when one is analyzing a large number of mass peaks and/or many tissue samples. Below, we will describe the IMS-linked PCA to compare the metabolite composition of the normal/ cancerous regions. Here, we will not describe the detailed mathematical theory due to the space limitation, but in brief, PCA is a statistical method that merges the data containing multiple elements into low-dimensional data. It reduces a large set of variables to a small set of variables called “principal components” which are linear combinations of the original variables. In the PCAcoupled IMS data analysis, spectra obtained by IMS are processed to peak detection and based on the generated peak list, PCA decomposition was performed. PCA images (i.e., 2D intensity map of principle component score on the tissue section) were often utilized to ind differences of molecular composition among regions/ tissues. PCA calculation results in several parameters and below, the component score and factor loading are particularly important for the interpretation of results. A component score is calculated for each mass spectrum; all are deined for each principal component (e.g., for PC1, PC2, etc.). Those component scores are often plotted two-dimensionally, to facilitate interpretation of the PCA results. In Figure 3.40, component scores for each principle component are plotted on the x- and yaxis, and each dot in the graph represents a spectrum

STATISTICAL PROCEDURES FOR IMS DATA ANALYSIS

75

Na+ − O O

N

80

+

O

OH

666.5

100

d18:1

O NH

60

NL 59

NL 183

O

C16:0

300

400

NL 59

542.5

NL 183

40

725.5

Relative intensity(%)

(A)

184

20 0

200

600

500

700 m/z

(B) (a)

Relative intensity(%)

100

557.2

MS/MS of m/z 616

616.2 59 Da N

N Fe

50

N

571.2 −73 Da HOOC

0

100

200

45 Da

−59 Da

−45 Da

COOH

300

400

600

500

m/z

(b) 100 Relative intensity(%)

73 Da

543.2

N

498.1

MS/MS of m/z 557 (MS3 of 616)

557.2 59 Da

50

0

100

200

300

400

500

600 m/z

FIGURE 3.39 MS/MS enables the molecular identiication of interested ions directly on the tissue surface. A. Product ion mass spectrum on the liver section of m/z 725. The NL of 59 and 124 u observed in the spectra is trimethylamine and cyclophosphate, indicating phosphocholine structure. This fragmentation occurred when alkali metal adducted to the precursor ion. The biomolecule of m/z 725 was suggested to be the sodiated molecule of SM(16:0). B. Product ion mass spectra on the liver section of m/z 616.2 (a) and 557.2 (b). The m/z value and fragment patterns indicate that the product ion of m/z 616 is heme B. Consecutive NLs of 73, 59, and 45 Da correspond to CH2CH2COOH, CH2COOH, and COOH, respectively. The molecular structure of heme B is shown as an inset in (a).

from a distinct data point on the tissue section. What is important to note is whether two (or several) populations of spectra (= dot) are obtained from distinct regions, for example, normal versus diseased, they are spatially separated on the graph, or not. If they are

separated (Figure 3.40A), it means that the molecular expression patterns of these two regions were statistically distinct from each other. If not, PCA failed to extract the statistical differences between the populations (Figure 3.40B).

76

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION Data point (spectrum with spatial infomation)

Sample A

(A)

Sample B

(B)

PC1 score

PC2 score

PC1 score

PC2 score

FIGURE 3.40 Example of data interpretation of IMS-linked PCA. In this study, dots seen in the 2D plot represent the case, that is, the spectrum from distinct data points. If dots from distinct sample are separated (A), it means that the molecular expression patterns of these two regions were statistically distinct from each other. If not, PCA failed to extract the statistical differences between the two populations (B).

Figure 3.41 shows the result of imaging mass spectrometry–principal component analysis (IMSPCA) for the colon cancer tissue. In this case, this unsupervised analysis revealed that the largest spectral difference (i.e., the largest difference in metabolite composition) was observed between the normal and the other tissue areas (i.e., normal vs. stroma/cancer area), and the second largest difference was observed between the stroma and normal/cancer area. The overall interpretation of PCA was shown in Table 3.6. In the graphs shown in Figure 3.41B, the circles indicate mass spectra obtained from the normal, stroma, and cancerous regions (colored white, gray, and black, respectively). Notably, the three populations are spatially separated by the component scores for PC2 and PC3, but not for PC1. This indicates that PC2 and PC3 particularly contain the statistical differences among these three regions. In detail, along with the PC2 scores (x-axis), the spectra from normal and the others are clearly separated. The PC2 image also demonstrates a large difference of PC2 score value between normal and the other regions (Figure 3.41C). On the other hand, along with the PC3 scores (y-axis), spectra from stroma and the other regions are separated and the PC3 image also shows much higher PC3 score value of stroma region than the other regions (Figure 3.41C).

FIGURE 3.41 IMS-linked PCA of colon cancer liver metastasis section revealed altered metabolite compositions among normal stroma and cancer regions. A. Optical images of HE-stained section after IMS measurement representing the normal, stroma, and cancer regions. B. Graphs in which principle component scores for PC1, PC2, and PC3 are plotted. C. Principal component images. According to the value of the principle component score calculated for the spectrum at each tissue location, pixels are indicated with gray value.

STATISTICAL PROCEDURES FOR IMS DATA ANALYSIS

77

TABLE 3.6 The Results and Interpretation of PCA of the Liver Section with Colon Cancer Metastasis Primary Contributing PCs Component 2

Component 3

Negative

m/z 616.2 (heme B)

Positive Negative Positive

m/z 744.4 m/z 722.0 m/z 760.4

m/z 828.4 c/t [PC(diacyl16:0/22:6)+Na]+) m/z 754.4 – m/z 766.4

3.4.3.5 Analysis of Loading Factor for Each Principle Component Facilitates Identiication of “Responsible” Molecules Which Differentiate Control and Diseased Samples As a next step, an analysis of the factor loading plot would identify peaks that were differentially expressed between regions. Since a component score deined for each spectrum is a sum of the value of the factor loading value, multiplied by peak intensity, when numbers (= m) of mass peaks were used in the analysis, the component score will be: m

ScorePC1( x, y) =

∑ load(n) × Int.(n), n =1

Interpretation of the Component m/z 844.4 c/t [PC(diacyl16:0/22:6)+Na]+) m/z 768.4 –

Representing altered metabolite composition between normal and other regions. Representing altered metabolite composition between stroma and other regions.

be speciically expressed by the normal liver cells, we thus picked up mass peaks with large negative loading values for PC2, and obtained their distribution image. As a result, we found that ion at m/z 616.2, which is already demonstrated as a normal region-speciic molecule, was statistically classiied into the normal region speciic category (Figure 3.42, lower part). Furthermore, by this procedure, other mass peaks corresponding to sodiated and potassiated molecules of PC(diacyl16:0/22:6) were identiied as the normal cell speciic metabolites (Figure 3.42, lower part). On the other hand, through the same procedure, we also successfully identiied the molecules which localized in the cancerous/ stroma region (Figure 3.42, upper part).

where ScorePC1(x,y) = component score against PC1, obtained from (x, y), load(n) = factor loading value against a mass peak for n, Int.(n) = mass peak intensity for n, m = number of mass peaks used for calculation. According to this equation, in the spectra from the normal tissue region, the mass peak with large negative value regarding PC2 factor loading is supposed to be intense. On the other hand, it was also indicated that peaks with large positive values for PC2 factor loading would be speciic molecules to the stroma/cancerous regions. In other words, such mass peaks with a radical absolute value for PC2 factor loading are suggested to be major contributors to differentiate these regions. 3.4.3.6 PC2 Represents Difference of Metabolite Composition between Normal versus Other Region In Figure 3.42, the factor loading values for PC1 and PC2 are plotted on the x- and y-axes, respectively. Each dot indicates a distinct mass peak. Such a graph makes it very easy to ind the peaks with the intended factor loading value against each PC. Since peaks that have negative loading values regarding PC2 are supposed to

3.4.3.7 PC3 Represents Difference of Metabolite Composition between Stroma versus Other Region In Figure 3.43, the factor loading values for PC1 and PC3 are plotted on the x- and y-axes, respectively. In this case, since mass peaks that have a large positive loading value regarding PC3 are supposed to be speciic molecules to the stroma region, we picked up the mass peak at m/z 722.0 and obtained a distribution image (Figure 3.43, upper part). As a result, we identiied ions at m/z 722.0 as stroma-speciic molecules, and on the other hand, with same procedure, we revealed that ions at m/z 760.4 and 766.4 almost disappeared in the stroma region.

3.4.4

Conclusion

Biomarkers are objective indicators of particular pathogenic processes, pharmacological responses, or normal biological states; they can involve any kind of molecule in living organs, for example, proteins, peptides, DNA, and/or metabolites. Biomarkers are essential for the diagnosis and prediction of diseases; IMS can provide distribution information regarding various biomolecules at the cell and tissue levels, and thus it is expected to become a powerful tool for in situ biomarker discovery.

78

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION

FIGURE 3.42 PC2 represents difference of metabolic composition between normal and other region. In the center graph, factor loading values for PC1 and PC2 are plotted on the x- and y-axes, respectively. Each dot indicates the distinct mass peak. The peaks with a large positive/negative value for loading factor 2 (i.e., major contributors to differentiation among the groups) were chosen and their distribution images are visualized. c/t, corresponding to.

STATISTICAL PROCEDURES FOR IMS DATA ANALYSIS

79

FIGURE 3.43 PC2 represents difference of metabolic composition between normal versus other region. In the center graph, factor loading values for PC1 and PC3 are plotted on the x- and y-axes, respectively. The mass peaks with large positive/negative value for loading factor 3 were chosen and their distribution images are visualized.

80

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION

In this chapter, we showed the identiication of potential biomarkers, which are molecules that differentiate among the normal, cancerous and even stroma cells in the colon cancer liver metastasis. For this purpose, we showed that the statistical strategy is quite effective to deal with the large volume data set of IMS. The volumes of IMS data sets continue to increase because of current improvements to IMS with regards to high-resolution [90], three-dimensional (3D) imaging [71], and reconstruction from 3D mass spectra containing ion drift times in ion mobility MS [29]. Data analysis of such large data sets will increasingly depend on the statistical analysis, and therefore the development and application of such analyses will be a more important issue.

REFERENCES 1. Garrett, T.J., et al. (2006) Imaging of small molecules in tissue sections with a new intermediate-pressure MALDI linear ion trap mass spectrometer. International Journal of Mass Spectrometry, 260, 11. 2. Khatib-Shahidi, S., Andersson, M., Herman, J.L., Gillespie, T.A., Caprioli, R.M. (2006) Direct molecular analysis of whole-body animal tissue sections by imaging MALDI mass spectrometry. Analytical Chemistry, 78, 6448–6456. 3. Stoeckli, M., Staab, D., Staufenbiel, M., Wiederhold, K.H., Signor, L. (2002) Molecular imaging of amyloid beta peptides in mouse brain sections using mass spectrometry. Analytical Biochemistry, 311, 33–39. 4. Chaurand, P., Norris, J.L., Cornett, D.S., Mobley, J.A., Caprioli, R.M. (2006) New developments in proiling and imaging of proteins from tissue sections by MALDI mass spectrometry. Journal of Proteome Research, 5, 2889– 2900. 5. Krause, E., Wenschuh, H., Jungblut, P.R. (1999) The dominance of arginine-containing peptides in MALDI-derived tryptic mass ingerprints of proteins. Analytical Chemistry, 71, 4160–4165. 6. Gharahdaghi, F., Kirchner, M., Fernandez, J., Mische, S.M. (1996) Peptide-mass proiles of polyvinylidene diluoridebound proteins by matrix-assisted laser desorption/ ionization time-of-light mass spectrometry in the presence of nonionic detergents. Analytical Biochemistry, 233, 94–99. 7. Annesley, T.M. (2003) Ion suppression in mass spectrometry. Clinical Chemistry, 49, 1041–1044. 8. Jones, J.J., Borgmann, S., Wilkins, C.L., O’Brien, R.M. (2006) Characterizing the phospholipid proiles in mammalian tissues by MALDI FTMS. Analytical Chemistry, 78, 3062–3071. 9. Lemaire, R., et al. (2006) MALDI-MS direct tissue analysis of proteins: Improving signal sensitivity using organic treatments. Analytical Chemistry, 78, 7145–7153.

10. Schwartz, S.A., Reyzer, M.L., Caprioli, R.M. (2003) Direct tissue analysis using matrix-assisted laser desorption/ ionization mass spectrometry: practical aspects of sample preparation. Journal of Mass Spectrometry, 38, 699–708. 11. Seeley, E.H., Oppenheimer, S.R., Mi, D., Chaurand, P., Caprioli, R.M. (2008) Enhancement of protein sensitivity for MALDI imaging mass spectrometry after chemical treatment of tissue sections. Journal of the American Society for Mass Spectrometry, 19, 1069–1077. 12. Mann, M., Jensen, O.N. (2003) Proteomic analysis of posttranslational modiications. Nature Biotechnology, 21, 255–261. 13. Yan, R., et al. (1999) Membrane-anchored aspartyl protease with Alzheimer’s disease beta-secretase activity. Nature, 402, 533–537. 14. Rohner, T.C., Staab, D., Stoeckli, M. (2005) MALDI mass spectrometric imaging of biological tissue sections. Mechanisms of Ageing and Development, 126, 177–185. 15. McDonnell, L.A., Heeren, R.M. (2007) Imaging mass spectrometry. Mass Spectrometry Reviews, 26, 606–643. 16. Groseclose, M.R., Andersson, M., Hardesty, W.M., Caprioli, R.M. (2007) Identiication of proteins directly from tissue: in situ tryptic digestions coupled with imaging mass spectrometry. Journal of Mass Spectrometry, 42, 254–262. 17. Shimma, S., Furuta, M., Ichimura, K., Yoshida, Y., Setou, M. (2006) A novel approach to in situ proteome analysis using chemical inkjet printing technology and MALDIQIT-TOF tandem mass spectrometer. Journal of the Mass Spectrometry Society of Japan, 54, 133–140. 18. Shimma, S., et al. (2008) Mass imaging and identiication of biomolecules with MALDI-QIT-TOF-based system. Analytical Chemistry, 80, 878–885. 19. Setou, M., Hayasaka,T., Shimma, S., Sugiura,Y., Matsumoto, M. (2008) Protein denaturation improves enzymatic digestion eficiency for direct tissue analysis using mass spectrometry. Applied Surface Science, 255, 1555–1559. 20. Woo, H.K., Northen, T.R., Yanes, O., Siuzdak, G. (2008) Nanostructure-initiator mass spectrometry: a protocol for preparing and applying NIMS surfaces for high-sensitivity mass analysis. Nature Protocols, 3, 1341–1349. 21. Patti, G.J., et al. (2010) Detection of carbohydrates and steroids by cation-enhanced nanostructure-initiator mass spectrometry (NIMS) for bioluid analysis and tissue imaging. Analytical Chemistry, 82, 121–128. 22. Liu, Q., Guo, Z., He, L. (2007) Mass spectrometry imaging of small molecules using desorption/ionization on silicon. Analytical Chemistry, 79, 3535–3541. 23. Taira, S., et al. (2008) Nanoparticle-assisted laser desorption/ionization based mass imaging with cellular resolution. Analytical Chemistry, 80, 4761–4766. 24. Harada, T., et al. (2009) Visualization of volatile substances in different organelles with an atmospheric-pressure mass microscope. Analytical Chemistry, 81, 9153–9157. 25. Holscher, D., et al. (2009) Matrix-free UV-laser desorption/ ionization (LDI) mass spectrometric imaging at the singlecell level: distribution of secondary metabolites of

REFERENCES

Arabidopsis thaliana and Hypericum species. The Plant Journal, 60, 907–918. 26. Jackson, S.N., Wang, H.Y., Woods, A.S. (2005) In situ structural characterization of phosphatidylcholines in brain tissue using MALDI-MS/MS. Journal of the American Society for Mass Spectrometry, 16, 2052–2056. 27. Jackson, S.N., Wang, H.Y., Woods, A.S. (2007) In situ structural characterization of glycerophospholipids and sulfatides in brain tissue using MALDI-MS/MS. Journal of the American Society for Mass Spectrometry, 18, 17–26. 28. Rujoi, M., Estrada, R., Yappert, M.C. (2004) In situ MALDI-TOF MS regional analysis of neutral phospholipids in lens tissue. Analytical Chemistry, 76, 1657–1663. 29. McLean, J.A., Ridenour, W.B., Caprioli, R.M. (2007) Proiling and imaging of tissues by imaging ion mobilitymass spectrometry. Journal of Mass Spectrometry, 42, 1099–1105. 30. Sugiura, Y., Setou, M. (2009) Imaging mass spectrometry for visualization of drug and endogenous metabolite distribution: toward in situ pharmacometabolomes. Journal of Neuroimmune Pharmacology, 5, 31–43. 31. Wang, H.Y., Jackson, S.N., Woods, A.S. (2007) Direct MALDI-MS analysis of cardiolipin from rat organs sections. Journal of the American Society for Mass Spectrometry, 18, 567–577. 32. Sugiura, Y., Setou, M. (2009) Selective imaging of positively charged polar and nonpolar lipids by optimizing matrix solution composition. Rapid Communications in Mass Spectrometry, 23, 3269–3278. 33. Han, X., Gross, R.W. (2001) Quantitative analysis and molecular species ingerprinting of triacylglyceride molecular species directly from lipid extracts of biological samples by electrospray ionization tandem mass spectrometry. Analytical Biochemistry, 295, 88–100. 34. Hsu, F.F., Turk, J. (2001) Structural determination of glycosphingolipids as lithiated adducts by electrospray ionization mass spectrometry using low-energy collisionalactivated dissociation on a triple stage quadrupole instrument. Journal of the American Society for Mass Spectrometry, 12, 61–79. 35. Piomelli, D., Astarita, G., Rapaka, R. (2007) A neuroscientist’s guide to lipidomics. Nature Reviews. Neuroscience, 8, 743–754. 36. Hitzemann, R.J., Johnson, D.A. (1983) Developmental changes in synaptic membrane lipid composition and luidity. Neurochemical Research, 8, 121–131. 37. Zerouga, M., Jenski, L.J., Stillwell, W. (1995) Comparison of phosphatidylcholines containing one or two docosahexaenoic acyl chains on properties of phospholipid monolayers and bilayers. Biochimica et Biophysica Acta, 1236, 266–272. 38. Ostrowski, S.G., Van Bell, C.T., Winograd, N., Ewing, A.G. (2004) Mass spectrometric imaging of highly curved membranes during Tetrahymena mating. Science, 305, 71–73.

81

39. Murakami, M., Nakatani, Y., Atsumi, G., Inoue, K., Kudo, I. (1997) Regulatory functions of phospholipase A2. Critical Reviews in Immunology, 17, 225–283. 40. Astarita, G., Ahmed, F., Piomelli, D. (2008) Identiication of biosynthetic precursors for the endocannabinoid anandamide in the rat brain. Journal of Lipid Research, 49, 48–57. 41. Sugiura, Y., et al. (2009) Visualization of the cell-selective distribution of PUFA-containing phosphatidylcholines in mouse brain by imaging mass spectrometry. Journal of Lipid Research, 50, 1776–1788. 42. Salem, N., Jr., Litman, B., Kim, H.Y., Gawrisch, K. (2001) Mechanisms of action of docosahexaenoic acid in the nervous system. Lipids, 36, 945–959. 43. Kim, H.Y. (2007) Novel metabolism of docosahexaenoic acid in neural cells. The Journal of Biological Chemistry, 282, 18661–18665. 44. Stubbs, C.D., Smith, A.D. (1984) The modiication of mammalian membrane polyunsaturated fatty acid composition in relation to membrane luidity and function. Biochimica et Biophysica Acta, 779, 89–137. 45. van Echten, G., Sandhoff, K. (1993) Ganglioside metabolism. Enzymology, topology, and regulation. The Journal of Biological Chemistry, 268, 5341–5344. 46. Sonnino, S., Chigorno, V. (2000) Ganglioside molecular species containing C18- and C20-sphingosine in mammalian nervous tissues and neuronal cell cultures. Biochimica et Biophysica Acta, 1469, 63–77. 47. Sambasivarao, K., McCluer, R.H. (1964) Lipid components of gangliosides. Journal of Lipid Research, 15, 103–108. 48. Schwarz, H.P., Kostyk, I., Marmolejo, A., Sarappa, C. (1967) Long-chain bases of brain and spinal cord of rabbits. Journal of Neurochemistry, 14, 91–97. 49. Jungalwala, F.B., Hayssen, V., Pasquini, J.M., McCluer, R.H. (1979) Separation of molecular species of sphingomyelin by reversed-phase high-performance liquid chromatography. Journal of Lipid Research, 20, 579– 587. 50. Palestini, P., Sonnino, S., Tettamanti, G. (1991) Lack of the ganglioside molecular species containing the C20-longchain bases in human, rat, mouse, rabbit, cat, dog, and chicken brains during prenatal life. Journal of Neurochemistry, 56, 2048–2050. 51. Palestini, P., Masserini, M., Sonnino, S., Giuliani, A., Tettamanti, G. (1990) Changes in the ceramide composition of rat forebrain gangliosides with age. Journal of Neurochemistry, 54, 230–235. 52. Mansson, J.E., Vanier, M.T., Svennerholm, L. (1978) Changes in the fatty acid and sphingosine composition of the major gangliosides of human brain with age. Journal of Neurochemistry, 30, 273–275. 53. Kotani, M., Kawashima, I., Ozawa, H., Terashima, T., Tai, T. (1993) Differential distribution of major gangliosides in rat central nervous system detected by speciic monoclonal antibodies. Glycobiology, 3, 137–146.

82

IMAGING MASS SPECTROMETRY (IMS) FOR BIOLOGICAL APPLICATION

54. Jackson, S.N., Wang, H.Y., Woods, A.S. (2005) Direct proiling of lipid distribution in brain tissue using MALDITOFMS. Analytical Chemistry, 77, 4523–4527. 55. Hayasaka, T., et al. (2008) Matrix-assisted laser desorption/ ionization quadrupole ion trap time-of-light (MALDIQIT-TOF)-based imaging mass spectrometry reveals a layered distribution of phospholipid molecular species in the mouse retina. Rapid Communications in Mass Spectrometry, 22, 3415–3426. 56. Chen, Y., et al. (2008) Imaging MALDI mass spectrometry using an oscillating capillary nebulizer matrix coating system and its application to analysis of lipids in brain from a mouse model of Tay-Sachs/Sandhoff disease. Analytical Chemistry, 80, 2780–2788. 57. Cornett, D.S., Frappier, S.L., Caprioli, R.M. (2008) MALDIFTICR imaging mass spectrometry of drugs and metabolites in tissue. Analytical Chemistry, 80, 5648–5653. 58. Jackson, S.N., et al. (2007) MALDI-ion mobility-TOFMS imaging of lipids in rat brain tissue. Journal of Mass Spectrometry, 42, 1093–1098. 59. Amantonico, A., Oh, J.Y., Sobek, J., Heinemann, M., Zenobi, R. (2008) Mass spectrometric method for analyzing metabolites in yeast with single cell sensitivity. Angewandte Chemie International Edition,, 47, 5382–5385. 60. Burrell, M., Earnshaw, C., Clench, M. (2007) Imaging matrix assisted laser desorption ionization mass spectrometry: a technique to map plant metabolites within tissues at high spatial resolution. Journal of Experimental Botany, 58, 757–763. 61. Benabdellah, F., Touboul, D., Brunelle, A., Laprevote, O. (2009) In situ primary metabolites localization on a rat brain section by chemical mass spectrometry imaging. Analytical Chemistry, 81, 5557–5560. 62. Knowles, J.R. (1980) Enzyme-catalyzed phosphoryl transfer reactions. Annual Review of Biochemistry, 49, 877–919. 63. Stoeckli, M., Staab, D., Schweitzer, A. (2006) Compound and metabolite distribution measured by MALDI mass spectrometric imaging in whole-body tissue sections. International Journal of Mass Spectrometry, 260, 195–202. 64. Umemura, A., Mabe, H., Nagai, H., Sugino, F. (1992) Action of phospholipases A2 and C on free fatty acid release during complete ischemia in rat neocortex. Effect of phospholipase C inhibitor and N-methyl-D-aspartate antagonist. Journal of Neurosurgery, 76, 648–651. 65. Rehncrona, S., Westerberg, E., Akesson, B., Siesjo, B.K. (1982) Brain cortical fatty acids and phospholipids during and following complete and severe incomplete ischemia. Journal of Neurochemistry, 38, 84–93. 66. Sugiura, Y., Shimma, S., Setou, M. (2006) Thin sectioning improves the peak intensity and signal-to-noise ratio in direct tissue mass spectrometry. Journal of the Mass Spectrometry Society of Japan, 54, 4. 67. Scherl, A., et al. (2005) Gold coating of non-conductive membranes before matrix-assisted laser desorption/

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80. 81.

82.

ionization tandem mass spectrometric analysis prevents charging effect. Rapid Communications in Mass Spectrometry, 19, 605–610. Chaurand, P., Schwartz, S.A., Caprioli, R.M. (2002) Imaging mass spectrometry: a new tool to investigate the spatial organization of peptides and proteins in mammalian tissue sections. Current Opinion in Chemical Biology, 6, 676–681. Stoeckli, M., Chaurand, P., Hallahan, D.E., Caprioli, R.M. (2001) Imaging mass spectrometry: a new technology for the analysis of protein expression in mammalian tissues. Nature Medicine, 7, 493–496. Goodwin, R.J., Pennington, S.R., Pitt, A.R. (2008) Protein and peptides in pictures: imaging with MALDI mass spectrometry. Proteomics, 8, 3785–3800. Andersson, M., Groseclose, M.R., Deutch, A.Y., Caprioli, R.M. (2008) Imaging mass spectrometry of proteins and peptides: 3D volume reconstruction. Nature Methods, 5, 101–108. Altelaar, A.F., et al. (2006) Gold-enhanced biomolecular surface imaging of cells and tissue by SIMS and MALDI mass spectrometry. Analytical Chemistry, 78, 734–742. Goodwin, R.J., Dungworth, J.C., Cobb, S.R., Pitt, A.R. (2008) Time-dependent evolution of tissue markers by MALDI-MS imaging. Proteomics, 8, 3801–3808. Murphy, R.C., LIPID MAPS Lipidomics Workshop. (2009) Future directions: tissue and cell imaging. Available at http://www.lipidmaps.org/resources/lipidmapspresentations/EB2009/MurphyEB2009.pdf. Sugiura, Y., Shimma, S., Setou, M. (2006) Two-step matrix application technique to improve ionization eficiency for matrix-assisted laser desorption/ionization in imaging mass spectrometry. Analytical Chemistry, 78, 8227–8235. Mock, K.K., Sutton, C.W., Cottrell, J.S. (1992) Sample immobilization protocols for matrix-assisted laserdesorption mass spectrometry. Rapid Communications in Mass Spectrometry, 6, 233–238. Yanagisawa, K., et al. (2003) Proteomic patterns of tumour subsets in non-small-cell lung cancer. Lancet, 362, 433–439. Ajiki, W., Tsukuma, H., Oshima, A. (2004) Survival rates of childhood cancer patients in Osaka, Japan. Japanese Journal of Clinical Oncology, 34, 50–54. Harkins, L., et al. (2002) Speciic localisation of human cytomegalovirus nucleic acids and proteins in human colorectal cancer. Lancet, 360, 1557–1563. Elsevier. Newell, G.R., Spitz, M.R., Sider, J.G. (1989) Cancer and age. Seminars in Oncology, 16, 3–9. Kuriki, K., et al. (2006) Risk of colorectal cancer is linked to erythrocyte compositions of fatty acids as biomarkers for dietary intakes of ish, fat, and fatty acids. Cancer Epidemiology, Biomarkers & Prevention, 15, 1791–1798. Huerta, S., et al. (2003) Gene expression proile of metastatic colon cancer cells resistant to cisplatin-induced apoptosis. International Journal of Oncology 22, 663–670.

REFERENCES

83. Demirev, P.A., et al. (2002) Detection of malaria parasites in blood by laser desorption mass spectrometry. Analytical Chemistry, 74, 3262–3266. 84. Rael, L.T., Ayala-Fierro, F., Bar-Or, R., Carter, D.E., Barber, D.S. (2006) Interaction of arsine with hemoglobin in arsine-induced hemolysis. Toxicological Sciences, 90, 142–148. 85. Dueck, D.A., et al. (1996) The modulation of choline phosphoglyceride metabolism in human colon cancer. Molecular and Cellular Biochemistry, 162, 97–103. 86. Dobrzyńska, I., Szachowicz-Petelska, B., Sulkowski, S., Figaszewski, Z. (2005) Changes in electric charge and phospholipids composition in human colorectal cancer cells. Molecular and Cellular Biochemistry, 276, 113–119. 87. Brasitus, T.A., Dudeja, P.K., Dahiya, R. (1986) Premalignant alterations in the lipid composition and luidity of colonic brush border membranes of rats administered 1, 2 dimethylhydrazine. Journal of Clinical Investigation, 77, 831–840. 88. Stark, D.D., Wittenberg, J., Butch, R.J., Ferrucci, J.T. (1987) Hepatic metastases: randomized, controlled comparison of detection with MR imaging and CT. Radiology, 165, 399–406. 89. Stark, D.D., et al. (1988) Superparamagnetic iron oxide: clinical application as a contrast agent for MR imaging of the liver. Radiology, 168, 297–301. 90. McDonnell, L.A., et al. (2005) Subcellular imaging mass spectrometry of brain tissue. Journal of Mass Spectrometry, 40, 160–168.

83

91. Chaurand, P., et al. (2004) Integrating histology and imaging mass spectrometry. Analytical Chemistry, 76, 1145–1155. 92. Astigarraga, E., Barreda-Gómez, G., Lombardero, L., Fresnedo, O., Castaño. F., et al. (2008) Proiling and imaging of lipids on brain and liver tissue by matrix-assisted laser desorption/ionization mass spectrometry using 2-mercaptobenzothiazole as a matrix. Analytical Chemistry, 80, 9105–9114. 93. Sugiura, Y., Shimma, S., Konishi, Y., Yamada, M.K., Setou, M. (2008) Imaging mass spectrometry technology and application on ganglioside study; visualization of agedependent accumulation of C20-ganglioside molecular species in the mouse hippocampus. PLoS ONE, 3, e3232. 94. Ageta, H., Asai, S., Sugiura, Y., Goto-Inoue, N., Zaima, N., Setou, M. (2009) Layer-speciic sulfatide localization in rat hippocampus middle molecular layer is revealed by nanoparticle-assisted laser desorption/ionization imaging mass spectrometry. Medical Molecular Morphology, 42, 16–23. 95. Cha, S., Yeung, E.S. (2007) Colloidal graphite-assisted laser desorption/ionization mass spectrometry and MSn of small molecules. 1. Imaging of cerebrosides directly from rat brain tissue. Analytical Chemistry, 79, 2373–2385. 96. Zhang, H., Cha, S., Yeung, E.S., (2007) Colloidal graphiteassisted laser desorption/ionization MS and MS(n) of small molecules. 2. Direct proiling and MS imaging of small metabolites from fruits. Analytical Chemistry, 79, 6575–6584.

4 METHODOLOGIES FOR IDENTIFYING MICROORGANISMS AND VIRUSES BY MASS CATALOGING OF RNAs George W. Jackson, Rafal Drabek, Mithil Soni, Roger McNichols, Richard C. Willson, and George E. Fox

4.1 INTRODUCTION: THE IMPORTANCE OF MICROBIAL GENOTYPING In light of growing concern over increasing microbial resistance to antibiotics [1–4], recent bioterror attacks, and the continuous emergence and reemergence of infectious diseases, several U.S. government agencies have identiied the development of new infectious disease diagnostics as a key to maintaining the health of the nation and the world as a whole [5–7]. The World Health Organization estimates that over 1600 people die each hour from infectious disease, half of whom are children under 5 years of age. Infectious diseases account for 26% of total global mortality [8] and are the third leading cause of death in the United States [9–10]. Sales of infectious disease diagnostic products in the United States are estimated to represent about $2.8 billion annually, including microbiology, virology, parasitology, and related infectious disease immunoassays [11]. Particularly disconcerting is the emergence and reemergence of antibiotic resistance especially in the hospital setting, where up to 70% of Staphylococcus species are found to be methicillin resistant (methicillinresistant Staphylococcus aureus or MRSA) [12–17]. Rapid, reliable genetic typing methods are therefore not only important for diagnosis and curtailing the overprescription of our best, broad-spectrum antibiotics, but genetic typing is also indispensable for epidemiological tracking.

4.1.1 The Need for Rapid Molecular Identiication Determinative bacteriology often still relies on culturebased methods involving time-consuming isolation, cultivation, and characterization of phenotypic traits. While in a few cases a rapid identiication can be made using phenotypic methods, the phylogenetic resolution of such methods is usually quite low [12]. Characterization of cells based on morphology, staining, and metabolic traits also takes days to weeks for unambiguous identiication [13, 18–19]. Perhaps most importantly, it is estimated that less than 1% of organisms are cultivable under laboratory conditions [20], and therefore, culturebased methods may not be applicable to an emerging pathogen or a bioengineered bacterium. In fact, cultureindependent surveys have determined that there are about 30 or 40 major bacterial phylogenetic groupings with very few or no cultured representatives in collection [21]. Finally, such methods are labor intensive, not amenable to automation, and require extensive “handson” time and interpretation by a trained microbiologist. In the “postgenome” era, molecular methods are rapidly supplanting phenotypic characterization.

4.1.2

Existing Molecular Methods

There are three commonly used approaches to molecular identiication of bacteria or other organisms: immunochemical methods, conventional sequencing, and

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

85

86

METHODOLOGIES FOR IDENTIFYING MICROORGANISMS AND VIRUSES BY MASS CATALOGING OF RNAs

partial sequence veriication by hybridization methods (including polymerase chain reaction [PCR] and DNA arrays). Immunochemical methods such as enzymelinked immunosorbent assay (ELISA) and others can be used for bacterial identiication, however it is not clear that enough organism-speciic antigens exist [22]. Furthermore, no generalized informational database can be leveraged to develop new immunological assays. On the other hand, the ever-increasing amount of sequence data from bacterial organisms has made nucleic acid-based identiication increasingly attractive. A comprehensive review of current molecular diagnostic techniques is beyond the scope of this chapter; however, a list of useful references for varied techniques has been provided [23–37]. Of particular interest in the identiication of bacterial species is the analysis of gene sequences encoding ribosomal RNA (rRNA). Due to evolutionary pressure to maintain a functioning 3D structure, rRNA sequences are highly conserved. Nevertheless, they also include local variable regions that are highly characteristic of the organism in which the ribosome resides. In prokaryotes, three RNAs comprise the ribosome—5S, 16S, and 23S rRNA. Since the early demonstration that phylogenetic information is obtainable from cataloging of 16S rRNA sequences, comparison of rRNA sequences has prevailed as the best single molecule for use in the phylogenetic characterization of nonviral organisms [38–42]. Due to this well-demonstrated utility in determining evolutionary relatedness, 16S rRNA gene sequences (rDNA) comprise the largest set of gene-speciic data for bacteria with over 1,000,000 sequences now publicly available [43–45]. The most straightforward PCR assays rely on identiication of at least one highly speciic primer or internal reporter probe, and therefore, are generally limited to identiication of a small number of species or even strains. Speciicity is therefore achieved at the cost of identifying a speciic probe or primer for every organism of interest. By their very nature, such sequences are derived from the most variable (organism-speciic) regions of the respective genomes. Therefore, these sequences are most sensitive to loss of speciicity or false-negative results due to mutation [46]. For instance, a TaqMan assay (Life Technologies, Carlsbad, CA) for West Nile Virus failed to detect 47% of possible singlenucleotide variations in the probe-binding site and was unable to detect any targets with more than two mutations [47]. Most DNA microarray designs follow the same approach of “probing” with highly organismspeciic sequences (albeit in a highly multiplexed fashion). While promising, most microarray experiments require at least 24 h per sample to obtain results. Microarray experiments would require enormous

amounts of equipment to parallelize as well as require the identiication of new probes as new strains emerge. Also, standardization is dificult due to the large number of steps involved [32–36, 48–50]. Finally, restriction fragment length polymorphism (RFLP) or conformational differences in universally ampliied PCR products can be used to distinguish species [51–52]; in all of these cases subsequent analysis by gel electrophoresis is required, the results of which can be dificult to standardize within and between laboratories [14, 17]. Complete sequence determination, alignment, and comparison of conserved genes such as 16S rRNA has long been the standard for ultimate determination of genetic relatedness. With the onset of genomics it is now becoming commonplace to develop phylogenies based on a concatenation of multiple genes. Unfortunately, even the fastest sequencing separations are time consuming compared with mass spectrometry, typically requiring 24 h for results [53–54]. Sequencing (and hybridization discussed above) requires a means for radioisotope, enzyme, or luor labeling, and for complex microbial mixtures, cloning of amplicons is required for “sorting” of the resultant library. Finally, with regard to sequencing, highly parallel array sequencers [55] and other next-generation sequencing approaches are now becoming available. These systems will produce prodigious amounts of data on a modest number of samples but with a 24–72 h turnover time, and they do not meet the rapid response needs of medical diagnostics or many other applications. Compared with most analytical techniques, mass spectrometry is generally faster, providing both separation of complex mixtures and a fundamental measurement simultaneously. Due to this analytical speed, mass spectrometry is also amenable to high-throughput analysis of very many samples. Therefore, mass spectrometry represents an important approach for analysis of biological exposure [56]. In addition to clinical applications, diagnostics are also needed for food and water safety, bioreactor analysis/sterility assurance, and environmental microbiology. 4.1.3 Mass Spectrometric Approaches to Sequencing and Compositional Characterization The number of different sequences using a four-letter alphabet (A, C, G, and T or U) increases as 4n, where n is the number of bases in the sequence. The number of different compositions is signiicantly less, (n + 4 − 1)!/ [n!(4 − 1)!] (i.e., the number of combinations of 4 elements chosen n at a time with replacement) [57]. For instance, the number of unique compositions (and hence unique fragment masses) for the complete set of possible 10-mers is 13!/(10! × 3!) or 286, much less than the 410 = 1,048,576 unique sequences. Despite this degener-

INFORMATICS: ENABLING ASPECTS OF 16S RRNA AND DATABASE CONSTRUCTION

acy, a number of methods for characterization of nucleic acids based on mass spectrometric analysis have been proposed. For discrimination of organisms or viruses based upon compositional information alone, either extremely accurate mass measurement of a large sequence region is necessary, or less precise measurement of multiple informative fragments is required. For instance, Ecker et al. have employed electrospray ionization Fourier transform ion cyclotron resonance (ESI/FTICR) to successfully type strains of streptococci by analyzing PCR products of ∼80–140 bp for their total base compositions (i.e., AwCxGyTz) [58–59]. Unfortunately, the very high resolution required for unambiguous compositional assignment (±1 ppm) of such large molecules requires instrumentation at least an order of magnitude more costly than a matrix-assisted laser desorption/ionization time-of-light (MALDI-TOF) instrument and out of reach for many laboratories. Instead, it may be more promising to introduce a fragmentation step, which will reduce the resolution requirements while retaining valuable information. Among mass spectrometric methods, MALDI-TOF mass spectrometry is the method of choice for measuring the mass of oligonucleotides, especially mixtures thereof [60]. The MALDI process results from absorption of laser light excitation by an aromatic organic acid and transfer of that energy to the cocrystallized analyte. Once ionized, the desorbed plume of matrix and analyte is accelerated in a vacuum through a long light tube to a time-of-light (TOF) mass analyzer. The light time of the analyte is proportional to the square root of the mass-to-charge ratio. When matrix-assisted laser desorption/ionization (MALDI) analysis of nucleic acids irst came into practice, it was quickly seen as a possible substitute for nucleic acid sequencers due to the rapidity of the separation versus capillary or gel electrophoresis. While MALDI-TOF mass spectrometry has previously been used for chain-termination sequencing [61–65], the maximum read length using such an approach is ∼56 nucleotides [61]. Consequently, there is now greater focus on fragmentation. For example, the endoribonuclease RNase T1 can be used to completely digest a single-stranded RNA after each G residue. Figure 4.1 contains a schematic representation of the resulting spectrum of a hypothetical 19-mer digested after every G residue. In contrast to a Sanger-type spectrum which would have 19 peaks, only four, lighter mass fragments are observed. The tradeoff, however, is that sequence information is lost. Despite this loss in sequence information, microbial identiication based upon base-speciic fragmentation patterns appears extremely promising. While successes have been reported [66–70], to our knowledge, ours is the irst work to perform systematic calculations

% 1 2 3 4 Intensity

87

CCCUUG/AUAG/CCG/CUACG + RNase T1 1. CCG 2. AUAG 3. CUACG 4. CCCUUG

m/z

FIGURE 4.1 Schematic representation of MALDI spectrum for a base-speciic digest of a hypothetical 19-mer RNA oligonucleotide.

assessing the absolute identifying power or classifying utility of base-speciic fragments from large sequence data sets [71–72], and no generalized commercial assay is yet available using such technique. More importantly, we have repeatedly demonstrated the feasibility of the base-speciic cleavage technique and developed the necessary computing infrastructure and methods to support routine sample identiication [71–74].

4.2 INFORMATICS: ENABLING ASPECTS OF 16S RRNA AND DATABASE CONSTRUCTION MALDI generates singly charged ions and is widely regarded as the mass spectrometric technology most applicable to analyzing the types of mixtures generated by nucleic acid fragmentation. In fact, the experimental feasibility of using a fragmentation strategy in combination with MALDI has already been established. For example Hahner et al. have described endoribonuclease digestion for the generation of RNase T1 fragments and MALDI characterization [66]. Hartmer et al. modiied the approach and applied it to the discovery of single nucleotide polymorphisms (SNPs) and identiication of bacteria using 16S rRNA regions [69].Von Wintzingerode et al. showed MALDI of base-speciic fragmentation patterns of 16S rDNA amplicons to be a viable method for microbial identiication and compared experimental to predicted masses for Bordetella species [67], and more recently Lefmann et al. have used base-speciic cleavage to discriminate mycobacteria [70]. Despite these demonstrations, our group presented the irst systematic calculations to assess the absolute identifying power or phylogenetic classifying utility of matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) when applied to RNase T1 or RNase A-generated fragments from large 16S rRNA sequence data sets [71–72].

88

4.2.1

METHODOLOGIES FOR IDENTIFYING MICROORGANISMS AND VIRUSES BY MASS CATALOGING OF RNAs

Hypothesis and Underlying Issues

We hypothesized that there is suficient discriminatory power in base-speciic rRNA fragment compositions alone for identiication of known and previously unknown organisms [71–72]. In order to investigate the occurrence of identifying compositions in a large data set, computational, base-speciic fragmentations of a large number of 16S rRNA sequences were performed and the identifying power of the fragments as distributed among the organisms were investigated. Preliminary issues involving genetic alphabet, information degeneracy, isotopes, and instrument resolution were considered. With regard to alphabet, obvious considerations are whether DNA or RNA would be used and whether the fragments observed would be single or double stranded. In nature, mass-modiied nucleotides are occasionally present in both DNA and RNA. However, posttranscriptionally modiied nucleotides in rRNA are few. Large fragments containing them are rare, and they are typically universal to large groupings of bacteria. Therefore, speciic modiication-produced masses are known [75]. Finally, such modiied bases are absent from in vitro runoff and ampliication products. Thus, it was only necessary to consider the usual four-letter alphabets (A, C, G, and T or A, C, G, and U). Compared with the traditional catalogs of oligoribonucleotide sequences, a catalog of fragment masses suffers from degeneracy due to the fact that multiple distinct permuted sequences generated by RNase digestion (e.g., AUCCG and UACCG) have the same composition, and hence, the same mass. This many-to-one relationship results in a loss of information potentially useful for microbial identiication. For RNA targets, this problem is exacerbated by the small (1 Da) mass difference between C and U, raising the question of whether it is possible to distinguish RNase T1 oligomers differing only in the number of Cs and Us. Unequivocal determination of oligoribonucleotide composition based on mass alone is obviously dependent on the resolution and accuracy of the mass spectrometer. Koomen et al. have published an extensive review of the requirements for accurately determining oligonucleotide compositions from measured mass [76]. They determined that theoretically, all monoisotopic compositions of DNA up to 13-mers could be accurately assigned at 5 ppm mass accuracy or better. Operating in relectron mode, employing proper sample preparation techniques, and including internal calibration standards, they were able to obtain 6 ppm or better accuracy on samples ranging from 5 to 13-mers. An accuracy of 6 ppm, for example would allow assignment to within 0.0234 Da at 3900 Da, allowing discrimination of a U to C difference in composition in a 50% purine oligoribo-

nucleotide with a length of approximately 13. MALDITOF mass spectrometry operated in linear mode with no added internal standards is a lower resolution, lower accuracy technique, typically yielding a resolution of m/Δm 500–1000 and accuracy of 0.05%–0.1% (500– 1000 ppm) [76]. To illustrate, at an accuracy of 1000 ppm on 5000 Da (roughly the mass of a single-stranded 16mer), a compositional assignment may be in error by 5 Da. This error suggests that, though any substitution involving a purine is readily detectable, up to ive substitutions of U for C or vice versa might be indistinguishable within this hypothetical 16-mer. To address this issue experimentally, studies relying on 16S rRNA RNase T1 fragments will either need to be done at the highest accuracy and resolution routinely attainable or an alternative protocol will be needed wherein the mass distinction between C and U is enhanced. Alternatively, a statistical approach based on the simultaneous observation of multiple mass fragments observed in combination may be able to unambiguously identify a microorganism even if the C/U distinction is not readily made. Oligoribonucleotides generated by RNase A will internally contain only G and A, which are readily distinguished, and hence, would not present a problem. A core question that faced the original RNase T1 fragment cataloging approach was the oligoribonucleotide length at which a match between two organisms relected actual sequence similarity as opposed to a chance occurrence. The number of fragments that exist for any length, n, increases as 4n. RNase T1 oligoribonucleotide digestion products, however, always contain a single G at the 3′ end of the sequence and no internal Gs. Hence, the number of possible products is 3(n−1), which for length six is 243 products (in contrast to 46 = 4096 possible 6-mers). This is signiicantly larger than the number of 6-mers generated by digestion of any single 16S rRNA, and hence an identical match of 6-mers (or any longer fragments) between two 16S rRNAs is likely to relect actual sequence identity rather than chance. This result was borne out by subsequent examination of complete sequences. However, the number of different mass compositions is signiicantly less than the number of unique sequences. Mass compositions are determined by the number of 4 elements chosen n at a time with replacement:  4 + n − 1 (n + 3)! 4 choose n (with replacement) =   = n ! 3!  n  (n + 3)(n + 2)(n + 1) , = 6 (4.1) where ! denotes factorial [57]. For instance, the number of unique compositions for the complete set of possible

INFORMATICS: ENABLING ASPECTS OF 16S RRNA AND DATABASE CONSTRUCTION

10-mers is 13!/(10! × 3!) or 286, which is much less than the 410 = 1,048,576 unique sequences. Because an RNase T1 oligoribonucleotide will always end in G, the number of possible compositions of length n can be expressed as: (n + 1)! n(n + 1) , = (n − 1)! 2 ! 2

(4.2)

where n is still the full length of the oligoribonucleotide. RNase A fragments will end in either U or C and contain only preceding internal As or Gs. The number of possible compositions (not sequences) in this case can be expressed as: 2n! = 2 n. (n − 1)!

(4.3)

4.2.2 Mass Cataloging of Bacterial 16S RNA Oligoribonucleotides In order to address the issues above regarding the occurrence of 16S-derived oligonucleotide fragments, isotopic distribution, required resolution, and organism discrimination, we examined attributes of virtual digests of a large number of 16S rRNA sequences [71]. 16S rRNA sequences from 7322 bacterial organisms were obtained from RDP Release 7.1 [43]; 1921 16S rRNA sequences met our selection criteria for length and quality and were used to generate endoribonuclease digestion products. Table 4.1 contains summary statistics of complete endoribonuclease digestions of the 1921 selected 16S rRNAs [71].

TABLE 4.1 Sequences

Attributes of the oligoribonucleotide catalog Length Range

Without isotopes: RNase T1 1–54 RNase A 1–21 With isotopes (see text): RNase T1 1–54 RNase A 1–21 b c d e f

As can be interpreted from Table 4.1, since RNase T1 cuts 16S rRNA less frequently than RNase A, a smaller number of oligoribonucleotides with a greater average length are generated. These longer oligoribonucleotides have more unique sequences and accordingly potentially unique masses. As a result, a mass catalog generated with RNase T1 digestion is more informative and thus more useful for bacterial identiication than an RNase A catalog. Isotopic distributions of fragment masses were also calculated. In this calculation, only isotopes of carbon and oxygen were considered and only peaks of more than 50% maximum relative intensity were cataloged. Table 4.1 also shows, as expected, that inclusion of isotopic elements increases the number of masses in the catalogs without changing the number of distinct oligoribonucleotide alphabetic compositions. While our group has performed a number of other calculations in addition to those presented in Table 4.1, a primary result of this ground-laying work was the determination of the minimal informative length of base-speciic fragments for organism discrimination [71]. This is exempliied by Figure 4.2. Speciically, it was found that oligonucleotide fragments generated by either RNase A or T1 digestion of length 5 or less are simply uninformative as they are shared by virtually all organisms. In contrast, at 6-mers or longer, there begin to occur fewer masses in the average organism’s catalog than theoretically possible. Thus, any ribonucleotide fragment of length 6 or longer (mass greater than approximately 1800 Da) is likely to be helpful in discriminating bacterial organisms from one another [71]. Additionally, at least among the 1921 rRNA sequences examined, a number of organisms could be uniquely identiied by the observance of just one, two, or three

Comparison between RNase T1 and RNase A 16S rRNA Catalogs Resulting from Virtual Digestion of 1921

Enzyme

a

89

a

b

Total Oligos

Distinct Oligosc

Distinct Massesd

898,494 1,225,481

8,601 1,994

858 227

128 81

77 50

898,494 1,225,481

8,601 1,994

2,404 644

128 81

159 88

Average Oligo Sequences per 16Se

The minimum and maximum length of oligoribonucleotides in catalogs. The number of all oligoribonucleotides generated by complete RNase digestion. The number of different oligoribonucleotide sequences in the catalogs. The number of different oligoribonucleotide masses in the catalogs. The average number of different oligoribonucleotide sequences generated by every 16S rRNA RNase digestion. The average number of different oligoribonucleotide masses generated by every 16S rRNA RNase digestion.

Average Masses per 16Se

90

METHODOLOGIES FOR IDENTIFYING MICROORGANISMS AND VIRUSES BY MASS CATALOGING OF RNAs 300 Actual 16S rRNA RNase T1 catalog Theoretical RNase T1 oligi set Actual 16S rRNA RNase A catalog Theoretical RNase A oligi set

Unique mass count

250

200

150

100

50

0 0

5

10

15

20

25

30

Oligoribonucleotide length

FIGURE 4.2 Comparison of numbers of unique polyisotopic oligoribonucleotide masses actually occurring in digests of 16S rRNA versus possible masses. The number of unique polyisotopic oligoribonucleotide masses in the actual RNase T1 and A catalogs are presented. The unique polyisotopic oligoribonucleotide masses in the theoretical sets are calculated from all possible RNase T1 and RNase A oligos and with consideration for the natural isotopic distribution. Only carbon and oxygen isotopes and the resultant oligoribonucleotide masses above a 50% maximum relative intensity are considered. The counts of RNase T1-generated oligoribonucleotide more than 26 nt long are not shown. Source: Zhang et al. BMC Bioinformatics 2006 7:117. doi: 10.1186/14712105-7-117 [71].

unique RNase T1 oligonucleotide masses [71]. Based on this promising theoretical groundwork, entire patterns of fragment masses universally obtainable from bacteria by utilization of PCR primers targeted at highly conserved regions of the 16S rRNA gene (16S rDNA) were quantitatively scored [72]. 4.2.3 Phylogenetic Trees Based on Mass Spectrometric Observables To further examine the discriminatory power of an RNase T1 cataloging approach by mass spectrometry, mass catalogs for a large number of organisms were generated and compiled into a structured, relational database. Genetic afinities deduced from observable mass fragment spectra were compared with phylogenies based on conventional 16S rRNA sequencing using neighbor joining [77] to construct distance-based trees. This approach allowed the extent to which unrelated organisms might be incorrectly identiied by mere coincidence of mass spectral patterns to be examined. Also, the degree (i.e., family, genus, or species level) to which the genetic afinity of various organisms may be resolved could be determined. Using a straightforward spectral comparison metric described below, distance matrices were derived. Neighbor-joining trees are then constructed to obtain insight into the limit of resolution of

the method when a single cleavage reaction is used. These in silico analyses are based on an experimental protocol which we have now used hundreds of times. First, universal primers (appended with 5′-RNA polymerase promoters) are used to amplify a homologous sequence region of the 16S rDNA from the organism(s) in a sample. Next, the DNA amplicons are transcribed to RNA and subjected to base-speciic fragmentation. The mass of the resulting products is then determined by MALDI-TOF mass spectrometry. Finally, measured masses are cataloged and compared with mass databases derived from rRNA sequence databases to determine the genetic afinity of the sample organism. 4.2.4

Comparison of Mass Spectra

In order to quantitatively intercompare mass spectral “ingerprints” produced by base-speciic fragmentation, we formulated the scalar or inner product deined by Equation 4.1. We deine a scalar product (often referred to as a “dot product”) of two mass spectra as: N1

M, M ′ = M ⋅ M ′ ≡

N2

∑ ∑ δ(m − m′ ), i

j

(4.4)

i =1 j =1

where mi are the masses of each of the N1 individual fragments in the spectrum for species 1 and m′j are the

INFORMATICS: ENABLING ASPECTS OF 16S RRNA AND DATABASE CONSTRUCTION

masses of each of the N2 fragments for species 2, and δ is the discrete (Kronecker) delta function deined as: k=0 1 δ(k ) =  . 0 otherwise 

(4.5)

It can be easily veriied that the following commutative, distributive, and positive deiniteness conditions for an inner product are satisied: M1 ⋅ M 2 = M 2 ⋅ M1 , (α 1 M 1 + α 2 M 2 ) ⋅ M 3 = α 1 M 1 ⋅ M 3 + α 2 M 2 ⋅ M 3 , M1 ⋅ M1 > 0∀M1 ≠ [0].

(4.6a–c)

to primer sequences shared by many organisms, primer pairs yielding amplicons of ∼500 bp or less are also of interest because the mass spectra acquired from such shorter regions will have minimal complexity. For example, transcription and subsequent RNase T1 cleavage of a 400 bp amplicon will typically yield only ∼22–25 distinct RNA fragment masses [71]. Table 4.2 contains a representative listing of primers which have been routinely employed in our laboratory for amplifying regions of 16S ribosomal DNA from a large fraction of all bacteria discovered thus far [72]. In particular, the “LaneAB” and “Lane-BC” primer pairs are routinely utilized to generate a ∼400 and ∼500 bp amplicon, respectively, from bacterial samples. 4.2.6

Using this inner product, we then deine the following metric or “coincidence function”:

cij = c(Mi , M j ) =

2 × Mi ⋅ M j . (Mi ⋅ Mi ) + (M j ⋅ M j )

(4.7)

This function provides a normalized (i.e., between 0 and 1) representation of the extent to which two spectra are similar. Using this metric, a coincidence (or similarity) matrix, C, with elements cij can be generated to tabulate the degree of similarity between the fragment catalogs of every pair of organisms. Likewise, a matrix of distances, D, with elements dij = (1 − cij) can be created, and used as input to conventional cluster analysis algorithms. 4.2.5 Universal Primers Directed at Bacterial 16S Ribosomal RNA (rRNA) Since the early demonstration that phylogenetic information is obtainable from catalogs of RNase T1 digests of 16S rRNA [38–42], comparison of 16S rRNA (or rDNA) sequences has become the most widely used single molecular method to characterize the phylogenetic position or genetic afinity of bacteria and archaea. Over 1,000,000 accessions containing partial or full rDNA sequences are available from the Ribosomal Database Project (as of RDP release 10.14, over 450,000 sequences are of length 1200 nucleotides or more [78]). Weisburg et al. [79] and Lane et al. [80] have previously reported “universal” primers yielding PCR products from various 16S rDNA (or rRNA) sequence regions for a large fraction of all known bacteria. When present, these primers yield almost full coverage of the approximately 1500 bp 16S sequence, and have thus been used routinely for phylogenetic study. In addition

91

Organisms of Greatest Interest

The National Institute of Allergy and Infectious Diseases (NIAID) has prioritized research on a number of organisms categorized as Category A, B, or C pathogens for the purposes of epidemic prevention and for biodefense [6–7, 81]. To examine to what extent mass spectrometric methods might resolve these organisms from one another (using the primer sets above), all the 16S rDNA sequences of all the bacteria listed in NIAID Categories A, B, or C were downloaded. In some cases an entire genus such as Salmonella, for example, is speciied. Based on initial experimental results, we expected to obtain at least genus-level resolution and, in many cases, species-level resolution of bacteria using masses cataloged from just a single cleavage reaction. Of course, since pathogenicity can arise from minor genomic changes (presence of plasmid, a single gene, changes in regulation, etc.), even complete 16S rRNA sequencing cannot always distinguish pathogenic strains from nonpathogenic strains. For example, pathogenic strains of Escherichia coli have the same rRNA sequence as nonpathogenic strains. Thus, the best possible outcome in these cases is to assign an unknown organism to a genus or species whose members include pathogens. With large databases of predicted mass-fragment catalogs in hand (typically ∼20–25 masses from a given “universally ampliiable” 16S rDNA sequence subregion), we calculated thousands of mass-spectral coincidences between all catalogs using Equation 4.7. We investigated the average value of the coincidence function at different phylogenetic levels for a number of organism lineages and compared coincidence values to the average entropy of several multisequence alignments. Finally, we used coincidence values to generate distance matrices and these distances were placed in a format acceptable for input into the freely available program MEGA 3.1 for generating neighbor-joining trees [82].

92 TABLE 4.2

Coverage and Amplicon Characteristics for the Adjacent Weisburg and Lane Universal Primer Sets in 47,257 Bacterial Sequences

Forward Primer

“Weisburg FWD” AGAGTTTGATCCTGGCTCAG AGAGTTTGATCATGGCTCAG AGAGTTTGATCCTGGCTTAG AGAATTTGATCTTGGTTCAG “Lane A” CAGCAGCCGCGGTAATAC CAGCAGCCGCGGTAATTC CAGCCGCCGCGGTAATAC CAGCCGCCGCGGTAATTC “Lane B” AAACTCAAAGGAATTGACGG AAACTCAAATGAATTGACGG AAACTTAAAGGAATTGACGG AAACTTAAATGAATTGACGG “Lane C” GTACACACCGCCCGT GCACACACCGCCCGT

Reverse Primer

Occurrence of Primer Pair in All Bacterial Sequences

Occurrence of Primer Pair in Enteric Sequences

Average Amplicon Length and Std. Dev. (All Bacteria)

“Lane A” CAGCAGCCGCGGTAATAC CAGCAGCCGCGGTAATTC CAGCCGCCGCGGTAATAC CAGCCGCCGCGGTAATTC

9,337/47,257 = 19.76%

333/1655 = 20.12%

518.78 ± 37.94 bp

“Lane B” AAACTCAAAGGAATTGACGG AAACTCAAATGAATTGACGG AAACTTAAAGGAATTGACGG AAACTTAAATGAATTGACGG “Lane C” GTACACACCGCCCGT GCACACACCGCCCGT

38,310/47,257 = 81.07%

1553/1655 = 93.84%

408.83 ± 7.35 bp

36,058/47,257 = 76.30%

1401/1655 = 84.65%

500.83 ± 13.94 bp

9,991/42,257 = 21.14%

373/1655 = 22.54%

123.62 ± 57.24 bp

“Weisburg REV” GGTTGGATCACCTCCTTA AAGTCGTAACAAGGTAACCGT AAGTCGTAACAAGGTAGCCGT AAGTCGTAACAAGGTATCCGT

INFORMATICS: ENABLING ASPECTS OF 16S RRNA AND DATABASE CONSTRUCTION

4.2.7 Generation of Trees Based on Separate Analysis of the “Lane-AB” or “Lane-BC” RNase T1 Mass Catalogs Based on the initial results above and the desire to create a “universal” bacterial assay, we decided to limit further analysis to fragment masses derived from only the “Lane-AB” or “Lane-BC” amplicons. Figure 4.3 contains a neighbor-joining tree of all explicitly named

93

NIAID Category A, B, or C bacterial pathogens as resolved by RNase T1 fragmentation of a virtual transcript from the Lane-AB amplicon and spectral distances derived by our coincidence function. Figure 4.3 indicates that all of the Category ABC pathogens listed are resolved from each other at the genus level by an RNase T1 mass spectrum of the Lane-AB sequence region of 16S rRNA.

FIGURE 4.3 Neighbor-joining tree of all explicitly named NIAID Category A, B, or C bacterial pathogens as resolved by basespeciic fragmentation of the “Lane-AB” amplicon and spectral distances derived from the presented “coincidence function.” Separation of some of the unresolved clusters may be improved by further mass spectrometric analysis of the “Lane-BC” sequence region.

94

METHODOLOGIES FOR IDENTIFYING MICROORGANISMS AND VIRUSES BY MASS CATALOGING OF RNAs

Of course, many of the near-phylogenetic neighbors to these pathogens were not included in the analysis shown in Figure 4.3, so further analysis of each of the presented clusters including near-neighbors is warranted. To investigate more closely the extent to which the Category ABC pathogens might be resolved from their innocuous relatives using just a single base-speciic cleavage reaction, the methods described above were used to analyze entire genera including pathogens of interest and their close relatives. Figure 4.4 illustrates that 20 Vibrio

cholerae strains can be distinguished based on an observable Lane-AB RNase T1 fragmentation pattern in the context of 457 other “universally” ampliiable Vibrio strains. For clarity, only V. cholerae strains are labeled by strain name in Figure 4.4. As can be seen, none of the 20 V. cholerae strains examined was found in branches containing any of the other 457 Vibrio species or strains under consideration. In other words, the pathogen is theoretically resolvable from its nearest known genetic neighbors by our mass spectrometric approach.

FIGURE 4.4 Demonstration of successful clustering of the pathogen Vibrio cholerae among 477 members of the genus Vibrio by mass spectrometric observables. The presented neighbor-joining (NJ) tree is based on mass-spectral coincidence analysis of the Lane-AB 16S rDNA sequence region of the genus Vibrio. Only the names of 20 strains of the pathogen V. cholerae are displayed out of 477 ampliiable Vibrio species or strains were used for generation of the NJ tree. The mass spectrometric observable distances from near-neighbors presented may be used to determine the limit of discrimination for a given pathogen and basespeciic cleavage reaction.

INFORMATICS: ENABLING ASPECTS OF 16S RRNA AND DATABASE CONSTRUCTION

4.2.8

Flavivirus Assay Development

While the number of applications for a universal bacterial genotyping system is quite large, we also sought to demonstrate feasibility of viral identiication. Of course, viruses are more genetically diverse, and the ability to ind pan-viral PCR primers which can be used to create a predictable product can vary from viral family to family. The ability to adapt our system to the broad viral genus Flavivirus was assessed [74]. The genus Flavivirus includes more than 70 singlestranded RNA viruses responsible for severe encephalitic, hemorrhagic, hepatic, and febrile illnesses in humans and other vertebrates [83]. Among the pathogens in this genus are the yellow fever, tick-borne encephalitis, Japanese encephalitis, St. Louis encephalitis, West Nile, and dengue viruses. Together, they cause considerable morbidity and mortality worldwide. Of particular interest are the four serotypes of the mosquito-borne dengue virus (DEN-1, -2, -3, and -4) that can cause dengue hemorrhagic fever (DHF) and dengue shock syndrome (DSS). Currently, there is no protective vaccine or speciic treatment available, and it is estimated that over 100 million cases of dengue occur annually worldwide [84–85]. Typing of dengue is of critical importance in order to distinguish endemic strains from new outbreak strains so that new outbreaks can be readily contained. Typing is also important in treatment because infection by a new serotype in a patient previously infected by one of the other three is associated with a greatly increased risk of developing DHF and/or DSS [86–87]. The current reference method for identifying and typing the various laviviruses is isolation of the virus in cell culture followed by immunoluorescence typing [88]. This time-consuming procedure requires signiicant expertise and cell culture facilities that may not be readily available. Alternative serological tests are relatively easy to use and can accommodate a large number of samples, both necessities when confronting an epidemic. These beneits, however, come at a cost; tests such as hemagglutination inhibition, immunoglobulin G enzyme-linked immunosorbent assay (IgG-ELISA), and immunoglobulin M antibody-capture enzyme-linked immunosorbent assay (MAC-ELISA) cannot easily distinguish dengue at the serotype level and are likely to misidentify other laviviruses as dengue [89–91]. In addition, in areas where a virus is endemic, an outbreak strain will need to be distinguished from any endemic strain which might belong to the same serotype. To adapt our mass cataloging approach from bacteria to lavivirus, “almost-universal” primer pairs for laviviruses in general and all four serotypes of dengue in particular were derived from the literature [83, 92–94].

95

The coverage of each primer pair was evaluated against large sequence data sets. The sequences that would be expected to be ampliied were “digested” with RNase T1 in silico, and the expected mass catalogs were calculated. The coincidence function described above was used to calculate distance matrices. Subsequently, cluster analysis was used to determine which viruses and serotypes would be expected to be resolvable. Based upon the successful clustering of the various laviviruses and serotypes of dengue in these predictive calculations, the approach was experimentally veriied on representative strains of DEN1 and DEN4. Scaramozzino et al., Drosten et al., and others have identiied primer sets useful for both broad and speciesspeciic ampliication of laviviral sequences [83, 92–94]. These primer sets were tested and extended as follows. For broad lavivirus ampliication, a large set of lavivirus complete genomes, partial sequences, and sequences for the ifth nonstructural protein NS5 were collected into a database. A CLUSTALW multiple sequence alignment [95] was performed and the degenerate positions were compared with the forward (“MAMD”) and reverse (“cFD2”) primers reported by Scaramozzino et al. targeted at the NS5 coding region in most laviviruses. Based on our alignments, variants of these degenerate forward and reverse primers were chosen. Speciically, the literal search strings or in silico “primers” (using standard International Union of Pure and Applied Chemistry [IUPAC] base degeneracies; N = A, C, G, or T; D = A, G, or T; H = A, T, or C; M = A or C; R = A or G; S = G or C; Y = C or T) were “ScaraFWD,” 5′-ACATSATGGGNAARMGDRAR-3′, and “ScaraREV,” 5′-GAYACMGCHGGHTGGGA-3′. (The latter reverse primer sequence must be reversecomplemented for actual PCR). As of July 2007, both primer sequences were simultaneously present in 287 out of 383 or 75% of complete lavivirus genomes downloaded from GenBank (Entrez nucleotide) [45]. In practice, these primers could be used under conditions of “permissive” or “mismatch-tolerant” PCR (or reverse transcription polymerase chain reaction [RT-PCR]), further extending their coverage. Similarly, the following dengue-speciic primers described by Drosten et al. [92] and targeted at the 3′ noncoding region of the dengue genome were used to create a database of expected base-speciic cleavage masses from the parent sequence database: DrostFWD, 5′-AGACC AGAGATCCTGCTGT-3′, and DrostREV, 5′-RCGCC RSAARATGGAWTG-3′. The sense-strand sequences of the resulting virtual amplicons were then fragmented by cleavage after each guanosine residue to simulate an RNase T1 digest of RNA transcribed from the cDNA product. As in previous work in which phylogenetic trees were developed

96

METHODOLOGIES FOR IDENTIFYING MICROORGANISMS AND VIRUSES BY MASS CATALOGING OF RNAs

Dengue type 1

Dengue type 3

Powassan, Tick-borne encephalitis, Karshi

Dengue type 4 Yellow Fever

Dengue type 2

Dengue type 3

Murray Valley

West Nile

Japanese encephalitis

FIGURE 4.5 Cluster analysis based on mass spectral patterns of laviviruses ampliied by universal PCR of the NS5 coding region and fragmented by RNase T1.

based on mass-spectrometric observables [72], a matrix of distances (1 – coincidence values, Equation 4.7) between the virtual mass spectra of all the ampliied laviviruses was constructed. As in the bacterial analyses above, the pairwise distances were then entered into MEGA v3.1 (Molecular Evolutionary Genetics Analysis) [82] to generate trees using the neighborjoining algorithm. Figure 4.5 contains the result. Despite some information loss in moving from analysis of primary sequences to compositions (masses), the viruses cluster quite well with no strains falling into the wrong serotype groups. The Scaramozzino primers were intended to identify a homologous genomic region from all members of the genus Flavivirus. The same methods for mass fragment-based cluster analysis have also been applied using the dengue-speciic primers described by Drosten et al. for amplifying the 3′ noncoding region from all four serotypes of dengue. When targeted at our

database of 383 complete lavivirus genomes, only dengue strains were “ampliied,” verifying the dengue speciicity of the Drosten primers.

4.3 EXPERIMENTAL METHODS: PCR, TRANSCRIPTION, AND ENZYMATIC FRAGMENTATION Regardless of sample type, the irst step in our process is nucleic acid ampliication. The “prestep” is therefore nucleic acid extraction and/or puriication and is highly dependent upon the sample type and matrix (blood, saliva, dirt, etc.) in which the organism(s) of interest is (are) found. Thus, nucleic acid sample preparation is somewhat outside the scope of our methods, but numerous kits and protocols are commercially available for various applications. Subsequent steps comprising PCR,

EXPERIMENTAL METHODS: PCR, TRANSCRIPTION, AND ENZYMATIC FRAGMENTATION

transcription, enzymatic fragmentation, and spotting of MALDI sample are described following a list of materials below. Materials Treatment with Exonuclease I PCR product Exonuclease I RNase-free water Transcription RNase-free water Linearized DNA template (0.05–1 µg) AmpliScribe T7 Flash 10X Reaction buffer (Epicentre, Madison, WI) 100 mM adenosine triphosphate (ATP) 100 mM cytidine triphosphate (CTP) 100 mM guanosine triphosphate (GTP) 50 mM Aminoallyl-UTP (Fermentas, Hanover, MD) 100 mM dithiothreitol (DTT) AmpliScribe T7 Flash Enzyme solution Ribonuclease Digestion Transcript 3-HPA Ribonuclease T1 MALDI ZipTip™ (Millipore, Billerica, MA) Wetting solution: 50% acetonitrile (ACN) in nuclease-free water Equilibration solution: 0.1 M triethylammonium acetate (TEAA), pH 7 Wash solution 1: 0.1 TEAA, pH 7 Wash Solution 2: Nuclease-free water Matrix: 9 parts of 50 mg/mL of 3-hydroxypicolinic acid (3-HPA) in 50% ACN/Water, 1 part of ammonium acetate in nuclease free water 4.3.1

97

standard PCR reaction (5 min denaturation at 95°C, 30 cycles of 95, 55, and 72°C for 30, 30, and 45 s, respectively, followed by a 7-min extension at 72°C totaling approximately 105 min). Following PCR, reaction mixtures are treated directly with 1 µL (20 units) of DNA exonuclease I (Epicentre) at 37°C for 5 min to digest any unincorporated single-stranded primers. Without this step, we have occasionally observed low yields of RNA transcript, presumably due to interaction of the RNA polymerase with primers or primer dimers instead of the double-stranded PCR product. Exonuclease I is then deactivated at 80°C for 5 min. 4.3.2

RNA Transcription

Approximately 2 µL (typically ∼1 µg DNA) of the resulting mixture was then used directly without puriication as template for in vitro transcription for 30 min using a T7-lash™ kit (Epicentre; see the “Materials” list) containing the manufacturer’s suggested T7 RNA polymerase and nucleotide concentrations, except that aminoallyl-UTP was used as a 100% substitute for the natural UTP substrate, thereby improving the cleavage product discrimination of the experiment. 4.3.3 Ribonuclease Digestion and Spotting for MALDI Finally, the transcripts were completely digested by RNase T1 at 37°C for 10 min with MALDI matrix (3HPA) added as a denaturant. Briely, 5 µL of RNA transcript (up to 20 µg) is added to 4 µL 3-HPA in 50% ACN/water and 1 µL of RNase T1 (1000 units) and reacted for 10 min and then placed on ice for MALDI preparation. For highest quality spectra, samples are desalted by reverse-phase puriication in ZipTips according to manufacturer’s directions for nucleic acid puriication [96]. The inal step in this process is elution of puriied RNA oligonucleotide fragments onto the MALDI target using 2 µL of the MALDI matrix itself. Samples (MALDI “spots”) are allowed to air dry or may be rapidly dried under vacuum.

Polymerase Chain Reaction

For subsequent PCR, the forward universal primers employed are generally 5′-appended to allow for subsequent T7 “runoff” transcription back to RNA. The T7 RNA polymerase is 5′-taatacgactcactataagg PRIMERSEQ-3′ (where lowercase indicates T7 RNA polymerase promoter). The reverse primer is generally 5′-appended with one or more sequences encoding masses for internal calibration of acquired mass spectra. Such an approach also serves as a conirmation that the RNA transcription was full length. For bacterial samples, typically 25 ng of total DNA is used as template in a

4.3.4

MALDI Acquisition

Mass spectra of the digests have been successfully acquired in linear, negative-ion mode on both a Voyager DE-STR MALDI-TOF (horizontal instrument,Applied Biosystems, now Life Technologies) as well as a Voyager DE-PRO (vertical instrument). These are some of the most commonly available MALDI instruments on the market, and all protocol described here should be applicable to instruments of other manufacturers. Figure 4.6 contains a typical spectrum acquired for two different bacterial samples using the “Lane-AB” primer pair as described above.

98

METHODOLOGIES FOR IDENTIFYING MICROORGANISMS AND VIRUSES BY MASS CATALOGING OF RNAs (A)

% Intensity

100 90 2072.24 80 70 1992.67 60 2016.67 50 40 30 1966.57 2042.40 2090.52 20 2008.54 2121.01 10 0 1900 2160

2409.05

Pseudomonas aeruginosa 2386.56

2737.85

2594.55

2269.71 2807.60 2344.90 2345.04 2502.50

3075.71

2483.11 2450.50

2658.73 2604.85

2420

2991.02 2962.79 3005.00

2766.49

2680

3094.30

2940

0 3200

Mass (m/z)

(B)

% Intensity

2.1E+4

100 90 80 70 60 50 40 30 20 10 0 1900

2738.44

9.5E+3 Vibrio proteolyticus

2016.64 2400.52

2045.25 2321.40

2372.44

2056.21 1969.58

2080.64 2066.23 2122.15 1994.77

2160

2628.43 2658.17

2344.55 2260.44

2325.15 2451.37

2420

2649.35

m/z

2680

2745.85 2754.50 2780.73 2831.63

2940

0 3200

FIGURE 4.6 Typical spectra acquired from the “Lane-AB” region of 16S rRNA from two different bacterial species.

All spectra are typically processed in an identical fashion as described in detail elsewhere [73], including noise iltering, mass calibration, centroiding, and thresholding. The resulting short peak lists are then subjected to automated comparison to large, predicted massfragment databases for the organism-type in question (bacterial or viral) and the PCR primers employed to derive the spectrum.

4.4 DATABASE COMPARISON AND RESULTS: TYPICAL RESULTS OF BACTERIAL AND VIRAL GENOTYPING For comparison of acquired masses to our database of predicted mass catalogs, the quantitative metric or “coincidence function” of Equation 4.7 was used. Typical results for ive other representative bacterial species are listed in Table 4.3 [73]. These results were obtained by spectral matching to a database representing over 47,000 bacterial species and strains. Note that in each case the result is in high agreement with that obtained by complete conventional sequencing (and BLAST of GenBank [97]) of the same genetic region. Similarly, high-idelity results for genotyping of lavivirus have been reported [74]. Figure 4.7 contains a typical spectrum acquired from a DEN1 sample using primers universal for the NS5 gene in laviviruses and

corresponds to the genomic region analyzed by the cluster analysis of Figure 4.4. For the purposes of universal Flavivirus identiication, an expanded database containing additional sequence information for the NS5 genes as well as complete lavivirus genomes was constructed. Comparing the spectral acquisition from Figure 4.7 to the database of predicted spectra for 338 laviviruses allowed a rapid, unambiguous typing of the dengue sample. Table 4.4 contains the identiication rank by coincidence analysis in the context of 338 other laviviruses strains that would have also been ampliied by the universal (Scaramozzino) primers. Identiication rank is based on normalizing the inal calculated mass spectral coincidence by the highest value. Notable in the table is that coincidence analysis with proper documentation gives a quantitative means to compare strains based on date, locale, and so on. The highest ranked 29 strains were all DEN1. The strains ranked between 30 and 90 alternated between DEN2 and DEN3 except for the conspicuous appearance of the Cambodian type 1 strain that ranked 37th. The “commingling” of the lower ranking types 2 and 3 can be explained at least semiquantitatively by reference to Figure 4.5, where it is noted that the DEN2 and DEN3 clusters are roughly equidistant from DEN1. Again by reference to the mass-based phylogenetic tree of Figure 4.5, West Nile virus and Japanese encephalitis are pre-

DATABASE COMPARISON AND RESULTS: TYPICAL RESULTS OF BACTERIAL AND VIRAL GENOTYPING

TABLE 4.3

99

Bacterial Genotyping of Five Different Species Compared with Conventional Sequencing

Comparison of Bacterial Identiication by Conventional Sequencing versus Base-Speciic Mass Spectrometric Coincidence Analysis Sequencing/BLAST

Score. “bits” (rank)

Mass spectrometric coincidence analysis

Combined AB/BC score (rank)

Sample: Escherichia coli (K-12, MG1655) E. coli K-12 MG1655 E. coli strain RW-29 16S . . . E. coli O157:H7 EDL933 E. coli 16S rRNA gene E. coli C2 16S rRNA E. coli 16S rRNA gene Escherichia albertii strain 10457 E. albertii strain 12502

930 (1) 930 (1) 930 (1) 930 (1) 930 (1) 930 (1) 930 (1) 930 (1)

E. coli: K-12: M87049 E. coli: K-12: U18997 E. coli: O157:H7: BA000007 E. coli: AU1713: AY043392 E. coli: L10328 E. coli: CCCO4: AF511430 Shigella lexneri 2a str. 301: AE005674 S. lexneri 2a str. 2457T: AE016989

0.476 (1) 0.476 (1) 0.476 (1) 0.476 (1) 0.476 (1) 0.476 (1) 0.476 (1) 0.476 (1)

Sample: Acinetobacter sp. (ATCC 33604) Uncultured Acinetobacter sp. 16S rRNA

759 (1)

0.303 (1)

Acinetobacter sp. H1 16S rRNA Acinetobacter sp. phenon 10 Uncultured clone ELB19-080 16S rRNA Acinetobacter junii 16S rRNA Acinetobacter sp. PAMU-1.11 Acinetobacter sp. phenon 3 Acinetobacter junii 16S rRNA

759 (1) 759 (1) 759 (1) 759 (1) 759 (1) 759 (1) 759 (1)

Uncultured bacterium: AY700608 (genus Acinetobacter) Acinetobacter calcoaceticus subsp. anitratus Acinetobacter sp. ATCC 31012: AF542963 Acinetobacter grimontii (T): AF509828 Acinetobacter johnsonii: 5B02: AF509831 Acinetobacter sp. PAMU-1.11: AB118222 Acinetobacter sp. ADP1: CR543861 Acinetobacter sp. ADP1: 93A2: AJ812656

0.298 (2) 0.288 (3) 0.283 (4) 0.267 (5) 0.254 (6) 0.248 (7) 0.248 (7)

Sample: Pseudomonas aeruginosa (ATCC 25102) 946 (1) P. aeruginosa gene for 16S rRNA 938 (2) Pseudomonas sp. pDL01 16S rRNA 938 (2) Pseudomonas sp. Bx1-1 938 (2) P. aeruginosa ATCC BAA-1006 938 (2) Pseudomonas sp. BWDY-42 16S rRNA 938 (2) P. aeruginosa partial 16S rRNA 938 (2) Pseudomonas sp. HY-7 16S rRNA 938 (2) Pseudomonas sp. LQG-3 16S

P. aeruginosa: AT10: AJ549293 Pseudomonas alcaligenes (T): LMG 1224T P. alcaligenes: M4-7: AY835998 P. aeruginosa: ATCC BAA-1006: P. aeruginosa: PAO1: AE004949 P. aeruginosa: ATCC 27853: P. aeruginosa: SCD-13: Pseudomonas sp. pDL01: AF125317

0.471 (1) 0.455 (2) 0.446 (3) 0.439 (4) 0.439 (4) 0.439 (4) 0.439 (4) 0.439 (4)

S. maltophilia: ATCC 19861T Uncultured beta proteobacterium: AF529323 P. aeruginosa: SCD-1: AF448038 P. aeruginosa: AF225956 Pseudomonas geniculata (T). ATCC 19374T MTBE-degrading bacterium PM1: AF176594 Uncultured beta proteobacterium: AJ422152 Uncultured bacterium: W33: AY770973

0.251 (1) 0.246 (2) 0.245 (3) 0.245 (3) 0.245 (3) 0.244 (6) 0.244 (6) 0.243 (8)

Vibrio sp. VI1067/44: X97989 Uncultured bacterium: PDC-OTU7 V. proteolyticus (T): ATCC15338T Vibrio alginolyticus. LA6: AF513447 Vibrio parahaemalyticus RIMD 2210633: O3:K6 Vibrio parahaemolyticus: ATCC 17802 Vibrio sp. NLEP97-1598: AF410778 Vibrio sp. NAP-4: AF064637

0.493 (1) 0.489 (2) 0.481 (3) 0.481 (3) 0.481 (3)

Sample (clone): Stenotrophomonas maltophilia 811 (1) S. maltophilia AY748889.1 811 (1) S. maltophilia AY748888.1 811 (1) S. maltophilia strain TKW2 811 (1) S. maltophilia strain B25R 811 (1) S. maltophilia strain B8R 811 (1) S. maltophilia DQ141193.1 811 (1) S. maltophilia AY360340.1 Uncultured bacterium clone PDB811 (1) OTU11 Sample: Vibrio proteolyticus (ATCC 15338T) V. proteolyticus (ATCC 15338T) Uncultured bacterium clone PDC-OTU7 Vibrio alginolyticus 16S rRNA Vibrio alginolyticus 16S rRNA Vibrio parahaemolyticus RIMD 2210633 Vibrio parahaemolyticus 16S rRNA Vibrio sp. NLEP97-1598 16S Vibrio sp. AB 16S rRNA gene

944 (1) 918 (2) 918 (2) 918 (2) 918 (2) 918 (2) 918 (2) 918 (2)

0.481 (3) 0.481 (3) 0.481 (3)

The top eight scores for each method are presented. Species-level “hits” of the sample organism for each method are shown in bold, regardless of rank.

100

METHODOLOGIES FOR IDENTIFYING MICROORGANISMS AND VIRUSES BY MASS CATALOGING OF RNAs 2353.03

100

4509.3

Relative Ion Abundance

90 Dengue Virus Type 1

80 70 60

1984.20 2048.30

50

2377.02 2802.14

40 30

2030.41

3675.76

1959.95 2122.26 1974.48 10 1940.87 2073.36

20

0 1906.0

3387.64

2441.14 2427.16 2450.95 2515.71 2352.8

3370.32 3450.68

2747.16 2876.31 3264.4

2799.6

3650.11

4004.97

3493.2

0 4140.0

m/z

FIGURE 4.7 Typical spectrum acquired for lavivirus identiication.

TABLE 4.4 Mass Spectral Typing of Dengue-1 against all other Flaviviruses by “C,” the Normalized Coincidence (Equation 4.7) ID Rank 1 1 1 4 25 29 30a 37b 64 90 109 109 117 129 158c 161 161 171 185 186 338

C

Virus Name

1.0000 1.0000 1.0000 0.9650 0.5938 0.5714 0.5313 0.5196 0.4454 0.3711 0.3571 0.3571 0.3090 0.2912 0.2184 0.1545 0.1545 0.1485 0.1456 0.0789 0.0742

Dengue type 1 rec. clone rDEN1mutF complete genome Dengue type 1 rec. clone rDEN1delta30 complete Dengue type 1 rec. clone rDEN1 complete genome Dengue type 1 genomic RNA complete genome strain 98901530 Dengue type 1 strain D1.Myanmar.059/01 Dengue virus type 1 genomic RNA complete genome Dengue type 3 strain BR74886/02, complete genome Dengue virus type 1 from Cambodia complete genome. Dengue type 2 strain New Guinea Dengue type 3 strain Singapore 8120/95, complete genome Tick-borne encephalitis strain MDJ-01 nonfunctional Edge Hill strain V366 NS5 protein gene partial cds. Usutu strain Vienna 2001 from Austria complete genome Yellow fever strain Ivory Coast 1999 complete genome Dengue type 2 strain Ven2 polyprotein gene complete cds. West Nile strain 02002684 NS5 gene partial cds. Kunjin gene for polyprotein (C prM E NS1 NS2A . . . Japanese encephalitis strain CC27-S8 complete genome Powassan 1427-62 NS5 gene complete cds. West Nile strain NY99-lamingo382-99 complete genome Japanese encephalitis FU strain complete genome

a

Highest ranking non-DEN1 strain. Lowest ranking DEN1 strain. c Lowest ranking dengue strain of any serotype. rec., recombinant; cds., coding region. b

dicted to have spectra that are the least similar to DEN1 and these are in fact ranked low in Table 4.4. As a benchmark to Table 4.4, we have conirmed agreement with conventional sequencing and BLAST searches of GenBank [74] (data not shown).

We also successfully typed DEN4 in the laboratory by our rapid mass cataloging approach [74]. In an analysis completely analogous to that in Table 4.4, Table 4.5 shows successful experimental typing of DEN4 using alternate PCR primers. Using the primers described by

REFERENCES

101

TABLE 4.5 Mass Spectral Typing of Dengue-4 with Dengue-Speciic Primers (from Drosten et al. [92]) by “C”, the Normalized Coincidence ID Rank 1 1 1 1 1 1 1 1 1 1 11 11 11 11 11 11 11 11 11 11 11 11 11

C 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 0.3000 0.3000 0.3000 0.3000 0.3000 0.3000 0.3000 0.3000 0.3000 0.3000 0.3000 0.3000 0.3000

Drosten et al. [92] in conjunction with base-speciic cleavage, 16 DEN4 strains all tied with a relative coincidence of 1.0. Serotypes 1, 2, and 3 all tied with much lower scores of 0.30.

4.5

Virus Name Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue Dengue

vector p4(Delta30) complete seq. type 4 vector p4 complete seq. type 4 strain ThD4_0485_01 complete genome type 4 strain Singapore 8976/95 complete genome type 4 strain 814,669 complete genome type 4 recombinant clone rDEN4del30 complete type 4 recombinant clone rDEN4 complete seq. type 4 recombinant clone 2Adel30 complete seq. type 4 recombinant clone 2A complete genome type 4 polyprotein precursor gene complete cds. type 3 strain Singapore 8120/95, complete genome type 3 strain BR74886/02, complete genome type 3 strain 80-2 complete genome type 3 isolate PF94/136116, complete genome type 2 vector p4(delta30)-4995 complete seq. type 2 vector p4(delta30)-4995 complete seq. type 2 vector p2(delta30) complete seq. type 2 strain ThNH81/93 complete genome type 2 strain New Guinea type 2 strain Cuba89/97 complete genome type 2 China isolate 04 complete genome type 2 (S1 vaccine strain) complete genome type 1 strain Fj231/04 complete genome

NCC-9-58 and its successor grant NNJ04HF34G) to RCW and GEF, the Institute of Space Systems Operations to GEF and a Small Business Innovation and Research Grant from the National Institutes of Health (2R44AI066425 to GWJ, BioTex Inc).

DISCUSSION REFERENCES

The methods utilized here have been developed to provide a broad-based or “universal” genotyping system for all bacteria. As we have demonstrated with the laviviruses, the approach is readily adapted to broad viral typing as well, if well-conserved regions can be identiied for PCR priming across the strains of interest. We are presently working on upstream steps for separation of complex bacterial communities as well as algorithmic approaches to nonclonal samples.

ACKNOWLEDGMENTS The work described herein was supported in part by grants from the Robert A. Welch Foundation (E-1264 and E-1451) and National Aeronautics and Space Administration (NASA; Cooperative Agreement-

1. Angulo, F.J., Nargund, V.N., Chiller, T.C. (2004) Evidence of an association between use of anti-microbial agents in food animals and anti-microbial resistance among bacteria isolated from humans and the human health consequences of such resistance. J. Vet. Med., 51(8–9), 374–379. 2. National antimicrobial resistance monitoring system: enteric bacteria CDC’s emerging infections program, http://www.cdc.gov/narms/. 3. Reinert, R.R., Reinert, S., van der Linden, M., Cil, M.Y., Al-Lahham, A., Appelbaum, P. (2005) Antimicrobial susceptibility of Streptococcus pneumoniae in eight European countries from 2001 to 2003. Antimicrob. Agents Chemother., 49, 2903–2913. 4. Treitman, A.N., Yarnold, P.R., Warren, J., Noskin, G.A. (2005) Emerging incidence of Enterococcus faecium among hospital isolates (1993 to 2002). J. Clin. Microbiol., 43, 462–463.

102

METHODOLOGIES FOR IDENTIFYING MICROORGANISMS AND VIRUSES BY MASS CATALOGING OF RNAs

5. Protecting the nation’s health in an era of globalization: CDC’s global infectious disease strategy (2002), http://www. cdc.gov/globalidplan/6-partnerships.htm. 6. NIAID biodefense research agenda for CDC category A agents (February 2002). Publication no. 03-5308. http:// www3.niaid.nih.gov/biodefense/research/ biotresearchagenda.pdf. 7. NIAID strategic plan for biodefense research. 2007 update. NIH publication No 03-5306 (2002). 8. Kindhauser, M., World Health Organization Communicable Diseases Cluster (2003) Communicable Diseases 2002: Global Defence against the Infectious Disease Threat. Geneva: World Health Organization. 9. Justiication narrative for iscal year 2005 President’s budget for NIAID, (2005) http://www.niaid.nih.gov/facts/facts.htm. 10. Arias, E., Anderson, R.N., Kung, H.C., Murphy, S.L., Kochanek, K.D. (2003) Deaths: inal data for 2001. Natl. Vital. Stat. Rep., 52(3), 1–115. 11. Wolfe, G., Wolfe, M. (2004) Microbiology growth trends and driving factors In: Clinical Lab Products Online. http:// www.clpmag.com/issues/articles/2004-05_01.asp. 12. Arakere, G., Nadig, S., Swedberg, G., Macaden, R., Amarnath, S.K., Raghunath, D. (2005) Genotyping of methicillin-resistant Staphylococcus aureus strains from two hospitals in Bangalore, South India. J. Clin. Microbiol., 43(7), 3198–3202. 13. Ghoshal, U., Prasad, K.N., Singh, M., Tiwari, D.P., Ayyagari, A. (2004) A comparative evaluation of phenotypic and molecular methods for the detection of oxacillin resistance in coagulase-negative staphylococci. J. Infect. Chemother., 10(2), 86–89. 14. Murchan, S., Kaufmann, M.E., Deplano, A., Rd, R., Stuelens, M., Zinn, C.E., Fussing, V., Salmenlinna, S., Vuopio-Varkila, J., Solh, N.E., Witte, C.W., Tassios, P.T., Legakis, N.L., van Leeuwen, W., van Belkum, A., Vindel, A., Laconcha, I., Garaizar, J., Haeggman, S., OlssonLijequist, B., Ransjo, U., Coombes, G., Cookson, B. (2003) Harmonization of pulsed-ield gel electrophoresis protocols for epidemiological typing of strains of methicillinresistant Staphylococcus aureus: a single approach developed by consensus in 10 European laboratories and its application for tracing the spread of related strains. J. Clin. Microbiol., 41, 1574–1585. 15. Enright, M.C., Day, N.P.J., Davies, C.E., Peacock, S.J., Spratt, B.G. (2000) Multilocus sequence typing for characterization of methicillin-resistant and methicillinsusceptible clones of Staphylococcus aureus. J. Clin. Microbiol., 38, 1008–1015. 16. Shopsin, B., Gomez, M., Montgomery, S.O., Smith, D.H., Waddington, M., Dodge, D.E., Bost, D.A., Riehman, M., Naidich, S., Kreiswirth, B.N. (1999) Evaluation of protein A gene polymorphic region DNA sequencing for typing of Staphylococcus aureus strains. J. Clin. Microbiol., 37(11), 3556–3563. 17. van Belkum, A., Leeuwen, W., Kaufmann, M.E., Cookson, B., Forey, F., Etienne, J., Goering, R., Tenover, F., Steward, C., O’Brien, F., Grubb, W., Tassios, P., Legakis, N., Morvan,

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

A., El Sohl, N., DeRyck, R., Struelens, M., Salmenlinna, S., Vuopio-Varkila, J., Kooistra, M., Talens, A., Witte, W., Verbrugh, H. (1998) Assessment of resolution and intercenter reproducibility of results of genotyping Staphylococcus aureus by pulsed-ield gel electrophoresis of SmaI macrorestriction fragments: a multicenter study. J. Clin. Microbiol., 36, 1653–1659. Ellepola, A.N., Hurst, S.F., Elie, C.M., Morrison, C.J. (2003) Rapid and unequivocal differentiation of Candida dubliniensis from other Candida species using species-speciic DNA probes: comparison with phenotypic identiication methods. Oral Microbiol. Immunol., 18(6), 379–388. Willis, R.C. (2004) Improved molecular techniques help researchers diagnose microbial conditions. Mod. Drug Discov., Feb., 36–42. Amann, R.I., Binder, B.J., Olson, R.J., Chisholm, S.W., Devereux, R., Stahl, D.A. (1990) Combination of 16S rRNA-targeted oligonucleotide probes with low cytometry for analyzing mixed microbial populations. Appl. Environ. Microbiol., 56, 1919–1925. Hugenholtz, P., Goebel, B.M., Pace, N.R. (1998) Impact of culture-independent studies on the emerging phylogenetic view of bacterial diversity. J. Bacteriol., 180(18), 4765– 4774. Andreotti, P.E., Ludwig, G.V., Peruski, A.H., Tuite, J.J., Morse, S.S., Peruski J, L.F. (2003) Immunoassay of infectious agents. BioTechniques, 35, 850–859. Ivnitski, D., O’Neil, D.J., Gattuso, A., Schlicht, R., Calidonna, M., Fisher, R. (2003) Nucleic acid approaches for detection and identiication of biological warfare and infectious disease agents. BioTechniques, 35(4), 862–869. Walker, P., Subasinghe, R. (1999) DNA-based Molecular Diagnostic Techniques: Research Needs for Standardization and Validation of the Detection of Aquatic Animal Pathogens and Diseases. Fisheries Technical Paper, United Nations Food and Agricultural Organization 395. U.S. Food and Drug Administration, Center for Biologic Evaluation and Research. Approved products. http:// www.fda.gov/cber/products. DeLong, E.F., Wickham, G.S., Pace, N.R. (1989) Phylogenetic stains: ribosomal RNA-based probes for the identiication of single microbial cells. Science, 243, 1360– 1363. ProbeBase, an on-line resource for rRNA-targeted oligonucleotide probes, http://www.microbial-ecology.net/ probebase/. Wagner, M., Horn, M., Daims, H. (2001) Fluorescence in situ hybridisation for the identiication and characterisation of prokaryotes. Curr. Opin. Microbiol., 6, 302– 309. Amann, R., Fuchs, B.M., Behrens, S. (2001) The identiication of microorganisms by luorescence in situ hybridization. Curr. Opin. Biotechnol., 12, 231–236. Amann, R.I., Krumholz, L., Stahl, D.A. (1990) Fluorescentoligonucleotide probing of whole cells for determinative, phylogenetic, and environmental studies in microbiology. J. Bacteriol., 172, 762–770.

REFERENCES

31. Fuchs, B.M., Syutsubo, K., Ludwig, W., Amann, R. (2001) In situ accessibility of the Escherichia coli 23S ribosomal RNA for luorescently labeled oligonucleotide probes. Appl. Environ. Microbiol., 67, 961–968.

44.

32. Ribosomal RNA-targeted oligonucleotide microarrays (PhyloChips) for analysis of microbial communities in environmental and clinical samples. University of Vienna, Institute of Ecology and Conservation Biology, http:// www.microbial-ecology.de/phylochips.html. 33. Busti, E., Bordoni, R., Castiglioni, B., Monciardini, P., Sosio, M., Donadio, S., Consolandi, C., Rossi Bernardi, L., Battaglia, C., De Bellis, G. (2002) Bacterial discrimination by means of a universal array approach mediated by LDR (ligase detection reaction). BMC Microbiol., 2, 27.

45. 46.

34. Chandler, D.P., Jarrell, A.E. (2005) Taking arrays from the lab to the ield: trying to make sense of the unknown. BioTechniques, 38(4), 591–600. 35. Chandler, D.P., Newton, G.J., Small, J.A., Daly, D.S. (2003) Sequence versus structure for the direct detection of 16s rrna on planar oligonucleotide microarrays. Appl. Environ. Microbiol., 69, 2950–2958. 36. Brazma, A., Hingamp, P., Quackenbush, J., Sherlock, G., Spellman, P., Stoeckert, C., Aach, J., Ansorge, W., Ball, C.A., Causton, H.C., Gaasterland, T., Glenisson, P., olstege, F.C., Kim, I.F., Markowitz, V., Matese, J.C., Parkinson, H., Robinson, A., Sarkans, U., Schulze-Kremer, S., Stewart, J., Taylor, R., Vilo, J., Vingron, M. (2001) Minimum information about a microarray experiment (MIAME)—toward standards for microarray data. Nat. Genet., 29, 365–371. 37. Kaposi-Novak, P., Lee, J.-S., Mikaelyan, A., Patel, V., Thorgeirsson, S.S. (2004) Oligonucleotide microarray analysis of aminoallyl-labeled cDNA targets from linear RNA ampliication. BioTechniques, 37, 580–588. 38. Woese, C.R., Fox, G.E., Zablen, L., Uchida, T., Bonen, L., Pechman, K., Lewis, B.J., Stahl, D. (1975) Conservation of primary structure in 16S ribosomal RNA. Nature, 254(5495), 83–86. 39. Fox, G.E., Pechman, K.R., Woese, C.R. (1977) Comparative cataloguing of 16S ribosomal RNA: molecular approach to prokaryotic systematics. International Journal of Systemic Bacteriology, 27, 44–57. 40. Woese, C.R., Fox, G.E. (1977) Phylogenetic structure of the prokaryotic domain: the primary kingdoms. Proc. Natl. Acad. Sci. U S A, 74(11), 5088–5090. 41. Fox, G.E., Stackebrandt, E., Hespell, R.B., Gibson, J., Maniloff, J., Dyer, T.A., Wolfe, R.S., Balch, W.E., Tanner, R.S., Magrum, L.J., Zablen, L.B., Blakemore, R., Gupta, R., Bonen, L., Lewis, B.J., Stahl, D.A., Luehrsen, K.R., Chen, K.N., Woese, C.R. (1980) The phylogeny of prokaryotes. Science, 209(4455), 457–463. 42. Woese, C.R., Stackebrandt, E., Macke, T.J., Fox, G.E. (1985) A phylogenetic deinition of the major eubacterial taxa. Syst. Appl. Microbiol., 6, 143–151. 43. Cole, J.R., Chai, B., Farris, R.J., Wang, Q., Kulam, S.A., McGarrell, D.M., Garrity, G.M., Tiedje, J.M. (2005) The

47.

48.

49.

50.

51.

52.

53. 54. 55.

103

Ribosomal Database Project (RDP-II): sequences and tools for high-throughput rRNA analysis. Nucleic Acids Res., 33 (Database issue): D294–D296. Cannone, J.J., Subramanian, S., Schnare, M.N., Collett, J.R., D’Souza, L.M., Du, Y., Feng, B., Lin, N., Madabusi, L.V., Muller, K.M., Pande, N., Shang, Z., Yu, N., Gutell, R.R. (2002) The comparative RNA Web (CRW) site: an online database of comparative sequence and structure information for ribosomal, intron, and other RNAs. BMC Bioinformatics, 3, 2 (Correction: BioMed Central Bioinformatics. 3:15.). National Center for Biotechnological Information, http:// www.ncbi.nlm.nih.gov/. Gardner, S., Kuczmarski, T., Vitalis, E., Slezak, T. (2003) Limitations of TaqMan(r) PCR for detecting divergent viral pathogens illustrated by hepatitis A, B, C, and E viruses and human immunodeiciency virus. J. Clin. Microbiol., 41, 2417–2427. Papin, J.F., Vahrson, W., Dittmer, D.P. (2004) SYBR green-based real-time quantitative PCR assay for detection of west nile virus circumvents false-negative results due to strain variability. J. Clin. Microbiol., 42(4), 1511– 1518. Harrison, A.P., Johnston, C.E., Orengo, C.A. (2007) Establishing a major cause of discrepancy in the calibration of Affymetrix GeneChips. BMC Bioinformatics, 8, 195. Hsiao, C.R., Huang, L., Bouchara, J.-P., Barton, R., Li, H.C., Chang, T.C. (2005) Identiication of medically important molds by an oligonucleotide array. J Clin Microbiol, 43(8), 3760–3768. Jackson, G.W., Karouia, F., Putonti, C., Fofanov, Y., Fox, G.E., Willson, R.C. (2005) Microarray designs and MALDI-TOF mass spectrometry for rapid microbial identiication. In: 229th National Meeting of the American Chemical Society, March 2005, San Diego, CA. Machouart-Dubach, M., Lacroix, C., De Chauvin, M.F., Le Gall, I., Giudicelli, C., Lorenzo, F., Derouin, F. (2001) Rapid discrimination among dermatophytes, Scytalidium spp., and other fungi with a PCR-restriction fragment length polymorphism ribotyping method. J. Clin. Microbiol., 39(2), 685–690. Kumeda, Y., Asao, T. (1996) Single-strand conformation polymorphism analysis of PCR-ampliied ribosomal DNA internal transcribed spacers to differentiate species of Aspergillus section Flavi. Appl. Environ. Microbiol., 62(8), 2947–2952. Applied Biosystems I: MicroSeq 16S rRNA Gene Kit. Applied Biosystems I: MicroSeq Full Gene 16S rDNA Bacterial Identiication Kit. Margulies, M., Egholm, M., Altman, W.E., Attiya, S., Bader, J.S., Bemben, L.A., Berka, J., Braverman, M.S., Chen, Y.-J., Chen, Z., Dewell, S.B., Du, L., Fierro, J.M., Gomes, X.V., Godwin, B.C., He, W., Helgesen, S., Ho, C.H., Irzyk, G.P., Jando, S.C., Alenquer, M.L.I., Jarvie, T.P., Jirage, K.B., Kim, J.-B., Knight, J.R., Lanza, J.R., Leamon, J.H., Lefkowitz, S.M., Lei, M., Li, J., Lohman, K.L., Lu, H., Makhijani, V.B.,

104

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

METHODOLOGIES FOR IDENTIFYING MICROORGANISMS AND VIRUSES BY MASS CATALOGING OF RNAs

McDade, K.E., MKenna, M.P., Myers, E.W., Nickerson, E., Nobile, J.R., Plant, R., Puc, B.P., Ronan, M.T., Roth, G.T., Sarkis, G.J., Simons, J.F., Simpson, J.W., Srinivasan, M., Tartaro, K.R., Tomasz, A., Vogt, K.A., Volkmer, G.A., Wang, S.H., Wang, Y., Weiner, M.P., Yu, P., Begley, R.F., Rothberg, J.M. (2005) Genome sequencing in microfabricated high-density picolitre reactors. Nature, 437, 376–380. Sauer, S., Kliem, M. (2010) Mass spectrometry tools for the classiication and identiication of bacteria. Nature Rev., 8(1), 74–82. Null, A.P., Muddiman, D.C. (2001) Perspectives on the use of electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry for short tandem repeat genotyping in the post-genome era. J. Mass. Spectrom., 36(6), 589–606. Ecker, D.J., Sampath, R., Blyn, L.B., Eshoo, M.W., Ivy, C., Ecker, J.A., Libby, B., Samant, V., Sannes-Lowery, K.A., Melton, R.E., Russell, K., Freed, N., Barrozo, C., Wu, J., Rudnick, K., Desai, A., Moradi, E., Knize, D.J., Robbins, D.W., Hannis, J.C., Harrell, P.M., Massire, C., Hall, T.A., Jiang, Y., Ranken, R., Drader, J.J., White, N., McNeil, J.A., Crooke, S.T., Hofstadler, S.A. (2005) Rapid identiication and strain-typing of respiratory pathogens for epidemic surveillance. Proc. Natl. Acad. Sci. U S A, 102(22), 8012–8017. Hofstadler, S.A., Sampath, R., Blyn, L.B., Eshoo, M.W., Hall, T.A., Jiang, Y., Drader, J.J., Hannis, J.C., SannesLowery, K.A., Cummins, L.L. (2005) TIGER: the universal biosensor. Int. J. Mass Spectrom. Ion Process., 242(1), 23–41. Crain, P.F., McCloskey, J.A. (1998) Applications of mass spectrometry to the characterization of oligonucleotides and nucleic acids. Curr. Opin. Biotechnol., 9(1), 25–34. Kwon, Y., Tang, K., Cantor, C., Koster, H., Kang, C. (2001) DNA sequencing and genotyping by transcriptional synthesis of chain-terminated RNA ladders and MALDI-TOF mass spectrometry. Nucleic. Acids Res., 29(3), E11. Roskey, M.T., Juhasz, P., Smirnov, I.P., Takach, E.J., Martin, S.A., Haff, L.A. (1996) DNA sequencing by delayed extraction-matrix-assisted laser desorption/ionization time of light mass spectrometry. Proc. Natl. Acad. Sci. U S A, 93(10), 4724–4729. Spottke, B., Gross, J., Galla, H.J., Hillenkamp, F. (2004) Reverse Sanger sequencing of RNA by MALDI-TOF mass spectrometry after solid phase puriication. Nucleic Acids Res., 32(12), e97. Koster, H. (1997) DNA diagnostic (sic) based on mass spectrometry. United States patent 5,605,798 and continuations. Koster, H. (2000) DNA diagnostics based on mass spectrometry. United States patent 6,043,031 and continuations. Hahner, S., Ludemann, H.C., Kirpekar, F., Nordhoff, E., Roepstorff, P., Galla, H.J., Hillenkamp, F. (1997) Matrixassisted laser desorption/ionization mass spectrometry

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

(MALDI) of endonuclease digests of RNA. Nucleic Acids Res, 25(10), 1957–1964. von Wintzingerode, F., Bocker, S., Schlotelburg, C., Chiu, N.H., Storm, N., Jurinke, C., Cantor, C.R., Gobel, U.B., van den Boom, D. (2002) Base-speciic fragmentation of ampliied 16S rRNA genes analyzed by mass spectrometry: a tool for rapid bacterial identiication. Proc. Natl. Acad. Sci. U S A, 99(10), 7039–7044. Krebs, S., Medugorac, I., Seichter, D., Forster, M. (2003) RNaseCut: a MALDI mass spectrometry-based method for SNP discovery. Nucleic Acids Res., 31(7), e37. Hartmer, R., Storm, N., Boecker, S., Rodi, C.P., Hillenkamp, F., Jurinke, C., 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. Lefmann, M., Honisch, C., Bocker, S., Storm, N., von Wintzingerode, F., Schlotelburg, C., Moter, A., van den Boom, D., Gobel, U.B. (2004) Novel mass spectrometrybased tool for genotypic identiication of mycobacteria. J. Clin. Microbiol., 42(1), 339–346. Zhang, Z., Jackson, G.W., Fox, G.E., Willson, R.C. (2006) Microbial identiication by mass cataloging. BMC Bioinformatics, 7, 117. Jackson, G.W., McNichols, R.J., Fox, G.E., Willson, R.C. (2006) Bacterial genotyping by 16S rRNA mass cataloging. BMC Bioinformatics, 7, 321. Jackson, G.W., McNichols, R.J., Fox, G.E., Willson, R.C. (2007) Universal bacterial identiication by mass spectrometry of 16S ribosomal RNA cleavage products. Intl. J. Mass Spectrometry, 261(2–3), pp. 218–226. Jackson, G.W., McNichols, R.J., Fox, G.E., Willson, R.C. (2008) Toward universal lavivirus identiication by mass cataloging. J. Mol. Diagn., 10(2), 135–141. McCloskey, J.A., Rozenski, J. (2005) The small subunit rRNA modiication database. Nucleic Acids Res., 33 (Database issue): D135–D138. Koomen, J.M., Russell, W.K., Tichy, S.E., Russell, D.H. (2002) Accurate mass measurement of DNA oligonucleotide ions using high-resolution time-of-light mass spectrometry. J. Mass Spectrom., 37(4), 357–371. Saitou, N., Nei, M. (1987) The neighbor-joining method: A new method for reconstructing phylogenetic trees. Mol. Biol. E, 4, 406–425. Cole, J.R., Wang, Q., Cardenas, E., Fish, J., Chai, B., Farris, R.J., Kulam-Syed-Mohideen, A.S., McGarrell, D.M., Marsh, T., Garrity, G.M., Tiedje, J.M. (2009) The Ribosomal Database Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Res., 37 (Database issue): D141–D145. Weisburg, W.G., Barns, S.M., Pelletier, D.A., Lane, D.J. (1991) 16S ribosomal DNA ampliication for phylogenetic study. J. Bacteriol., 173(2), 697–703. Lane, D.J., Pace, B., Olsen, G.J., Stahl, D.A., Sogin, M.L., Pace, N.R. (1985) Rapid determination of 16S ribosomal RNA sequences for phylogenetic analyses. Proc. Natl. Acad. Sci. U S A, 82(20), 6955–6959.

REFERENCES

81. NIAID Category A, B, & C priority pathogens (2005) http:// www3.niaid.nih.gov/biodefense/bandc_priority.htm. 82. Kumar, S., Tamura, K., Nei, M. (2004) MEGA3: Integrated software for Molecular Evolutionary Genetics Analysis and sequence alignment. Brief. Bioinform., 5, 150–163. 83. Scaramozzino, N., Crance, J.-M., Jouan, A., DeBriel, D.A., Stoll, F., Garin, D. (2001) Comparison of lavivirus universal primer pairs and development of a rapid, highly sensitive heminested reverse transcription-PCR assay for detection of laviviruses targeted to a conserved region of the NS5 gene sequences. J. Clin. Microbiol., 39(5), 1922–1927. 84. Kao, C.L.K.C., Chao, D.Y., Wu, H.L., Chang, G.J.J. (2005) Laboratory diagnosis of dengue virus infection: current and future perspectives in clinical diagnosis and public health. J. Microbiol. Immunol. Infect., 38, 5–16. 85. Monath, T.P. (1994) Dengue: the risk to developed and developing countries. Proc. Natl. Acad. Sci. USA, 91, 2395–2400. 86. Shu, P.-Y., Chen, L.-K., Chang, S.-F., Su, C.-L., Chien, L.-J., Chin, C., Lin, T.-H., Huang, J.-H. (2004) Dengue Virus serotyping based on envelope and membrane and nonstructural protein NS1 serotype-speciic capture immunoglobulin M enzyme-linked immunosorbent assays. J. Clin. Microbiol., 42(6), 2489–2494. 87. Halstead, S.B. (1988) Pathogenesis of dengue: challenge to molecular biology. Science, 239, 476–481. 88. Teles, F.R., Prazeres, D.M., Lima-Filho, J.L. (2005) Trends in dengue diagnosis. Rev. Med. Virol., 15(5), 287–302. 89. Putonti, C., Chumakov, S., Mitra, R., Fox, G.E., Willson, R.C., Fofanov, Y. (2006) Human-blind probes and primers for dengue virus identiication—exhaustive analysis of subsequences present in the human and 83 dengue genome sequences. FEBS J., 273, 398–408.

105

90. Kao, C.L., King, C.C., Chao, D.Y., Wu, H.L., Chang, G.J.J. (2005) Laboratory diagnosis of dengue virus infection: current and future perspectives in clinical diagnosis and public health. J. Microbiol. Immunol. Infect., 38, 5–16. 91. De Paula, S.O., da Fonseca, B.A.L. (2004) Dengue: a review of the laboratory tests a clinician must know to achieve a correct diagnosis. Braz. J. Infect. Dis., 8, 390–398. 92. Drosten, C., Gottig, S., Schilling, S., Asper, M., Panning, M., Schmitz, H., Gunther, S. (2002) Rapid detection and quantiication of RNA of Ebola and Marburg viruses, Lassa virus, Crimean-Congo hemorrhagic fever virus, Rift Valley fever virus, dengue virus, and yellow fever virus by realtime reverse transcription-PCR. J. Clin. Microbiol., 40, 2323–2330. 93. Lanciotti, R.S., Calisher, C.H., Gubler, D.J., Chang, G.J., Vorndam, A.V. (1992) Rapid detection and typing of dengue viruses from clinical samples by using reverse transcriptase-polymerase chain reaction. J. Clin. Microbiol., 30, 545–551. 94. Fulop, L., Barrett, A.D.T., Phillpotts, R., Martin, K., Leslie, D., Titball, R.W. (1993) Rapid identiication of laviviruses based conserved NS5 gene sequences. J. Virol. Methods, 44, 179–188. 95. Thompson, J.D., Higgins, D.G., Gibson, T.J. (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-speciic gap penalties and weight matrix choice. Nucleic Acids Res., 22(22), 4673–4680. 96. Technical Note, TN225, “Sample Preparation of Oligonucleotides Prior to MALDI-TOF MS Using ZipTipC18 and ZipTipµ-C18 Pipette Tips” (2000) www.millipore.com/ziptip. 97. Altschul, G.F., Gish, W., Miller, W., Myers, E.W., Lipman, D.J. (1990) Basic local alignment search tool. J. Mol. Biol., 215, 403–410.

SECTION II PHARMACEUTICAL

5 PRECLINICAL PHARMACOKINETICS: INDUSTRIAL PERSPECTIVE Ayman El-Kattan and Manthena Varma

5.1

INTRODUCTION

In parallel to bioanalytical revolution, the drug discovery revolution continues inexorably forward. Today, we not only beneit from revolutionary new medicines that have dramatically improved our quality of life, but we expect even more to come. Compound collections have bal­ looned through combinatorial library synthesis and advancements in biological science mapped the human genome while we impatiently await the promise of genet­ ically tailored medicines that will treat every ailment. This pharmaceutical revolution could not have been achieved without the undergirdment of advanced ana­ lytical instrumentation. Liquid chromatography com­ bined with mass spectrometry (LC/MS) and tandem mass spectrometry (LC/MS/MS) has enabled the char­ acterization of novel potential drugs as well as quantita­ tive measurement in increasingly complex milieu with an incredibly rapid throughput rate. Quantitation of drugs in biological media such as blood (usually plasma or serum) and tissues has become more or less routine, with much improved sensitivity. The role of mass spec­ trometry in enabling this quantitation is now a straight­ forward process and is taken for granted. However, signiicant challenges to bioanalysis of potential new chemical entities (NCEs) still lurk in the background, waiting to catch the unsuspecting biologist unaware, and helping to keep analytical chemists gainfully employed. The main purposes of this chapter are, irst, to intro­ duce the fundamentals of pharmacokinetics derived

from the estimated concentrations in the biological matrices (e.g., plasma and urine), and second, to present a prospective view on the role of LC/MS/MS method­ ologies to address the bioanalytical needs in the phar­ macokinetic assessment of new drug entities (NCE) at the drug discovery stage.

5.2 A PHARMACOKINETICS PRIMER Pharmacokinetics is the science that describes the time course of drug concentration in the body resulting from administration of a certain drug dose. In comparison, pharmacodynamics is the science that describes the relationship of the time course of drug concentration and the drug effects in the body [1]. Therefore, pharma­ cokinetics can be considered a biomarker of drug ex­ posure as well as a marker of eficacy and safety. Key determinants of the pharmacokinetics of a drug include absorption, distribution, metabolism, and elimination (ADME) [2]. Discovering novel therapeutic agents is an increas­ ingly time­consuming and costly process. Most estimates indicate that it takes approximately 10–15 years and more than $800 million to discover and develop a suc­ cessful drug product [3]. It is well established in the literature that poor drug pharmacokinetics is one of the leading causes of compound failure in preclinical and clinical development [4]. Compounds with poor pharmacokinetic proile tend to have low oral systemic

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

109

PRECLINICAL PHARMACOKINETICS: INDUSTRIAL PERSPECTIVE











To select compounds with the maximum potential of reaching the target. To determine the appropriate route of administra­ tion to deliver the drug (typically oral). To understand how the blood levels relate to efi­ cacy or toxicity in order to choose eficacious and safe doses. To decide on the frequency and duration of dosing in order to maintain adequate drug concentration at target for disease modiication. To accurately predict the human pharmacokinetic proile prior to clinical studies.

A pharmacokinetic study is composed of three phases, namely the in­life phase, bioanalysis, and data analysis. The in­life phase includes administering the compound to animals or humans and collecting samples from the appropriate matrix of interest such as blood or urine at predetermined time intervals for bioanalysis. The bioanalytical phase involves analysis of a drug and/ or its metabolite(s) concentration in blood, plasma, serum, or urine. This analysis typically involves sample extraction and detection of analytes using sensitive methods like LC/MS/MS. The third phase is data analy­ sis using noncompartmental or compartmental pharma­ cokinetic computational methods. In drug discovery, preliminary pharmacokinetic studies are usually conducted in rodents to evaluate the extent of drug exposure in vivo. These rodent studies are commonly followed by pharmacokinetic studies in larger animals such as dog or monkey to better charac­ terize the pharmacokinetic proile of the compound and to support safety studies. Pharmacokinetic scaling, or allometry, is a discipline that is used to predict human pharmacokinetic proile using preclinical data and is widely used in predicting the drug human half­life, dose, and extent of absorption. Accurate prediction of human pharmacokinetic proile is imperative to minimize drug failure in development due to pharmacokinetic liability. A detailed description of methods in predicting human pharmacokinetics is beyond the scope of this chapter, but can be found in many excellent reviews [5–8]. A more in­depth discussion of various pharmacokinetic concepts and their applications can be found in various references [9,10].

5.3 PARAMETERS THAT DEFINE PHARMACOKINETIC PROFILE 5.3.1 Area under the Curve (AUC) The irst step in a pharmacokinetic experiment is to dose animals or humans with NCE and collect blood samples at predeined time points. Animals are gener­ ally dosed intravenously (IV) and/or orally (po). After sample preparation and quantiication, a concentration­ time proile is generated (Figure 5.1). Mathematically, area under the plasma (or blood) concentration­time curve (AUC) can be calculated from the obtained concentration­time proile by: ∞



AUC = C ⋅ dt.

(5.1)

o

AUC is a primary measure of the extent of drug avail­ ability to the systemic circulation (i.e., the total amount of unchanged drug that reaches the systemic circulation following IV or extravascular administration). The unit for AUC is concentration per unit time (e.g., nanogram hour per milliliter [ng*h/mL]). AUC is determined using simple integration method as shown in Equation 5.1 or a linear trapezoidal method, which is the most widely used approach (Figure 5.1). The area of each trapezoid is calculated using the following equation: AUC t1 → t2 =

(C 2 + C 1 ) × (t 2 − t1 ). 2

(5.2)

Cmax Drug Plasma a Concentration

plasma exposure and high interindividual variability, which limits their therapeutic utility. Therefore, a better understanding of the pharmacokinetic proile early on enables the discovery of compounds with drug­like properties. In discovery settings, the main outcomes of pharmacokinetic and pharmacodynamic evaluations are as follows:

C2



110

AUC = ∫C ⋅ dt 0

C1

Clast T1T2 Tmax

Tlast Time

FIGURE 5.1 Estimation of the area under the plasma concentration­time curve (AUC).

PARAMETERS THAT DEFINE PHARMACOKINETIC PROFILE

The extrapolated area from tlast to ∞ is estimated as: AUC tlast →∞ = C last / K e,

(5.3)

where Clast is the last observed concentration at tlast and Ke is the slope obtained from the terminal portion of the curve, representing the terminal elimination rate constant. The total AUC (AUC0→∞) is determined as:

In general, a drug is either eliminated unchanged through excretion in the urine and/or bile, or by meta­ bolic conversion into more polar metabolite(s) that can be readily excreted in urine and/or bile. Therefore, total body CL is a sum of all CLs by various mechanisms and can be expressed mathematically as shown in Equation 5.6: CL tot = CL hep + CL ren + CL bil ,

AUC 0→∞ = AUC 0→ tlast + AUC tlast →∞ .

Clearance

Clearance (CL) is a primary pharmacokinetic parame­ ter that describes the process of irreversible elimination of a drug from the systemic circulation. It is deined as the volume of blood or plasma that is totally cleared of its content of drug per unit time. Thus, it measures the removal of drug from blood or plasma. However, CL does not indicate the extent of drug that is being removed, but instead represents the rate and thus the units are given as milliliter per minute (mL/min) or mil­ liliter per minute per kilogram (mL/min/kg) (normal­ ized to body weight). The most widely used approach to evaluate plasma (total) CL involves IV administration of a single dose and measuring its plasma concentration at different time points (Figure 5.2). In this manner, the calculated CL (Equation 5.5) will not be confounded by complex absorption and distribution phenomena which com­ monly occur during oral dosing [9].

Drug Plasma Concentration

CL tot = DoseIV / AUC IV .

(5.6)

(5.4)

AUC is used in the calculation of clearance (CL), appar­ ent volume of distribution (Vd), and bioavailability (see following sections) and relects the general extent of exposure over time. 5.3.2

111

(5.5)

CLlot = DoseIV/AUCIV

where CLtot is the total body CL from all different organs and mechanisms, CLhep is the hepatic CL, CLren is the renal CL, and CLbil is the biliary CL. It is interesting to note that around three­quarters of the top 200 prescribed drugs in the United States are primarily cleared by hepatic metabolism [11]. The hepatic extraction ratio (Eh) is a pharmacokinetic parameter that is widely used to assess the liver’s ability to extract drug from the systemic circulation [10]. Eh is deined as the fraction of a drug in the blood that is cleared (extracted) on each passage through the liver and is a function of hepatic blood CL (CLhep) and the hepatic blood low (Q) [10]: Eh =

CL hep . Q

(5.7)

If the predominant CL mechanism for a compound is via hepatic metabolism, then it is reasonable to assume that the CLtot is equal to CLhep. Thus, Eh =

CL hep CL tot = . Q Q

(5.8)

Compounds that undergo hepatic metabolism can be classiied according to their Eh. Compounds with Eh > 0.7 are considered high extraction drugs, whereas com­ pounds with Eh < 0.3 are considered low extraction drugs. Eh has a major impact on oral drug bioavailability (see below). 5.3.3

Plasma versus Blood CL

Calculation of Eh from drug CL in blood requires the determination of drug concentration in whole blood. Since determination of drug concentration is usually performed in plasma or serum, knowledge of the blood/ plasma concentration ratio is necessary to estimate the blood CL. Blood CL is calculated using this equation: Time

FIGURE 5.2 IV dosing.

Plasma concentration time proile following

Plasma Clearance Blood Concentration (C b ) = Blood Clearance Plasma Concentration (C p )

(5.9)

112

PRECLINICAL PHARMACOKINETICS: INDUSTRIAL PERSPECTIVE

5.3.4 Apparent Volume of Distribution (Vd) Volume of distribution (Vd) is a proportionality factor that relates the amount of a drug in the body to its blood or plasma concentrations: Amount of drug in the body at time t = Vd × C plasma at time t.

(5.10)

Following IV dosing and at t = 0, the amount of drug in the body is equal to the administered IV dose. Vd at t = 0 is termed volume of the central compartment (Vc). Similar to CL, Vd is a primary pharmacokinetic parameter and its unit is volume (e.g., liters per kilo­ gram [L/kg]). Vd is used to assess the extent of drug distribution in the body. This is usually achieved by com­ paring the drug Vd to the total body water. If the drug has a Vd that is smaller than the total body water (human total body water = 42 L per 70 kg human body weight), then this would suggest that the drug has limited tissue distribution (e.g., naproxen has a Vd = 11 L per 70 kg human body weight) [12]. On the other hand, if a drug has a Vd larger than the total body water, then this would suggest that the drug is able to distribute to body tissues (e.g., olanzapine has a Vd = 1120 L per 70 kg human body weight) [13]. In the literature, Vd ranges from 3 to more than 40,000 L per 70 kg human body weight. Therefore, the term “apparent volume of distri­ bution” is usually used.

Use of loading dose is important especially for those drugs in which it is desirable to immediately or rapidly reach the steady­state plasma concentration (Css) (e.g., anticoagulant, antiepileptic, antiarrhythmic, and antimi­ crobial therapy). 5.3.6

Half­life (t1/2) is the time that is required for the amount (or plasma concentration) of a drug to decrease by one­ half. It is calculated by the following equation: t 1/ 2 =

Vdss is the volume of distribution that is determined when plasma concentrations are measured at steady state and in equilibrium with the drug concentration in the tissue compartment.

Vdss

Amount of drug in the body at equilibrium conditions . (5.11) = Steady-state plasma concentrations (C ss )

Although Vdss is a steady­state parameter, it can be calculated using nonsteady state data as: Vdss = CL × MRT,

(5.12)

where MRT is the drug mean residence time (see below). Furthermore, Vdss is used in the calculation of a loading dose as: Loading dose =

Vdss × C ss . F

(5.13)

0.693 × Vd . CL

(5.14)

Half­life (t1/2) is a dependent pharmacokinetic param­ eter that is determined by both CL and Vd, which are independent primary pharmacokinetic parameters. Therefore, t1/2 is increased by a decrease in CL or increase in Vd and vice versa. t1/2 is the most widely reported pharmacokinetic parameter since it may con­ stitute a major determinant of the duration of action after single and multiple dosing. In addition, it plays a key role in determining the time that is required to reach steady state following multiple dosing and the frequency with which doses can be given. The unit for t1/2 is time (e.g., hour). 5.3.7

5.3.5 Apparent Volume of Distribution at Steady State (Vdss)

Half-Life (t1/2)

Bioavailability (F%)

According to the European Medicines Evaluation Agency (EMEA), bioavailability (F%) is “the rate and extent to which an active moiety is absorbed from a pharmaceutical form, and becomes available in the sys­ temic circulation.” As a parameter, there are two types of bioavailability: 1. Absolute bioavailability, which refers to the frac­ tion of the extravascular (e.g., oral) dose that reaches the systemic circulation unchanged in ref­ erence to an IV dose. It is usually determined by calculating the respective AUC after oral and IV administration: Absolute bioavailability =

AUC po DoseIV × . AUC IV Dosepo (5.15)

2. Relative bioavailability, which refers to the frac­ tion of a dose of drug reaching the systemic circu­ lation relative to a reference product. It is usually calculated as:

PARAMETERS THAT DEFINE PHARMACOKINETIC PROFILE

Relative bioavailability =

AUC test Doseref × . AUC ref Dosetest (5.16)

Oral bioavailability is determined by the fraction of dose absorbed (Fa) in the gastrointestinal tract and frac­ tion of dose that does not undergo metabolism in the intestinal tract (Fg) and liver (Fh). It is mathematically expressed by the following equation: F = Fa ⋅ Fg ⋅ Fh .

(5.17)

Fh is calculated using the following equation: Fh = 1 − Eh = 1 −

CL h . Q

(5.18)

Thus, if a drug has a high hepatic extraction (Eh > 0.7), then its extent of bioavailability will be low when it is given orally (F ≤ 0.3). On the other hand, if a drug has low hepatic extraction (Eh < 0.3), then the extent of bio­ availability will be high provided that it is completely absorbed and not signiicantly metabolized by the intestine. 5.3.8

MRT

MRT is the average time for all drug molecules to exist in the body. MRT is another measure of drug elimina­ tion and its unit is time (e.g., hour). Following IV dosing, MRT is calculated as:

113

5.3.9 Maximum Plasma Concentration (Cmax) and Time of Maximum Concentration (Tmax) Maximum plasma concentration (Cmax) is deined as the maximum observed drug concentration in the plasma concentration­time proile following IV or oral dosing. Most commonly, Cmax is obtained by direct observation of the plasma concentration­time proile (Figure 5.1). For some drugs, the pharmacological effect is dependent on the Cmax. For example, aminoglycosides, which are widely used antibiotics, needs to achieve a Cmax that is at least eight­ to 10­fold higher than the minimum inhib­ itory concentration (MIC) to obtain a clinical response ≥90% [14,15]. The unit of Cmax is concentration unit (e.g., nanogram per milliliter [ng/mL]). Time of maximum concentration (Tmax) is the time required to reach Cmax. As with Cmax, Tmax is usually determined from direct observation of the plasma concentration­time proile. The unit of Tmax is time (e.g., hour). 5.3.10

Noncompartmental Pharmacokinetics

Various pharmacokinetic parameters such as CL, Vd, F%, MRT, and t1/2 can be determined using noncom­ partmental methods. These methods are based on the empirical determination of AUC and AUMC described above. Unlike compartmental models (see below), these calculation methods can be applied to any other models, provided that the drug follows linear pharmacokinetics. However, a limitation of the noncompartmental method is that it cannot be used for the simulation n of different plasma concentration­time proiles when there are alterations in dosing regimen or when multiple dosing regimens are used.



AUMC MRT = = AUC

∫ C ⋅ t ⋅ dt 0





5.3.11 ,

(5.19)

C ⋅ dt

0

where AUMC is the area under the irst moment versus time curve from time t = 0 to ∞ and calculated using trapezoidal rule similar to AUC. It should be emphasized that in some cases, MRT can be a better parameter to assess drug elimination com­ pared to half­life. This can be attributed to the greater analytical sensitivity shown with various analytical systems such as LC/MS/MS; the lower drug concentra­ tions measured following drug administration appeared to yield longer terminal half­lives which are not related to the drug’s pharmacologically relevant half­life. In a case like this, it would be recommended to measure MRT rather than half­life to assess drug elimination.

Compartmental Pharmacokinetics

Compartmental models of pharmacokinetic analysis are widely used to describe drug distribution and disposi­ tion. In these models, the body is assumed to be com­ posed of one compartment or more and the drug kinetics can be deined by differential equations generally of irst­order process. These compartments do not have any physiological signiicance. However, they may rep­ resent a group of tissues or organs with similar distribu­ tion characteristics. For example, highly blood perfused body organs like the liver, lungs, and kidney often have different drug distribution than fat tissue. Compartmental models are usually arranged in a mammillary format, such that there are one or more compartments that feed from a central compartment. 5.3.11.1 One-Compartment Open Model In the one­compartment model, the body is assumed to be a

114

PRECLINICAL PHARMACOKINETICS: INDUSTRIAL PERSPECTIVE

In Plasma Concentration

In Plasma Concentration

C° = D/Vc

Slope = –Ke

A B Slope = α Slope = β

C = C°e–Ke2 Time C = Ae−α⋅t + Be−β⋅t

FIGURE 5.4 Two­compartment open model. Time

FIGURE 5.3 One­compartment model.

homogenous unit where the drug is rapidly distributing throughout the body, and once eliminated it follows a monoexponential decline (Figure 5.3). Following IV dosing, the plasma drug concentration can be calcu­ lated as: C = C° ⋅ e − Ke t ,

(5.20)

where C° is the plasma drug concentration immediately after IV dosing. C° is also calculated as: C° = D / Vc .

(5.21)

Unlike other compartmental models, there is only one Vd, where Vc = Vdss. 5.3.11.2 Two-Compartment Open Model When the drug concentration­versus­time proile demonstrates a biexponential decline following IV dosing, a two­ compartment model that is the sum of two irst­order processes (distribution and elimination) will better describe the data (Figure 5.4). A drug that follows the pharmacokinetics of a two­compartment model does not rapidly distribute throughout the body as evident in the one­compartment model. In the two­compartment model, the drug is assumed to distribute into two com­ partments, the central and the tissue compartments. The central compartment represents the highly per­ fused body organs where the drug distributes rapidly and uniformly. On the other hand, in the tissue compart­ ment, the drug distributes more slowly. For a drug that follows the two­compartment model, the rate of drug plasma concentration change following IV dose can be determined as: C = Ae− α⋅t + Be− β⋅t ,

where A and B are functions of the administered dose, and α and β are the irst­order constants for the distribu­ tion and elimination phase, respectively. In this chapter, only the one­ and two­compartment models following IV dosing were described. Other models with extravascular dosing have an additional compartment with an absorption rate constant describ­ ing input into the central compartment. Models with three or greater compartments may be used if the drug concentration versus time may be described better with additional exponential terms. However, these models have greater complexity.

(5.22)

5.4 MODELING TO PREDICT SINGLE- AND MULTIPLE-DOSE PHARMACOKINETIC PROFILES As previously discussed, compartmental models can be effectively used to project plasma concentrations that would be achieved following different dosage regimen and/or multiple dosing. However, for these projections to be accurate, the drug pharmacokinetic proile should follow irst­order kinetics where various pharmacoki­ netic parameters such as CL, Vd, t1/2, and F% do not change with dose.

5.4.1

Linear and Nonlinear Pharmacokinetics

Drug metabolism, renal tubular secretion, or biliary secretion are usually mediated by metabolizing enzymes or transporter proteins. These protein systems usually have good substrate selectivity with inite capacities, which are described by the Michaelis–Menten equation, ν=

Vmax ⋅ C , Km + C

(5.23)

MODELING TO PREDICT SINGLE­ AND MULTIPLE­DOSE PHARMACOKINETIC PROFILES

5.4.2 Allometric Scaling Allometric scaling (allometry) is the discipline that pre­ dicts human pharmacokinetics using preclinical data [16]. This approach is based on empirical observations that various physiological parameters are functions of body size. The most widely used equation in allometry is a one­term power function: y = aBx , log Y = x log B + log a,

(5.24) (5.25)

where B is any independent variable such as animal body weight, and y is any dependent variable. For example, in pharmacokinetics, y is usually, CL, Vdss, or t1/2. The expo­ nent of the allometric equation (x) determines the slope of a double logarithmic plot (Figure 5.5). It is interesting to note that interspecies CL fre­ quently scales with an exponent of 0.75. Whereas the exponent for interspecies Vdss frequently scales with an exponent of 1.0, the half­life usually scales with an expo­

logY = blogX + loga

log Y

Where C is the drug plasma concentration, Vmax is the maximum elimination or transport rate, and Km is the Michaelis constant. The values of Vmax and Km are dependent on the nature of the drug and enzymatic process involved. This equation implies that when the drug plasma concentration is lower than Km, no satura­ tion of the enzymes or transporters protein occurs. Therefore, various pharmacokinetic parameters such as CL, Vd, t1/2, and F% remain constant with respect to dose and time and the drug is considered to follow linear pharmacokinetics. This is a desirable property in that prediction of the plasma exposure following various dosing regimens and over multiple dosing can be more easily achieved. However, when the drug plasma con­ centration is larger than Km, saturation of the enzymes or transporter proteins occurs and the rate of elimina­ tion or transport rate is maximized and approaches that of Vmax. As a result, drug elimination or secretion becomes a zero­order process and the pharmacokinetic parameters such as CL, Vdss, and t1/2 become a function of the administered dose or plasma concentration. Furthermore, drugs that demonstrate saturable meta­ bolism may exhibit less than expected oral irst­pass metabolism, resulting in a higher bioavailability. Con­ sequently, there is a greater fractional increase in Css (or AUC) than the corresponding fractional increase in the rate of drug administration. Drugs with this character­ istic will likely require more careful monitoring when dosage adjustment is made in order to achieve the desired therapeutic effects and to minimize the poten­ tial for adverse effects.

115

log X

FIGURE 5.5 The linear transform of the simple allometric equation.

nent of 0.25. Allometry sometimes fails to predict human pharmacokinetic proile in particular human CL. This is attributed to interspecies differences in the metabolic enzyme. Therefore, it is widely accepted to use in vitro metabolic data obtained from multiple species to correct for the human pharmacokinetic parameters obtained from allometry. 5.4.3 Physiology-Based Pharmacokinetic (PBPK) Modeling PBPK modeling is a quantitative description of pharma­ cokinetics using the drug substance properties and the species­speciic anatomical and physiological informa­ tion. Mathematically, the model constitutes of multiple compartments, each corresponding to different body organs or tissues and linked together based on their anatomical placement with respect to blood low. The vital physiological information in building these models are the tissue volumes, blood low to the tissue, and the tissue composition. Meanwhile, the compound­ speciic information like tissue to plasma partition (Kp) and the intrinsic CL are provided as inputs. The major advantages of such modeling are twofold. First, unlike the empirical models (e.g., allometry), the obtained pharma­ cokinetic parameters are physiologically relevant and second, concentration­time proiles of each tissue can be obtained simultaneously. With the exploratory relation­ ships between tissue concentration­proiles and the phar­ macological or toxicological effects, PBPK modeling provides a framework for mechanistic pharmacokinetics/ pharmacodynamics (PK/PD) modeling. PBPK models are also the most reliable in dose–response and tissue exposure assessments under various physiological condi­ tions (e.g., age, disease condition) [17].

116

PRECLINICAL PHARMACOKINETICS: INDUSTRIAL PERSPECTIVE

5.4.4

PK/PD Modeling

Pharmacokinetics describes the drug concentration­ time courses in body luids resulting from administra­ tion of a certain drug dose, while pharmacodynamics describes the observed response resulting from a certain drug concentration. The rationale for PK/PD modeling is to link pharmacokinetics and pharmacodynamics so as to establish the dose–concentration–response rela­ tionships and subsequently to predict the effect–time courses following drug administration. Several rela­ tively simple pharmacodynamic models, which comprise the ixed­effect model, the linear model, the long­linear model, the Emax model, and the sigmoid Emax­model can describe concentration–response relationships [9,18,19]. The Emax model is described by a hyperbolic equation, E=

Emax ⋅ C , EC 50 + C

(5.26)

where E is the effect, intensity, or response, Emax is the maximum effect; EC50 is a constant and represents the concentration which corresponds to 50% of the maximum effect; and C is the plasma drug concentra­ tion. More complex integrated PK/PD models are nec­ essary to link and account for a possible temporal dissociation between the plasma concentration and the observed effect. Four basic attributes may be used to characterize PK/PD models: First, the link between measured concentration and the pharmacological response mechanism that mediates the observed effect (direct versus indirect link); second, the response mech­ anism that mediates the observed effect, (direct versus indirect response); third, the information used to estab­ lish the link between measured concentration and observed effect (hard versus soft link); and fourth, the time dependency of the involved pharmacodynamic parameters (time variant versus time invariant) [20–23]. The expanded and early use of PK/PD modeling in drug discovery and development is highly beneicial for increasing the success rate of drug discovery and devel­ opment and will most likely improve the current state of applied therapeutics.

5.5 ROLE OF LC/MS/MS IN PHARMACOKINETIC ASSESSMENT IN DRUG DISCOVERY Drug discovery demands for the optimization of chem­ istry space with the in­house decisions around target potency, in vivo pharmacology,ADME/pharmacokinetics optimization, and dose range inding to guide the suc­

cessful selection of lead candidate compounds. Such decisions­in early drug discovery should allow defects to be identiied and corrected prior to time­consuming and expensive preclinical and clinical development stages. This process is called lead optimization [24–26]. From a pharmacokinetic perspective, the parameters that are investigated and optimized include half­life, CL, Vd, and bioavailability. In the past, signiicant numbers of molecules have failed in the early stages of clinical development due to inappropriate pharmacokinetic properties, and the aim of lead optimization should be to reduce this attrition rate. Until recently, the time taken to conduct a preclinical pharmacokinetic study was too slow to enable the pharmacokinetic screen to be successfully used for lead optimization. The bioana­ lytical method was usually the bottleneck that limited the timely delivery of the data, and therefore, decision making. Recent advances in the use of LC/MS/MS in bioanalysis have transformed the discovery process and allowed rapid pharmacokinetics optimization for effec­ tive lead optimization. With the rapid advancements in developing sensitive bioanalytical methods along with automation to handle large number of samples per week, rapid pharmacokinetic proiling is now a reality. In essence, there must be a balance between compound throughput and the depth of information gained. For example, a number of pharmaceutical companies used N­in­One dosing (i.e., cassette dosing), which involves the simultaneous administration of several compounds to a single animal followed by rapid sample analysis by LC/MS/MS, to increase the throughput of in vivo pharmacokinetic screening and enable rapid lead optimization. The speciicity of the LC/MS/MS bioanal­ ysis enables a plasma concentration­time proile and associated pharmacokinetic parameters to be obtained for each individual compound. Albeit there are known advantages of cassette dosing, which include less resour­ ces, higher throughput, and providing valid/rapid/useful pharmacokinetic screening information, there are pos­ sible complications associated with this approach, such as the potential for drug–drug interactions. To minimize these interactions, the administered compounds doses are kept low and a standard compound with a known pharmacokinetic proile is included in each cocktail to monitor such interactions and ensure the quality and reproducibility of the data. Therefore, it should be stressed that the liability of drug–drug interactions is ever present, and the method should be validated for any compound chemical class prior to use in a decision­ making role. Another approach that can further increase the productivity of in vivo screening can be achieved by incorporating the limited time­point approach into N­in­One studies. If studies indicated that potential drug–drug interactions would limit the use of cassette

REFERENCES

dosing, then some gains in throughput can still be obtained by modifying the study design of usual cassette­ dosing assays. In this study design, animals only receive one compound in a dose, but plasma samples from animals that have received different compounds are pooled prior to LC/MS/MS assay. While the in vivo workload is not reduced, the sample processing and LC/ MS/MS time are signiicantly reduced, leading to greater eficiency. Indeed, there is still a chance of interactions during LC/MS/MS analysis, but these are usually both predictable and avoidable [27,28]. For more detailed characterization of the pharmaco­ kinetic parameters of lead candidates, conventional studies that are based on running both IV and oral dosing with extensive blood sampling are used. However, these studies are relatively time consuming and resource intensive. They are recommended to differentiate within a limited group of lead candidates, where small differ­ ences in the pharmacokinetic proiles could be signii­ cant in selecting lead compounds moving forward. For the selection of the best molecules during a lead opti­ mization program, it may only be necessary to rank molecules according to predetermined speciic pharma­ cokinetic parameters. For example, if high oral bioavail­ ability is required for a series of compounds that is known to be well absorbed, then it is not necessary to dose by both oral and IV routes. A simple IV screen for low CL will identify the best and worst compounds. Furthermore, to rank compounds according to CL, an estimate of plasma concentrations at two, or at most three, time points may be all that is needed. While an AUC based on these limited points would not give an accurate measure of CL, it will, with the right study design, correlate with CL and hence be a reliable ranking tool. Both of these approaches reduce analytical and in vivo workload, and if terminal bleeds are involved animal numbers are also reduced [29]. In summary, there are a number of ways to increase pharmacokinetic throughput by using novel study designs of N­in­One dosing, N­in­One analysis, and limited sample time. The decision on whether to use any of these, or to use conventional study design, must be made on the balance between throughput on the one hand and accuracy and conidence in the pharmacoki­ netic parameters on the other.

REFERENCES 1. Meibohm, B., Derendorf, H. (1997) Basic concepts of pharmacokinetic/pharmacodynamic (PK/PD) modelling. Int. J. Clin. Pharmacol. Ther., 35(10), 401–413. 2. Lin, J., Sahakian, D.C., de Morais, S.M., Xu, J.J., Polzer, R.J., Winter, S.M. (2003) The role of absorption, distribution,

3.

4.

5.

6.

7.

8.

9. 10. 11.

12.

13.

14.

15.

16.

17.

117

metabolism, excretion and toxicity in drug discovery. Curr. Top. Med. Chem., 3(10), 1125–1154. DiMasi, J.A., Hansen, R.W., Grabowski, H.G. (2003) The price of innovation: new estimates of drug development costs. J. Health Econ., 22(2), 151–185. van De Waterbeemd, H., Smith, D.A., Beaumont, K., Walker, D.K. (2001) Property­based design: optimization of drug absorption and pharmacokinetics. J. Med. Chem., 44(9), 1313–1333. Raunio, H., Taavitsainen, P., Honkakoski, P., Juvonen, R., Pelkonen, O. (2004) In vitro methods in the prediction of kinetics of drugs: focus on drug metabolism. Altern. Lab. Anim., 32(4), 425–430. Miners, J.O., Smith, P.A., Sorich, M.J., McKinnon, R.A., Mackenzie, P.I. (2004) Predicting human drug glucuroni­ dation parameters: application of in vitro and in silico modeling approaches. Annu. Rev. Pharmacol. Toxicol., 44, 1–25. Poggesi, I. (2004) Predicting human pharmacokinetics from preclinical data. Curr. Opin. Drug Discov. Devel., 7(1), 100–111. Obach, R.S., Baxter, J.G., Liston, T.E., Silber, B.M., Jones, B.C., MacIntyre, F., Rance, D.J., Wastall, P. (1997) The pre­ diction of human pharmacokinetic parameters from pre­ clinical and in vitro metabolism data. J. Pharmacol. Exp. Ther., 283(1), 46–58. Rowland, M., Tozer, T. (1995) Clinical Pharmacokinetics, 3rd ed. Philadelphia: Lippincott Williams & Wilkins. Gibaldi, M., Perrier, D. (1982) Pharmacokinetics, 2nd ed. New York: Marcel Dekker, Inc. Hosokawa, M., Maki, T., Satoh, T. (1990) Characterization of molecular species of liver microsomal carboxylesterases of several animal species and humans. Arch. Biochem. Biophys., 277(2), 219–227. Wells, T.G., Mortensen, M.E., Dietrich, A., Walson, P.D., Blasier, D., Kearns, G.L. (1994) Comparison of the phar­ macokinetics of naproxen tablets and suspension in chil­ dren. J. Clin. Pharmacol., 34(1), 30–33. Callaghan, J.T., Bergstrom, R.F., Ptak, L.R., Beasley, C.M. (1999) Olanzapine. Pharmacokinetic and pharmacody­ namic proile. Clin. Pharmacokinet., 37(3), 177–193. Schentag, J.J. (1999) Antimicrobial action and pharmacokinetics/pharmacodynamics: the use of AUIC to improve eficacy and avoid resistance. J. Chemother., 11(6), 426–439. MacGowan, A.P. (2001) Role of pharmacokinetics and pharmacodynamics: does the dose matter? Clin. Infect. Dis., 33(Suppl. 3), S238–S239. Ritschel, W.A., Vachharajani, N.N., Johnson, R.D., Hussain, A.S. (1992) The allometric approach for interspecies scaling of pharmacokinetic parameters. Comp. Biochem. Physiol. C, 103(2), 249–253. Peters, S.A., Ungell, A.L., Dolgos, H. (2009) Physiologically based pharmacokinetic (PBPK) modeling and simulation: applications in lead optimization. Curr. Opin. Drug Discov. Devel., 12(4), 509–518.

118

PRECLINICAL PHARMACOKINETICS: INDUSTRIAL PERSPECTIVE

18. Derendorf, H., Hochhaus, G., Meibohm, B., Mollmann, H., Barth, J. (1998) Pharmacokinetics and pharmacodynamics of inhaled corticosteroids. J. Allergy Clin. Immunol., 101(4 Pt 2), S440–S446. 19. Beierle, I., Meibohm, B., Derendorf, H. (1999) Gender differences in pharmacokinetics and pharmacodynamics. Int. J. Clin. Pharmacol. Ther., 37(11), 529–547. 20. Lees, P., Cunningham, F.M., Elliott, J. (2004) Principles of pharmacodynamics and their applications in veterinary pharmacology. J. Vet. Pharmacol. Ther., 27(6), 397–414. 21. Aarons, L. (1999) Software for population pharmacokinet­ ics and pharmacodynamics. Clin. Pharmacokinet., 36(4), 255–264. 22. Steimer, J.L., Ebelin, M.E., Van Bree, J. (1993) Pharmacokinetic and pharmacodynamic data and models in clinical trials. Eur. J. Drug Metab. Pharmacokinet., 18(1), 61–76. 23. Danhof, M., Mandema, J.W., Hoogerkamp, A., Mathot, R.A. (1993) Pharmacokinetic­pharmacodynamic model­ ling in pre­clinical investigations: principles and perspec­ tives. Eur. J. Drug Metab. Pharmacokinet., 18(1), 41–47.

24. Jemal, M., Xia, Y.Q. (2006) LC­MS Development strate­ gies for quantitative bioanalysis. Curr. Drug Metab., 7(5), 491–502. 25. O’Connor, D. (2002) Automated sample preparation and LC­MS for high­throughput ADME quantiication. Curr. Opin. Drug Discov. Devel., 5(1), 52–58. 26. Shou, W.Z., Zhang, J. (2006) Recent development in high­ throughput bioanalytical support for in vitro ADMET proiling. Expert Opin. Drug Metab. Toxicol., 6(3), 321– 336. 27. Hsieh, Y., Korfmacher, W.A. (2006) Increasing speed and throughput when using HPLC­MS/MS systems for drug metabolism and pharmacokinetic screening. Curr. Drug Metab., 7(5), 479–489. 28. Smith, N.F., Raynaud, F.I., Workman, P. (2007) The applica­ tion of cassette dosing for pharmacokinetic screening in small­molecule cancer drug discovery. Mol. Cancer Ther., 6(2), 428–440. 29. Singh, S.S. (2006) Preclinical pharmacokinetics: an approach towards safer and eficacious drugs. Curr. Drug Metab., 7(2), 165–182.

6 LC-MS IN DRUG METABOLISM AND PHARMACOKINETICS: A PHARMACEUTICAL INDUSTRY PERSPECTIVE Wenying Jian, Wilson Shou, Richard W. Edom, Naidong Weng, and Mingshe Zhu

6.1

INTRODUCTION

Drug metabolism and pharmacokinetics (DMPK) is a ield of research dedicated to the study of the metabolism and disposition of drugs administered to animals or humans. Pharmacokinetics (PK) involves characterization of the time-dependent disposition of drugs throughout the body which is determined by the rate and extent of absorption, distribution, metabolism, and excretion (ADME). The fundamental purpose of PK investigations is to interpret the quantitative relationship between the dose and the measured concentrations of drugs in biological luids (e.g., plasma or blood), which is correlated to the pharmacological and/or toxicological effects elicited. Drug metabolism (DM) is a process by which drugs are converted into more hydrophilic species by metabolizing enzymes to facilitate their elimination from the body. By altering the chemical structure of the drug, the pharmacological activity, duration of activity, and toxicity may be changed. DM studies investigate how drugs are metabolized, how drugs and/ or their metabolites are absorbed, distributed, and eliminated, what enzymes and/or transporters are involved in these processes, and how drugs and metabolites affect metabolizing enzymes and drug transporters. Unfavorable DMPK properties are a prominent reason for drug candidate failure. Therefore, the modern pharmaceutical industry allocates signiicant resources to optimize the properties of candidate drugs starting from the early stages of drug discovery and continuing

throughout the entire drug development process (Figure 6.1) [1–6]. All DMPK studies can be categorized into three main functional types. 1. Quantitative Analysis of Drugs, Metabolites, or Substrates of Drug-Metabolizing Enzymes or Drug Transporting Proteins for In Vitro Screening Systems. Typical experiments include: metabolic stability, cytochrome P450 (CYP) inhibition/ induction, permeability, protein binding, transporter substrate/inhibition, and reaction phenotyping. The ever-growing number of new chemical entities (NCEs) generated during early drug discovery demands high-speed methodology to evaluate their DMPK properties. In vitro methods offer a rapid means to screen thousands of drug candidates for potential DMPK problems. The information provided by these studies can assist medicinal chemists to optimize and to select compounds with desirable DMPK properties for further development. 2. Quantitative Analysis of Drugs and Their Metabolites in Biological Fluids such as Plasma, Blood, or Urine, from Animals and Humans, Namely Bioanalysis. Bioanalytical support is required for almost all in vivo studies throughout the entire drug discovery and development process, including early discovery PK property screening, toxicokinetic (TK) studies, human PK or pharmacodynamic (PD) studies, drug–drug interaction

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

119

120

LC-MS IN DRUG METABOLISM AND PHARMACOKINETICS: A PHARMACEUTICAL INDUSTRY PERSPECTIVE

Discovery

ƒ Metabolic stability ƒ CYP inhibition ƒ CYP induction ƒ PAMPA ƒ Caco-2

IND FIH

Development

NDA

In Vitro ADME and DDI ƒ Definitive CYP inhibition and induction ƒ P-gp studies ƒ Protein binding

ƒ Definitive metabolizing enzyme reaction phenotyping ƒ Transporter inhibition and substrate analysis ƒ 14C-ADME in animals ƒ Tissue distribution ƒ 14C- ADME in humans

Biotransformation

ƒ Metabolic soft spots ƒ In vitro metabolism ƒ Reactive metabolites across species ƒ Profiling and quantitative estimation of ƒ BDC rat ADME plasma metabolites in humans and Tox species

Discovery PK ƒ PK screening ƒ PK optimization PK/Bioanalysis

Preclinical PK ƒ Clinical candidate selection and characterization ƒ Specie comparisons and multiple dose ƒ Formulation selection and food effect test

Clinical PK ƒ Safety evaluation ƒ PK/PD evaluation ƒ Drug−drug interaction ƒ Special populations ƒ Liver or kidney impairment studies

Toxicokinetics in Tox studies

FIGURE 6.1 A summary of DMPK studies in drug discovery and development. IND, investigational new drug; FIH, irst in human; NDA, New Drug Application.

studies (DDI), and so on. Concentration data for drugs and their metabolites generated from bioanalysis play an important role in the evaluation of PK parameters, formulation optimization, safety assessment, and interpretation of eficacy and toxicological observations. 3. The Study of Drug Biotransformation and Disposition By Proiling and Structural Characterization of In Vitro and In Vivo Metabolites and Determination of Their Concentrations Relative to the Parent Drug in Incubations, Plasma, Urine, Bile and Feces. In early discovery, in vitro systems, mainly liver microsomes or hepatocytes, are used to address biotransformation related issues in lead optimization, such as reactive metabolite screening and metabolic soft-spot analysis (Figure 6.1). The information from these assays can be used by medicinal chemists to rationally synthesize compounds to be more metabolically stable or to have less potential to form reactive metabolites. In the late discovery and development stages, metabolic pathways elucidated from in vitro incubations across species, and from animals in vivo, provide a scientiic basis for selection of toxicological species and prediction of metabolism and PK in human. After the drug candidate enters clinical studies, comparison of the metabolic proiles in humans with those in animals provides information regarding metabolite exposure and

safety assessment of the metabolites. Finally, radioactivity analysis in tissue distribution and radiolabeled ADME studies in animals and humans elucidates distribution and disposition of the drug in the body. It is a challenging task to detect, characterize, and quantitatively analyze drugs and their metabolites in biological matrices due to the excess amounts of protein, lipids, and other endogenous components that might interfere with bioanalytical assays. In the past, liquid chromatographic (LC) techniques coupled with ultraviolet (UV), luorescence, or electrochemical detection were the predominant methods utilized in DMPK studies [7]. However, these detection methods were usually low throughput and had limited sensitivity and poor speciicity. In comparison, mass spectrometry (MS) provides superior sensitivity and selectivity. MS-based techniques also afford the capability to determine the molecular weight (MW) as well as fragmentation patterns which can provide valuable information for structure elucidation of analytes. Gas chromatography (GC)-MS has been a popular technique for analyzing trace components in complex samples. However, in order to be compatible for GC analysis, the analytes must be volatile and thermally stable, which often requires a time-consuming and tedious process of chemical derivatization of nonvolatile compounds. Since the early 1990s, interfacing of LC technologies with MS

INSTRUMENTATION UTILIZED FOR DMPK STUDIES

has been revolutionized by the application of atmospheric pressure ionization (API) techniques, and liquid chromatography–mass spectrometry (LC-MS) analysis has played a predominant role in the pharmaceutical industry for the qualitative and quantitative analysis of drugs and their metabolites in support of DMPK studies [8–12]. 6.2 INSTRUMENTATION UTILIZED FOR DMPK STUDIES 6.2.1

Mass Spectrometry

Currently, four types of mass analyzers are commonly utilized for qualitative and quantitative DMPK applications: triple quadrupole, ion trap and linear ion trap, triple quadrupole linear ion trap, and high-resolution mass spectrometers (Table 6.1) [13–15]. Scan functions, major applications, and signiicant limitations of these instruments are included in Table 6.1.

TABLE 6.1

121

6.2.1.1 Triple Quadrupole Mass Spectrometry Quadrupole mass spectrometers are widely used in the pharmaceutical industry for quantitative studies. The Q1 mass analyzer ilters ions with the desired mass-tocharge ratio (m/z) so that these ions are fragmented within Q2 via collision-induced dissociation (CID), and their product ions are resolved by Q3 before reaching the detector [15]. Triple quadrupole instruments provide superior selectivity and sensitivity against background noise originating from chemical and biological matrices via selected reaction monitoring (SRM), also called multiple reaction monitoring (MRM), which speciically monitor selected product ions generated via CID from targeted molecular ions. SRM is routinely used for in vitro ADME screening studies and quantitative bioanalysis [16,17]. Precursor ion (PI) scanning generates a mass spectrum which is comprised only of molecular ions that can generate a certain desired product ion. Neutral loss (NL) spectrum consists of molecular ions that can undergo the loss of a speciic neutral

Mass Spectrometers That Are Used in Support of DMPK Studies

Instrument

Scan Function

Triple quadrupole

• • •

MRM PI NL

Major Application •



Ion trap and linear ion trap

Triple quadrupole linear ion trap

• •

• • • • •

High resolution

• •

Full scan Datadependent MSn

MRM PI NL MIM Datadependent MS2 or MS3 Full scan Datadependent MS2 or MSn



• •





MRM for quantitative analysis PI and NL for detection of uncommon metabolites Metabolite identiication and structure elucidation

Advantage •

MRM is highly sensitive and selective

Disadvantage •

• •

• • •

High scan speed Relatively generic data acquisition methods MSn capability for structural elucidation





Quantitative analysis Fast proiling and identiication of various metabolites



Multiple scan functions suitable for both quantiication and metabolite identiication



Comprehensive metabolite detection and structure elucidation Fast quantitative analysis



Accurate mass and high-resolution information suitable for metabolite detection and structure elucidation

• •

MRM needs knowledge of MW and fragmentation Relatively slow scan PI and NL not suitable for fast metabolite proiling Low sensitivity of quantitative analysis due to high background noise Not suitable in detecting uncommon metabolites No high-resolution function for elemental composition and structure elucidation

Expensive Relatively complicated maintenance and operation

122

LC-MS IN DRUG METABOLISM AND PHARMACOKINETICS: A PHARMACEUTICAL INDUSTRY PERSPECTIVE

Modern triple quadrupole instruments can rapidly switch between many SRM transitions within a single acquisition cycle (MRM), enabling the measurement of multiple analytes simultaneously. It is critical to choose a unique molecular ion and product ion pair (the “SRM transition”) for speciic monitoring of the targeted analyte. An example of liquid chromatography-multiple reaction monitoring (LC-MRM) analysis of midazolam and its two major metabolites, 1′-hydroxymidazolam and 4-hydroxymidazolam, is shown in Figure 6.2 [18].

fragment [9]. Ten years ago, PI and NL scanning with a triple quadrupole instrument was routinely employed for detection of drug metabolites, especially phase II conjugate metabolites such as glutathione (GSH) or N-acetyl cysteine conjugates, which exhibit predictable fragmentation patterns regardless of their MW. Recently, ion trap, triple quadrupole linear ion trap and high-resolution mass spectrometers have played more important roles in drug metabolite proiling and identiication.

3000

OH

2500 Intensity, cps

1¢‐hydroxymidazolam m/z 342 Æ 203

3.36

  N

2000

N 1500

297

1000 Cl

N

203

F

500 0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Intensity, cps

Time, min

1000 900 800 700 600 500 400 300 200 100 0

2.88

4‐hydroxymidazolam m/z 342 Æ 297 

Interference from 1¢‐hydroxymidazolam

3.36

N N OH N

Cl

297 F 3.53

2.50 2.58

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Time, min

3.43

6.4e4 6.0e4

N

Intensity, cps

5.0e4

291

4.0e4 3.0e4

Midazolam m/z 326 Æ 291

N N

Cl

2.0e4

F

1.0e4 0.0

0.5

1.0

1.5

2.0 2.5 Time, min

3.0

3.5

4.0

FIGURE 6.2 Mass chromatograms, structure, and MS/MS fragmentation of 1′-hydroxymidazolam (m/z 342 → 203), 4-hydroxymidazolam (m/z 342 → 297), and midazolam (m/z 326 → 291) from the analysis of an extracted monkey plasma sample [18].

INSTRUMENTATION UTILIZED FOR DMPK STUDIES

123

The parent drug, midazolam, is monitored by the transition m/z 326 → 291. The metabolites, 1′-hydroxymidazolam and 4-hydroxymidazolam, are isobaric and are monitored by the transitions m/z 342 → 203 and m/z 342 → 297, respectively. Since 1′-hydroxymidazolam also generates the product ion m/z 297, it produces a chromatographic peak in the MRM transition for 4-hydroxymidazolam, which could cause interference with quantitation of 4-hydroxymidazolam, if they were not chromatographically separated. Triple quadrupole instruments are the most commonly used type of mass spectrometer for quantitative analysis of drugs and metabolites for in vitro ADME screening and in vivo bioanalysis owing to their high sensitivity and selectivity in SRM mode for targeted detection and quantitation of known analytes. As depicted in Figure 6.3, quantitative analysis of drugs and their metabolites for in vitro ADME screening and in vivo bioanalysis share some common procedures, including sample preparation, chromatographic separation, MS analysis, and data analysis. Qualitative analysis of drugs and their metabolites in biotransformation studies involves much more diversiied instrumentation and platforms and will be discussed in detail in Section 6.5 [13,19].

scan and MS/MS scan) similar to an ion trap mass spectrometer [21,22]. Most importantly, MS, MRM, PI, and NL scans on Q-Trap instruments can serve as survey scans to trigger the information-dependent acquisition (IDA) of enhanced product ion (EPI) spectra with polarity switching, which provide product ion mass spectra with rich fragmentation information [23–27]. Multiple reaction monitoring–enhanced product ion (MRM-EPI) is able to perform as many as 200–500 MRM transitions to trigger recording of MS/ MS spectra, which has been utilized for detection and identiication of drug metabolites [24]. As an alternative, multiple ion monitoring–enhanced product ion (MIMEPI) derived from MRM-EPI was recently developed for screening of in vitro metabolites with sensitivity and duty cycle capabilities similar to those of MRM-EPI [25,28]. In MIM analysis, the same ion is monitored in both Q1 and Q3, and CID in Q2 is kept to a minimal value. The IDA methods, including PI-EPI, NL-EPI, MRM-EPI, and MIM-EPI, have been widely applied to reactive metabolite screening and metabolic soft-spot determination. The Q-Trap mass spectrometer has been considered a unique LC-MS platform, and it is well suited for both quantitative analysis and structural elucidation for DMPK studies.

6.2.1.2 Ion Trap and Linear Ion Trap Mass Spectrometers Ion trap and linear ion trap mass spectrometers are two of the cornerstone instruments for qualitative metabolite identiication [13,20]. Ion trap instruments feature full MS scanning and data-dependent tandem mass spectrometry (MSn ) acquisition, including list-, intensity-, and isotope pattern-dependent acquisitions. Ion trap instruments have higher scan speed and more sensitive full MS scanning than triple quadrupole instruments because of their ability to accumulate ions within the trap without ion iltering. Unlike triple quadrupole MS, ion trap instruments are not capable of recording either PI spectra or data relating to NLs. Ion trap instruments were mainly developed for drug metabolite proiling and identiication. Due to the recent advance of high-resolution mass spectrometers with respect to technology, operation, and price, highresolution instruments are promising to replace ion trap or linear ion trap instruments as a major MS platform for drug metabolite identiication in the pharmaceutical industry.

6.2.1.4 High-Resolution Mass Spectrometry Highresolution mass spectrometry (HR-MS), which can provide resolution greater than 10,000 (full width at half maximum height) and accurate mass measurement less than 10 ppm, has become the most important LC-MS platform in drug metabolite identiication. The exact mass measurement of metabolite molecules and their fragment ions is highly useful for conirmation of elemental composition [19]. In combination with postacquisition data mining tools, especially mass defect iltering (MDF), they provide a highly eficient and speciic tool for both detection and structural characterization of drug metabolites in complex biological matrices [29]. There are three main types of high-resolution mass spectrometers in modern DMPK laboratories.

6.2.1.3 Triple Quadrupole Linear Ion Trap Mass Spectrometry Hybrid triple quadrupole linear ion trap (Q-Trap,AB Sciex, Foster City, CA) mass spectrometers provide scan functions identical to those of classical triple quadrupole (MRM, PI, and NL scanning). However, the Q-Trap also has scan functions (full MS

1. Quadrupole Time-of-Flight (Q-TOF): Q-TOFs combine the Q1 mass ilter and Q2 collision cell from triple quadrupole mass spectrometers with a time-of-light (TOF) region as the second mass analyzer. These instruments can operate as tandem mass spectrometers with the advantage of providing accurate mass measurements of the product ions. Q-TOF instruments can conduct datadependent scanning for MS and MS2 (product ion scan), but not MSn. 2. Fourier Transform Ion Cyclotron Resonance Mass Spectrometry (FTICR-MS): FTICR-MS is based

124

LC-MS IN DRUG METABOLISM AND PHARMACOKINETICS: A PHARMACEUTICAL INDUSTRY PERSPECTIVE

in vivo Bioanalysis

in vitro ADME Screening

Sample Generation In Vitro Incubation

Sample Collection

Animal Dosing

Sample Extraction

Chromatographic Separation

Quad Mass Filter (Q3)

Quad Mass Filter (Q1)

Mass Spectrometry Analysis

Octopole 1

Lens 1 and 2

2.34 2851

100 % 0

2.32 109918

100 %

60 50 40 30 20 10 0

4.00

IS

Cmax

0 Analyte Area / IS Arear

• Metabolic stability • Permeability • CYP inhibition/induction • Protein binding …...

Analyte

2.00

2.00 Time

Plasma Con ncentration

Data Analysis

10 KV Detector

Collision Cell

4.00

Calibration Curve y = 0.06x + 0.0012

Absorption phase Elimination phase (t 1/2) Cmin AUC Tmax

0

1000 2000 3000 4000 Analyte Concentration

Time

5000

FIGURE 6.3 The in vitro ADME screening and in vivo bioanalytical process.

upon the different frequencies of cyclotron (circular) motion by ions with different m/z ratios in a strong ield and is one of the most sensitive ion detection methods with resolution typically ranging from 105 to 106 and mass accuracy B direction, whereas bidirectional assays yield values in both A > B and B > A directions. Since many active transporter proteins are expressed in the Caco-2 cell line, the ratios of PA>B to PB>A in bidirectional assays can be used to assess transporter involvement in drug permeability. If PA>B >> PB>A, then the compound is likely to be actively transported for uptake; if PB>A/PA>B > ∼3, then the compound is likely a substrate for eflux transporters. Additional assays using the same bidirectional format can be used to further identify whether the test compounds are substrates or inhibitors of the transporters, therefore providing liability assessments of transporter-related DDI potential as either victims (substrates) or perpetrators (inhibitors) [77]. Due to the high cost of cell culture, Caco-2 assays are usually used as a follow-up to PAMPA in ADME screening [78], and as a result, the sample burden for bioanalysis is not as heavy as for some irst-line assays, such as metabolic stability. There have been a number of reports in the literature that use automated optimization and single LC-MS/MS for sample analysis for Caco-2 assay support [46,79–81]. Nevertheless, Caco-2 samples pose a unique bioanalytical challenge. Unlike plasma or microsomal samples rich in proteins that help solubilize compounds and prevent adsorptive loss, Caco-2 samples are essentially aqueous buffer samples with very little protein. As a result, compounds with low solubility and/ or adsorption problems tend to exhibit poor recoveries in the assay due to precipitation and adsorptive losses [82,83]. An effective solution to this problem is the use of organic solvent to “catch” compounds immediately after incubation, but prior to analysis, in order to maintain solubility and prevent adsorptive loss to container surfaces. Another approach involves the addition of some protein such as bovine serum albumin (BSA) to the assay buffer system, thus reducing compound loss/ precipitation and improving recoveries [84]. 6.3.2.3 CYP Inhibition and Induction DDI are a major ADME liability, and therefore, they are increasingly examined for drug candidates in early discovery. The CYP family of enzymes is responsible for the metabolism of most drugs, and either CYP inhibition or, to a lesser extent, CYP induction, has been implicated in many observed DDI cases in the clinical setting. As

131

a result, CYP DDI assays, especially CYP inhibition assays, are widely used as a major component of in vitro ADME liability screening [85,86]. CYP inhibition assays are conducted by coincubating the test compound along with a probe substrate speciic for a particular CYP enzyme in either HLM or a recombinant CYP enzyme. Fluorescent substrates are typically used with recombinant enzymes in a 384- or 1536-well format as a irst-line high-throughput assay since the detection can be made with a plate reader [87,88]. However, luorescence CYP assays have some disadvantages such as interference from test compounds and lack of concordance with results generated by native substrates in microsomes for some compounds. On the other hand, CYP inhibition assays using HLM have been proven to have the best connectivity with in vivo results, however, they require speciic and sensitive analysis of probe metabolites by LC-MS/MS, and therefore, they do not have as high a throughput as luorescence assays [89,90]. Several in vitro systems are used to evaluate CYP induction potential. Hepatic cell lines transfected with a reporter gene are commonly used for high-throughput screening to assess CYP induction liability. Another approach for induction assessment is to use similar probe molecules as inhibition assays with an LC-MS/MS readout, but using fresh or cryopreserved human hepatocytes. The use of human hepatocytes for CYP induction assays pose both cost and availability challenges, and therefore they are primarily used as a follow-up assay for a small number of compounds. Bioanalytical support for DDI assays with LC-MS is distinctively different from support for most other ADME screens in that only the probe metabolite (instead of all test compounds) is monitored in all samples. As a result, no MS/MS method optimization (other than that for the probe metabolite) is needed. On the other hand, the LC-MS/MS sample burden for CYP inhibition assays is probably the heaviest among all in vitro ADME screening assays, as CYP assays are typically conducted in 384-well plates and involve full IC50 (half maximal inhibitory concentration) curves, and many times in a time-dependent fashion to measure inhibition potentials for both the test compound and its metabolite(s). As a result, bioanalytical research for CYP inhibition support has mostly focused on improving sample analysis speed and throughput. (Data handling and report generation are now mostly automated and do not represent a bottle neck.) Several assay and bioanalytical approaches used either individually or in combination have been reported for CYP inhibition sample analysis. One approach is to incubate multiple substrates for multiple isozymes in the same well, and use an SRM method with multiple MS/MS transitions to analyze all the metabolites and assess the inhibition

132

LC-MS IN DRUG METABOLISM AND PHARMACOKINETICS: A PHARMACEUTICAL INDUSTRY PERSPECTIVE

of all isozymes. This “cocktail” assay approach can dramatically reduce the number of analytical samples; however, it could also lead to dificulty in data interpretation due to interactions between the probes, as well as nonlinear enzyme kinetics resulting from the ixed concentration ratios of various isozymes in HLM. A variant to this approach is to still perform the assays separately, but pool the samples together postincubation for LC-MS/MS analysis. As long as the bioanalytical sensitivity is suficient to overcome the dilution factor due to pooling, this approach can effectively reduce sample numbers without negatively impacting assay quality. Yet another approach, called the “double cocktail,” uses mixed probes and mixed recombinant CYPs. Unlike cocktail assays using HLM, in this approach the relative abundances of recombinant enzymes can be optimized to ensure linear kinetics, therefore it can achieve a signiicant sample reduction without sacriicing assay data quality. Just like support for all other ADME screens, multiplexed LC-MS/MS [91,92], and more recently online SPE/MS/MS [93,94], are used to achieve an injection-to-injection speed as fast as 5∼15 s/sample to support CYP inhibition in a throughput approaching that of a luorescent plate reader but with far superior quality. In bioanalysis of CYP assay samples, the test compounds could sometimes cause signal suppression to the probe metabolite if they happen to coelute, which would in turn result in an apparent inhibition of metabolite production (false positive). Therefore, it is highly recommended to use a stable isotope-labeled (STIL) internal standard for each probe metabolite in order to minimize the effect of this potential analytical bias due to MS signal suppression. 6.3.2.4 Protein Binding The binding of NCEs to serum proteins is one of the important determinants of both PK and PD properties, since it is believed that only the free, or unbound, fraction of a drug can interact with protein targets and access various clearance pathways. Protein binding data is also essential in projecting the human dose from in vivo animal studies [95,96]. Two in vitro assay approaches, namely ultrailtration and equilibrium dialysis, are commonly used in ADME screening to determine the protein binding of drug candidates. Just as its name suggests, ultrailtration uses centrifugation at a speed of several thousand g to separate unbound drug from that bound to serum proteins through a membrane with a high MW cutoff (such as 30 k). The concentration of the drug in both the original serum and the ultrailtrate are then determined by LC-MS/MS to calculate the (%) free value [97,98]. Ultrailtration can be conducted quickly in a 96-well plate format, and therefore, it is amenable to high-throughput ADME screening. However, sometimes nonspeciic binding of

test compounds to the membrane can result in unreliable results due to poor recovery. The other major protein binding assay approach is equilibrium dialysis, whereby the serum sample is allowed to reach equilibrium with an isotonic buffer in a two-compartment dialysis chamber separated by a high MW cutoff membrane. The unbound drug can move freely between the two chambers, whereas the bound drug cannot cross the membrane from the serum side into the buffer side. After the concentrations of the drug in both the serum and buffer sides are determined by LC-MS/MS, the (%) free value can be calculated. Equilibrium dialysis does not suffer from recovery problems and is considered the gold standard method for protein binding determination. However, traditional dialysis methods require a large sample volume and are not easy to automate. More recently, several 96-well equilibrium dialysis devices have been developed to enable higherthroughput determination of protein binding in the ADME screening setting (Figure 6.6C) [99,100]. Bioanalytical support for protein binding has all the common characteristics of ADME bioanalysis; however, there are a few factors that require special consideration. First, serum samples from protein binding assays are in a more complex matrix than other in vitro sample matrices such as microsomes. Therefore, better sample extraction and chromatography methods are required to minimize matrix suppression and interference problems. Second, drugs may exhibit nonspeciic binding to the incubation container especially on the buffer side. Caution needs to be taken to avoid or mitigate this issue, such as choosing containers made of the proper material, or adding organic solvent to recover the drug in the end of incubation. Third, protein binding samples are unique in that two samples from different matrices (serum and buffer) exist for each compound assayed, and therefore, two calibration curves (one in each matrix) are usually required to accurately determine the concentrations of both samples (because the analyte response factors are usually different in the two matrices due to differential matrix effects in each). One ingenious solution to this problem is to use a mixed matrix dilution scheme. In this approach, a single calibration curve is irst constructed in a 9:1 serum : buffer matrix. Serum samples are then diluted with blank buffer, and buffer samples are diluted with blank serum in such a ratio that the inal composition of all diluted samples consists of 90% serum and 10% of buffer [100]. One further variant of this approach is to use mixed sera from multiple species (such as human, rat, mouse, etc.) instead of a single-species serum, so that serum and buffer samples from protein binding assays in multiple species can be all quantiied together using a single calibration curve [59]. This approach maintains the quality

LC-MS FOR IN VIVO BIOANALYSIS IN SUPPORT OF DMPK

of sample analysis, while achieving signiicant time and cost savings in preparing and injecting calibration curves in multiple matrices. Finally, for highly protein bound drugs, the drug concentration on the buffer side may be extremely low, and therefore, a high sensitivity assay may be required.

6.4 LC-MS FOR IN VIVO BIOANALYSIS IN SUPPORT OF DMPK 6.4.1

Bioanalytical Procedures

Bioanalysis is a ield of research dedicated to the quantitation of drugs and their metabolites in biological matrices, such as plasma, blood, urine, and tissue. In collaboration with various research functions, bioanalysis departments support many types of in vivo studies to provide concentration data for drugs and their metabolites, which plays an important role in the evaluation of PK parameters, formulation optimization, safety assessment, and interpretation of eficacy and toxicological observations. The general process of bioanalysis is depicted in Figure 6.3. Following animal dosing and biological sample collection, the typical steps in the bioanalytical procedure are sample preparation, chromatographic separation, MS detection, and data analysis, all of which are critical in determining assay quality [101–103]. 6.4.1.1 Sample Collection and Storage In general, plasma samples are the most commonly collected sample type for LC-MS/MS quantitation of drugs and their metabolites in preclinical and clinical studies. The procedure to obtain plasma samples involves drawing of the blood from animals or human subjects into vacutainers containing a suitable anticoagulant, separating plasma from blood cells by centrifugation, and transferring the plasma to a separate tube. The plasma samples are then kept frozen in a −20 or −70°C freezer. If transportation of the samples to a different site is required, the samples are shipped on dry ice to ensure the frozen condition. Whole blood samples are sometimes also needed to be analyzed in rare cases such as constant redistribution of drug or metabolite from blood cells into plasma. In addition, for drugs targeting blood cells, it is often requested to collect the cells and analyze the intracellular concentration of the drugs. There are specially designed collection tubes commercially available for collection of the blood cells. It is critical to ensure ex vivo stability of the targeted analytes in the biological matrices during sample collection and storage in order to accurately measure the concentrations at the time of collection. Instability of a drug or metabolite can be caused by chemical reaction,

133

enzymatic reaction, redistribution between plasma and blood cells, or a combination of the above factors. A very common example of chemical instability can be seen in acyl glucuronide conjugates [104,105]. A carboxyl group-containing drug may undergo phase II metabolism to form an acyl glucuronide conjugate, which can be unstable and undergo hydrolysis and/or positional rearrangement under neutral or slightly basic pH. Hydrolysis would lead to underestimation of the glucuronide conjugate, and more seriously, overestimation of the parent drug concentration. This reaction can be controlled by lowering the pH to less than 4 using citric acid buffer or ammonium formate buffer. Other common examples of chemical instability include oxidation of catechols, reduction of N-oxides, cis/trans isomerization, or reaction of sulfhydryl (–SH) compounds with another of its own molecule to form a dimer, or with endogenous components in biological matrix such as GSH or Cys to form mixed disulides [106]. Antioxidants, such as ascorbic acid or sodium metabisulite, are often used to prevent oxidation reactions for catechols. To prevent reduction of N-oxides, high temperature, strong acid, or strong base should be avoided in sample handling and processing. Isomerization can be prevented by controlling exposure to UV wavelengths of light, and by temperature control. Sulfhydryl compounds are often stabilized by reacting with a derivatizing agent to prevent its reaction with another −SH group [107]. In addition to chemical reactions, there are a variety of enzymes in biological matrix that can catalyze degradation of drugs and metabolites. For example, esterases can hydrolyze esters, lactones (which is the closed ring form of ester), or phosphoamides, and the reaction can be controlled by enzyme inhibitor such as NaF, PMSF (phenylmethylsulfonyl luoride), eserine, and so on. Finally, blood cells can act as a source of free drug or metabolite by releasing them into plasma, leading to artiicially increased concentrations. In this case, it is critical to separate plasma from blood cells immediately after blood collection. Prevention of hemolysis (rupturing of red blood cells) is also critically important. Overall, based on the cause and extent of instability, different procedures can be taken to prevent instability of the targeted analyte, including temperature control, pH control, antioxidant, enzyme inhibition, derivatization, and so on. Carefully designed experiments need to be conducted to evaluate the effectiveness of the stabilizing procedure in conditions experienced by the study samples. It is always preferable to evaluate stability using fresh plasma or blood because enzymatic or cellular systems denature slowly in aged plasma or blood, which may lead to underestimation of the instability. If instability is found to be severe, then it may be

134

LC-MS IN DRUG METABOLISM AND PHARMACOKINETICS: A PHARMACEUTICAL INDUSTRY PERSPECTIVE

necessary to precharge the stabilizer into blood collection vacutainers. Under this situation, it is important to conirm that vacuum was maintained during storage of the vacutainers before sample collection and that effectiveness of the stabilizer was preserved as well. Urine samples are often collected to elucidate renal clearance of the drug or metabolite. A unique challenge of urine assays is nonspeciic binding where a compound(s) adsorbs to the container wall [108]. The adsorption happens frequently in urine assays because urine lacks proteins and lipids that can bind to the analytes or solubilize lipophilic analytes. Therefore, urine bioanalytical assays tend to suffer from analyte losses more often than plasma assays. A variety of antiadsorptive agents including organic solvent, BSA, zwitterionic detergents such as CHAPS (3-[(3-Cholamidopropyl) dimethylammonio]-1-propanesulfonate), sodium dodecyl benzene sulfonate, β-cyclodextrin, Tween 80 (polysorbate 80), and Tween 20 (polysorbate 20) can be used to block the adsorption of the compound to the surface of the container. Adsorption should always be evaluated for a urine assay prior to the urine collection so that a correct urine collection procedure can be incorporated into the study protocol. 6.4.1.2 Sample Preparation Normally, biological samples are not directly compatible with LC-MS/MS analysis, and they need to be processed before delivery to the LC-MS system. Effective sample preparation procedures can “clean up” the sample and concentrate the analytes of interest. Conventional sample preparation is often the most labor-intensive and time-consuming step in a liquid chromatography–tandem mass spectrometry (LC-MS/MS) analytical method. Introduction of automated 96-well extraction techniques, and advances in sample preparation techniques, have greatly improved the eficiency of sample preparation. For matrices that do not contain a large amount of cellular components and protein, such as urine and bile, samples can often be directly injected onto the LC column following simple dilution, centrifugation, and/or iltration. For biological samples that contain higher amounts of protein, such as plasma, the protein can be precipitated by an organic solvent such as acetonitrile or methanol, followed by centrifugation and/or iltration to remove the precipitate. However, neither the “dilute and shoot” approach nor protein precipitation (PPT) may suficiently remove endogenous compounds such as salts, phospholipids, and fatty acids, resulting in signiicant matrix interference. Therefore, it is recommended to dilute the sample and inject a small volume to avoid matrix interference and column deterioration. On the other hand, PPT and “dilute and shoot” approaches are most simple, quick, cost-effective, and

readily automated in 96-well plate format. In addition, they are often the method of choice for metabolite proiling work because drug-related components are less likely to be lost or to decompose in the extraction process. SPE provides sample cleanup by selectively retaining the targeted analyte on the solid sorbent by hydrophobic, hydrophilic, or ionic interactions, while undesired matrix components are washed away. The most commonly used SPE sorbents are bonded silica-based materials. In addition, SPE packed with cross-linked polymers or graphitized carbon is also available. Currently, SPE in 96-well plate format is widely used in bioanalysis labs due to its compatibility with automation. Depending on the packing materials and the speciic extraction procedure, SPE mechanisms include reversed phase, normal phase, hydrophilic interaction chromatography (HILIC), ion-exchange, and mixed mode. In general, there are four steps for SPE extraction: (1) conditioning of the sorbent bed; (2) loading of the sample onto the sorbent to allow selective retention of the analytes; (3) washing the sorbent to remove the undesired matrix components; and (4) eluting the retained compounds using a strong solvent, which usually needs to be dried and reconstituted using appropriate solvent for injection on the LC-MS/MS system. SPE often provides better sample cleanup than PPT, but may be more time consuming and more expensive. Liquid–liquid extraction (LLE) gives excellent sample cleanup and has been shown to remove phospholipids, which are believed to be the major cause of MS detection interference, more effectively than PPT and SPE [109]. In the LLE process, the aqueous sample is mixed vigorously with an immiscible organic solvent (e.g., ethyl acetate, methyl tert-butyl ether, hexane), which exhibit preferential solubility to the targeted analyte. The organic layer containing the extracted analytes is then separated from the aqueous phase by centrifugation, transferred, and evaporated to dryness. The samples are then reconstituted into an appropriate solvent for injection onto the LC-MS/MS system. A key step in LLE is to neutralize the chemical charge on the analytes by pH manipulation to achieve preferable distribution of the analytes into the organic solvent. LLE can be conducted in 96-well plate format using automation. In this case, careful attention needs to be taken to avoid cross-contamination between wells, which can readily happen. LLE is more cost-effective than SPE but it may generate larger amounts of organic waste. Supported liquid–liquid extraction (SLE) is a newly marketed alternative to LLE. In SLE, the sample is absorbed into an inert solid support and the organic solvent is then passed through the support to extract the analyte. SLE is formatted in 96-well plates and

LC-MS FOR IN VIVO BIOANALYSIS IN SUPPORT OF DMPK

is more expensive than regular LLE due to the material cost. For extraction of proteineous samples such as plasma, it is critical to completely disrupt the drug–protein binding. Otherwise, only the free form or a partial fraction of the total drug content is extracted. For PPT extraction of highly protein bound drugs, a higher ratio of organic solvent/plasma should be used (e.g., 10:1). Acid can be added to the organic solvent to facilitate release of the drug as protein structure can be disrupted by altering pH. When SPE or LLE procedure is used, buffer is often added prior to sample extraction to disrupt the drug–protein binding. Because SPE or LLE conditions without addition of buffer are usually milder than PPT, the latter is often more favored as a generic method for highly protein bound drugs in order to fully extract the total fraction of the drugs. In ofline sample preparation, decomposition of certain chemical structures on the analyte may take place during extraction and evaporation, which may lead to loss of analyte, artiicial generation of metabolites, or artiicial regeneration of parent drug. Examples include conversion between basic nitrogens and their N-oxides, hydrolysis of esters, conversion between lactones and the open hydroxy acid compound generation of esters from carboxylate compounds when alcohols are used during extraction, hydrolysis and rearrangement of acyl glucuronides, generation of N-glycoside from primary amines, and so on [110]. In addition, analytes can also be lost due to adsorption to the surfaces of the container. Therefore, it is important to carefully choose an appropriate extraction procedure to avoid these issues. If possible, elimination of the evaporation step can provide the advantage of avoiding analyte conversion and/or adsorptive loss, as well as saving signiicant time. Heating conditions or extreme pH should also be avoided as much as possible to prevent chemical reactions. Another common issue is carryover if the liquid handler used to aliquot and transfer samples has ixed tips (nondisposable). Carefully optimized tip wash procedures need to be adopted to address this problem. Sample preparation can also be conducted by direct online extraction, which offers advantages over ofline extraction in that it involves minimized sample preparation, reduces loss of the analyte during extraction, and decreases the chance of artifacts, contamination, and dilution during the extraction [16,103]. A typical coniguration for online extraction consists of an autosampler, two sets of binary pumps, an extraction column, an analytical column, and a divert valve. In the loading/ extraction stage, a sample is injected onto the extraction column, and the analytes are retained on the head of the extraction column, while the undesired matrix components such as proteins and salts are removed and

135

diverted to waste. In the elution stage, the valve is switched so that the extraction column is in line with the analytical column. The analytes are eluted to the analytical column for separation and subsequent detection by the mass spectrometer. Commonly used on-line extraction procedures include the following: 1. Turbulent Flow Chromatography (TFC): Under turbulent low, a low MW analyte can rapidly transfer to the stationary phase while large molecules are lushed away. High low rates (4–6 mL/min) on narrow bore columns (typically, 1 × 50 mm) packed with large diameter particles (typically, 30–50 µm) are required to generate turbulent low while maintaining manageable system back pressure. 2. SPE Cartridges: Multiple-use or single-use SPE extraction cartridges based on reversed phase, ionexchange, or mixed-mode mechanisms are commercially available. The single-use cartridge system has the advantage of being less prone to column clogging and carryover. 3. Restricted Access Materials (RAM): Large molecules are excluded from retaining on the surface of RAM either by a physical diffusion barrier (pore size) or by a chemical diffusion barrier (bonded polymer/protein network), while small molecules are retained via hydrophobic interaction to the interior of the phase. 6.4.1.3 Chromatographic Separation It was once argued that extensive chromatographic separation is not necessary for bioanalytical methods because tandem mass spectrometric detection offered suficient selectivity. However, in vivo bioanalytical samples usually contain excessive amounts of endogenous materials such as salts, proteins, and lipids, and exogenous components such as dosing vehicles (e.g., polyethelyene glycol, cyclodextrin, Tween), all of which may artiicially suppress or enhance ionization if not suficiently separated from the analyte [111,112]. In addition, coeluting drugrelated components, such as prodrugs or metabolites, may compete with the parent drug for ionization and cause ion suppression. Conjugate metabolites (e.g., acyl glucuronide, sulfate conjugate), N-oxides, and prodrugs are also prone to undergo in-source conversion to generate ions that are identical to that of the parent drug, and therefore, lead to overestimation of the concentration of the parent drug [102,113]. Therefore, it is very important to achieve adequate chromatographic separation between target analytes and potential interference components in order to ensure the performance of a bioanalytical assay. Reversed-phase liquid chromatography (RPLC) is the most commonly used separation technique in

136

LC-MS IN DRUG METABOLISM AND PHARMACOKINETICS: A PHARMACEUTICAL INDUSTRY PERSPECTIVE

pharmaceutical analysis because it is suitable for compounds with a variety of properties. However, it is very challenging to retain polar compounds under traditional RPLC conditions, because analytes are retained based on hydrophobic interaction. HILIC is complimentary to RPLC in that it retains analytes based on hydrophilicity [114]. Therefore, HILIC can provide good retention and unique selectivity for polar compounds [115,116]. The highly volatile organic mobile phase used in HILIC can provide increased sensitivity with electrospray ionization mass spectrometry. In addition, the organic extracts from PPT, LLE, or SPE are directly compatible with the high organic content mobile phase in HILIC and can often be directly injected onto a HILIC column [117– 119]. This eliminates the time-consuming evaporation and reconstitution steps during preparation of biological samples. Normal-phase liquid chromatography (NPLC) can effectively retain and separate relatively nonpolar analytes such as fatty acids and hormones. In addition, NPLC is extensively used for chiral separations. Ion-exchange chromatography (IEC) and sizeexclusion chromatography (SEC) are two important modes of separation for large molecules, but so far, these approaches have seen limited application in DMPK studies for small molecule drug candidates. With the recent emergence of biomolecules such as peptides, proteins, and oligonucleotides as therapeutic agents, IEC and SEC have the potential to be the future “methods of choice” in support of DMPK studies. Chiral separation is often required for analysis of steroisomeric drugs. According to the Food and Drug Administration (FDA) guideline issued in 1992 for the development of new stereoisomeric drugs, in order to evaluate the PK of a single enantiomer, or a mixture of enantiomers, manufacturers should develop quantitative assays for the individual enantiomers in in vivo samples early in drug development [120]. Stereoisomers have identical MWs, and therefore, they cannot be differentiated by MS in most cases. Chromatographic separation must be achieved for the individual enantiomers, which will allow the assessment of potential interconversion, and the activity, toxicity, and DMPK properties of each alone. Chiral separation is achieved by the different interaction exhibited by isomeric analytes with the single enantiomers (chiral selector) immobilized on the chiral stationary phase (CSP) [121]. The most popular chiral selectors used for pharmaceutical application are polysaccharide-based derivatives due to their versatility, durability, and loading capacity [122]. Chiral separations using polysaccharide columns are typically performed using normal-phase conditions because the interactions involved in enantiomeric resolution are stronger in normal phase. Reversed-phase

chiral columns such as macrocyclic CSP, polysaccharidebased columns, and protein-based columns are also commercially available [123]. Column dimension, particle size, low rate, gradient system, injection solvent, and carryover are important factors to be considered for successful development of a bioanalytical assay: 1. Column Dimension and Particle Size: Narrowbore (∼2.0 mm diameter) columns with column lengths of 20–100 mm and packed with particle beads of 3–5 µm are routinely used in current bioanalytical labs. The small column diameter provides a low volumetric low rate and sample preconcentration, and therefore, improves ionization eficiency. The short column length gives reduced run times, typically between 2 and 8 min depending on the low rate and gradient system utilized. Smaller particles ( 184 is used to monitor the elution of PC and LPC to show if they are separated from the analyte of interest. This approach is based on the fact that all PC and LPC, upon high energy in-source CID, generate a common fragment ion of m/z 184, which is selected in Q1, passed through Q2 with low collision energy to minimize further fragmentation, and then detected in Q3. To manage the impact of matrix effects on the performance of bioanalytical assays, an often adopted approach is to use STIL internal standards, which are expected to experience ion suppression or enhancement to the same extent as the analytes, and therefore, compensate for matrix effects. However, STIL internal standards are expensive and time consuming to synthesize, and therefore, they are rarely used in the discovery stage of the drug development process. APCI as an alternative mode of ionization to ESI is known to be less subject to matrix effect [130]. However, one needs to keep in mind that APCI is not feasible for thermally liable compounds or metabolites that break down in the ion source. Alternatively, separation of the matrix effectcausing components from the targeted analyte may be achieved by changing the chromatographic conditions, such as mobile phase additives or pH. Finally, extensive sample cleanup is the ultimate solution for matrix effects. For example, LLE have been demonstrated to be more eficient than SPE and PPT to remove phospholipids. PPT ilter plates that can selectively remove phospholipids are also commercially available and will potentially provide improved eficiency for eliminating matrix effects [132]. However, their relatively higher cost is a disadvantage of these specially manufactured ilter plates. Another commonly observed issue for MS detection is in-source conversion of metabolites [102]. In either ESI or APCI mode of ionization, especially the latter, certain metabolites can undergo in-source CID to generate the molecular ion of the parent drug, thereby providing the same SRM transition as that of the drug. Examples of the conversion include lactone to hydroxy acid and vice versa, glucuronide to parent drug, sulfate to parent drug, N-oxide or S-oxide to parent drug, and disulide to sulfhydryl. Unless there is chromatographic separation between the metabolite and the parent drug, the amount of parent drug could be overestimated. In addition, isomeric or isobaric metabolites share identical SRM transitions and, if not separated, may cause

140

LC-MS IN DRUG METABOLISM AND PHARMACOKINETICS: A PHARMACEUTICAL INDUSTRY PERSPECTIVE

interference for each other in quantitation. Furthermore, metabolites with masses which are lower by 1 or 2 Da compared with the parent drug may cause interference due to M+1 or M+2 isotopic contributions if the metabolites are present in suficient concentration relative to the parent drug. Alternatively, metabolites higher by 1 or 2 Da could be interfered with by the isotopic distribution of the parent drug. Therefore, it is very important to systematically evaluate the bioanalytical assay for metabolite interference using study samples. A recommended approach is to analyze PPT processed pooled plasma samples using a comprehensive list of SRM transitions covering all the potential metabolites, especially phase I and phase II metabolites, that may undergo insource CID and those with masses lower than that of the parent drug by 1 or 2 Da. The SRM list can be easily created if the mass spectrometric system is equipped with a metabolite SRM table builder; otherwise, the table can be created manually. 6.4.1.5 Data Analysis Internal standard calibration is the most commonly used strategy for bioanalytical quantiication based on LC-MS/MS. In this strategy, an analog of the analyte of interest is used as the internal standard and is added in a constant amount into all samples, including calibration standard samples, which are spiked with differing amounts of analyte. After the LC-MS/MS process, the chromatographic peaks of the analyte and internal standard, at their expected retention times, are integrated by software to obtain peak areas. The ratios of the peak area of the analyte to that of the internal standard from the calibration standard samples are plotted against concentrations of the spiked analyte to generate a calibration curve. Most often, a linear or quadratic regression with a weighting factor of 1/x or 1/x2 is used to obtain best itting. The concentration of the analyte from an unknown sample can be calculated from its analyte : internal standard ratio using the calibration curve. An internal standard should be a compound whose structure is closely related to that of the analyte so that it will behave similarly during the bioanalysis process. Inclusion of an internal standard can compensate for variability during bioanalytical workup, such as potential loss of analyte during sample preparation, variation in injection volume, signal suppression/enhancement, and instrument response shift. An ideal internal standard is a STIL analog (such as 2H, 13C, or 15N labeled), which will be almost identical to the unlabeled analyte in chemical and physical properties. Use of a STIL internal standard has been shown to signiicantly improve the accuracy and precision of bioanalytical assays. Synthesis of STIL compounds is expensive and time consuming. Therefore, a structural analog is often employed as a more cost-effective alter-

native, especially in the early drug discovery stage. However, one needs to be aware that structural analog internal standards may not fully mimic the behavior of the analyte, which may result in unacceptable quantiication data. It is also noteworthy that STIL internal standards will not compensate for analyte loss due to instability or adsorption to the container prior to sample preparation, or due to low recovery caused by tight protein binding. Occasionally, STIL internal standards have been shown to chromatographically separate from the analyte, at least partially, which may cause them to experience different matrix effects [133]. Chromatographic separation is more common with 2 H-labeled analogs, especially when many deuterium atoms are incorporated into the structure to overcome the presence of halogens, such as chlorine, in the natural analog. Whenever a bioanalytical method is used to support preclinical or clinical studies for regulatory submission purposes, a method validation needs to be conducted to demonstrate that the method used for quantitative measurement of analytes in a given biological matrix is reliable and reproducible for the intended use. Method validation involves evaluation of accuracy, precision, linearity, selectivity, sensitivity/speciicity, reproducibility, stability, matrix effects, recovery, and carryover using calibration standards and quality control (QC) samples that are prepared by spiking a known amount of analyte into the intended blank biological matrix. FDA has issued a guidance document and a white paper which describe the procedures and acceptance criteria for bioanalytical methods [134,135]. In recent years, there is increasing concern that the performance of calibration standards and QCs may not adequately mimic that of the study samples, such as metabolites converting to parent drugs, protein-binding differences in patient samples, recovery issues, sample inhomogeneity, analyte instability, and matrix effects. Therefore, a more rigorous way to test the accuracy and reproducibility of a bioanalytical assay is to conduct a study sample reproducibility (ISR) evaluation, which now is recommended by FDA and has become a mandatory part of bioanalytical assay validation [136]. 6.4.2

Discovery and Development Bioanalysis

As a drug candidate progresses from discovery, to preclinical development, and on to clinical development, bioanalytical support plays a pivotal role in providing important data that allow decisions to be made at each critical stage. There are unique demands, challenges, and regulatory requirements for bioanalytical support at each stage of drug discovery and development process, which are summarized in Table 6.4.

LC-MS FOR IN VIVO BIOANALYSIS IN SUPPORT OF DMPK

TABLE 6.4

141

Comparison of Bioanalytical Support for Discovery, Preclinical, and Clinical Studies Discovery

Preclinical

Clinical

Regulatory requirement LC-MS/MS method strategy

Small set of samples per compound; very quick turnaround time Non-GLP

Moderate numbers of samples for dozen of compounds; relatively slow turnaround time GLP

Generic and fast; usually low sensitivity

Speciic for the compound; usually require extensive method development

Validation strategy

Minimal, abbreviated

Sample analysis strategy

Cassette dosing, sample pooling, automation, and multiplexing for high throughput

Extensive for multiple species and matrices Complicated automation and instrument coniguration not routinely used

Large numbers of samples for very few compounds; quick turnaround time Same practice and standards as GLP Speciic to the compound; Very sensitive method may be needed; usually require extensive method development but can leverage preclinical development Extensive but limited to human Automation, multiplexing, and high-speed analysis for quick release of data

Demand

In the discovery stage, a large number of drug candidates need to be rapidly screened for their drug-like properties. The ultimate goal is to generate data as quickly as possible to enable the rapid elimination of poor candidates. One strategy to increase the throughput of sample analysis has been to dose and analyze multiple compounds simultaneously (cassette dosing). Another strategy is to dose each compound separately, but pool the samples after collection. Both methods take advantage of the selectivity afforded by unique MS/MS fragmentation. The sample preparation technique and LC-MS/MS conditions in discovery bioanalysis are usually generic, which streamlines the automation and application to many drug candidates. Complicated instrumentation for improvement of analysis eficiency, such as multiplexing conigurations, is not unusual in discovery bioanalytical support. Validation of the method is usually minimal, and data review by a Quality Assurance unit are not needed. As the drug candidate moves from the discovery stage to preclinical development, the priority in bioanalytical switches from high throughput to high quality. Method reproducibility and method speciicity are essential for successful preclinical bioanalytical support. The bioanalytical activities and the data generated during assay validation and sample analysis are conducted under strict GLP (Good Laboratory Practice) regulations and are subject to rigorous scrutiny by regulatory agencies. A unique method needs to be developed and validated for each drug candidate to ensure the speciic requirements in the laboratory’s standard

operating procedures (SOPs), as well as in the regulatory guidance, are met. Automation and complicated instrument settings are used with caution and only after these technologies meet the challenges of vigorous validation. When a drug candidate moves to the clinical development stage, high throughput becomes essential due to the large number of samples and the demand for quick turnaround. Automation, online extraction, multiplexing, and higher speed LC (such as UHPLC) are frequently used for faster sample processing and analysis. Methods with very high sensitivity are often necessary because of the low initial doses used in human. Extensive method development may be needed to improve sensitivity and to overcome potential issues often associated with high sensitivity assays such as carryover and contamination. Although not strictly required, the same standards for bioanalytical conduct used for GLP studies are usually applied to clinical bioanalytical activities as well. In recent years, increased emphasis has been placed on quantitative characterization of drug metabolites during the drug discovery and development process. The FDA MIST (metabolites in safety testing) guidance and ICH (International Conference on Harmonisation) guidance regarding safety testing of drug metabolites have emphasized the need for comparison of exposure levels of major metabolites in human to those in animal species in order to avoid any potential safety risk associated with inadequate metabolite safety testing [137, 138]. It has become a consensus in the pharmaceutical

142

LC-MS IN DRUG METABOLISM AND PHARMACOKINETICS: A PHARMACEUTICAL INDUSTRY PERSPECTIVE

community that quantitation of metabolites would be best conducted using a lexible “tiered” approach [135]. It allows metabolite measurement in early drug development to be conducted using bioanalytical methods qualiied using criteria established a priori (e.g., accuracy, precision, stability) which are scientiically it for the purpose. As a drug candidate moves to further stages of development, the bioanalytical method will use increasingly more rigorous acceptance criteria, and will be fully validated in accordance with FDA guidance for late development stages. This tiered approach allows a better allocation of bionalytical resources to later stages in development when there is a greater likelihood of drug success. There are unique chromatographic challenges associated with metabolite quantitation. First of all, metabolites are commonly more polar than the parent drug and may show little retention and poor peak shape on reversed-phase columns, leading to compromised performance of bioanalytical assays. Second, multiple metabolites from a drug may possess an identical MW (isobaric), which, if not separated, may interfere with each other in LC-MS/MS measurement. In addition, incomplete chromatographic resolution between a parent drug and major metabolites, or between different metabolites, could cause ion suppression, compromising the sensitivity and accuracy of the metabolite assay. Therefore, in order to ensure accurate estimation of the concentration of targeted metabolites, it is critical to develop chromatographic conditions with a suficient separation factor. A favored choice is HILIC due to its resolution eficiency to separate polar compounds. Phase II conjugate metabolites are usually excluded from quantitative evaluation because they are generally considered to be pharmacologically inactive and readily excreted from the body. However, there are cases when conjugate metabolites are pharmacologically active (such as the 6-O-glucuronide of morphine or the sulfate conjugate of minoxidil) or pose toxicological concerns by forming reactive intermediates (such as acyl glucuronides), which lead to interest in quantitative evaluation of these conjugates [139]. An analytical standard of a conjugate is often dificult to synthesize due to issues such as instability. In the absence of a standard, a conjugate can be measured using an indirect method. In such a method, parent drug concentration in the samples is measured before and after hydrolyzing the conjugate using either enzymatic cleavage (e.g., glucuronidase or sulfatase) or a chemical method. The amount of conjugate can be estimated by subtracting the concentration obtained before hydrolysis from that after hydrolysis. Indirect methods are time consuming, and sometimes unreliable, because the hydrolysis can be incomplete due to interference by components in the matrix.

6.4.3 Case Study: Quantitation of Midazolam and Its Metabolites 1′-Hydroxymidazolam and 4-Hydroxymidazolam in Monkey Plasma 6.4.3.1 Sample Preparation Calibration standards and QCs were made from separate stock solutions of midazolam, 1′-hydroxymidazolam, and 4hydroxymidazolam (all at 100 µg/mL in 50:50 methanol : water). Pooled standards at midazolam/1′hydroxymidazolam/4-hydroxymidazolam concentrations of 0.1, 0.2, 0.5, 2, 10, 40, 80, and 100 ng/mL each were prepared in blank monkey plasma. QCs at levels of 0.3, 15, and 70 ng/mL each were also prepared. All standards and QCs were aliquoted into 2 mL polypropylene vials and stored at −20°C. Midazolam, 1′-hydroxymidazolam, and 4hydroxymidazolam were extracted from monkey plasma using an LLE method. An aliquot (200 µL) of blank plasma, the calibration standards, QC samples and study samples were irst transferred to a 16 × 125 mm glass screw-top tube. Internal standard, triazolam (50 µL, 100 ng/mL in 50:50 methanol : water), was then added to all tubes except the blanks. After addition of 50 µL of 2% ammonium hydroxide (v/v), 70:30 diethyl ether : hexane (v/v, 2 mL) was added to each tube. All tubes were vortex-mixed on a lat-bed vortexer (VWR International, West Chester, PA) for 2 min. All tubes were then centrifuged at 3000 rpm for 5 min on a Beckman Coulter J2-HS centrifuge (Beckman Coulter, Fullerton, CA). The aqueous layer of each tube was frozen in a dry ice/acetone bath and the organic layer was decanted into a set of 10 mL conical glass tubes. The organic solvent was dried under a gentle stream of nitrogen in a TurboVap™ concentrator (Zymark, Hopkinton, MA) set at 40°C and the residues were reconstituted in 100 µL of acetonitrile for injection onto the LC-MS/MS system. 6.4.3.2 Chromatographic Separation The LC system was a Shimadzu (Columbia, MD) series 10AD VP equipped with binary pumps, a degasser, and an SIL-HT autosampler. A Betasil silica column (50 × 3 mm, 5 µm) from Keystone Scientiic (Bellefonte, PA) operated under HILIC mode was used. The isocratic mobile phase was composed of 5:95 (v/v) A : B, where A was 0.05% triluoroacetic acid (TFA) in water (v/v) and B was 0.05% TFA in acetonitrile (v/v). The low rate was 0.6 mL/min and the injection volume was 10 µL. Separation was performed at ambient temperature. 6.4.3.3 MS Detection The experiments were conducted on an API 3000 triple quadrupole mass spectrometer (Applied Biosystems, Foster City, CA) equipped with its TurboIonspray interface operated

LC-MS FOR IN VIVO BIOANALYSIS IN SUPPORT OF DMPK ϴ ϳ

Concentration (ng/mL)

under positive mode. The Ionspray needle was maintained at 2.0 kV. The turbo gas temperature was 400°C and the auxiliary gas low was 8.0 L/min. Nebulizing gas, curtain gas, and collision gas lows were at instrument settings of 12, 8, and 4, respectively. The DP and focusing potential (FP) were maintained at 36 and 200 V, respectively. The mass spectrometer was operated under SRM mode with a collision energy of 38 eV. Both quadrupoles were maintained at unit resolution (0.7 mass unit width at half height). The transitions monitored were m/z 326 → 291 for midazolam, m/z 342 → 203 for 1′-hydroxymidazolam, m/z 342 → 297 for 4hydroxymidazolam, and m/z 343 → 308 for triazolam with a dwell time of 50 ms for each pair.

143

ϲ

midazolam PK Curve

ϱ ϰ

1’-hydroxymidazolam PK Curve

ϯ

4-hydroxymidazolam PK Curve

Ϯ ϭ Ϭ

6.4.3.4 Data Analysis Chromatographic peaks of the analyte and internal standard were integrated and regressed using Analyst® 1.4.1 software (Applied Biosystems). For each calibration standard, the peak area ratio of analyte : internal standard was determined. A linear regression describing the calibration curve was then calculated using the regression equation: y = m(x) + b, where y is equal to the peak area ratio of analyte : internal standard, m is equal to the slope of the calibration curve, x is equal to the concentration of analyte, and b is equal to the y-intercept of the calibration. Regression weighting (1/x2) was used to improve the goodness of it. For each QC and study sample, the peak area ratio of analyte : internal standard was determined and the concentration was calculated using the equation derived above. 6.4.3.5 Result and Discussion In this case study, midazolam, 1′-hydroxymidazolam and 4hydroxymidazolam were extracted from monkey plasma using an LLE method for maximum sample cleanup. An analog internal standard, triazolam, was employed and added at the beginning the extraction steps. Ammonium hydroxide was added to adjust the pH so that the anlaytes were neutralized for chemical charge, followed by extraction using organic solvent. The samples were reconstituted in 100% acetonitrile, which has a weaker elution strength than that of the isocratic mobile phase (95% acetonitrile containing 0.05% TFA) under the HILIC chromatographic conditions used in this study. This step facilitated on-column focusing of the analytes and avoided peak distortion. Separation of the three analytes was achieved using HILIC because they are relatively polar compounds. 1′-hydroxymidazolam and 4-hydroxymidazolam are isobaric and share a common SRM transition from m/z 342 to m/z 297, which was used to monitor 4-hydroxymidazolam (Figure 6.2). As a result, 1′hydroxymidazolam gives rise to an interference peak in

Ϭ

ϱ

ϭϬ

ϭϱ

ϮϬ

Ϯϱ

ϯϬ

Time after dose (h)

FIGURE 6.7 The PK proiles of midazolam, 1′-hydroxymidazolam and 4-hydroxymidazolam from a cynomolgus monkey following a single oral midazolam dose of 0.5 mg/kg [18].

the 4-hydroxymidazolam chromatogram. Therefore, it is important to achieve separation between the two metabolites in order to avoid overestimating the concentration of 4-hydroxymidazolam. As demonstrated in Figure 6.2, 4-hydroxymidazolam eluted at 2.88 min using the isocratic mobile phase on a silica column. 1′-hydroxymidazolam eluted at 3.36 min, which is baseline separated from 4-hydroxymidazolam. The calculated concentrations of the analytes were used to construct a concentration-time proile. Figure 6.7 shows concentration-time proiles of midazolam, 1′hydroxymidazolam, and 4-hydroxymidazolam from a cynomolgus monkey following a single oral midazolam dose of 0.5 mg/kg. The metabolites demonstrated a parallel concentration proile to the parent drug. The proiles were used to determine the pharmacokinetic parameters such as maximum concentration (Cmax), half-life (t1/2), area under the curve (AUC), and so on [18]. 6.4.4 Bioanalysis of Peptide and Protein Drugs Using LC-MS/MS With the advances of molecular biology and biochemistry, emerging large molecule medicines, such as peptides and proteins, are becoming an increasingly larger proportion of the development portfolio in the pharmaceutical industry. Currently, the bioanalytical methods for quantitation of peptides and proteins in support of pharmacokinetic studies are predominantly based on immunoassay methodologies. Although these antibodybased techniques are exceptionally sensitive (typically in the picogram per milliliter range), issues such as

144

LC-MS IN DRUG METABOLISM AND PHARMACOKINETICS: A PHARMACEUTICAL INDUSTRY PERSPECTIVE

cross-reactivity, long development time, and high variability have stimulated interest in an equally sensitive, though more selective and controlled assay platform. The proven capability of targeted MS/MS analysis, considered the “golden standard” for small molecule quantitation, offers an attractive solution to the limitations associated with antibody-based detection [140–142]. LC-MS/MS based assays have several features that distinguish them from conventional immunoassays [143]. First, MS provides much higher structural speciicity, which is a potential quality advantage in cases when an immunoassay is subject to cross-reactivity or interference. For example, the presence of endogenous antibodies directed against the therapeutic protein could interfere with an antibody-based assay, while an LC-MS/ MS-based assay would be less prone to this issue. Second, LC-MS/MS assays require less development time. Immunoassays require a speciic antibody for every new protein analyte. The antibody production and selection process is itself resource and time consuming, contributing to long assay development time which may take up to 5 months for a typical validated enzymelinked immunosorbent assay (ELISA) used for toxicological or clinical studies. In contrast, LC-MS/MS-based assays usually take days to weeks to develop, which is very attractive for discovery stage activities before committing the very substantial time and resources required to create an immunoassay. Third, LC-MS/MS-based assays can provide excellent assay linearity over several orders of magnitude, whereas ligand-binding assays are often limited in dynamic range and often show nonlinear responses. This is because immune-based detection relies on the binding interaction of two macromolecules, while LC-MS/MS detection is direct and only depends on the chemical nature of the analytes. Last, the ability to assay multiple MRM transitions in a single LC-MS/ MS experiment is another strength not offered by standard immunoassays. Multiple proteins can be analyzed in a single LC-MS/MS run for higher throughput. Proteins and large peptides are usually too large to be directly quantiied using standard LC-MS/MS assays. A typical worklow for the quantitation of peptides and proteins by targeted MS/MS involves the enzymatic cleavage of the large molecule followed by the LC-MS/ MS analysis of the one or more peptides as a surrogate (Figure 6.8). Commonly, trypsin is used to digest a protein into its constituent peptides. Trypsin cleaves proteins’ C-terminal of arginine and lysine, except when blocked by an adjacent C-terminal proline residue [144]. The average tryptic peptide is ideal for LC-MS/MS analysis due to its relatively small size and single Cterminal located basic residue. Positive-mode ESI of tryptic peptides will normally produce M+2H+ and M+3H+ molecular ions with m/z values between 400 and

1000, and upon CID, will show a predominantly Cterminal y- and N-terminal b-fragment ion series (Figure 6.8) [145]. Several peptides (three to ive) which are predicted to uniquely and stoichiometrically represent the targeted protein, so-called signature peptides, are monitored by MRM for highly speciic detection. A potential drawback of “signature peptide” approach is the lack of differentiation between intact protein/peptide and in vivo degraded/metabolized fragments which may still contain the same amino acids sequences. One such example is PEGylated (polyethylene glycolyated) proteins and their de-PEGylated components. In one study, a full-length PEGylated protein (conjugated with PEG at C-terminus) was analyzed by immunocapture using an anti-PEG antibody, followed by trypsin digestion, and LC-MS/MS analysis of a signature peptide located at the N-terminus [146]. Complexity of the plasma proteome, in which the dynamic range of protein abundance ranges 11 orders of magnitude, presents a great challenge to the quantitation of selected proteins in plasma. Enrichment procedures prior to trypsin digestion are often needed for reducing interferences and improving sensitivity for LC-MS/MS analysis of a targeted protein drug, especially those circulating at low concentrations. One such enrichment procedure is depletion of high-abundance serum proteins. It is estimated that 99% of serum’s total protein mass is due to the top 20 most abundant protein species [147]. Albumin, the most abundant protein, is present at ~60 mg/mL. Depletion kits using chemical afinity or immune afinity for removal of human serum albumin have been shown to reduce protein content by 50%. In addition to depletion, speciic or nonspeciic enrichment strategies can be applied, such as capture by antibodies for the target protein drug or signature peptides, protein A afinity enrichment for capture of antibody drugs containing the Fc region of immunoglobulin G (IgG), or capture by antiPEG antibody for PEGylated proteins, and so on [148,149]. In some cases, gel electrophoresis separation or SEC/iltering is necessary to fractionate the protein components for further reduction of sample complexity before subsequent digestion and chromatographic separation. Reversed-phase chromatographic conditions have been commonly used for separation of peptides [140]. Alternatively, HILIC conditions can provide good retention for hydrophilic peptides and improved ionization eficiency. Two-dimensional LC using an orthogonal separation mechanism such as ion-exchange chromatography–reversed-phase liquid chromatography (IEC-RPLC) or reversed-phase liquid chromatography– hydrophilic interaction chromatography (RPLC-HILIC) have also been utilized to extensively clean up the sample and thus improve the sensitivity and reproducibility of the assays [150]. The column particle size and

LC-MS FOR IN VIVO BIOANALYSIS IN SUPPORT OF DMPK

^LJŶƚŚĞƐŝnjĞŝƐŽƚŽƉĞ ͲůĂďĞůĞĚƐŝŐŶĂƚƵƌĞ ƉĞƉƟĚĞĂƐŝŶƚĞƌŶĂů ƐƚĂŶĚĂƌĚ

ĞĮŶĞƐŝŐŶĂƚƵƌĞ ƉĞƉƟĚĞƐĨŽƌƉƌŽƚĞŝŶ ĐĂŶĚŝĚĂƚĞ

[

[

H2N

CH

C

R1

D

\ ]

O H N

CH

C

D F

] O

H N

H C

R2

E

Authentic peptide

\ ]

O

Spike internal standard

[

\

C

D F

Trypsin digestion O

H N

H C

R3

E

145

C

OH

R4

>ͲDZDĂŶĂůLJƐŝƐ

E F

FIGURE 6.8 The basic process of quantitative analysis of protein using LC-MS/MS strategy and nomenclature of common ion types after fragmentation of peptides by MS/MS. Cleavage at the amide bonds generates b- and y-ions, containing the N- or Cterminus, respectively, plus one or more additional amino acid residues. Ion pairs formed after cleavage at other positions in the backbone can be a- and x-ions or c- and z-ions.

dimensions commonly used for peptide quantitation are generally the same as those for small molecule applications. Capillary columns (75 µm internal diameter [ID]) have been used for peptide chromatography for increased ionization eficiency of ESI by providing a low volumetric low rate (a typical low rate for a 75 µm ID column is 250 nL/min) and a higher analyte concentration in the column eluent. A drawback of capillary LC is the need for complex instrumentation for accurate delivery of low at the nL/min level. Conventional LC packing materials with 60–150 Å pore sizes are not generally suitable for large molecule separations, as the analytes are not able to access the surface area within these pores. In addition, the smaller pores can become fouled with strongly retained matrix components. LC particles with larger pore sizes (∼300 Å) have been specially developed for separation of large molecules for better retention and column robustness. Ion-pairing agents, most often TFA, are commonly used as mobile phase additives for peptide chromatographic separations due to their ability to improve peak shape by suppressing undesirable interaction between analytes and the stationary phase. In addition, they promote retention of peptides on RPLC by ion-pairing with analytes, thereby increasing their hydrophobicity. Different factors might inluence the ionization eficiency of peptides and thus effect the sensitivity, accu-

racy, and precision of the assay. Sometimes, small alternations in matrix interferences, mobile phase composition, or mass spectrometer operating parameters can signiicantly inluence the abundance of speciic ion types (i.e., different charge states or adduct ions). Measuring the sum of multiple peptides, each with multiple MRM transitions for different charge states, may reduce the variability in ionization eficiency, and therefore, improve the reproducibility of the assay. TFA, the ion-pairing agent most widely used for separation of peptides, is known to suppress ESI signal intensity due to its ability to form gas phase ion pairs with positively charged ions. Postcolumn infusion of propionic acid and isopropanol has been shown to alleviate ion suppression caused by TFA [151]. Alternatively, acetic acid (0.5%) or propionic acid (1%) directly added to the TFAcontaining mobile phase can also effectively reduce ion suppression [152]. A well-designed internal standard strategy is critical for the successful development of an LC-MS/MS-based protein quantitation assay because of the high variability usually exhibited by peptides and proteins during sample processing, chromatography, and ionization. STIL signature peptides have been commonly used as internal standards, and they are typically added at an early stage in sample processing (e.g., during enzymatic digestion; Figure 6.8). However, variability in the recovery

146

LC-MS IN DRUG METABOLISM AND PHARMACOKINETICS: A PHARMACEUTICAL INDUSTRY PERSPECTIVE

of the targeted proteins from the enrichment steps and variability in the digestion eficiency in the plasma matrix are not tracked by the internal standard in this case. If a signature peptide is not detected in an MRM assay for a certain sample, then it is often unclear if this is because of truly low concentration, or poor recovery from enrichment steps, or poor enzymatic digestion [153]. Employment of a STIL recombinant whole protein internal standard added at the start of sample processing, or even during sample collection, will provide better control over all the steps of the entire worklow and improve the accuracy and precision of the bioanalytical assay for protein drug quantitation. In the absence of an isotope-labeled whole protein, an analog protein may also be used as internal standard. In summary, despite the signiicant challenges associated with LC-MS/MS-based assays for protein drug quantitation, these assays still show promise as an emerging complementary technique for protein bioanalysis.

6.5 LC-MS IN DRUG METABOLITE PROFILING AND IDENTIFICATION In general, drug metabolite identiication using LC-MS techniques involves three steps: (1) detection of drug metabolite ions in a biological matrix using a special scanning method followed by data processing, (2) acquisition of their product ion spectra, and (3) structural elucidation of metabolites based on their MWs (in some cases, elemental compositions are needed) and interpretation of product ion spectra. From the LC-MS analysis perspective, drug metabolites can be categorized into common (predictable) and uncommon (unpredictable) metabolites. Common metabolites are those formed via known biotransformation reactions as displayed in Table 6.5. Their molecular masses (m/z values) can be readily predicted by mass shifts from the parent drugs. For example, nefazodone (NEF) is metabolized to multiple oxidative metabolites in HLM (Figure 6.9, and Table 6.6). Most of these metabolites are common metabolites, including three monohydroxy metabolites (M2, M5, and M7) and three dihydroxy metabolites (M1, M3, and M6). MWs of the monohydroxy and dihydroxy metabolites are 16 and 32 Da greater than that of their parent drugs, respectively. Uncommon metabolites are those formed via unconventional or multiplestep biotransformation reactions. Molecular masses of uncommon metabolites are not easily predicted (Table 6.7). For example, oxidation of NEF in HLM generated a few uncommon metabolites, including the triazoledione metabolite (M4) that was formed via monohydroxylation followed by a rearrangement reaction (Figure

6.9 and Table 6.6). M10 and M13 are also considered uncommon metabolites since their MWs are not easily predicted (Figure 6.9 and Table 6.6). Once a metabolite ion is found either using a speciic scanning method or a postacquisition data mining process, product ion spectral data of this metabolite can be easily acquired via product ion scanning or a datadependent MS/MS acquisition. Figure 6.10 displays product ion spectra of NEF and M4 acquired via product ion scanning on a triple quadrupole mass spectrometer. By interpretation and comparison of the two spectra, the location of metabolic modiication on M4 can be determined. Spectral interpretation is a time-consuming and dificult process, which is heavily dependent on the skill and knowledge of individual scientists and is out of scope of this chapter. Detection of low levels of uncommon metabolites in a complex matrix is a challenging task since their mass shifts are typically dificult to predict from the parent drugs and these metabolites are not displayed in total ion chromatograms (TIC) due to low abundance. The following sections will focus on scanning methods and postacquisition datamining methodologies for drug metabolite detection using various mass spectrometers and data-mining processes. 6.5.1 Detection and Identiication of Common Metabolites by Various LC-MS Instruments Acquisition of a full MS scan data set followed by extracted ion chromatographic (EIC) analysis based on their predicted m/z values is the most effective approach, and this can be performed by all types of LC-MS instruments. The m/z values of common metabolites can be easily calculated based on their mass shifts from their parent drugs as shown in Table 6.5. Mass spectral data processing software packages from most venders provide m/z values of common metabolites derived from a speciic drug, which can be employed for performing EIC analysis. For example, to determine common metabolites of NEF in HLM, an incubation sample was analyzed by full-scan MS analysis on a triple quadrupole instrument. As shown in Figure 6.11 (panel A), the TIC of the analysis only displays M9 and M1, while NEF and most metabolites are not visible. EIC processing of the full MS data ile, which targeted NEF and metabolites from monohydroxylation, dihydroxylation, and N-dealkylation, clearly revealed the presence of NEF and the formation of corresponding NEF metabolites in the HLM incubation sample (Figure 6.11B–E). The EIC analysis was able to detect most common metabolites of NEF. However, three uncommon metabolites—M4, M10, and M13—were not

LC-MS IN DRUG METABOLITE PROFILING AND IDENTIFICATION

TABLE 6.5 Drug [159]

147

Summary of Common Metabolism Reactions That Lead to Oxidative Metabolites Similar to the Parent

Biotransformation Reaction RSH → RSO3H 3 × (RH → ROH) 2 × (RH → ROH) RSR′ → RSO2R′ RH → ROH → ROCH3 RCH3 → RCOOH Quinone formation RCH=CHR′ → RCH2CH(OH)R′ Epoxide hydrolysis RH → ROH Epoxidation S- or N-oxidation RCH2CH3 → RCOOH RCH2NH2 → RCOOH Methylation RCH2R′ → RCOR′ RCH2OH → RCOOH 2 × (RCH=CHR′ → RCH2CH2R′) RCOR′ → RCH(OH)R′ RCH=CHR′ → RCH2CH2R′ RCH3 → ROH RCH2NH2 → RCH2OH N-oxide → N-hydroxy R-CH(OH)CH3 → RCOOH RCHNH2R′ → RCOR′ RCH2CH2R′ → RCH=CHR′ RCH(OH)R′ → RCOR′ RCH2NH2 → RCN RCOOH → RCH2OH Demethylation RSOR′ → RSR′ RNO2 → RNH2 ROCH3 → RH

Mass Shift (Da)

Molecular Formula Change

Mass Defect Shift (mDa)

+48

+O3

–0.0153

+32

+O2

–0.0102

+30

+OCH2 +O2, −H2

+0.0106 –0.0259

+18

+H2O

+0.0106

+16

+O

–0.0051

+15 +14

+O2, −CH4 +O2, −NH3 +CH2 +O, −H2

–0.0415 –0.0367 +0.0157 –0.0208

+H4 +H2

+0.0313 +0.0157

+O, −CH2 +O, −NH +H +O, −CH4 +O, −NH3 –H2

–0.0208 –0.0160 +0.0078 –0.0364 –0.0317 –0.0157

–H4 +H2, −O –CH2 –O +H2, −O2 –CH2O

–0.0313 +0.0208 –0.0157 +0.0051 +0.0259 –0.0106

+4 +2

+1 0 –1 –2 –4 –14 –16 –30

detected by this approach (Table 6.6). In addition to the full-scan mass spectrometry–extracted ion chromatographic analysis (MS-EIC) approach, data-dependent MS/MS scanning has been widely employed in detection and MS/MS acquisition of common metabolites using ion trap, Orbitrap, Q-TOF, and triple quadrupolelinear ion trap (Q-Trap) instruments (Table 6.1) [23,154– 156]. In list-dependent MS/MS acquisition, predicted m/z values of common metabolites are listed in the full MS scanning. If a detected ion has the same m/z value as a listed ion, and its intensity rises above a preset threshold, product ion scanning of this ion is automatically triggered. Therefore, data-dependent acquisition increases throughput of detection and identiication of common metabolites.

6.5.2 Detection and Identiication of Uncommon Metabolites by Various LC-MS Instruments Detection of uncommon metabolites is normally accomplished using speciic scan methods, postacquisition data-mining methods, or combined approaches, depending on the availability of scan functions and data processing tools of the mass spectrometer being used. PI scan and NL scan on triple quadrupole are classical LC-MS methods screening for uncommon metabolites based on their predicted fragmentation patterns. Recently introduced triple quadrupole linear ion trap and various HR-MS provide broader capabilities in drug metabolite proiling and identiication. For example, Table 6.8 summarizes the scanning and data

148 TABLE 6.6

Detection of NEF Metabolites by MDF, NL Scan and PI Scan Approaches [159]

246 −224

(A)

O

N N

N

N

O

Cl N

−196 274

MH+ = 470.2323 (C25H33O2N5Cl)

(B)

O

N N

N

N

Mass Defect Shift from Template (mDa)

Metabolite Identity

MH+ (Molecular Formula + H)

EIC Analysis

PI Scanc

NL Scand

M1 M2 M3 M4 M5 M6 M7 NEF M8 M9

–0.0102 –0.0051 –0.0102 –0.0364 –0.0051 –0.0102 –0.0051 0 –0.0051 0

502.2221 (C25H33O4N5Cl) 486.2272 (C25H33O3N5Cl) 502.2221 (C25H33O4N5Cl) 458.1959 (C23H29O3N5Cl) 486.2272 (C25H33O3N5Cl) 502.2221 (C25H33O4N5Cl) 486.2272 (C25H33O3N5Cl) 470.2323 (C25H33O2N5Cl) 376.2349 (C19H30O3N5) 360.2339 (C19H30O2N5)

+ + + – + + + + – +

290 274, 246 274, 246 – 290 290 274, 246 274, 246 290 274, 246

212 212 – 196 196 – 212 196 – –

M10

–0.0207

374.2192 (C19H28O3N5)







M11 M12

+0.0156 –0.0051

Dihydroxy NEF Monohydroxy NEF Dihydroxy NEF Triazoledione Monohydroxy NEF Dihydroxy NEF Monohydroxy NEF The parent drug Monohydroxy M9 Dealkylated metabolite Ketone derivative of M9 Alcohol derivative Carboxylic acid derivative

292.1661 (C15H22O3N3) 306.1454 (C15H20O4N3)

+ +

246 246

– –

M13

–0.0051

Monohydroxy m-CPP

213.0795 (C10H14ON2Cl)







m-CPP

197.0845 (C10H14N2Cl)

+





Metabolite Detected by MDFb

MDF Template for Metabolite Searcha

NH

O

MH+ = 360.2399 (C19H30O2N5) (C)

O

N N

N

CHO

O

MH+ = 290.1505 (C15H20O3N3) Cl

(D) N

N

MH+ = 197.0845 (C10H14N2Cl)

M14

0

An LC-MS data ile of NEF metabolites in HLM was processed using both drug and core structure templates. Each of the four MDFs used was deined in two dimensions: a mass defect range of ±0.040 Da from a template and a nominal mass range of ±50 Da from a template. b NEF metabolites are displayed in the MDF-processed proiles present in Figure6.9. c PI scan was monitored for product ions at m/z 274, 246, and 290 (274+O). d NL scan monitored NLs of 196, 224, and 212 (196+O) Da. a

LC-MS IN DRUG METABOLITE PROFILING AND IDENTIFICATION

149

OH N N

N

O

M10

N

NH

O

O

N

O

N N O

N

O

M9

N N

N

NH

M8

Cl N

N N

NH O

O

N

N

N

M2

OH

Cl N N

N

O

Cl N

O

N N

N O

NEF

N

OH

N

Cl N

O

N

N N

O

N

Cl N N

N

M5 O

M13 N

N N

COOH

O

O

M12

N

O

CHO

O

OH

O

Dihydroxylated metabolites (M1, M3, M6)

N

M7

M14

HN

N

O

Cl HN

OH

O

N

N N

O CH 2OH

N

Cl NH N

N

N

O

O

M11

M4

FIGURE 6.9 Metabolic pathways and metabolite structures of NEF which are formed in HLM. M4, M10, and M14 are considered as uncommon metabolites because their MWs are dificult to predict. Other metabolites are common metabolites, whose MWs can be calculated based on mass shifts from NEF [159].

processing functions of HR-MS technology for detection and structural characterization of drug metabolites. All of the methodologies except for EIC analysis have a common feature in that detection of drug metabolites is not based on predicted MWs. It should be mentioned that a majority of these methods are capable of detecting both common and uncommon drug metabolites. 6.5.2.1 Triple Quadrupole Mass Spectrometry PI and NL scanning methods on triple quadrupole mass spectrometers are classical approaches for detecting uncommon metabolites by LC-MS [13,19], which are performed based on predicated fragmentation patterns of metabolites. For detecting conjugated metabolites, PI and NL scans that target product ions or NLs are routinely employed (Table 6.9). For example, positive NL scanning of 129 Da or negative PI scanning of product ion at m/z 272 was used for screening for GSH adducts [157,158]. For detecting oxidative metabolites, PI and NL scanning methods are carried out following fragmentations of the parent drugs. For example, NEF generates two major product ions at m/z 274 and 246 via NLs of 196 and 224 Da, respectively (Figure 6.10A)

[159]. Therefore, PI scans of m/z 274 and 246 and NL scans of 196 and 224 Da, can be employed for searching for oxidative metabolites that are similar to NEF. As summarized in Table 6.6, the PI and NL scans on a triple quadrupole LC-MS were able to detect most oxidative metabolites, including both common and uncommon metabolites of NEF in HLM (Table 6.6). For example, an uncommon metabolite M4 that underwent an NL of 196 Da was detected by an NL scan of 196 Da (Figure 6.10B and Table 6.6). PI and NL scans are not capable of detecting oxidative metabolites that have fragmentation patterns completely different from those of their parent drugs, which is a major limitation of the triple quadrupole instrument-based approach. For example, two N-dealkylated metabolites of NEF, M13, and M14, were not detected by NL and PI scans because of this reason (Figure 6.9 and Table 6.6). 6.5.2.2 Triple Quadrupole Linear Ion Trap Mass Spectrometry Hybrid triple quadrupole linear ion trap mass spectrometry (Q-Trap) can perform NL- and PIdependent MS/MS acquisition (neutral loss–enhanced product ion scan [NL-EPI] and precursor ion–enhanced

150

LC-MS IN DRUG METABOLISM AND PHARMACOKINETICS: A PHARMACEUTICAL INDUSTRY PERSPECTIVE

TABLE 6.7

Metabolic Reactions that Lead to Metabolites Signiicantly Smaller Than the Parent Drugs [159]

Biotransformation Reaction

Filter Template

Dealkylation

RCH2XCH2R′ → RCH2XH + R′CHO (X = NH, O, or S) (RCH2XH or R′CHO > 50 Da) RC(=O)OCH2R′ → RCOOH + R′CH2OH RC(=O)NHCH2R′ → RCOOH + R′CH2NH2 RSC(=O)R′ → RSH + R′COOH RCH2OR′ → RCOOH + R′OH RN=NR″ RS-SR′ → RSH + R′SH

Cleavage of ester, amide or, ether

Reduction of azo and disulide

(A)

RCH2XH, R′CHO RCOOH, R′CH2OH RCOOH, R′CH2NH2 RSH, R′COOH RCOOH, R′OH RNH2, R′NH2 RSH, R′SH

246 −224

CI

N N

N

O

N

N −196

O 274 274.29

% relative intensity

100 80 60 40

246.26

20

470.40

180.27

0 100

200

300

500

400

600

m/z 234

(B)

O NH N

N

O

−196

O 262 262.1

100 % relative intensity

N

80 60 168.3 40 176.1 196.9

20 0

50

100

150

200

458.4 234.2

293.7

250

300

350

400

450

500

m/z

FIGURE 6.10 Product ion spectra and major fragmentations of NEF (A) and its metabolite M4 (B), which were acquired with product ion scanning on a triple quadrupole [159].

LC-MS IN DRUG METABOLITE PROFILING AND IDENTIFICATION

151

FIGURE 6.11 Full-scan analysis followed by extracted ion chromatograms (EIC) of NEF metabolites in HLM. Panel A is the TIC. Panels B (m/z 470), C (m/z 486), D (m/z 502), and E (m/z 292) are the EIC [159].

product ion scan [PI-EPI]), which allow metabolite detection and MS/MS spectral recording in the same LC-MS injection [21,22]. NL-EPI and PI-EPI are capable of detecting some uncommon metabolites based on their fragmentation patterns. In addition, the Q-Trap has MRM-EPI and MIM-EPI scanning functions that are not available in other types of mass spectrometers [24–26,28]. As demonstrated in Figure 6.12, MRM is much more sensitive than NL and PI scans [24]. However, detection of metabolites by MRM is based on predicted MWs and fragmentation patterns, which greatly limit the utility of MRM-EPI in analysis of unknown metabolites if neither their MWs nor fragmentation patterns are predictable. One way to improve the effectiveness of MRM in detection of various types of metabolites is to set multiple MRM scans to cover an entire m/z range instead of speciic m/z values of predicted individual metabolites. The recently introduced Q-Trap 5500 instrument allows the performance of 450 or more MRM-EPI transitions in metabolite screening without a signiicant loss of sensitivity. Therefore, this MRM-EPI approach can detect certain types of metabolites even though their MWs are unpredictable [24]. Another way to broaden the application of MRMEPI is to use the MIM-EPI approach. Analysis of in vitro metabolites using MIM-EPI has shown that this approach can detect unknown metabolites regardless of their fragmentations. This approach is similar to full-

scan EPI-based data mining and allows for highthroughput drug metabolite screening without prior knowledge of the MWs and fragmentation patterns of the metabolites. The advantage of MIM-EPI over fullscan EPI is its better sensitivity and selectivity due to isolation of ions in Q1. MIM-EPI was also used to acquire product ion spectra of plasma metabolites for setting MRM transitions in subsequent quantiication of these metabolites [25]. In addition to superior speed and sensitivity, Q-Trapbased information-dependent MS/MS acquisitions in combination with polarity switching can improve the quality of results. For example, GSH adducts generate several common products from the GSH moiety, including m/z 272, under negative ion mode, regardless of GSH adduct structures, while product ion spectra of GSH adducts often display compound-dependent fragments in positive ion mode, which are much more useful in structural elucidation than their corresponding negative product ion spectra (Table 6.9) [27,160]. Thus, a Q-Trap-based worklow was developed for fast and sensitive screening of GSH-trapped reactive metabolites. In the analysis, PI scanning of m/z 272 in negative ion mode is used as the survey scan to trigger MS/MS acquisition of product ion spectra in positive ion mode. In a single LC-MS injection, both detection and recording of unknown GSH adducts can be accomplished via PI-EPI scans and polarity switching. In a similar fashion, NL

152

LC-MS IN DRUG METABOLISM AND PHARMACOKINETICS: A PHARMACEUTICAL INDUSTRY PERSPECTIVE

TABLE 6.8 Data Acquisition and Postacquisition Data Mining Methods Commonly Employed in Drug Metabolite Identiication by HR-MS Method Data acquisition method Intensity-dependent MS/MS acquisition with or without an inclusive list [155,156]

Operation or Metabolite Detection Mechanism Full MS scan to trigger MS/MS acquisition of ions above preset intensity and/or listed ions

Applications and Limitations



• •

“All-in-one” scan (MSE scan) or related methods [171,175]

A derived full MS scan method, in which full MS scan with alternated low and high collision energy is performed.



• •

Isotope pattern-dependent MS/ MS acquisition [172,173]

Full MS scan to trigger MS/MS acquisition of ions that display a speciic isotope pattern





Mass defect-dependent MS/MS acquisition

Full MS scan to trigger MS/MS acquisition of ions, whose mass defects are in a speciic range





Data mining method Extracted ion chromatography (EIC)

Predicted MW

• •

Mass defect ilter (MDF) [159,163]

Predicted mass defect window

Product ion ilter (PIF) and neutral loss ilter (NLF) [174,175]

Predicted fragmentation

Isotope pattern ilter (IPF) [170,172]

Predicted isotope pattern

• •

• •

• •

Background subtraction (BS) [167–169] or control sample comparison (CSC) [171,189]

Metabolite ion that is not present in control sample

• •

Generic acquisition method suited for fast proiling of in vitro metabolite High sensitivity of list-dependent MS/MS acquisition Not suited for low levels of unknown metabolites Generic acquisition method suited for fast proiling of in vitro metabolites Generation of pseudo MS/MS spectra May not be suited for trace metabolites coeluted with large amounts of interference ions Capable of triggering MS/MS acquisition of minor metabolites with a speciic isotope pattern Compound-dependent acquisition method not suited for various metabolites Capable of triggering MS/MS acquisition of minor metabolites in complex sample Compound-dependent acquisition method not suited for fast metabolite proiling Sensitive detection of expected metabolites Not suited for detecting metabolite whose MW is unpredictable Suitable for all types of metabolites Sensitivity and selectivity depend on matrix Sensitive detection of unexpected metabolites Not suited for metabolites that do not generate signiicant predictable fragmentation Sensitive detection of unexpected metabolites Not suited for metabolites that have no unique isotope pattern Suitable for all types of metabolites Sensitivity and selectivity dependent on the control sample applied

LC-MS IN DRUG METABOLITE PROFILING AND IDENTIFICATION

TABLE 6.9

153

Metabolic Conjugative Reactions That Lead to Conjugates

Biotransformation Reaction

Mass Shift (Da)

Molecular Formula Change

Mass Defect Shift (mDa)

GSH conjugation

+305.682

+10C+15H+3N+6O+S

68.2

Glucuronidation Glycine conjugation Methylation N-acetylcysteine conjugation Acetylation Cysteine conjugation Taurine conjugation

+176.0321 +57.0215 +14 +161.0147

+6C+8H+7O +2C+3H+N+O +C+2H +6C+8H+N+3O+S

32.1 21.5 15.7 14.7

+42.0106 +119.0041 +107.0041

+2C+2H+O +3C+5H+N+2O+S +2C+5H+N+2O+S

10.6 4.1 4.1

Sulfate conjugation

+80.0432

+3O+S

–43.2

Common Fragmentation NL of 129 Da (+) and 307 Da (+) Precursors of m/z 272 (–) NL of 176 Da (+/–)

NL of 129 Da (–) in negative ion mode

Precursors of m/z 126 (+) or m/z 124 (–) NL of 80 Da (+/–)

FIGURE 6.12 Analysis of GSH adducts of mefenamic acid formed in HLM using Q-Trap 4000. (A) TIC of NL scan, PI scan, and MRM scan. (B) Pathways of reactive metabolite formation of mefenamic acid [24].

scans of 129 Da, followed by acquisition of positive MS/ MS spectra via polarity switching, is utilized for fast analysis of N-acetylcysteine conjugates in incubations and urine (Table 6.9) [26]. 6.5.2.3 Ion Trap and Linear Ion Trap Mass Spectrometry Full-scan MS experiments followed by data-dependent MSn acquisition with three-dimensional ion trap or linear ion trap mass spectrometers are commonly carried out to identify metabolites [9]. This

approach not only allows for sensitive detection of common metabolites via EIC data processing, but also acquires MSn spectra for structural elucidation in one or two injections. Although ion trap instruments are not capable of performing NL or PI scans to directly detect drug metabolites, NL or product ion iltering of MS/ MS data acquired by data-dependent acquisition have been utilized to search for uncommon metabolites based on their fragmentation patterns. One example is to identify reactive metabolites trapped by STIL GSH.

154

LC-MS IN DRUG METABOLISM AND PHARMACOKINETICS: A PHARMACEUTICAL INDUSTRY PERSPECTIVE

In the analyses, isotope-dependent MS/MS acquisition methods with or without polarity switching are employed to record MS/MS spectra of analytes that display isotope patterns similar to those predeined. Then, NL iltering of 129 and 307 Da in the positive ion mode, or product ion iltering of m/z 272 in the negative ion mode, are carried out to identify GSH adducts [161,162]. 6.5.2.4 HR-MS Traditionally, HR-MS was employed as a supplemental tool for metabolite identiication by LC-MS [9]. Its major application is the determination of molecular formulae and accurate mass fragments of metabolites to assist in structural characterization. The major limitation of HR-MS in metabolite identiication was the lack of capability to perform true PI scan or NL scan analysis for detecting uncommon metabolites. Recently, an MDF technique (also referred to as “fractional mass iltering” in some publications) was developed for detecting both common and uncommon drug metabolites via postacquisition processing of highresolution LC-MS data [29,159,163], which, together with other accurate mass data acquisition and data processing technologies (Table 6.8), has fundamentally changed the role of HR-MS instruments in drug metabolite proiling and identiication [29]. As shown in Table 6.5, mass shifts of oxidative metabolites from the parent drugs are between +50 and −50 Da, while mass defect

shifts of these metabolites from the parent drugs are less than 50 mDa. The predicted narrow ranges of the mass defect values of oxidative metabolites, whose structures are similar to parent drugs (Table 6.5), are utilized for metabolite detection via MDF of accurate mass fullscan MS data. For example, MDF of full-scan MS data of the NEF incubation sample using NEF as a ilter template (NEF mass defect ±40 mDa over a mass window of NEF ±40 Da; Table 6.6) was able to detect NEF and its metabolites, M1–M7 (Figure 6.13), which are similar to NEF (Figure 6.9 and Table 6.6) [159]. In addition to the detection of oxidative metabolites that are similar to the parent drugs, MDF is capable of inding metabolites that are signiicantly smaller than the parent drugs using core structure MDF templates. These metabolites are usually formed via N, O, or Sdealkylation reactions, cleavage of ester, amide or ether linkages, and reduction of azo and disulides [159,164] (Table 6.7). Figure 6.13 displays the detection of all of the oxidative metabolites of NEF, including multiple N-dealkylated metabolites, using multiple MDFs [159]. The core structure MDF templates applied for detection of these N-dealkylated metabolites are displayed in Table 6.6. Similarly, MDF has been employed for detecting conjugated metabolites, such as GSH adducts [164,165] and glucuronides, using conjugate ilter templates (Table 6.10).

(A) % relative intensity

Scan ES+ BPC NEF 5.29e3

M1

100

M4/M5 M2 M3 M6

M7

0 5.00

10.00

15.00

20.00

(E) M13

Scan ES+ BPC 2.19e4

M9/M14

% relative intensity

100

25.00

M10 M11 M8

M1 M12 NEF M2M4/M5 M3 M6

0 5.00

10.00

15.00

20.00

25.00

Time (min)

FIGURE 6.13 MDF of NEF metabolites formed in HLM incubations. Full-scan MS data were acquired by Orbitrap. (A) MDFprocessed metabolite proile of NEF using the parent drug as a ilter template. (B) MDF-processed metabolite proile of NEF using multiple MDF. Metabolite structures are displayed in Figure 6.9. MDF templates are displayed in Table 6.7 [159].

TABLE 6.10

Summary of Key In Vitro and In Vivo Biotransformation Experiments in Drug Discovery and Development

Experiment

Objective of Experiment

Model of Experiment

Metabolic soft-spot analysis



Determine major metabolite structures in liver microsomes



Liver microsomes and NADPH

Reactive metabolite screen



Determine if a test compound forms reactive metabolites Determine the structure of the reactive metabolite trapped by GSH Estimate quantity of reactive metabolite formed Determine metabolite proiling and structures across species



Liver microsomes, GSH, and NADPH



Liver microsomes and NADPH Hepatocytes





In vitro metabolism across species





Experimental Conditions

Sample for Analysis

Incubations with a test compound (5–10 µM) for 5–30 min at pH 7.4 and 37°C Incubations with a test compound (10–50 µM) for 30–60 min at pH 7.4 and 37°C

Incubation mediums after removing proteins



LC-UV/MS



Provide scientiic basis for chemists to design metabolically stable drugs

Incubation mediums after removing proteins



LC-MS



Provide scientiic basis to design or select lead compounds that have no or minimal potential to form reactive metabolites

Incubation mediums after removing proteins or cells



LC-UV/MS LC/radio detection/MS



Provide scientiic basis for predicting human PK from in vitro metabolic rates or from animal PK Provide scientiic basis for selection of animal species for toxicology studies

Incubation of liver microsomes (15– 45 min) or hepatocytes (0–4 h) with a test compound (10–30 µM) at pH 7.4 and 37°C

Analytical Method



Utility of Results



(Continued)

155

156 TABLE 6.10

(Continued)

Experiment Metabolism in BDC rats

Metabolite proiling in radiolabeled ADME study

Objective of Experiment •

• •





Metabolite proiling in irst human clinical study





Proiling, identiication and quantitative estimation of metabolites in plasma, urine, and bile

Determine mass balance Determine exposure levels of drug-related compounds in humans and animals Elucidate drug elimination pathways in humans Elucidate metabolite structures Determine metabolite proiles and structures in plasma Determine concentrations of metabolites relative to parent drug

Model of Experiment •

• •



BDC rats

Healthy human subjects Animals

Healthy human subjects

Experimental Conditions

Sample for Analysis

A single dose of test Plasma, urine, and compound to bile duct bile samples cannulated (BDC) rats. Plasma samples up to 24 h and urine and bile samples up to 48 h are collected. Plasma, urine, feces, A single dose of and bile samples radiolabeled pooled across compound (5 µL) is interchangeable in order to adjust the injection volume. For injection volumes below 5 µL, internal loops should be used. The typical injection volume is 1–20 µL onto the column. To

connect the pump, injection valve, column, and MS system, connecting tubes with internal diameters of 0.1–0.25 mm and external diameters of 1/16 inch are commonly used. The length of the connections should be as small as possible. Finger-tight ittings are recommended to make the connections that can withhold the necessary high pressures [3]. A guard column is recommended to be placed between the injector and the analytical column to protect the latter from clogging due to the trace denatured proteins and other residues in the reconstituted sample extracts. Reversed-phase LC has been widely used for quantitative LC-MS/MS. With increase of organic solvent composition in the mobile phase, the retention of the analyte of interest normally decreases. One should be aware of a possible bimodal retention of certain analytes on the reversed-phase column due to the presence of silanol groups on the stationary phase, which might cause a shift in retention during the run or poor assay method reproducibility [4]. Polar compounds often have poor retention on a reversed-phase column. In order to retain the polar compounds on a reversedphase column, high aqueous mobile phase becomes necessary. Unfortunately, the high aqueous mobile phase often results in detrimental matrix effects and unacceptable ionization eficiency of the analyte(s) in electrospray ionization (ESI) in MS source due to the presence of coeluted polar matrix components and high percentage of water. Ion-pair chromatography (IPC) and hydrophilic interaction chromatography (HILIC) are two excellent alternatives [4,5]. Ultrahigh performance liquid chromatography (UPLC), a relatively new technology, offers signiicant advantages in resolution, speed, and sensitivity for analytical determinations, particularly when coupled with mass spectrometers capable of high-speed data acquisitions. By using sub-2 µm particle columns and operating at pressures signiicantly greater than conventional HPLC columns, dramatic improvements in chromatographic separation can be expected that often lead to superior selectivity and sensitivity [7–9]. 7.2.1.2 Mass Spectrometer (MS) MS is the most sensitive method of molecular analysis. A mass spectrometer is a highly computerized instrument. It basically consists of ive parts: sample introduction (typically in conjunction with an LC system), ionization, mass analysis, ion detection, and data handling [10]. The principle of MS is the production of ions that are subsequently separated or iltered according to their mass-to-charge (m/z) ratio, followed by the detection of charged ion (m/z). The resolution of a mass spectrometer is qualitatively deined as its ability to discriminate between adjacent ions in a spectrum [11]. Typically, the

METHODOLOGY

resolution is set to “unit mass,” indicating, for instance, m/z = 100 and m/z = 101 can be distinguished. “Unit mass” is suitable for the determination of the nominal masses of a compound and its fragment ions. Instrument accuracy to ±0.5 m/z of a nominal mass measurement is often suficient to provide the level of accuracy required for an analytical application. The resulting mass spectrum is a plot of the relative abundance of the ions generated as a function of the m/z ratio. Today, electrospray ionization (ESI) and atmospheric pressure chemical ionization (APCI) are considered the standard ionization techniques for LC-MS/MS due to their predominant advantages in quantitative analysis of drug molecules in various sample matrices with high sensitivity, selectivity, reliability, robustness, and ease of operation. Other techniques, for example, atmospheric pressure photoionization (APPI), electron capture atmospheric pressure chemical ionization (EC-APCI), and high-ield asymmetric waveform ion mobility mass spectrometry (FAIMS) serve as complements to the established ESI and/or APCI technical platforms whenever necessary for an enhanced sensitivity and/or selectivity of a bioanalytical assay [4,5]. ESI [12,13] is a well-known soft ionization process by which analyte molecules in solution delivered from a syringe pump or HPLC pass through the electrospray capillary with a high potential difference (compared with the counter electrode), typically in a range of 1–6 kV. The high potential forces the spraying of charged droplets from the capillary with a surface charge of the same polarity to the charge on the capillary. The multiple-charged droplets are repelled from the capillary toward the source sampling cone or oriice. As the droplets travel the space between the capillary tip and the cone/oriice, solvent evaporation occurs and the droplet shrinks until it reaches a state that its surface tension can no longer sustain the charge (the Rayleigh limit). A “Coulombic explosion” occurs and the droplet is ripped apart, producing smaller droplets that can repeat this process and produce charged analyte molecules for MS or tandem mass spectrometry (MS/MS) detection. A general ESI source consists of a capillary interface for liquid introduction and a metal capillary needle to apply static voltage for ionization. APCI [14] is an analogous ionization method to chemical ionization (CI); a signiicant difference between the two is that APCI occurs at atmospheric pressure and has its primary applications for the ionization of low mass compounds. APCI is not suitable for the analysis of thermally labile compounds. In APCI, the analyte solution is introduced into a pneumatic nebulizer and desolvated in a heated quartz tube before interacting with the corona discharge. The corona discharge produces primary N2+ and N4+ by electron ioniza-

173

tion. These primary ions collide with the vaporized solvent molecules to form secondary reactant gas ions, for example, H3O+ and (H2O)nH+. These reactant gas ions then undergo repeated collisions with the analyte, resulting in the formation of analyte ions. A higher frequency of collision results in a higher ionization eficiency and thermalization of the analyte ions with a mass spectrum of predominantly molecular species and adducts ions with very little fragmentation. A general APCI source consists of a capillary interface for liquid introduction, a heated nebulization system, and a highvoltage corona discharge needle. Ionization polarity should be chosen in conjunction with the mobile phase because pH can inluence ease of ionization in positive or negative ion modes. Basic or neutral compounds are readily ionized in positive ion mode at a pH below 7. Since many potential drug candidates contain amine moieties, the majority of LC-MS/ MS methods are conducted in positive ion mode. Since mass spectrometers can determine atomic and molecular isotope ratios, a compound will have a distribution of individual isotopic molecular masses based on the relative abundance of its constituent ions. In small molecule MS application, the most important concept is the monoisotopic mass. It is the sum of the masses of the atoms in a molecule using the unbound, groundstate, rest mass of the principal (most abundant) isotope for each element instead of the isotopic average mass [15]. For typical organic compounds, monoisotopic mass is calculated using the lightest isotope. Under a positive ionization mode, a protonated molecular ion [M+H]+ at monoisotopic mass typically gives out the highest intensity in a mass spectrum. The triple quadrupole analyzer (QqQ) is the most widely used MS system for quantitative analysis. The QqQ typically has an upper dynamic limit of m/z 1250– 4500. To determine the correct mass spectrometer operation conditions for a given compound, a neat reference standard solution (containing certain modiiers, e.g., acetic acid, formic acid, or ammonium hydroxide, etc.) can be introduced into the ion source by syringe pump infusion at a low low rate (e.g., 10 µL/min) with or without mixing with the column efluent. After selecting the target molecular ions (protonated, deprotonated, or protonated + adduct, e.g., [M+NH4]+ for cyclosporine, etc.) in the irst mass analyzer (Q1) via Q1 scan, the next step is to induce fragmentation of the molecular ions by applying a collision energy and a collision gas (Ar, N2, or He) for collision-induced dissociation (CID) or collision-associated dissociation (CAD) in the collision cell (Q2). The most abundant product ion, corresponding to a portion of the molecular ion after cleavage, is detected in the third quadrupole (Q3) via product ion scan. To ensure the MS detection selectivity, the selected

174

QUANTITATIVE MASS SPECTROMETRY IN SUPPORT OF PHARMACOKINETIC STUDIES

product ion should, in general, represent a signiicant loss (e.g., greater than 20 Da) from the precursor ions (or molecular ions), which can differentiate the background ion that gives out only a smaller loss, for example, loss of water (18 Da). Once the precursor and product ions are selected, other MS instrument parameters (Micromass [Milford, MA] instruments: cone voltage, collision energy, desolvation gas low, cone gas low, and collision gas low, etc.; Sciex API [Concord, Ontario, Canada] systems: low of desolvation gas, nebulization gas, curtain gas and collision gas, potentials of entrance, declustering, focusing and collision cell exit, ionspray voltage, and desolvation temperature, etc.; and ThermoScientiic [San Jose, CA] systems: needle voltage, sheath gas pressure, auxiliary gas pressure, ion sweep gas pressure, ion transfer capillary temperature, tube lens offsets, and source CID offsets, etc.) can be optimized by monitoring the precursor–product ion mass transitions for greatest and most stable product ion intensity for the analyte of interest. Monitoring a special collisioninduced reaction signiicantly improves the selectivity, resulting in greatly improved signal-to-noise ratios and sensitivity. The same process should also be conducted for the internal standard (IS) (stable isotopelabeled or analogous) and other analytes and their corresponding ISs for simultaneous determination via multiple reaction monitoring (MRM). The MRM mass transitions with all the optimal MS conditions that yield the best assay selectivity and sensitivity will be used in the LC-MS/MS method. In order for more eficient quantitative LC-MS/MS method setup, some venders have developed automatic MRM method development software, for example, AutoScan™ by Sciex, QuanOptimize™ by Micromass and QuickQuan™ by ThermoScientiic. Greater details can be found in the literature [16]. For MRM experiments, the discrete collection time per data point is called the dwell time and is simply the length of time in which each discrete m/z ion signal is integrated or the time interval before another point is collected. Appropriate dwell time setting is important to balance the MRM sampling points and the chromatographic peaks. A common approach is to set the dwell time such that at least 20 data points are collected for any chromatographic peak. The typical dwell time settings range from 0.05 to 0.3 s. An LC-MS/MS assay run generates a threedimensional data array, retention time(s), MRM response(s) (m/z of precursor ion → m/z of product ion), and peak intensity. The data array is generated by repeating the MRM signals during the analysis time and storing the intensity data. The data array can be handled in various ways. The most important type of output is the LC-MS/MS chromatogram for integration [3].

The fundamental approach to quantitative LC-MS/ MS relies on the mathematical processing of analyte concentration in the calibration standards and LC-MS/ MS response (peak area ratio to the IS) via regression analysis. Thus, a set of calibration standard samples must be included in each assay run on a daily basis no matter how the ruggedness of LC-MS/MS has been increased with the improved source design and sample preparation procedure. The calibration standard samples are normally prepared by spiking a series of standard working solutions with predeined concentration into the blank matrix that was prescreened to ensure no interference. These standards are prepared in the same manner as for the intended study samples. In most cases, a linear regression is used in an LC-MS/MS quantiication with a dynamic range of three to four orders of magnitude for either ESI or APCI technique. Sometimes at the higher concentrations, the curve may deviate from the linear response. This deviation might be explained by the signal saturation in the MS instrument or ion suppression at higher analyte concentrations, leading to signal intensities lower than those theoretically predicted. If the analyte LC-MS/MS response (peak area ratio to the IS) of a given sample is greater than that obtained from the highest concentration standard, then the sample is deined as above the upper limit of quantiication (ULOQ or ULQ) and needs to be diluted until its LC-MS/MS response falls within the calibration curve range. The calculated concentration is then corrected with a dilution factor. On the other hand, the lower limit of quantiication (LLOQ) is deined at a signal-to-noise ratio of at least 10 compared with blank response and analyte peak (response) should be identiiable, discrete, and reproducible with a precision of ≤20% and accuracy of 80%–120% obtained from six replicate quality control (QC) samples assayed with a set of calibration standards. The samples with LC-MS/ MS response below LLOQ are generally referred to below the lower limit of quantiication (BLLOQ or BLQ) [16]. In LC-MS/MS bioanalytical method development, the most common challenges are matrix effects and stability. These issues, if not assessed and controlled properly, can lead to signiicant errors in assay results. Matrix effect has been deined as the analyte ionization suppression or enhancement in the presence of the matrix components that could originate from the endogenous components, metabolites, and/or coadministered drugs and their metabolites. The matrix effect could also derive from dosing vehicles, mobile phase additives, and plastic components. Employing a suitable IS that is structurally the same as or similar to the analyte of interest is a very common approach in quantitative LC-MS/MS to compensate for any possible inconsisten-

METHODOLOGY

cies in sample preparation, shift in retention time, and signal suppression or enhancement in MS/MS detection. Ideally, the mass of the stable isotope-labeled (13C, 2H and/or 15N, etc.) ISs should be separated from the analyte of interest by at least 1%–2% difference in daltons so that the naturally occurring isotopes of the parent analyte do not interfere with the response of the IS and vice versa. The more detailed sources and possible resolutions to the above issues can be found in the literature [4,5,16–22]. Stability is a prominent preanalytical criterion for the determination of analytes in biological samples. Inaccuracies resulting from losses of analytes during sample storage, transportation, and processing might occur before a quantitative LC-MS/MS method is in place. Failure to assess instability of an analyte in a given biological matrix would lead to inaccuracies of analyte concentration results and, ultimately, a possible repeat study. At the early discovery stage, extensive studies of drug stability are not required during analytical method development because of time constraints. However, stability assessments for drug candidates under various conditions are obligatory at the drug development stage. Although the source of instability is, in general, compound dependent, the exact cause falls into one or more of the following: natural degradation, light sensitivity, enzymatic cleavage/decomposition, hydrolysis, container interaction, hydrolysis, volalitility, nucleophilic interaction, and so on. Stability assessment should be conducted by speciic analytic methods including the parent compounds and their major metabolites to establish a clear pattern of alternations of drug concentration and decomposition. Appropriate preventive approaches must be taken if instability of an analyte is conirmed [4,5,23]. 7.2.1.3 Sample Preparation Because large amounts of proteins are present in biological samples (except urine), conventional HPLC columns will not tolerate the direct introduction of these samples for quantitative analysis. Most bioanalytical assays have a sample preparation step to remove the bulk proteins from the samples [2]. In addition, there are other important reasons for a sample preparation step when developing LC-MS/MS methods. These include the reduction of matrix components from the samples and minimization of ion suppression (also called “matrix effects”) in the mass spectrometric detection [18]. Once a bioanalytical method has been developed, the method performance must remain consistent over the duration of the study. The results generated based on a validated method procedure should be free from systematic error and any other characterized errors and meet the predeined acceptance criteria. Sample preparation is used to

175

ensure that a method maintains the basic elements of ruggedness and consistency. For bioanalytical applications, protein precipitation (PP), liquid–liquid extraction (LLE), and solid-phase extraction (SPE) are the most common sample preparation procedures [24], the choice of which is dependent on the composition of the matrix to be handled and the chemical properties of a given analyte, and so on. Protein Precipitation (PP) Simple organic solvents such as methanol or acetonitrile are effective for the removal of more than 98% of the proteins from plasma or serum when used in a ratio of 2:1 or greater [25] and are compatible with LC-MS/MS systems. Therefore, organic solvents are the most popular denaturing solvents for PP. After addition of the denaturing solution to the sample, PP is facilitated by vortexing and centrifugation of the plate or tubes. The denaturation process causes the analytes to be released from the proteins and remain in the supernatant liquid. This process is essentially a phase separation. The supernatant could then be evaporated to dryness under a stream of nitrogen followed by reconstitution using the starting mobile phase prior to injection onto the LC-MS/MS system for analysis. In some cases, a small aliquot (e.g., 1–5 µL) of the resulting supernatant might be injected directly onto the LC-MS/MS for analysis. Although PP is a simple and often useful sample preparation approach, this method is ineffective for the removal of lipids and other lipophilic components from the samples. Lipophilic compounds and other residues can accumulate on the column head and eventually clog and alter the retention characteristics of the column if a suitable postcolumn wash and the associated column reequilibration is not implemented during each injection cycle. PP is also ineffective for the removal of salts from the samples, which can cause matrix effects in the MS detection if column switch is not utilized to divert those unwanted column efluents to waste [24]. Semiautomated or automated PP procedures in 96well format are now commonly used to improve assay throughput and to reduce matrix effects via the more effective removal of phospholipids [4,5]. Liquid–Liquid Extraction (LLE) The principle of LLE is based on the differential solubility and partitioning equilibrium of drug molecules between aqueous (sample) and organic phases. In many cases, pH adjustment with appropriate acids, bases, or buffer is necessary to neutralize the analyte molecules and enables a more eficient extraction. An immiscible organic solvent is added to the aqueous sample matrix, followed by vortex mixing to facilitate equilibrium partitioning of analyte molecules between the two phases. When the

176

QUANTITATIVE MASS SPECTROMETRY IN SUPPORT OF PHARMACOKINETIC STUDIES

phases are separated via centrifugation, the organic phase is evaporated to dryness and reconstituted with the starting mobile phase or some other solvent prior to injection onto the LC-MS/MS system. An advantage of LLE is the lexibility of choosing an extraction solvent that has a similar polarity of the analytes of interest. The more polar solutes dissolve preferentially in the more polar solvent. In contrast, the nonpolar analytes preferentially dissolve in the nonpolar solvents. A fair recovery should be obtained for a given analyte with a meaningful distribution ratio or partition coeficient between the two phases when the organic solvent-tosample volume ratio is greater than three fold. LLE is simple for implementation, however, this method is quite labor intensive with many steps of manual manipulation and possible formation of emulsions, which could be a source of cross-contamination. Thus, semi- or fully automated LLE procedures based on commercially available liquid handling workstations (TomTec [Hamden, CT], Hamilton [Bonaduz, Switzerland], or Multiprobe [Shelton, CT], etc.) have been increasingly used [4,26,27]. Supported liquid extraction (SLE) is an alternative to conventional LLE, but is a low-through technique. The technique uses specially cleaned and sized porous adsorbents, including diatomaceous earth, as the supporting media where an eficient LLE takes place between the very thin aqueous stationary ilm formed after sample loading and a waterimmiscible organic solvent. Because a gravity low process is used, formation of emulsions, if any, can be prevented. Furthermore, SLE helps alleviate the liquid handling issues that are associated with conventional LLE and is, therefore, more amenable to highthroughput assays. The technique has been increasingly used for high-throughput quantitative analysis of drug molecules in plasma samples. SLE in combination with HILIC can further enhance sample analysis throughput by eliminating drying step of sample preparation [28]. Solid-Phase Extraction (SPE) SPE generally provides the cleanest sample extract among all sample preparation techniques in terms of selectivity. SPE uses the afinity of analytes in a liquid phase for a stationary phase, through which the samples pass with the analytes of interest retained on the stationary phase. The unwanted portion that passes through the stationary phase is discarded. The analytes retained on the stationary phase can then be removed for collection. The stationary phase is normally in the form of a packed column, a packed cartridge, or a 96-well plate [4,29], each of which can be mounted on a speciic type of extraction manifold. The manifold allows multiple samples to be processed simultaneously. Application of vacuum (negative mode) or pressure (positive mode)

speeds up the extraction process by pulling/pushing the liquid sample through the stationary phase. The analytes are collected in sample tubes or collection plates inside or below the manifold after they pass through the stationary phase. SPE method development and optimization is relatively time consuming compared with PP and LLE. The initial step typically involves the conditioning of the sorbent bed with a small volume of methanol or acetonitrile followed by water or buffer. The aqueous sample is then loaded onto the sorbent and the sample is allowed to pass through the bed via gravity, or centrifugal force or being pulled/pushed through by vacuum/ pressure if necessary. A series of wash steps are implemented using different solvents to remove salts and certain unwanted matrix components. The inal step is to elute the compound from the stationary phase with an appropriate solvent. The elution solution can either be directly injected onto the LC-MS/MS or evaporated to dryness followed by reconstitution using the starting mobile phase prior to injection onto LC-MS/MS. The most popular automated SPE procedures are set up in a 96-well format using 96-well workstations (e.g., TomTec, etc.). Compared with individual sample processing in series, the 96-well format with a parallel sample processing can dramatically improve the assay eficiency and reduce the turnaround time. Another approach for automation is online SPE, which tends to yield a lower throughput than the 96-well format procedure, requires a relatively complex instrument setup (e.g., multiple columns and switching valves), and often presents challenges associated with carryover [4].

7.2.2

Method Validation

As a drug candidate progresses through the development pipeline, the characteristics of a suitable bioanalytical method may change. Whereas a “quick and dirty” method may satisfy early compound screening, fully validated methods will be needed to support Good Laboratory Practice (GLP) toxicology studies. Further reinement of the method for increased sensitivity and higher throughput is often necessary in order to support large-scale clinical trials. Validation is the process and documentation of the speciic laboratory investigations that demonstrate an analytical method is suitable and reliable for its intended applications [31]. A validated bioanalytical method must demonstrate accuracy, precision, selectivity, sensitivity, reproducibility, and stability in order to support GLP toxicology studies and clinical trials. Below are the principles and acceptance criteria in each area.

METHODOLOGY

7.2.2.1 Selectivity and Speciicity Selectivity is the ability of an analytical method to differentiate and quantify the analyte in the presence of other components in the sample. A minimum of two (for animal matrices) or six (for human matrices) different batches of biological matrix (e.g., plasma), with at least three replicates of blank samples from each batch, should be prepared and analyzed for potential endogenous interferences at the retention time of analyte. At least one batch of the matrix, with at least three replicates of blank samples spiked with IS at working concentration, should be prepared and analyzed for potential contribution of the IS at the retention time of analyte. At least one batch of the matrix, with at least three replicates of blank samples spiked with the analyte at the ULOQ level, should be prepared and analyzed for potential contribution of the analyte to the IS. The mean response of interfering peaks at the retention time of the analyte or IS should be less than 20% of the mean response of the LLOQ or ≤5% of the mean response of the IS, respectively. If the method is intended to simultaneously quantify multiple analytes, then neat solutions of each analyte working solution at ULOQ and the corresponding IS working solution(s) should be analyzed to ensure that the impurities from one of the analytes does not interfere with the response of other analyte(s) and the IS(s) and vice versa. 7.2.2.2 Matrix Effect and Recovery For LC-MS/ MS-based methods, the signal suppression or enhancement of the analyte due to the presence of the matrix interferences (matrix effects) in MS/MS detection should be evaluated by comparing the response (peak area) of the analyte and the IS from the extracted blank samples post-fortiied with the analyte and the IS with the response of neat solutions with both the analyte and the IS at the same concentrations as above. Matrix effects should be evaluated in one pooled batch of animal matrix or in at least three different batches of human matrix, using three replicates at a minimum of three QC concentrations (e.g., low quality control [LQC], medium quality control [MQC], high quality control [HQC]) with IS at working concentration. The coeficient of variation (CV%) of the matrix effect variability should be 15 QC sample results could be obtained for each concentration level to assess the between-run (interday, interrun) accuracy and precision. The same acceptance criteria as above should be applied for the between-run (interrun or interday) accuracy and precision. 7.2.2.4 Sensitivity LC-MS/MS method sensitivity refers to the LLOQ with analyte peak reliably differentiated from the background noise (e.g., a signal-to-noise ratio of >5). The LLOQ represents the lowest amount of an analyte in a sample that can be quantitatively determined with suitable precision (≤20% CV ) and accuracy (±20% bias) as recommended in the Food and Drug Administration (FDA) Guidance or in the European Medicines Agency (EMEA) Guideline [30,31]. A common approach to establishment of the LLOQ is to prepare several concentrations near the estimated LLOQ

178

QUANTITATIVE MASS SPECTROMETRY IN SUPPORT OF PHARMACOKINETIC STUDIES

and to measure them in replicate along with the calibration curve. The lowest concentration to yield the desired acceptable level of accuracy and precision as above would therefore represent the LLOQ and should also be established as the lowest concentration in the calibration curve. The greater the number of concentrations tested, the more accurate the established LLOQ will be. In practice, the LLOQ should be predetermined based on anticipated analyte concentration and the needs of pharmacokinetics or toxicokenetics. Often an LLOQ at 5% of the anticipated peak concentration from the lowest dose group study samples would sufice (see acceptance criteria in the section “Accuracy and Precision”). 7.2.2.5 Calibration/Standard Curve A calibration (standard) curve is the relationship between the LC-MS/ MS response and known concentrations of the analyte. A calibration curve should be prepared in the same biological matrix as the samples in the intended study by spiking the matrix with known amounts of the analyte. A calibration curve should consist six to eight nonzero standard samples covering the expected dynamic range, including LLOQ and ULOQ. All nonLLOQ/ULOQ standard concentrations should be adequately distributed throughout the dynamic range in order to appropriately deine the curve. The results from blank sample (matrix sample processed without IS) and zero sample (matrix sample processed with IS) should not used as part of calibration curve. At least 75% of or a minimum of six nonzero standard results should be within ±15% of the nominal concentration values, except the LLOQ, which should be within ±20% of the nominal values. Standard curve itting is determined by applying the simplest model that adequately describes the concentration–response relationship using appropriate weighting and statistical tests for goodness of the it. Values that fall outside the above acceptance criteria can be discarded provided they do not change the established calibration model. 7.2.2.6 Stability Procedures should be in place to evaluate the analyte stability in the intended matrix during sample collection and handling, after short-term (bench top, room temperature) or long-term storage (frozen at the intended storage temperature), and after going through freeze and thaw cycles. Stability of the analyte in stock solutions and working solutions in deined storage condition and in reconstituted sample extracts in predeined autosampler condition (postpreparative stability) should also be evaluated. Conditions used in stability assessment should relect the situations likely to be encountered during actual sample handling and analysis.

All stability in matrix and in reconstituted sample extracts (postpreparative stability or autosampler stability) should be assessed at least in three replicates and at two QC, that is, LQC and HQC concentration levels. •





Freeze and Thaw Stability: Analyte stability in QC samples after at least three freeze and thaw cycles should be assessed. At least three aliquots of each of the low and high QC samples should be stored at the intended storage temperature (95% of protein molecules are modiied at Cys-17 with no covalent modiications at any other site. This alkylated form of IFN-β1a (NEMIFN) has approximately 50% lower activity compared to the intact protein and renders itself as an attractive model of protein degradation, whereby any change in protein activity and conformation must necessarily originate from the single covalent modiication. The diminished biological activity of NEM-IFN is likely mediated by changes in the protein higher-order structure. However, classical biophysical tools provide little information on conformational consequences of IFN alkylation. Alkylation of IFN with NEM did result in a detectable reduction of SEC elution time (consistent with the increase of hydrodynamic radius). In addition, analytical ultracentrifugation analysis indicated a slight reduction (of about 0.6%) in the sedimentation coeficient of the NEM-IFN monomer in comparison to the native IFN monomer, and luorescence measurements yielded a small blue shift [14], consistent with the notion of less stable tertiary structure of NEM-IFN. However, the paucity of the observed changes makes them unsuited for gauging the integrity of the higherorder structure. In the case of the far-UV CD spectra for NEM-IFN, a small reduction in the total helical content was evident, but the amplitude decrease of the negative bands at 208 and 222 nm, which are signatures of α-helical structure, was not very signiicant and not well suited to serve as a marker of the structure loss, especially in a situation when stress-affected IFN molecules constitute only a fraction of the entire protein population [14]. Ellipticity changes in the near-UV region indicated a partial loss of the protein’s tertiary structure upon alkylation [14], but the intrinsic weakness of the near-UV CD signal makes it hardly suitable for reliable detection of unfolding, especially if structurally compromised species represent only a fraction of the entire ensemble of protein molecules. Overall, the classical biophysical measure-

CONFORMATIONAL PROPERTIES OF PROTEIN DRUGS PROBED BY HDX MS

233

+9

+14 100

+13 Relative abundance

75

+10 +15

+8

50

+12

22250 22500 22750 23000 Mass, Da

25

+16

+11

0 1000

1500

2000

2500

m/z

FIGURE 10.4 ESI mass spectra of intact (darker trace) and NEM-alkylated IFN (lighter trace) buffer-exchanged to aqueous 100 mM ammonium acetate prior to MS analysis. The inset shows limited heterogeneity of IFN due to the presence of several glycoforms. Black trace shows an ESI mass spectrum of intact IFN acquired under strongly denaturing conditions (50% methanol, 6% acetic acid). Reproduced with permission from Kaltashov et al. [16].

ments provided little insight as to where these changes were occurring on IFN and in some cases were questionable as to their signiicance. Evidence of partial unfolding of IFN triggered by its alkylation was readily provided by electrospray ionization mass spectrometry (ESI-MS) [14]. Ions of intact IFN exhibited a narrow charge state distribution with low charge density, as expected for compact, tightly folded macromolecular species, while the charge state distribution of the NEM-IFN ions was clearly bimodal (Figure 10.4). Presence of the higher-charge density ions in ESI mass spectrum of NEM-IFN clearly signaled that a fraction of the protein molecules failed to maintain the compact native structure. Although the presence of the NEM-alkylated residue within IFN could be easily deduced from the mass shift of NEM-IFN ions, the detection of partial unfolding was based entirely on charge, rather than mass, measurements and, therefore, did not depend on the ability to detect any changes in covalent structure concurrently with (or prior to) the conformational analysis. In fact, this technique can detect changes in protein compactness triggered by nonenzymatic PTMs that do not alter the protein mass (e.g., disulide scrambling). It must be said that the extent of multiple charging of the NEM-IFN species representing the less compact conformations (charge states +11 through +14 in Figure 10.4) still implies the presence of some residual structure. Indeed, the average charge density of IFN ions in

ESI mass spectrum acquired under denaturing conditions (black trace in Figure 10.4) was noticeably higher. This suggests that the solvent accessible surface area of “noncompact” conformers of NEM-IFN, whose presence under near-native conditions is revealed by the protein ion charge state distribution analysis, is lower compared to the fully denatured species of IFN [32,33]. However, this technique yields only a global measure of conformational “disorder” and, therefore, cannot provide any information that would allow these unfolding events to be localized within the protein structure. This gap can be easily illed by HDX MS, which provides local information on backbone dynamics when the exchange reaction in solution is supplemented by proteolysis under the slow exchange conditions and MS analysis. HDX MS identiied several segments in NEM-IFN whose stability was greatly compromised by alkylation of Cys-17 [14]. One example is shown in Figure 10.5, where the evolution of the isotopic distribution of a peptic fragment [L88-L102] is traced over irst 120 s of exchange in solution. Very slow uptake of deuterium is exhibited by the fragment derived from intact IFN (darker trace), while the protein alkylation results in a dramatic acceleration of the exchange kinetics of this segment. Presence of several such segments whose exchange kinetics was signiicantly altered by the alkylation event clearly signaled a change in conformation and/or dynamics induced by the alkylation, consistent with the conclusions of the protein ion charge state distribution analysis.

HYDROGEN/DEUTERIUM EXCHANGE MASS SPECTROMETRY (HDX MS)

234 100

Unlabeled peptide

MSYNLLGFLQ RSSNFQCQKL LWQLNGRLEY030

0 875 100

880

885

890

m/z

CLKDRMNFDI PEEIKQLQQF QKEDAALTIY060

HDX 10 sec

EMLQNIFAIF RQDSSSTGWN ETIVENLLAN090 0 875 100

880

885

890

m/z

HDX 30 sec

VYHQINHLKT VLEEKLEKED FTRGKLMSSL120

0 875 100

880

885

890

m/z

HLKRYYGRIL HYLKAKEYSH CAWTIVRVEI150

HDX 120 sec

LRNFYFINRL TGYLRN166

0 875

880

885

890

m/z

FIGURE 10.5 Evolution of isotopic distributions of peptic fragments (88–102) derived from intact (darker trace) and NEMalkylated (lighter trace) IFN throughout the course of HDX. The end point of the exchange reaction is indicated with a gray trace (isotopic distribution of a fully exchanged peptide). Location of this peptide within the amino acid sequence of IFN is shown on the right. Adapted with permission from Bobst et al. [14].

Importantly, these measurements did not require that the protein be transferred to an ESI-MS-friendly buffer prior to the isotopic labeling, thereby allowing the conformational analysis to be carried out under relevant conditions (e.g., in the formulation buffer), since isolation of the protein drug material from the commercial product is often viewed as a step capable of altering its structural integrity [34]). Furthermore, the ability to localize the unfolding events within the protein sequence provides valuable information that can be used to predict the functional consequences of such partial unfolding. Just like the charge state distribution analysis, characterization of conformation and dynamics by HDX MS does not require any prior knowledge of the nonenzymatic PTMs that may or may not be present in the protein drug. In fact, identiication of all peptic fragments carried out prior to HDX MS analysis is likely to detect most PTMs within the protein. Even if some of them escape the detection, this would obviously make no impact on the validity of the HDX MS analysis of protein conformation and dynamics. The detailed information on the backbone protection of IFN and NEM-IFN provided by HDX MS measurements allows the conformational properties of intact and alkylated forms of the protein to be compared directly [14,16]. Perhaps the most important conclusion derived from the HDX MS work is the dramatic destabilization of helix D (residues 91–106) by the PTM event which affects a remote site in the amino acid

sequence of this protein. Although this segment is very distant from the residue affected by alkylation (Cys-17) within the protein sequence, it is proximal to the Cys-17 side chain within the three-dimensional structure of the protein (Figure 10.6). Analysis of the crystal structure of IFN (PDB id 1AU1 [35]) makes it clear that modiication of Cys-17 with NEM introduces a signiicant steric clash in the protein interior, which distorts helix packing, thereby adversely affecting the stability of the protein higherorder structure. One extreme scenario of “relieving” this steric clash is presented in Figure 10.6, where a total loss of secondary structure within helix D converts it to a lexible loop, thereby allowing the bulky NEM group to be accommodated within the protein structure. Another plausible scenario involves conversion of helix A to a lexible loop (this element of the secondary structure appears to be highly dynamic even in the absence of alkylation [14] and may also explain the possibility of forming a disulide-linked IFN dimer despite sequestration of Cys-17 side chain in the protein interior [31]). Since the alkylation increases the dynamic character of both helices (A and D), NEM-IFN is likely to transiently sample conformations where either or both of these elements of secondary structure are compromised either through more frequent local structural luctuations or via cooperative unfolding of the entire elements. Even transient unfolding of either helix A or D (or both) would inevitably increase the aggregation

PROTEIN INTERACTION WITH PHYSIOLOGICAL PARTNERS AND THERAPEUTIC TARGETS PROBED BY HDX MS

A

B

C

D

FIGURE 10.6 Location of the (88–102) segment within the crystal structure of IFN (1AU1, panel A) with respect to Cys17 (both are highlighted). Alkylation of Cys-17 inevitably leads to steric clashes within the native structure, which can be removed by unfolding of the helix D containing the (88– 102) segment (panel B). Side chains of hydrophobic residues within helix D are sequestered in the protein interior (highlighted in panel C), but become exposed to solvent upon unfolding of this structural element (panel D). Reproduced with permission from Kaltashov et al. [16].

propensity of IFN by exposing the hydrophobic residues, which are sequestered in the interhelical interfaces in the native structure of IFN (Figure 10.6). In fact, the diminished stability of NEM-IFN in solution compared to intact IFN became evident in the course of working with this protein. Furthermore, analysis of the backbone lexibility maps also provides some clues regarding the molecular mechanism of IFN inactivation following its alkylation. Indeed, the conformational “hot spots” are localized within a receptor binding interface, suggesting that alkylation may have signiicant impact on the ability of IFN to bind one of its receptors, a notion that was recently conirmed in our laboratory by monitoring IFN association with its cognate receptors using direct ESI-MS [16].

10.4 PROTEIN INTERACTION WITH PHYSIOLOGICAL PARTNERS AND THERAPEUTIC TARGETS PROBED BY HDX MS: MECHANISTIC ASPECTS OF TRANSFERRIN-RECEPTOR INTERACTION REVEALED BY HDX MS The HDX MS analysis of conformational consequences of IFN alkylation considered in the previous section shows an example of how this technique can be used to

235

detect and characterize changes in conformation and dynamics of biopharmaceutical products. Furthermore, the ability of HDX MS to localize protein segments with anomalous protection levels is very appealing as a means of mapping interaction sites, and is now actively evaluated for the purpose of optimizing the process of screening small molecule drug candidates [36–38]. Since protein drugs exert their therapeutic action via highly speciic interactions with their physiological targets, HDX MS can lend itself very useful in aiding the design of new therapies by providing detailed information on how biopharmaceuticals interact with their speciic receptors. An example of such work is shown in Figure 10.7, where HDX is used to probe conformational dynamics of an iron transporter transferrin (Tf) alone and in the receptor-bound form. Tf (an 80 kDa glycoprotein) is one of the very few plasma proteins that can enter the cell in the process of receptor-mediated endocytosis. Since the rapidly growing malignant cells have dramatically elevated requirements for iron consumption, they overexpress transferrin receptor (TfR), making Tf very attractive as a vehicle for selective delivery of cytotoxic agents to cancer cells [39–41]. Since the ability of Tf or any Tf-conjugate drug to enter the cells critically depends on its recognition by the TfR, detailed information on Tf/TfR interaction at the molecular level is extremely important for the drug development process [42]. Unfortunately, the large size and the presence of paramagnetic ferric ion (Fe3+) places the Tf/TfR complex out of reach of NMR, while it’s dynamic character and high carbohydrate content make acquisition of highresolution X-ray diffraction data highly problematic, leaving HDX MS as the only technique capable of providing structural information on Tf/TfR interaction. Both Tf-metal and Tf-TfR complexes readily dissociate under the slow exchange conditions prior to MS analysis; therefore, the protein mass evolution in each case relects solely deuterium uptake in the course of exchange in solution. The extra protection afforded by the receptor binding to Tf persists over an extended period of time (Figure 10.7), and it may be tempting to assign it to shielding of labile hydrogen atoms at the protein–receptor interface. However, this view is overly simplistic, as the conformational effects of protein binding are frequently felt well beyond the interface region. The difference in the backbone protection levels of receptor-free and receptor-bound forms of Fe2Tf appears to grow during the initial hour of the exchange, relecting signiicant stabilization of Fe2Tf higher-order structure by the receptor binding. Indeed, while the fast phase of HDX is typically ascribed to frequent local luctuations (transient perturbations of higher-order structure) affecting relatively small protein segments,

236

HYDROGEN/DEUTERIUM EXCHANGE MASS SPECTROMETRY (HDX MS)

ϳϭͲϴϭ

ϲϭϮͲϲϮϭ

618

1056 1060 1064 1068

ϭϭϯͲϭϯϰ

ϭϬŵŝŶ

614

616

770

772

ϭŵŝŶ

774 m/z

ϯϵϲͲϰϬϴ

666

668

670 m/z

ϮϵϱͲϯϬϵ

ϳϲϵϬϬϳϳϬϬϬ ϳϳϭϬϬ DĂƐƐ͕Ă 577 578 579 580 m/z

FIGURE 10.7 Left: HDX MS of Fe2Tf in the presence (darker trace) and the absence (lighter trace) of the cognate receptor. The exchange was carried out by diluting the protein stock solution 1:10 in exchange solution (100 mM NH4HCO3 in D2O, pH adjusted to 7.4) and incubating for a certain period of time as indicated on each diagram followed by rapid quenching (lowering pH to 2.5 and temperature to near 0°C). The black trace shows unlabeled protein. Right: Localizing the inluence of the receptor binding on backbone protection of Fe2Tf using bottom-up HDX MS on a physiologically relevant timescale. The panels show isotopic distributions of representative peptic fragments derived from the protein subjected to HDX in the presence (darker traces) and the absence (lighter traces) of the receptor and followed by rapid quenching. Dotted lines indicate deuterium content of unlabeled and fully exchanged peptides. Colored segments within the Fe2Tf/receptor complex show location of the peptic fragments. Adopted with permission from Kaltashov et al. [13].

the slower phases of HDX usually relect relatively rare, large-scale conformational transitions (transient partial or complete unfolding). This is why global HDX MS measurements similar to those presented in Figure 10.7 can be used to obtain quantitative thermodynamic characteristics for protein interaction with a variety of ligands, ranging from metal ions [43] and small organic molecules [44] to other proteins [45] and oligonucleotides [46]. Evolution of deuterium content of various peptic fragments in Figure 10.7 reveals a wide spectrum of protection, which is distributed very unevenly across the protein sequence. While some peptides exhibit nearly complete protection of backbone amides (e.g., segment [396–408] sequestered in the core of the protein C-lobe), exchange in some other segments is fast (e.g., peptide

[612–621] in the solvent-exposed loop of the C-lobe). The inluence of the receptor binding on the backbone protection is also highly localized. While many segments appear to be unaffected by the receptor binding, there are a few regions where exchange kinetics noticeably decelerates (e.g., segment [71–81] of the N-lobe, which contains several amino acid residues that form Tf/ receptor interface according to the available model of the complex based on low-resolution cryo-electron microscopy (cryo-EM) data [47]). Although the increased protection of backbone amides proximal to the protein/receptor binding interface is hardly surprising, HDX MS data also reveal a less trivial trend: acceleration of exchange kinetics in some segments of the protein as a result of receptor binding (such behavior is illustrated in Figure 10.7, with

CHALLENGES AND FUTURE DIRECTIONS

segment [113–134], a part of the N-lobe that is distal to the receptor). Therefore, in addition to mapping binding interface regions, HDX MS also provides a means to localize the protein segments that are affected by the binding indirectly via allosteric mechanisms. 10.5 CHALLENGES AND FUTURE DIRECTIONS 10.5.1 Challenging Targets: Large and Highly Heterogeneous Systems Biopharmaceutical products cover a very wide molecular weight range, but the larger protein systems are obviously more challenging targets. An increase in size of a protein drug also inevitably translates into elevated frequency of nonenzymatic PTMs, which often makes precise mapping of these changes a gargantuan task. Therefore, availability of a robust, easy-to-use, and reasonably highthroughput method to probe conformational integrity of such large systems directly would obviously be a boon to the biopharmaceutical industry. Implementation of HDX MS for conformational analysis of large (>50 kDa) protein drugs faces several challenges. It was already mentioned that nonspeciicity of pepsin cleavage combined with the large physical size of the protein typically results in a large number of fragment peptides, which must be identiied in order to extract meaningful HDX information. This problem is magniied by the frequent occurrence of glycosylation and disulide bonds in protein drugs, which both multiply the sheer number of candidate peptides that it a given mass and also make it more dificult to obtain meaningful sequence information using classical MS/ MS approaches. Furthermore, even complete identiication of all observed peptic fragments does not necessarily result in complete sequence coverage of a large protein. Indeed, the necessity to minimize the sample handling time prior to MS analysis as a means to limit the extent of back exchange typically translates to very short (and, therefore, crowded) liquid chromatography (LC) runs, while utilization of more eficient chromatographic schemes (such as ultrahigh performance liquid chromatography, UPLC) may improve the separation, it results in more extensive back exchange [48]. While MS detection itself (and particularly high-resolution MS) solves the problem of detection of multiple coeluting peaks, signal suppression is likely to result in loss of some of the peptides, thereby leaving gaps in sequence coverage. Despite these dificulties, signiicant progress was made recently in this ield, as demonstrated by the successful use of HDX MS to probe conformational properties of larger protein therapeutics, such as 63 kDa β-glucocerebrosidase [38], 73 kDa (nonredundant mass)

237

monoclonal antibody [15], as well as our own work with transferrin (80 kDa), an example of which is shown in the preceding section. Another serious challenge to the characterization of protein drugs is posed by the high degree of structural heterogeneity exhibited by a large fraction of these species. While glycosylation is an inherent (enzymatic) PTM that was originally the major source of heterogeneity in many protein therapeutics, the second generation of biopharmaceuticals often employs “designer” PTMs to enhance their therapeutic properties. This enhancement can be achieved by either engineering elaborate and extensive glycosylation patterns [49], conjugating the protein drug to a synthetic polymer [50] or to a carrier protein [51]. Some preliminary work carried out in our laboratory indicates that HDX MS can be successfully applied to probe the conformational stability of these systems. Perhaps the ultimate success of MS as a tool to probe the conformation of protein drugs and its acceptance in this capacity will be determined not only by its ability to deal with very large and highly heterogeneous systems, but by its adaptability to the speciic needs of the biopharmaceutical industry. Several important questions have to be addressed in order for MS to become an integral part of the analytical routine in assessing the conformation and dynamics of biopharmaceutical products. Can MS be used to probe the integrity of the higher-order structure in a true high-throughput fashion? Can the entire procedure (from sample handling to data processing) be automated and commercialized in a form of an easy-to-use “turnkey” instrument? What is the reproducibility and robustness of these measurements if they are carried out in different laboratories and/or using different LC-MS platforms? What is the sensitivity of this technique in detecting small conformational changes and/or alterations of the higherorder structure that affect only a small fraction of the protein molecules? Addressing these questions will require extensive concerted efforts of academic and industrial researchers, but the end result will certainly be beneicial and worthwhile. Indeed, adoption of MS by the biopharmaceutical industry in this new role will not only become a boon to analytical characterization, but is also certain to greatly catalyze the development of new and improvement of existing potent therapies. 10.5.2 Limitations of the Bottom-Up HDX MS Strategies: The Top-Down Approach to HDX MS: HDX MS/MS and Inclusion of Ion Fragmentation in the Bottom-Up HDX MS Schemes As has been mentioned above, spatial resolution in HDX MS measurements is determined by the number

238

HYDROGEN/DEUTERIUM EXCHANGE MASS SPECTROMETRY (HDX MS)

and the size of proteolytic fragments. In some favorable cases spatial resolution in HDX MS of small proteins (4)-1-thio-β-Dglucopyranoside (lactose aryl diazirine) on diamond. Surf. Interface Anal., 31, 457–464. 71. Wagner, M.S., Horbett, T.A., Castner, D.G. (2003) Characterizing multicomponent adsorbed protein ilms using electron spectroscopy for chemical analysis, timeof-light secondary ion mass spectrometry, and radiolabeling: capabilities and limitations. Biomaterials, 24, 1897–1908. 72. Davies, J., Nunnerley, C.S., Paul, A.J. (1996) A correlative study of the measurement of protein adsorption to steel, glass, polypropylene, and silicone surfaces using ToFSIMS and dynamic contact angle analyses. Colloids Surf. B, 6, 181–190. 73. Heard, P.J., Feeney, K.A., Allen, G.C., Shewry, P.R. (2001) Determination of the elemental composition of mature wheat grain using a modiied secondary ion mass spectrometer (SIMS). Plant J., 30(2), 237–245. 74. Pradier, C.M., Costa, D., Rubio, C., Compère, C., Marcus, P. (2002) Role of salts on BSA adsorption on stainless steel in aqueous solutions. I. FT-IRRAS and XPS characterization. Surf. Interface Anal., 34, 50–54. 75. Poleunis, C., Rubio, C., Compère, C., Bertrand, P. (2002) Role of salts on the BSA adsorption on stainless steel in aqueous solutions. II. ToF-SIMS spectral and chemical mapping study. Surf. Interface Anal., 34, 55–58. 76. Aoyagi, S., Hiromoto, S., Hanawa, T., Kudo, M. (2004) TOF-SIMS investigation of metallic material surface after culturing cells. Appl. Surf. Sci., 231–232, 470– 474. 77. Tyler, B., Ryal, G., Castner, D.G. (2007) Multivariate analysis strategies for processing ToF-SIMS images of biomaterials. Biomaterials, 28, 2412–2423. 78. Vaidyanathan, S., Fletcher, J.S., Goodacre, R., Lockyer, N.P., Mickleield, J., Vickerman, J.C. (2008) Subsurface biomolecular imaging of streptomyces coelicolor using secondary ion mass spectrometry. Anal. Chem., 80, 1942–1951. 79. Tempez, A., Schultz, J.A., Della-Negra, S., Depauw, J., Jacquet, D., Novikov, A., Lebeyec, Y., Pautrat, M., Caroff, M., Ugarov, M., Bensaoula, H., Gonin, M., Fuhrer, K., Woods, A. (2004) Orthogonal time-of-light secondary ion mass spectrometric analysis of peptides using large gold clusters as primary ions. Rapid Commun. Mass Spectrom., 18, 371–376. 80. Zhou, C., Qi, K., Wooley, K.L., Walker, A.V. (2008) Timeof-light secondary ion mass spectrometry, luorescence microscopy and scanning electron microscopy: combined tools for monitoring the process of pattering and layerby-layer assembly of synthetic and biological materials. Colloids Surf. B Biointerfaces, 65(1), 85–91.

REFERENCES

81. Lau, Y.T., Schultz, J.M., Weng, L.T., Ng, K.M., Chang, C.M. (2009) Control of the fold surface conformation of the lamellae of an oligomer. Langmuir, 25, 8263–8267. 82. Leufgen, K., Mutter, M., Vogel, H., Szymczak, W. (2003) Orientation modulation of a synthetic polypeptide in self-assembled monolayers: a TOF-SIMS study. J. Am. Chem. Soc., 125, 8911–8915. 83. Coullerez, G., Baborowski, J., Viornery, C., Chevolot, Y., Xanthopoulos, N., Ledermann, N., Muralt, P., Setter, N., Mathieu, H.J. (2003) Imaging by time-of-light secondary ion mass spectrometry of plasma patterned metal and oxide thin ilms. Appl. Surf. Sci., 203, 527–531. 84. Xia, N., Castner, D.G. (2003) Preserving the structure of adsorbed protein ilms for time-of-light secondary ion mass spectrometry analysis. J. Biomed. Mater. Res., 67A, 179–190. 85. Xia, N., May, C.J., McArthur, S.L., Castner, D.G. (2002) Time-of-light secondary ion mass spectrometry analysis of conformational changes in adsorbed protein ilms. Langmuire, 18, 4090–4097. 86. Arlinghaus, H.F., Ostrop, M., Friedriches, O., Feldner, J., Gunst, U., Lipinsky, D. (2002) DNA sequencing with ToFSIMS. Surf. interface Anal., 33, 35–39. 87. Hashimoto, H., Nakamura, K., Takase, H., Okamoto, T., Yamamoto, N. (2004) Quantitative TOF-SIMS imaging of DNA mircroarrays produced by bubble jet printing technique and the role of TOF-SIMS in life science industry. Appl. Surf. Sci., 231–232, 385. 88. Samuel, T., Castner, D.G. (2004) ToF-SIMS characterization hybridization onto self-assembled single-stranded DNA monolayers. Appl. Surf. Sci., 231–232, 397. 89. Malmberg, P., Nygren, H. (2008) Methods for the analysis of the composition of bone tissue, with a focus on imaging mass spectrometry (TOF-SIMS). Proteomics, 8, 3755–3762. 90. Kim, J.H., Kim, J.H., Ahn, B.J., Park, J.-H., Shon, H.K., Yu, Y.S., Moon, D.W., Lee, T.G., Kim, K.-W. (2008) Label-free calcium imaging in ischemic retinal tissue by TOF-SIMS. Biophys. J., 94, 4095–4102. 91. Chandra, S. (2001) Studies of cell division (mitosis and cytokinesis) by dynamic secondary ion mass spectrometry ion microscopy: LLC-PK1 epithelial cells as a model for subcellular isotopic imaging. J. Microsc., 204, 150–165.

257

95. Malmberg, P., Nygren, H., Richter, K., Chen, Y., Dangardt, F., Friberg, P., Magnusson, Y. (2007) Imaging of lipids in human adipose tissue by cluster ion TOF-SIMS. Microsc. Res. Tech., 70, 828–835. 96. Benabdellah, F., Seyer, A., Quinton, L., Touboul, D., Benabdellah, A., Laprévote, O. (2010) Mass spectrometry imaging of rat brain sections: nanomolar sensitivity with MALDI versus nanometer resolution by TOF-SIMS. Anal. Bioanal. Chem., 396, 151–162. 97. Börner, K., Nygren, H., Hagenhoff, B., Malmberg, P., Tallarek, R., Månsson, J.E. (2006) Distribution of cholesterol and galactosylceramide in rat cerebella white matter. Biochim. Biophys. Acta, 1761, 335–344. 98. Mathieu, H.J. (2001) Bioengineered material surfaces for medical applications. Surf. Interface Anal., 32, 3–9. 99. Aubagnac, J.L., Enjalbal, C., Drouot, C., Combarieu R., Martinez, J. (1999) Imaging time-of-light secondary ion mass spectrometry of solid-phase peptide syntheses. J. Mass Spectrom., 34, 749–754. 100. Aoyagi, S., Hayama, M., Hasegawa, U., Sakai, K., Tozu, M., Hoshi, T., Kudo, M. (2004) Estimation of protein adsorption on dialysis membrane by means of TOFSIMS imaging. J. Memb. Sci., 236(1–2), 91–99. 101. Aoyagi, S., Kawashima, Y., Kudo, M. (2005) TOF-SIMS imaging technique with information entropy. Nucl. Instrum. Methods Phys. Res. B, 232, 146–152. 102. Chatterjee, R. (2004) 84. Application of TOF-SIMS to monitor coating processes on biological and organic surfaces. Appl. Surf. Sci., 231–232, 437–441. 103. Lloyd, K.G., O’Keefe, D.P. (2004) Re-discovering surface mass spectrometry: chemical mapping from micro to macro. Appl. Surf. Sci., 231–232, 207–216. 104. Hagenhoff, B. (1995) Surface mass spectrometry— application to biosensor characterization. Biosens. Bioelectron., 10, 885–894. 105. Delcorte, A., Hermans, S., Devillers, M., Lourette, N., Aubriet, F., Muller, JF, Bertrand, P. (2004) Desorption/ ionization of molecular nanoclusters: SIMS versus MALDI. Appl. Surf. Sci., 231–232, 131–135. 106. McDonnell, L.A., Piersma, S.R., Maarten Altelaar, A.F., Mize, T.H., Luxembourg, S.L., Verhaert, P.D., van Minnen, J., Heeren, R.M. (2005) Subcellular imaging mass spectrometry of brain tissue. J. Mass Spectrom., 40(2), 160.

92. Tanji, N., Okamoto, M., Katayama, Y., Hosokawa, M., Takahata, N., Sano, Y. (2008) Investigation of the cosmetic ingredient distribution in the stratum corneum using NanoSIMS imaging. Appl. Surf. Sci., 255, 1116. 93. Poleunis, C., Rubio, C., Compère, C., Bertrand, P. (2003) ToF-SIMS chemical mapping study of protein adsorption onto stainless steel surfaces immersed in saline aqueous solutions. Appl. Surf. Sci., 203–204, 693-697.

108. Kollmer, F., Bourdos, N., Kamischke, R., Benninghoven, A. (2003) Nonresonant Laser-SIMS and TOF-SIMS analysis of sub-[mu] m structures. Appl. Surf. Sci., 203–204, 238–243.

94. Okamoto, M., Tanji, N., Katayama, Y., Okada, J. (2006) TOF-SIMS investigation of the distribution of a cosmetic ingredient in the epidermis of the skin. Appl. Surf. Sci., 252, 6805.

109. Kubota, N., Hayashi, S. (2008) Application of resonant laser postionization SNMS for quantitative depth proiling in stainless steel with oxide ilm. Appl. Surf. Sci., 255, 1516–1518.

107. Murayama, Y., Komatsu, M., Kuge, K., Hashimoto, H. (2006) Enhanced peptide molecular imaging using aqueous droplets. Appl. Surf. Sci., 252(19), 6774–6776.

258

TOF-SIMS APPLICATIONS TO BIOIMAGING AND BIOMOLECULE EVALUATION METHODS

110. Kubota, N., Hayashi, S. (2008) State selective detection of sputtered Al neutrals by resonant laser ionization SNMS. Appl. Surf. Sci., 255, 834–836. 111. Koizumi, M., Sakamoto, T. (2008) Resonance enhanced multi-photon ionization of neutral atoms sputtered with Ga-FIB. Appl. Surf. Sci., 255, 901–904. 112. Csik, A., Vad, K., Langer, G.A., Katona, G.L., Tóth-Kádár, E., Péter, L. (2009) Analysis of Co/Cu multilayers by SNMS reverse depth proiling. Vacuum, 84(1), 141–143.

113. Nakata, Y., Honda, Y., Ninomiya, S., Seki, T., Aoki, T., Matsuo, J. (2009) Matrix-free high-resolution imaging mass spectrometry with high-energy ion projectiles. Mass Spectrom., 44(1), 128–136. 114. Nakata, Y., Yamada, H., Honda, Y., Ninomiya, S., Seki, T., Aoki, T., Matsuo, J. (2009) Imaging mass spectrometry with nuclear microprobes for biological applications. Nucl. Instrum. Methods Phys. Res. B, 267, 2144– 2148.

12 ACCELERATOR MASS SPECTROMETRY IN PHARMACEUTICAL DEVELOPMENT Benjamin J. Stewart, Graham Bench, Bruce A. Buchholz, Kurt W. Haack, Michael A. Malfatti, Ted J. Ognibene, and Kenneth W. Turteltaub

12.1 INTRODUCTION TO BIOLOGICAL ACCELERATOR MASS SPECTROMETRY 12.1.1 Overview of Biological Accelerator Mass Spectrometry Accelerator mass spectrometry (AMS) is an analytical method for the measurement of rare, long-lived isotopes with extremely high sensitivity. In contrast to traditional liquid scintillation counting (LSC), which relies on nuclear decay events to quantify radioactive material, AMS counts atoms of a rare isotope independent of decay by measuring the mass ratio of the radioisotope of interest relative to a stable isotope of the element. AMS is currently the most sensitive technique available for measuring 14C, with the ability to quantify labeled material over six orders of magnitude down to the attomole (10−18) level [1–4]. The exquisite sensitivity of AMS has enabled the use of human subjects in a number of biomedical tracer studies that would not otherwise be possible, since experiments can be designed to produce negligible radiological exposure to the experimental subjects without perturbing the normal biology of the system under consideration [2,3]. AMS was initially applied in isotope dating in the ields of archaeology and earth science, and more recently has been applied to the realm of experimental biology, including the areas of pharmacology and toxicology [3] Applications of AMS in the biological sciences include investigation of tissue turnover rates [5,6], pharmacokinetic parameters of drugs [7,8], toxicants

[9], and nutrients [10], including absorption, distribution, metabolism, and excretion, as well as protein and DNA binding of drugs and reactive metabolites [11,12] and risk assessment [13,14]. The ability to administer and accurately measure extremely low levels of a compound of interest using AMS is especially promising for application in the realm of pharmaceutical development. Tracer experiments using AMS to elucidate the disposition and metabolism of investigational compounds in phase I studies can provide valuable data that will help determine whether an investigational compound warrants further development. AMS is also well -suited for determining molecular targets of a compound of interest in order to identify pharmacological sites of action or toxicological mechanisms. Subtherapeutic or “microdosing” experiments have made increasing use of AMS in recent years, allowing investigational compounds to be introduced into humans earlier than would be possible using more traditional techniques [8,15]. Because of the sensitive, quantitative, and precise nature of the technique, AMS can provide important information regarding the disposition and metabolism of the test compound, and could potentially become an important early screening tool in drug development. Some situations in which AMS may be the most appropriate measurement technique include: • •

compounds with low bioavailability, compounds with low systemic distribution,

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

259

260 • •



ACCELERATOR MASS SPECTROMETRY IN PHARMACEUTICAL DEVELOPMENT

highly potent compounds, compounds created by dificult synthesis procedures where only limited quantities can be produced, and experiments where a tracer must be measured for long periods of time (weeks or months).

12.1.2

Capabilities and Limitations of AMS

AMS is most suitable for the measurement of longlived, low natural abundance radioisotopes. It is important that the radioisotope of interest have a low natural occurrence so that the labeled compound can be discriminated from background levels while maintaining a low dose of the labeled material. This precludes the use of stable isotopes such as 13C, which is naturally present in relatively high abundance and comprises approximately 1% of naturally occurring carbon. In addition to 14 C, many other radioisotopes are suitable for measurement by AMS, but are generally of limited utility to the biological scientist. Biological AMS studies to date have made use of 3H [16], 41Ca [17], and 10Be [18]. AMS is unlike other forms of mass spectrometry in that the spectrometer is designed speciically for quantitation of the isotopes of a single element. AMS does not provide any structural information, and samples must be properly and adequately deined prior to AMS analysis in order to ensure that results are meaningful. Although a typical sample can be measured by AMS in a matter of a few minutes, the rate of sample preparation limits the throughput of the technique. Samples intended for AMS measurement must be converted to a form that is easily conductive in the ion source, and which does not allow for isotopic fractionation. Currently most spectrometers use solid graphite on a conductive core such as cobalt or iron [19,20]. Thus, in order to measure a single high performance liquid chromatography (HPLC) trace, it may be necessary to prepare up to 100 samples for AMS by graphitization [21]. Because of the sample preparation time, it can require several days between the time the samples were collected and the time when they are measured by AMS. AMS is not suitable for all labeling experiments, particularly those in which sample material is plentiful and radiological dose is not an important consideration. Other techniques that may be considered in place of AMS include LSC, isotope ratio mass spectrometry using 13C, 13C-nuclear magnetic resonance (NMR), or in some cases luorescent or chemical tagging of the compound or molecule of interest. The level of sensitivity, importance of maintaining a low radiological dose, number of samples, complexity of samples, and cost of analysis are major considerations in determining the appropriateness of using AMS.

TABLE 12.1. Basic Performance Characteristics for the Biological AMS Instrument at Lawrence Livermore National Laboratory Instrumental Performance

Approximate Value

Speciicity (background signal with no 14C present) Stability (loss of 14C during sample preparation) Linear range of measurement Instrument carryover between samples Reproducibility Precision

Two parts per 1015

Lower limit of quantitation Upper limit of quantitation Signal recovery

12.1.3

CV < 3% 10−2–103 Modern None 300°C), high carbon content, and low 14C content. A typical addition of carrier carbon occurs by adding 1 or 2 µL of tributyrin in capillary tubes to the sample. It is always advisable to analyze each new type of sample for carbon content prior to graphitization to ensure that a proper carbon inventory is maintained. 12.2.1.4 Sample Analysis and Interpreting the Results AMS reports results in Modern, which is a deined unit of 14C/C isotope concentration equal to 1.18 × 10−12 14C/C. Conversion of this number to a meaningful result requires careful inventory of the carbon sources in the sample. For the simplest case in which a sample contains suficient carbon so as not to require the addition of carrier carbon, this conversion is facile. In many experiments, it is necessary to add carrier to make up the total amount of carbon necessary for the graphitization process to occur. In these experiments, the fraction Modern and the carbon mass of both the sample and the carrier must be considered in the calculation. Correctly accounting for all sources of carbon in the analyzed sample is essential, as a low estimate of carbon will artiicially increase the apparent quantity of tracer, and an overestimate of carbon will delate the apparent amount of tracer present. The measured ratio is a composite value that includes the total amount of 14C and the total carbon present in the measured sample. It is necessary to know the quantity of carbon and 14C in the sample tissue, as well as the amount of carbon due to the tracer, in order to solve for the 14C due to the tracer. If the only sources of carbon in the measured sample are the tracer and the tissue, determining the 14C due to the tracer proceeds as mentioned in the succeeding paragraphs. In most cases, the carbon due to the tissue is much greater than the carbon due to the tracer, so that the tracer carbon can be neglected in the calculation:

C measured 14 C tracer + 14 C tissue = , (12.1) C measured C tracer + C tissue 14 C tracer + 14 C tissue (12.2) = . C tissue 14

R measured = R measured

Because the measured fraction Modern is a composite of the contribution of the tissue and the contribution of the tracer, rearranging to solve for the tracer 14C gives: 14

C tracer = C tissue × (R measured − R tissue ),

(12.3)

where: 14

R tissue =

C tissue . C tissue

(12.4)

If carrier carbon is also present, then the general equation is as follows: 14

R measured =

C tracer + 14 C tissue + 14 C carrier . C tracer + C tissue + 14 C carrier

(12.5)

In this case, the carrier carbon is usually much greater than the carbon due to the tissue or the tracer, allowing these quantities to be neglected in the calculation: C tracer + 14 C tissue + 14 C carrier , C carrier

(12.6)

C tracer = C carrier × (R measured − (R tissue + R carrier ).

(12.7)

14

R measured = which becomes: 14

Additional example calculations can be found in Section 12.3.

12.2.2

Speciic Applications of AMS

AMS excels at detecting very low levels of labeled compounds of interest, and has been applied to a variety of types of experiments applicable to drug development. The common feature of these experiments is the ability to trace the fate of a labeled compound in a biological system. This can include determination of pharmacokinetic parameters, biotransformation of the labeled compound, formation of covalent adducts with cellular macromolecules, drug– ligand interactions, and tissue or subcellular localization of the compound of interest. Such experiments can provide valuable information about the pharmacological and toxicological characteristics of investigational compounds. AMS may likewise be used to investigate metabolism of endogenous, naturally occurring compounds within a

APPLICATIONS OF AMS IN DRUG DEVELOPMENT

biological system without perturbing the normal physiology of the system. The types of experiments listed below are by no means exhaustive, but are intended to give the reader a sense of the potential value of AMS in the drug development process. 12.2.2.1 Pharmacokinetics Several studies have compared the pharmacokinetic proiles of drugs using standard therapeutic doses and subtherapeutic doses, commonly known as “microdosing” [8,15,29]. In these experiments, a subtherapeutic level of a 14C-labeled investigational compound is administered to animals or human volunteers, and the plasma levels of the compound are measured over time to determine the clearance rate and other pharmacokinetic parameters of the compound. These experiments are essentially the same as traditional “dose and measure” pharmacokinetic experiments, but the investigational compound can be administered and measured at much lower levels than are possible using standard techniques. This allows for human studies to be performed without the risk of harmful effects that could occur with the administration of higher doses. While the utility of such microdosing experiments as a replacement for standard pharmacokinetic techniques remains a matter of discussion, several reports indicate that microdosing may be useful as a screening tool for drug candidates [8]. An understanding of the pharmacokinetic parameters of a drug candidate may disqualify the compound for further investigation based on a poor pharmacokinetic proile or may provide guidelines for predicting optimal doses within the therapeutic range. In planning and conducting pharmacokinetics studies using AMS as a measurement technique, it is essential that an appropriate dose of the labeled material be administered to the test subject in order to allow accurate measurement at the selected time points. Assuming that the quantity of drug to be administered is signiicantly below the level required to elicit a pharmacological or toxic effect, dosing can be calculated based on the level of 14C required to detect an adequate signal in the collected sample. If higher levels of drug are to be administered, then the dose should be calculated based on the therapeutic dose range in addition to the radiological dose. The general steps involved in performing a pharmacokinetics study using AMS are as follows: •



Determine the speciic activity of the 14C-labeled compound of interest to be used in the pharmacokinetics study. Determine the mode(s) of administration of the compound of interest (injection, oral, dermal, subcutaneous, etc.). If appropriate, estimate the bioavailability of the compound.











265

Determine the anticipated length of time the compound should be followed, the number of samples to be collected (blood, urine), and the sampling interval. Determine the minimum target fraction Modern to be achieved in the sample (blood, urine) at the inal time point. In most cases, this level should be at least 10% above 1 Modern in order to allow appropriate resolution. Calculate the initial dose based on the estimated number of biological half-lives of the compound to produce the desired ending fraction Modern. Determine the total radiological exposure to be achieved using the calculated initial dose, assuming that all radioactive decay energy will be absorbed by the subject. Consult with a health physicist to ensure that this level of exposure is acceptable based on current regulations and requirements. For purposes of assessing the radiological burden, a human can be assumed to have a normal 14C level of 1 Modern—which for a 70-kg person corresponds to an energy deposit of 110 nJ/h—and assuming that the radiological dose is uniformly distributed throughout the body, this corresponds to 1.6 nSv/h [19]. If no information about the compound’s bioavailability and biological half-life is available, then calculate the initial dose based on the target highest amount of label to be present in the sample, and proceed to calculate the total radiological exposure. The initial samples collected should be measured by LSC to ensure that the calculations were correct and that the level of dosing was correct for AMS measurement.

12.2.2.2 Biotransformation and Conjugate Formation Identiication of metabolites of a drug candidate or other molecule of interest can be greatly facilitated by AMS. Tracing of a 14C label allows unequivocal identiication of compounds originating from the 14C-labeled parent compound, including metabolites and degradation products, even if the structures of these compounds are not previously known. This type of experiment may be performed by collecting blood and urine and separating components by HPLC or other separation techniques, and identifying the parent compound as well as any metabolites or conjugates. Appropriate separation techniques are essential for the identiication of metabolites. Although the presence of label introduced as the parent compound unambiguously demonstrates the source of the label, it is important to ensure that the collected material is reasonably

266

ACCELERATOR MASS SPECTROMETRY IN PHARMACEUTICAL DEVELOPMENT

pure to allow further characterization of novel metabolites. In many instances, the metabolites of a compound of interest may be unknown or incompletely characterized. A typical metabolism study consists of the following components: •











Characterize chemical purity, radiopurity, and speciic activity of the 14C-labeled compound of interest. Administer the labeled material and collect samples (blood, urine, tissue, etc.). Perform mass balance to account for all labeled material. Separate metabolites by standard techniques (HPLC, electrophoresis, etc.). Remove any solvents used in separations processes and analyze the sample fractions by AMS, adding carrier as necessary. In the case of HPLC fractions, the amount of carbon present in each sample is usually negligible compared to the amount of carrier carbon. Analysis occurs by comparison of detector peaks (for example, UV) with AMS results to identify fractions containing the parent compound, its metabolites, and conjugates (e.g., glutathione conjugates).

12.2.2.3 Protein and DNA Adduct Measurement for Determination of Reactive Metabolites Incorporation of label into protein or DNA is a useful indicator of adduct formation or noncovalent binding of the moiety of interest. This approach has been used very successfully in the evaluation of known or potential carcinogens that exhibit DNA-binding activity [11,12,30,31]. This type of experiment may be particularly valuable in assessing the potential toxicity, mutagenicity, or carcinogenicity of a compound of interest. For example, AMS was used to detect DNA adducts in the colon after humans were exposed to a dietary-relevant dose of the cooked food carcinogen PhIP [11]. It may also be important to determine the quantitative fate of compounds that can form reactive intermediates capable of binding to cellular macromolecules, for example, compounds that can form immunogenic haptens or glutathione conjugates. Experiments of this type can also be designed to distinguish covalent and noncovalent interactions. As in the types of experiments described in previous sections, it is absolutely essential that the sample be thoroughly characterized and as pure as is reasonably achievable to ensure that results are not confounded by the presence of contaminating materials. These experiments include the following parameters:

• •

• •



Determine the tissue to be sampled Calculate the dose of compound to administer and the quantity of tissue to collect Administer the dose and collect sample material Separate and quantify the biomolecule of interest (protein, DNA) Perform AMS measurement on the sample

12.2.2.4 Drug–Ligand Binding and Subcellular Localization Tracing the radiolabel into tissues, cells, and even subcellular compartments is possible using AMS [32,33]. These experiments may be of great importance in determining the site of action and speciic molecular target(s) of a drug or other compound of interest. Identifying the site of action or the molecular target of a drug candidate is very important in characterizing its biological activity and elucidating its mechanism of action. Such experiments could be a valuable screening tool to help in the selection of lead compounds for development as investigational drugs, or to indicate that a compound has undesirable binding or localization characteristics early in the drug development process. As with other types of experiments, the sample material must be thoroughly characterized prior to AMS analysis. The steps involved in these types of experiments are similar to those described in the previous section, but may include cellular fractionation in addition to isolation of target biomolecules.

12.3 EXAMPLE CALCULATION OF DOSED TISSUE CONCENTRATION The conversion of a measured isotope ratio of a dosed tissue on the spectrometer to a biologically meaningful compound concentration in neat tissue is straightforward. The procedure below gives a concentration of molecular equivalents of parent compound in the tissue. The measured isotope ratio Rmeasured can be expressed as shown previously in Equation 12.1: 14

R measured =

C tracer + 14 C tissue , C tracer + C tissue

(12.8)

where: 14

Ctracer is the amount of 14C from the labeled compound, 14 Ctissue is the endogenous level of 14C in the tissue, Ctracer is the amount of carbon from the labeled compound, and Ctissue is the amount of carbon in the tissue.

REFERENCES

Equation 12.8 is simpliied in that Ctissue ≫ Ctracer and can thus be neglected, as previously demonstrated in Equation 12.9: 14

C tracer = (R measured − R tissue ) × C tissue,

(12.9)

where Rtissue is the 14C/C isotope ratio of the undosed tissue, obtained from the predosed control tissue. Consider the results of the following example using a labeled compound (59 nCi, 58 Ci/mol) detected in plasma after an oral dose:

267

cellular localization. AMS can also be applied to variations of these experiments, as well as additional types of experiments. This chapter has described the fundamentals of biological AMS and provides a starting point for researchers interested in performing studies that involve AMS measurements. Researchers who have not previously conducted AMS experiments are strongly encouraged to consult with scientists experienced in AMS prior to beginning their own studies.

REFERENCES Rmeasured = 2.50 Modern (1 Modern = 97.89 amol 14C/ mg C) Rtissue = 1.08 Modern Ctissue = 0.042 mg C/mg plasma The excess 14C per unit mass of tissue is calculated by application of Equation 12.4: 14

C tracer /mg tissue = (2.50 − 1.08) × 0.042 × 97.89 = 5.84 amol 14 C/mg plasma.

The excess 14C/mass tissue can be converted to a compound concentration in the tissue, knowing that 1 14 C/C per molecule has a speciic activity of 62.4 Ci/mol and that the density of plasma is about 1.02 g/mL: 62.44 Moles trace compound Excess 14C × ρplasma = × Volume tissue Gram tissue Sp. Act 62.4 = 5.84 × × 1.02 × 10 3 58 = 6.41 × 103 amol compound/mL plasma = 6.41 fmol compound/mL plasma.

12.4

SUMMARY

AMS is a powerful measurement technique offering exquisite sensitivity and precise quantitation of isotopically labeled compounds of interest. The primary use of AMS is in the measurement of very low levels of 14 C-labeled tracer compounds. Valuable information regarding absorption, metabolism, distribution, excretion, and toxicity of experimental compounds can be obtained through properly designed AMS experiments. The types of studies discussed include pharmacokinetics, biotransformation and conjugate formation, protein and DNA adduct measurement for determination of reactive metabolites, and drug–ligand binding and sub-

1. Brown, K., Dingley, K.H., Turtelatub, K.W. (2005) Accelerator mass spectrometry for biomedical research. Methods in Enzymology, 402, 423–443. 2. Brown, K., Tompkins, E.M., White, I.N.H. (2006) Applications of accelerator mass spectrometry for pharmacological and toxicological research. Mass Spectrometry Reviews, 25, 127–145. 3. Turteltaub, K.W., Vogel, J.S. (2000) Bioanalytical applications of accelerator mass spectrometry form pharmaceutical research. Current Pharmaceutical Design, 6, 991–1007. 4. White, I.N.H., Brown, K. (2004) Techniques: the application of accelerator mass spectrometry to pharmacology and toxicology. Trends in Pharmacological Sciences, 25(8), 442–447. 5. Spalding, K.L., Arner, E., Westermark, P.O., Bernard, S., Buchholz, B.A., Bergmann, O., Blomqvist, L., Hoffstedt, J., Näslund, E., Britton, T., Concha, H., Hassan, M., Rydén, M., Frisén, J., Arner, P. (2008) Dynamics of fat cell turnover in humans. Nature, 453(7196), 783–787. 6. Bergmann, O., Bhardwaj, R.D., Bernard, S., Zdunek, S., Barnabé-Heider, F., Walsh, S., Zupicich, J., Alkass, K., Buchholz, B.A., Druid, H., Jovinge, S., Frisén, J. (2009) Evidence for cardiomyocyte renewal in humans. Science, 324(5923), 98–102. 7. Boddy, A.V., Sludden, J., Grifin, M.J., Garner, C., Kendrick, J., Mistry, P., Dutreix, C., Newell, D.R., O’Brien, S.G. (2007) Pharmacokinetic investigation of imatinib using accelerator mass spectrometry in patients with chronic myeloid leukemia. Clinical Cancer Research, 13(14), 4164–4169. 8. Lappin, G., Kuhnz, W., Jochemsen, R., Kneer, J., Chaudhary, A., Oosterhuis, B., Drijfhout, W.J., Rowland, M., Garner, R.C. (2006) Use of microdosing to predict pharmacokinetics at the therapeutic dose: experience with 5 drugs. Clinical Pharmacology and Therapeutics, 80(3), 203–215. 9. Buchholz, B.A., Fultz, E., Haack, K.W., Vogel, J.S., Gilman, S.D., Gee, S.J., Hammock, B.D., Hui, X., Wester, R.C., Maibach, H.I. (1999) HPLC-accelerator MS measurement of atrazine metabolites in human urine after dermal exposures. Analytical Chemistry, 71(16), 3519–3525. 10. Buchholz, B.A., Arjomand, A., Dueker, S.R., Schneider, P.D., Clifford, A.J., Vogel, J.S. (1999) Intrinsic erythrocyte labeling and attomole pharmacokinetic tracing of

268

ACCELERATOR MASS SPECTROMETRY IN PHARMACEUTICAL DEVELOPMENT 14

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

C-labeled folic acid with accelerator mass spectrometry. Analtyical Biochemistry, 269, 348–352. Malfatti, M.A., Dingley, K.H., Nowell-Kadlubar, S., Ubick, E.A., Mulakken, N., Nelson, D., Lang, N.P., Felton, J.S., Turteltaub, K.W. (2006) The urinary metabolite proile of the dietary carcinogen 2-aminon-1-methyl-6phenylimidazo[4,5-b]pyridine is predictive of colon DNA adducts after a low-dose exposure in humans. Cancer Research, 66(21), 10541–10547. Brown, K., Tompkins, E.M., Boocock, D.J., Martin, E.A., Farmeer, P.B., Turteltaub, K.W., Ubick, E., Hemingway, D., Horner-Glister, E., White, I.N.H. (2007) Tamoxifen forms DNA adducts in human colon after administration of a single [14C]-labeled therapeutic dose. Cancer Research, 67(14), 6995–7002. Creek, M.R., Mani, C., Vogel, J.S., Turteltaub, K.W. (1997) Tissue distribution and macromolecular binding of extremely low doses of [14C]-benzene in B6C3F1 mice. Carcinogenesis, 18(12), 2421–2427. Dingley, K.H., Curtis, K.D., Nowell, S., Felton, J.S., Lang, N.P., Turteltaub, K.W. (1999) DNA and protein adduct formation in the colon and blood of humans after exposure to a dietary-relevant dose of 2-amino-1-methyl-6phenylimidazo[4,5-b]pyridine. Cancer Epidemiology, Biomarkers, and Prevention, 8(6), 507–512. Sandhu, P., Vogel, J.S., Rose, M.J., Ubick, E.A., Brunner, J.E., Wallace, M.A., Adelsberger, J.K., Baker, M.P., Henderson, P.T., Pearson, P.G., Baillie, T.A. (2004) Evaluation of microdosing strategies for studies in preclinical drug development: demonstration of linear pharmacokinetics in dogs of a nucleoside analog over a 50-fold dose range. Drug Metabolism and Disposition, 32, 1254–1259. Dingley, K.H., Roberts, M.L., Velsko, C.A., Turteltaub, K.W. (1998) Attomole detection of 3H in biological samples using accelerator mass spectrometry: application in low-dose, dual-isotope tracer studies in conjunction with 14C accelerator mass spectrometry. Chemical Research in Toxicology, 11(10), 1217–1222. Fitzgerald, R.L., Hillegonds, D.J., Burton, D.W., Grifin, T.L., Mullaney, S., Vogel, J.S., Deftos, L.J., Herold, D.A. (2005) 41Ca and accelerator mass spectrometry to monitor calcium metabolism in end stage renal disease patients. Clinical Chemistry, 51(11), 2095–2102. Chiarappa-Zucca, M.L., Finkel, R.C., Martinelli, R.E., McAninch, J.E., Nelson, D.O., Turteltaub, K.W. (2004) Measurement of beryllium in biological samples by accelerator mass spectrometry: applications for studying chronic beryllium disease. Chemical Research in Toxicology, 17(12), 1614–1620. Vogel, J.S., Love, A.H. (2005) Quantitating isotopic molecular labels with accelerator mass spectrometry. Methods in Enzymology, 402, 402–422. Ognibene, T.J., Bench, G., Vogel, J.S. (2003) A highthroughput method for the conversion of CO2 obtained from biochemical samples to graphite in septa-sealed vials

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

for quantiication of 14C via accelerator mass spectrometry. Analytical Chemistry, 75, 2192–2196. Buchholz , B.A., Dueker, S.R., Lin, Y., Cliffors, A.J., Vogel, J.S. (2000). Methods and applications of HPLC-AMS. Nuclear Instruments and Methods in Physics Research. Section B, 172, 910–914.. Ognibene, T.J., Bench, G., Brown, T.A., Vogel, J.S. (2004) The LLNL accelerator mass spectrometry system for biochemical 14C-measurements. Nuclear Instruments and Methods in Physics Research. Section B, 223–224, 12–15. McIntyre, C.P., Sylva, S.P., Roberts, M.L. (2009) Gas chromatograph-combustion system for 14C-accelerator mass spectrometry. Analytical Chemistry, 81(15), 6422–6428. Uhl, T., Kretschmer, W., Luppold, W., Scharf, A. (2004) Direct coupling of an elemental analyzer and a hybrid ion source for AMS measurements. Radiocarbon, 46(1), 65–75. Flarakos, J., Liberman, R., Tannenbaum, G., Skipper, S.R., Skipper, P.L. (2008) Integration of continuous-low accelerator mass spectrometry with chromatography and massselective detection. Analytical Chemistry, 80(13), 5079–5085. Buchholz, B.A., Freeman, S.P.H.T., Haack, K.W., Vogel, J.S. (2000) Tips and traps in the 14C bio-AMS preparation laboratory. Nuclear Instruments and Methods in Physics Research. Section B, 172, 404–408. Getachew, G., Kim, S.H., Burri, B.J., Kelly, P.B., Haack, K.W., Ognibene, T.J., Buchholz, B.A., Vogel, J.S., Modrow, J., Clifford, A.J. (2006) How to convert biological carbon into graphite for AMS. Radiocarbon, 48(3), 325–336. Zermeno, P., Kurdyla, D.K., Buchholz, B.A., Heller, S.J., Kashgarian, M., Frantz, B.R. (2004) Prevention and removal of elevated radiocarbon contamination in the LLNL/CAMS natural radiocarbon sample preparation laboratory. Nuclear Instruments and Methods in Physics Research. Section B, 223–334, 293–297. Miyaji, Y., Ishizuka, T., Kawai, K., Hamabe, Y., Miyaoka, T., Oh-Hara, T., Ikeda, T., Kurihara, A. (2009) Use of an intravenous microdose of 14C-labeled drug and accelerator mass spectrometry to measure absolute oral bioavailability in dogs; cross-comparison of assay methods by accelerator mass spectrometry and liquid chromatography-tandem mass spectrometry. Drug Metabolism and Pharmacokinetics, 24(2), 130–138. Coldwell, K.E., Cutts, S.M., Ognibene, T.J., Henderson, P.T., Philips, D.R. (2008) Detection of Adriamycin-DNA adducts by accelerator mass spectrometry at clinically relevant Adriamycin concentrations. Nucleic Acids Research, 36(16), 1–10. da Costa, G.G., Marques, M.M., Beland, F.A., Freeman, J.P., Churchwell, M.I., Doerge, D.R. (2003) Quantiication of tamoxifen DNA adducts using on-line sample preparation and HPLC-electrospray ionization tandem mass spectrometry. Chemical Research in Toxicology, 16, 357–366.

REFERENCES

32. Beumer, J.H., Garner, R.C., Cohen, M.B., Galbraith, S., Duncan, G.F., Grifin, T., Beijnen, J.H., Schellens, J.H.M. (2007) Human mass balance study of the novel anticancer agent ixabepilone using accelerator mass spectrometry. Investigational New Drugs, 25, 327–334.

269

33. Vuong, L.T., Ruckle, J.L., Blood, A.B., Reid, M.J., Wasnich, R.D., Synal, H., Dueker, S.R. (2008) Use of accelerator mass spectrometry to measure the pharmacokinetics and peripheral blood mononuclear cell concentrations of zidovudine. Journal of Pharmaceutical Sciences, 97(7), 2833–2843.

SECTION III CLINICAL ANALYSIS

13 MASS SPECTROMETRY IN CLINICAL ANALYSIS: SCREENING FOR INBORN ERRORS IN METABOLISM Donald H. Chace

13.1

INTRODUCTION

As of 2010, nearly all (>98%) infants in several countries (United States, Australia, New Zealand, Germany, and others) have been screened using tandem mass spectrometry (MS/MS) for at least one inherited metabolic disease. The disorder most commonly and uniquely screened using MS/MS is medium-chain acyl CoA dehydrogenase (MCAD) deiciency, a disorder of fatty acid metabolism. This genetic disorder has been shown to account for many previously unexplained deaths in unscreened newborns and infants. Inherited diseases, similar to MCAD deiciency, are also detected by MS/ MS with minimal additional analysis time or effort. That is why, for most newborn screening (NBS) laboratories, the metabolic panel consist of screening more than 50 metabolites to detect over 30 diseases. These disorders often include phenylketonuria (PKU), a disorder of amino acid metabolism, propionic academia (PA), a disorder of organic acid metabolism, and very long-chain acyl CoA dehydrogenase (VLCAD) deiciency in addition to other disorders of fatty acid, organic acid, and amino acid metabolism. From a clinical chemistry perspective, MS/MS is one of the most common mass spectrometry (MS)-based applications found in metabolic and public health labs. As noted earlier, several million infants are tested worldwide per year and this number continues to grow. This chapter will focus primarily on the practical appli-

cation of MS/MS in the analysis of the metabolites used to detect a wide array of metabolic diseases in NBS. This will include the history, early validation, sample handling and preparation, MS/MS operation, data generation, interpretation of results, and approaches to quality assurance and harmonization. Special topics integrated throughout the text will include the importance of stable isotope standards and one of the most unique aspects of newborn and population screening, the dried ilter paper blood spot (DBS). 13.1.1 The History and Background of Newborn and Metabolic Screening Using MS/MS NBS and the use of dried blood spot (DBS) began almost 50 years ago, when, in the early 1960s, Dr. Robert Guthrie developed a bacterial inhibition assay [1] for the measurement of phenylalanine (Phe). This biological assay utilized cultured bacteria that could only grow in the presence of Phe. This analysis was suficiently sensitive to measure elevated concentrations of Phe from the DBS of newborns in the irst 2–3 days of life. Although a somewhat imprecise assay, it was suficiently accurate to reliably detect PKU, an inherited disorder of Phe metabolism. Untreated PKU results in profound mental retardation and possible institutionalization. Early and continued treatment by dietary intervention throughout life prevents mental retardation and substantially improves the health of affected individuals.

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

273

274

MASS SPECTROMETRY IN CLINICAL ANALYSIS: SCREENING FOR INBORN ERRORS IN METABOLISM

NBS became a state-mandated public health measure to ensure all infants were screened in part due to the savings achieved by avoiding institutionalization of affected and untreated individuals. NBS expanded somewhat over the next two decades with a few additional assays for amino acid, steroid, carbohydrate, and congenital hormone deiciencies. MS did not have a role in NBS until the 1990s, when a DBS analysis of amino acids was developed [2]. Comparison studies of MS/MS with other NBS methods demonstrated its equivalence and versatility for detection of PKU. It was shown that it was more clinically sensitive and accurate in the detection of PKU than other methods such as luorometry [3]. A reduction of the false result rates without the addition of several new separate assays, that is, tyrosine (Tyr) analysis, demonstrated the advantages of using an MS/MS approach. To better understand metabolic disease, its detection, consider Figure 13.1, in which screening for PKU is illustrated. From the illustration, the advantages in using a multiple metabolite approach of MS/MS versus single analyte for luorometric assays should be clear. Clinical diagnostics in the study of metabolism are irmly rooted in gas chromatography–mass spectrometry (GC-MS) [4,5]. Today, as it has been since the 1980s, a diagnosis of metabolic disease should always utilize GC-MS analyses of urine organic acids. Disorders of fatty acid, organic acid, amino acids, nucleic acids, Krebs cycle intermediates, and many other small molecules produced by metabolic aberrations produced abnormal metabolites that are detected in urine using GC-MS. Much of the current knowledge in metabolism is based on clinical research and clinical practice in affected patients using this MS approach. Knowledge of fatty acid metabolism ultimately led to a better understanding of the role of fatty acid esters of L-carnitine (free carnitine [FC], unesteriied carnitine), commonly referred to as fatty acylcarnitines or simply acylcarnitines (ACs). These important compounds involved in fatty acid and organic acid metabolism are quaternary ammonium compounds and therefore have a permanently charged cation (positive ion). Analysis by GC-MS could be performed without extensive sample preparation and complex chemistry to make ACs volatile and stable for the gas chromatographic analysis. Clearly, liquid chromatography–mass spectrometry (LC-MS) might be a more appropriate method for analyzing this class of compounds. The structures of L-carnitine and other common ACs are presented in Figure 13.2. Millington, Roe, and colleagues explored new LC-MS ionization technology developed in the 1980s—fast atom bombardment (FAB) [6,7]—to detect carnitine and ACs in plasma and urine [8]. These and other early liquid chromatography (LC) ionization techniques such as ther-

Phenylalanine

Phenylalanine Hydroxylase

Dietary Protein

Normal Phe = 70* Tyr = 70 Phe / Tyr = 1 Cellular Protein

Tyrosine

Phenylalanine

Phenyylalanine Hydroxylase

Dietary Protein

Phenylketonuria Phe = 300* Tyr = 50 Phe / Tyr = 6 Cellular Protein

Tyrosine

Phenylalanine

IV Amino Acids

Phenylalanine Hydroxylase

IV Nutrition Phe = 240* Tyr = 200 Phe / Tyr = 1.2 Cellular Protein

Tyrosine FIGURE 13.1 Schematic illustrating the metabolism of phenylalanine (Phe) normal, PKU, and infants receiving IV amino acids (TPN, total parenteral nutrition). The major exogenous source of Phe and Tyr is from dietary protein (normal diet) or an enteral or parenteral mixture of amino acids (TPN). A secondary source is from the breakdown of cellular protein from tissues/muscles. In normal individuals, Phe is converted to Tyr by phenylalanine hydroxylase (PAH). Phe and Tyr are incorporated into proteins during an anabolic state and can be released back into the plasma pool following protein turnover by cells in a catabolic state. In infants with PKU, the activity of the enzyme PAH is 10% or less. Therefore, the conversion of Phe to Tyr is impaired, resulting in a signiicant elevation of Phe and a reduction in Tyr. Note, however, that Tyr is also obtained in the diet and the reduction in Tyr levels may be mild. In infants receiving high doses of protein (amino acids) via IV feeding (very low birthweight premature infants), a moderate to signiicant elevation of Phe and Tyr are often observed. Sometimes this is due to blood contamination of the TPN mixture being administered if the blood is collected improperly. Commonly encountered concentrations of Phe and Tyr (micromolar) are provided as reference. Note that multiple metabolite MS/MS detection provides all three data points, whereas a luorometric-based screen provides only Phe (marked with an asterisk). Note that a typical abnormal result for PKU is when both the concentration of Phe is greater than 180 µ mol/L and the Phe : Tyr ratio is greater than 2.5.

INTRODUCTION

>ͲĂƌŶŝƟŶĞ н ;,ϯͿϯE

13.1.2

K, KK,

ĐLJůĐĂƌŶŝƟŶĞƐ ^ŚŽƌƚŚĂŝŶ ŝ ;ĐĞƚLJů͕ϮͿ

н ;,ϯͿϯE

KƌŐĂŶŝĐĐŝĚ ŝ  ŝĚ ;/ƐŽǀĂůĞƌLJů͕ϱͿ

KK,ϯ KK,

н ;,ϯͿϯE

DĞĚŝƵŵŚĂŝŶ ;KĐƚĂŶŽLJů͕ϴͿ

н ;,ϯͿϯE

KK,Ϯ,;,ϯͿϮ KK,

sĞƌLJ>ŽŶŐŚĂŝŶ ;WĂůŵŝƚŽLJů͕ϭϲͿ

KK;,ϮͿϲ,ϯ KK,

н ;,ϯͿϯE

KK;,ϮͿϭϰ,ϯ KK,

hŶƐĂƚƵƌĂƚĞĚ ;KůĞŽLJů͕ϭϴ͗ϭͿ

н ;,ϯͿϯE

KK;,ϮͿϳ;,с,Ϳ;,ϮͿϳ;,Ϳϯ KK,

ŝĐĂƌďŽdžLJůŝĐ ĐŝĚ ;'ůƵƚĂƌŝĐ͕ϱͿ

н ;,ϯͿϯE

KK;,ϮͿϰKK, KK,

ϯͲ,LJĚƌŽdžLJĐŝĚƐ ;ϯͲ,LJĚƌŽdžLJƉĂůŵŝƚŽLJů͕ϭϲK,Ϳ

н ;,ϯͿϯE

KK;,ϮͿ;,K,Ϳ;,ϮͿϭϰ,ϯ KK,

FIGURE 13.2 Chemical structures of L-carnitine (levocarnitine) and several fatty ACs of different chain length, saturation, or functional groups.

mospray [9] led to a simpler method for detection of ACs by MS with liquid chromatographic separation. Near the turn of the decade, in 1990, MS/MS was used to detect these ACs without the need for chromatography [7,10,11]. The analysis was suficiently sensitive and selective to provide clinical data that demonstrated the importance of AC as biomarkers of beta oxidation and mitochondrial metabolism. The analysis using fast atom bombardment tandem mass spectrometry (FAB MS/MS) became a diagnostic test in plasma for diseases of beta oxidation such as MCAD, VLCAD, and long-chain 3-hydroxyacylcoenzyme A dehydrogenase (LCHAD) deiciencies, as well as organic acidemias such as propionic, isovaleric, and glutaric acidemias. An alternative method for detection of diseases of fat metabolism using different biomarkers—urinary acylglycines—by LC-MS was also developed [12]. Together, these two techniques led to a rapid understanding of fatty acid and organic acid metabolism disorders. Clinical studies strongly suggested that early infant screening could result in a decrease in morbidity and mortality of the affected infants. The potential for a routine NBS test was recognized [11].

275

Early Developments in NBS by MS/MS

The optimization of MS/MS for amino acids in DBS was the irst step in the development and validation of NBS [2]. The reason that an amino acids newborn screen was developed irst rather than an AC screen was the requirement for validation of using a DBS for analysis and quantiication by MS/MS. By choosing a compound that is “MS friendly” (i.e., positive ion, small molecule, previous MS analyses) and is also a key metabolite in NBS, it was more likely that validation of the assay would be achieved and that validation accepted. Phe is that “MS-friendly” metabolite used in the detection of PKU and screened for by all 50 U.S. states. It was likely that development of the Phe assay using MS/MS would be similar to that for ACs, since chemically speaking they are not that different in terms of molecular weight and the fact that they both have basic and acid groups. Sample extraction and preparation are also likely to be similar. So the initial attempts at an assay could mirror that used for plasma ACs. In the initial MS/MS experiments, during a full-scan analysis of a DBS, multiple peaks in addition to Phe and its stable isotope internal standard (i.s.) were present. The atomic mass unit (amu) differences were 14 and 16 daltons (Da), which suggested methyl and hydroxyl functional groups were present in these compounds. To the mass spectrometrists, this is an interesting inding because these are mass differences that are often observed for molecules that have additional methyl and hydroxyl groups. By analyzing a premixed set of reference standards of amino acids, many of these same peaks were present. Conirmation was made by investigating the mass spectra of individual alpha amino acids. Using MS/ MS, a common fragment molecule of alpha amino acids was discovered (see below). This common neutral loss (NL) of 102 Da for the butyl ester led to the development of an amino acid panel from DBS extracts using MS/MS. The impact of multiple analytes for amino acids in NBS cannot be overestimated. By measuring both Phe and Tyr, the detection of PKU was improved without the need for separate analyses. Comparison to other methods demonstrated that MS/MS improved selectivity and sensitivity (both analytically and clinically) by not only measuring two single metabolites simultaneously, Phe and Tyr, but by the ability to express the concentration of these two metabolites as a molar ratio, Phe : Tyr. As shown in a subsequent publication, Phe : Tyr could reduce false results and improve clinical speciicity [3]. Furthermore, calculation of ratios of Phe to other amino acids such as leucine (Leu) enabled differentiation of a generalized elevation of amino acids due to intravenous (IV) nutrition (TPN, total parenteral nutrition) from PKU [13].

276

13.1.3

MASS SPECTROMETRY IN CLINICAL ANALYSIS: SCREENING FOR INBORN ERRORS IN METABOLISM

Consolidation of Assays for NBS by MS/MS

The clinical diagnostic assay for plasma ACs utilized methyl ester derivatization. However, methyl ester derivatization could not be used in sample preparation for Phe since its use results in a compound which interferes with the analysis of Phe. Butyl ester derivatives of AC produced similar results as their respective methyl esters derivatives, albeit with a mass shift due to molecular weight differences. In addition, both AC and amino acids were eficiently extracted by methanol from DBS. Comparison of methyl ester analysis of plasma samples with blood from the same patient helped validate the AC portion of the newborn screen. The ability to analyze both ACs and amino acids in a single assay was a powerful test for many reasons. The analysis could be used to replace existing NBS assays for Phe, Tyr, Leu, and methionine (Met), and add a new series of clinical screening for fatty acid and organic acid metabolism by AC analysis. Clinical validation still required the actual analysis for more than 45 disorders in a large population of newborns [14]. It would take many years to achieve more than a half-million analyses either from statesponsored pilot projects in North Carolina or its use by a private screening lab in western Pennsylvania to achieve the appropriate number of newborns to detect diseases for which few statistics were available due to lack of screening and predicted incidences of 1 in 100,000 to 1 in 200,000 for some very rare disorders. It was not until 2000, 10 years after the irst validation studies, that states began to mandate the additional disorders detected by MS/MS as part of their routine NBS panel. This discussion of the history of NBS by MS/MS is important as it may serve as one model for the development of future screening and clinical assays. A lowchart of key events in its validation (Figure 13.3) may be helpful as a quick reference to the development of new screening assays. With the background, the remainder of the chapter will emphasize the assay details and will focus on key analytical features, including result interpretation and maintaining good quality analysis from the irst sample of the day to the last.

13.1.4

Dried Blood Specimens

From a purely clinical or analytical chemistry perspective, the most unique aspect of NBS is the DBS. This special purpose ilter paper was originally developed by Dr. Robert Guthrie to collect blood from the heel of an infant. It was developed as part of a bacterial inhibition screening assay for PKU [15]. The DBS is also known as the Guthrie card or the PKU card. It is prepared from

100% pure cotton ibers and has very high speciications as to the quantity of blood absorbed per unit area of paper given a deined hematocrit and volume of blood applied to the card. Manufacturers of this ilter paper must meet the speciications provided in the CLSI (Clinical Laboratory and Standards Institute) document LA4-A5 [16]. Figure 13.4 shows a typical card used in NBS. It contains a strip of ilter paper for blood collection marked with circular dashes to target the blood application. The strip of paper is attached to a form for patient/newborn information. Blood is applied to the paper as a free-lowing droplet from the heel of an infant’s foot, which ensures the most consistent results in terms of blood volume. Clinical diagnostic analysis of liquid specimens (plasma, urine) requires accurate volume measurement. For DBS, the volume of blood, at this time, is estimated. The quantity of dried blood on the ilter paper is based on the absorptive characteristics of the paper, the method of collection, the volume of blood and method of application to the paper, and the characteristics of the blood such as hematocrit. Samples are obtained for analysis by automated or hand punch usually with a diameter of 1/8 or 3/6 inch. The volume of blood in these punches is calculated based on the CLSI standard. Features of the DBS and ilter paper are shown in Figure 13.4. It is important to recognize that the volume of blood present on the ilter paper is based on a standard but that in the actual practice of NBS, the volume of blood may be highly variable. A high hematocrit will make blood slightly more viscous, will spread less per unit volume, and cause the actual blood concentration to be higher than the CLSI standard. Lower hematocrit produces greater spreading of blood and a lower blood volume per unit area. Aggressive squeezing of blood from the heel may cause mixing with interstitial luid, reducing hematocrit or produce serum rings. Specimens collected from IV lines or syringes/test tubes containing anticoagulants such as ethylene diamine tetra-acetic acid (EDTA) can contaminate a specimen. Capillary tubes could also result in altered estimated blood volumes if used improperly such as scratching the paper during application or applying multiple small “dots” of blood versus a illed dashed circle. Although there is considerable variation from specimen to specimen, careful collection has been shown to produce suficiently reliable results for NBS purposes, as demonstrated by its use for nearly 50 years. From a practical viewpoint, the theoretical volume of blood per square inch is used in concentration calculations until a method for estimating actual blood volume in each individual spot is developed. Based on a 55% hematocrit

INTRODUCTION

277

sĂůŝĚĂƟŽŶŽĨD^ͬD^ĨŽƌƌŝĞĚůŽŽĚ ^ƉŽƚƐ ŶĂůLJƐŝƐ

EĞǁŝŽŵĂƌŬĞƌŝƐĐŽǀĞƌLJ  ĐLJůĐĂƌŶŝƟŶĞƐĂŶĚĨĂƩLJ ĂĐŝĚ ŽdžŝĚĂƟŽŶ ĚŝƐŽƌĚĞƌƐ

ϭϵϴϱ

ϭϵϴϬ ƐƐĂLJĞǀĞůŽƉŵĞŶƚ ĞǀĞůŽƉŵĞŶƚŽĨĂŶD^ͲďĂƐĞĚĂƐƐĂLJƚŽŵĞĂƐƵƌĞďŝŽŵĂƌŬĞƌƐŝŶ ƉůĂƐŵĂŽĨĂīĞĐƚĞĚƉĂƟĞŶƚƐ

ůŝŶŝĐĂůZĞůĞǀĂŶĐĞŽĨŝŽŵĂƌŬĞƌƐ ďŶŽƌŵĂůĐŽŶĐĞŶƚƌĂƟŽŶƐŽĨĂĐLJůĐĂƌŶŝƟŶĞƐĨŽƵŶĚŝŶĂīĞĐƚĞĚ ƉĂƟĞŶƚƐǁŝƚŚŵĞƚĂďŽůŝĐĚŝƐĞĂƐĞƐƵĐŚĂƐDĚĞĮĐŝĞŶĐLJ

ŝĂŐŶŽƐƟĐƐƐĂLJZĞĮŶĞŵĞŶƚ ůŝŶŝĐĂůĂŶĂůLJƐŝƐŽŶĂƌĞƐĞĂƌĐŚďĂƐŝƐǁŝƚŚŽƚŚĞƌǀĂůŝĚĂƚĞĚƚĞƐƚ ;ŽƌŐĂŶŝĐĂĐŝĚƐďLJ'ͲD^Ϳ

ŝĂŐŶŽƐƟĐƐƐĂLJsĂůŝĚŝƚLJ ZĞƐƵůƚƐĨƌŽŵĐůŝŶŝĐĂůƚĞƐƟŶŐƉƌŽĚƵĐĞďĞŶĞĮƚƚŽĂīĞĐƚĞĚ ƉĂƟĞŶƚƐ

WƌĞǀĞŶƟŽŶ ůŝŶŝĐĂůƌĞƐĞĂƌĐŚƐƚƵĚŝĞƐƐŚŽǁĞĂƌůLJĚŝĂŐŶŽƐŝƐƉƌĞǀĞŶƚƐĚŝƐĞĂƐĞ ĂĚǀĞƌƐĞĞīĞĐƚƐ͘

^ĐƌĞĞŶŝŶŐƐƐĂLJŽŶƐŝĚĞƌĂƟŽŶƐ ƌŝĞĚďůŽŽĚƐƉŽƚĂŶĚŶĞǁďŽƌŶƐĐƌĞĞŶŝŶŐůĂďƐŵŽƐƚƐƵŝƚĞĚĨŽƌ ƉŽƉƵůĂƟŽŶƚĞƐƟŶŐ ƉŽƉƵůĂƟŽŶ

ϭϵϵϬ džŝƐƟŶŐŝŽŵĂƌŬĞƌƐĂŶĚ^ĐƌĞĞŶŝŶŐ WŚĞŶLJůĂůĂŶŝŶĞďŝŽŵĂƌŬĞƌŝŶĂĚƌŝĞĚďůŽŽĚƐƉŽƚŝƐĂŵĂƌŬĞƌĨŽƌ ƉŚĞŶLJůŬĞƚŽŶƵƌŝĂ

ƐƐĂLJĞǀĞůŽƉŵĞŶƚĂŶĚŽŵƉĂƌŝƐŽŶ hƐĞD^ͬD^ƚŽĚĞƚĞĐƚƉŚĞŶLJůĂůĂŶŝŶĞŝŶĂĚƌŝĞĚďůŽŽĚƐƉŽƚ

ůŝŶŝĐĂůŽŵƉĂƌŝƐŽŶĂŶĚƐƐĂLJsĂůŝĚĂƟŽŶ ŽŵƉĂƌĞƐĐƌĞĞŶŝŶŐƵƐŝŶŐĞdžŝƐƟŶŐŵĞƚŚŽĚŝŶĂŶE^ůĂďǁŝƚŚ D^ͬD^ĂŶĂůLJƐŝƐ

ŽŶƐŽůŝĚĂƟŽŶŽĨdĞƐƚƐ ĚĚŶĞǁĂĐLJůĐĂƌŶŝƟŶĞĂŶĂůLJƐŝƐƚŽƚŚĞǀĂůŝĚĂƚĞĚĂŵŝŶŽĂĐŝĚ ĂŶĂůLJƐŝƐ͘ĞŵŽŶƐƚƌĂƚĞŶĞǁŵĞƚĂďŽůŝƚĞƐĐĂŶďĞĐŽƐĐƌĞĞŶĞĚ͘

ϭϵϵϱ dĞƐƚdžƉĂŶƐŝŽŶ YƵĂůŝƚLJƐƐƵƌĂŶĐĞ͕,ĂƌŵŽŶŝnjĂƟŽŶ͕^ƵƉƉŽƌƚ ^ŝŵƉůŝĨLJĂŶĂůLJƐŝƐĂŶĚƉƌŽǀŝĚĞƐƵƉƉŽƌƚĂŶĚŝŵƉƌŽǀĞŵĞŶƚ

Wŝů ^ WŝůŽƚ^ĐƌĞĞŶŝŶŐ^ƚƵĚŝĞƐ ŝ ^ Ěŝ ZĞƐĞĂƌĐŚƐĐƌĞĞŶŝŶŐƚŽĂƐƐĂLJƐĞŶƐŝƟǀŝƚLJ͕ƐĞůĞĐƟǀŝƚLJ͕ĂŶĚ ĂĐĐƵƌĂĐLJŽĨƐĐƌĞĞŶŝŶŐƵƐŝŶŐD^ͬD^

ϮϬϬϬ EĞǁdĞƐƚƐ͕ dĞƐƚƐ͕ EĞǁůŝŶŝĐĂůƉƉůŝĐĂƟŽŶƐ ůŝŶŝĐĂů ƉƉůŝĐĂƟŽŶƐ

ϮϬϭϬ

ĚĚŶĞǁƚĞƐƚƐĨŽůůŽǁŝŶŐŇŽǁĐŚĂƌƚĚĞƐĐƌŝďĞĚĂďŽǀĞŽƌƵƐĞĚ ĐƵƌƌĞŶƚƐĐƌĞĞŶŝŶŐĨŽƌŶĞǁĐůŝŶŝĐĂůĂƉƉůŝĐĂƟŽŶƐ

dǁĞŶƚLJͲ&ŝǀĞzĞĂƌƐŽĨD^ͬD^ŶĂůLJƐŝƐŝŶDĞƚĂďŽůŝĐŝƐĞĂƐĞ FIGURE 13.3 History of MS/MS in NBS.

and a nearly full circle as marked by the dashed zones, a blood of volume of 7.6 µL per 3/16-inch punched dot is often used. It is noteworthy that plasma and urine can also be collected on the same ilter paper. In a related application using MS, plasma from dialysis patients is analyzed [17]. Application of plasma or urine is more reproducible because the issue of variable hematocrit is eliminated. However, plasma is somewhat more viscous than

urine and there are small variations in the plasma of normal patients and those recently undergoing dialysis. DBSs are marginally safer than liquid specimens in terms of handling and infectious disease, take up less storage space, and can be sent via regular mail to the clinical laboratory, saving signiicant cost of screening. For these reasons, the collection of dried specimens (blood, urine, plasma, tissue extracts) is being used in applications beyond NBS [17].

278

899901

SUBMITTER KEEP TOP YELLOW COPY FOR YOUR RECORDS

HOSPITAL OF BIRTH M BIRTHDATE BABYS LAST NAME

FIRST

DRAW DATE TIME

AM CHECX HERE IF BABY IS LESS PM THAN 24 HRS. OLD MOTHERS LAST NAME FIRST

SEX F GESTATION (WEEKS) TRANSPUSED?

L 8422008

ADDRESS DATE _______

DRAWN BY

BIRTHTIME BABY’S UED. PEC. NO. AM

PM BIRTHWEIGHT BIRTH … SINGLE (GRAMS)

… SM VOL … EXCHANGE

… OTHER _______

…A …B …C SPECIMEN IF REPEAT … INITIAL

… REPEAT

BABY’S RACE

… WHITE … BLACK

… AS … AMER INDLAN … OTHER

HISPANIC? … YES … NO

… RECUESTED

… ROUTINE

(PREVIOUS CARD #)

CITY STATE. ZIP

BABY’S PHYSICLAN

SUBMITTER

PHONE (MOTHER)

PHONE (PHYSICLAN)

ADDRESS IF OTHER THAN BIRTH FACILITY

903TM LOT # W-961 (Rov. 9/96)

A D D R E SA SR OE GA R A P H

899901

MASS SPECTROMETRY IN CLINICAL ANALYSIS: SCREENING FOR INBORN ERRORS IN METABOLISM

FIGURE 13.4 A typical ilter paper card and attached form. The area marked with four dashed circles is the ilter paper made from 100% cotton. The larger form to the left contains newborn information. Below this illustration is a schematic of the process of applying a droplet of blood to the card and punching the spots into a well of a microtiter plate for further processing.

13.1.5

Metabolite Extraction

The primary metabolites of interest in the MS/MSbased NBS panel are α-amino acids, carnitine, and ACs. Figure 13.2 provides generic chemical structures of many ACs while Figure 13.5 provides structures of key amino acids measured in the NBS panel using MS/MS. Chemically and structurally speaking, amino acids and ACs are quite similar in that they both have basic nitrogen and carboxylic acid functional groups and a side chain that contains various chemical groups. In the case of amino acids, the alpha carbon is attached to the basic amine, carboxylic acid, and the “R” group. The “R”

group for amino acids is quite variable such that they can be aliphatic, aromatic, basic, acidic, and so on. ACs also have an “R” group attached to an aliphatic alcohol. These R groups are fatty acids and organic acid acyl esters. The fatty acids range from the two-carbon acetyl to the 18-carbon octadecanoylcarnitine. These fatty acids may be unsaturated such as tetradecenoylcarnitine, contain an alcohol such as 3-OH isovaleric acid, or contain a second carboxylic acid group such as glutaric acid (Figure 13.2). Unesteriied carnitine (FC) has a free hydroxyl group where the fatty acids normally form esters. Table 13.1 contains a list of the most common amino acids and ACs detected in NBS, their common abbreviations, and the mass-to-charge ratio (m/z) values

INTRODUCTION ůƉŚĂĐĂƌďŽŶ

ůƉŚĂŵŝŶŽĐŝĚƐ

E,Ϯ

,



Z

KK, E,Ϯ ,

E,Ϯ



,

,Ϯ



KK, WŚĞŶLJůĂůĂŶŝŶĞ E,Ϯ ,



KK,

E,Ϯ ,ϯ

KK,

,



E,Ϯ ,

,

,

,Ϯ

,ϯ

KK,

sĂůŝŶĞ

>ĞƵĐŝŶĞ

;,ϮͿϮ ^

,ϯ

,



;,ϮͿϯ

KE,Ϯ

E,

KK,

DĞƚŚŝŽŶŝŶĞ

ŝƚƌƵůůŝŶĞ

E,Ϯ



,ϯ

E,Ϯ

KK,

,



,ϯ

E,Ϯ



K,

dLJƌŽƐŝŶĞ

,ϯ

KK,

ůĂŶŝŶĞ

,

,Ϯ

E,Ϯ ;,ϮͿϮ

KK,

KK,

,

'ůƵƚĂŵŝĐĐŝĚ



E, ;,ϮͿϯ

KK,



E,

E,Ϯ

ƌŐŝŶŝŶĞ

FIGURE 13.5 Chemical structures of selected α-amino acids (common part) of the newborn screen by MS/MS.

TABLE 13.1

279

for both derivatized (butyl esters) and underivatized (free acid) ACs. Both families of compounds are highly soluble in methanol. Methanol is an appropriate solvent because it can extract both hydrophobic and hydrophilic organic compounds but not inorganic salts and proteins (which it likely denatures). From a mass spectrometric perspective, the absence of inorganic salts in the extraction solvent reduces the likelihood of subsequent adduct formation with sodium and potassium during the ionization process of MS/MS. Microtiter plates are the preferred vessel for extraction of ACs and amino acids in the DBS. The shape and design of the plate is important for optimal extraction and removal of extracted metabolites as well as reduced interaction of extracted metabolites with the plastic wells. A standard polypropylene 96-well microtiter plates with a lat bottom coniguration has been used most successfully in NBS laboratories. This well accommodates DBS punches up to 1/4 inch diameter and volumes of 300–400 µL. Many laboratories utilized an automated liquid handling system to provide more eficient batch processing. The lat bottom wells work best in this case because DBSs lay lat on the bottom of the well and do not to contact the needle used in dispensation of methanol or the transfer of solvent extract to another well. Choice of the plate design is based on the blood spot diameter, extraction volume, and the liquid handling system utilized.

Most Common ACs, Abbreviations, and Pre Ions m/z Values

AC L-Carnitine (FC) Acetylcarnitine Propionylcarnitine Butyrylcarnitine Isovalerylcarnitine 3-OH Butyrylcarnitine Hexanoylcarnitine 3-OH Isovalerylcarnitine Octanoylcarnitine Malonylcarnitine Decenoylcarnitine Decanoylcarnitine Glutarylcarnitine Dodecanoylcarnitine Tetradecenoylcarnitine Tetradecanoylcarnitine Hexadecanoylcarnitine 3-OH Hexadecanoylcarnitine Octadecenoylcarnitine Octadecanoylcarnitine

Abbreviation

Butyl Ester (m/z)

Unesteriied (m/z)

FC, C0 C2 C3 C4 C5 C4OH C6 C5OH C8 C3DC C10:1 C10 C5DC C12 C14:1 C14 C16 C16OH C18:1 C18

218 260 274 288 302 304 316 318 344 360 370 372 388 400 426 428 456 472 480 482

162 204 218 232 246 248 260 262 288 248 316 318 276 344 370 372 400 416 424 426

280

MASS SPECTROMETRY IN CLINICAL ANALYSIS: SCREENING FOR INBORN ERRORS IN METABOLISM

13.1.6 Pseudo Isotope Dilution and Internal Standards Quantiication of metabolites extracted from a DBS requires stable isotope references standards. These standards are comprised of isotope-labeled ACs and amino acids. Many standards may be obtained individually, but most laboratories used either a premixed set of standards or as part of a kit containing reagents, calibrators, and so on. ACs are labeled with deuterium at one or more of the methyl groups attached to the quaternary ammonium nitrogen. This is one reason why AC standards are either 2H3, 2H6, or 2H9 enriched. The lexibility in labeling enables choosing the standard that occurs at m/z values that are free from interferences and other ACs. In addition, the position of the isotope standard is chosen such that it can be better quantiied after the fragmentation process (see discussion on MS/MS below). Amino acids are primarily deuterium (2H) or 13 C enriched although two amino acid standards contain 15 N. Stable labels are typically found on the R group, the alpha carbon or amino group. 13C is used instead of deuterium at potentially exchangeable positions such as aromatic hydrogen. Isotope dilution MS is based on the principle of the addition of stable isotope internal standard to a liquid specimen. The standard is chemically most like the unlabeled substrate such that it is goes through the same processes of sample preparation or modiication. The only real differences between the standard and unlabeled substrate is mass. MS uniquely exploits the mass differences and is one of the reasons why it is considered the gold standard in biomarker quantiication. Figure 13.6 illustrates this point for some of the most common internal standards used in NBS. It is noteworthy that inappropriate use of enrichments can alter the chemical and physical properties of the labeled internal standard. For example, excessive deuterium enrichment or the use of a deuterium at an exchangeable reactive position can result in poor quantiication or loss of enrichment. With DBSs, the internal standard is present in the extraction solvent and can only equilibrate with the endogenous metabolites that are extracted from the DBS. Therefore, isotope dilution principles do not apply to the extraction step since they are not in the blood (blood spot) to begin with. However, subsequent steps such as derivatization, sample transfer, and so on do indeed follow the principles of isotope dilution. To note the differences between analyzing liquid versus DBSs using stable isotope standards and the fact that quantiication is less accurate using DBSs, the term “pseudoisotope dilution” is often used.

K,

н ;,ϯͿϯE

KK,

Ϯ

,ϵ Ͳ>ͲĐĂƌŶŝƟŶĞ 'D

>ͲĐĂƌŶŝƟŶĞ

,ϯ н ϯ E

KK;,ϮͿϲ,ϯ KK,

н ;,ϯͿϯE

K,

н ;ϯͿϯE

KK,

KK;,ϮͿϲ,ϯ KK,

,ϯ

KĐƚĂŶŽLJůĐĂƌŶŝƟŶĞ

Ϯ

,ϯ ͲŽĐƚĂŶŽLJůĐĂƌŶŝƟŶĞ

'D E,Ϯ

E,Ϯ ,



,Ϯ

сϭϮ

,



сϭϯ

KK,

KK,

WŚĞŶLJůĂůĂŶŝŶĞ

,Ϯ

ϭϯ

ϲ ͲWŚĞŶLJůĂůĂŶŝŶĞ 'D

E,Ϯ ,



E,Ϯ ,ϯ





KK,

KK,

ůĂŶŝŶĞ

ϯ

Ϯ

,ϰ ͲůĂŶŝŶĞ

'D

FIGURE 13.6 Common stable isotope-labeled internal standard used in NBS. Both unlabeled and labeled compounds are shown. Positions of the stable isotopes are shown at either D for deuterium or as bullets where 13C is substituted for 12C. Mass shifts are also listed.

The concentrations of the internal standards are prepared such that the key metabolites are at approximately the same concentration as the median endogenous metabolite concentrations, at the concentration of the decision point for normal or abnormal, or at the concentrations of similar metabolites. For example in one set of AC standards, the relative concentration of the AC internal standards FC : C2 : C3 : C4 : C5 : C8 : C14 : C16 is 20: 5:1:1:1:1:1:2. There are many ideal approaches for setting the concentrations of internal standards. The reason for the choice above was that for many ACs, the endogenous concentrations are very low at or below the signal-to-noise ratio for detection. The internal standards were set to be close to the cutoff for abnormal/ normal and at a similar concentration to other standards. Further, keeping these relative concentrations is helpful in the interpretation and harmonization of the assay as they have been used for many years and in several publications. Amino acids are at concentrations more closely similar to endogenous levels rather than the same concentration of other amino acid standards.

FURTHER PREPARATION FOR ANALYSIS

13.2

FURTHER PREPARATION FOR ANALYSIS

In the early years of the MS/MS analysis of ACs, liquid secondary ionization (LSI), also known as either FAB or fast ion bombardment (FIB), was used. This ionization technique required esteriication of the carboxylic acid groups in order to make the AC suficiently lipid soluble so that it could rise to the surface of a glycerol/ methanol matrix where it was exposed to the ionization beam. Unesteriied (nonderivatized) samples were not suficiently ionized by FAB or FIB to be detectable by MS/MS. For the clinical analysis used in the late 1980s through the mid-1990s, methyl esteriication was utilized (the NBS methyl esters could not be used due to interferences for Phe measurement and dificulty analyzing glutarylcarnitine, which was 1 Da between the mass of octanoylcarnitine and its internal standard). By utilizing a different ester, speciically butyl ester, these problems were removed due to the mass shift produced by the addition of two butyl esters. Derivatization also produced a more lipophilic compound which enhanced ionization of formerly more acidic dicarboxylic acid ACs and amino acids. With the introduction of electrospray ionization (ESI), ionization eficiency of underivatized samples was improved dramatically. Some labs have adopted this approach to analysis. A discussion of the pros and cons of derivatization is described in a subsequent section. However, for the purposes of this chapter, unless otherwise stated, butyl ester derivatives were utilized. Following an incubation period of 30 min, the methanol extract (of metabolites from the blood spots) is transferred to another microtiter plate. The remaining blood spot is either discarded or utilized for another analysis such as succinylacetone [18,19] for detection of tyrosinemia type I. Due to the highly acidic derivatization reagents that may digest the ilter paper, the solvent extract must be transferred to another well. Approximately 90% of the solvent is transferred using automated liquid handling but this can vary from 5% to 15% based on the liquid handling device, whether a pipette is utilized, and the type of microtiter well design (round, conical, or lat bottom) [20]. Following transfer of the extract, methanol is removed under a stream of nitrogen and gentle heating. It is strongly recommended that air not be used in the process since the oxygen and water may cause metabolite oxidation and hydrolysis. The butyl esteriication reagent that is added to the dried metabolites extracted from the blood spot is comprised of 3 N HCl in n-butanol. Due to the signiicant water content of concentrated HCl, this reagent cannot be prepared in a laboratory by simply mixing concen-

281

trated hydrochloric acid and n-butanol together. The 3 N butanolic HCl is prepared by passing hydrochloric acid gas in n-butanol to saturation which results in a 3 N concentration. Other reagents prepared from acetyl chloride are also available, although this reagent is less volatile than butanolic HCl and may take longer drying times for its removal. It may be noteworthy that one advantage of using the highly acidic reagent is that any residual ibers or particles transferred from the original well containing the ilter paper blood spot are likely degraded further. This may prevent excessive clogging in the LC injector due to a reduction in the size of foreign particles. Some labs centrifuge plates to accomplish this same result or use inline ilters. Plates are covered with a nonreactive material (usually the original plastic covers) to prevent loss of the derivatization reagent and contamination of the oven used in heating the wells. Most commonly, the process of derivatization occurs at 65°C for 15 min with variations of 5°C and 5 min in some labs, that is, 60°C for 20 min. It is critical that the reaction be stopped as quickly as possible following the incubation period by immediately drying under heated nitrogen to remove the derivatization reagent, which is quite volatile. The plate is ready for MS/MS analysis and is quite stable in the dried state for several hours. It is recommended that the inal reconstitution solvent, that is, that solvent which will be injected into the mass spectrometer, not be added until just before MS/MS analysis. For those labs that do not derivatize, these procedures and solvent compositions may be different. Direct sampling of the methanol or other solvent extract of the DBS can be utilized or the procedure followed above with the exception of addition of the derivatization reagents (sample is transferred to a new plate and dried). The solvent used for reconstitution is often the same as the mobile phase used in the liquid chromatography– tandem mass spectrometry (LC-MS/MS) analysis. Either Telon seals (mats) or aluminum foil has been used to cover plates to prevent evaporation of solvent prior to injection of sample. A schematic of the entire sample preparation process is summarized in Figure 13.7. 13.2.1

MS/MS Analysis: Sample Introduction

The earliest applications of MS/MS that utilized LSI methods (FAB or FIB) required placing a few microliters of a viscous glycerol/methanol matrix on a stainless steel probe, insertion into the fast-atom bombardment/ fast-ion bombardment (FAB/FIB) source housing, removal of air from the source, then further insertion of the probe to expose the sample and matrix to the atom or ion beam. Although this procedure was tedious, a

282

MASS SPECTROMETRY IN CLINICAL ANALYSIS: SCREENING FOR INBORN ERRORS IN METABOLISM

'%6 GLVFDUGHG RUUHXVHG



96-Well Plate

^ƚŽĐŬ /͘^͘

3XQFK

$GG0H2+ LV

6ROYHQW7UDQVIHU

ϯE,ů EͲƵ

1 'U\

$GG 0RELOH3KDVH

Ě njĞ ĂƟ ƌŝǀ ĚĞ Ŷ ;ŶŽ

Ϳ

'U

;ŶŽŶĚĞƌŝǀĂƟnjĞĚ

$GG0H2+ LV

E ,ϮϬ

1HZ3ODWH

E ,ϮϬ

D^ͬD^ FIGURE 13.7 Schematic of sample preparation showing the steps for preparing derivatized and underivatized samples.

sample could be analyzed every 3 min. Approximately 100,000 samples were analyzed in this manner in the early to mid-1990s. A low injection method of this ionization method (low FAB, FIB) with automated injection was investigated [21] but was impractical because sample carryover between injections was signiicant due to the viscous glycerol component of the mobile phase. In 1994, ESI was utilized which dramatically improved low injection proiles due to the fact that viscous solvents were not utilized and the spray itself was ionized eficiently. Details of the ESI methods used in this application are in a subsequent section. Microtiter plates (96 well conigurations) are used in autoinjectors of various designs and capacities and may have a lat, round, or conical bottom. The reconstitution solvent is approximately 50–100 µL, from which 10– 20 µL is injected depending on the laboratory. The injector loops may be ixed to the volume of injection. Alternatively, the loop may be much larger and the volume of sample injected depends on the precise volume delivered by the injector syringe into the loop.

However, the accuracy of the loop injection is not critical to quantiication because of the stable isotope internal standards. The important component to the system do relate to luidics and the quality of tubing and connections, low rates, and ilters which may affect the quality of the solvent low as observed in the low injection proile. One approach to setting up the low injection is to irst consider source design. Using a Sciex API 3000 (AB Sciex, Foster City, CA), it was possible to thread deactivated fused silica from the injector port through the source component and out to the end of the electrospray probe. In this way, the sample did not contact any metal surface and the capillary tip could be adjusted (distance from the end of the steel probe by 1 or 2 mm and trimmed if contaminated or damaged). This coniguration eliminated dead volume and thereby minimized peak tailing at low low rates (15–20 µL/min). A suficient length was cut to provide suficient back pressure for the LC pump to operate properly. This setup prevents clogs and air bumbles in the transfer line and

FURTHER PREPARATION FOR ANALYSIS

283

3.86e6 cps

100 90

Ion counts per second

80

Total Ion Intensity (%)

1 scan period 70 60

End of acquisition 1.4 minutes

50 40

6 full scans

30 20 10

Injector recycle time

Baseline 1.0

0.5

1.5

Time (minutes)

Injection/Acquisition Start

1 scan period = 5 scan functions = 2 full scans, 3 SRM scans

FIGURE 13.8 Flow injection proile from an MS/MS acquisition of AC and amino acids. The vertical axis represents ion intensity and the horizontal axis is time in minutes. Sample injection started at 0 min and sample eluted between 0.4 and 1.3 min as shown by the shaded box. The total number of scan periods was six in this region. A scan period is deined as the period of the total number of scan functions in one complete scan (i.e., 1 Pre ion, 1 NL, multiple SRMs). The MS/MS acquisition stopped at 1.4 min but the injector was allowed to recycle for the next 0.6 min (needle rinse, obtain next sample etc.). Each injection is separated by 2 min.

source region and tended to occur at the capillary injector port interface. However, this might occur for one or two samples rather than an entire run since the air bubble was often lushed out after an injection. This setup is not always possible with different source designs. Another approach commonly used is polyetheretherketone (PEEK) tubing that connects the injector to the source inlet. To eliminate dead volume or retention on PEEK tubing, higher low rates could be utilized at the beginning of the injection and at the end of analysis to get the sample to the injector more quickly and to remove the remainder of the sample, preventing peak tailing and carryover. During the elution of a peak, the LC low rates are reduced. For example, an injection can be made using 20 µL/min during the elution of a peak and then followed by 50 µL/min to lush the sample. Ultimately the desire is to have an elution proile where 80% of the peak occurs in approximately 30 s and is comprised of ive or more data points. An example of an elution proile from a Sciex API 3000 instrument using the capillary tubing coniguration is shown in Figure 13.8. With a between sample injection time of 2 min, a microtiter plate containing 96 samples should be completed in a little more than 3 h (192 min).

13.2.2

Ionization

Currently, ESI is the primary ionization technique for NBS of ACs and amino acids from dried blood specimens using LC-MS/MS. Ion sources are designed to enhance the formation of charged species in nebulized droplets from the lowing stream of mobile phase and sample. Positive ion generation is the desired result for ACs and amino acids. Details of electrospray and ionization can be found elsewhere. However, approaches to enhancing ionization of ACs and amino acids are described below. ACs and amino acids compete for ionization with hundreds of organic compounds extracted from the blood spot. To achieve the maximum sensitivity, it is important to enhance the ionization of AC and amino acids while minimizing it for other compounds. Ionization eficiency is related to detection eficiency, which is ultimately represented as analytical sensitivity. A high sensitivity is needed for some ACs present at very low concentrations. One approach to enhancing ionization eficiency is to obtain as much of the metabolite as possible for ionization. But this is not necessarily achieved by analyzing larger blood spots because with additional ACs or amino

284

MASS SPECTROMETRY IN CLINICAL ANALYSIS: SCREENING FOR INBORN ERRORS IN METABOLISM

acids comes additional interfering compounds. What is desired is to obtain more ACs and amino acids as compared to other components in blood. The DBS can be used in a similar manner as a chromatographic cartridge or column extraction. In other words, the DBS can selectively retain compounds that may not be of interest while allowing those desired compounds (ACs) to be extracted. Proteins, inorganic salts, and very ionic or hydrophilic compounds are not removed when using methanol as the extraction solvent. ACs and amino acids are very soluble in methanol and extracted easily. If other compounds that are left behind are desired, then a secondary extraction may be performed as demonstrated for the succinylacetone assay described at the end of the chapter. Another approach is to tailor the chemistry to the positive mode of ionization. ACs and amino acids both contain basic amine (positive) and carboxylic acid (negative) groups and both charges are present in aqueous solution. In addition, the R groups of amino acids and fatty acid conjugates of ACs may have carboxylic acids, amines, and other polar groups. Depending on the dissociation constant, the distribution of negatively and positively charged ions will vary by pH. In solution, the distribution of charge will be more negative. ACs and amino acids exist in many different charged states in solution. Figure 13.9 shows some of these. Chemistry can be applied (derivatization, different pH) such that the distribution of the charged states is altered as described below. In the original MS/MS analysis using FAB/FIB ionization, it was important to ensure that no negative charges were present for the greatest ionization eficiency. Therefore, these carboxylic acid sites were esteriied using 3 N HCl in n-butanol. For this reason, the second step in the preparation of ACs and amino acids is derivatization of the carboxylic acid functional groups. By forming esters with the acid, the percentage of ACs and amino acids that may have a net negative charge is reduced substantially. This is also illustrated in Figure 13.9. Electrospray is a more eficient form of ionization than FAB/FIB so that nonderivatized ACs and amino acids are detected in the analysis. However, the number of ions that would be produced by dicarboxylic acid specimens is still reduced as compared to the butyl ester derivative. Another approach used by some labs is to acidify the inal mobile phase to shift the balance of ionization in favor of protonation and neutralization of acidic groups. The concern, here, however, is that a highly acidic environment could cause hydrolysis of some ACs. This is already known to occur during the derivatization process because of the highly acidic 3 N HCl in n-butanol.

Finally, source design and interface can affect how samples are ionized. Since this is highly dependent on the instrument choice and design, it would be inappropriate to describe speciic settings needed to optimize ionization. There are, however, some issues common to all instruments that are worth noting. Source cleanliness and operation is important to maximize sensitivity. Fortunately, the analysis of newborn blood spot extracts produces a relatively “clean” analysis and a source should operate for relatively long periods of time as compared to many other clinical analyses performed in the laboratory. It is recommended that labs try to use a speciic instrument primarily for NBS, otherwise problems in sensitivity and contamination are more likely from other assays, especially those that use organic buffers or whose sample preparation results in high concentrations of many compounds. High low rates may also affect cleanliness by exposing the MS to more solvent. The lowest possible low rates that produces excellent sensitivity is recommended. It is worth noting that a routine test should be put in place to check for sensitivity at the beginning of a long run of specimens, that is, an injection of a tune standard, and a criteria set (achieve a certain count threshold) that will ensure quality analysis for 300 or 400 samples. 13.2.3

MS/MS of ACs and Amino Acids

Following spray ionization, charged molecules enter the mass analyzers regions of the mass spectrometer where MS/MS is performed. The MS/MS is essentially two mass analyzers, MS1 and MS2, separated by a collision cell (CC) where molecules may be fragmented. Nearly all mass spectrometers used in AC and amino acid analyses are quadrupole mass spectrometers. In fact, there may be more than two sets of quadrupoles in addition to MS1 and MS2; one set may act as a preilter of ions with no mass separation while the other is part of the CC where fragmentation occurs. These additional quadrupoles enable more focused and abundant ion transmission. Optimizing a mass spectrometer for the best MS/MS analysis may seem complex at irst, but is simple in practice. Some strategies for this optimization as it pertains to NBS are described in a subsequent chapter. Although the details of how MS/MS analysis works are described elsewhere, it is helpful to briely review the major types of MS/MS analysis used in NBS. The irst mass analyzer in an MS/MS system (known as MS1) can be set to scan a range of masses or focus on a particular mass value. These ions are known as precursor (Pre) ions (historically called parent ions), since they come before fragmentation (the ions of origin) which

FURTHER PREPARATION FOR ANALYSIS &RPSRXQG

6KRUWKDQG,OOXVWUDWLRQ

WŽƐŝƟǀĞ

KK, E,Ϯ

,

&KDUJH, O O XVWU DWL RQ

Z

KZ

н ;,ϯͿϯE



EĞƚŚĂƌŐĞ

Z

нE;, Ϳ ϯ ϯ

ŽƌE,Ϯ

KK,

ĐĞƚLJůĂƌŶŝƟŶĞ ;ZсK,ϯͿ ůĂŶŝŶĞ ;Zс,ϯͿ 'ůƵƚĂƌLJů ĂƌŶŝƟŶĞ ;ZсK;,ϮͿϰKK,Ϳ 'ůƵƚĂŵŝĐ ĐŝĚ ;Zс;,ϮͿϮKKƵͿ ůĂŶŝŶĞ ƵƚLJůĞƐƚĞƌ ;Zс,ϯͿ 'ůƵƚĂƌLJů ĂƌŶŝƟŶĞ ƵƚLJůĞƐƚĞƌ ů ;ZсK;,ϮͿϰKK,Ϳ 'ůƵƚĂŵŝĐ ƵƚLJůĞƐƚĞƌ ;Zс;,ϮͿϮKƵͿ

Ϭ

нϭ

Ϭ

нϭ

Ϭ

нϭ

Ϭ

нϭ

285

EĞƵƚƌĂů

KK,

EĞŐĂƟǀĞ

͵ϭ

͵ϭ

͵ϭ

͵Ϯ

͵ϭ

нϭ

нϭ

нϭ

FIGURE 13.9 Illustration of potential charge states of ACs and amino acids in aqueous solution. The general chemical structure of AC and amino acids is shown on the top left. An illustration showing the functional groups as a triangle is shown in the top center and three symbols are used to represent charges. The net charge of a compound is shown as an integer in the center of each triangle. Each triangle point represents the functional group with a charge. In the center and bottom of the chart are the triangular illustrations for selected ACs and amino acids. The net charge state is the sum of the positive (+) and negative (−) charges. Several possible charge states are shown. For acetylcarnitine, there are two charge states of either 0 or +1. Derivatization represented by the solid area on the ends of the triangle result in only a +1 charge. Comparison of derivatized and underivatized species shows a shift of potential charge states from 4 possible species (0, +1, −1, −2) in the case of glutarylcarnitine to +1 for the butyl ester derivative. Only positively charged species are detected. Completely neutral charge states are not shown.

occurs in the CC. The CC is illed with a gas, commonly argon and nitrogen, such that a collision occurs between the Pre ion and the gas, resulting in fragmentation. The extent of fragmentation is often controlled by the amount of energy imparted on the molecules from the “nonmass separating” CC. These ions now called product ions (historically called daughter ions) are then separated by full scan or focused scans in the second mass analyzer (MS2). The products ions that pass from MS2 are then collected and counted at the detector. The ion signal is measured either as counts or voltage produced by the rate and number of ions striking the detector. The

detector, MS2, CC, and MS1 are all in sync and communicate in such a way that enables unique MS/MS settings such as Pre ion and NL scans. It is noteworthy that for NL and Pre ion MS/MS scans that although the ions detected are product ions, the mass spectra shows the Pre/parent/molecular ions m/z values. Product ion scans actually show the product ions. One setting that can be utilized as well is MS only, where essentially, the MS/MS operates as an MS-only system (one of the mass analyzers simply passes all ions through it and only one mass separation/selection is performed). The functions of MS/MS are illustrated in Figure 13.10.

286

MASS SPECTROMETRY IN CLINICAL ANALYSIS: SCREENING FOR INBORN ERRORS IN METABOLISM WĂƐƐƚŚƌŽƵŐŚ

^ĐĂŶŶĞĚ

WĂƌŬĞĚ

&ƌĂŐŵĞŶƚĂƟŽŶ

D^ͲKŶůLJ ^ĐĂŶ

WƌŽĚƵĐƚ/ŽŶ^ĐĂŶ

WƌĞĐƵƌƐŽƌ/ŽŶ^ĐĂŶ

DĂƐƐKīƐĞƚ^ĐĂŶ DŝŶƵƐyĂůƚŽŶƐ EĞƵƚƌĂů>ŽƐƐ^ĐĂŶ

^ĞůĞĐƚĞĚZĞĂĐƟŽŶDŽŶŝƚŽƌŝŶŐ^ĐĂŶ

FIGURE 13.10 Illustration of MS/MS scan functions. Full scanning of a mass range is represented as the box containing the mass spectrum. Single ion scanning (parked) is shown as the single line in the hexagonal box.

The MS/MS characteristics/features of ACs and amino acids (as butyl esters) were irst determined by running a product ion scan of a particular AC or amino acid. Two examples are the product ion scans of Phe and palmitoylcarnitine shown in Figure 13.11. In the original research, the product ion scans for ACs and amino acids show common fragment ions or molecules. In the case of ACs, that common fragment was an ion with an m/z of 85. For amino acids, the common fragment was a molecule with no charge and a mass of 102 Da. Ions that

strike the detection in an NL 102 scan are 102 Da less than the Pre ions. The fragmentation pathways of these compounds are included in Figure 13.11. Note that some amino acids and FC may produce more abundant fragment ions or neutral molecules at different mass values. FC produces a more abundant 103 Da fragment ion in addition to the fragment ion at m/z 85. Basic amino acids have two neutral fragments produced. One fragment is the loss of ammonia (−17 Da) and the other is an amino group and butyl formate (−119 Da). For

FURTHER PREPARATION FOR ANALYSIS н

WƌŽĚƵĐƚƐŽĨϮϮϮ

120

90

,

80

,Ϯ

KKƵ

70 % Intensity

ŵͬnjϮϮϮ

E,ϯ



287

60

>±'D QHXWUDOPROHFXOH@

50

,KKƵ ;ϭϬϮĂͿ

40 30 ŵͬnjϭϮϬ

20 222

50

100

150 m/z, amu

85

200

,



,Ϯ

250

WƌŽĚƵĐƚƐŽĨϰϱϲ

90 80 70 % Intensity

н E,

Ϯ

10

;,ϯͿϯE

KK;,ϮͿϭϰ,ϯ KKƵ

60 50

ϰ,ϴ

>'D SURGXFWLRQ@

40

KK;,ϮͿϭϰ,ϯ ;,ϯͿϯE

30 20 10

456 57

100

200 300 m/z, amu

н ,Ϯ

KK,

400

FIGURE 13.11 Product ion scans for Phe and Palmitoyl carnitine butyl esters. amu, atomic mass units.

example, citrulline can be observed in an NL 102 spectrum at m/z 215 even though protonated molecular ion is at 232 Da. The loss of ammonia occurs in the source of the mass spectrometer prior to mass separation so that the actual Pre ion is of mass 215 and not 232. It is interesting to note that both AC butyl esters and “underivatized” free acid produce the same fragment ion at m/z 85, an important concept to note since any underivatized ACs will appear at a different mass (less 56 amu). For underivatized amino acids, the NL is formic acid (46 Da). As mentioned above, unless otherwise noted, the analysis of ACs and amino acids is focused on the butyl ester derivatives. Many MS/MS analyses utilized three different scan functions in their routine analysis. These are NL 102, Pre 85, and several selected reaction monitoring (SRM) and an example of a normal newborn proile are shown in Figure 13.12. As observed from the product ion spectra (Figure 13.11), a Pre ion scan at 85 Da is utilized. Note that for many MS/MS systems, the collision energy is ramped from low to high m/z since more energy is required to fragment a larger, very long-chain AC (C16) than that for short-chain acetylcarnitine (C2)

or FC. The collision energy increases from low to high mass for amino acids. However, the collision energy difference in this NL of 102 Da scan is often less than that for ACs in part due to the lower mass range (130–275 for amino acids vs. 210–500 for ACs). Several SRMs are often used in addition to the full-scan Pre 85 and NL 102. Utilization of SRM scans conserve the scan times primarily by not wasting precious scan time of masses that are not of interest. SRM is also used for improved sensitivity since longer dwell times per mass can be used. For example, a full scan of carnitine and AC could range from 215 to 500 Da (which includes FC through C18OH). This is 285 mass values and data points if one point per mass is acquired. By eliminating FC from a full-range scan and starting the analysis at m/z 255, nearly 40 mass units are not acquired. An SRM can be set up for FC that sets a single Pre ion and product ion for each transition. More than one SRM can be programmed into the analysis. Although an analysis of only SRM of important masses can be acquired, most labs utilize full-scan data so that visual spectra can be observed. This is important for interpretation and quality assessment.

288

MASS SPECTROMETRY IN CLINICAL ANALYSIS: SCREENING FOR INBORN ERRORS IN METABOLISM SRM (Short-Chain Acylcarnitines)

SRM (Free Carnitine)

C2

2H -FC 9

FC

2H -C2 3

13C

() (C2) + Glu

2H FC 3

from hydrolyzed acylcarnitine i.s.

218/103

221/103

260/85

227/103

2H -C3 3

C3

261/85 263/85

274/85

277/85

Full-Scan Precursor 85 (Acylcarntines)

Ion Intensity

2H -C16 3

C16

2H -C14 9 2H -C4 3

2H -C8 3

2H -C5 9

C18:1 C18

2H -C3 3

C3 280

300

320

340

360

380

Pro

2H -Ala 4

Gly

Ala Ser 150

420

440

460

480

500

SRM (Basic Amino Acids) 2H -Cit 2

2H -Leu 3

Leu/Ile Val

Phe

Met 200

2H -Orn 2

13C -Phe 6 13C -Tyr 6

2H -Met 3

15N,13C-Gly y

Full-Scan Neutral Loss 102 (Amino Acids)

400

Tyr

2H -Glu 3 Glu

250

Orn Cit 189/70

232/113

Arg

2H 13C-Arg 4

236/76

mass/charge (m/z)

FIGURE 13.12 Five scans that comprise one MS/MS scan period in a typical MS/MS analysis. These include three SRM scans and two full scans. Stable isotope internal standards are represented as stars.

This same full-scan and SRM approach used for ACs is also done for amino acids. Basic amino acids such as citrulline and arginine are characterized by loss of 102 Da and ammonia or other basic amino side chain. Selective reaction monitoring (SRM) is used rather than full scan and is based on NL of 119 for citrulline (102 + 17, where 17 is ammonia) and NL of 161 for arginine (102 + 59, where 59 is the amino side chain). These SRMs are often grouped together in the visual spectrum as shown in Figure 13.12. Note that citrulline can also be acquired in an NL 102 scan since sourceinduced dissociate may cause the ammonia to be lost and hence detection of m/z 215 (MH+ minus 17) versus m/z 232. Details of this fragmentation are described elsewhere. With the MS/MS analysis of a mixture of full scan and SRMs, an analysis of 1 min may have from 7 to 15 scan periods depending on the dwell times for each mass and how many points per mass are acquired. Each scan contains acquisition of the one scan for the full scans and SRM acquisitions called a scan period. It is important that the elution of the low injection lasts

suficiently long to get enough scans for quantiication or there may be higher imprecision. Flow rates have to be kept narrow enough so that one sample can be acquired every 2 min (MS analysis time 30 s) and baseline achieved but also not so narrow as to result in only two or three complete scans per analysis. There are many issues in setting up an ideal MS/MS analysis than can impact on the quality of results. It is important for the user to understand many fundamental issues in MS and the instrument they are using. For example, insuficient dwell time due to too fast of a scan could cause reduced sensitivity and an unexpected mass shift (apparent loss of calibration). To avoid this, typically an AC scan is approximately 2 s per period or about 0.01 s per unit mass. Each period may be 3–5 s, which includes amino acids and SRM analysis. It is important to note that some users acquire data differently. For example, mass spectrometers may be set to acquire one point per mass value while others use one-tenth point per mass. For acquisitions of one point per mass (or dalton), it is important that the masses acquired shift with increasing mass based on calibration

DATA PROCESSING

and other methods, and that actual mass of the molecule (not integer value often used to describe the mass of compounds). This is often known as the mass offset. Carbon 12 has a mass of 12.000 and the decimal mass doesn’t change with an increase in carbon number. However, hydrogen has an exact mass of 1.008 and there can be many hydrogen atoms in larger molecules. A compound with 20 additional hydrogen atoms will have an increase in the decimal mass value of 0.16 Da. For the analysis of ACs it is not unusual to ind an analysis of acetylcarntine at 260.3 and an analysis of C16 at 456.5 This is from the increase of 14 carbons (from C2 to C16) and as such an additional 28 hydrogen atoms for a total mass increase in addition to integer mass of 0.224 Da. Some users set their mass spectrometers scan every tenth of a mass unit. Therefore, the exact mass at the max peak height for each mass is always acquired. Data processing software will integrate each peak producing a single value representing the area or peak height. Note that SRM analysis is always one point per peak since it is deined as one mass transition between Pre ion and product ion.

13.3

DATA PROCESSING

Data is processed in a manner that is unique to each MS/MS software package. There are, however, common features of these software packages used in NBS. Data are obtained as ion intensity for each m/z value and each scan function. Figure 13.13 illustrates this process. The elution proile is integrated as shown in Figure 13.8 using either the top 50% or 90% of the peak depending on how many periods deine the peak. The m/z values may be either single points per integer mass or fractions of mass. If 10 points per mass are acquired, then software may be utilized to baseline subtract, smooth, and integrate the peak area or height. Some compounds have the same molecular mass but are quantitatively unique. For example, m/z 260 is the butyl ester of acetylcarnitine and glutamic acid. However, each was acquired using different scan functions (Pre 85 and NL 102, respectively). This illustrates the unique abilities of MS/MS and why the scan function is associated with each m/z value and ion intensity. Data are exportable in a spreadsheet compatible format. Special software to further process data can be obtained from the vendor as part of the MS/MS system, commercially available from software designers, or produced in house. The data processing systems include calculation of ion ratios based on two m/z values, which in turn is used to calculate concentration based on userdeined values. Furthermore, many programs allow

289

additional calculations such as molar ratios (ratios of concentrations) and input of abnormal reference ranges. It is important to note the differences between a molar ratio and an ion ratio. An ion ratio is simply the ratio of two ion intensities while a molar ratio is a ratio of two concentration values. 13.3.1

From Data to Report

There are three types of results that can be obtained from an NBS MS/MS analysis. These are illustrated in Figure 13.13. The most basic of results is ion intensity. As described previously, every m/z data point is paired with an ion intensity value. In terms of reporting, however, usually the ion intensity of one or two of m/z values for selected internal standards is listed in a report or spreadsheet. This provides the screening lab an ability to assess the quality of the analysis by determining whether the intensity of an internal standard is suficiently high to ensure a sensitive analysis and conirm the mass spectrometer is operating properly or the sample was prepared properly. For example, the ion intensities of the internal standards for Phe and octanoylcarnitine are usually obtained. Ion counts which fall below a particular threshold for a few samples in a run require an examination of the full proile and analysis and may trigger a reanalysis or second sample preparation. Ion counts that fall below the threshold for multiple samples or plates may suggest the mass spectrometer requires maintenance to restore sensitivity to the instrument. The second type of result that can be generated is the concentration of metabolites. The calculation is based on ion ratios. Concentrations are most often calculated from the ratio of the ion intensity of a metabolite (AC or amino acid) to its respective stable isotope-labeled internal standard. The internal standard is most ideally an isotopomer. For example, to obtain the concentration for octanoylcarnitine, the ion ratio of the raw ion intensity of octanolycarnitine at m/z 344 to the raw ion intensity of 2H3-octanoylcarnitine at m/z 347 is calculated. This ratio becomes a concentration after interpolation from a standard curve or using response factors and constants that include the concentration of the internal standard, which is known. If it was assumed the extraction of ACs was 100% and the detection eficiency for both labeled and unlabeled AC was identical, then the calculation of concentration would be calculated simply using Equation 13.1. Note that I indicates ion intensity and C indicates concentration for x (metabolite of interest) and s (internal standard or reference standard): I x /Is = C x /C s or C x = I x /Is × C s.

(13.1)

290

MASS SPECTROMETRY IN CLINICAL ANALYSIS: SCREENING FOR INBORN ERRORS IN METABOLISM

&ƌŽŵĂƚĂƚŽZĞƉŽƌƚ

WŚĞ сϮϮϮͬϮϮϳ×ϳϲ× ϭ͘ϭ ϴсϯϰϰͬϯϰϳ×ϭ͘Ϭ×ϭ͘ϭ &сϮϭϴͬϮϮϳ×ϮϬ×ϭ͘ϭ

DĞƚĂďŽůŝƚĞ

ŽŶĐ͘;mDͿ

WŚĞ

ϳϰ͘ϯ



ϰ͘ϰ

&

ϲ͘ϲ

^ĐĂŶ

ŵͬnj

/ŽŶŽƵŶƚ

E>ϭϬϮ

ϮϮϮ

ϰϬ͕ϬϬϬ

E>ϭϬϮ

ϮϮϴ

ϰϱ͕ϬϬϬ

WƌĞϴϱ

ϯϰϰ

ϭϮ͕ϬϬϬ

WƌĞ ϴϱ

ϯϰϳ

ϯ͕ϬϬϬ

^ZD

ϮϭϴͬϭϬϯ

ϲ͕ϬϬϬ

^ZD

ϮϮϳͬϭϬϯ

ϮϬ͕ϬϬϬ

ϴͬ&сϰ͘ϰͬϲ͘ϲсϬ͘ϲϳ

DĞƚ,ŝŐŚ WŚĞ ϭϴϬ ϴϬ͘ϯϱ &ϭϬϬ ϴͬ&Ϭ͘Ϭϱ

DĞƚĂďŽůŝƚĞ ŽŶĐ͘

ƵƚŽī

EŽƚĞ

WŚĞ

ϳϰ͘ϯ

ϭϴϬ

K
Žǁ

ϴͬ&

Ϭ͘ϲϳ Ϭ ϲϳ

Ϭ͘Ϭϱ Ϭ Ϭϱ

,ŝŐŚ

WŚĞ ŝ͘Ɛ͘

ϰϱϬϬϬ

ϮϬϬϬϬ

K
ĞƵ

ůĂ

DĞƚ

Wͬd



'ůƵ

dLJƌ

WŚĞ

DĞƚ

>ĞƵ

ůĂ

'ůƵ

dLJƌ

WŚĞ

>ĞƵ

DĞƚ

ůĂ

WŚĞ

dLJƌ

Wͬd

>ĞƵ

DĞƚ

нннн

ŽŬ

ннннн

ŽŬ

ŽŬ

;WŝǀĞƌŝƐĞĂƐĞ ŚŽůĞƐƟ ƚĂƐŝƐ dŽƚĂůWĂƌĞŶƚĞƌĂů EƵƚƌŝƟŽŶ ;/sŵŝŶŽĐŝĚƐͿ

DĞĚŝƵŵͲŚĂŝŶĞĨ DƵůƟƉůĞĞĨ͘ ^ŚŽƌƚͲŚĂŝŶĞĨ͘ sĞƌLJ>ŽŶŐͲŚĂŝŶĞĨ͘ DĞĚŝƵŵͲŚĂŝŶdƌŝŐůLJĐĞƌŝĚĞƐ ;DdKŝůͿ sĂůƉƌŽŝĐ ĐŝĚ ;ŶƟĐŽŶǀƵůƐĂŶƚͿ

FIGURE 13.16 Summary of the multiplex, multiple metabolites, and single-assay capability of MS/MS used in NBS. One MS/MS can measure two families of compounds, amino acids and ACs. These proiles include more than 65 different metabolites, of which some are listed in the igure with their mass values under each proile type. These metabolites can than be used to detect several different diseases (more than 45 distinct genetic diseases have been reported). One metabolite may be helpful in detecting more than one disease (genetic, nongenetic, iatrogenic, age, nutrition). Genetic diseases for NBS are shown with solid arrows, whereas non genetic conditions that may be helpful in ruling out false results or to be used in clinical assays and nongenetic screening are shown with dashed arrows.

13.3.4

Expansion of NBS

MS/MS is clearly a versatile instrument in that it can measure different metabolite classes in a single assay, that is, ACs and amino acids. Structurally similar compounds that may also be analyzed include thyroxine (T4). Thyroxine is derived from Tyr and is an iodinated hormone. Due to the fact that T4 retains the alpha amino groups, it can also be detected as an NL of 102 for the butyl ester. This recently published assay [22] has yet to be widely used in part because metabolite is present in very low concentrations (more than 100-fold less than some amino acids), and the disorder screened

294

MASS SPECTROMETRY IN CLINICAL ANALYSIS: SCREENING FOR INBORN ERRORS IN METABOLISM

(congenital hypothyroidism) produces a low T4 compared to normal, and hence analytical sensitivity must be even greater. Newer instrumentation does improve the detection limits and the assay compares well to a commercial T4 immunoassay. Further, many labs screen for thyroid-stimulating hormone (TSH) levels rather than T4 and it is believed that this is a better analyte. Arguments can be made that measuring both T4 and TSH would be ideal and using MS/MS could save considerable dollars by eliminating one of these assays. In addition to adding metabolites that are coprepared with amino acids and ACs, the DBS can be reused, as is the case with succinylacetone. As shown in Figure 13.7, a previously methanol extracted blood spot could be reused of a different assay such as succinylacetone. The leftover spot is derivatized using hydrazine and the resultant derivative is recombined with the derivatized ACs and amino acids for a single injection into the MS/MS [18,19]. Succinylacetone is the key metabolite for detecting tyrosinemia type I and for this reason it is being used with greater frequency in NBS laboratories. Finally, new assays can be developed using other blood spot punches available from the DBS card. The analysis of 17-OH progesterone is being investigated and shown to be equivalent to the analysis using immunoassays. MS/MS can analyze other steroids in the same analysis which has shown to reduce false-positive rates. However the analysis requires LC and takes more than 2 min to perform. Solutions are being developed, however, to enhance this throughput. Currently, the assay is primarily used for second-tier conirmation analysis rather than irst-tier primary screening analysis. Other new assays included an enzyme assay for lysosomal storage diseases (LSDs), where the substrates metabolized by the enzymes produce products that are easily analyzed by MS/MS and may be recombined in the inal injection solvent that contains amino acids and ACs [23,24]. 13.3.5

certain key metabolites such as Phe and C8 internal standards could be used to assess satisfactory performance. Injection of a tune standard might be used for this purpose. For example, achievement of 1 × 105 ion counts for the peak intensity for Phe internal standard may be an indicator of performance for the next 200– 400 samples. Participating in proiciency testing of actual blood spots with enriched metabolites is also recommended to ensure that both the analysis, calculation of concentrations, and reporting are satisfactory. Many proiciency testing programs such as that administered by the Centers for Disease Control and Prevention not only provide results submitted by each lab but a blind comparison to other laboratories performing the same test. Blood spot testing should also be performed in each plate to ensure quality analysis per run. Although MS/MS is a robust and accurate system, it can produce false results from loss of calibration, sensitivity, or other tuning failures.

13.4

SUMMARY

MS/MS analysis of DBS is one model of clinical analysis that has been utilized on millions of infants for 15 years. Each year since its irst clinical use in the mid-1990s, there have been substantial improvements such as improved ionization, more sensitive instruments, and a reduction of the actual size of instruments. The proile has expanded to include T4, succinylacetone, and in a few labs, LSD metabolites and steroids. Many of these analyses can be used in other clinical applications and research such as nutrition studies in premature infants or postmortem screening on sudden unexplained deaths in infants or carnitine status in dialysis patients with end-stage renal disease. As additional clinical assays are developed and used, it might be wise for these labs to consider the models of MS/MS, NBS, and the DBS as a reference.

Ensuring Quality Analysis in Screening

Having an adequate quality assurance program in an NBS lab is not an option but an essential requirement. Assuring quality, MS/MS analysis can be done by participating in proiciency testing through various programs in the United States and abroad. From the laboratory perspective, it is important that the MS/MS system receives proper and regular maintenance by certiied engineers. However, this alone is insuficient. Pre ion and NL scans are not utilized much in clinical and pharmaceutical analysis. Therefore, AC and amino acid tune standards and check samples should be prepared regularly after each maintenance and major cleaning/ adjustments of the instrument. An ion count “goal” for

REFERENCES 1. Guthrie, R. (1992) The origin of newborn screening. Screening, 1, 5–15. 2. Chace, D.H., Millington, D.S., Terada, N., Kahler, S.G., Roe, C.R., Hofman, L.F. (1993) Rapid diagnosis of phenylketonuria by quantitative analysis for phenylalanine and tyrosine in neonatal blood spots by tandem mass spectrometry. Clin. Chem., 39, 66–71. 3. Chace, D.H., Sherwin, J.E., Hillman, S.L., Lorey, F., Cunningham, G.C. (1998) Use of phenylalanine-to-tyrosine ratio determined by tandem mass spectrometry to improve newborn screening for phenylketonuria of early

REFERENCES

4.

5. 6.

7.

8.

9.

10.

11.

12.

13.

discharge specimens collected in the irst 24 hours. Clin. Chem., 44, 2405–2409. Niwa, T. (1986) Metabolic proiling with gas chromatography–mass spectrometry and its application to clinical medicine. J. Chromatogr., 379, 313–345. Gerhards, P. (1999) GC/MS in Clinical Chemistry. Weinheim, Germany and New York: Wiley-VCH. Norwood, D.L., Kodo, N., Millington, D.S. (1988) Application of continuous-low liquid chromatography/ fast-atom bombardment mass spectrometry to the analysis of diagnostic acylcarnitines in human urine. Rapid Commun. Mass Spectrom., 2, 269–272. Millington, D.S., Norwood, D.L., Kodo, N., Roe, C.R., Inoue, F. (1989) Application of fast atom bombardment with tandem mass spectrometry and liquid chromatography/ mass spectrometry to the analysis of acylcarnitines in human urine, blood, and tissue. Anal. Biochem., 180, 331–339. Millington, D.S., Roe, C.R., Maltby, D.A. (1984) Application of high resolution fast atom bombardment and constant B/E ratio linked scanning to the identiication and analysis of acylcarnitines in metabolic disease. Biomed. Mass Spectrom., 11, 236–241. Yergey, A.L., Liberato, D.J., Millington, D.S. (1984) Thermospray liquid chromatography/mass spectrometry for the analysis of L-carnitine and its short-chain acyl derivatives. Anal. Biochem., 139, 278–283. Millington, D., Kodo, N., Terada, N., Roe, D., Chace, D. (1991) The analysis of diagnostic markers of genetic disorders in human blood and urine using tandem mass spectrometry with liquid secondary ion mass spectrometry. Int. J. Mass Spectrom. Ion Process., 111, 211–228. Millington, D.S., Kodo, N., Norwood, D.L., Roe, C.R. (1990) Tandem mass spectrometry: a new method for acylcarnitine proiling with potential for neonatal screening for inborn errors of metabolism. J. Inherit. Metab. Dis., 13, 321–324. Rinaldo, P., O’Shea, J.J., Welch, R.D., Tanaka, K. (1990) Diagnosis of medium chain acyl-CoA dehydrogenase deiciency by stable isotope dilution analysis of urinary acylglycines: retrospective and prospective studies, and comparison of its accuracy to acylcarnitine identiication by FAB/mass spectrometry. Prog. Clin. Biol. Res., 321, 411–418. Clark, R.H., Chace, D.H., Spitzer, A.R. (2007) Effects of two different doses of amino acid supplementation on growth and blood amino acid levels in premature neonates admitted to the neonatal intensive care unit: a randomized, controlled trial. Pediatrics, 120, 1286–1296.

295

14. Chace, D.H., Kalas, T.A., Naylor, E.W. (2003) Use of tandem mass spectrometry for multianalyte screening of dried blood specimens from newborns. Clin. Chem., 49, 1797–1817. 15. Guthrie, R., Susi, A. (1963) A simple phenylalanine method for detecting phenylketonuria in large populations of newborn infants. Pediatrics, 32, 338–343. 16. Hannon, W.H. (2007) Blood Collection on Filter Paper for Newborn Screening Programs; Approved Standard—Fifth Edition. NCCLS Document LA4-A5, p. 27. 17. Reuter, S.E., Evans, A.M., Faull, R.J., Chace, D.H., Fornasini, G. (2005) Impact of haemodialysis on individual endogenous plasma acylcarnitine concentrations in end-stage renal disease. Ann. Clin. Biochem., 42, 387– 393. 18. Chace, D.H., Lim, T., Hansen, C.R., De Jesus, V.R., Hannon, W.H. (2009) Improved MS/MS analysis of succinylacetone extracted from dried blood spots when combined with amino acids and acylcarnitine butyl esters. Clin. Chim. Acta, 407, 6–9. 19. Turgeon, C., Magera, M.J., Allard, P., Tortorelli, S., Gavrilov, D., Oglesbee, D., et al. (2008) Combined newborn screening for succinylacetone, amino acids, and acylcarnitines in dried blood spots. Clin. Chem., 54, 657–664. 20. Chace, D.H., Lim, T., Hansen, C.R., Adam, B.W., Hannon, W.H. (2009) Quantiication of malonylcarnitine in dried blood spots by use of MS/MS varies by stable isotope internal standard composition. Clin. Chim. Acta, 402, 14–18. 21. Chace, D., Millington, D. (1994) Neonatal screening for inborn errors of metabolism by automated liquid secondary ion tandem mass spectrometry. In New Horizons in Neonatal Screening, Proceedings of the 9th International Neonatal Screening Symposium, edited by Farriaux, J., Dhont, J. Amsterdam, The Netherlands: Elsevier, pp. 373–377. 22. Chace, D.H., Singleton, S., Diperna, J., Aiello, M., Foley, T. (2009) Rapid metabolic and newborn screening of thyroxine (T4) from dried blood spots by MS/MS. Clin. Chim. Acta, 403, 178–183. 23. Dajnoki, A., Muhl, A., Fekete, G., Keutzer, J., Orsini, J., Dejesus, V., et al. (2008) Newborn screening for Pompe disease by measuring acid alpha-glucosidase activity using tandem mass spectrometry. Clin. Chem., 54, 1624– 1629. 24. DeJesus, V., Rios, I., Davis, C., Chen, Y., Calhoun, D., Zakeri, Z., et al. (2002) Induction of apoptosis in human replicative senescent ibroblasts. Exp. Cell Res., 274, 92–99.

14 MASS SPECTROMETRY FOR STEROID ANALYSIS William J. Griffiths, Michael Ogundare, Anna Meljon, and Yuqin Wang

14.1

INTRODUCTION

Mass spectrometry (MS) in combination with the chromatographic separation techniques of gas chromatography (GC) or liquid chromatography (LC) have become the “gold standard” methods for the analysis of steroids [1–3]. These methods allow the identiication of novel steroids and in combination with reference standards quantiication. Biological MS stretches back to the 1950s, when steroids was one of the irst class of biomolecules analyzed by MS [4–9]. MS involves the vaporization and ionization of analyte in the ion source and subsequent analysis according to mass-to-charge ratio (m/z) in the analyzer prior to detection. Two mass (m/z) analyzers may be arranged in series either “in space” or “in time” to allow tandem mass spectrometry (MS/MS or MS2). Ion trap instruments, where mass (m/z) analyzers are arranged in series “in time,” are also capable of “third-order” tandem mass spectrometry (MS/MS/MS or MS3) extending up to MS6 and beyond (i.e., MSn), depending on the availability of sample. In the early studies of steroids, electron ionization (EI) was the dominant ionization mode [4–9], although chemical ionization (CI) was later exploited [10]. EI and CI are readily coupled to GC as both operate with gas phase analytes. However, many steroids require derivatization before they are suitable for gas chromatography–mass spectrometry (GC-MS) analysis [1–3,9,11]. Following the introduction of fast atom bombardment (FAB) ionization it became possi-

ble to analyze steroids and their conjugates directly by fast atom bombardment mass spectrometry (FAB-MS) without derivatization [12,13]. This is similarly true for the atmospheric pressure ionization (API) methods of electrospray ionization (ESI) [14], atmospheric pressure chemical ionization (APCI) [15], and desorption electrospray ionization (DESI) [16], and also for vacuum or AP matrix-assisted laser desorption/ ionization (MALDI) [17]. Steroids are important biomolecules with structural, signaling, and regulatory functions (Figure 14.1) [1,9]. For example, in vertebrates cholesterol is an important constituent of cellular membranes and is enriched in microdomains such as lipid rafts [18]. The classical steroid hormones are ligands to nuclear receptors, while a number of sterol metabolites are now being identiied as ligands to once “orphan” and subsequently “adopted” nuclear receptors such as liver X receptors (LXRs) α and β, pregnane X receptor (PXR), and farnesoid X receptor (FXR) [19]. Latest research has also revealed bile acids and bile alcohols (oxysterols) to be ligands to G protein-coupled receptors [20,21]. Steroids can also posttranslationally modify protein [22]. The continued discovery of new properties of steroids and their wide structural variation in biology dictates the need of methods for their analysis which are unbiased, quantitative, and simple. In the following sections we will discuss the various MS approaches which attempt to reach these goals, including sample preparation strategies before MS analysis.

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

297

298

MASS SPECTROMETRY FOR STEROID ANALYSIS

FIGURE 14.1 Structures of the main classes of steroids found in vertebrates. Useful structure drawing tools can be found at http://www.lipidmaps.org/.

14.2

SAMPLE PREPARATION

The simplest method of sample preparation from biological luids is by reversed-phase solid-phase extraction (RP-SPE). Traditionally this has been performed on commercially available C18 cartridges/columns [12,13] or small packed columns, which are loaded, washed, and steroids inally eluted with methanol or ethanol.

Extraction can similarly be performed on an RP highperformance liquid chromatography (HPLC) column. Nonpolar sterols will not be captured from highly aqueous solutions in which they are dispersed rather than dissolved [23], in which case they should be irst extracted with ethanol which is then diluted to ∼70% ethanol and passed through the RP-SPE column. The low through is further diluted to 35% ethanol and reap-

GAS CHROMATOGRAPHY–MASS SPECTROMETRY

plied to the column. Following further dilution and recycling the sterols are fully extracted on the column at 15% ethanol. Following a wash step with 10% ethanol they can be eluted with 99.9% ethanol or 100% methanol. For the most polar steroids, the inal ethanol dilution and wash should be at 5% ethanol. Steroids can be similarly extracted from tissue samples [24,25]. Generic protocols for the extraction of sterols, bile acids/alcohols, and conjugated steroids from urine, plasma, and tissue are given below (see the section “Extraction Protocols”). It should be emphasized that each protocol should be validated for the analytes of interest by the dedicated researcher. Steroids can then be analyzed by direct infusion electrospray ionization mass spectrometry ESI-MS (shotgun steroidomics), liquid chromatography–mass spectrometry (LC-MS) or GC-MS following derivatization. Alternatively, the steroid fraction can be further fractionated according to acid strength by ion-exchange chromatography [9,23,26,27].

14.3 GAS CHROMATOGRAPHY–MASS SPECTROMETRY GC-MS offers the high chromatographic separation of gas–liquid chromatography with the mass (more correctly m/z) selective detection of MS. Ionization is usually performed via EI, in which case structural information is forthcoming via informative and often characteristic fragmentation, also allowing library-based searches. Alternatively, CI spectra are dominated by ions indicative of molecular weight. However, for the majority of steroid analysis, derivatization is required prior to injection on the GC column [28]. Derivatization is designed to enhance analyte volatility and thermal stability. Once derivatized, positional isomers including stereoisomers can often be separated by GC and differentiated by electron ionization mass spectrometry (EI-MS) [1].

14.3.1 Derivatization for Gas Chromatography– Electron Ionization Mass Spectrometry (GC-EI-MS) The most important derivatization reactions are (1) conversion of alcohol groups to trialkylsilyl ethers, often trimethylsilyl ethers (TMSO), (2) methylation of carboxylic acid groups, and (3) conversion of oxo groups alkyl oximes [1–3,11] (Figure 14.2). There are numerous reagents for the formation of trialkylsilylethers (see table 1 in [1]), for example, reaction with dry pyridine : hexamethyldisilazane : trimethylchlorosilane 3:2:1 or 9:3:1 (by volume) with or without heating. However, if oxo groups are present in the steroid, then these reac-

299

tions can yield enol TMS ethers and multiple products. This artifact can be avoided by converting the oxo group into an oxime, usually a methyloxime (MO). To perform this conversion, the sample is dissolved in 100 µL of pyridine with 10 mg methoxyammonium chloride and heated for 30–60 min at 60°C. The preferred derivatization of a carboxyl group is via methylation. Methylation can be achieved by reaction with fresh diazomethane in ether added to a methanol solution of the sample. This method gives the highest speciicity for carboxyl groups and side reactions are few. Formation of methyl ethers is generally not a problem and is further prevented by low reaction temperature and short reaction times. Oxo groups may react to a minor extent with diazomethane. The drawback of using diazomethane is its toxicity in preparation and use, and this reagent requires care in handling. An alternative reagent, trimethylsilyldiazomethane in hexane, is commercially available and appears useful for preparation of bile acid methyl esters [29]. A generic protocol for bile acid and sterol derivatization for GC-MS is given in the section “Bile Acids for GC-MS.” Derivatization is also advantageous for gas chromatography–chemical ionization mass spectrometry (GC-CI-MS) studies. For example, alcohol groups can be derivatized to perlouroacylesters and oxo groups to pentaluorobenzyloximes, both of which are utilized in electron-capture (EC) CI [10]. Gas chromatography– atmospheric pressure chemical ionization mass spectrometry (GC-APCI-MS) is now available on modern instruments and can be utilized in both the positive and negative ion modes. However, the thermal lability and involatility of many steroids still dictates the requirement of derivatization to allow passage through the GC column even though ionization is at atmospheric pressure. 14.3.2

EI Fragmentation

A number of the common fragment ions and losses seen in mass spectra of methylester-trimethylsilylether derivatives of bile acids and of trimethylsilylether derivatives of C19 and C21 steroids are shown in Tables 14.1 and 14.2. The majority of fragment ions are formed as a result of: (1) the loss of methyl groups, angular or from the TMSO group; (2) the elimination of water or the TMS alcohol; and (3) the loss of the side chain and part or all of the D-ring. Spectra are complicated as these fragmentations may proceed successively, leading to fragments of fragments. C27 and C24 bile acids and bile alcohols (oxysterols), if substituted in the side chain, also show cleavage α to a TMSO group. EI spectra are reproducible and allow the differentiation of isomers, including stereoisomers.

300

MASS SPECTROMETRY FOR STEROID ANALYSIS

FIGURE 14.2 Some examples of derivatization reactions for GC-MS. (A) Lithocholic acid is dissolved in methanol to which solutions of toluene and trimethylsilyldiazomethane in hexane are added. After a short reaction time (30 min, 40°C), the resulting methyl ester is dried and derivatized to the TMS ether by the addition of dry pyridine : hexamethyldisilazane : trimethylchlorosilane (3:2:1 or 9:3:1, v/v/v) with or without heating (e.g., 60°C, 30 min). (B) 3β-Hydroxy-7-oxochol-5-en-24-oic acid is methylated as in (A), and then the dried sample is converted to the MO by reaction with methoxyammonium chloride in pyridine at 60°C for 30 min. The product is dried under a stream of nitrogen, and the TMS ether is formed by addition of trimethylsilylimidazole (TMSI) and heating. TMSI is involatile but can be removed by adding 1 mL of cyclohexane to the hot reaction tube followed by 0.5 mL of water. After vortexing and centrifugation, the upper layer is suitable for transfer to the GC injection vile. See the section “Bile Acids for GC-MS.”

TABLE 14.1 Observed Ion (m/z)

Diagnostic Ions and Fragmentations Observed in Bile Acid Methyl Ester-Trimethylsilyl Ether Derivatives Origin

129 131 143/159 161/162

A-ring fragment (C-1 to C-3 and -OTMS) Glycine side-chain fragment Possible C-22-23 plus trimethylsilanol C-ring cleavage after loss of C-7 trimethylsilanol

174/175

C-ring cleavage after loss of C-7 trimethylsilanol

175 181 181 182 195

Cleavage alpha to a C-22 trimethylsilanol A-ring fragment after loss of C-4 trimethylsialnol Cleavage across C-9,10 and C-5,6 bonds (m/z 271–90) B-ring fragmentation to form A-ring minus –OTMS A-ring + C-6 and -OTMS, derived from m/z 285 (−90)

208 211 213 215 217

Cleavage of C-11,12 and C-8,14 bonds with -OTMS D-ring cleavage, loss of side chain and C-15 to C-17 D-ring cleavage, loss of side chain and C-15 to C-17 D-ring cleavage, loss of side chain and C-15 to C-17 A-ring fragment (1,3-bis-trimethylsiloxy structure)

Comments on Probable Structure Intense for 3β-hydroxy-Δ5 structure Intense ion in glycine-conjugated Me-TMS ethers Observed in C-23 hydroxy-cholanoates Diagnostic for 7-hydroxy-3-oxo-Δ4 structure, or Δ4,6 structure Diagnostic for 7-hydroxy-3-oxo-Δ4 structure, or Δ4,6 structure Base peak in 22-hydroxy-UDCA Base peak in 3α,4β-hydroxy structures Observed in C-2 hydroxy structures Observed in C-1 and C-2 hydroxy structures Intense in 3,6,7-trihydroxy, seen in C-2 hydroxy structures Feature of 12-hydroxy group Low abundance in trihydroxy cholanoates Observed in dihydroxy cholanoates Prominent in monohydroxy cholanoates Usually base peak in 1β-hydroxy structures

GAS CHROMATOGRAPHY–MASS SPECTROMETRY

TABLE 14.1 Observed Ion (m/z)

(Continued) Origin

217 217 224 229

Cleavage of C-20,22, loss of side chain and 22-OTMS Fragmentation of glucuronide moiety A-ring after cleavage of C-7,8 and C-9,10 bonds D-ring plus side chain and C-15 to C-17 carbons

231

Loss of trimethylsilanol from m/z 321 fragment

243 243 247

Cleavage of AB-rings (C-3 to C-7 and bis-OTMS) Side chain plus C-15 to C-17 Cleavage through B-ring

249

Cleavage through B-ring

249

CD-rings plus side chain

251 253 255 257 261 262 269

ABCD-rings after loss of substituents and side chain ABCD-rings after loss of substituents and side chain ABCD-rings after loss of substituents and side chain ABCD-rings after loss of -OTMS group and side chain CD-ring fragment after cleavage of B-ring CD-ring and C-7 with side chain ABCD-rings after loss of side chain and trimethylsilanol Cleavage across C-9,10 and C-5,6 bonds plus –OTMS [M-90+115]– loss of side chain and trimethylsilanol

271 271 279-282 cluster 281 281 283 283 284 285

Loss of side chain and migration of hydrogens

331

ABCD-rings and C-20,21 after allylic cleavage ABCD-rings and C-20,21 after allylic cleavage ABCD-rings and C-20,21 after allylic cleavage ABCD-rings and C-20,21 after allylic cleavage Loss C-22 to C-24 and -OTMS with hydrogen transfer B-ring ission yielding A-ring + C-6 and two -OTMS groups B-ring cleavage of C-5,6 and C-9,10 bonds Retro-Diels–Alder after loss of A-ring -OTMS Retro-Diels–Alder after loss of A-ring -OTMS 70 Da loss from m/z 386 Retro-Diels–Alder after loss of A-ring –OTMS D-ring cleavage with C5 side chain and methyl ester Formed by cleavage across C-5,6, C-8,9 and C-12,13 Fragment after loss of substituents, side chain and C-17 BCD-ring and side chain after loss of 3β-hydroxy-Δ5

343

Loss of side chain and two trimethylsilanol groups

345

ABCD-ring after loss of side chain and trimethylsilanol ABCD-rings plus side chain after loss of 3 -OTMS groups

292/293 314 316 316 318 321 323 329

368

301

Comments on Probable Structure Feature of C-22 hydroxy C27 bile acid Me-TMS ether of glucuronide with m/z 204and m/z 317 Indicative of 7-hydroxy-3-oxo-Δ4 structure Observed in many 3-oxo-6- or 3-oxo-7-hydroxy structures Typical for 12-oxo-3-hydroxy structures, can be base peak Feature of 3,7-dihydroxy, intense in 2-hydroxy structures Feature of 15-hydroxylated cholanoates Formed when there is Δ6 structure, for example, 3,12-dihydroxy-Δ6 Formed when there is Δ6 structure, for example, 3-hydroxy-Δ6 Seen in dihydroxy-cholanotes, and 3β-hydroxy-Δ5 structure Usually indicative of tetrahydroxy structure Trihydroxy-, or dihydroxy with oxo- group, or Δ bond Dihydroxy-, or monohydroxy- with oxo-group, or Δ bond Monohydroxy structure Observed in 7,12-dihydroxy structures Distinguishes 3,7- from 3,6- or 3,12-dihydroxy structures Feature of C24, C27 acids and sterols with 3-oxo-Δ4 structure Feature of C-2 hydroxy structures Intense in 3-oxo bile acids with a C-6, C-7 or C-12hydroxyl Observed in Δ24 and in Δ25 trihydroxy-C27 acids Feature of 24-hydroxy-C27 tetrahydroxy-bile acid Base peaks in Δ23 in C27 trihydroxy-bile acid Feature of 24-hydroxy-C27 trihydroxy-bile acid Base peaks in Δ23 in C27 trihydroxy-bile acid Prominent in 22-hydroxy-CDCA Characteristic of 3,6,7-trihydroxy structures Distinguishes 7-oxo-3-hydroxy- from 3-oxo-7-hydroxyConirms two -OTMS groups in BCD-rings Observed in dihydroxy-cholanoates Low intensity ion in 3-oxo-monohydroxy structures Observed in monohydroxy-cholanoates Characteristic of 12-oxo-3-hydroxy-C24, with m/z 231 ion Observed in 3,6-dihydroxy structures Inluenced by C-16 hydroxy structure, also in C-21 hydroxy Intense in 3β-hydroxy-5-cholenoate, not with C-7 -OTMS In all trihydroxy-cholanoates, but m/z 343 > 368 with a 5α-H In all dihydroxy-cholanoates, but m/z 345 > 370 with a 5α-H Characteristic fragment in all trihydroxy-C24 bile acids

(Continued)

302

MASS SPECTROMETRY FOR STEROID ANALYSIS

TABLE 14.1 Observed Ion (m/z) 370 372 373 386 405 407 410 412 412

(Continued) Origin ABCD-rings plus side chain after loss of 2 -OTMS groups ABCD-rings plus side chain after loss of trimethylsilanol [M-129] from loss of C-3 to C-7 with trimethylsilanol Loss of trimethylsilanol from molecular ion at m/z 476 Fragmentation across C-1 and C-4 with loss of -OTMS Intact glucuronyl methyl ester-persilyl moiety ABCD-ring and side chain after loss of 3 trimethylsilanols ABCD-ring and side chain after loss of 2 trimethylsilanols [M-90] from m/z 502 due to loss in trimethylsilanol

456 [M-90] from m/z 546 due to loss in trimethylsilanol 458 [M-90] from m/z 548 due to loss in trimethylsilanol 460 [M-90] loss in dihydroxy-cholanoates 461 Loss of –CH3 from molecular ion, m/z 476 500 [M-90] from m/z 590 due to loss in trimethylsilanol Observed loss (daltons) −15 Loss of −CH3 from trimethylsilanol or angular methyl Loss of H2O −18 −31 −59 −70 −89

Loss of methoxy group Loss of a carboxymethyl group Retro-Diels–Alder elimination of A-ring with 3-oxo group Loss of trimethylsilanol without hydrogen transfer Loss of trimethylsilanol group, or n groups

−90 or [−90]n −101 −103 −121/122

Loss of C4 side chain methyl ester Loss of primary trimethylsilanol Loss of trimethylsilanol and methoxy of methyl ester

−113

Loss of C5 side chain methyl ester with Δ bond

−115 −129 −129 −155 −155 −157 −229 −245 −245

Loss of C5 side chain methyl ester Loss of C6 side chain methyl ester Loss of C-1 to C-3 with trimethylsilanol D-ring cleavage with C5 side chain methyl ester Loss of C8 side chain methyl ester with a double-bond Loss of C8 side chain methyl ester Loss of C10 dicarboxylic acid side chain Loss of side chain, C-15 and trimethylsilanol Loss of C8 side chain with trimethylsilanol substituent

Comments on Probable Structure Characteristic fragment in all dihydroxy-C24 bile acids Prominent in monohydroxy-C24 bile acids, especially if 5β-H Base peak in 3β-hydroxy-5-cholestanoate Characteristic of oxo-monohydroxy-C24 structure Feature of C-6 hydroxy bile acids Low intensity in bile acid glucuronides Prominent ion in THCA Prominent feature of DHCA Prominent ion in 3β-hydroxy-Δ5 C27 monohydroxy structure Base peak in 7α-hydroxy-cholesterol Base peak in 3β,7α-dihydroxy-Δ5-C24 Intense ion in 3,7β-dihydroxy structure Base peak in 6-oxo-lithocholic acid Base peak in 3β,7α-dihydroxy-5-cholestanoate Often highest mass in the absence of molecular ion Seen throughout fragmentation sequence of oxo bile acids Observed in methyl ethers and in C29-dicarboxylic acid Seen as [M-59] in C-25 hydroxylated cholestanoic acids Characteristic loss in 3-oxo bile acids Indicates vicinal hydroxyls, for example, 2,3- 3,4- or 6,7-hydroxy Indicates presence and number of derivatized hydroxyls Indicates norcholanoate(C23) acid structure May indicate C-18, C-19, C-21, or C-26 hydroxyl Diagnostically signiicant for 7-hydroxy-3-oxo-Δ4 structure Unsaturation in side chain or presence of lost substituent Conirms cholanoic acid (C24) structure Conirms homocholanoic acid (C25) structure Characteristic loss in 3β-hydroxy-Δ5 structures Characteristic of 12-oxo cholanoates Indicates cholestanoic side chain with Δ bond or -OTMS Conirms cholestanoic acid (C27) structure Characteristic of C29-dicarboxylic side chain structure Establishes C-ring substitution Observed in C-24 hydroxy cholestanoic acids

Table prepared in collaboration with Professor Jan Sjövall, Karolinska Institutet, Stockholm, Sweden. Δx, double bond at carbon number x; UDCA, ursodeoxycholic acid (5β-BA-3α,7β-diol); CDCA, chenodeoxycholic acid (5β-BA-3α,7α-diol); THCA, trihydrocholestanoic acid (5β-CA-3α,7α,12α-triol); DHCA, dihydroxycholestanoic acid (5β-CA-3α,7α-diol).

GAS CHROMATOGRAPHY–MASS SPECTROMETRY

303

TABLE 14.2 Common Fragment Ions and Losses Observed in the Mass Spectra of Trimethylsilyl and MO Derivatives of C19 and C21 Steroids Fragment Ion or Loss −29 −30 −43 −44 −46 −47 −56 −59 −71 −85,86 −86 87 −99–86 100,87,70 103, −103 116,117 116 or 117 117 −117 −116 124 125 125, 137, 153 126 129, −129 −131 −131 133 138 142, 143 −142 143 −143 −144 −145 147 −147 −152 156, 184, 199 156, 188 156, 158, 188 157, 159, 172, 186 158 158,174 −159 161 169 169, 182 170, 201 −171 172

Fragment Composition or Origin

Structures Giving Designated Ions

C2H5 CH2O (C-18 or -19) Side chain C11,12 D-ring A-ring C3H4O (C-1,2,3) Side chain D-ring – Side chain D-ring – – Side chain D-ring C-16,-17 Side chain D-ring C-16,-17 CH2OTMS – – Side chain – C2H3OTMS C8H12O – A-ring Side chain D-ring C3H4OTMS – Side chain – A-ring, C-19 C4H5(6) OTMS – – – – – (CH3)2SiOTMS – – – Side chain D-ring Side chain D-ring Side chain D-ring

3,6-OTMS; 3-OTMS-6-one 19-OTMS-3-one; 18-OTMS-17-one 20-one 15-or 20-OTMS-11-one 11- or 16-OTMS-17-MO 6-OTMS-4-ene-3-MO (C19C21) 3-OTMS-5-ene-17-, 11- or 20-one (C21 ,C19); 6-OTMS-4-ene-3-one (C21) 7- or 8-ene-20-MO 4-ene-3-one-16-OTMS 20-one 15-OTMS-17-MO 17-MO (C19) 20-MO 20-MO 18-, 19-, or 21-OTMS (C19, C21) 17-ol-20,21-OTMS; 4-ene-3-one-16-OTMS 15- or 16-OTMS-17-one 20-OTMS 17,20-OTMS 1- or 2-OTMS-4-ene-3-one 4-ene-3-one 3,11-OTMS-17-MO 4-ene-3-MO 7-or 8-ene-20-MO 5-ene-3-OTMS; 17-OTMS(C19); 2,3-OTMS; some 3,6-OTMS; 3-OTMS-11,17-one 7,17-OTMS or 18-OTMS-17-one(C19); 15,16-OTMS-17-one; 12,17-OTMS(C19) 21-OTMS-20-one 1-OTMS-4-ene-3-MO; 15- or 16-OTMS-17- MO(C19); 17-OTMS-16-MO(C19) 5-3,6-MO 4-ene-3-OTMS; 3-enol-TMS;2,3-OTMS; 1,3-OTMS 4-ene-3-OTMS 3,11-OTMS; 7-or 8-ene-3-OTMS; 15- or 16-OTMS-17-one 4-ene-3,6-OTMS; 6-OTMS-3-one 3-OTMS-11-one (C21); 15 or 16-OTMS-17-one; 17-OTMS-16-one (C19) 3,6()-OTMS Di- and poly-OTMS (vicinal) 15-OTMS-17-MO 18,21-OTMS-20-MO 3,11-OTMS-17-one 16-OTMS-20-MO 17-OTMS-20-MO (C21) 16-OTMS-20-one

D-ring D-ring Side chain C16,17 – – – Side chain D-ring D-ring D-ring

18-or19-OTMS-17-one(C19); 17-OTMS-16-MO(C19) 16-OTMS-17-MO (C19) 15,16-OTMS-20-one 3,6-OTMS(C21); 11,21-OTMS 18-OTMS-17-one 11,17-OTMS(C19) 15-OTMS-20-MO; 20-OTMS-16-MO 15-OTMS-20-one 15- or 16-OTMS-20 one

(Continued)

304

MASS SPECTROMETRY FOR STEROID ANALYSIS

TABLE 14.2 Fragment Ion or Loss

(Continued) Fragment Composition or Origin

Structures Giving Designated Ions

174, −174

D-ring

174, −174 175, 188 −187 −188 191 −193 196 196,271 205 −205 217,(218, 219) 223 234a 243 −247 258, 289 260 276 276, 246, 244 −307 333

D-ring Side chain D-ring Side chain D-ring Side chain D-ring CH(OTMS)2 D-ring – – – Side chain C3H3(OTMS)2

16-OTMS-17-MO; 17-OTMS-16-MO; 16,18-OTMS-17-MO(C19); 11,21-OTMS20-one; 15,21-OTMS-20-one 21-OTMS-20-MO 21-OTMS-20-MO 15,21-OTMS-20-MO (11,15,16,17,18)-di(tri)-OTMS(C21); C21-poly-OTMS 16,17-OTMS (C19) 16-OTMS-17-one 1,3-OTMS-11-one(C21) 16,17-OTMS(C19) 17,20,21-OTMS 1,3-OTMS; 15,(18),17-OTMS(C19)

C,D rings C9H22O3Si2 C5H5(OTMS)2 Side chain C,D rings Side chain D-ring Side chain D-ring Side chain D-ring Side chain D-ring – Side chain C,D rings

6,17-OTMS(C19) 17,20-OTMS-21-COOMe 17,20,21-OTMS; 3,7-OTMS; 3,5-OTMS 15,17,20-OTMS 15,17-OTMS-20-MO; 15,21-OTMS-20-MO 15,21-OTMS-20-one;16,21-OTMS-20-one 16,21-OTMS-20-MO 17,21-OTMS-20-MO 15,17,20-OTMS 15,17,20-OTMS

Table prepared in collaboration with Professor Jan Sjövall, Karolinska Institutet, Stockholm, Sweden. MO is a methoximated carbonyl. OTMS is a trimethylsilylated hydroxyl. “-one” is an underivatized carbonyl. Common fragmentations -90,-15 and -31 not included. Unless speciied otherwise, it is generally assumed that all steroids have a 3-OTMS. a Base peak in methylesters of cortoic acids. Fragment formed by McLafferty rearrangement.

Here we mention just a few of the more important structurally diagnostic fragment ions observed in the EI spectra of steroids. The intact ABCD ring system can provide a diagnostic ion (Tables 14.1 and 14.2, Figures 14.3 and 14.4). The m/z of this ion is determined by loss of all the hydroxyls or other substituents with an addition of 14 Da per additional methyl or oxo group. The A-ring 3-oxo-4-ene function will give a fragment ion at m/z 124, which is elevated to 153 in the MO, while the TMSO of a 3β-hydroxy-5-ene will give an ion at m/z 129 or a [M-129]+ ion. The additional presence of an allylic 7-TMSO group is often characterized by an abundant [M-90]+ ion. In bile acids esters and bile alcohols-bearing TMSO groups on the side chain, dominant cleavages are α to the TMSO group; this leads to ions at m/z 173 in C-22substituted bile alcohols, 145 and [M-43]+ in C-24 bile alcohols, and an ion at m/z 131 in C-25-substituted bile alcohols.

14.4 ESI-MS AND ELECTROSPRAY IONIZATION TANDEM MASS SPECTROMETRY (ESI-MS/MS) ESI-MS was irst used for steroid analysis in the late 1980s [30,31]; however, a decade earlier, Shackleton and colleagues exploited a forerunner of this soft ionization method (i.e., FAB), for steroid analysis [12,13]. FAB-MS was used extensively for the analysis of bile acids and steroid conjugates in urine and plasma [12,13], and was very effective for the diagnosis of inborn errors of bile acid biosynthesis [32,33]. Urine and plasma samples could be analyzed directly following RP-SPE. Today, we are using direct-inlet ESI (shotgun steroidomics) rather than FAB and are able to exploit the high resolution and exact mass capabilities of modern Fourier transform (FT) MS instruments (e.g., FT-ion cyclotron resonance and FT-Orbitrap) to partially characterize steroid conjugates and bile acids (Figure 14.5). Bile acids and

FIGURE 14.3 Some important steroid fragment ions observed in EI spectra.

Average of 17.629−17.783 min.: SWAGRI022-SG-GCEI-2.D/data.ms 372.4

Abundance 20,000

215.2

75.0

18,000 16,000 14,000 12,000 10,000 55.0 8000

257.3 93.0

6000 161.2 121.1

4000

145.1

2000

188.2

39.0 m/z

0

20

40 60

231.2

267.3 299.3

323.3 344.3

447.4 389.1415.1431.3 463.4 490.4 511.5.528.5

80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 1: TOF MS AP+ 462.352 0.10 Da 4.51e4

4568_1035 10.91

100

%

B

0 8.00 8.25 8.50

8.75

9.00 9.25 9.50

9.75 10.00 10.25 10.50 10.75 11.00 11.25 11.50 11.75 12.00 12.25 12.50 12.75 13.00 C27H47O3Si (−0.2 mDa, −0.4 ppm) 372.294

100

C28H50O3Si (−3.1 mDa, −6.7 ppm)

%

462.350 373.309

MS 159.085

0

463.361 447.329 257.229 230.204 371.280 374.318 464.360 258.234 341.282 216.182 375.321431.329 299.278

m/z 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 600 4568_1035 695 (10.922) Cm (695-(693+699)) 2: TOF MS AP+ 215.154 106.063 7.13e3 100 147.109 MSE 159.102 133.096 95.082 161.124 %

173.128 175.143 67.054

0 60

216.166 257.215 229.187 230.199

357.253

372.285 323.261 373.303 258.236 301.218 341.283 405.287 447.329 476.351190.341 535.401

m/z 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 600

FIGURE 14.4 Lithocholic acid methyl ester TMS ether. (A) EI-MS. The [M]+ ion is at 462, [M-CH3]+ at 447, [M-TMSOH]+ at 372, [M-CH3-TMSO]+ at 357, ABCD+ ion at 257, and the ABC+ ion at 215. (B) GC-APCI-MS proile. (C) Upper trace GC-APCIMS and lower trace GC-APCI-MSE. The spectrum in (A) was recorded on a single quadrupole instrument, while (B) and (C) were recorded on a quadrupole time-of-light (Q-TOF)-type instrument. The MSE spectrum was recorded by raising the collision energy across the collision cell but without mass selection.

Spectrum Peak Measured (m/z)

Elemental Composition

5A/5B 5A 5A

238.07792− 246.07552− 465.2489−

C21H32O8S22− C21H32O9S22− C25H37O8−

5A

481.2430−

C25H37O9−

5A

525.2692−

C27H41O10−

5A 5A/5C

539.2496− 541.2646−

C27H39O11− C27H41O11−

5D 5D 5D 5D 5D 5D/E 5D 5D 5D 5D

254.62292− 262.62022− 270.61762− 279.61412− 287.61162− 295.60872− 469.2260− 485.2214− 512.2682 527.2693

C26H39NO7S2− C26H39NO8S2− C26H39NO9S2− C26H41NO8S22− C26H41NO9S22− C26H41NO10S22− C24H37O7S− C24H37O8S− C26H42NO7S− C27H43O8S−

5A 5A

255.2332 283.2643

C16H31O2− C18H35O2−

Suggested Structure (See Appendix 1 and 2; Other Isomers Are Possible) Pregn-5-ene-3β,20α-diol (P5-3β,20S-diol) disulfate Pregn-5-ene-3β,17α,20-triol (P5-3β,17α,20-triol) disulfate Androsterone (5α-A-3α-ol-17-one) and etiocholanolone (5β-A-3α-ol-17-one) glucuronide 11β-Hydroxyandrosterone (5α-A-3α,11β-diol-17-one) and 11βhydroxyetiocholanolone (5β-A-3α,11β-diol-17-one) glucuronide Tetrahydro-11-deoxycortisol (5β-P-3α,17α,21-triol-20-one) and 5α/βtetrahydrocorticosterone (P-3α,11β,21-triol-20-one) glucuronide Tetrahydrocortisone (5β-P-3α,17α,21-triol-11,20-dione) glucuronide 5α/β-Tetrahydrocortisol (P-3α,11β,17α,21-tetrol-20-one) and 5α/β-cortolone (P-3α,17α,20,21-tetrol-11-one) glucuronide 3β-Hydroxychol-5-enoylglycine (BA5-3β-ol, 24-G) 3-sulfate 3β,7α-Dihydroxychol-5-enoylglycine (BA5-3β,7α-diol, 24-G) 3-sulfate 3β,7α,12α-Trihydroxychol-5-enoylglycine (BA5-3β,7α,12α-triol, 24-G) 3-sulfate 3β-Hydroxychol-5-enoyltaurine (BA5-3β-ol, 24-T) 3-sulfate 3β,7α-Dihydroxychol-5-enoyltaurine (BA5-3β,7α-diol, 24-T) 3-sulfate 3β,7α,12α-Trihydroxychol-5-enoyltaurine (BA5-3β,7α,12α-triol, 24-T) 3-sulfate 3β,7α-Dihydroxychol-5-enoic acid (BA5-3β,7α-diol) 3-sulfate 3β,7α,12α-Trihydroxychol-5-enoic acid (BA5-3β,7α,12α-triol) 3-sulfate 3β,7α,12α-Trihydroxychol-5-enoyltaurine (BA5-3β,7α,12α-triol, 24-T) 3β,7α,12α-Trihydroxycholest-5-en-26-oic acid (CA5-3β,7α,12α-triol) 3-sulfate Fatty acids Palmitic acid Stearic acid

G, glycine; T, taurine.

FIGURE 14.5 Direct-inlet negative ion ESI-MS and ESI-MS/MS spectra of steroids and bile acids in urine (i.e., shotgun steroidomics). (A) ESI-MS spectrum of pooled “control” urine. The inset shows the m/z region populated by doubly charged steroid disulfates. (B,C) ESI-MS/MS of the peaks at 238.082− and 541.26− from a urine sample with elevated steroid levels. (D) ESI-MS spectrum of urine from an infant with neonatal hepatitis of unknown etiology. The pattern of bile acid precursors suggests an inborn error of bile acid biosynthesis. (E) ESI-MS/MS of the peak at 295.612− from (D). Spectra were recorded on an LTQ-Orbitrap XL, (Thermo Fisher Scientiic, San Jose, CA). ESI-MS scans utilize the Orbitrap-FT analyzer for ion detection, which gives mass accuracy of better than 5 ppm. Suggested structures based on measured mass are given above. Steroid and bile acid peaks are indicated by an asterisk in ESI-MS spectra. From the measured mass values, elemental compositions can be determined but isomers can’t be differentiated. MS/MS spectra are likely to represent composite spectra of multiple isomers. The MS/MS spectra in (C) and (E) were obtained with the Orbitrap-FT analyzer and the elemental composition of fragment ions were determined. The spectrum in (B) was recorded on the LTQ ion trap analyzer giving unit mass accuracy. %RA, % relative abundance.

308

MASS SPECTROMETRY FOR STEROID ANALYSIS





− − −





− − − −

− −

FIGURE 14.5 (Continued)

ESI-MS AND ELECTROSPRAY IONIZATION TANDEM MASS SPECTROMETRY (ESI-MS/MS)









FIGURE 14.5 (Continued)

309

310

MASS SPECTROMETRY FOR STEROID ANALYSIS

steroid sulfates and glucuronides are best analyzed in the negative-ion ESI mode, where they give [M-H]− ions. Structural information can then be obtained by performing MS/MS (Figure 14.5). However, steroids and bile acids exist in biological samples as part of complex mixtures and an [M-H]− ion is likely to represent a number of isomers. These can, however, be separated by LC prior to ESI [27] and partially characterized by MS/MS [34]. Bile acids, alcohols, and side-chain-shortened steroids have been extensively analyzed by liquid chromatography electrospray ionization mass spectrometry (LCESI-MS) [1–3]. The mobile phase usually features water/ methanol or water/acetonitrile containing an organic acid or buffer. As ESI is a concentration-dependent process, maximum sensitivity is obtained with highconcentration, low-volume samples run at low low rate [26,27]. We have extensively used low-low-rate LCESI-MS to achieve maximum sensitivity for steroid analysis [26], and with the advent of reliable nanoscale liquid chromatography (nano-LC) equipment, this form

TABLE 14.3

Characteristic ESI-MS/MS Fragmentations of Bile Acid and Steroid Conjugates

Bile acid

Conjugating Group Sulfate

GlcA

Bile acid [M-H]−

–80 (SO3); 97− (HSO4)−

–176 (C6O6H8)

Steroid/sterol [M-H]−

–80 (SO3); 80− (SO3)−; 97− (HSO4)− –80 (SO3); −98 (H2SO4)

–176 (C6O6H8)

Steroid/sterol [M+H]+

of analysis may become more popular in the future. However, nano-LC is not ideal for fast chromatography, which may be more important for metabolomic studies. A more appropriate approach may be the use of ultra performance liquid chromatography (UPLC) or ultrahigh performance liquid chromatography (UHPLC), which lends well to fast chromatography, giving sharp chromatographic peaks [35,36]. See the section below on LC-MS and MS/MS or MS2 protocols for a full description of LC-ESI-MS methods. Electrospray ionization tandem mass spectrometry (ESI-MS/MS) has been extensively exploited in steroid analysis, particularly with methods that feature multiple reaction monitoring (MRM) and precursor and neutral loss scan modes. The MRM methods are made possible by characteristic fragmentations, especially for steroid and bile acid conjugates. Some characteristic fragmentations are listed in Table 14.3. Most fragmentations are derived by loss of the conjugating group, but in the case of 3-oxo-4-enes fragmentation, is also observed in the A and B rings (Table 14.4 and Figure 14.6) [37].

Glc

GlcNAc

Taurine

Glycine

–162 (C6O5H10)

–203 (C8O5NH13)

–125 (NH2C2H4SO3H); 80− (SO3)−; 107− (CH2CHSO3)−; 124− (NH2C2H4SO3)−

–44 (CO2); −62 (H2CO3);74− (NH2CH2CO2)−

–176 (C6O6H8) −194 (C6O7H10)

GlcA, glucuronic acid; Glc, glucose; GlcNAc, N-acetylglucosamine.

TABLE 14.4 Characteristic Fragment Ions Observed in the CID Spectra of [M+H]+ and [M]+ Ions from Protonated and Derivatized 3-oxo-4-ene (3-oxo-Δ4) Steroids Structure 4

3-Oxo-Δ -steroid 3-Oxo-Δ4-steroid 3-oxime 3-Oxo-Δ4-steroid 3-GT and 3-GP hydrazones a

Precursor Ion +

[M+H] [M+H]+ [M]+ ([M-59]+/[M-79]+)b; [M]+ ([M-87]+/[M-107]+)c

Fragment Ion 97 (b1-12) 112 (b1-12) 151b 123c (b1-12)

109a (b3-C2H4) 124a (b3-C2H4) 163a,b 135a,c (b3-C2H4)

123 (b2) 138 (b2) 177b 149c(b2)

Fragment ion elevated by 16 Da when additional hydroxylation at position 7. See Figure 14.6. Fragment ions observed in MS/MS spectra recorded on Q-TOF and tandem quadrupole instruments, and in [M]+, [M-59]+ and [M]+, [M-79]+ MS3 spectra recorded on ion traps from GT and GP hydrazones, respectively. c Fragment ions observed in MS/MS spectra recorded on Q-TOF and tandem quadrupole instruments, and in [M]+, [M-87]+ and [M]+, [M-107]+ MS3 spectra recorded on ion traps from GT and GP hydrazones, respectively. b

FIGURE 14.6 MS/MS fragmentation of 3-oxo-4-ene steroids exempliied by testosterone. (A) MS/MS fragmentation of [M+H]+ ions. Oxime and methyl oxime derivatives give an identical pattern of fragment ions but elevated in mass by 15 (NH) and 29 (NHCH2) Da, respectively. (B) High mass fragment ions seen in MS/MS (MS2) spectra of Girard P (GP) and Girard T (GT) derivatives. (C) Low mass fragment ions observed in MS/MS spectra of GP and GT derivatives recorded on tandem quadrupole and Q-TOF instruments and in MS3 spectra recorded on ion traps (see Table 14.4).

312

MASS SPECTROMETRY FOR STEROID ANALYSIS

Although one of the major advantages of ESI is that derivatization is not prerequisite, it can very often be advantageous. For example Johnson et al. devised a derivatization method to enhance the analysis of C27 bile acids based on acetylation of alcohol groups and conversion of carboxylic acids to dimethylaminoethylesters prior to analysis by positive ion ESI and detection of the [M+H-(CH3CO2H)n]+ fragment by MRM [38] (Figure 14.7). We have used an alternative approach where we derivatize carboxylic acids with an aminosulfonic acid and thereby enhance gas phase acidity and [M-H]− ion current [27,39]. Neutral steroids can also be derivatized to enhance ionization. For example, sterols can be converted to sulfate esters to enhance [M-H]− ion currents [39–41], while oxosteroids can be converted to oximes or methyl oximes, both of which are more basic than the oxo group and thus enhance [M+H]+ ion formation [26,42,43]. An alternative method of derivatization of oxo groups is conversion to charged hydrazones [44,45]. We have also exploited this idea as described below.

14.5 APPLICATION OF MS TO STEROID ANALYSIS 14.5.1

In Brain

The brain is a steroidogenic organ. About 25% of the body cholesterol in mammals is found in brain [46]. As cholesterol will not cross the blood–brain barrier, essentially all brain cholesterol is derived from in situ synthesis from acetyl CoA via the mevalonate pathway [46]. In recent years, C19, C21, and C27 steroids have been found in brain [9,24–26,47–51], and also C27 and C24 bile acids [52,53]. C19 and C21 steroids synthesized in brain are called neurosteroids, while the term “neuroactive steroid” is used for steroids which are active in the nervous system, whether synthesized in the central nervous system (CNS) or not. For example, allopregnanolone and pregnenolone sulfates are neuroactive steroids in that they will interact with the γ-aminobutyric acid receptor A (GABAA) and modulate neuronal transmission [47]. Although once reported to be present in rodent brain [54], pregnenolone sulfate and also dehydroepiandrosterone (DHEA) sulfate are probably not present in detectable quantities [24,26,55]. This inding contrasts with the situation with human brain where these steroid sulfates have been detected [24,55] (see the section “Nano-LC-MS for Steroid Sulfates”). While neurosteroids are present in brain at picograms per milligram levels, oxysterols are present at far higher levels (nanograms per milligram).

Oxysterols are formed in the irst steps of all cholesterol metabolism. The most abundant oxysterol in mammalian brain is 24S-hydroxycholesterol, also known as cerebrosterol. Cerebrosterol is present at a level of 5–50 ng/mg in wet brain and is formed from cholesterol in a cytochrome P450 46A1 (CYP46A1)catalyzed reaction [56]. CYP46A1 is only expressed in nervous tissue. CYP46A1 also has 25- and 26-hydroxylase activity [57] and low levels of both 25- and 26-hydroxycholesterols are found in brain (1–5 ng/mg) [48,58]. Oxysterol 7α-hydroxylase (CYP7B1) is also expressed in nervous tissue [59] and its substrates 25- and 26-hydroxycholesterols are 7αhydroxylated by cells of the nervous system and their 7α-hydroxylated products found to be present in brain itself [49,60]. To deep-mine the brain steroidome, we have devised an analytical methodology suitable for the high sensitivity analysis of steroids with a 3β-hydroxy-5-ene, 3βhydroxy-5α-hydrogen, or oxo-containing function. As a irst step, we perform our steroid extractions in ethanol. The most nonpolar steroids are separated from oxysterols and more polar steroids by SPE, either straightphase SPE using hexane : dichloromethane (2:8, v/v) to load the steroids and ethyl acetate as the eluting solvent for oxysterols after prior elution of less polar sterols by hexane : dichloromethane (2:8, v/v) [49]; or RP-SPE using 70% ethanol to load, and then elute oxysterols and the more polar steroids in the low through, then 99.9% ethanol to subsequently elute cholesterol and more hydrophobic sterols (Figure 14.8; see the section “Tissue [Mouse Brain] for ESI-MS”) [58]. Next after reconstitution of the eluate in propan-2-ol the separate fractions are treated with cholesterol oxidase and then with Girard P reagent (GP) [49,58] (Figure 14.9). The GP reagent contains a pyridine group linked through its quaternary nitrogen group to a hydrazine group via a carboxymethyl function. The hydrazine group reacts with oxo groups which are either present in the initial steroid or introduced by a microchemical reaction catalyzed by cholesterol oxidase (Figure 14.9). Following removal of excess derivatization reagent on an RP-SPE column, the derivatized steroids are suitable for ESI-MS analysis (see the section “3β-Hydroxy-5-ene- (3βhydroxy-5α-hydrogen) and Oxo-Steroids for LC-MS”). Incorporation of the charged GP moiety on the steroid in a process we call “charge-tagging” greatly improves ESI sensitivity (by two to three orders of magnitude) [49]. In addition, the GP tag directs fragmentation in MS/MS, leading to structurally informative spectra. The MS2 and MS3 spectra of some of the GP “chargetagged” steroids identiied in rodent brain are shown in Figure 14.10.

APPLICATION OF MS TO STEROID ANALYSIS

313

FIGURE 14.7 Derivatization for ESI-MS/MS. (A) Preparation of the dimethylaminoethylester of dihydroxycholestanoic acid bis acetate. (B) Preparation of aminoethanesulfonate derivative of chenodeoxycholic acid. (C) Formation of sulfate ester of cholesterol. (D) Preparation of oxime of testosterone.

314

MASS SPECTROMETRY FOR STEROID ANALYSIS

FIGURE 14.8 Sample preparation protocol for the analysis of oxysterols in brain, CSF, or plasma.

.

FIGURE 14.9 Charge-tagging of oxysterols with GP reagent.

APPLICATION OF MS TO STEROID ANALYSIS

)

( 2

315

3

FIGURE 14.10 LC-ESI-MS chromatograms and MS and MS spectra of GP charge-tagged oxysterols isolated from newborn mouse brain. 24S-Hydroxycholesterol is formed from cholesterol in a CYP46A1-catalyzed reaction while 24S,25-epoxycholesterol is formed in parallel to cholesterol via a shunt mechanism of the mevalonate pathway. CYP46A1 also has 27-hydroxylase activity, as does CYP27A1, which is also expressed in brain and either enzyme may be responsible for formation of 26-hydroxycholesterol from cholesterol. 26-Hydroxycholesterol can be metabolized further by CYP27A1 and CYP7B1 to 3β,7α-dihydroxycholest-5-en26-oic acid. (A) Reconstructed ion chromatogram (RIC) of m/z 534.4054, 550.4003, 564.3796, and 564.4160 ± 5 ppm. (B) MS2 534.41→ and MS3 534.41→455.36→ spectra of the peak eluting at 7.41 min and corresponding to GP charge-tagged 24Shydroxycholesterol. Peaks at 7.41 min and 7.71 min in (A) are syn and anti conformers of the GP hydrazone derivative. (C) MS2 534.41→ and MS3 534.41→455.36→ spectra of the peak eluting at 7.90 min and corresponding to GP charge-tagged 26-hydroxycholesterol. (D) MS2 550.40→ and MS3 550.40→471.36→ spectra of the peak eluting at 3.95 min and corresponding to GP charge-tagged 24,25-dihydroxycholesterol. Peaks at 3.95 min and 4.63 min in (A) are syn and anti conformers of the hydrazone derivative. 24,25-Dihydroxycholesterol may be formed enzymatically by CYP46A1 or may be a hydrolysis product of endogenous 24S,25epoxycholesterol. (E) MS2 550.40→ and MS3 550.40→471.36→ spectra of the peak eluting at 4.50 min and corresponding to GP charge-tagged 24,26 or 20,22-dihydroxycholesterol. (F) MS2 564.38→ and MS3 564.38→485.34→ spectra of the peak eluting at 5.95 min and possibly corresponding to GP charge-tagged 3β,7α-dihydroxycholest-5-en-26-oic acid. The level of this acid is an order of magnitude lower than the oxysterols. (G) MS2 564.42→ and MS3 564.42→485.37→ spectra of the peak eluting at 6.17 min and corresponding to GP charge-tagged hydroxymethoxycholesterol derived from methanolysis of 24S,25-epoxycholesterol during GP derivatization in acidic methanol solution. Peaks at 6.17 min and 6.61 min in (A) are syn and anti conformers of the hydrazone derivative. The equivalent of 60 µg of brain was injected on column in 20 µL of 60% methanol, 0.1% formic acid. Chromatographic separation of GP-tagged steroids was achieved using a Hypersil Gold RP column (1.9 µm particles, 50 × 2.1 mm, Thermo Scientiic). Mobile phase A was 33.3% methanol, 16.7% acetonitrile containing 0.1% formic acid, and mobile phase B was 63.3% methanol, 31.7% acetonitrile containing 0.1% formic acid. After 1 min at 20% B, the proportion of B was raised to 80% B over the next 7 min and maintained at 80% B for a further 5 min before returning to 20% B in 6 s and reequilibrating for a further 3 min 54 s. The low rate was 200 µL/min. The eluent was transferred to the ESI source of an LTQ-Orbitrap XL, hybrid linear ion trap (LIT) FT-Orbitrap mass spectrometer (Thermo Scientiic). Mass spectra were recorded at 30,000 (full width at half maximum height, FWHM) resolution in the Orbitrap analyzer over the m/z range 400–610 or 300–800. The maximum ion ill time was 500 ms and MS2 and MS3 events were performed in the LIT with ill times of 200 ms simultaneously to acquisition of the mass scan in the Orbitrap. See also the protocol “LC-MS for GP-Derivatized Steroids.”

316

MASS SPECTROMETRY FOR STEROID ANALYSIS

FIGURE 14.10 (Continued)

APPLICATION OF MS TO STEROID ANALYSIS

FIGURE 14.10 (Continued)

317

318

MASS SPECTROMETRY FOR STEROID ANALYSIS

FIGURE 14.10 (Continued)

NEW APPLICATIONS

24S-Hydroxycholesterols represent an initial metabolite in one of the pathways of bile acid biosynthesis. Although bile acid biosynthesis predominantly occurs in the liver, this irst step in the 24S-hydroxylase pathway is extrahepatic [61]. Interestingly, Ferdinandusse et al. [52] and Mano et al. [53] have been able to identify C27 and C24 bile acids in brain by negative ion ESI-MS. Like Mano et al., we have also been able to identify cholic acid in rodent brain [62]. Mano et al. identiied both cholic and deoxycholic acids in rat brain cytoplasmic extracts; however, when these extracts were treated with guanidine to reduce protein noncovalent interactions, high levels of chenodeoxycholic acid were also found.

14.5.2

In Plasma

Steroids, bile acids, and bile alcohols can be identiied in plasma following a simple RP-SPE step (see the section “Plasma for ESI-MS”). The steroid levels may, however, be low (from picograms per milliliter to micrograms to milliliter), although cholesterol is present at a level of 2 mg/mL, and further separation by LC may be required prior to MS analysis. Nevertheless, steroid sulfates can be observed in simple shotgun negative ion ESI-MS analysis of plasma (Figure 14.11). By application of MS/MS, some structural information is also forthcoming, but it is unclear if the [M-H]− signal represent the sum of numerous isomers or a single component. LC separation is required to resolve this question. In cases of hepatobillary disease, including inborn errors of bile acid biosynthesis, the levels of bile acid precursors in plasma may be greatly elevated [27,32,33]. The elevated levels of bile acids in plasma may allow a rapid diagnosis of the defective enzyme by direct-inlet ESI-MS as an elevated abundance of the metabolites prior to the defective enzyme will be found in plasma (Figure 14.11). Oxysterols in plasma are usually analyzed by GC-MS. Dzeletovic et al. developed a classical methodology which is still widely adopted today [63]. Oxysterols are found in plasma mostly esteriied with fatty acids but also with sulfuric acid and as an acetal with glucuronic acid [64,65]. The protocol developed by Dzeletocvic et al. focused on fatty acids esters, so an initial base hydrolysis step was performed with KOH in ethanol. Following neutralization with phosphoric acid and a Folch-like extraction into CHCl3, oxysterols were separated from other less polar sterols on a silica cartridge. Derivatization was performed with pyridine : hexamethyldisilazane : trimethylchlorosilane 3:2:1 (v/v/v) and analysis of the TMS ethers by GC-EI-MS. Nine oxysterols were identiied in plasma using this protocol includ-

319

ing 24-, 25-, 26-, and 7α-hydroxycholesterols and cholesterol-5,6-epoxides. Lütjohann et al. modiied this method to include solvolysis of oxysterol sulfates [65]. Solvolysis was performed following the base hydrolysis step by acidiication with mineral acid. For 24hydroxycholesterol, Lütjohann et al. found about 10% of the total oxysterol to be esteriied with sulfuric acid, while about 70% was esteriied with fatty acids. Oxysterols can also be glucuronidated, and Sjövall and colleagues found 24S-hydroxycholesterol by liquid chromatography electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS) to be both sulfated and glucuronidated [64]. We also have also found oxysterols in plasma from a child with neonatal hepatitis to be both sulfated and glucuronidated (Figure 14.11). Cholestenoic acids are also found in plasma. Using a GC-MS protocol similar to that used by Dzeletovic et al., but also including methylester formation, Axelson et al. identiied 3β-hydroxycholest-5-en-26-oic, 3β,7α-dihydroxycholest-5-en-26-oic, and 7α-hydroxy-3oxocholest-4-en-26-oic acids in plasma in the 35–85 ng/ mL range [66]. We have also identiied these acids and their oxysterol precursors in plasma using a chargetagging LC-ESI-MSn approach (see the section “LC-MS for GP-Derivatized Steroids”) [67,68]. C24 bile acids are also present in plasma [66]. These bile acids have been extensively studied by GC-MS [1,2] and are currently being studied using LC-MS technologies [35].

14.6

NEW APPLICATIONS

One area that is not normally covered in general steroid studies is secosterol analysis. When cholesterol is reacted with either ozone or singlet oxygen, it will undergo ring opening to give a secosterol that can then recyclize to give a ive-membered B ring via an aldol condensation reaction (Figure 14.12) [69,70]. Both the secosterol and its aldol have been implicated with neurodegenerative diseases, perhaps through posttranslationally modifying lysine residues and encouraging amyloidogenesis [71,72]. Using our charge-tagging methodology, we have been able to identify the secosterol and aldol both in brain and plasma [49,67]. Vitamins D2 and D3 are both secosterols formed via UV-induced ring opening of ergosterol and 7-dehydrocholesterol, respectively. They are metabolized in liver to 25-hydroxyvitamins D2 and D3, which are the metabolites found in plasma [73]. We have been able to derivatize the secosterols and enhance their analysis by ESI-MSn by treating these metabolites with cholesterol oxidase and the GP reagent (Figure 14.13).

320

MASS SPECTROMETRY FOR STEROID ANALYSIS

Spectrum

Peak Measured (m/z)

11A/B/C 11A/B

367.1587− 369.1756−

C19H27O5S− C19H29O5S−

11A/B

383.1531−

C19H27O6S−

11A/B 11A/B 11A/B 11D 11D 11D/F 11D

395.1860− 397.2031− 413.2016− 254.62242− 279.61392− 280.12362− 287.61082−

C21H31O5S− C21H33O5S− C21H33O6S− C26H39NO7S2− C26H41NO8S22− C27H44O8S22− C26H41NO9S22−

11D 11D 11D 11D/E 11D

288.12082− 296.11812− 328.16102− 336.15852− 344.15552−

C27H44O9S22− C27H44O10S22− C33H52O11S2− C33H52O12S2− C33H52O13S2−

11D 11D 11D 11D 11D

465.3027− 481.2973− 497.2923− 513.2873− 527.2675−

C27H45O4S− C27H45O5S− C27H45O6S− C27H45O7S− C27H43O8S−

11D 11D/E 11D

657.3278−a 673.3198−a 689.3035−a

C33H53O11S− C33H53O12S− C33H53O13S−

A/D A/D

255.2323 283.2643

C16H31O2− C18H35O2-

a

Elemental Composition

Suggested Structure (See Appendices 2–4; Other Isomers are Possible) Dehydroepiandrosterone (A5-3β-ol-17-one) 3-sulfate Androsterone (5α-A-3α-ol-17-one), epiandrosterone (5α-A-3β-ol-17-one), androst-5-ene-3β,17β-diol (A5-3β,17βdiol) sulfate 7α-Hydroxydehydroepiandrosterone (A5-3β,7α-diol), 16α-hydroxydehydroepiandrosterone (A5-3β,16α-diol) sulfate Pregnenolone (P5-3β-ol-20-one) 3-sulfate Pregn-5-ene-3β,20α-diol (P5-3β,20S-diol) sulfate Pregn-5-ene-3β,17α,20α-triol (P5-3β,17α,20S-triol) sulfate 3β-Hydroxychol-5-enoylglycine (BA5-3β-ol, 24-G) 3-sulfate 3β-Hydroxychol-5-enoyltaurine (BA5-3β-ol, 24-T) 3-sulfate Cholest-5-ene-3β,x-diol (C5-3β,x-diol) disulfate 3β,7α-Dihydroxychol-5-enoyltaurine (BA5-3β,7α-diol, 24-T) 3-sulfate Cholest-5-ene-3β,x,y-triol (C5-3β,x,y-triol) disulfate Cholest-5-ene-3β,x,y,z-tetrol (C5-3β,x,y,z-tetrol) disulfate Cholest-5-en-3β,x-diol (C5-3β,x-diol) sulfate, glucuronide Cholest-5-en-3β,x,y-triol (C5-3β,x,y-triol) sulfate, glucuronide Cholest-5-en-3β,x,y,z-tetrol (C5-3β,x,y,z-tetrol) sulfate, glucuronide Cholesterol (C5-3β-ol) sulfate Cholest-5-en-3β,x-diol (C5-3β,x-diol) sulfate Cholest-5-en-3β,x,y-triol (C5-3β,x,y-triol) sulfate Cholest-5-en-3β,x,y,z-tetrol (C5-3β,x,y,z-tetrol) sulfate 3β,x,y-Trihydroxycholest-5-en-26-oic acid (CA5-3β,x,y-triol) sulfate Cholest-5-en-3β,x-diol (C5-3β,x-diol) sulfate, glucuronide Cholest-5-en-3β,x,y-triol (C5-3β,x,y-triol) sulfate, glucuronide Cholest-5-en-3β,x,y,z-tetrol (C5-3β,x,y,z-tetrol) sulfate, glucuronide Fatty acids Palmitic acid Stearic acid

Low abundance ions measured with reduced mass accuracy.

FIGURE 14.11 Direct inlet negative ion ESI spectra of plasma from a healthy child and from a child with neonatal hepatitis of unknown etiology. (A) ESI-MS spectrum of plasma from a healthy child. (B) m/z region occupied by C19 and C21 steroid sulfates. (C) ESI-MS/MS of the peak at 367.16− from spectrum (A). (D) ESI-MS spectrum of plasma from an infant with neonatal hepatitis of unknown etiology. (E) ESI-MS/MS of the peaks at 673.32− and 336.162− from spectrum (D). (F) ESI-MS/MS of the peak at 280.122− from spectrum D. Spectra were recorded on an LTQ-Orbitrap XL. ESI-MS utilized the Orbitrap-FT analyzer for ion detection which gives mass accuracy of better than 5 ppm. Suggested structures based on measured mass are given above. Steroid peaks are indicated by an asterisk. From the measured mass values elemental compositions can be determined but isomers can’t be differentiated. MS/MS spectra are likely to represent composite spectra of multiple isomers. The MS/MS spectrum in (C) was recorded using the ion trap in PQD (pulsed Q collision-induced dissociation) mode with ion trap detection. Spectra in (E) and (F) utilized the Orbitrap-FT analyzer, allowing determination of elemental composition of fragment ions.

NEW APPLICATIONS

FIGURE 14.11 (Continued)

321

322

MASS SPECTROMETRY FOR STEROID ANALYSIS







FIGURE 14.11 (Continued)

NEW APPLICATIONS



− −





− − −







FIGURE 14.11 (Continued)

323

324

MASS SPECTROMETRY FOR STEROID ANALYSIS

(

)

FIGURE 14.12 Cholesterol is oxidized by ozone (O3) or singlet molecular oxygen (1Δq) to 3β-hydroxy-5-oxo-5,6-secocholestan6-al, which isomerizes to 3β,5β-dihydroxy-B-norcholestane-6β-carboxaldehyde [69,70]. Both compounds are identiied in rodent brain homogenate by LC-ESI-MS/MS following charge-tagging with GP reagent [49]. (A) Cholesterol oxidation with O3 or O2 (1Δg). (B) LC-ESI-MS RIC of the GP charge-tagged aldehydes (m/z 552.4160 ± 5 ppm) isolated from rodent brain homogenate. (C) MS2 552.42→ spectrum of the peak eluting at 9.23 min, and (D) MS2 552.42→ spectrum of the peak eluting at 10.03 min. Spectra were recorded under similar conditions to those described in Figure 14.10, however in this case mobile phase A was 50% methanol, 0.1% formic acid, and mobile phase B was 95% methanol containing 0.1% formic acid.

NEW APPLICATIONS

FIGURE 14.12 (Continued)

325

326

MASS SPECTROMETRY FOR STEROID ANALYSIS

FIGURE 14.13 Charge-tagging of 25-hydroxyvitamin D3. (A) Cholesterol oxidase converts the 3-hydroxy-5-ene to a 3-oxo-4-ene group. The C7–C8 double bond rearranges to C6–C7. (B) Fragmentation pathway of GP charge-tagged 25-hydroxyvitamin D3. (C) MS3 532.39→435.34→ spectrum of GP charge-tagged 25-hydroxyvitamin D3. The spectrum was recorded on the LTQ-OrbitrapXL as described in Figure 14.10.

PROTOCOLS

327

FIGURE 14.13 (Continued)

14.7

PROTOCOLS

Extraction Protocols Urine for ESI-MS 1. Add urine (100 µL) from sample vial dropwise to an Eppendorf tube (Cambridge, UK) containing 1 mL of HPLC-grade water. Check pH, should be 6–7. 2. Set-up a Certiied Sep-Pak C18 (200 mg, Waters Corporation, Elstree, UK) cartridge/column and attach a Leur-lock syringe (BD Biosciences, Franklin Lakes NJ) to the end of the cylinder. 3. Precondition the sorbent in the cartridge by irst rinsing the bed with 4 mL 99.9% ethanol followed by another rinse of 4 mL methanol and 4 mL HPLC water. 4. Add urine solution (1 mL) to the column, and allow to low at 0.25 mL/min (if necessary apply back pressure via the Leur-lock syringe). Discard the efluent. Wash with 3 mL of HPLC-grade water. Elute bile acids and steroids in 4 mL methanol. Reduce volume to 0.5 mL for direct analysis by negative ion ESI-MS (see Figure 14.5).

Plasma for ESI-MS 1. Add plasma (50 µL) from sample vial dropwise to an Eppendorf tube containing 1.05 mL of 99.9% analytical reagent-grade ethanol in an ultrasonic bath. Ultrasonicate for 5 min. 2. Dilute to 70% ethanol by adding 0.40 mL of HPLC grade water to the Eppendorf tube. Ultrasonicate for a further 5 min and centrifuge at 14,000 × g at 4°C for 30 min. 3. Set-up a Certiied Sep-Pak C18 (200 mg, Waters Corporation) cartridge and attach a Leur-lock syringe to the end of the cylinder. 4. Precondition the sorbent in the cartridge by irst rinsing the bed with 4 mL 99.9% ethanol, followed by a rinse of 6 mL 70% ethanol. 5. Add plasma in 70% ethanol (1.5 mL) to the column, and allow to low at 0.25 mL/min (if necessary, apply back pressure via the Leur-lock syringe). Collect the low through and combine with column wash of 1.5 mL of 70% ethanol in a 15 mL polypropylene centrifuge tube (Corning Inc. Amsterdam, The Netherlands). Steroids, oxysterols, and bile acid elute in this fraction (SPE1Fr-1). Elute less polar oxysterols from the column

328

6.

7.

8. 9. 10.

11.

12. 13. 14.

15. 16. 17. 18.

MASS SPECTROMETRY FOR STEROID ANALYSIS

in an additional 4 mL 70% ethanol. Elute residual sterols in 4 mL of 99.9% ethanol. Proceed to step 7 for direct ESI-MS analysis or dry down and reconstitute in propan-2-ol for charge-tagging with GP reagents (see the section “3β-Hydroxy-5-ene- (3β-hydroxy-5α-hydrogen) and Oxo-Steroids for LC-MS”). Precondition a second Certiied Sep-Pak C18 (200 mg, Waters Corporation) cartridge. Wash with 6 mL 100% methanol, then with 6 mL 10% methanol followed by 4 mL 70% ethanol. Load fraction SPE1-Fr-1 onto the column and collect efluent. Wash the sample lask with 1 mL 70% methanol. Apply to the column and collect. Add an equal amount (4 mL) of HPLC-grade water to the combined efluent and wash from the column (now ∼8 mL of 35% methanol : ethanol). Wash the column with 1 mL 35% methanol and collect and add to the collected efluent above (now ∼9 mL of 35% methanol : ethanol). Apply the resultant 9 mL of 35% methanol : ethanol to the column. Collect the efluent. Add 9 mL of HPLC-grade water to the efluent (now ∼18 mL 17.5% methanol : ethanol). Wash the column with 1 mL 17% methanol and collect and add to efluent above (now 19 mL 17.5% methanol : ethanol). Apply 19 mL 17.5% methanol/ethanol to the column and collect the efluent in one container. Wash the column with 6 mL 10% methanol and collect the efluent. Using 4 × 1 mL methanol, elute steroids. Reduce volume to 100 µL for direct negative ion ESI-MS analysis (see Figure 14.11).

Tissue (Mouse Brain) for ESI-MS 1. Lyse/extract mouse brain (1 g) in 10.5 mL 99.9% ethanol using mechanical mixing (rod mixer Polytron PT1200, Fisher Scientiic, Loughborough, UK) and ultrasonicate at 12 microns peak to peak for 2 min at room temperature. Add 4.5 mL of HPLC grade water and sonicate for another minute. 2. Centrifuge at 50,000 × g at 4°C for 1 h and collect the supernatant. 3. Set-up a Certiied Sep-Pak C18 (1 g, Waters Corporation) cartridge/column. 4. Rinse the column 20 mL 99.9% ethanol and condition with 30 mL 70% ethanol. 5. Add the supernatant from above (15 mL) to the cartridge/column (low rate about 0.25 mL/min).

Collect the low through and combine with four column washes of 5 mL of 70% ethanol. Cholesterol will be retained on the Sep-Pak even after four washes, while more hydrophilic molecules like bile acids, steroids, and oxysterols will elute in the low through and two irst washes. To ensure high recovery of steroids, all four washes and low through are combined. 6. Cholesterol is eluted with 10 mL of 99.9% ethanol. 7. More hydrophobic sterols are eluted with a second 10 mL portion of 99.9% ethanol. 8. Fractions are dried down and reconstituted in methanol or ethanol for direct ESI-MS analysis, or propan-2-ol for charge-tagging with GP reagents (see the section “3β-Hydroxy-5-ene- (3β-hydroxy5α-hydrogen) and Oxo-Steroids for LC-MS”).

Derivatization Protocols Bile acids for GC-MS (see Figure 14.2, Scheme A) 1. All steps should be performed in glass vials. All solvents/reagents/glassware should be dry. 2. Dry lipid sample and redissolve in methanol (100 µL). 3. Working in a fume hood, add 400 µL toluene, followed by 25 µL of 2.0 M trimethylsilyldiazomethane in hexane (Sigma-Aldrich, Gillingham, UK). Leave the solution at 40°C for 30 min. 4. Dry under a stream of nitrogen (the sample should be completely dry). 5. Add 350 µL of a freshly prepared solution of pyridine : hexamethyldisilazane : trimethylchlorosilane (Sigma-Aldrich) (3:2:1,v/v/v) to the dried sample. Incubate at 60°C for 30 min. 6. Dry under a stream of nitrogen (take care to avoid contamination with water or alcohol). 7. The sample should be reconstituted in 250 µL hexane, and the supernatant transferred to a fresh glass vial for GC-MS analysis (see Figure 14.4). 3β-Hydroxy-5-ene- (3β-hydroxy-5α-hydrogen) and Oxo-Steroids for LC-MS 1. The following protocol has been used in the analysis of steroids extracted from cerebrospinal luid (CSF), plasma, brain (see the sections “Plasma for ESI-MS” and “Tissue [Mouse Brain] for ESI-MS”), and cell culture [58,67,74]. Volumes and column sizes have been optimized for steroid extracts from 500 µL CSF, 100 µL plasma, and 50 mg rat brain. Isotope-labeled internal

PROTOCOLS

2.

3. 4. 5.

6. 7.

8. 9.

10. 11. 12. 13.

14.

15.

16. 17. 18.

standards (e.g., [26,26,26,27,27,27-2H6]24(R/S) -hydroxycholesterol [Avanti Polar Lipids, Alabaster, AL] 40 ng/500 µL CSF, 20 ng/100 µL plasma, and 200 ng/100 mg brain) should be added to 99.9% ethanol during the initial step in sterol extraction (see above “Extraction Protocols”). The sterol fractions eluted from the initial Certiied Sep-Pak C18 (200 mg, Waters Corporation) column (SPE1; see above “Extraction Protocols”) are subdivided into fractions (A) and (B) and dried down (see Figure 14.8). Add 100 µL of propan-2-ol to each subfraction; vortex thoroughly (2 min). Add 1 mL of 50 mM KH2PO4 buffer, pH 7, to each subfraction. Add 3 µL cholesterol oxidase from Streptomyces sp. (2 mg/mL in H2O, 44 units/mg protein; SigmaAldrich) to subfraction (A) and 3 µL of H2O to subfraction (B). Incubate at 37°C for 1 h. Add 2 mL of methanol containing 10–50 ng of isotope-labeledstandard[25,26,26,26,27,27,27-2H7]3βhydroxycholest-5-en-7-one. 3β-Hydroxycholest5-en-7-one is not oxidized by cholesterol oxidase, but is derivatized at C-7 with the GP reagent. Add 150 µL of glacial acetic acid; vortex (2 min). Add 150 mg of GP reagent (TCI Europe, Zwijndrecht Belgium), vortex 2 min. Incubate overnight at room temperature (RT) in the dark (see Figure 14.9). Precondition a second Certiied Sep-Pak C18 (200 mg, Waters Corporation) cartridge (SPE2). Wash with 6 mL 100% methanol, then 6 mL 10% methanol followed by 4 mL 70% methanol. Load the irst derivatized subfraction (i.e., SPE1Fr-1A) onto the column and collect efluent. Wash the sample lask with 1 mL 70% methanol. Apply to the column and combine with efluent above. Add an equal amount (4 mL) of HPLC-grade water to the combined efluent and wash from the column (now ∼8 mL of 35% methanol). Wash the column with 1 mL 35% methanol and collect and add to the collected efluent above (now ∼9 mL of 35% methanol). Apply the resultant 9 mL of 35% methanol to the column. Collect the efluent. Add 9 mL of HPLC-grade water to the efluent (now ∼18 mL 17.5% methanol). Wash the column with 1 mL 17% methanol and collect and add to efluent above (now 19 mL 17.5% methanol).

329

19. Apply 19 mL 17.5% methanol to the column and collect the efluent in one container. 20. Wash the column with 6 mL 10% methanol and collect the efluent. 21. Elute the derivatized sterols in 3 × 1 mL fractions of 100% methanol. 22. Elute any remaining sterols in 1 mL 99.9% ethanol. 23. Repeat steps 12–22 for each subfraction using a fresh Certiied Sep-Pak C18 for each subfraction. 24. Derivatized C19 and C21 steroids, C24 and C27 bile acids, and most oxysterols elute in the irst 2 mL of methanol from SPE2. Derivatized cholesterol elutes over the irst 3 mL of methanol. 25. The derivatized fractions are ready for LC-ESIMS analysis in the positive ion mode (see Figure 14.10 and 14.12). 26. Quantiication of steroids with a 3β-hydroxy-5ene function can be performed against [26,26,26,27,27,27-2H6]24(R/S)-hydroxycholesterol internal standard, and those with an oxo function against [25,26,26,26,27,27,27-2H7]3β-hydroxycholest5-en-7-one. Cholesterol and other nonpolar sterols with a 3β-hydroxy-5-ene can be quantiied against [25,26,26,26,27,27,27-2H7]cholesterol (Avanti Polar Lipids), which can be added in 99.9% ethanol during the initial extraction. 27. Only steroids possessing a natural oxo group will be derivatized in subfraction (B); those possessing a natural oxo group and those oxidized to contain one will be derivatized in subfraction (A). Quantitative differences between (A) and (B) will reveal the levels of metabolites with a 3β-hydroxy-5-ene and 3β-hydroxy-5α-hydrogen function.

LC-MS and MS/MS or MSn Protocols LC-MS for C24 and C27 Bile Acids/Alcohols (without Derivatization) 1. The following protocol is appropriate for mixtures of free and conjugated C24 and C27 bile acids/ alcohols. Details are given for capillary column (300 µm internal diameter [i.d.]) LC; low rates should be scaled up (2.1 mm i.d.) for conventional or down for nano-LC ( 6 test that measures the ratio of testosterone (T) and epitestosterone (E) [20,172–174]. GC-IRMS has been accepted by the International Olympic Committee (IOC) Medical

GUIDELINES FOR THE INSTALLATION AND IMPLEMENTATION

Commission (1997) as a viable technique for distinguishing between exogenous and endogenous steroid metabolites (Fourel, as cited by Phillips et al. [23]). Cawley and Flenker [175] provide an overview of the application of IRMS to doping control. The authors discuss the beneits of measuring carbon isotope ratios of urinary steroids to conirm their synthetic origin based on the carbon isotope ratio content. The authors also address some of the issues that can be encountered in such analyses and the importance of caution when interpreting the data. Cawley et al. [176] and Saudan et al. [177] discuss the successful use of GC-IRMS for the analysis of markers in the screening and conirmation of endogenous steroids in urine. Meier-Augenstein and Liu [9] discuss the potential of tagging substances with selected stable isotopes during their manufacture to assist with sample differentiation. The authors use the example of tagging drugs that are likely to be used for illicit purposes. In this scenario, IRMS could be used to differentiate samples from different sources and to identify the source of the drug and its precursors. Cawley et al. [178] demonstrate the application of GC-IRMS (in combination with GC-MS) to conirm androstenedione (ADIONE) administration. ADIONE is a synthetic steroid sometimes administered by athletes to increase systemic levels of testosterone during training. Sottas et al. [179] discuss the evaluation of a range of testosterone abuse detection protocols against a number of scenarios that may be encountered by antidoping laboratories. The aim of the evaluation was to determine the best process for interpretation of the T : E ratio in order to enhance the detection of testosterone abuse with minimal impact on inancial and administrative costs to sports authorities. The evaluation included IRMS and the application of a Bayesian interpretation of the T : E-time proile. Saudan et al. [180] evaluated the use of GC-IRMS for the measurement of carbon isotope values for the detection of exogenous gamma-hydroxybutyric acid (GHB) in human urine. Preliminary results indicate possible differentiation between endogenous and exogenous GHB. Aguilera et al. [181] demonstrate the application of GC-IRMS for steroid conirmation through the use of a 2007 Pan American Games case study (Rio de Janeiro, Brazil). A select sample of urine samples were analyzed for a range of steroids utilizing GC-IRMS. The carbon isotope value results from a set of suspicious samples were conclusive in determining whether illegal doping had occurred.

363

15.5 GUIDELINES FOR THE INSTALLATION AND IMPLEMENTATION OF IRMS INTO AN OPERATIONAL FORENSIC LABORATORY 15.5.1

Installation

Like any other piece of scientiic instrumentation, isotope ratio mass spectrometers (IRMS) must be installed as per the manufacturer’s speciications. The end user should ensure that the instrument manufacturer conducts a preinstallation site inspection to ensure that the proposed site is suitable for the reliable operation of the instrument. The following summary addresses some of the key considerations that should be addressed in preparation for the installation of an IRMS instrument. These examples are derived from the author’s experience with the installation of the DELTAplusXP IRMS instrument (ThermoFinnigan) during 2003–2004 and should not replace or override a particular manufacturer’s preinstallation requirements. A number of these issues were recommendations made along the way by various experienced users, as well as the instrument manufacturer. Meier-Augenstein [182] also addresses issues for consideration in the setup of a laboratory for IRMS analyses. 1. Location: Ensure appropriate loor space and bench space is available for the instrument and all ancillary equipment (i.e., gas chromatograph, elemental analyzers, and interfaces) and also future equipment that may be purchased to interface with the IRMS, for example, a liquid chromatograph. Consideration should also be given to the required space for conducting maintenance on the instrument and easy access to all components. 2. Environment: Ensure that there is minimal human trafic to the area where the instrument will be installed. It is best to have a dedicated room/area for the IRMS. There should be no signiicant power sources nearby, for example, strong magnets. The room temperature should be stable, for example, 23 ± 1°C per hour [182]. 3. Gases: Ensure that there is appropriate space for the storage of the required gases. Refer to the relevant standard that addresses the storage and handling of gases in cylinders (e.g., Australian Standard AS 4332-2004). Consideration should be given to the appropriate storage of gases relevant to their intended use (e.g., utilised as carrier or true reference gases). According to some users,

364

FORENSIC APPLICATIONS OF ISOTOPE RATIO MASS SPECTROMETRY

it is best practice to store the reference gases (i.e., carbon monoxide, nitrogen, carbon dioxide, and hydrogen) in a temperature-controlled environment. According to other users, cylinders can be stored outside as long as the isotopic composition of the cylinder remains constant for the duration of an analytical sequence, or at least between two sets of solid standards in a sequence (which are run every 10–18 samples, that is, approximately every 2 h). According to ThermoFinnigan [183], reference gases in the liquid phase inside a tank should be stored in a temperature-controlled room (i.e., CO2). Due to the high consumption of helium (as the carrier gas), it is recommended to have two cylinders installed with a pressure-dependent toggle switch (e.g., when the pressure in the irst cylinder drops below 500 psi, the regulator will automatically switch to the second cylinder) [182]. Ensure that there are release valves connected to all regulators that allow the gas lines to be purged with gas when installing a new cylinder and prevent air entering the gas lines and ultimately the instrument. Ensure that all lines, regulators, and ittings are made from stainless steel to ensure that the materials do not degrade and result in air leaks and contamination. All gas cylinders should be installed with relatively short lines to the instrument, to minimize leaks and, with regards to reference gases, minimize the potential for fractionation. Ensure that all the gases are of the required purity/grade (these are an expensive ongoing cost to the user and often have a signiicant lead time from the supplier). It is important to liaise very closely with the gas supplier to establish procedures to minimize lead times. There must also be a source of contaminantfree compressed air. 4. Power: Ensure that a reliable uninterrupted power supply (UPS) is installed that guards against power outages and also surges/spikes in power. Ensure that the appropriate power supply, as speciied by the manufacturer, is available. 5. Exhaust: Installation of exhaust lines from the interfaces to remove any potential traces of CO and H2 and also to remove exhaust from the vacuum pumps. 6. Gas Cylinder Pressure Monitors: This will minimize the potential for air to reach the lines and ultimately the instrument by alarming when the gas cylinders require replacing, that is, drop to a certain cylinder pressure (e.g., 500 psi).

7. Gas Sensors: H2 and CO gas monitors should be installed in the room in case there is a leak from the lines or from the interface. Sensors for any gases that are installed indoors should also be installed.

15.5.2

Implementation

Relative to other analytical instruments (e.g., FTIR or GC-MS), IRMS instruments are resource intensive to maintain and operate. Consumables, for example, furnaces and high purity gases, are an expensive ongoing cost. Maintenance contracts are also an added ongoing expense that requires consideration. In addition to the standard considerations that must be given to the implementation of a new instrument into a forensic laboratory (i.e., inancial sustainability, resources to operate and maintain, development and validation of standard procedures), the following summarizes a number of key requirements speciically relating to the implementation of an IRMS instrument: 1. Instrumentation/Equipment: Appropriate installation of IRMS (refer to Section 15.5.1 on “Installation”); analytical scales capable of weighing down to 0.001 mg; desiccators for the storage of samples and standards. 2. IAEA Standards/Certiied Reference Materials: To allow calibration of laboratory standards, for measurement as QC standards, and to allow interlaboratory comparisons and data exchange (comparability of results from different laboratories). 3. Appropriate Laboratory (Working) Standards: To routinely measure samples of interest and determine the true value of the samples relative to certiied reference materials. If performed correctly, then the use of reference materials will allow intraas well as interlaboratory comparisons. Laboratory standards are required that are appropriate for the individual target materials. 4. Validation: To understand the performance of the instrument, including limitations. Measurement uncertainty for the instrument, and speciically for individual target materials, is required to assist in interpretation, that is, to determine what constitutes a “signiicant” difference between samples. The validation protocol should be developed in line with guidelines provided by the local accreditation body (e.g., the National Association of Testing Authorities, Australia, NATA [184,185]) while taking into consideration the relevance of

REFERENCES

5.

6.

7.

each performance characteristic to the IRMS technique. Standard Procedures, Maintenance Schedule, and Methods. Speciic considerations in method development include: ensure that the entire peak is detected; ensure there is no interference from other gases; sample peak height should be relatively the same height as the reference peak height; samples should be the same weight so that linearity effects are not introduced; and ensure that complete combustion of samples and standards is achieved. Ensure that a consistent analytical sequence is used with blanks, standards, and samples. Werner and Brand [4] also recommend maintaining performance charts to monitor the long-term performance of the instrument and to assist with predicting required maintenance on the instrument. Analytical Protocols: IRMS should be utilized in casework following analysis and identiication of the samples of interest utilizing traditional forensic techniques. The technique is destructive; hence, the technique should generally be utilized at the end of an analytical sequence, particularly if limited sample is available. Data Treatment: Develop algorithms/protocols to address the following as required [4]: • Drift Correction: Can be observed in the results as a function of time (or sample number) and can be caused by a range of factors, including isotopic change in the reference gas, buildup of water or other contaminants during analysis, changing MS conditions, deterioration of the ion source, and so on; • Linearity Correction: May observe a drift with size, that is, the measured ratio is a function of the quantity of sample; • Memory Effects: Carryover from previously measured samples; and • Isobaric Interferences: Interfering ion currents from other species hitting the same FC detectors, for example: 17 O correction for δ13C determination (e.g., in the ThermoFinnigan Isodat software, one would utilize the Craig or Santrock correction); H3+ Correction: H3+ factor should be determined at the beginning of every sequence (ensure H3+ factor is low and stable, that is, contribution from the tail of the 4He+ peak into the m/z = 3 FC is minimal). Data Interpretation: Interpretation requires databases and/or demonstrated ability to compare 



8.

365

results interlaboratory (i.e., require a validated referencing regime against certiied reference materials). The resources to develop and maintain such databases are considerable and require signiicant ongoing collaboration with relevant industries. Interpretation also requires an understanding of the instrument performance (i.e., estimate of measurement uncertainty), and the potential for isotopic fractionation to occur to determine whether an observed difference between two samples is “signiicant.”

15.6

SUMMARY

The overview provided in this chapter highlights the fundamental principles of the IRMS technique and the signiicant value that it can play in complex forensic investigations. This chapter also addresses some of the key considerations for the installation and implementation of the technique into an operational forensic laboratory. The current status of the technique for a number of the forensic applications discussed in this chapter allow for the use of IRMS results for intelligence purposes; however for others, continued research is required to ensure the generation of court reportable results and also to continue the path toward broader forensic casework application.

REFERENCES 1. Benson, S., Lennard, C., Maynard, P., Roux, C. (2006) Forensic applications of isotope ratio mass spectrometry—a review. Forensic Sci. Int., 157(1), 1–22. 2. Barrie, A., Prosser, S.J. (1996) Automated analysis of light-element stable isotopes by isotope ratio mass spectrometry. In Mass Spectrometry of Soils, edited by Boutton, T.W., Yamasaki, S. New York: Marcel Dekker Inc, pp. 1–46. 3. Fourel, F. (2002) Forensic applications in continuous low IRMS using the IsoPrime. Presentation at the Forensic Isotope Ratio Mass Spectrometry (FIRMS) Conference, September 16–17, 2002, http://www.forensic-isotopes. rdg.ac.uk/conf/conf2002.htm (accessed November 20, 2011). 4. Werner, R.A., Brand, W.A. (2001) Referencing strategies and techniques in stable isotope ratio analysis. Rapid Commun. Mass Spectrom., 15, 501–519. 5. Gröning, M. (2004) International stable isotope reference materials. In Handbook of Stable Isotope Analytical Techniques, Vol. 1, edited by de Groot, P.A. Amsterdam, The Netherlands: Elsevier, pp. 874–906. 6. Peters, K.E., Moldowan, J.M. (1993) The Biomarker Guide Interpreting Molecular Fossils in Petroleum and

366

7.

8.

9.

10.

11. 12.

13. 14.

15.

16.

17.

18.

19.

20.

FORENSIC APPLICATIONS OF ISOTOPE RATIO MASS SPECTROMETRY

Ancient Sediments. Upper Saddle River, NJ: Prentice Hall, pp. 122–130. International Organization for Standardization (2009) ISO guide 34:2009 general requirements for the competence of reference material producers. http:// infostore.saiglobal.com/store/Details.aspx?ProductID= 1378940 (accessed November 20, 2011). Barrie, A., Davies, J.E., Park, A.J., Workman, C.T. (1989) Continuous-low stable isotope analysis for biologists. Spectroscopy, 4(7), 42–52. Meier-Augenstein, W., Liu, R.H. (2004) Forensic applications of isotope ratio mass spectrometry. In Advances in Forensic Applications of Mass Spectrometry, edited by Yinon, J. Boca Raton, FL: CRC Press, pp. 149–180. Brand, W. (1996) High precision isotope ratio monitoring techniques in mass spectrometry. J. Mass Spectrom., 31, 225–235. Muccio, Z., Jackson, G.P. (2009) Isotope ratio mass spectrometry. Analyst, 134(2), 213–222. Hoogewerff, J. (2004) Origin of trace element and isotope variation, relevance for forensic studies. Presentation at the NITECRIME Workshop, Wellington, New Zealand, April 2, 2004. Galimov, E.M. (1985) The Biological Fractionation of Isotopes. Maryland Heights, MO: Academic Press. Vanhaecke, F., Heumann, K.G. (2004) Precise isotope ratio measurements as a unique tool in modern analytical chemistry. Anal. Bioanal. Chem., 378, 227–228. Schmidt, T.C., Zwank, L., Berg, M.E.M., Meckenstock, R.U., Haderlein, S.B. (2004) Compound-speciic stable isotope analysis of organic contaminants in natural environments: a critical review of the state of the art, prospects, and future challenges. Anal. Bioanal. Chem., 378, 283–300. Coplen, T.B., Böhlke, J.K., De Bievre, P., Ding, T., Holden, N.E., Hopple, J.A., Krouse, H.R., Lamberty, A., Peiser, H.S., Revesz, K., Rieder, S.E., Rosman, K.J.R., Roth, E., Taylor, P.D.P., Vocke, R.D. Jr., Xiao, Y.K. (2002) Isotopeabundance variations of selected elements (IUPAC technical report). Pure Appl. Chem., 74(10), 1987–2017. Rapley, C., Grassl, H. The Global Network of Isotopes in Precipitation (GNIP) Booklet. Website: http://wwwnaweb.iaea.org/napc/ih/documents/userupdate/ description/1stpage.html (accessed November 20, 2011). Weilacher, T., Gleixner, G., Schmidt, H.-L. (1996) Carbon isotope pattern in purine alkaloids a key to isotope discriminations in C1 compounds. Phytochemistry, 41(4), 1073–1077. Besacier, F., Guilluy, R., Brazier, J.L., Chaudron-Thozet, H., Girard, J., Lamotte, A. (1997) Isotopic analysis of 13C as a tool for comparison and origin assignment of seized heroin samples. J. Forensic Sci., 42(3), 429–433. Meier-Augenstein, W. (1999) Review: applied gas chromatography coupled to isotope ratio mass spectrometry. J. Chromatogr. A, 842, 351–371.

21. Desage, M., Guilluy, R., Brazier, J.L., Chaudron, H., Girard, J., Cherpin, H., Jumeau, J. (1991) Gas chromatography with mass spectrometry or isotope-ratio mass spectrometry in studying the geographical origin of heroin. Anal. Chim. Acta, 247, 249–254. 22. Palhol, F., Lamoureux, C., Naulet, N. (2003) 15N isotopic analyses: a powerful tool to establish links between seized 3,4-methylenedioxymethamphetamine (MDMA) tablets. Anal. Bioanal. Chem., 376, 486–490. 23. Phillips, S.A., Doyle, S., Philp, L., Coleman, M. (2003) Proceedings: network developing forensic applications of stable isotope ratio mass spectrometry conference 2002. Sci. Justice, 43(3), 153–160. 24. Dautraix, S., Guilluy, R., Chaudron-Thozet, H., Brazier, J.L., Lamotte, A. (1996) 13C Isotopic analysis of an acetaminophen and diacetylmorphine mixture. J. Chromatogr. A, 756, 203–210. 25. Schmitt, J., Glaser, B., Zech, W. (2003) Amount-dependent isotopic fractionation during compound-speciic isotope analysis. Rapid Commun. Mass Spectrom., 17, 970–977. 26. Urey, H.C. (1948) Oxygen isotopes in nature and in the laboratory. Science, 108, 489–496. 27. Hilkert, A.W., Douthitt, C.B., Schlüter, H.J., Brand, W.A. (1999) Isotope ratio monitoring gas chromatography/ mass spectrometry of D/H by high temperature conversion isotope ratio mass spectrometry. Rapid Commun. Mass Spectrom., 13, 1226–1230. 28. Thermo Fisher Scientiic, Waltham, MA (2002) ThermoFinnigan DELTAPlusXP Operating Manual. Ident. No. 114 90 80, Issue 06/2002. 29. Barsdale, R.J., Dugdale, R.C. (1965) Rapid conversion of organic nitrogen to N2 for mass spectrometry: an automated Dumas procedure. Anal. Biochem., 13, 1–5. 30. Preston, T., Owens, N.J.P. (1983) Interfacing an automatic elemental analyser with an isotope ratio mass spectrometer: the potential for fully automated total nitrogen and nitrogen-15 analysis. Analyst, 108, 971–977. 31. Thermo Fisher Scientiic, Waltham, MA (2001) ThermoFinnigan ConFlo III Operating Manual. Ident. No. 111 0041, Issue 5/2001. 32. Oesselmann, J., Hilkert, A., Douthitt, C.B. (2001) Trends in stable isotope analyses and applications. ThermoFinnigan DeltaPlus MAT 252 Application Presentation, 2001. 33. Thermo Fisher Scientiic, Waltham, MA (2001) ThermoFinnigan High Temperature Conversion Elemental Analyzer (TC/EA) Operating Manual. Ident. No. 112 76 01, Issue 11/2001. 34. Sano, M., Yotsui, Y., Abe, H., Sasaki, S. (1976) New technique for the detection of metabolites labelled by the isotope 13C using mass fragmentography. Biomed. Mass Spectrom., 3, 103. 35. Jasper, J.P., Edwards, J.S., Ford, L.C., Corry, R.A. (2002) Putting the arsonist at the scene: “DNA” for the ire investigator? Gas chromatography/isotope ratio mass spectrometry. Fire Arson Invest., 51(2), 30–34.

REFERENCES

36. Thermo Fisher Scientiic, Waltham, MA (2001) ThermoFinnigan High Temperature Conversion Interface (GC/TC) Operating Manual. Ident. No. 112 76 31, Issue 04/2001. 37. Thermo Fisher Scientiic, Waltham, MA (2002) ThermoFinnigan GC Combustion Interface III (GC/ CIII) Operating Manual. Ident. No. 111 00 71, Issue 06/2002. 38. Krummen, M., Hilkert, A.W., Juchelka, D., Duhr, A., Schlüter, H.-J., Pesch, R. (2004) A new concept for isotope ratio monitoring liquid chromatography/mass spectrometry. Rapid Commun. Mass Spectrom., 18, 2260–2266. 39. Godin, J.-P., Hopfgartner, G., Fay, L. (2008) Temperatureprogrammed high-performance liquid chromatography coupled to isotope ratio mass spectrometry. Anal. Chem. 80(18), 7144–7152. 40. Godin, J.-P., Breuillé, D., Obled, C., Papet, I., Schierbeek, H., Hopfgartner, G., Fay, L.-B. (2008) Liquid and gas chromatography coupled to isotope ratio mass spectrometry for the determination of 13C-valine isotopic ratios in complex biological samples. J. Mass Spectrom., 43(10), 1334–1343. 41. Thermo Electron Corporation. (2003) Finnigan™ LC IsoLink Catalogue. 42. Hettmann, A., Brand, W.A., Gleixner, G. (2007) Improved isotope ratio measurement performance in liquid chromatography/isotope ratio mass spectrometry by removing excess oxygen. Rapid Commun. Mass Spectrom., 21, 4135–4141. 43. Browne, T.R. ed. (1997) Stable Isotopes in Pharmaceutical Research. Amsterdam, The Netherlands: Elsevier Science. 44. Jasper, J.P., Fourel, F., Eaton, A., Morrison, J., Phillips, A. (2004) Stable isotopic characterization of analgesic drugs. Pharm. Technol., August, 60–67. 45. Jasper, J.P., Westenberger, B.J., Spencer, J.A., Buhse, L.F., Nasr, M. (2004) Stable isotopic characterization of active pharmaceutical ingredients. J. Pharm. Biomed., 35(1), 21–30. 46. Bommer, P., Moser, H., Stichler, W., Trimborn, P., Vetter, W. (1976) Determination of the origin of drugs by measuring natural isotope contents: D/H and 13C/12C ratios of some diazepam samples. Z. Naturforsch., 31c, 111–114. 47. Santamaria-Fernandez, R., Hearn, R., Wolff, J.-C. (2009) Detection of counterfeit antiviral drug Heptodin™ and classiication of counterfeits using isotope amount ratio measurements by multicollector inductively coupled plasma mass spectrometry (MC-ICPMS) and isotope ratio mass spectrometry (IRMS). Sci. Justice, 49(2), 102–106. 48. Koziet, J., Rossmann, A., Martin, G.J., Ashurst, P.R. (1993) Determination of carbon-13 content of sugars of fruit and vegetable juices. Anal. Chim. Acta, 271, 31–38. 49. Morrison, J., Fourel, F. (1999) Application of 13C EA/ IRMS to the detection of adulteration of honey. Micromass application brief, AB8.

367

50. Boner, M., Forstel, H. (2002) Examination of the 13C/12C isotopes in sparkling and semi-sparkling wine with the aid of simple on-line sampling. Micromass Application Note 516. 51. Lofthouse, J., Rhodes, C., Heaton, K., Burwood, C., Brereton, P., Hird, S. (2002) The use of pyrolysis-isotope ratio mass spectrometry (Py-IRMS) to combat fraud in the food industry: some lessons to be learnt for applications in forensics. Poster at the Forensic Isotope Ratio Mass Spectrometry (FIRMS) Conference, September 16–17, 2002, http://www.forensic-isotopes.rdg.ac.uk/conf/ conf2002.htm (accessed November 20, 2011). 52. Kelly, S., Hoogewerff, J. (2004) Detection of food adulteration. Presentation at the NITECRIME Workshop, Wellington, New Zealand, April 2, 2004. 53. Krueger, D.A., Krueger, H.W. (1983) Carbon isotopes in vanillin and the detection of falsiied “natural” vanillin. J. Agric. Food Chem., 31, 1265–1268. 54. Krueger, D.A., Krueger, H.W. (1985) Detection of fraudulent vanillin labelled with 13C in the carbonyl carbon. J. Agric. Food Chem., 33(3), 323–325. 55. Doner, L.W., Krueger, H.W., Reesman, R.H. (1980) Isotopic composition of carbon in apple juice. J. Agric. Food Chem., 28, 362–364. 56. Calderone, G., Naulet, N., Guillou, C., Reniero, F. (2004) Characterization of European wine glycerol: stable carbon isotope approach. J. Agric. Food Chem., 52, 5902–5906. 57. Jamin, E., Martin, F., Martin, C.G., Blanch-Cortes, A.I., Christoph, N., Ciambotti, A., Giraudon, S., Guerin, R., Guillou, C., Kelly, S., Lamoureux, C., Mazac, J., Naulet, N., Pavanello, F., Retif, M., Rossmann, A., Versini, G. (2004) Determination of the C-13/C-12 ratio of ethanol derived from fruit juices and maple syrup by isotope ratio mass spectrometry: collaborative study. J. AOAC Int., 87(3), 621–631. 58. Fink, K., Richling, E., Heckel, F., Schreier, P. (2004) Determination of H-2/H-1 and C-13/C-12 isotope ratios of (E)-methyl cinnamate from different sources using isotope ratio mass spectrometry. J. Agric. Food Chem., 52(10), 3065–3068. 59. Pilgrim, T.S., Watling, R.J., Grice, K. (2008) Application of trace element and stable isotope signatures to determine the provenance of tea (Camellia sinensis) samples. Food Chem., 118(4), 921–926. doi:10.1016/j.foodchem. 2008.08.077. 60. Papesch, W., Horacek, M. (2009) Forensic applications of stable isotope analysis: case studies of the origins of water in mislabelled beer and contaminated diesel fuel. Sci. Justice, 49, 138–141. 61. Luykx, D.M.A., van Ruth, S.M. (2008) An overview of analytical methods for determining the geographical origin of food products. Food Chem., 107, 897–911. 62. Evershed, R.P., Arnot, K.I., Collister, J., Eglinton, G., Charters, S. (1994) Application of isotope ratio monitoring gas chromatography–mass spectrometry to the analy-

368

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

FORENSIC APPLICATIONS OF ISOTOPE RATIO MASS SPECTROMETRY

sis of organic residues of archaeological origin. Analyst, 119, 909–914. Tripp, J.A., Hedges, R.E.M. (2004) Single-compound isotopic analysis of organic materials in archaeology. LC GC N. Am., 22(11), 1098. Fourel, F., Merren, T., Morrison, J., Wassenaar, L., Hobson, K. (2000) Application of EA Pyrolysis—IRMS δD and δ18O analysis to animal migration patterns. Micromass Application Note 511. Merren, T., Morrison, J., Dudd, S., Hobson, K., Wassenaar, L. (2001) Investigation of whale migration patterns using stable isotope variations in baleen growth bands. Micromass Application Note 515. Royle, J.A., Rubenstein, D.R. (2004) The role of species abundance in determining breeding origins of migratory birds with stable isotopes. Ecol. Appl., 14(6), 1780–1788. Stelling, M.A., Van der Peijl, G.J.Q. (2001) Forensic methods for criminal investigations related to endangered species. Netherlands Forensic Institute Report, December 2001. Bowen, G.J., Wassenaar, L.I., Hobson, K.A. (2005) Global application of stable hydrogen and oxygen isotopes to wildlife forensics. Oecologia, 143, 337–348. Wassenaar, L.I., Hobson, K.A. (2006) Stable-hydrogen isotope heterogeneity in keratinous materials: mass spectrometry and migratory wildlife tissue subsampling strategies. Rapid Commun. Mass Spectrom., 20, 2505–2510. Kelly, A., Thompson, R., Newton, J. (2008) Stable hydrogen isotope analysis as a method to identify illegally trapped songbirds. Sci. Justice, 48(2), 67–70. Renou, J.-P., Bielicki, G., Deponge, C., Gachon, P., Micol, D., Ritz, P. (2004) Characterisation of animal products according to geographic origin and feeding diet using nuclear magnetic resonance and isotope ratio mass spectrometry. Part II. beef meat. Food Chem., 86, 251–256. Renou, J.-P., Deponge, C., Gachon, P., Bonnefoy, J.-C., Coulon, J.-B., Garel, J.-P., Vérité, R., Ritz, P. (2004) Characterisation of animal products according to geographic origin and feeding diet using nuclear magnetic resonance and isotope ratio mass spectrometry: cow milk. Food Chem., 85, 63–66. Boner, M., Förstel, H. (2004) Stable isotope variation as a tool to trace the authenticity of beef. Anal. Bioanal. Chem., 378, 301–310. Brescia, M.A., Monfreda, M., Buccolieri, A., Carrino, C. (2005) Characterisation of the geographical origin of buffalo milk and mozzarella cheese by means of analytical and spectroscopic determinations. Food Chem., 89, 139–147. Pye, K. (2004) Isotopic and trace element characterization of human teeth and bones for identiication and provenance purposes. Presentation at the NITECRIME Workshop, Wellington, New Zealand, April 2, 2004. Bol, R., Plieger, C. (2002) Stable isotope (C-13, N-15 and S-34) analysis of the hair of modern humans and their

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

domestic animals. Rapid Commun. Mass Spectrom., 16(23), 2195–2200. Bell, L.S., Lee-Thorp, J.A., Dobney, K. (2006) Mapping human movement using stable oxygen isotopic ratio mass spectrometry: potential application to forensic science demonstrated by a modern horse-human study. Can. Soc. Forensic Sci. J., 39(2), 47–54. Fraser, I., Meier-Augenstein, W., Kalin, R.M. (2006) The role of stable isotopes in human identiication: a longitudinal study into the variability of isotopic signals in human hair and nails. Rapid Commun. Mass Spectrom., 20(7), 1109–1116. Fraser, I., Meier-Augenstein, W. (2007) Stable 2H isotope analysis of modern-day human hair and nails can aid forensic human identiication. Rapid Commun. Mass Spectrom., 21, 3279–3285. Fraser, I., Meier-Augenstein, W., Kalin, R.M. (2008) Stable isotope analysis of human hair and nail samples: the effects of storage on samples. J. Forensic Sci., 53(1), 95–99. Meier-Augenstein, W., Fraser, I. (2008) Forensic isotope analysis leads to identiication of a mutilated murder victim. Sci. Justice, 48(3), 153–159. Rauch, E., Rummel, S., Lehn, C., Büttner, A. (2007) Origin assessment of unidentiied corpses by use of stable isotope ratios of light (bio-) and heavy (geo-) elements—a case report. Forensic Sci. Int., 168, 215–218. Bull, I.D., Berstan, R., Vass, A., Evershed, R.P. (2009) Identiication of a disinterred grave by molecular and stable isotope analysis. Sci. Justice, 49, 142–149. Wakelin, D., Doyle, S., Andrews, C., Mountford, S., Daeid, N.N. (2008) Network developing forensic applications of stable isotope ratio mass spectrometry conference 2005. Sci. Justice, 48(2), 79–90. Carter, J.F., Hill, J.C., Doyle, S., Lock, C. (2009) Results of four inter-laboratory comparisons provided by the Forensic Isotope Ratio Mass Spectrometry (FIRMS) network. Sci. Justice, 49(2), 127–137. Thermo Fisher Scientiic, Waltham, MA (1995) Finnigan mat, 15N/14N and 13C/12C by EA/IRMS forensic studies using the ConFlo II interface. Application Flash Report No. 15. 12/1995 PL 0/1177. Farmer, N., Meier-Augenstein, W., Lucy, D. (2009) Stable isotope analysis of white paints and likelihood ratios. Sci. Justice, 49(2), 114–119. Phillips, S.A., Doyle, S., Mountford, S. (2003) Proceedings of the Forensic Isotope Ratio Mass Spectrometry (FIRMS) Workshop, May 14, 2003, http://www.forensicisotopes.rdg.ac.uk/papers/workshop.pdf (accessed November 20, 2011). Horacek, M., Min, J.-S., Heo, S., Park, J., Papesch, W. (2008) The application of isotope ratio mass spectrometry for discrimination and comparison of adhesive tapes. Rapid Commun. Mass Spectrom., 22(11), 1763–1766. Carter, J.F., Grundy, P.L., Hill, J.C., Ronan, N.C., Titterton, E.L., Sleeman, R. (2004) Forensic isotope ratio mass

REFERENCES

91.

92.

93.

94.

95.

96.

97.

98.

99.

100.

101.

102.

103.

spectrometry of packaging tapes. Analyst, 129, 1206– 1210. Taylor, E., Carter, J.F., Hill, J.C., Morton, C., Daeid, N.N., Sleeman, R. (2008) Stable isotope ratio mass spectrometry and physical comparison for the forensic examination of grip-seal plastic bags. Forensic Sci. Int., 177(2–3), 214–220. van Es, A., de Koeijer, J., van der Peijl, G. (2009) Discrimination of document paper by XRF, LA-ICP-MS and IRMS using multivariate statistical techniques. Sci. Justice, 49, 120–126. Van der Peijl, G., Montero, S., Wiarda, W., de Joode, P. (2004) Isotopic and other investigations in relation to a safe burglary. In Proceedings of the 17th International Symposium on the Forensic Sciences Handbook and Programme, Wellington, New Zealand, March 28–April 2, 2004. Farmer, N.L., Meier-Augenstein, W., Kalin, R.M. (2005) Stable isotope analysis of safety matches using isotope ratio mass spectrometry—a forensic case study. Rapid Commun. Mass Spectrom., 19, 3182–3186. Farmer, N.L., Ruffell, A., Meier-Augenstein, W., Meneely, J., Kalin, R.M. (2007) Forensic analysis of wooden safety matches—a case study. Sci. Justice, 47, 88–98. Farmer, N., Curran, J., Lucy, D., Daeid, N.N., MeierAugenstein, W. (2009) Stable isotope proiling of burnt wooden safety matches. Sci. Justice, 49(2), 107–113. Croft, D.J., Pye, K. (2003) The potential use of continuouslow isotope-ratio mass spectrometry as a tool in forensic soil analysis: a preliminary report, Rapid Commun. Mass Spectrom., 17(23), 2581–2584. Pye, K., Blott, S.J., Croft, D.J., Carter, J.F. (2006) Forensic comparison of soil samples: assessment of small-scale spatial variability in elemental composition, carbon and nitrogen isotope ratios, colour, and particle size distribution. Forensic Sci. Int., 163, 59–80. Roelofse, F., Horstmann, U.E. (2008) A case study on the application of isotope ratio mass spectrometry (IRMS) in determining the provenance of a rock used in an alleged nickel switching incident. Forensic Sci. Int. 174(1), 64–67. Ohuchi, T., Muehlenbachs, K. (2003) Isotopic ingerprinting of market products: carbon and hydrogen isotope ratios of polymer materials. In Proceedings from the Canadian Society of Forensic Science (CSFS) 50th Anniversary Meeting, Vancouver BC, March 25–29, 2003. Nissenbaum, A. (1975) The distribution of natural stable isotopes of carbon as a possible tool for the differentiation of samples of TNT. J. Forensic Sci., 20, 455–459. McGuire, R.R., Lee, C.G., Velsko, C.A., Raber, E. (1995) Application of stable isotope ratios to the analysis of explosive residues. In Proceedings of the Fifth International Symposium on the Analysis and Detection of Explosives, Washington, DC, December 4–8, 1995. Wakelin, D. (2000) Isotopic ratio analysis of explosives traces—a new type of evidence. United Kingdom Defence

369

Evaluation and Research Agency Report DERA/CES/ CS/CR000294, December 2000. 104. Wakelin, D. (2001) Isotopic ratio analysis of explosives traces—a new type of evidence Part 2. United Kingdom Defence Evaluation and Research Agency Report DERA/ DSTL/CR00569, April 2001. 105. Beardah, A.M. (2002) Stable isotope ratio mass spectrometry of explosives—a new type of evidence (Part 3). United Kingdom Defence Science and Technology Laboratory Report DSTL/CR05104, October 2002. 106. Lott, M.J., Howa, J., Ehleringer, J.R. (2002) Locating the origins of explosives through stable isotope ratio analysis, Poster at the Forensic Isotope Ratio Mass Spectrometry (FIRMS) Conference, September 16–17, 2002, http:// www.forensic-isotopes.rdg.ac.uk/conf/conf2002.htm (accessed November 20, 2011). 107. Motzer, W.E., Mohr, T.K.G., McCraven, S., Stanin, P. (2006) Stable and other isotope techniques for perchlorate source identiication. Environ. Forensics, 7(1), 89–100. 108. Pierrini, G., Doyle, S., Champod, C., Taroni, F., Wakelin, D., Lock, C. (2007) Evaluation of preliminary isotopic analysis (13C and 15N) of explosives A likelihood ratio approach to assess the links between semtex samples. Forensic Sci. Int., 167(1), 43–48. 109. Bernstein, A., Ronen, Z., Adar, E., Nativ, R., Lowag, H., Stichler, W., Meckenstock, R.U. (2008) Compoundspeciic isotope analysis of RDX and stable isotope fractionation during aerobic and anaerobic biodegradation. Environ. Sci. Technol., 42(21), 7772–7777. 110. Lock, C.M., Meier-Augenstein, W. (2008) Investigation of isotopic linkage between precursor and product in the synthesis of a high explosive. Forensic Sci. Int., 179(2–3), 157–162. 111. Meier-Augenstein, W., Kemp, H.F., Lock, C.M. (2009) N2: a potential pitfall for bulk 2H isotope analysis of explosives and other nitrogen-rich compounds by continuouslow isotope-ratio mass spectrometry. Rapid Commun. Mass Spectrom., 23, 2011–2016. 112. Widory, D., Minet, J.-J., Barbe-Leborgne, M. (2009) Sourcing explosives: a multi-isotope approach. Sci. Justice, 49, 62–72. 113. Quirk, A.T., Bellerby, J.M., Carter, J.F., Thomas, F.A., Hill, J.C. (2009) An initial evaluation of stable isotopic characterisation of post-blast plastic debris from improvised explosive devices. Sci. Justice, 49(2), 87–93. 114. Benson, S.J., Lennard, C.J., Maynard, P., Hill, D.M., Andrew, A.S., Roux, C. (2009) Forensic analysis of explosives using isotope ratio mass spectrometry (IRMS)— discrimination of ammonium nitrate sources. Sci. Justice, 49(2), 73–80. 115. Benson, S.J., Lennard, C.J., Maynard, P., Hill, D.M., Andrew, A.S., Roux, C. (2009) Forensic analysis of explosives using isotope ratio mass spectrometry (IRMS)— preliminary study on TATP and PETN. Sci. Justice, 49(2), 81–86.

370

FORENSIC APPLICATIONS OF ISOTOPE RATIO MASS SPECTROMETRY

116. Philp, R.P. (2007) The emergence of stable isotopes in environmental and forensic geochemistry studies: a review. Environ. Chem. Lett., 5, 57–66. 117. Jasper, J.P., Edwards, J.S., Ford, L.C. (2001) A novel method for arson accelerant analysis: gas chromatography/isotope ratio mass spectrometry, http:// www.naturesingerprint.com/pdfs/Arson%20Stable%20 Isotope%20Analysis.pdf (accessed November 20, 2011). 118. Sandercock, P.M.L., DuPasquier, E. (2004) Chemical ingerprinting of gasoline. 3. Comparison of unevaporated automotive gasoline samples from Australia and New Zealand. Forensic Sci. Int., 140(1), 71–77. 119. Blessing, M., Jochmann, M.A., Schmidt, T.C. (2008) Pitfalls in compound-speciic isotope analysis of environmental samples. Anal. Bioanal. Chem., 390(2), 591– 603. 120. Turner, N., Jones, M., Grice, K., Dawson, D., IoppoloArmanios, M., Fisher, S. (2006) δ13C of volatile organic compounds (VOCS) in airborne samples by thermal desorption-gas chromatography-isotope ratio-mass spectrometry (TD-GC-IR-MS). Atmos. Environ., 40, 3381– 3388. 121. Sturchio, N.C. (2004) Use of table chlorine and carbon isotope analysis in environmental forensic investigations of groundwater contamination. Abstracts from PITTCON—Analytical Chemistry and. Applied Spectroscopy Conference Chicago, USA (March 7–12, 2004). 122. Silva, S.R., Ging, P.B., Lee, R.W., Ebbert, J.C., Tesoriero, A.J., Inkpen, E.L. (2002) Forensic applications of nitrogen and oxygen isotopes in tracing nitrate sources in urban environments. Environ. Forensics, 3, 125–130. 123. Whittaker, M., Pollard, S.J.T., Fallick, A.E., Preston, T. (1996) Characterisation of refractory wastes at hydrocarbon-contaminated sites-II. Screening of reference oils by stable carbon isotope ingerprinting. Environ. Pollut., 94(2), 195–203. 124. Kaplan, I.R., Galperin, Y., Lu, S.-T., Lee, R.-P. (1997) Forensic environmental geochemistry: differentiation of fuel types, their sources and release time. Org. Geochem., 27, 289–317. 125. Kelley, C.A., Hammer, B.T., Cofin, R.B. (1997) Concentrations and stable isotope values of BTEX in gasoline-contaminated groundwater. Environ. Sci. Technol., 31, 2469–2472. 126. Dempster, H.S., Lollar, B.S., Feenstra, S. (1997) Tracing organic contaminants in groundwater: a new methodology using compound-speciic isotopic analysis. Environ. Sci. Technol., 31(11), 3193–3197. 127. Harrington, R.R., Poulson, S.R., Drever, J.I., Colberg, P.J.S., Kelly, E.F. (1999) Carbon isotope systematics of monoaromatic hydrocarbons: vaporization and adsorption experiments. Org. Geochem., 30, 765–775. 128. Saber, D., Mauro, D., Sirivedhin, T. (2006) Environmental forensics investigation in sediments near a former manufactured gas plant site. Environ. Forensics, 7, 65–75.

129. Smallwood, B.J., Philp, R.P., Allen, J.D. (2002) Stable carbon isotopic composition of gasolines determined by isotope ratio monitoring gas chromatography mass spectrometry. Org. Geochem., 33, 149–159. 130. Ahad, J.M.E., Lollar, B.S., Edwards, E.A., Slater, G.F., Sleep, B.E. (2000) Carbon isotope fractionation during anaerobic biodegradation of toluene: implications for intrinsic bioremediation. Environ. Sci. Technol., 34, 892–896. 131. Mansuy, L., Philp, R.P., Allen, J. (1997) Source identiication of oil spills based on the isotopic composition of individual components in weathered oil samples. Environ. Sci. Technol., 31, 3417–3425. 132. Harris, S.A., Whiticar, M.J., Eek, M.K. (1999) Molecular and isotopic analysis of oils by solid phase microextraction of gasoline range hydrocarbons. Org. Geochem., 30, 721–737. 133. Rogers, K.M., Savard, M.M. (1999) Detection of petroleum contamination in river sediments from Quebec City region using GC–IRMS. Org. Geochem., 30, 1559–1569. 134. Wang, Z., Fingas, M., Page, D.S. (1999) Review: oil spill identiication. J. Chromatogr. A, 843, 369–411. 135. Harris, S.A., Whiticar, M.J., Fowler, M.G. (2003) Classiication of Duvernay sourced oils from central and southern Alberta using compound speciic isotope correlation (CSIC). B. Can. Petrol. Geol., 51(2), 99–125. 136. Jasper, J.P., Edwards, J.S., Ford, L.C., Corry, R.A. (2002) An initial report of the arson stable-isotope analysis project. Poster at the Forensic Isotope Ratio Mass Spectrometry (FIRMS) Conference, September 16–17, 2002, http://www.forensic-isotopes.rdg.ac.uk/conf/conf2002. htm (accessed November 20, 2011). 137. Smallwood, B.J., Philp, R.P., Burgoyne, T.W. (2001) The use of stable isotopes to differentiate speciic source markers for MTBE. Environ. Forensics, 2, 215–221. 138. Oudijk, G. (2008) Compound-speciic stable carbon isotope analysis of MTBE in groundwater contamination ingerprinting studies: the use of hydrogeologic principles to assess its validity. Environ. Forensics, 9(1), 40–54. 139. Philp, R.P., Allen, J., Kuder, T. (2002) The use of the isotopic composition of individual compounds for correlating spilled oils and reined products in the environment with suspected sources. Environ. Forensics, 3, 341–348. 140. Pond, K.L., Huang, Y., Wang, Y., Kulpa, C.F. (2002) Hydrogen isotopic composition of individual n-alkanes as an intrinsic tracer for bioremediation and source identiication of petroleum contamination. Environ. Sci. Technol, 36, 724–728. 141. Oudijk, G. (2005) Fingerprinting and age-dating of gasoline releases—a case study. Environ. Forensics, 6, 91–99. 142. Horita, J., Vass, A.A. (2003) Technical note: stable-isotope ingerprints of biological agents as forensic tools. J. Forensic Sci., 48(1), 1–5. 143. Kreuzer-Martin, H.W., Lott, M.J., Ehleringer, J.R. (2002) Locating the geographic origin of microbial cultures through stable isotope ratio analysis. Poster at the

REFERENCES

144.

145.

146.

147.

148.

149.

150.

151.

152.

153.

154.

155.

156.

Forensic Isotope Ratio Mass Spectrometry (FIRMS) Conference, September 16–17, 2002, http://www.forensicisotopes.rdg.ac.uk/conf/conf2002.htm (accessed November 20, 2011). Kreuzer-Martin, H.W., Chesson, L.A., Lott, M.J., Dorigan, J.V., Ehleringer, J.R. (2004) Stable isotope ratios as a tool in microbial forensics-part 1. microbial isotopic composition as a function of growth medium. J. Forensic Sci., 49(5), 954–960. Kreuzer-Martin, H.W., Chesson, L.A., Lott, M.J., Dorigan, J.V., Ehleringer, J.R. (2004) Stable isotope ratios as a tool in microbial forensics. Part 2. Isotopic variation among different growth media as a tool for sourcing origins of bacterial cells and spores. J. Forensic Sci., 49(5), 1–7. Kreuzer-Martin, H.W., Chesson, L.A., Lott, M.J., Ehleringer, J.R. (2005) Stable isotope ratios as a tool in microbial forensics-Part 3. Effect of culturing on agarcontaining growth media. J. Forensic Sci., 50(6), 1372–1379. Ihle, E., Schmidt, H.-L. (1996) Multielement isotope analysis on drugs of abuse. Possibility for their origin assignment. Isotopes Environ. Health Stud., 32, 226–228. Ehleringer, J.R., Cooper, D.A., Lott, M.J., Cook, C.S. (1999) Geo-location of heroin and cocaine by stable isotope ratios. Forensic Sci. Int., 106, 27–35. Besacier, F., Chaudron-Thozet, H. (1999) Chemical proiling of illicit heroin samples. Forensic Sci. Rev., 11(2), 106–119. Zhang, D., Sun, W., Yuan, Z., Ju, H., Shi, X., Wang, C. (2005) Origin differentiation of a heroin sample and its acetylating agent with 13C isotope ratio mass spectrometry. Eur. J. Mass Spectrom. (Chichester, Eng.), 11, 277–285. Casale, J.F., Ehleringer, J.R., Morello, D.R., Lott, M.J. (2005) Isotopic fractionation of carbon and nitrogen during the illicit processing of cocaine and heroin in South America. J. Forensic Sci., 50(6), 1–7. Casale, J., Casale, E., Collins, M., Morello, D., Cathapermal, S., Panicker, S. (2006) Stable isotope analyses of heroin seized from the merchant vessel Pong Su. J. Forensic Sci., 51(3), 603–606. Ehleringer, J.R., Casale, J.F., Lott, M.J., Ford, V.L. (2000) Tracing the geographical origin of cocaine. Nature, 408, 311–312. Galimov, E.M., Sevastyanov, V.S., Kulbachevskaya, E.V., Golyavin, A.A. (2005) Isotope ratio mass spectrometry: δ13C and δ15N analysis for tracing the origin of illicit drugs. Rapid Commun. Mass Spectrom., 19(10), 1213– 1216. Idoine, F.A., Carter, J.F., Sleeman, R. (2005) Bulk and compound-speciic isotopic characterisation of illicit heroin and cling ilm. Rapid Commun. Mass Spectrom., 19(22), 3207–3215. Liu, J.H., Lin, W.-F., Fitzgerald, M.P., Saxena, S.C., Shieh, Y.N. (1979) Technical note: possible characterization of samples of Cannabis sativa L. by their carbon isotopic distributions. J. Forensic Sci., 24, 814–816.

371

157. Denton, T.M., Schmidt, S., Critchley, C., Stewart, G.R. (2001) Natural abundance of stable carbon and nitrogen isotopes in Cannabis sativa relects growth conditions. Aust. J. Plant Physiol., 28, 1005–1012. 158. Shibuya, E.K., Sarkis, J.E.S., Neto, O.N., Moreira, M.Z., Victoria, R.L. (2006) Sourcing Brazilian marijuana by applying IRMS analysis to seized samples. Forensic Sci. Int., 160(1), 35–43. 159. Shibuya, E.K., Sarkis, J.E.S., Neto, O.N., Ometto, P.H.B. (2007) Multivariate classiication based on chemical and stable isotopic proiles in sourcing the origin of marijuana samples seized in Brazil. J. Braz. Chem. Soc., 18(1), 205–214. 160. West, J.B., Hurley, J.M., Ehleringer, J.R. (2009) Stable isotope ratios of marijuana. I. carbon and nitrogen stable isotopes describe growth conditions. J. Forensic Sci., 54(1), 84–89. 161. Lott, M.J., Howa, J., Ehleringer, J.R., Jauregui, J.F., Douthitt, C. (2002) Detecting the manufacturing origins of pseudoephedrines through stable isotope ratio analysis. Poster at the Forensic Isotope Ratio Mass Spectrometry (FIRMS) Conference, September 16–17, 2002, http:// www.forensic-isotopes.rdg.ac.uk/conf/conf2002.htm (accessed November 20, 2011). 162. Mas, F., Beemsterboer, B., Veltkamp, A.C., Verweij, A.M.A. (1995) Technical note: determination of “common-batch” members in a set of coniscated 3,4-(methylendioxy)-methylamphetamine samples by measuring the natural isotope abundance: a preliminary study. Forensic Sci. Int., 71, 225–231. 163. Palhol, F., Chabrillat, M., Naulet, N., Fourel, F., Dudd, S. (2002) Tracking the origin of illicit MDMA using 15 N/14N stable isotope ratios. Micromass Application Note 514. 164. Carter, J.F., Titterton, E.L., Murray, M., Sleeman, R. (2002) Isotopic characterisation of 3,4-methylenedioxyamphetamine and 3,4-methylenedioxymethylamphetamine (Ecstasy). Analyst, 127, 830–833. 165. Palhol, F., Lamoureux, C., Chabrillat, M., Naulet, N. (2004) 15N/14N isotopic ratio and statistical analysis: an eficient way of linking seized Ecstasy tablets. Anal. Chim. Acta, 510, 1–8. 166. Billault, I., Courant, F., Pasquereau, L., Derrien, S., Robins, R.J., Naulet, N. (2007) Correlation between the synthetic origin of methamphetamine samples and their 15 N and 13C stable isotope ratios. Anal. Chim. Acta, 593(1), 20–29. 167. Buchanan, H.A.S., Daeid, N.N., Meier-Augenstein, W., Kemp, H.F., Kerr, W.J., Middleditch, M. (2008) Emerging use of isotope ratio mass spectrometry as a tool for discrimination of 3,4-methylenedioxymethamphetamine by synthetic route. Anal. Chem., 80(9), 3350–3356. 168. de Korompay, A., Hill, J.C., Carter, J.F., Daeid, N.N., Sleeman, R. (2008) Supported liquid-liquid extraction of the active ingredient (3,4-methylenedioxymethylamphetamine) from Ecstasy tablets for isotopic analysis. J. Chromatogr. A, 1178(1–2), 1–8.

372

FORENSIC APPLICATIONS OF ISOTOPE RATIO MASS SPECTROMETRY

169. Kurashima, N., Makino, Y., Sekita, S., Urano, Y., Nagano, T. (2004) Determination of origin of ephedrine used as precursor for illicit methamphetamine by carbon and nitrogen stable isotope ratio analysis. Anal. Chem., 76, 4233–4236. 170. Kurashima, N., Makino, Y., Urano, Y., Sanuki, K., Ikehara, Y., Nagano, T. (2009) Use of stable isotope ratios for proiling of industrial ephedrine samples: application of hydrogen isotope ratios in combination with carbon and nitrogen. Forensic Sci. Int., 189, 14–18. 171. Schneiders, S., Holdermann, T., Dahlenburg, R. (2009) Comparative analysis of 1-phenyl-2-propanone (P2P), an amphetamine-type stimulant precursor, using stable isotope ratio mass spectrometry. Presented in part as a poster at the 2nd meeting of the Joint European Stable Isotope User Meeting (JESIUM), Giens, France, September 2008. Sci. Justice, 49(2), 94–101. 172. Aguilera, R., Becchi, M., Casabianca, H., Hatton, C.K., Catlin, D.H., Starcevic, B., Pope, H.G. Jr. (1996) Improved method of detection of testosterone abuse by gas chromatography/combustion/isotope ratio mass spectrometry analysis of urinary steroids. J. Mass Spectrom., 31(2), 169–176. 173. Phillips, A. (1999) Gas chromatography combustion isotope ratio mass spectrometry for conirming testosterone administration from urinary androstanediols, Micromass Application Note 501. 174. Shackleton, C.H.L., Phillips, A., Chang, T., Li, Y. (1997) Conirming testosterone administration by isotope ratio mass spectrometric analysis of urinary androstanediols. Steroids, 62, 379–387. 175. Cawley, A.T., Flenker, U. (2008) The application of carbon isotope ratio mass spectrometry to doping control. J. Mass Spectrom., 43(7), 854–864. 176. Cawley, A.T., Hine, E.R., Trout, G.J., George, A.V., Kazlauskas, R. (2004) Searching for new markers of endogenous steroid administration in athletes: “looking outside the metabolic box”. Forensic Sci. Int., 143, 103–114. 177. Saudan, C., Baume, N., Mangin, P., Saugy, M. (2004) Urinary analysis of 16(5α)-andronsten-3α-ol by gas chromatography/combustion/isotope ratio mass spectrometry: implications in anti-doping analysis. J. Chromatogr. B, 810, 157–164.

178. Cawley, A.T., Trout, G.J., Kazlauskas, R., George, A.V. (2008) The detection of androstenedione abuse in sport: a mass spectrometry strategy to identify the 4-hydroxyandrostenedione metabolite. Rapid Commun. Mass Spectrom., 22(24), 4147–4157. 179. Sottas, P.-E., Saudan, C., Schweizer, C., Baume, N., Mangin, P., Saugy, M. (2008) From population- to subjectbased limits of T/E ratio to detect testosterone abuse in elite sports. Forensic Sci. Int., 174(2–3), 166–172. 180. Saudan, C., Augsburger, M., Mangin, P., Saugy, M. (2007) Carbon isotope ratio analysis by gas chromatography/ combustion/isotope ratio mass spectrometry for the detection of gamma-hydroxybutyric acid (GHB) administration to humans. Rapid Commun. Mass Spectrom., 21, 3956–3962. 181. Aguilera, R., Chapman, T.E., Pereira, H., Oliveira, G.C., Illanes, R.P., Fernandes, T.F., Azevedo, D.A., Neto, F.A. (2009) Drug testing data from the 2007 Pan American Games: δ13C values of urinary androsterone, etiocholanolone and androstanediols determined by GC/C/IRMS. J. Steroid Biochem. Mol. Biol., 115(3–5), 107–114. 182. Meier-Augenstein, W. (2004) Laboratory set-up for GC-MS and continuous-low IRMS. In Handbook of Stable Isotope Analytical Techniques, Vol. 1, edited by de Groot, P.A. Amsterdam, The Netherlands: Elsevier B.V, pp. 1038–1042. 183. Thermo Fisher Scientiic, Waltham, MA (2003) Service Information Number IS-099. Subject: Reference Gases of all IRMS Peripherals. Issue date: 27 Feb 2003. Approved by Peter Haubold. 184. National Association of Testing Authorities, Australia (NATA). (2009) Guidelines for the validation and veriication of chemical test methods. Technical Note 17, April 2009, http://www.nata.com.au/phocadownload/publications/ Technical_publications/Technotes_Infopapers/technical_ note_17.pdf (accessed November 20, 2011). 185. National Association of Testing Authorities, Australia (NATA). (2009) Guidelines for estimating and reporting measurement uncertainty of chemical test results. Technical Note 33, December 2009, http://www.nata. com.au/phocadownload/publications/Technical_ p u b l i c a t i o n s / Te c h n o t e s _ I n f o p a p e r s / t e c h n i c a l _ note_33.pdf (accessed November 20, 2011).

16 ANALYSIS OF TRIACETONE TRIPEROXIDE EXPLOSIVE BY MASS SPECTROMETRY Michael E. Sigman and C. Douglas Clark

16.1

INTRODUCTION

In 1895, Wolfenstein irst synthesized 3,3,6,6,9,9hexamethyl-1,2,4,5,7,8-hexaoxacyclononane, otherwise known as triacetone triperoxide (TATP) [1]. TATP is classiied as a primary high explosive, meaning that it is highly sensitive to heat and friction. Owing to its instability, the peroxide is of little or no industrial or military use and it is not possible to purchase bulk quantities from commercial sources. Dilute solutions of analytical standards are available commercially. Despite the hazards associated with the synthesis and handling of TATP, the material is readily prepared from commercially available chemicals. Recipes of varying levels of detail for the preparation of TATP can be found on the Internet. The reader should note that the preparation and possession of TATP without proper licensing and knowledge can be hazardous and illegal. TATP found increased use by terrorists as an improvised explosive at the end of the twentieth century and early in the twenty-irst century. The increased use of such a dangerous material may be attributed to several factors, including the availability of Internet recipes, commercially available starting materials, and tightened controls on commercial and military high explosives which makes their acquisition increasingly dificult. A brief review of incidents that involve TATP worldwide prior to 2001 has been given by Oxley [2]. Over the past 10 years there have been a number of reports

in the popular press of incidents involving TATP; some of the more noteworthy ones are briely discussed here. On August 9, 2001, a suicide bomber outside of a Sbarro’s pizza restaurant in Jerusalem detonated a TATP explosive belt, killing 15 people and injuring in excess of 90 people [3]. Later that same year, Richard Reid failed to detonate a shoe illed with explosives by lighting a TATP detonator [4,5]. On July 7, 2005, attacks were carried out on the London underground subway and bus system using TATP detonators [5]. The attacks resulted in 55 deaths, including 52 innocent passengers and three suicide bombers [6]. A University of Oklahoma student killed himself on October 1, 2005, with a TATP improvised explosive device [7]. In 2006, one man was killed and another man severely injured in a TATP manufacturing mishap in Texas City, Texas [8]. In September of the following year, German police disrupted a terrorist cell linked to an Islamic Jihad militant group. The cell had amassed 1500 lb. of hydrogen peroxide to be used in the preparation of improvised explosive devices [9]. In May 2008, TATP was detected on two workers at the Oskarshamn nuclear power plant in Sweden [10]. Afghan immigrant Najibullah Zazi was arrested in September 2009, on suspected terrorism charges in connection with the attempted synthesis of TATP. Mr. Zazi pleaded guilty in early 2010 [11]. On Christmas day, in 2009, Umar Farouk Abdulmutallab attempted to detonate a device constructed from pentaerythritol tetranitrate (PETN) with TATP initiation onboard an international light from Amsterdam, The

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

373

374

ANALYSIS OF TRIACETONE TRIPEROXIDE EXPLOSIVE BY MASS SPECTROMETRY

Netherlands, to Detroit, Michigan, in the United States [12]. The importance of detection and forensic analysis of TATP is abundantly clear from the nature and number of incidents involving the material. Details of the synthesis and chemical characteristics of TATP will be presented prior to a discussion of methodologies for the analysis of this improvised explosive.

16.2

BACKGROUND

16.2.1 TATP Synthetic Mechanism The irst published synthesis of TATP was in 1895 and involved the reaction of an uncatalyzed mixture of acetone and hydrogen peroxide [1]. In the late 1950s, Milas and Golubovic´ published several studies on the acid catalyzed reaction of hydrogen peroxide with acetone and several other ketones [13–15]. The reactions generated a mixture of products, some of which were isolable. Tentative structures were assigned using the analytical structure proof tools available at the time, including infrared spectroscopy and derivative melting point determination. The acetone product mixture was proposed to contain 2,2-dihydroperoxypropane, 2,2′-dihydroperoxydipropylperoxide, the cyclic trimer (TATP), and other higher mass polymeric peroxides terminated on each end of the chain as hydroperoxides. Sauer et al. utilized ultraviolet (UV) absorption and nuclear magnetic resonance (NMR) spectroscopy in the early 1970s to examine the catalyzed and uncatalyzed reactions of acetone with hydrogen peroxide [16,17]. His studies led to the proposal of reaction intermediates 2,2′-bis(hydroperoxy)propane and 2,2′-bis(hydroperoxy)2,2′-diisopropylperoxide. A mechanism for the uncatalyzed reaction was proposed which would account for the formation of the cyclic dimmer, diacetone diperoxide (DADP), TATP, and hydroperoxy-terminated oligoperoxides of varying chain length [18]. Figure 16.1 illustrates a detailed mechanism for the production of TATP which is based on the work of Hiatt. Key steps in the reaction include the uncatalyzed addition of hydrogen peroxide or an organic hydroperoxide to acetone in lines 1 and 3. Acid-catalyzed condensation of hydrogen peroxide with hydroxyl groups to form hydroperoxides occurs in lines 2, 4, 5, and 8. Two intermediates were proposed to undergo intramolecular cyclization to form DADP (line 5) and TATP (line 7). It is important to note that alternative mechanistic steps could be envisioned. For example, lines 3 and 4 could be replaced by an acid catalyzed condensation of two moles of 2-hydroxy-2hydorperoxypropane and a subsequent catalyzed condensation with hydrogen peroxide. Additionally, the

cyclic tetramer has been observed experimentally, though it is not accounted for by the scheme in Figure 16.1. In 2009, Jensen reported the rate-determining step in the acid-catalyzed synthesis was ring closure to form TATP from the open-chain intermediate [19]. The linear hydroperoxide-terminated oligoperoxides have also been observed as in the ozonolysis of tetramethylethylene [20]. The oligoperoxides were identiied as their ammonium adducts by electrospray ionization mass spectrometry (ESI-MS) [21]. The exact nature of the open-chain intermediates and the conditions under which they are formed remain the subject of investigation in several laboratories. The results from synthesis and structure proof studies demonstrate that the product resulting from a TATP synthesis may be a complex mixture. Furthermore, the products are rich in peroxide functionality and make them relatively unstable, and therefore, unsafe. The thermal-induced decomposition of TATP has been studied and is reported to produce primarily acetone, carbon dioxide, and ozone as decomposition products [22–25]. TATP from syntheses catalyzed with sulfuric acid, methanesulfonic acid, and perchloric acid was reported to spontaneously decompose over time to yield DADP, even at low temperatures (i.e., 0°C) [26]. While it may be suficient in many cases to simply identify the presence of TATP, forensic examination may proit from a more complete analysis of the product mixture that comprises unpuriied synthetic TATP. Unpuriied TATP may be encountered in terrorism investigations, industrial accidents, and other events that involve the possible formation of organic peroxides from acetone and related ketones. 16.2.2 Analytically Important Physical and Chemical Properties of TATP 16.2.2.1 TATP and Oligoperoxide–Cation Binding Computational chemistry has been a valuable asset in understanding the gas phase behavior of TATP, especially the interaction with cations which play an important role in the observed behavior in a mass spectrometer. Mass spectral analysis of TATP and the oligomeric peroxides is facilitated by the formation of cation adducts, as depicted in Figure 16.2. The cation complexes result from donation of electron density from the highest occupied molecular orbital (HOMO) of the peroxide to electron-deicient cation. Figure 16.2A illustrates the calculated (density functional theory [DFT] B88LYP/ DVZP) HOMO of TATP, with the electron density primarily located on the peroxide oxygen atoms. Three oxygens are on each side of a plane passing through the three methylene carbons of the ring. The three methyl groups above and below this plane may lead to steric

BACKGROUND

FIGURE 16.1 peroxide.

375

Synthetic mechanism for the production of TATP by acid-catalyzed reaction of acetone with hydrogen

FIGURE 16.2 (A) DFT (B88LYP/DVZP) calculated HOMO of TATP, (B) calculated [TATP + NH4]+ complex structure shown as a space-illing model, and (C) the [TATP + NH4]+ complex depicted as a ball-and-cylinder model viewed from the side.

376

ANALYSIS OF TRIACETONE TRIPEROXIDE EXPLOSIVE BY MASS SPECTROMETRY

FIGURE 16.3 DFT (BLYP/6-31G**+CRENBL ECP) calculated bond lengths (Å) and partial charge (δ) for (A) TATP and (B) [TATP + Cu]+. Data adapted from Reference [32].

interactions that inhibit complexation with larger cations. Figure 16.2B illustrates the [TATP + NH4]+ complex as a space-illing model and Figure 16.2C illustrates the complex depicted as a ball-and-cylinder model viewed from the side. Complex formation between TATP and sodium cation has been calculated to proceed with a decrease in the σO–O bond order from 0.9697 to 0.9675 and associated O–O bond length increase upon complex formation from 1.462 to 1.471 Å [27]. The calculated binding energy for the [TATP + Na]+ complex was calculated to be 40.2 kcal/mol, which is signiicantly higher than the 30.5 kcal/mol calculated (DFT B3LYP/6-31 +G(d) level) dissociation energy of the O–O bond in the isolated peroxide. The calculation results suggested that under collision-induced dissociation (CID), the sodium cation would remain associated with the fragmented TATP, rather than simply dissociating and leaving the TATP ring intact. Different levels of theory give different results for bond angles and distances. For example, the peroxide bond length in TATP has been calculated to be 1.303, 1.557, 1.459, and 1.529 Å, respectively, by semiempirical AM1 and PM3, Hartree–Fock/6-31G(d,p), and DFT/DZP BLYP levels of theory [28,29]. In a related report, the interaction energies for TATP with ammonium and sodium ions were calculated to be 26 and 42 kcal/mol, respectively, by DFT at the B88LYP level with the DZVP basis set [30]. Unlike in the case of sodium, the interaction energy for ammonium ion with TATP was calculated to be less than the peroxide bond strength, 36.5 kcal/mol, which manifests itself in the behavior of the complex under CID (vide infra). As an interesting note, the experimental O–O bond lengths from X-ray studies, 1.483 ± 0.005 Å, lie between the Hartree–Fock/6-31G(d,p) and the DFT/DZP BLYP values [28]. More recent X-ray analysis and higherlevel computational modeling of TATP by DFT with the B3LYP correlation functional approximation and the correlation-consistent polarized valence doubleξ (cc-pVDZ) basis set gave 1.422 Å O–O bond lengths, in comparison to the X-ray measured values of

1.418 ± 0.001 Å [24]. Subsequent X-ray studies have found TATP to be polymorphic [31]. The binding energy of a series of cations with TATP has also been calculated with and without a zero-point energy correction using a variety of methods, including Hartree–Fock and DFT [32]. The ordering of the binding energies varied somewhat from one computational method to the next. As an example, the ordering obtained from DFT/BeLYP/6-31G**+CRENBL ECP was Cu+ < Li+ < Cd2+ < Zn2+ < In3+ < Sb3+ < Sc3+ < Ti4+. The reader is directed to the appropriate reference for more information on the computational methods. The stronger binding cations were found to lead to calculated cleavage of the TATP ring structure. Some of the calculated changes upon complex formation with Cu+ ion are shown in Figure 16.3 and the general trends observed from these calculated results are illustrated [32]. The bond lengths are given in angstroms (Å) and shown with dashed arrows, and the partial charges on select atoms are designated as δ and shown by solid arrows. Figure 16.3A illustrates the bond lengths and partial charges calculated for TATP without complexing with a cation. Figure 16.3B illustrates that upon complexation with Cu+, the peroxide bond undergoes a 0.015 Å increase in length, while the bond between the ring carbon and the oxygen complexing to oxygen undergoes an even greater increase in length, 0.035 Å. Yet the bond length between the ring carbon and the oxygen that is not complexing with the cation is relatively unchanged in length. This result was observed across all TATP–cation complexes calculated and suggests that complex formation weakens the bonds between the three-ring carbons and the oxygen atoms that interact with the cation. Binding with the cation leads to a more negative partial charge on the oxygens complexing the cation and a less negative partial charge on the oxygens not interacting directly with the metal. A signiicantly larger increase in partial positive charge is observed on the methyl carbons upon association with the cation, a Δδ = 0.1 for TATP-Cu+ complex formation, indicating a signiicant redistribution of the cation posi-

ANALYTICAL METHODS

tive charge. Stronger binding was shown for smaller cations and for cations with higher formal charges. Furthermore, the oxygen–cation distances in the complexes were seen to correlate with the ionic radius of the cations [32]. Similar calculations have given the interactions energies for [H(OOCH2)nOOH + NH4]+ (n = 2, 3, 4) as 39, 45, and 54 kcal/mol, respectively. The calculated interaction energies for [H(OOCH2)nOOH + Na]+ (n = 2, 3, 4) were reported to be somewhat higher at 56, 77, and 80 kcal/mol, respectively [33]. These complexes are observed by ESI- and atmospheric pressure chemical ionization–mass spectrometry (APCI-MS), as discussed below.

16.3 ANALYTICAL METHODS 16.3.1

Gas Chromatography–Mass Spectrometry

16.3.1.1 Electron Ionization Zitrin et al. reported one of the earliest mass spectra of TATP, which was dominated by a mass-to-charge ratio (m/z) 43(100%) ion [34]. In a later report on the forensic analysis of a TATP sample using a linear quadrupole instrument, the electron ionization (EI) spectrum of the TATP contained ions at m/z 43 (100%), 58, 59 and 75 [35]. In a more extensive report, TATP was synthesized, puriied, and subsequently characterized by gas chromatography– mass spectrometry (GC-MS), gas chromatography– Fourier transform infrared spectroscopy (GC-FTIR), Raman spectroscopy, and NMR spectroscopy [36]. The following ions were reported observed from the gas chromatography–mass spectrometry with electron ionization (GC-EI-MS) analysis of TATP: m/z 43(100%), 57, 58, 59, 73, 74, 75, 101, and 117, and the molecular ion at 222(0.07%). A stronger molecular ion has been observed in the EI spectrum of TATP produced in a supersonic expansion of helium [37]; however, under more commonly encountered conditions in commercial spectrometers, the molecular ion is generally weak. A linear quadrupole EI spectrum of TATP is shown in Figure 16.4A [30]. Although the parent ion is often weak, as reported by several groups, a subnanogram limit of detection (LOD) can be obtained by GC-EI-MS analysis of liquid samples of TATP. A 50 pg LOD for TATP has been reported using single ion monitoring (SIM) of m/z 43, 59, 75, 101, and 222, under conditions where the GC-MS analyses was carried out at an injector temperature of 120°C and EI source temperature of 200°C [36]. Linear quadrupole and quadrupole ion trap instrument have been reported to yield 50 and 100 pg LOD based on extracted ions m/z 43, 59, and 75 under optimized conditions of injection

377

port temperature and source temperature of 110 and 100°C, respectively [30]. Muller et al. have reported 1.6 and 10 ng LOD values for analysis of chloroform solutions of TATP using injection port and source temperatures of 250 and 120°C, respectively, on an ion trap and 180 and 230°C, respectively, on a linear quadrupole [38]. The LOD from liquid injection samples is generally optimized by lowering the injection port and detector source temperatures to around 100°C. The optimized GC-MS parameters for TATP analysis from one report are given in Table 16.1 [30]. In addition to liquid injections, vapor sampling has been demonstrated to give low LOD values using EI. A 100 pg LOD was reported by Stambouli et al. for direct injection of a vapor sample of TATP, based on the m/z 43 extracted ion chromatogram [39]. Detection of TATP by solid-phase microextraction (SPME) has been reported for 64 ng/L of analyte in nylon bags [38]. Similarly, Kende et al. reported a 5 ng optimized LOD for a 10-min sampling at 60°C by SPME (100 µm polydimethyl siloxane) sampling of TATP [40]. 16.3.1.2 Chemical Ionization The absence of a prominent molecular ion in EI mass spectra may favor the use of chemical ionization methods coupled with GC separations. The spectra resulting from different chemical ionization modes, reagent gases, and mass analyzers can vary signiicantly. Positive ion chemical ionization (PICI) of TATP has been reported using methane, isobutene, and ammonia. The methane PICI spectrum of TATP obtained with a linear quadrupole mass analyzer has been reported to contain ions of m/z 103, 117, 133, and 223, corresponding to the pseudomolecular ion [TATP + H]+ [35]. A fullscan methane PICI spectrum of TATP has also been reported to contain ions of m/z 43(100%), 59, 74, 75, 91, and 223(0.5 µm >0.5 µm

0.05–1 0.1–2 1–30 1–10 15–25 1–10 10–20

Source: Mayer, K., Wallenius, M., Ray, I. (2005) Nuclear forensics––a methodology providing clues on the origin of illicitly traficked nuclear materials. Analyst, 130(4), 433–441.

cles, in addition to scientiic reports and guidelines published by governmental organizations, such as the International Atomic Energy Agency (IAEA), the U.S. Nuclear Regulatory Commission (NRC), and DHS’s Domestic Nuclear Detection Ofice, serve as a guide for the present and future use of MS for the detection of nuclear and radiological materials.

20.5

EXPLOSIVE THREATS

20.5.1 Background on Explosives and Explosive Threat Detection The analysis of explosives has been investigated since the 1970s. The rise of global terrorism in the past few decades has increased the urgency for more selective and sensitive explosive detection methods. The analysis of explosives by MS has been reviewed according to the type and application of each technique [123–124]. A list of basic explosive compounds is compiled by the U.S. Bureau of Alcohol, Tobacco, and Firearms [125]. This list indicates the wide array of compounds that are available to create an explosion. However, the most common explosive compounds are listed in Table 20.11 along with their chemical and common names, chemical formula, MW, and common use. In addition to these explosive compounds, several previously unusual explosives are now being used more often in IEDs [125–127]. The sample collection, ionization, and analysis of explosives can be challenging [128]. There are several physical limitations on the detection of explosives

EXPLOSIVE THREATS

TABLE 20.11 Typical Uses

461

Common Explosives with Abbreviated Common Name, Molecular Formula, Molecular Weight (MW), and

Explosive Compound Organic aromatics 2,4,6-Trinitrotoluene 2,4,6-Nitrophenol Organic aliphatics Nitroglycerin Pentaerythritol tetranitrate Ethylene glycol dinitrate 1,3,5-Trinitro-1,3,5-triazine 1,3,5,7-Tetranitro-1,3,5,7tetraazacyclooctane Nitroguandine Cyclic peroxides Triacetone triperoxide Hexamethylene triperoxide diamine Inorganic Ammonium nitrate

Common Name

Molecular Formula

MW (g/mol)

Use – Castable explosives Munitions

TNT Picric acid

C7H5N3O6 C6H3N3O7

227.15 229.00

NG PETN

C3H5N3O9 C5H8N4O12

227.09 316.14

EGDN RDX HMX

C2H4N2O6 C3H6N6O6 C4H8N8O8

152.10 222.03 269.16



C1H4N4O2

104.03

Dynamite, industrial explosive Detonators, sheet explosives, Semtex, blasting caps NG/EGDN mixtures C-4, sheet explosives, Semtex Artillery shells, rocket propellants Triple-based propellants

TATP HMTD

C9H18O6 C6H12N2O6

222.11 210.19

Homemade explosive Homemade explosive

AN

NH4NO3

80.02

Homemade explosive, industrial explosive

Sources: Cooper, P.W., Kurkowski, S.R. (1996) Introduction to the Technology of Explosives. Wiley-VCH; and Yinon, J., Zitrin, S. (1993) Modern Methods and Applications in Analysis of Explosives. Wiley.

including low vapor pressures for the nitramine and nitroaromatic explosives, the amount of sample available for collection, intentional concealment of the explosive material, and matrix interferences from the sample and sample collection media [128]. These challenges must be addressed and managed in any analysis of explosives. 20.5.2

GC-MS for Explosive Threat Detection

GC-MS is used to detect both common and emerging explosive compounds. A review of GC-MS methods used to detect organic explosive compounds is available [129]. Two common GC-MS sample introduction techniques are solid-phase microextraction (SPME) and headspace vapor collection [130–133]. These sample introduction methods are employed in the analysis of water and soil samples with suspected explosive residue contamination [134–137]. The U.S. EPA Method 8095, “Explosives by Gas Chromatography,” is recommended as a resource for sample preparation of soil and water samples analyzed for the common nitroaromatic, nitramine, and nitrate ester explosive compounds [138]. GC-MS is also used to identify and characterize vapors from trace residue produced by various explosive compounds [12]. Elemental sulfur in explosive powders is determined by this method and is 400 times more sensitive than a standard chemical color test for

sulfur [139]. GC-MS is used to verify the identity of explosives in combination with Fourier transform infrared spectroscopy (FTIR) and high performance thin layer chromatography (HPTLC) [140–142]. GC-MS is also useful in studying the effect of storage and aging on common explosive compounds [137, 143]. GC-MS is used to characterize emerging explosive compounds including cyclic organic peroxides such as triacetone triperoxide (TATP) and hexamethylene triperoxide diamine (HMTD); this method may be chosen over similar techniques due to the higher vapor pressures of cyclic peroxides over the more common nitratecontaining explosives [144–149]. 20.5.3

LC-MS for Explosive Threat Detection

LC-MS is another widely used analytical method for the analysis of explosives. Several articles review LC-MS for explosives analysis [150–152]. In particular, Yinon and his coworkers have signiicantly contributed to the development of explosives analysis by LC-MS. Zhao and Yinon identiied nitrate ester explosives by LC-MS using both ESI and APCI [153]. In both ionization studies, postcolumn additives are introduced to increase speciicity and sensitivity of the product ions produced by the explosives pentaerythritol tetranitrate (PETN), nitroglycerin (NG), and ethylene glycol dinitrate (EGDN) to form stable adduct ions (as shown in Figure

462

HOMELAND SECURITY

EGDN m/z 214

NG

m/z 289 PETN

m/z 378 0.0

1.0

2.0

3.0 4.0 5.0 Retention Time (min)

6.0

7.0

FIGURE 20.15 Extracted ion chromatograms from the LCESI-MS analysis of three explosives in a mixture with addition of ammonium nitrate after the LC separation: 2 ng/mL EGDN, 10 pg/mL NG, and 10 pg/mL PETN. Source: Zhao, X., Yinon J. (2002) Identiication of nitrate ester explosives by liquid chromatography–electrospray ionization and atmospheric pressure chemical ionization mass spectrometry. Journal of Chromatography A, 977, 59–68.

both ESI and APCI sources for LC-MS explosives analyses was compared [162]. Variations of ESI were used to analyze explosives by LC-MS including DESI, SESI, and extractive electrospray ionization (EESI). DESI, with the addition of adduct-forming additives, was used to enhance the sensitivity and selectivity of MS and MS/MS analyses for explosives including 2,4,6-trinitrotoluene (TNT), 1,3,5-trinitroperhydro-1,3,5-triazine (RDX), and 1,3,5,7-tetranitro-1,3,5,7-tetraazacyclooctane (HMX) [163]. DESI was also used to detect trace amounts of TNT, RDX, HMX, and PETN on human skin using a methanol– water spray solution doped with sodium chloride as an adduct-forming additive [164]. A SESI source was coupled to an SCIEX API-365 triple quadrupole MS, an SCIEX APR-5000 triple quadrupole MS, and an SCIEX Q-Star QTOFMS to obtain low detection limits for PETN and TNT at 0.2 and 0.4 ppt, respectively. Atmospheric pressure photoionization (APPI) is coupled to a LC-MS system to achieve greater sensitivity and gains in dynamic range over GC-MS or LCAPCI-MS techniques [165–166]. API sources are also coupled to LC systems to achieve selective detection of semivolatile and thermolabile compounds. 20.5.4

20.15). In the ESI-MS studies, ammonium nitrate, sodium nitrite, propionic acid, and ammonium chloride are used as additives; dichloromethane, chloroform, carbon tetrachloride, and ammonium chloride were used in the APCI-MS studies. Zhao and Yinon achieved the lowest detection limits with ESI by adding ammonium nitrate after the LC separation step. The authors also reported detection limits for PETN, NG, and EGDN at 5 pg/µL, 5 pg/µL, and 2 ng/µL, respectively. MS/MS with CID is used to conirm the identity of the adduct ions found by each method. They also noted that EGDN was dificult to analyze by LC-ESI-MS or LC-APCI-MS due to its high vapor pressure and was not detected by either method without the addition of a postcolumn additive. LC-MS is also used to detect cyclic organic peroxides, a class of compounds that contain the liquid explosives TATP, and HMTD [154–155]. Analysis of cyclic organic peroxides is problematic due to their thermolability and easy degradation, making any analysis challenging for both LC-MS and GC-MS. LC-MS/MS identiies degradation products from these compounds [154]. Considerable work is still needed to successfully detect these compounds and their degradation products by MS methods [154–155]. ESI is used in many LC-MS and LC-MS/MS systems for explosives analysis [9, 156–161]. The performance of

IMS for Explosive Threat Detection

20.5.4.1 Background on IMS for Explosive Threat Detection IMS is the major chemical trace detection technique used for explosives detection since the 1970s. The detection of TNT in picogram quantities by IMS (called plasma chromatography at the time) was irst demonstrated in 1974 [44]. Since then, a community of researchers has developed the technique so that an IM system can analyze a sample in less than 5 s, is easy to operate in the ield, has sub-part per billion detection limits for many explosives and other chemicals, and operates at atmospheric pressure with air as the carrier gas. The irst commercial ion mobility spectrometer was manufactured and marketed by Martin Cohen at PCP, Inc., in 1975 (West Palm Beach, FL, now part of SAES Getters [Milan, Italy]) [167]. These instruments were designed for research lab use and for coupling with a mass spectrometer. The advent of smaller, more portable ion mobility spectrometers came after the bombing of Pan Am light 103 over Lockerbie, Scotland, in 1988 [168]. After this airplane bombing (where a small amount of Semtex explosive was smuggled aboard inside luggage), transportation security initiatives in the United States and elsewhere created the need for trace explosive detectors at airport security checkpoints. Table 20.12 shows a list of several current commercial IMS systems available for trace explosive detection. All of the systems use IMS to detect explosive particles and

EXPLOSIVE THREATS

463

TABLE 20.12 Survey of Commercially Available Trace Detection Systems Using Ion Mobility Spectrometry (IMS) for Explosives Detection Explosive Detection System

Instrument Type

Manufacturer

EntryScan series Easytec-XP EGIS defender GA 2100 Itemiser 3 Itemiser DX Ionscan 400B Ionscan Document Scanner Sabre 4000 Sentinel II MO-2D MO-2M MobileTrace Quantum Sniffer QS-H150 VaporTracer

Personnel portal Field portable Portable (GC-IMS) Portable Portable Portable Portable Portable document scanner Field portable Personnel portal Portable document scanner Field portable Field portable Field portable Field portable

Morpho Detection, Inc. Ion Applications, Inc. Thermo Scientiic, Inc. Excellims Corporation Morpho Detection, Inc. Morpho Detection, Inc. Smiths Detection, Inc. Smiths Detection, Inc Smiths Detection, Inc. Smiths Detection, Inc Sibel Ltd. Sibel Ltd. Morpho Detection, Inc. Implant Sciences Corporation Morpho Detection, Inc.

NB: This is a sampling of the most widely available devices in 2010 and is not an exhaustive list. Source: FEMA. (2009) Responder’s knowledge base. U.S. Department of Homeland Security (accessed June 10, 2010).

vapors (explosive trace residue) at trace levels but are conigured as part of a ield portable, portable, or personnel portal setup. Field portable units are usually handheld devices with self-contained power systems and onboard air sample collectors for explosive vapors. Portable units have an instrument footprint of a desktop computer and plug into typical building wall outlets. Samples are typically collected with small pieces of cloth, paper, or similar by wiping of surfaces to pick up any trace explosive residue. These systems may also be used to scan packages and documents for mail bombs. Personnel portals are designed to collect trace explosive residue from people and clothing by sampling large volumes of air through IMS systems; these systems are typically deployed at transportation security checkpoints and other controlled access areas [169–170]. Consequently, commercial IMS systems in various conigurations are deployed across the world at airport and other security checkpoints. 20.5.4.2 Various Ionization Methods for Trace Explosive Threat Detection by IMS Several articles and books review IMS for trace explosives detection [18, 45, 48]. In particular, a review article, describes many of the product ion formation reactions that create IMS product ions from explosive compounds [45]. The review also contains extensive tables of reduced mobility (K0) values, which are used to identify ions in IMS, for the most common explosive compounds. Traditionally, negative mode nickel-63 is the ionization source of choice for many commercial IMS trace

explosive detection systems because it requires no external power supply, little or no maintenance, and is stable and reliable over long periods of time [18]. The majority of explosives are detected in the negative ionization mode because of their relatively high electronegativities [45]. This source produces quickly thermalized β− particles, which interact with the IMS neutral drift gas to form reactant ions. Typical negative ionization mode reactant ions are thermalized electrons in nitrogen drift gas and O2− or CO2− in air drift gas [171]. Often, dichloromethane is introduced as a neutral vapor into the IMS ion–molecule reaction region to produce chloride reactant ion chemistry [172–173]. These secondary reactant ions interact with the explosive sample vapors by gas-phase ion–molecule reactions. Chloride ion chemistry is used to form predictable product ions from explosive vapors through proton abstraction, charge transfer, or adduct formation reactions [174]; other halide reactant ion chemistries have also been investigated [175–176]. However, even with chloride reactant ion chemistry, the nickel-63 source causes unacceptable rates of false-positive and -negative responses in commercial systems. This problem, coupled with regulatory, inancial, and organizational concerns about using a radioactive source in commercial systems, has led to recent research and development of alternate ionization sources for the detection of explosives by IMS. Atmospheric pressure CDI is used as an alternate ionization source for explosives detection in IMS [177–181]. ESI and its variation, SESI, are also used to

464

HOMELAND SECURITY

detect explosives with IMS [182–184]. A commercial, portable IMS with an ESI source is now available from Excellims Corporation (Acton, MA). ESI does not require volatilization (or desorption) of the explosive sample prior to ionization since ESI is capable of ionizing nonvolatile and inorganic analytes. ESI and SESI may also be a useful way to introduce chloride reactant ion chemistry into the IMS rather than as a vapor, thereby decreasing the complexity of the overall apparatus. A novel distributed plasma ionization source (DPIS) is also used to detect explosives with IMS and could be used as an alternative to nickel-63 sources in the future due to its low power requirements [185]. 20.5.5

IM-MS for Explosive Threat Detection

While IMS and MS are both widely used for explosives analysis, hybrid IM-MS instruments have recently been applied to detect explosives [186]. The lack of HLSrelated research with IM-MS systems may be in part due to the availability of commercial IM-MS systems [183, 187]. However, this hybrid technique offers a distinct advantage over both IMS and MS alone: the ability to simultaneously separate samples by both mobility and mass. This twofold separation mechanism greatly decreases the likelihood of a mass or mobility interferent masking the signal of the analyte of interest. In complex, real-world samples where matrix effects may signiicantly inhibit the detection of trace amounts of explosive material, the ability to separate in two dimensions (2D) is extremely powerful [188]. Figure 20.16 shows a typical 2D IM-MS plot of black powder with

mobility drift time (µs)

12,000 11,000 10,000 9000 8000 7000 6000 50

100 m/z

150

200

FIGURE 20.16 Two-dimensional ion mobility-mass spectra of black powder with mass spectra (top) from 10 to 200 Th and mobility spectra (right) with a drift time range of 6000 µs (6 ms)—12,000 µs (12 ms). Source: Crawford, C.L., et al. (2010) Analysis of black powder by ion mobility-time-of-light mass spectrometry. Analytical Chemistry, 82(1), 387–393.

the mobility drift time on the vertical axis and the m/z range on the horizontal axis. The corresponding mobility spectrum is on the right side of the plot with the mass spectrum on the top of the plot. An IM-quadMS was the irst hybrid instrument used to detect nitrotoluene compounds with a radioactive ionization source [189]. Later work uses laser desorption ionization (LDI) to detect TNT, DNT, RDX, and HMX with an IM-quadMS [190]. Recently, the formation reaction mechanism of RDX response ions in an atmospheric pressure CDI source is now better understood with the use of an IM–triple quadrupole mass spectrometer [179]. 20.5.5.1 Differential Mobility–MS for Explosive Threat Detection Differential mobility spectrometry (DMS), also known as FAIMS, is a technique closely related to IMS [191–192]. In this system, the ratio of the electric ield strength (E, V/cm) applied to the electrodes to the drift gas number density (N, cm−3) is increased to a level beyond that used in DT-IMS (the most common coniguration of IMS systems) so that the mobility of the ion (K) is no longer constant but is dependent on the strength of E/N (Townsend [Td]) [192]. DMS separation is essentially an atmospheric pressure ion ilter and is analogous to a quadrupole mass analyzer’s separation of ions at reduced pressure [193– 194]. Typical instruments are constructed out of two conducting parallel plates (DMS) or concentric tube electrodes, also known as FAIMS. Figure 20.17 shows a diagram of the simplest parallel plate design where an ion (shown as a circle with a cross) enters in between the parallel metal plates via the sample gas. Ions oscillate between the two parallel plates as the electric ield switches back and forth from a high voltage (+4000 V) to a low voltage (−2000 V) for different periods of time [193]. The voltage is held at +4000 V for half as long (thigh) as the time the voltage is held at −2000 V (tlow). Thus, the ion travels very quickly but for a short time period when the voltage is high, and travels more slowly but for a longer time period when the voltage is low. The voltage difference between 0 V and Vhigh is called the dispersion voltage (DV) because it serves to “disperse” or separate ions with different changes in their mobility. Repetitive cycles of alternating the electric ield eventually will cause the ion to hit the lower plate; this behavior is illustrated by the ion’s path between the plates as shown in Figure 20.17. If the ion strikes the lower plate before it reaches the collecting electrode for detection, then the ion will neutralize and not be detected. A much weaker electric ield is added to the DV to prevent the ion from striking the lower plate; this electric ield is called the compensation voltage (CV). The CV “com-

EXPLOSIVE THREATS 0.4 µs +4000 V DV

thigh 0.8 µs

0V –2000 V tlow

FIGURE 20.17 Diagram of an ion’s motion as it travels between two parallel plates in a FAIMS analyzer. An illustration of the asymmetric waveform (top) illustrates how the application of the waveform to the upper electrode inluences the motion of the ion. Source: Guevremont, R. (2004) Highield asymmetric waveform ion mobility spectrometry: a new tool for mass spectrometry. Journal of Chromatography A, 1058, 3–19.

pensates” or corrects for the ion’s travel toward the lower plate by increasing the ion’s energy so that it has just enough velocity to hit the collecting electrode for detection. If the CV is scanned so that it increases from a low value to a higher value, ions with differing changes in mobility are collected for detection [195]. A proof-of-concept FAIMS-IT-MS is used to analyze TNT, trinitrobenzene (TNB), Tetryl, RDX, HMX, PETN, and NG, which are introduced as part of a continuous vapor stream [196]. A commercially available DMS, the Sionex microDMX (Bedford, MA), is coupled to an MS; the DMS technology has a lower cost and greatly decreases analysis times over LC-MS [197]. 20.5.6 Miscellaneous MS Methods for Explosive Threat Detection A wide variety of other MS techniques are used to detect explosives. Two notable techniques are Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR-MS) and CE-MS. FT-ICR-MS is used to probe pseudomolecular ion formation of RDX, PETN, and TNT using several ionization sources including LDI, EI, electron capture ionization (EC), and chemical ionization (CI). Analyses are performed both in the positive and negative ionization mode, and identities are assigned to the major pseudomolecular ion peaks seen in the spectra from each explosive [198]. The composition of several explosive compounds from postblast residue is assessed with FT-ICR-MS by identifying the explosive and inactive ingredients in a smokeless powder, TNT,

465

and C-4. These investigations have led to more accurate assignments of ion identities generated from the above compounds [156]; however, analysis costs remain extremely high for samples analyzed with FT-ICR-MS. An Agilent 3D CE coupled to a Bruker Esquire 3000+ quadrupole ion trap (QIT)-MS with a sulfobutylethercyclodextrin-ammonium acetate separation buffer at pH 6.9 is used to analyze TNT, TNB, RDX, HMX, and CL-20 [199]. CE coupled to a mass spectrometer with ESI enables better resolution than LC separation and thus enhanced identiication of moderate polarity nitramine explosives and their degradation products from soil and water samples. Various ionization sources are employed in the analysis of explosives by MS. APCI is used with GC-MS, LC-MS, and LC-MS/MS systems for explosives analysis [200–204]; this source is chosen because lower LOD can be found with APCI over EI sources. However, EI remains the ionization source used most often in GC-MS analyses of explosives [12, 134–137, 139–143, 205]. DESI, desorption atmospheric pressure chemical ionization (DAPCI), and direct analysis in real time (DART) sources are used to detect explosives, including emerging explosive threats like TATP and as part of ielddeployable MS systems [163, 206–209]. Other less widely used ionization sources have been employed for explosives analysis including singlephoton laser ionization (SPLI) [210–211], cluster SIMS [212], and dielectric barrier discharge ionization [213]. 20.5.7 Miniaturized Mass Analyzers for Explosive Threat Detection The future of explosives analyses by MS, and CBRNE threat detection in general, may be in the miniaturization of MS systems. Reviews on the development of portable instrumentation for the detection of explosives and illicit drugs are available and include a portable GC-MS system [214]. A community of researchers interested in “deploying mass spectrometers outside the typical lab setting” organize workshops on harshenvironment MS [215]. This community provides Internet resources for the most up-to-date advances and developments in MS miniaturization [215]. The past decade has seen signiicant commercial development of miniaturized MS (mini MS) systems [215]. One such system, the Guardion®-7 by Torion Technologies, Inc. (American Fork, UT), couples a small-scale gas chromatograph to a miniaturized toroidal ion trap mass spectrometer (TMS) for HLS threat detection. This product features an instrument footprint about the size of a briefcase, weighs about 30 lb and is battery operated. The entire system, not only the GC and mass analyzer, but also the vacuum, power, and electronic systems, is

466

HOMELAND SECURITY

Analyzer r0 (radial) dimension = 2 mm Einzel lens/gate

Cross section of the toroid trapping field

Filament end cap

Filament ca. 5 cm

Inner ring Outer ring Detector end cap

Continuous dynode electron multiplier

ca. 10 cm

FIGURE 20.18 Diagram of a miniaturized toroidal ion trap mass analyzer. Source: Lammert, S.A., et al. (2006) Miniature toroidal radio frequency ion trap mass analyzer. Journal of the American Society for Mass Spectrometry, 17, 916–922.

made on a smaller scale for greater portability. The miniature toroidal ion trap mass analyzer, as shown in Figure 20.18, consists of an EI source, an einzel lens for ion injection, a toroidal ion trap with a trapping region radius of 2 mm, and a continuous dynode electron multiplier for detection [216]. There are several key reasons why an ion trap is miniaturized over other mass analyzers. The size of an ion trap is reduced because of its straightforward design, toleration of higher operating pressures (10−3 Torr), less rigorous alignment of ion optics, high sensitivity, and MS/MS capability [217]. A toroidal ion trap is used because it provides good ion storage capacity for its size; in general, miniaturizing ion traps reduces their ion storage capacity [217]. Power consumption by the mass analyzer is also reduced by decreasing the trapping volume radii to 2 mm from a typical volume radii of 10 mm [217]. Literature reports on MS miniaturization include a “tiny TOF” system that detects particulate samples of explosives and illicit drugs at security screening checkpoints [218]. This mini MS system features a MALDI ionization source for simultaneous sample desorption and ionization and a 20-cm light length. The system detects the explosive RDX at sub-nanogram levels. An APCI source is available for use in an MS system for explosive trace detection [219]. This source features high ionization eficiency of nitrogen-containing compounds and low maintenance requirements. DART and DESI sources are coupled to a prototype mobile MS system, providing proof-of-concept analyses of explosives [206]. This system features the already mentioned atmospheric pressure ionization sources as well as a

small cylindrical ion trap that requires lower voltage and vacuum requirements than its larger counterparts. The entire system weighs less than 45 kg and has a 0.1-m3 footprint [206]. The future miniaturization of MS will depend on decreasing the size, power, high voltage, and vacuum requirements of the mass analyzers. Further reductions in the size and complexity of electronic control units including data processing and output modules may lead to the realization of mass spectrometers used outside the laboratory. The continued improvement of systems with miniature mass analyzers will help to increase performance, resolving power, and ease of use of these systems.

20.6

CONCLUSION

A wide variety of MS techniques are used to detect CBRNE threats. CWAs are commonly analyzed by LC-MS or LC-MS/MS methods using either an APCI or an ESI source. The further development of MS techniques for CWA detection should involve MS miniaturization and coupling of sample collection and introduction devices that can analyze air and water samples. The past decade has seen the advent of BWA detection by MS. In particular, the detection of ricin has already received considerable method development. Environmental research to detect and identify biological toxins in natural waters may be applied to the analysis of drinking waters for HLS applications. The future of MS detection of BWAs lies in the related ield of biological MS. This rapidly advancing area of MS

LIST OF ABBREVIATIONS

467

method and instrumental development may be easily applied for the analysis and detection of pathogenic bacteria and biological toxins. Further work is needed to characterize pathogenic bacteria using new and established MS proteomic techniques; continued reinement of MALDI sources and improved proteomic databases will help in these efforts. The detection of radiological and nuclear threats by MS will beneit from more rapid sample preparation and better ionization sources. TIMS is replacing SIMS, ICP-MS, and GD-MS for the most sensitive, accurate, and precise isotopic ratio determinations; however, sample preparation is on the order of 15–30 days [93]. Further MS method and instrument development (that will provide lower LOD and better isotopic ratio determinations) will beneit nuclear forensic analyses and counterproliferation efforts. Explosives are commonly analyzed by GC-MS, LC-MS, and now IM-MS methods. Typically, MS is used to conirm the identity of an explosive in forensic appli-

cations. The coupling of IMS to MS has enabled the separation and detection of explosives by mobility and mass, which helps to combat matrix effects from the sample. The analysis of explosives by MS would beneit from more sensitive and selective detection of explosives, further miniaturization of MS systems, and new methods to detect emerging explosives found in IEDs. Overall, the MS analysis of each CBRNE threat type would beneit from further miniaturization of MS systems along with breakthroughs in ease of use and advances in data processing and interpretation. In the future, MS may leave the lab environment and ind use as a rapid, on-site detector for HLS threats.

LIST OF ABBREVIATIONS

FAIMS

AMS APCI

FTIR

Accelerator mass spectrometry Atmospheric pressure chemical ionization API Atmospheric pressure ionization APPI Atmospheric pressure photoionization CBRNE Chemical, biological, radiological, nuclear, and explosive CE-MS Capillary electrophoresis–mass spectrometry CI Chemical ionization CID Collision-induced dissociation CIR Chemical ionization reaction Cluster SIMS Cluster secondary ion mass spectrometry CV Compensation voltage CWA Chemical warfare agent CWC Chemical Weapons Convention DESI Desorption electrospray ionization DMS Differential mobility spectrometry, see also FAIMS DPIS Distributed plasma ionization source DT-IMS Drift tube ion mobility spectrometry DV Dispersion voltage E Electric ield (V/cm) EC Electron capture ionization EESI Extractive electrospray ionization EI Electron impact ionization ELISA Enzyme-linked immunosorbent assay Er Erbium ESI Electrospray ionization EU European Union

ACKNOWLEDGMENT This work was funded in part by grant #0731306 from the U.S. National Science Foundation.

FT-ICR-MS GA GB GC-MS GD GD-MS GF HD HEU HLS HMTD HMX HPLC-UV HPTLC IAEA ICP-MS ICP-OES ICP-SFMS ID ID-MS

Field asymmetric ion mobility spectrometry, see also DMS Fourier transform infrared spectroscopy Fourier transform ion cyclotron resonance mass spectrometry Tabun Sarin Gas chromatography–mass spectrometry Soman Glow discharge mass spectrometry Cyclosarin Mustard gas Highly enriched uranium Homeland security Hexamethylene triperoxide diamine 1,3,5,7-Tetranitro1,3,5,7-tetraazacyclooctane High performance liquid chromatography with an ultraviolet detector High performance thin layer chromatography International Atomic Energy Agency Inductively coupled plasma mass spectrometry Inductively coupled plasma optical emission spectroscopy Inductively coupled plasma sector ield mass spectrometry Identiication Isotope dilution–mass spectrometry

468

HOMELAND SECURITY

INAA IMS IM-MS IT-MS K L LA LC-MS LDI MALDI Mini MS MS/MS-CID N O PETN ppb ppt Pu Py-MS

Instrumental neutron activation analysis Ion mobility spectrometry Ion mobility mass spectrometry Ion trap mass spectrometry Ion mobility constant (cm2/V•s) Length (cm) Laser ablation Liquid chromatography–mass spectrometry Laser desorption ionization Matrix-assisted laser desorption/ ionization Miniaturized mass spectrometry Tandem mass spectrometry–collisioninduced dissociation Number density (cm−3) Oxygen Pentaerythritol tetranitrate Parts per billion Parts per trillion Plutonium Pyrolysis mass spectrometry

quadMS Q-TOF-MS RDD RDX RF SIMS SESI SPLI SPME TATP td TIC TIMS TMS TNB TNT TOF-MS U V VX

REFERENCES 1. U.S. Homeland Security Council. (2007) National strategy for homeland security, http://www.dhs.gov/xlibrary/ assets/nat_strat_homelandsecurity_2007.pdf (accessed October 18, 2011). 2. U.S. Department of Homeland Security. (2008) One team, one mission, securing our homeland: strategic plan, http://www.dhs.gov/xlibrary/assets/DHS_StratPlan_ FINAL_spread.pdf (accessed October 18, 2011) 3. U.S. Ofice of the President. (2001) Establishing the ofice of homeland security and the homeland security council, 51812–51817, U.S. Ofice of the Federal Register, pp. 51812–51817.

9.

10.

11.

4. Archick, K., et al. (2006) European approaches to homeland security and counterrorism. 51, U.S. Congressional Research Service. 5. U.S. Federal Aviation Administration. (2008) The Federal Aviation Administration: A Historical Perspective, 1903– 2008, http://www.faa.gov/about/history/historical_perspective/ (accessed October 18, 2011). 6. U.S. Nuclear Regulatory Commission.(2003) Dirty bombs: fact sheet, United States Nuclear Regulatory Commission: Ofice of Public Affairs. 7. John, H., et al. (2008) LC-MS-based procedures for monitoring of toxic organophosphorus compounds and veriication of pesticide and nerve agent poisoning. Analytical and Bioanalytical Chemistry, 391, 97–116. 8. DelVecchio, V.G., Mujer, C.V. (2008) Bioterrorism. In Mass Spectrometry: Instrumentation, Interpretation, and Applications, edited by Ekman, R., Silberring, J.,

12.

13.

14.

15.

Quadrupole mass spectrometry Quadrupole time-of-light mass spectrometry Radiological dispersion device 1,3,5-Trinitroperhydro-1,3,5-triazine Radio frequency Secondary ion mass spectrometry Secondary electrospray ionization Single-photon laser ionization Solid-phase microextraction Triacetone triperoxide Time, ion drift (s) Toxic industrial chemical Thermal ionization mass spectrometry Toroidal ion trap mass spectrometry Trinitrobenzene 2,4,6-Trinitrotoluene Time-of-light mass spectrometry Uranium Voltage (V) O-ethyl S-(2-diisopropylaminoethyl) methylthiophosphonate

Westman-Brinkmalm, A., and Kraj, A. Hoboken, NJ: John Wiley & Sons, pp. 267–273. Yinon, J., Xiaoming, Z., Gapeev, A. (2004) Tracking the terrorists. Identiication of explosive residues in postexplosion debris by LC/MS methods. NATO Science Series, II: Mathematics, Physics and Chemistry, 167 (Vapour and Trace Detection of Explosives for AntiTerrorism Purposes), 51–62. Packer, A.P., et al. (2007) Validation of an inductively coupled plasma mass spectrometry (ICP-MS) method for the determination of cerium, strontium, and titanium in ceramic materials used in radiological dispersal devices (RDDs). Analytica Chimica Acta, 588(2), 166–172. D’Agostino, P.A., et al. (1990) Capillary column gas chromatography-mass spectrometry and gas chromatography-tandem mass spectrometry detection of chemical warfare agents in a complex airborne matrix. Journal of Chromatography A, 504, 259–268. Burns, D., et al. (1995) Analysis and characterization of nitroglycerin-based explosives by gas chromatographymass spectrometry. Analytica Chimica Acta, 307, 89–95. Yinon, J., Hwang, D.G. (1983) High-performance liquid chromatography-mass spectrometry of explosives. Journal of Chromatography A, 268, 45–53. Dworzanski, J.P., et al. (2004) Identiication of bacteria using tandem mass spectrometry combined with a proteome database and statistical scoring. Analytical Chemistry, 76, 2355–2366. Scarlatos, A., et al. (2005) Methods for detecting botulinum toxin with applicability to screening foods against biological terrorist attacks. Journal of Food Science, 70, R121–R130.

REFERENCES

16. Wallenius, M., Mayer, K. (2000) Age determination of plutonium material in nuclear forensics by thermalionization mass spectrometry. Fresenius’ Journal of Analytical Chemistry, 366, 234–238. 17. Smiths Detection. (2010) Rapid trace detection of explosives: ionscan 500DT, http://www.smithsdetection.com/ IONSCAN_500DT.php (accessed October 18, 2011). 18. Eiceman, G., Karpas, Z. (1994) Ion Mobility Spectrometry. Boca Raton, FL: CRC Press. 19. Winder, C. (2001) The toxicology of chlorine. Environmental Research Section A, 85, 105–114. 20. D’Agostino, P.A. (2005) Recent advances and applications of LC-MS for the analysis of chemical warfare agents and their degradation products––a review. Trends in Chromatography, 1, 23–42. 21. Haigh, T.A. (2007) Chemical warfare agents and their hydrolysis products from the US EPA standardized analytical methods and GC-MS analytical method for the analysis of chemical warfare agent degradation products listed in the EPA standardized analytical methods. In NEMC Conference. Cambridge, MA. 22. John, H., et al. (2009) Toxicokinetics of chemical warfare agents: nerve agents and vesicants. In Handbook of Toxicology of Chemical Warfare Agents, edited by Gupta, R.C. New York: Elsevier. 23. Organization for the Prohibition of Chemical Weapons. (1997) Annexure IV: Schedules of Chemicals. 24. Organization for the Prohibition of Chemical Weapons. (2009) Status of Participation in the Chemical Weapons Convention as of 21 May 2009. 25. Broad, W.J. (1998) Sowing death: a special report; how japan germ terror alerted world. The New York Times. 26. Black, R.M. (2010) History and perspectives of bioanalytical methods for chemical warfare agent detection. Journal of Chromatography B, 878, 1207–1215. 27. Smith, J.R., Shih, M.L. (2001) Analysis of the degradation compounds of chemical warfare agents using liquid chromatography/mass spectrometry. Journal of Applied Toxicology, 21, S27–S34. 28. D’Agostino, P.A., et al. (2003) Mass spectrometric analysis of chemical warfare agents and their degradation products in soil and synthetic samples. European Journal of Mass Spectrometry, 9, 609–618. 29. Noort, D., et al. (2002) Biomonitoring of exposure to chemical warfare agents: a review. Toxicology and Applied Pharmacology, 184, 116–126. 30. Carol-Visser, J., et al. (2008) Development of an automated on line pepsin digestion–liquid chromatography– tandem mass spectrometry coniguration for the rapid analysis of protein adducts of chemical warfare agents. Journal of Chromatography B, 870, 91–97. 31. Wensing, M.W., et al. (2000) Determination of chemical warfare degradation products and toxins by mass spectrometric techniques. In Abstracts of Papers, 220th ACS National Meeting.

469

32. Mawhinney, D.B., et al. (2007) Enhancing the response of alkyl methylphosphonic acids in negative electrospray ionization liquid chromatography tandem mass spectrometry by post-column addition of organic solvents. Journal of the American Society for Mass Spectrometry, 18, 1821–1826. 33. Swaim, L.L., et al. (2008) Quantiication of organosphosphorus nerve agent metabolites using a reduced-volume, high-throughput sample processing format and liquid chromatography-tandem mass spectrometry. Journal of Analytical Toxicology, 32, 774–777. 34. D’Agostino, P.A., et al. (2006) Liquid chromatography electrospray tandem mass spectrometric and desorption electrospray ionization tandem mass spectrometric analysis of chemical warfare agents in ofice media typically collected during a forensic investigation. Journal of Chromatography A, 1110, 86–94. 35. Owens, J., Koester, C. (2009) Quantitative analysis of chemical warfare agent degradation products in beverages by liquid chromatography tandem mass spectrometry. Journal of Agricultural and Food Chemistry, 57, 8227–8235. 36. Kroening, K.K., et al. (2009) Screening hydrolysis products of sulfur mustard agents by high-performance liquid chromatography with inductively coupled plasma mass spectrometry detection. Analytical and Bioanalytical Chemistry, 393, 1949–1956. 37. Kubachka, K.M., et al. (2008) Detection of chemical warfare agent degradation products in foods using liquid chromatography coupled to inductively coupled plasma mass spectrometry and electrospray ionization mass spectrometry. Journal of Chromatography A, 1202, 124–131. 38. Wilbur, S. (2004) Applications of ICP-MS in homeland security, Agilent Technologies, http://www.chem.agilent. com/Library/applications/5989-0741EN.pdf (accessed October 18, 2011). 39. Kanamori-Kataoka, M., Seto, Y. (2008) Laboratory identiication of the nerve gas hydrolysis products alkyl methylphosphonic acids and methylphosphonic acid, by gas chromatography-mass spectrometry after tertbutyldimethylsilylation. Journal of Health Science (Tokyo, Japan), 54(5), 513–523. 40. Pardasani, D., et al. (2005) Gas chromatography/mass spectrometric analysis of methyl esters of N,Ndialkylaminoethane-2 sulfonic acids for veriication of the Chemical Weapons Convention. Rapid Communications in Mass Spectrometry, 19, 3015–3020. 41. Dubey, D.K., et al. (2008) Microsynthesis and gas chromatography/electron ionization mass spectrometric and tandem mass spectrometric analysis of cyclic alkylphosphonates for veriication of the Chemical Weapons Convention. Rapid Communications in Mass Spectrometry, 22, 2526–2532. 42. Borrett, V.T., et al. (1996) Veriication of the united nations chemical weapons convention: the application of electrospray mass spectrometry. Rapid Communications in Mass Spectrometry, 10, 114–118.

470

HOMELAND SECURITY

43. D’Agostino, P.A., et al. (2001) Electrospray mass spectrometry of chemical warfare agents. Advances in Mass Spectrometry, 15, 297–316.

59.

44. Karasek, F.W., Denney, D.W. (1974) Detection of 2,4,6-trinitrotoluene vapours in air by plasma chromatography. Journal of Chromatography, 93, 141–147. 45. Ewing, R.G., et al. (2001) A critical review of ion mobility spectrometry for the detection of explosives and explosive related compounds. Talanta, 54, 515–529. 46. Mason, E.A., Schamp, H.W. (1958) Mobility of gaseous ions in weak electric ields. Annales de Physique, 4, 233–270. 47. Mason, E.A., McDaniel, E.W. (1988) Transport Properties of Ions in Gases. New York: Wiley and Sons. 48. Eiceman, G.A., Stone, J.A. (2004) Ion mobility spectrometers in national defense. Analytical Chemistry, 76(21) 390A–397A.

60.

61.

62.

49. McHugh, V. (2010) Commercial CWA detection systems, personal communication

63.

50. Bartlet-Hunt, S.L., et al. (2008) A review of chemical warfare agent simulants for the study of environmental behavior. Critical Reviews in Environmental Science and Technology 38, 112–136.

64.

51. Bocos-Bintintan, V., et al. (2002) Characterization of the chemical warfare agent simulant methyl salicylate by ion mobility spectrometry/mass spectrometry (IMS/MS) at ambient temperature. Chemia, 47, 149–160.

65.

52. Steiner, W.E., et al. (2003) Secondary ionization of chemical warfare agent simulants: atmospheric pressure ion mobility time-of-light mass spectrometry. Analytical Chemistry, 75, 6068–6076.

66. 67.

53. Steiner, W.E., et al. (2006) Detection of aqueous phase chemical warfare agent degradation products by negative mode ion mobility time-of-light mass spectrometry [IM(tof)MS]. Journal of the American Society for Mass Spectrometry, 17, 241–245. 54. Steiner, W.E., et al. (2005) Detection of a chemical warfare agent simulant in various aerosol matrixes by ion mobility time-of-light mass spectrometry. Analytical Chemistry, 77, 4792–4799. 55. Kolakowski, B.M., et al. (2007) Analysis of chemical warfare agents in food products by atmospheric pressure ionization-high ield asymmetric waveform ion mobility spectrometry-mass spectrometry. Analytical Chemistry, 79, 8257–8265.

68.

69.

70. 71.

56. Lagarrigue, M., et al. (2006) Separation and identiication of isomeric acidic degradation products of organophosphorus chemical warfare agents by capillary electrophoresis-ion trap mass spectrometry. Journal of Chromatography A, 1137, 110–118.

72.

57. Lagarrigue, M., et al. (2007) CE-MS-MS for the identiication of chemical warfare agent degradation products. LCGC North America, 25, 292–301.

73.

58. Lagarrigue, M., et al. (2009) Analysis of nerve agent degradation products in high-conductivity matrices by tran-

74.

sient ITP preconcentration and CZE separation coupled to ESI-MS. Electrophoresis, 30, 1522–1530. Martin, A.N., et al. (2007) Single-particle aerosol mass spectrometry for the detection and identiication of chemical warfare agent simulants. Analytical Chemistry, 79, 6368–6375. Palit, M., Mallard, G. (2009) fragmentation energy index for universalization of fragmentation energy in ion trap mass spectrometers for the analysis of chemical weapon convention related chemicals by atmospheric pressure ionization-tandem mass spectrometry analysis. Analytical Chemistry, 81, 2477–2485. Syage, J.A., et al. (2006) Direct sampling of chemical weapons in water by photoionization mass spectrometry. Analytical Chemistry, 78, 2967–2976. Cotte-Rodriguez, I., et al. (2006) Analysis of gaseous toxic industrial compounds and chemical warfare agent simulants by atmospheric pressure ionization mass spectrometry. Analyst, 131, 579–589. Cordell, R.L., et al. (2007) Detection of chemical weapon agents and simulants using chemical ionization reaction time-of-light mass spectrometry. Analytical Chemistry, 79, 8359–836. Sekiguchi, H., et al. (2006) On-site determination of nerve and mustard gases using a ield-portable gas chromatograph-mass spectrometer. Forensic Toxicology, 24, 17–22. CDC. (2010) Bioterrorism agents and diseases––by category, http://www.bt.cdc.gov/agent/agentlist-category.asp (accessed October 18, 2011). Corporate. (2008) Case not closed. Nature, 454, 917. Stephenson, J. (2003) EPA’s cleanup was successful; opportunities exist to enhance contract oversight (U.S. Government Accounting Ofice, ed.). Maizels, M., Buddle, W.L. (2004) A LC/ MS method for the determination of cyanobacteria toxins in water. Analytical Chemistry, 76(5), 1342–1351. Krock, B., et al. (2009) Pectenotoxins and yessotoxin from Arica Bay, North Chile as determined by tandem mass spectrometry. Toxicon, 54(3), 364–367. Messineo, V., et al. (2009) Cyanobacterial toxins in Italian freshwaters. Limnologica, 39, 95–106. Brix, R., et al. (2009) Identiication of disinfection byproducts of selected triazines in drinking water by LCQ-ToF-MS/MS and evaluation of their toxicity. Journal of Mass Spectrometry, 44, 330–337. Farre, M., et al. (2001) Determination of drugs in surface water and wastewater samples by liquid chromatographymass spectrometry: methods and preliminary results including toxicity studies with Vibrio ischeri. Journal of Chromatography A, 938, 187–197. Pietsch, J., et al. (2001) Analytical determination of algae toxins (hepato- and neurotoxins) by HPLC/MS/MS. Biologische Abwasserreinigung, 16, 35–50. Pietsch, J., et al. (2001) Simultaneous determination of cyanobacterial hepato- and neurotoxins in water samples

REFERENCES

75.

76. 77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

by ion-pair supported enrichment and HPLC-ESI-MSMS. Chromatographia, 54, 339–344. Smith, G.A., et al. (2003) Liquid chromatography mass spectrometry (LC-MS) based method for the analysis of algal toxins. In Proceedings—Water Quality Technology Conference, 554–558 Rutenber, E., et al. (1991) Crystallographic reinement of ricin to 2.5 A. Proteins, 10, 240–250. Becher, F., et al. (2007) Detection of functional ricin by immunoafinity and liquid chromatography-tandem mass spectrometry. Analytical Chemistry, 79, 659–665. Kalb, S.R., Barr, J.R. (2009) Mass spectrometric detection of ricin and its activity in food and clinical samples. Analytical Chemistry, 81, 2037–2042. Oestin, A., et al. (2007) Solvent-assisted trypsin digestion of ricin for forensic identiication by LC-ESI MS/MS. Analytical Chemistry, 79, 6271–6278. Fredriksson, S.-A., et al. (2005) Forensic identiication of neat ricin and of ricin from crude castor bean extracts by mass spectrometry. Analytical Chemistry, 77, 1545– 1555. Mouser, P., et al. (2007) Fatal ricin toxicosis in a puppy conirmed by liquid chromatography/mass spectrometry when using ricinine as a marker. Journal of Veterinary Diagnostic Investigation, 19, 216–220. Duriez, E., et al. (2008) Detection of ricin in complex samples by immunocapture and matrix-assisted laser desorption/ionization time-of-light mass spectrometry. Journal of Proteome Research, 7, 4154–4163. Seto, Y., Kanamori-Kataoka, M. (2005) Mass spectrometric strategy for the determination of natural and synthetic organic toxins. Journal of Health Science, 51, 519–525. Yan, Y., et al. (2006) Application of peptide mass ingerprint in ricin identiication with matrix-assisted laser desorption ionization time-of-light-mass spectrometry. Hua Xue Xue Bao, 34, 187–190. Jiang, J., et al. (2007) An immunoafinity tandem mass spectrometry (iMALDI) assay for detection of Francisella tularensis. Analytica Chimica Acta, 605, 70–79. Krishnamurthy, T., et al. (1996) Detection of pathogenic and non-pathogenic bacteria by matrix-assisted laser desorption/ionization time-of-light mass spectrometry. Rapid Communications in Mass Spectrometry, 10, 883–888. Basile, F., et al. (1998) Pathogenic bacteria: their detection and differentiation by rapid lipid proiling with pyrolysis mass spectrometry. TRAC: Trends in Analytical Chemistry, 17, 95–109. Wilkes, J.G., et al. (2006) Pyrolysis mass spectrometry for distinguishing potential hoax materials from bioterror agents. Rapid Communications in Mass Spectrometry, 20, 2383–2386. Potter, W.C. (1995) Before the deluge? Assessing the threat of nuclear leakage from the post-Soviet states. Arms Control Today, 25, 9–16.

471

90. Nowak, D. (2010) Georgia conirms highly enriched uranium seizure. The New York Times. 91. Deutch, J. (1996) The threat of nuclear diversion. U.S. Special Council on Governmental Affairs, Ed. 92. Wallenius, M., et al. (2007) Nuclear forensic investigations with a focus on plutonium. Journal of Alloys and Compounds, 444–445, 57–62. 93. Mayer, K., Wallenius, M. (2009) Atomic detectives––an introduction to nuclear forensics. Strahlenschutz Praxis, 15, 28–36. 94. Smith, D.K., et al. (2008) Documentation of a model action plan to deter illicit nuclear traficking. Journal of Radioanalytical and Nuclear Chemistry, 276, 415–419. 95. Becker, J.S. (2005) Recent developments in isotope analysis by advanced mass spectrometric techniques. Journal of Analytical Atomic Spectrometry, 20, 1173–1184. 96. Mayer, K., et al. (2005) Nuclear forensics––a methodology providing clues on the origin of illicitly traficked nuclear materials. Analyst, 130, 433–441. 97. Mayer, K., et al. (2009) Unveiling the history of seized plutonium through nuclear forensic investigations. Radiochimica Acta, 97, 261–264. 98. Schwantes, J.M., et al. (2009) Nuclear archeology in a bottle: evidence of pre-trinity U.S. weapons activities from a waste burial site. Analytical Chemistry, 81, 1297–1306. 99. Entwistle, J.A., et al. (2003) Identiication and characterization of radioactive “hot” particles in Chernobyl fallout-contaminated soils: the application of two novel approaches. Mineralogical Magazine, 67, 183–204. 100. Ammann, A.A. (2007) Inductively coupled plasma mass spectrometry (ICP MS): a versatile tool. Journal of Mass Spectrometry, 42, 419–427. 101. Burger, S., Riciputi, L.R. (2009) A rapid isotope ratio analysis protocol for nuclear solid materials using nanosecond laser-ablation time-of-light ICP-MS. Journal of Environmental Radioactivity, 100, 970–976. 102. Stefanka, Z., et al. (2008) Laser ablation assisted ICP-MS as a tool for rapid categorization of seized uranium oxide materials based on isotopic composition determination. Journal of Analytical Atomic Spectrometry, 23, 1030–1033. 103. Lloyd, N.S., et al. (2009) Precise and accurate isotopic analysis of microscopic uranium-oxide grains using LAMC-ICP-MS. Journal of Analytical Atomic Spectrometry, 24, 752–758. 104. Tuniz, C., et al. (2003) Sherlock Holmes counts the atoms. Nuclear Instruments and Methods in Physics Research. Section B, Beam Interactions with Materials and Atoms, 213, 469–475. 105. Whitney, S.M., et al. (2007) Analyzing nuclear fuel cycles from isotopic ratios of waste products applicable to measurement by accelerator mass spectrometry. Nuclear Science and Engineering, 157, 200–209. 106. Brown, T.A., et al. (2004) Developments in the measurement of actinides and 129I at LLNL by accelerator mass spectrometry. In Annual Meeting Proceedings of the

472

107.

108.

109.

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

HOMELAND SECURITY

Institute of Nuclear Materials Management 398/391-398/398. Brown, T.A., et al. (2002) Measurement of actinides in environmental samples at micro-becquerel levels by accelerator mass spectrometry.In Annual Meeting Proceedings of the Institute of Nuclear Materials Management, 121–128. Vogel, J.S., Love, A.H. (2005) Quantitating isotopic molecular labels with accelerator mass spectrometry. Methods in Enzymology, 402, 402–422. Tamborini, G. (2004) SIMS analysis of uranium and actinides in microparticles of different origin. Microchimica Acta, 145, 237–242. Betti, M., et al. (1999) Use of secondary ion mass spectrometry in nuclear forensic analysis for the characterization of plutonium and highly enriched uranium particles. Analytical Chemistry, 71, 2616–2622. Pajo, L., et al. (2001) Investigation of the oxygen isotopic composition in oxidic uranium compounds as a new property in nuclear forensic science. Fresenius’ Journal of Analytical Chemistry, 371, 348–352. Tamborini, G., et al. (2002) Oxygen isotopic measurements by secondary ion mass spectrometry in uranium oxide microparticles: a nuclear forensic diagnostic. Analytical Chemistry, 74, 6098–6101. Tamborini, G., et al. (2000) SIMS characterization of inhouse monodispersed isotopic standard uranium oxide particles. 995–998. Wallenius, M., Mayer, K. (2000) Age determination of plutonium material in nuclear forensics by thermalionization mass spectrometry. Fresenius’ Journal of Analytical Chemistry, 366(3), 234–238. Pajo, L., et al. (2001) A novel isotope analysis of oxygen in uranium oxides: comparison of secondary ion mass spectrometry, glow discharge mass spectrometry and thermal ionization mass spectrometry. Spectrochimica Acta, 56B, 541–549. Buerger, S., et al. (2007) A high eficiency cavity ion source using thermal ionization mass spectrometry (TIMS) for nuclear forensic analysis. Journal of Alloys and Compounds, 444–445, 660–662. Chartier, F., et al. (1999) Determination of erbium in nuclear fuels by isotope dilution thermal ionization mass spectrometry and glow discharge mass spectrometry. Journal of Analytical Atomic Spectrometry, 14, 1461– 1465. Betti, M. (2005) Isotope ratio measurements by secondary ion mass spectrometry (SIMS) and glow discharge mass spectrometry (GDMS). International Journal of Mass Spectrometry, 242(2–3), 169–182. Bacon, J.R., et al. (2006) Atomic spectrometry update. Atomic mass spectrometry. Journal of Analytical Atomic Spectrometry, 21, 785–818. Wallenius, M., et al. (2006) Nuclear forensic investigations: two case studies. Forensic Science International, 156, 55–62.

121. Colella, M., Reinhard, M., Tuniz, C., Thomson, S. (2007) Unmasking the illicit traficking of nuclear and other radioactive materials. In Radionuclide Concentrations in Food and the Environment, edited by Leo, M. Nollet, L., Poschl, M. Boca Raton, FL: CRC Press, pp. 333–365. 122. Hou, X., et al. (2005) Analytical atomic spectrometry for nuclear forensics. Applied Spectroscopy Reviews, 40, 245–267. 123. Steinfeld, J.I., Wormhoudt, J. (1998) Explosives detection: a challenge for physical chemistry. Annual Reviews in Physical Chemistry, 49, 203–232. 124. Moore, D.S. (2004) Instrumentation for trace detection of high explosives. The Review of Scientiic Instruments, 75, 2499–2512. 125. Yoder, C. (2002) Commerce in explosives; list of explosive materials. Federal Register, 67, 20864–20866. 126. Disley, J., Dowdney, H. (2005) 7/7: War on Britain: explosive link to Al Qaeda. The Mirror, London, England. 127. Jordan, L.J. (2007) FBI Work to thwart kitchen sink bombs. The Associated Press. 128. Steinfeld, J.I., Wormhoudt, J. (1998) Explosives detection: a challenge for physical chemistry. Annual Reviews in Physical Chemistry, 49, 203–232. 129. Sigman, M.E., Ma, C.-Y. (2001) Detection limits for GC/ MS analysis of organic explosives. Journal of Forensic Sciences, 46, 6–11. 130. Harper, R.J., et al. (2005) Identiication of dominant odor chemicals emanating from explosives for use in developing optimal training aid combinations and mimics for canine detection. Talanta, 67(2), 313–327. 131. Jezova, V., et al. (2004) Determination of low concentration trinitrotoluene content in water samples. In New Trends in Research of Energetic Materials, 158–164. 132. Halasz, A., et al. (2002) Detection of explosives and their degradation products in soil environments. Journal of Chromatography, 963, 411–418. 133. Stambouli, A., et al. (2004) Headspace-GC/MS detection of TATP traces in post-explosion debris. Forensic Science International, 146, S191–S194. 134. Baez, B., et al. (2004) Transport of explosives I. TNT in soil and its equilibrium vapor. In Proceedings of SPIE–– The International Society for Optical Engineering, 5415 (Pt. 2, Detection and Remediation Technologies for Mines and Minelike Targets IX), pp. 1389–1399. 135. Baez, B., et al. (2005) Transport of explosives II: use of headspace-SPME/GC-ECD and TEEM GC/MS for detection of TNT vapors from sand buried samples. In Proceedings of SPIE––The International Society for Optical Engineering, 5794, 1263–1271. 136. Correa, S.N., et al. (2005) Immersion mode SPME/ECD/ GC and TEEM-GC/MS for analysis of explosives buried in sand. In Proceedings of SPIE––The International Society for Optical Engineering, 5794, 1272–1280. 137. Xu, T., et al. (2004) Characterization of migrated substances from JOB-9003 explosive. Huozhayao Xuebao, 27, 80–83.

REFERENCES

138. U.S. EPA. (2000) Method 8095: explosives by gas chromatography. 139. Bradley, K.S. (2005) Determination of elemental sulfur in explosives and explosive residues by gas chromatography-mass spectrometry. Journal of Forensic Sciences, 50, 96–103. 140. Sharma, S.P., Lahiri, S.C. (2005) Characterization and identiication of explosives and explosive residues using GC-MS, an FTIR microscope, and HPTLC. Journal of Energetic Materials, 23, 239–264. 141. Sharma, S.P., Lahiri, S.C. (2005) GC-MS and HPTLC analysis of constituents of an unexploded bomb. Journal of the Indian Chemical Society, 82, 131–133. 142. Sharma, S.P., Lahiri, S.C. (2008) Preparation of data base for the determination and estimation of high explosives and a comparative study of different methods of estimation of RDX. Journal of the Indian Chemical Society, 85, 273–279. 143. Zuo, Y., et al.(2008) Interaction of nitrate ester explosive and polyurethane binder. Huaxue Yanjiu Yu Yingyong, 20, 691–695. 144. Pena, A.J., et al. (2005) Characterization and differentiation of high energy cyclic organic peroxides by GC/ FT-IR, GC-MS, FT-IR, and Raman microscopy. In Proceedings of SPIE––The International Society for Optical Engineering, 5778 (Pt. 1, Sensors, and Command, Control, Communications, and Intelligence (C3I) Technologies for Homeland Defense IV), pp. 347–358. 145. Pena-Quevedo, A.J., et al.(2006) Synthesis, characterization, and differentiation of high energy amine peroxides by MS and vibrational microscopy. In Proceedings of SPIE––The International Society for Optical Engineering, 6201 (Sensors, and Command, Control, Communications, and Intelligence (C31) Technologies for Homeland Security and Homeland Defense V): 62012E/162012E/10. 146. Muller, D., et al. (2004) Improved method for the detection of TATP after explosion. Journal of Forensic Sciences, 49, 935–938. 147. Sigman, M.E., et al. (2006) Analysis of triacetone triperoxide by gas chromatography/mass spectrometry and gas chromatography/tandem mass spectrometry by electron and chemical ionization. Rapid Communications in Mass Spectrometry, 20, 2851–2857.

473

151. Wood, M., et al. (2006) Recent applications of liquid chromatography-mass spectrometry in forensic science. Journal of Chromatography A, 1130, 3–15. 152. Yinon, J. (2006) LC-MS in forensic chemistry. In Applications of LC-MS in Toxicology, edited by Polettini A. London: Pharmaceutical Press, pp. 257–269. 153. Zhao, X., Yinon, J. (2002) Identiication of nitrate ester explosives by liquid chromatography–electrospray ionization and atmospheric pressure chemical ionization mass spectrometry. Journal of Chromatography A, 977, 59–68. 154. Xu, X., et al. (2004) Trace analysis of peroxide explosives by high performance liquid chromatography-atmospheric pressure chemical ionization-tandem mass spectrometry (HPLC-APCI-MS/MS) for forensic applications. Journal of Forensic Sciences, 49, 1230–1236. 155. Crowson, A., Beardah, M.S. (2001) Development of an LC/MS method for the trace analysis of hexamethylenetriperoxidediamine (HMTD). Analyst, 126, 1689–1693. 156. Wu, Z., et al. (2002) Composition of explosives by electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. Analytical Chemistry, 74, 1879–1883. 157. Mathis, J.A., McCord, B.R. (2005) The analysis of high explosives by liquid chromatography/electrospray ionization mass spectrometry: multiplexed detection of negative ion adducts. Commmunications in Mass Spectrometry, 19, 99–104. 158. Zhang, M. (2006) Analysis of mixture of RDX and HMX by HPLC-ESI MS. Huozhayao Xuebao, 29, 77–80. 159. Pan, X., et al. (2006) Liquid chromatography/electrospray ionization tandem mass spectrometry analysis of octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine (HMX). Rapid Communications in Mass Spectrometry, 20, 2222–2226. 160. Schmidt, A.-C., et al. (2006) Investigation of the ionization and fragmentation behaviour of different nitroaromatic compounds occurring as polar metabolites of explosives using electrospray ionization tandem mass spectrometry. Rapid Communications in Mass Spectrometry, 20, 2293–2302. 161. Schmidt, A.-C., et al. (2006) Identiication and quantiication of polar nitroaromatic compounds in explosivecontaminated waters by means of HPLC-ESI-MS-MS and HPLC-UV. Chromatographia, 63, 1–11.

148. Yamaguchi, S., et al. (2009) Gas chromatography/time-oflight mass spectrometry of triacetone triperoxide based on femtosecond laser ionization. Rapid Communications in Mass Spectrometry, 23, 3101–3106.

162. Gapeev, A., Yinon, J. (2004) Application of spectral libraries for characterization of oxidizers in post-blast residues by electrospray mass spectrometry. Journal of Forensic Sciences, 49, 227–237.

149. Kende, A., et al. (2008) Trace level triacetone-triperoxide identiication with SPME-GC- MS in model systems. Microchimica Acta, 163, 335–338.

163. Cotte-Rodriguez, I., et al. (2005) Desorption electrospray ionization of explosives on surfaces: sensitivity and selectivity enhancement by reactive desorption electrospray ionization. Analytical Chemistry, 77, 6755–6764.

150. Yinon, J. (2004) Analysis of explosives by LC/MS. In Advances in Forensic Applications of Mass Spectrometry, edited by Yinon, Y. Boca Raton, FL: CRC, Press, pp. 231–274.

164. Justes, D.R., et al. (2007) Detection of explosives on skin using ambient ionization mass spectrometry. Chemical Communications (Cambridge, England), 21, 2142–2144.

474

HOMELAND SECURITY

165. Song, L., Bartmess, J.E. (2009) Liquid chromatography/ negative ion atmospheric pressure photoionization mass spectrometry: a highly sensitive method for the analysis of organic explosives. Rapid Communications in Mass Spectrometry, 23, 77–84. 166. Crescenzi, C., et al. (2007) On-line strategies for determining trace levels of nitroaromatic explosives and related compounds in water. Journal of Chromatography A, 1153, 186–193. 167. Cohen, M.J., et al. (1984) The ion mobility spectrometer for high explosive vapor detection. Nuclear Materials Management, 13, 220–225. 168. The Guardian. (2010) Lockerbie plane bombing. The Guardian, http://www.guardian.co.uk/uk/lockerbie (accessed October 21, 2011). 169. Parmeter, J.E., et al. (1998) Explosives detection portal for high-volume personnel screening. In Proceedings of SPIE 3575, pp. 384–391. 170. Wu, C., et al. (2002) Construction and characterization of a high-low, high-resolution ion mobility spectrometer for detection of explosives after personnel portal sampling. Talanta, 57, 123–134. 171. Daum, K.A., et al. (2001) The role of oxygen in the formation of TNT product ions in ion mobility spectrometry. International Journal of Mass Spectrometry, 214, 257– 267. 172. Lawrence, A.H., Neudorl, P. (1988) Detection of ethylene glycol dinitrate vapors by ion mobility spectrometry using chloride reagent ions. Analytical Chemistry, 60, 104–109. 173. Lawrence, A.H., et al. (2001) The formation of chloride adducts in the detection of dinitrocompounds by ion mobility spectrometry. International Journal of Mass Spectrometry, 209, 185–195. 174. Spangler, G., et al. (1985) Recent advances in ion mobility spectrometry for explosives vapor detection. Journal of Testing Evaluation, 13, 234–240. 175. Daum, K.A., et al. (2001) Resolving interferences in negative mode ion mobility spectrometry using selective reactant ion chemistry. Talanta, 54, 299–306. 176. Daum, K.A., et al. (2001) Formation of halide reactant ions and effects of excess reagent chemical on the ionization of TNT in ion mobility spectrometry. Talanta, 55, 491–500. 177. Ewing, R.G., Miller, C.J. (2001) Detection of volatile vapors emitted from explosives with a handheld ion mobility spectrometer. Field Analytical Chemistry And Technology, 5, 215–221. 178. Ross, S.K., Bell, A.J. (2004) Corona ionization source. United States patent 6,822,226. 179. Ewing, R.G., Waltman, M.J. (2009) Mechanisms for negative reactant ion formation in an atmospheric pressure corona discharge. International Journal of Ion Mobility Spectrometry, 12, 65–72. 180. Khayamian, T., et al. (2003) Analysis of 2,4,6trinitrotoluene, pentaerythritol tetranitrate and cyclo-

181.

182.

183.

184.

185.

186.

187. 188. 189.

190.

191.

192.

193.

194.

195.

196.

1,3,5-trimethylene-2,4,6-trinitramine using negative corona discharge ion mobility spectrometry. Talanta, 59, 327–333. Tabrizchi, M., Abedi, A. (2002) A novel electron source for negative ion mobility spectrometry. International Journal of Mass Spectrometry, 218, 75–85. Asbury, G.R., et al. (2000) Analysis of explosives using electrospray ionization/ion mobility spectrometry (ESI/ IMS). Talanta, 50, 1291–1298. Krueger, C.A., et al. (2009) High resolution electrospray ionization ion mobility spectrometry. International Journal of Ion Mobility Spectrometry, 12, 33–37. Tam, M., Hill, H. (2004) Secondary electrospray ionization-ion mobility spectrometry for explosive vapor detection. Analytical Chemistry, 76, 2741–2747. Waltman, M.J., et al. (2008) Characterization of a distributed plasma ionization source (DPIS) for ion mobility spectrometry and mass spectrometry. Talanta, 77, 249–255. Crawford, C.L., et al. (2010) Analysis of black powder by ion mobility-time-of-light mass spectrometry. Analytical Chemistry, 82, 387–393. Mukhopadhyay, R. (2008) IMS/MS: its time has come. Analytical Chemistry, 80, 7919–7920. Kanu, A.B., et al. (2008) Ion mobility-mass spectrometry. Journal of Mass Spectrometry, 43, 1–22. Spangler, G. (1978) Principles underlying performance of ionization detectors for explosive vapor detection. In Proceedings of the New Concept Symposium and Workshop on Detection and Identiication of Explosives, p. 103. Huang, S.D., et al. (1987) Detection of explosives using laser desorption in ion mobility spectrometry/mass spectrometry. Applied Spectroscopy, 41, 1371–1376. Krylov, E.V. (2000) Comparison of the planar and coaxial ield asymmetrical waveform ion mobility spectrometer (FAIMS). International Journal of Mass Spectrometry, 225, 39–51. Shvartsburg, A.A. (2009) Differential Ion Mobility Spectrometry: Nonlinear Ion Transport and Fundamentals of FAIMS. Boca Raton, FL: CRC Press. Guevremont, R. (2004) High-ield asymmetric waveform ion mobility spectrometry: a new tool for mass spectrometry. Journal of Chromatography A, 1058, 3–19. Guevremont, R., Purves, R.W. (1999) Atmospheric pressure ion focusing in a high-ield asymmetric waveform ion mobility spectrometer. The Review of Scientiic Instruments, 70, 1370–1383. Viehland, L.A., et al. (2000) Comparison of high-ield ion mobility obtained from drift tubes and a FAIMS apparatus. International Journal of Mass Spectrometry, 397, 123–130. Wu, C.-H.A., Yost, R.A. (2008) Separation and characterization of explosive compounds using high-ield asymmetric waveform ion mobility spectrometry/mass spectrometry. In Abstracts of Papers, 236th ACS National Meeting, Philadelphia, PA.

REFERENCES

197. Anderson, A.G., et al. (2008) DMS-IMS2, GC-DMS, DMS-MS: DMS hybrid devices combining orthogonal principles of separation for challenging applications. In Proceedings of SPIE, 6954. 198. Giam, C.S., et al. (1994) Pseudo-molecular ion formation of explosives in FT-ICR-MS. In Proceedings of SPIE–– The International Society for Optical Engineering, 2092 (Substance Detection Systems), pp. 227–237. 199. Groom, C.A.H. (2005) Detection of nitroaromatic and cyclic nitramine compounds by cyclodextrin assisted capillary electrophoresis quadrupole ion trap mass spectrometry. Journal of Chromatography A, 1072, 73–82. 200. Collin, O.L., et al. (2009) Fast gas chromatography negative chemical ionization tandem mass spectrometry of explosive compounds using dynamic collision-induced dissociation. International Journal of Mass Spectrometry, 279, 93–99. 201. Calderara, S., et al. (2004) Organic explosives analysis using on column-ion trap EI/NICI GC-MS with an external source. Journal of Forensic Sciences, 49, 1005–1008. 202. Holmgren, E., et al. (2005) Determination and characterization of organic explosives using porous graphitic carbon and liquid chromatography-atmospheric pressure chemical ionization mass spectrometry. Journal of Chromatography A, 1099, 127–135. 203. Tachon, R., et al. (2007) Use of porous graphitic carbon for the analysis of nitrate ester, nitramine and nitroaromatic explosives and by-products by liquid chromatography-atmospheric pressure chemical ionization-mass spectrometry. Journal of Chromatography A, 1154, 174–181. 204. Perr, J.M., et al. (2005) Gas chromatography positive chemical ionization and tandem mass spectrometry for the analysis of organic high explosives. Talanta, 67, 430–436. 205. Kirchner, M., et al. (2007) Fast GC and GC-MS analysis of explosives. Petroleum and Coal, 49, 72–79. 206. Wells, J.M., et al. (2008) Implementation of DART and DESI ionization on a ieldable mass spectrometer. Journal of the American Society for Mass Spectrometry, 19, 1419–1424. 207. Pena-Quevedo, A.J., et al. (2007) Characterization and differentiation of high energy amine peroxides by direct analysis in real time TOF/MS. In Proceedings of SPIE–– The International Society for Optical Engineering, 6538 (Sensors, and Command, Control, Communications, and Intelligence (C3I) Technologies for Homeland Security Defense VI), 653828/1-653828/12.

475

208. Cotte-Rodrıguez, I., et al. (2006) Rapid trace detection of triacetone triperoxide (TATP) by complexation reactions during desorption electrospray ionization. Chemical Communications, 2006, 953–955. 209. Cotte-Rodrıguez, I., et al. (2008) In situ trace detection of peroxide explosives by desorption electrospray ionization and desorption atmospheric pressure chemical Ionization. Analytical Chemistry, 60, 1512–1519. 210. Mullen, C., et al. (2006) Detection of explosives and explosives-related compounds by single photon laser ionization time-of-light mass spectrometry. Analytical Chemistry, 78, 3807–3814. 211. Schramm, E., et al. (2009) Trace detection of organic compounds in complex sample matrixes by single photon ionization ion trap mass spectrometry: real-time detection of security-relevant compounds and online analysis of the coffee-roasting process. Analytical Chemistry, 81, 4456–4467. 212. Gillen, G., et al. (2006) Characterization of high explosive particles using cluster secondary ion mass spectrometry. Rapid Communications in Mass Spectrometry, 20, 1949– 1953. 213. Na, N., et al. (2007) Direct detection of explosives on solid surfaces by mass spectrometry with an ambient ion source based on dielectric barrier discharge. Journal of Mass Spectrometry, 42, 1079–1085. 214. Fetterolf, D.D., et al. (1994) Portable instrumentation: new weapons in the war against drugs and terrorism. In Proceedings of SPIE-The International Society for Optical Engineering 2092. 215. Kibelka, G., Gardner, B. (2010) Workshops on harshenvironment mass spectrometry, http://hems-workshop. org/index.html (accessed October 18, 2011) 216. Lammert, S.A., et al. (2006) Miniature toroidal radio frequency ion trap mass analyzer. Journal of the American Society for Mass Spectrometry, 17, 916–922. 217. Contreras, J.A., et al. (2008) Hand-portable gas chromatograph-toroidal ion trap mass spectrometer (GC-TMS) for detection of hazardous compounds. Journal of the American Society for Mass Spectrometry, 19, 1425–1434. 218. Bryden, W.A.B., et al. (1995) Tiny-TOF-MALDI mass spectrometry for particulate drug and explosives detection. In Proceedings of SPIE––The International Society for Optical Engineering, 2511, pp. 153–164. 219. Takada, Y. (2007) Explosives trace detection by mass spectrometry. Journal of the Mass Spectrometry Society of Japan, 55, 91–94.

21 MASS SPECTROMETRY IN HOMELAND SECURITY Yasuaki Takada

21.1

INTRODUCTION

Due to various changes in world circumstances, the threat of terrorism has become a serious problem for all countries. For example, military explosives are traded on the black market, and the general public has easy access to information on the assembly of improvised explosives via the Internet. In Japan, a high school student attacked a train station using an improvised explosive device (IED) several years ago. Therefore, to maintain a safe society, detection technologies for hidden explosive devices are in high demand [1–3]. In this chapter, the background, overview, and challenges in using mass spectrometry (MS) to detect various explosives are described.

21.2 DETECTION METHODOLOGY: BULK DETECTION AND TRACE DETECTION Two methods are primarily used to detect hidden explosives, as shown in Figure 21.1: bulk detection, which determines the existence of suspicious objects such as knives, irearms, and explosive devices from their shapes; and trace detection, which detects the presence of explosive contaminants by chemical analysis of vapor from objects. In bulk detection, physical imaging technology, which features the use of X-rays, neutron beams, and electromagnetic waves, for example, is applied to identify explosives and explosive devices by exposing the size, shape, and weight of the suspicious object in an image.

However, it is dificult to determine the identity of the object because the molecules of explosives consist of carbon, nitrogen, oxygen, and hydrogen, and furthermore, the elemental characteristics of the explosives are similar to those of food and common goods. Therefore, the false-positive rate of bulk detection is relatively high, namely, 10% or more. In trace detection, on the other hand, chemical analysis methods are applied to check for the existence of trace contamination of explosives on a passenger’s body, clothes, and luggage [1]. Trace detection methods have higher selectivity capabilities than bulk detection methods, and the false-positive rate is typically below 1%. However, even if an explosive contamination is conirmed or detected on a bag or passenger, the bag or passenger may not actually be carrying the explosive device at that time. The bulk and trace detection methods have different characteristics, and they complement each other’s weak points. Therefore, the combined use of bulk detection and trace detection is effective to improve security at important facilities. Bulk detection is a popular method that has been widely used, but trace detection technologies have become increasingly necessary as security concerns have grown. For example, wiping tests for luggage using ion mobility spectroscopy or chemiluminescence detection have been implemented as part of aviation security. However, the current trace detection methods mentioned above suffer from several problems. For example, the selectivity of ion mobility spectroscopy is not adequate, resulting in false-positive alarms in some cases.

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

477

478

MASS SPECTROMETRY IN HOMELAND SECURITY

Bulk detection

X-ray inspection

Discrimination of materials by X-ray absorption

Trace detection

Handling of explosives

Contaminant residue adheres to personal effects

Detection of explosives contamination

FIGURE 21.1 Detection methods for hidden explosive devices: bulk detection and trace detection.

On the other hand, the selectivity of chemiluminesence detection is suficiently high, but this method requires gas chromatography (GC) separation, and it is dificult to obtain quick responses for on-line use. Of the trace detection technologies, MS is a promising tool because it offers high sensitivity, high selectivity, and a short detection interval. Therefore, MS applications have been in high demand for practical security uses.

21.3

INSTRUMENTATION

For the ion source and the mass spectrometer, atmospheric pressure chemical ionization (APCI) with an ion trap mass spectrometer (ITMS) is mainly used. The ion source and the mass spectrometer used in our laboratory for detection of explosives are described below. 21.3.1

Ion Source

APCI is a well-known technique for ionizing explosive molecules. A high voltage of about −2 kV is applied to a needle electrode to produce a negative corona discharge. The typical ionization process of the explosive molecules (M) consists of two steps: O2 + e− → O2− (primary ionization by negative corona discharge), then O−2 + M → (M-H)− + HO2 (secondary ionization by ion-molecule reaction). However, nitrogen monoxide (NO) is also produced by the corona discharge, and NO reacts with O2− to produce NO3−. This reaction reduces the concentration of O2− and affects the ionization eficiency of the sample molecules.

1HHGOH &RURQDGLVFKDUJH UHJLRQ HOHFWURGH ±N9 6HFRQGDU\LRQL]DWLRQUHJLRQ 7RPDVVVSHFWURPHWHU $SHUWXUH PPLG *DV  /PLQ

&RXQWHUHOHFWURGH ±N9

FIGURE 21.2 Atmospheric pressure chemical ionization ion source with “counterlow introduction.”

To eliminate neutral NO molecules from the secondary ionization region, a novel APCI ion source with “counterlow introduction (CFI)” has been developed in our laboratory [4,5]. Figure 21.2 shows a schematic diagram of the newly developed ion source. In this ion source, the sample gas lows into the secondary ionization region, then lows to the corona discharge region, where negative ions are generated. The negative ions are extracted in the direction opposite to the gas low by the electric ield. In our new ion source, neutral NO molecules produced by the corona discharge are eliminated from the secondary ionization region by the gas low. 2,4-Dichlorophenol (DCP) was selected as a model compound to evaluate the new ion source because it is easily ionized using the negative APCI mode. Typical mass spectra of DCP obtained with a conventional ion source and the new ion source are shown in Figure 21.3A,B, respectively. The observed ion intensity of the DCP ions was greatly improved by using the new ion source, although the ion intensity of NO3− was reduced. Figure 21.4 shows plots that illustrate the relationship

INSTRUMENTATION

5HODWLYHLQWHQVLW\ 

$

479

%



12±

0±+ í







0± + í

12í





P]

P]



FIGURE 21.3 Comparison of mass spectra for dichlorophenol. (A) Conventional APCI ion source and (B) APCI ion source with “counterlow introduction.”

,RQLQWHQVLW\

25,000 20,000 O2 – – CO3 – NO3 (M-H)–

15,000 10,000 5,000 0 0

200

400

600

&RXQWHUJDVIORZUDWH P/PLQ FIGURE 21.4 Dependence of ion intensity on counter gas low rate.

between the gas low rates and observed ion intensities of various ions. The intensities of the O2− and DCP ions increased as the gas low rate increased, but that of NO3− decreased. 21.3.2

ITMS

Various mass spectrometers such as a single quadrupole, triple quadrupole, and quadrupole ITMSs are used for the detection of explosives. However, the ITMS has several advantages for homeland security applications. For example, the quadrupole ion trap has high sensitivity with a full scan mode and has tandem mass spectrometry (MS/MS) capabilities. Furthermore, the quadrupole ion trap is robust, and complicated calibrations are not necessary with this instrument format. We have primarily used the quadrupole ITMS in our laboratory for homeland security applications [5]. Recently, a linear ITMS with wire electrodes (wire-LIT)

was developed in our laboratory [6]. Figure 21.5A illustrates the wire-LIT format. The wire-LIT consists of an inlet lens, quadrupole rods, a trap wire, an extraction wire, and extraction lenses. The rod length of the wire-LIT is 200 mm, and the rod radius (R) is 5 mm. Two phases of trapping radio frequency (RF) voltage (834 kHz) are applied to rod pairs. The wire electrodes are placed between the quadrupole rods. The trap wire and extraction wire are aligned orthogonally. The trap wire conines ions, and the extraction wire is responsible for the extraction of excited ions in an axial fashion. The extraction lenses are aligned parallel to the extraction wire. Helium buffer gas of 0.3 mTorr is introduced into the wire-LIT. The measurement sequence for the analysis of negative ions is shown in Figure 21.5B. The sequence consists of accumulation, cooling, scanning, and emptying periods. During the accumulation period, a negative DC bias is applied to the inlet lens and the trap wire, which produces a trapping DC potential in the axial direction, whereas the quadrupole rods produce a radical potential. The trapping eficiency was 40% at an injection energy of 20 eV. During the cooling period, ions are cooled by collisions with buffer gas. During the scan period, a positive DC bias is applied to the extraction wire, and a negative DC bias is applied to the trap wire. A supplemental AC voltage (5–50 V0-peak, 360 kHz) is applied to the excitation lenses. The speciic m/z ions are resonantly excited in the direction parallel to the extraction wire (x-direction) by the supplemental AC voltage. Ions with large x-motions pass over the trap wire, while the presence of ions with small x-motions is conirmed. Ions that pass over the trap wire are extracted axially from the wire-LIT by the extraction wire. When the trap RF amplitude is scanned, ions are mass-selectively ejected sequentially and detected by an electron multiplier. During the emptying period, all ions in the LIT

480

MASS SPECTROMETRY IN HOMELAND SECURITY

(A) Extraction lens

Extraction wire Extraction lens

Quadrupole rod

Extraction wire

Inlet lens To detector

Ions Ion excitation direction

Ion extraction direction

Extraction lens

Trap wire

Trap wire

Extraction lens

(B) Inlet lens

Trap wire

Extraction wire

Trap RF

Supplemental AC Scanning Accumulation

Emptying

Cooling

FIGURE 21.5 Novel linear ion trap mass spectrometer with wire electrodes (wire-LIT). (A) Schematic diagram of the wire-LIT and (B) the measurement sequence for the analysis of negative ions.

21.4 21.4.1

DETECTION OF EXPLOSIVES BY MS Sampling of Explosives Vapors and Particles

Figure 21.6 illustrates the vapor pressure of various explosives. To detect explosives with higher vapor pressure such as 2,6-dinitrotoluene (DNT), 2,4,6-trinitrotoluene (TNT), and glycerol trinitrate (NG), vapors from the explosives are directly introduced into the ion source by a pump. Figure 21.7 shows a prototype of a vapor detector. A lexible tube is connected to the APCI ion source. The end of the tube is moved toward objects to be checked such as hands, clothes, and luggage. The sample vapor is collected by the lexible tube and introduced into the APCI ion source. The temperature of the inner

(WK\OHQHJO\FROGLQLWUDWH (*'1



9DSRUSUHVVXUH 3D

are emptied outside the trap by shutting off the trap RF voltage. The sensitivity of our novel wire-LIT is one order of magnitude higher than that of the quadrupole ITMS. The wire-LIT is applied for a walkthrough detection of improvised explosives, mentioned in Section 21.5 in this chapter.

7ULDFHWRQHWULSHUR[LGH 7$73

 

'LQLWURWROXHQH '17

 ± 

*O\FHUROWULQLWUDWH 1* 'LQLWURWROXHQH

 ±   ±   ± 

7ULQLWURWROXHQH 717 7ULQLWURWULD]DF\FORKH[DQH 5';

3HQWDHU\WKULWROWHWUDQLWUDWH 3(71







0ROHFXODUZHLJKW FIGURE 21.6 Vapor pressures of various explosives.

surface of the lexible tube is set at 100°C to prevent the absorption of explosive vapors onto the surface. The ions produced at the ion source are then analyzed by a single quadrupole mass spectrometer. Figure 21.8A,B indicates the detection of TNT adhering to a hand. The TNT-contaminated hand was moved

DETECTION OF EXPLOSIVES BY MS

$ Mass spectrometer

6DPSOLQJ

㻜㻚㻟 

Flexible tube

FIGURE 21.7 Prototype of a vapor detector.

close to the sampling probe, and a strong signal was detected on the mass chromatogram even after attempts to remove contamination by wiping off the hand with paper or cloth (Figure 21.8A). After carefully washing the hand several times in soap and water, the TNT signal was still detected as shown in Figure 21.8B because nitro-compounds are easily absorbed by the hands when a terrorist handles explosives. To detect lower vapor pressure explosives such as 1,3,5-trinitro-1,3,5-triazacyclohexane (RDX) and pentaerythritol tetranitrate (PETN), on the other hand, ine particles of explosives adhering to luggage, clothes, and the body are wiped with a wiping sheet. Then, the wiping sheet is set on a heating unit consisting of a tray and a heated metal box. The temperature of the heating unit is set at 150–200°C. Chemical molecules vaporized from the wiping sheet are introduced into the ion source and analyzed. 21.4.2

Organic Acids as Sensitive Dopant

In general, impurities degrade the analytical performance of conventional mass spectrometers because they cause very complicated mass spectra and affect the ionization eficiencies of the chemicals to be detected. A potential problem with APCI, where ionized molecular species can react eficiently with other neutral species, is charge competition between an explosive molecule and an interfering molecule, which may result in reduced sensitivity to explosives when an excessive quantity of interfering molecules is present. To avoid this problem in negative ionization, a chloride gas is often added as a dopant in the ionization region to stabilize explosives detection by forming an adduct ion of explosives and dopant molecules.

㻜㻚㻢

㻜㻚㻥 

㻝㻚㻞

㻝㻚㻡 

㻞㻚㻝 

7LPH PLQ

%

㼠㼕㼙㼑㻌㻔㼙㼕㼚㻕

6DPSOLQJ

,RQ LQWHQVLW\

Ion source

P] 

,RQ LQWHQVLW\

Data processor

481

㻜 

㻝

㻞 

㻟

7LPH PLQ 㼠㼕㼙㼑㻌㻔㼙㼕㼚㻕

FIGURE 21.8 Detection of TNT vapor from hands. (A) After wiping hands with cloth and (B) after washing hands with soap and water.

We found a new phenomenon in which lactic acid (LA), a major component of impurities, increases the ionization eficiency of explosives by producing ions formed by the addition of LA to explosives [7]. Figure 21.9 shows a typical mass spectrum of the explosive RDX with and without a dopant. The ions of m/z 46 and 268, which correspond to NO2− and (M+NO2)−, respectively, are clearly observed in Figure 21.9A–C, which shows the mass spectra with dichloromethane and LA gases as the respective dopants. The ions of m/z 257 and 311 are newly observed in Figure 21.9B,C, respectively. When a dopant gas such as dichloromethane and LA is introduced into the ion source, the gas is ionized by negative APCI processes. The ions produced from the dopant gas react with the explosive molecules to produce adduct ions of the explosive molecules. The m/z 89 signal observed in Figure 21.9C corresponds to a deprotonated LA molecule, and the m/z 311 signal is an adduct ion of the deprotonated LA molecule and RDX (namely, [RDX+LA–H]−). Figure 21.10A,B plots the dependence of the dopant concentration on the observed signal intensities. In Figure 21.10A, the ion signal of m/z 268 decreases with an increasing concentration of dichloromethane, and the ion signal of m/z 257 increases with an increasing concentration of dichloromethane and becomes a dominant signal over 1000 ppm of dichloromethane. On the

482

MASS SPECTROMETRY IN HOMELAND SECURITY

(NO2) − m/z 46

(B)

(RDX + NO2) − m/z 268

Ion intensity (a.u.)

Ion intensity (a.u.)

(A)

10

5

0 0

100

200

300

(NO2) − m/z 46

(RDX + NO2) − m/z 268 (RDX + Cl) − m/z 257

10

5

0

400

0

m/z

100

200

300

400

m/z (La −H) − m/z 89

(C)

(RDX + La − H) − m/z 310

Ion intensity (a.u.)

(NO2) − m/z 46 10

(RDX + NO2) − m/z 268

5

0 0

FIGURE 21.9 acid dopant.

100

200 m/z

300

400

Mass spectra of RDX. (A) Without any dopant gas, (B) with dichloromethane dopant, and (C) with lactic

(A)

(B)

RDX+NO2 RDX+Cl

1.0E+08

Ion intensity (Counts)

Ion intensity (Counts)

1.0E+08

1.0E+07

1.0E+06

1.0E+05 10

RDX+NO2 RDX+La

100

1000

10,000

Concentration of chlorine dopant (ppm)

1.0E+07

1.0E+06

1.0E+05 0

10

20

30

Concentration of lactic acid dopant (ppm)

FIGURE 21.10 Dependence of ion intensity on concentration of dopant gases. (A) With dichloromethane dopant and (B) lactic acid dopant.

other hand, LA was tested as a dopant gas as illustrated in Figure 21.10B. The ion signal of m/z 311 becomes a dominant signal with a low concentration of LA, and 15 ppm of LA is enough to obtain a strong signal of the adduct ions.

LA is a major component of impurities. In practical wiping tests of explosives, a suitable amount of LA is collected by wiping luggage, and it is not necessary to add extra LA. During security applications in the ield, the heating unit and the ion source of the explosives

DETECTION OF EXPLOSIVES BY MS

Detection of Black Powder

Black powder has been used in many improvised bombing incidents. In Japan, in particular, the black powder contained in ireworks has often been used for criminal bombing attacks because military and industrial explosives are strictly controlled. It is therefore important to detect black powder to improve security against improvised bombs. Despite this importance, a practical instrument for detecting black powder has not yet been reported, although sulfur clusters have been observed in a mass spectrum of black powder using glow discharge ionization under vacuum [8]. In this section, we show that a mass spectrometer with APCI can be used as a practical explosives detector for black powder [9]. Typical black powder consists of charcoal, sulfur, and potassium nitrate. A few particles of black powder used for propellants were placed in the heating unit set at 150°C, and the mass spectra of the gas evaporated from the black powder were recorded under the negative APCI mode. Figure 21.11 shows a typical mass spectrum. The ion signal m/z 96 is dominant, and signals for m/z 64, 98, and 128 are also observed. Characteristics of the APCI mass spectrum were similar to the result of the glow discharge ionization MS. Figure 21.12 illustrates the fragmentation of the ions with m/z 96. From m/z 96, the ions with m/z 32, 64, and 80 were produced

Ion intensity

S 3−(partly

16000 16,000 14000 14,000 12000 12,000 10000 10,000 8000 8000 6000 6000 4000 4000 2000 2000 00 20 20

96

96

SS 22−

64 64 70 70 m/z

12,000 12000 10,000 10000

98

98

SS 4 4−

128

128

120120

FIGURE 21.11 Typical mass spectrum of black powder.

96 96

8000 6000 4000 2000 0 20

S2 O2− )

Ion intensity

21.4.3

by collision-induced dissociation. We thus conclude that the ions with m/z 96 were mainly trimer ions of sulfur (S3−). In addition, part of the m/z 96 ions includes oxygen molecule adduct ions (S2O2−) because the fragment ions m/z 80 most likely correspond to S2O−. Thus, in the negative APCI mode, a series of sulfur cluster ions such as S2− (m/z 64), S3− (m/z 96), and S4− (m/z 128), are mainly observed in the mass spectrum from black powder. We selected the S3− ion (m/z 96) and its isotope ion (m/z 98) as the markers for detection because the signal-to-noise ratio (S/N) of the signals was suficiently high. The detection limits of an explosives detector depend on the alarm threshold level. To suppress false-positive alarms under practical conditions, we adjusted the alarm threshold level according to the wiping test results for 100 hand luggage items, which were chosen from dailyuse bags at the authors’ place of work. Figure 21.13 shows the intensity distribution of the maximum ion signals from the hand luggage items in the detection period (0–10 s after the wiping sheet has been inserted) for (A) m/z 96 and (B) m/z 98, respectively. Almost all of the signals detected are below 4 × 10−10 A, but the maximum signal, detected at m/z 96, is 1.7 × 10−9 A. The average of the m/z 96 signal distribution is 2.0 × 10−10 A, and the standard deviation (σ1) is 1.6 × 10−10 A as observed in Figure 21.13A. The average of the m/z 96 signal distribution is 1.7 × 10−10 A, and the standard deviation (σ2) is 5.5 × 10−11 A as observed in Figure 21.13B. The threshold for m/z 96 was set at 3 × 10−9 A (18 σ1) and the threshold for m/z 98 was

Ion intensity

detector are heavily contaminated by a large concentration of impurities caused by sweat, food residue, oils, cosmetics, or other daily goods. An impurity can be used to assist ionization as mentioned above. The “impurityassisted ionization” enables accurate detection of explosives during long operation times without requiring frequent adjustments or cleaning. The recommended calibration interval of the explosives detector that was developed in our laboratory is only once a week.

483

1000 900 800 700 600 500 400 300 200 100 0

40

60

80

100

120

80

100

120

64 64 32 32 20

40

60

m/z FIGURE 21.12 MS/MS analysis from m/z 96.

484

MASS SPECTROMETRY IN HOMELAND SECURITY

set at 1 × 10−9 A (15 σ2). The estimated false-positive rate (which is the probability that both signals will accidentally exceed the threshold level at the same time) is negligible. With these thresholds, the detection limits of our system were evaluated. An ethanol solution of sulfur

(A) Number of times

50 40 Threshold 3 × 10−9A

Max: 1.6 × 10−9A

30 20

(5 ng/µL) was dropped onto the wiping sheets, and the wiping sheets were placed in the heating unit. Under the alarm threshold level mentioned above, the detection limit of sulfur was 50 ng, which corresponds to about 500 ng of black powder. The selectivity of our system was determined by using some everyday goods, listed in Table 21.1. The results for a hot spring bath additive (300 µg of powder) and some hair dyes are listed in Table 21.2. No falsepositive alarms were detected out of four samples. Therefore, the system deemed to have suficient selectivity for practical use. Figure 21.14 shows the results of a wiping test of chemicals used in ireworks. The chemicals tested were

10 Signal 0

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

m/z 98

Ion signal ( × 10 −10 A) (B)

10

Threshold m/z 96 m/z 98

50 −8

Ion current (A)

Threshold 1 × 10 −9 A Number of times

m/z 96

−7

40 30 Max: 3.7 × 10

20

−10

A

10

10− 9 10−10

10 0

10 −11 − 20

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

−10

0

Ion signal ( × 10 −10 A)

TABLE 21.1

Everyday Goods Containing Sulfur

No.

Everyday Goods Hot-spring bath additive Hair dye (A) Hair dye (B)

TABLE 21.2 No.

1 2 3 4

20

Time (s)

FIGURE 21.13 Intensity distribution of maximum ion signals from hand luggage items in the detection period (0– 10 s after the wiping sheet has been inserted) for (A) m/z 96 and (B) m/z 98.

1 2 3

10

Placement of wiping sheet

FIGURE 21.14 Wiping test result of the chemicals in ireworks. Several ingerprints of the same inger were obtained and the tenth print was tested.

Name of Product and Manufacturer

Remarks

Tabi no yado (Noboribetsu), Kanebo, Ltd. Wella treatment color intensive, Wella Japan, Ltd.

Two hair dyes (A and B) were mixed before use.

Detection Results of Everyday Goods Containing Sulfur Sample

Hot-spring bath additive Hair dye (A) Hair dye (B) Hair dye (A + B) Alarm threshold

Ion Signal (A) m/z = 96

m/z = 98

9.6 × 10−11 (BG level) 2.9 × 10−9 1.8 × 10−9 2.2 × 10−9 3 × 10−9

6.8 × 10−11 (BG level) 2.0 × 10−10 (BG level) 7.5 × 10−10 3.3 × 10−10 (BG level) 1 × 10−9

Detection Result None None None None

Note: BG level means that the detected signal was almost the same as the background level from the environment, so no signiicant increase in the ion signal was observed.

DETECTION OF EXPLOSIVES BY MS

TABLE 21.3 Collected No. 1 2 3

Fireworks from Which Chemicals Were

21.4.4

Type of Firework

Name (in Japanese)

Amount of Chemicals per Firework

Firecracker Sparkler Rocket

Junrikouhou Hoka Shunrai

Less than 0.04 g 0.08 g 0.6 g

collected from the ireworks listed in Table 21.3. All of these ireworks are made in China and are available to the general public at toy shops in Japan. Small amounts of samples were observed to adhere to the index inger of one of the authors. Several ingerprints from that inger were then transferred onto plastic surfaces, and the tenth ingerprint was wiped with a wiping sheet. As shown in Figure 21.14, the signals for m/z 96 and 98 exceeded the set threshold levels. We repeated the wiping tests 10 times, and in seven of the tests, both the signals m/z 96 and 98 exceeded the alarm threshold levels for sulfur. In the remaining three tests, the signals did not exceed the alarm levels at the same time. When the ifth ingerprint in each of these three tests was rewiped, both the signals m/z 96 and 98 exceeded the alarm threshold levels. Consequently, our test results showed that the APCIMS system is suitable for detecting black powder in ireworks using the wiping test. Speciically, good sensitivity (500 ng for black powder) was obtained, and there were no false positives from everyday goods containing sulfur.

Detection of Improvised Explosives

Recently, the threat of IEDs has become more serious. Peroxides such as triacetone triperoxide (TATP) and hexamethylene triperoxide diamine (HMTD) are often used in criminal bombing attacks because they can be easily synthesized in the home. Detection methods and procedures regarding improvised explosives are almost the same as for those used to detect military explosives; however, a positive APCI mode is more suitable for detecting TATP and HMTD. Positive APCI mass spectra of TATP and HMTD are shown in Figure 21.15A,B, respectively. A negative APCI mass spectrum of their precursor, hydrogen peroxide, is also shown in Figure 21.15C. As shown in Figure 21.15, the improvised explosives are also detected by APCI-ITMS. 21.4.5

Detection of Postblast Residues

Identiication of postblast residues of explosives is also an important analytical challenge [9]. In the forensic science ield, the postblast residues adhering to samples are extracted using an organic solvent and the solution is analyzed by GC/MS. However, faster identiication of the explosives used in a bombing would lead to faster apprehension of the perpetrators. We conducted tests to detect postblast residues by using the MS-based explosives detector without any pretreatment of the samples. Small pieces of metal, acryl, and cloth were tested, which were remnants of explosions using black powder, irecrackers, dynamite, and electric detonators (see (B)

(A) 8000 7000 6000 5000 4000 3000 2000 1000 0 30 40 50 60 70 80 90 100

(C) 18,000 16,000 14,000 12,000 10,000 8000 6000 4000 2000 0 20

FIGURE 21.15 peroxide.

485

40

14,000 12,000 10,000 8000 6000 4000 2000 0

60

0

100

80

200

300

400

100

Mass spectra of improvised explosives and their precursor. (A) TATP, (B) HMTD, and (C) hydrogen

486

MASS SPECTROMETRY IN HOMELAND SECURITY

(A)

(B)

Metal Cloth Acryl

Acryl

Cloth

Metal

Iron pipe (explosive device) FIGURE 21.16 (A) Setup for explosion and (B) samples produced in the explosion.

$

%

21

1

Ion intensity

2 12

(B) m/z 227 (NG)



,RQLQWHQVLW\

(A) m/z 214 (EGDN)

  





P]

FIGURE 21.18 (A) Molecular structure of DDNP and (B) APCI mass spectrum of DDNP.

(C) m/z 289 (NG) Cloth placed in unit

(A)

0

100

200

300

Time (s) FIGURE 21.17 Signals obtained from a piece of cloth produced in an explosion of dynamite. (C) 181

183

151

0

100

200 m/z

181

Ion intensity

Figure 21.16). The samples were placed in the heating unit without any pretreatment. The temperature of the heating unit was set at 150°C for metal, but only 100°C for the acryl and cloth in order to reduce the thermal decomposition of the sample materials. Figure 21.17 shows the detection results of the postblast residues adhering to a cloth sample produced by the explosion of dynamite. As shown in Figure 21.17, the components of dynamite (i.e., ethylene glycol dinitrate [EGDN] and NG) were detected from the tested sample. Figure 21.18 illustrates the molecular structure and a typical mass spectrum of diazodinitrophenol (DDNP). DDNP is a primary explosive used for an electric detonator that is only made in Japan. Therefore, we can conclude that if DDNP was detected, the electric detonator that was made in Japan was used in the bombing. Figure 21.19A shows a photograph of a typical metal fragment produced by the explosion of the electric deto-

Ion intensity

(B)

300

151 121

0

100

(M-N2-H)–

200 300 m/z

FIGURE 21.19 (A) Piece of metal from an electric detonator produced in the explosion of the detonator, and mass spectra obtained from metal pieces (B) a few days after explosion and (C) 2 months after explosion.

nator. Figure 21.19B,C illustrates the temporal changes in the mass spectra obtained from both “fresh” blast and “old” blast fragments. From the fresh fragments, obtained a few days after the explosion, the DDNP signal m/z 183 is clearly observed, as shown in Figure 21.19B. However, in the case of the old fragments, 2

SYSTEM INTEGRATION

months after the explosion, m/z 181 is the dominant signal, as evident in Figure 21.19C. Consequently, we assumed that the m/z 181 signal is that of the decomposed product of DDNP. As shown above, quick identiication of postblast residues using the explosives detector will be a powerful tool for counterterrorism measures.

21.5

SYSTEM INTEGRATION

A high-throughput walkthrough detection system that features the use of MS is currently under development that aims to prevent future terrorist attacks in public transportation areas, sports stadiums, shopping malls, and other high-density areas [10]. Figure 21.20A,B shows the schematic images of two versions of the explosives detec-

(A)

tion system. When terrorists prepare an explosive device, their hands, clothes, luggage, and other everyday belongings are contaminated with trace amounts of the explosive they handled. In the case of the “portal version” (e.g., targeting use at airports), shown in Figure 21.20A, when terrorists pass through the portal, vapors from the explosive contaminants on their body, clothes, and luggage are detected. The target throughput of this system is 1200 persons per hour. In contrast, the “automatic ticket gate” version, shown in Figure 21.20B, is designed to check for explosive contamination on hands, tickets, integrated circuit (IC) cards, and so on. The target throughput of this system is 3600 persons per hour. A prototype of the portal version is shown in Figure 21.21. In preliminary trials, vapors emitted from a small amount of a common improvised explosive, TATP, were clearly detected within only 2 s of a test subject with

(B)

Air blower

Wireless communication

Vapor sampling ports

Detector

Vapor sampling ports

Detector

FIGURE 21.20 Images of walkthrough detection systems. (A) Portal version and (B) automatic tick gate version.

(A)

(B)

Vapor sampling ports

Detector

487

Air blower

FIGURE 21.21 (A) Prototype system of portal version and (B) ield test at the train station.

488

Ion intensity

MASS SPECTROMETRY IN HOMELAND SECURITY

Time

FIGURE 21.22 Walkthrough detection of TATP. The arrows show characteristic signals of TATP.

Video monitoring

Platform Explosives detector

Bulk detector Trace detector

Ticket gate

FIGURE 21.23 Image of a future train station with detection system in place.

TATP passing through the portal, as shown in Figure 21.22. The arrows in the igure highlight the characteristic signals of TATP. This system will be useful to prevent suicide attacks of terrorists. Figure 21.23 shows an image of a future train station. High-throughput explosives detectors are installed in many places throughout the terminal and are connected to video monitoring systems. A passenger detected with traces of explosive contaminants is identiied by an alarm on the explosives detector and automatically tracked by the video monitoring systems.

cation of MS in counterterrorism efforts is of increased importance as a sensitive and selective method to detect explosives and improve homeland security.

21.6

REFERENCES

CONCLUSION

In this chapter, the challenges involved in detecting explosives using MS are described. In trace detection, ion mobility instruments are most widely used. However, terrorists are becoming shrewder. Consequently, appli-

ACKNOWLEDGMENTS This work was partly supported by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. The author expresses his thanks to Dr. Toshihiko Ohta of NOF Corporation for his helpful comments.

1. Woodin, R.L. (2007) Trace Chemical Sensing of Explosives. Hoboken, NJ: John Wiley & Sons. 2. Schubert, H., Kuznetsov, A. (2006) Detection and Disposal of Improvised Explosives. Dordrecht: Springer.

REFERENCES

3. Schubert, H., Kuznetsov, A. (2008) Detection of Liquid Explosives and Flammable Agents in Connection with Terrorism. Dordrecht: Springer. 4. Kojima, K., Sakairi, M., Takada, Y., Nakamura, J. (2000) Vapor detection of TNT and RDX using atmospheric pressure chemical ionization mass spectrometry with counter-low introduction (CFI). J. Mass Spectrom. Soc. Jpn., 48, 360. 5. Takada, Y., Nagano, H., Suga, M., Hashimoto, Y., Yamada, M., Sakairi, M., Kusumoto, K., Ota, T., Nakamura, J. (2002) Detection of military explosives by atmospheric pressure chemical ionization mass spectrometry with counter-low introduction. Propell. Explos. Pyrot., 27, 224. 6. Sugiyama, M., Hasegawa, H., Hashimoto, Y. (2009) Massselective axial ejection from a linear ion trap with a direct current extraction ield. Rapid Commun. Mass Spectrom., 23, 2917.

489

7. Nagano, H., et al. (2006) Proceedings of the 4th International Aviation Security Technology Symposium, November 27–December 1, 2006, Washington, DC, p. 63. 8. McLuckey, S.A., et al. (1989) Proceedings of the 3rd International Symposium on the Analysis and Detection of Explosives, p. 25. 9. Takada, Y., et al. (2004) Proceedings of the 8th International Symposium on the Analysis and Detection of Explosives, June 6–10, 2004, Ottawa, Canada, p. 353. 10. Takada, Y., Nagano, H., Suzuki, Y., Sugiyama, M., Nakajima, E., Hashimoto, Y., Sakairi, M. (2011) High-throughput walkthrough detection portal for counter terrorism: detection of triacetone triperoxide (TATP) vapor by atmospheric-pressure chemical ionization ion trap mass spectrometry. Rapid Commun. Mass Spectrom., 25, 2448.

22 MEASUREMENTS OF SURFACE CONTAMINANTS AND SORBED ORGANICS USING AN ION TRAP SECONDARY ION MASS SPECTROMETER Gary S. Groenewold, Anthony D. Appelhans, Garold L. Gresham, and John E. Olson

22.1

INTRODUCTION

The outermost molecular layer of any solid surface consists of adsorbed molecules that originate from the luid environment. The luid may be the ambient atmosphere, a covering liquid, or even a fairly aggressive vacuum. And in each of these cases, adsorbates will cover the surface of a solid. Thus, a solid surface can in many cases concentrate compounds in two dimensions, which can make them easier to detect, compared with analytical strategies in which the analytes are dispersed into three dimensions. Usually molecules that have signiicant functional groups are most aggressively adsorbed, and in many instances, these molecules are of high analytical interest. The functionalized molecules can be bound strongly, which can defeat attempts to remove them by solvent extraction. This dificulty and the need for direct analytical strategies motivated the development of direct surface analysis, one of which is the ion trap secondary ion mass spectrometer (IT-SIMS). In addition to surface adsorption, contaminant molecules can also be absorbed, which occurs by an initial adsorption event, followed by permeation into the bulk of the solid. Absorption can occur by permeation of soluble contaminant molecule into the bulk, or by migration of molecules into cavities of porous or cracked solids. It can be argued that the latter case is really a type of adsorption to the inner surfaces of the pores; however, disposition of this sort results in contaminant

molecules in a location that is not amenable to a direct surface analysis. Molecules in this realm present a special case for which there is a need for an alternative, direct analysis method. In this protocol, direct surface analyses using ITSIMS will be described from the perspective of analyses that were conducted to measure chemical warfare agents and related compounds (precursors, degradation products, and surrogates) that were adsorbed on samples from exposed environments. During the course of the research that is described in this chapter, the basic instrumental design of the IT-SIMS was modiied for the analysis of absorbed compounds, and so the utilization of the instrument for these purposes is also described. 22.1.1 Initial Studies of Molecular Contaminants Using Static Secondary Ion Mass Spectrometer (SIMS) Initial studies of molecular contaminants present on solid surfaces were conducted using SIMS operated in what is termed the static mode. In the experiment, the surface is bombarded with an energetic primary ion, usually having kinetic energies on the order of several kilovolts to ∼15 kV. Impact produces a wide range of species, including atoms and molecules that can be either neutral or charged and abundant electrons. The ionic species may be positively or negatively charged, and either polarity may be measured on the basis of

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

491

492

MEASUREMENTS OF SURFACE CONTAMINANTS AND SORBED ORGANICS

their mass-to-charge ratio (m/z). By way of contrast, we note that under high ion luence conditions, the surface is seriously damaged and atomic ions are the predominant species that are produced: this is referred to as a dynamic SIMS experiment. In the static SIMS regime, the total primary ion dose does not exceed 1012–1013 ions/ cm2, and while higher doses do not necessarily atomize everything on the surface, they do cause alteration of the molecular structure of the adsorbed molecules. Since this can confuse the identiication of molecular surface contaminants, it is desirable to keep doses below this level. Benninghoven and coworkers were the irst to employ static SIMS to interrogate material surfaces [1–3], and discovered that the positive secondary ion mass spectrum contained abundant species having the general formula CxHy+, which was derived from organic contaminants on the surface. Using 3 keV Ar+ as a primary ion, parafinic cations were produced (exclusively even-electron species), while conversely, intact molecular ions were not produced. However, the overarching conclusion from these studies was that the mass spectra were dominated by ions originating from the surface adsorbates, and contained very little information about the underlying solid material. In contrast, the negative-ion mass spectrum showed abundant ions corresponding to chromium and vanadium oxide materials that were being analyzed. 22.1.2 Enhancing the Production of Intact Molecular Ions: Polyatomic Projectiles and Managing Surface Charging Shortly after Benninghoven’s pioneering research, additional studies were undertaken by several groups that showed that when polyatomic cluster ions were used as projectiles, more intense intact molecular cations and clusters were produced off an atomic ion [4]. The ReO4− anion (perrhenate) is used in the IT-SIMS as a bombarding projectile and has been shown to produce superior molecular ion abundance, compared, for example, with Cs+ [5]. Primary ions are generated by heating a ilament coated with a Ba(ReO4)2/Eu or Y oxide ceramic, which results in volatilization of perrhenate that can be accelerated for use as a SIMS projectile [6]. The small size of the primary ion (perrhenate) gun makes it compatible with the compact nature of the IT-SIMS, and in addition, it operates robustly, putting out stable primary ion currents of hundreds of picoamperes, and introduces a zero gas load to the vacuum system. Another technological advance that was a key to the development of the IT-SIMS, particularly for analysis of electrically insulating surfaces, is the control of surface charging. As noted above, when primary ions impact

surfaces, a signiicant quantity of electrons are emitted, which can result in an accumulation of charge on the surface. The charge, which is normally positive, can be large enough to repel incoming positive ions and will adversely affect the emission of secondary ions. The latter can be severe, because they are generally formed with low kinetic energies. Hence, the development of a means to manage the surface charge has been a focus. Initial studies showed that by altering the ratio of cation/ anion extraction time, charge buildup could be signiicantly mitigated. Further studies indicated that even better charge management could be achieved using selfcharge-stabilizing ion optics [7]. A version of this arrangement was utilized in the initial versions of the IT-SIMS. Eventually, this system was supplemented by fabrication of a semicircular ilament that surrounded the sample probe. By maintaining an electrical current through the ilament, electrons were generated, some of which impacted the target surface, thereby neutralizing the surface charging. By running the current higher, the instrument could be operated in an electron ionization mode (instead of SIMS), which enabled measurement of neutrals that were volatilizing from the sample surface. 22.1.3

Overcoming the Chemical Background

The parafinic “background” that is observed in the majority of static SIMS spectra poses a problem to direct surface analysis, namely, these organic ions can confound measurement of trace levels of surface adsorbates because they dominate ion production, and the resulting isobaric interferences can confuse the interpretation of the surface spectra that are generated. Since the entire analysis occurs in the mass spectrometer, prior chromatographic cleanup or compound separation is not performed. Tandem mass spectrometry (MS/MS) is an alternative approach for the measurement of speciic analytes in the presence of high chemical background, and thus, MS/MS is highly complementary to direct surface analysis. The use of MS/MS allows ion isolation and fragmentation steps to be added to the direct surface analysis procedure to result in the following overall analytical process: (1) sputter desorption, (2) ion isolation, (3) ion fragmentation, and (4) ion scan out with m/z and ion intensity measurement. There are many commercially available MS/MS instruments. In the chemical weapon applications described herein, there is strong motivation for employing an MS/MS platform that is lightweight and easy to operate, and has a modest cost. The quadrupole ion trap (IT) has most of these attributes [8] and, in addition, is particularly amenable to SIMS analysis of environmental surfaces, because it operates effectively under modest

INSTRUMENTAL DESCRIPTION Deflector

End caps

Collector ReO4− ion gun

493

Leakage ring

Dynode

Filament Re

Repeller

Ring electrode

Direct insertion probe

FIGURE 22.1 Schematic diagram of the IT-SIMS.

FIGURE 22.2 Photographs of the IT-SIMS instrument based on the Teledyne ion trap. (A) Vacuum unit situated on top of a cart, next to a monitor. Pumping, electronics, and computer are shown below the cart platform. (B) Top view (glass cover removed). The perrhenate ion gun is the lens stack on the left side, the curved channeltron is in the center, and the end caps and ring electrode are on the far right side. The small ruler on the edge of the vacuum chamber is 15 cm.

vacuum conditions (on the order of 10−4–10−5 Torr). Thus, samples that undergo signiicant outgassing (e.g., soil particles or vegetation) will not adversely impact the operation of the mass spectrometer, which is not the case with many other types of mass analyzers.

22.2

INSTRUMENTAL DESCRIPTION

The IT-SIMS instruments were custom designed and fabricated in-house at the Idaho National Laboratory. All instruments were based on modiications of commercial gas chromatography/mass spectrometry (GC/ MS) instruments that were altered by removing the gas chromatography (GC) and inlet, and replacing it with a direct insertion probe, a modiied conversion dynode and a venetian blind electron multiplier, and a perrhenate primary ion gun. The primary ion gun and the direct insertion probe are located coaxial with the IT,

while the detector components are located off-axis on the ion gun side, as shown schematically in Figure 22.1. During the course of the research, IT-SIMS instruments were successfully fabricated, starting with ITs procured from Finnigan (San Jose, CA), Teledyne (Mountain View, CA), and Varian (Walnut Creek, CA). An implementation of this basic design is shown in Figure 22.2B, which is a top view of an IT-SIMS based on the Teledyne IT. The lens stack of the primary ion gun is on the left-hand side of the vacuum chamber, a curved channeltron detector appears in the middle, and the end caps and ring electrode are separated by Telon spacers on the right-hand side. The direct insertion probe approaches the IT from the right-hand side and is illustrated in Figure 22.2A on the front of the vacuum unit to the right of the monitor. This instrument its easily on a cart having a surface area of approximately 120 × 75 cm, with all electronics, computer, and pumping mounted below (Figure 22.2A).

494

MEASUREMENTS OF SURFACE CONTAMINANTS AND SORBED ORGANICS

FIGURE 22.3 Top view of loaded sample holders. Clockwise from the upper left: a ingernail clipping, with much of the nail head exposed; cotton ibers; soil particles; and glass ibers.

The analysis involved attaching the solid sample to the direct insertion probe, which was generally accomplished using double-sided tape. A small piece of tape was attached to the end of a small inishing nail (Figure 22.3) that functioned as a sample holder. The end of the direct insertion probe contained a cylindrical sleeve with a set screw that was used to attach the loaded sample holder. Prior to attachment, the sample holder was tapped to remove any loosely held material. The direct insertion probe was then placed into the insertion lock, which was evacuated. This evacuation procedure normally required several minutes; however, for samples such as damp soils, up to 5 min was sometimes allowed. Pressure in the forelock dropped into the millitorr regime, after which a ball valve was opened and the probe was inserted into the vacuum chamber of the ITSIMS. This caused a momentary rise in the pressure inside the IT-SIMS. The ultimate vacuum pressure achievable in the ITSIMS was on the order of 5 × 10−8 Torr; however, the instrument was never run at this pressure. He buffer gas is needed to stabilize ion trajectories in the trap and was admitted directly into the IT volume prior to inserting the sample. Typical PHe values used are on the order of 1 × 10−5 Torr as measured by an ion gauge located outside the IT itself. Since the IT volumes tend to be tight, the pressure inside the IT is somewhat higher, probably on the order of 1 × 10−4 Torr. Depending on the background pressure, PHe values two to three times higher were used at times; higher values tended to work better for cations, while lower values worked better for anions. These PHe values were lower than those used in conventional quadrupole ITs but were satisfactory for

trapping ions. Use of higher values resulted in attenuation of the primary ion beam, with attendant loss of secondary ions. The IT-SIMS, like all IT mass spectrometers, manipulates and measures ions using temporal sequences during which voltages and frequencies are applied to achieve the desired analytical outcome. A preliminary broadband pulse clears ions from the trap, and then the primary ion beam is directed through the center of the trap onto the sample target using a delector electrode (Figure 22.1). The primary ion beam can be on target anywhere from a few milliseconds all the way up to several seconds; however, a few tens to hundreds of milliseconds are typical. The secondary ions produced are all trapped in the IT during this time period, which ends when the perrhenate beam is pushed away from the axial apertures by applying a larger voltage to the delector electrode. After a few milliseconds settle time, the trapped ions can be either scanned out of the trap (generating an MS-1 spectrum) or isolated. Ion isolation is achieved using a iltered noise ield, in which frequencies are applied to the ring electrode to eject all ions except those that are wanted for further study; in practice, the frequency spectrum is applied at all frequency values except those corresponding to the secular frequencies of the ions of interest in the IT-SIMS. The amplitude is large enough to eject ions from the trap. The iltered noise ield is applied for on the order of 50–100 ms. Subsequent to ion isolation, MS/MS experiments are then performed by applying yet another frequency to the ring electrode, corresponding to the ions of interest, only with more modest amplitude. The kinetic energy of the ions increases and results in an increase in the internal ion energy resulting from what are now hyperthermal collisions with the He bath gas. The increased ion internal energy eventually produces fragmentation. As in the case of ion isolation, the duration of this ion fragmentation step and the amplitude can be varied in order to achieve optimum conversion of intact parent ions to fragments. After ion fragmentation, ions remaining in the IT are scanned and recorded.

22.3 ANALYSIS OF ALKYL METHYLPHOSPHONIC ACIDS (AMPAS) 22.3.1

Background for the Analysis of AMPAs

One of the irst direct surface analyses conducted using the IT-SIMS was the measurement of AMPAs on soil and vegetation. The AMPAs are the hydrolysis products of nerve agents (Scheme 22.1) and have proven dificult to detect on account of their tendency to strongly adsorb

ANALYSIS OF ALKYL METHYLPHOSPHONIC ACIDS (AMPAS)

495

SCHEME 22.1 Hydrolysis reactions of sarin (GB), soman (GD), and VX, forming isopropyl methylphosphonic acid (IMPA), pinacolyl methylphosphonic acid (PMPA), and ethyl methylphosphonic acid (EMPA), respectively.

to surfaces. The pKa for the AMPAs is around 3.3, and so at ambient pH values, they tend to exist as their conjugate bases and form strong ionic complexes with metal cations on surfaces. This tendency results in species that are resistant to solvent extraction. Detection is made more dificult by their tendency to undergo a second hydrolysis reaction, forming the diacidic methylphosphonic acid, which binds even more strongly.

22.3.2 Protocol for the Analysis of Soil Samples Dosed with AMPAs 1. Samples were generated to evaluate the eficacy of the IT-SIMS for the detection of the AMPAs [9,10], which involved exposing soil samples to solutions at concentrations of 100 ppm and allowed to dry. 2. Exposed soil samples were afixed to a sample holder for analysis using double-sided tape. In this case, the sample holder consisted of a #18 × 7/8″ inishing nail, with a head diameter of 2.7 mm. The average mass of the soil particles analyzed was 2.6 mg. 3. The instrument used in these analyses was a modiied version of the Finnigan ion trap mass spectrometer (ITMS) running ICMS software and coupled with a Teledyne HST-1000 iltered noise ield system. The end caps of the IT were modiied with larger apertures (5 mm) that were covered with spot-welded 90% pass grids. The instrument was operated in the negative-ion mode.

4. System base pressure was 5 × 10−8 Torr, and 9 × 10−5 Torr with the He bath gas was added. 5. The primary ion gun was operated at 5 kV at a current of 5 × 1010 A, as measured using picoammeter attached to a probe with an electron collector instead of a sample. Samples were irradiated for 0.9 s/scan, and 10 scans were collected and averaged. Depending on the persistence of the secondary ion signal, multiple experiments might be performed, but in all instances, the total primary ion dose to which the sample was exposed was ∼1 × 1012 ions/cm2, which is less than the commonly accepted static SIMS limit [11]. 6. After trapping sputtered ions, MS-1 spectra were acquired by scanning ions out of the IT [8], or MS-2 spectra were produced by irst isolating the ions of interest, energizing them to induce dissociation and then scanning the fragment ions out of the IT-SIMS. 7. Ion isolation was achieved using a iltered noise ield applied to the end caps to achieve capture of a mass range of approximately 10 u centered on the ion of interest. The voltage of the ield was adjusted such that unwanted ions were ejected, without unduly compromising the intensity of the ions of interest. 8. Ion dissociation was achieved by subsequent application of a voltage having a frequency equivalent to the secular frequency of the parent ion in the IT-SIMS. A typical voltage value was a few hundred millivolts.

496

MEASUREMENTS OF SURFACE CONTAMINANTS AND SORBED ORGANICS

22.3.3 Results of IT-SIMS Analyses of Soil Samples Dosed with AMPAs Negative-ion analysis using the IT-SIMS produces mass spectra with a low abundance ion corresponding to the intact conjugate base [AMPA–H]−, together with more abundant ions at m/z 95, 79, and 63 that corresponded to [CH3(OH)P(=O)O]−, [PO3]−, and [PO2]−, respectively (Figure 22.4). The latter ions arise from the fragmentation of the conjugate base as a result of the energy deposited into the desorbing ion during the sputtering event. They tend to be signiicantly more abundant than the intact conjugate base and are good indicators of past exposure to nerve agent. The limitation of these ions is that they are not indicative of which nerve agent was used. Thus, the intact conjugate base must be identiied. The ion at m/z 179, the pinacolyl (C6H13) version of [AMPA–H]−, is a good example, in that while it is observed in the mass spectrum, it is normally not above background (Figure 22.4). However, isolation of m/z 179 followed by fragmentation produces an abundant fragment ion at m/z 95, corresponding to the elimination of C6H12. The signal-to-noise ratio (S/N) of this ion is signiicantly greater than that observed for the AMPA diagnostic ions in the MS-1 spectrum, enabling unequiv-

ocal identiication of the degradation products at very low surface concentrations. The example shown in Figure 22.4 indicates a surface concentration of 0.02 molecular layer. Subsequent examples will elaborate on the correlation between mass applied and surface concentration.

22.4 ANALYSIS OF BLISTER AGENTS 22.4.1 Background for Analysis of Mustard Blister Agents An understanding of whether or not SIMS could be used to identify bis(2-chloroethyl)sulide was desired because of the widespread prevalence of this compound and related derivatives in environments where chemical weapons were used or stored. Like the nerve agents, characterization can be complicated by hydrolysis reactions, and by the formation of sulfonium conjugates that result from the condensation of the initially formed intermediate with excess sulide compounds in the vicinity (Scheme 22.2). Initial studies utilized one-armed mustard, 2chloroethyl ethyl sulide (CEES), which is a close sur-

60

[PMPA-H]−

Ion intensity

40 20 0 40

60

80

100

120

140

30

160 ×3

180

200

160

180

200

20 −C6H12

10 0 40

60

80

100

120

140

m/z

FIGURE 22.4 Mass spectra of pinacolyl methylphosphonic acid (PMPA) on soil particles, at a concentration of 0.02 molecular layer. Top: MS-1; bottom: MS-2, following isolation and fragmentation of the conjugate base at m/z 179.

SCHEME 22.2 Degradation reactions of mustard occurring in the environment. Elimination of Cl− produced the initially formed sulfonium ion (A), which can undergo hydrolysis to form the hemi alcohol and then thiodiglycol, or can condense with a second H molecule to form sulfonium ion (B).

ANALYSIS OF BLISTER AGENTS

rogate to mustard but is signiicantly less toxic, and did not require the support of a surety facility. Using CEES as a surrogate, both qualitative and quantitative detectability were evaluated, as was detection of sulfonium ions on soil particles. Subsequent studies demonstrated the detection of mustard in collaboration with Dugway Proving Ground. 22.4.2

Protocol for Analysis of Blister Agents

1. Sample Preparation. Soil samples having known amounts of surrogate or agent compounds were prepared by irst weighing a quantity of a soil having a known speciic surface area (measured using the Brunauer, Emmet, Teller [BET] method) [12], which enabled the calculation of the total surface area of the soil sample Asoil. For most of the soil samples studied, the speciic surface areas were on the order of 2–60 m2/g. The soil sample was then exposed to a speciic volume of a solution having a known concentration of surrogate or agent compounds, and the solvent was allowed to evaporate. The area occupied by a single adsorbate molecule was not known but was estimated as a circular area with a radius equal to one-half the distance estimated for the length of the molecule in its most distended conformation on the surface, assuming 1 Å per molecular bond. For example, in the case of CEES, the length was estimated as 6 Å, and thus the molecular surface area is estimated at 30 Å2. Multiplying the total number of adsorbate molecules by the estimated molecular surface area provided Aadsorbate, the total area occupied by the adsorbate molecules assuming uniform nonoverlapping distribution on the surface. The ratio of these two quantities (Aadsorbate/Asoil) is then the number of molecular layers present on the soil surface, assuming the surface area of the container was negligible compared with the soil (usually a good assumption). 2. Sample Mounting. MS-1 and MS-2 were executed as described for the AMPAs above, with the excep-

497

tion that the mass spectrometer was operated in the positive-ion mode. 3. MS-3 analyses were performed by adding a second isolation/fragmentation sequence to the overall analysis (see the example below for CEES, in which m/z 89 → m/z 61 → MS-3 production ions). Thus, the overall temporal sequence was: a. bombardment/ionization b. isolation (e.g., of m/z 89) c. fragmentation (m/z 89 → m/z 61, 55) d. isolation (e.g., m/z 61) e. fragmentation (m/z 61 → m/z 35) f. scan out and detection. The second isolation/fragmentation periods were performed in a manner analogous to the initial isolation/ fragmentation periods (see Section 22.3.2, displayed list numbers 6–8). 22.4.3

Results of Surface Analysis of Blister Agents

22.4.3.1 Analysis of CEES as a Surrogate for H CEES is not an acidic compound and hence does not produce an abundant negative-ion mass spectral signature. Positive-ion analysis of samples exposed to CEES in methylene chloride solutions showed that an ionized form of the intact molecule was not detected in the positive-ion MS-1 spectrum; however, an ion at m/z 89 was identiied above background that corresponded to [CEES–Cl]+ (Scheme 22.3) [13]. The m/z 89 ion was not a unique identiier for CEES, as it is derived from a variety of hydrocarbon sources. Note that the m/z 89 ion appears in the mass spectra of both exposed and unexposed soil samples (Figure 22.5). Thus, MS/MS was used for conirmation and to determine the minimum detectable quantity. Isolation and fragmentation of m/z 89 produced loss of C2H4 forming m/z 61, and loss of H2S forming m/z 55 (Figure 22.6A,B), both of which are unique for samples that were exposed to CEES, but were not observed in the MS/MS analysis of samples that had not been exposed. Third-order tandem mass spectrometry (MS3) analysis of the m/z 61 ion further

SCHEME 22.3 Formation of the [CEES–Cl]+ and MS/MS fragmentation reactions.

498

MEASUREMENTS OF SURFACE CONTAMINANTS AND SORBED ORGANICS 140 A.

120 100 80 Ion abundance (counts)

60 40 20 0 50

60

70

80

140

90

100

90

100

m/z

120

B.

100 80

89

60 40 20 0 50

60

70

80 m/z

FIGURE 22.5 MS-1 spectrum of (A) unexposed soil and (B) soil exposed to 0.03 monolayers CEES.

30

A. isolation, m/z 89

20

89

10 0 40 30 20 10 0

80

100

80

100

80

100

80

100

61 55

40 30

60

B. exciation, m/z 89

60

C. isolation, m/z 61

61

20 10 0 40 30 20 10 0

60

D. exciation, m/z 61

61

35

40

60 m/z

FIGURE 22.6 MSn spectra of ions recorded in the positive-ion analysis of soil particles exposed to CEES. (A) MS/MS isolation of m/z 89. Note that m/z 91 appears more intense as a result of high intensity in the MS-1 spectrum. (B) Fragmentation of m/z 89, producing m/z 61 and 55. (C) MS3 isolation of m/z 61. (D) Fragmentation of m/z 61, producing m/z 35.

ANALYSIS OF BLISTER AGENTS

conirmed that the ion was sulfur containing, as it underwent elimination of C2H2 to form H3S+ at m/z 35 (Figure 22.6C,D). This result constituted strong evidence that the ion at m/z 89 was indeed derived from CEES, as was m/z 61, which is commonly associated with organosulide compounds, but not hydrocarbons. The absolute ion intensities of the MS/MS-produced fragment ions at m/z 61 and 55 were related to the molecular surface coverage. In order to make this correlation, the total number of data system counts was normalized versus ionization time; the latter parameter was varied to ensure good S/N across a range of surface concentrations. For the higher concentrations, a few tens of milliseconds were suficient, while for the lowest surface concentrations, several hundred milliseconds of ionization time was needed. Plotting the normalized MS/MS ion signals versus surface concentration showed a strong correlation (Figure 22.7), which was remarkable given the heterogeneity of the soil particles. These results showed that (1) CEES could be detected down to 1 × 10−3 molecular layers; (2) the response, while relatively imprecise on a sample-to-sample basis, was roughly linear; and (3) the response lattened off once the surface concentration exceeded one molecular layer. An ancillary point was that the low detection values were only achieved by combining the direct surface analysis using the perrhenate ion gun with the MS/MS capability of the IT-SIMS. The soil used in this case had a speciic surface area of ∼15 m2/g, and this meant that the absolute quantity of CEES in the analysis was on the order of 10–20 ng, given a sample of ∼2 mg on the end of the probe.

m/z 61 m/z 55

Counts/ms

100

10

1

1E-3

0.01

0.1

Monolayer

FIGURE 22.7 Plot of MS/MS ion intensities (in counts per millisecond ionization time) versus CEES surface concentration, in monolayers.

499

22.4.3.2 Analysis of Sulfonium Ion Conjugates A second interesting application that may have wider applicability was the detection of sulfonium ions formed from the hydrolysis and condensation of mustard-type blister agents [14]. The mustard compounds will undergo hydrolysis to produce thiiranium ions as represented by structure A in Scheme 22.2. These species are relatively aggressive Lewis acids, however, and react readily with water to form irst the hemiglycol and then thiodiglycol, or with adjacent sulides, to form trigonal sulfonium complexes (e.g., ion B). Further condensation reactions are possible and result in larger H polymers that are the source of mustard “skin” noted in many studies. The IT-SIMS was used to evaluate conjugate sulfonium ion detectability by application of a hydrolyzed CEES solution to soil. Hydrolysis was initiated by using methanol and H2O as a solvent. Intact sulfonium samples were detected on the surface using the IT-SIMS that corresponded to structures similar to ion B but having OH and OCH3 substituents in addition to Cl. The sulfonium ions are readily detected because they contain a ixed charge and are fairly surfactant. 22.4.3.3 Analysis of H and HN Agents The studies using CEES provided a basis for examining samples contaminated with mustard, bis(2-chloroethyl)sulide, or H. To evaluate these samples, the experiments had to be conducted at a surety facility (Dugway Proving Ground), and initial studies with mustard were conducted using IT-SIMS derived from a modiied Varian Saturn 2000 IT system that was mounted inside a negative pressure enclosure inside a mobile laboratory [15]. The small footprint and power requirements of the ITSIMS are highly compatible with analysis in this type of nonlaboratory setting. The analysis of soil samples contaminated with ∼0.5 molecular layer of H did not display ions characteristic of an intact agent (Figure 22.8A), but as in the case of CEES, [H–Cl]+ was observed at m/z 123 and 125 (containing the 35Cl and 37Cl isotopes, respectively). The spectra displayed signiicant chemical background, such that the identiication even at this relatively high surface concentration was somewhat equivocal. Isolation of the [H–Cl]+ ion envelope followed by excitation produced signiicant fragment ions at m/z 95, 63, and 61, which corresponded to the elimination of C2H4, C2H4S, and C2H3Cl, respectively (Scheme 22.4). The loss of C2H4 was not unique compared with the fragmentation of the background ion at m/z 123; however, the elimination of both C2H4S and C2H3Cl were unequivocal in identifying ion signatures. Detection down to 0.07 molecular layers was demonstrated using MS/MS. Interestingly, the analysis of headspace above these samples using GC/ MS did not indicate the presence of mustard, indicating

500

MEASUREMENTS OF SURFACE CONTAMINANTS AND SORBED ORGANICS

1.0 × 103

A.

123 125

Ion abundance (counts/s)

5.0 × 102 0.0 50 3.0 × 10

60

70

80

90

100

110

3

120

130

123

B.

2.0 × 103

125

1.0 × 103 0.0 50

2.0 × 103

60

70

80

90

100

110

120

125

C. 1.0 × 10

130

61

3

63

95

0.0 50

60

70

80

90

100

110

120

130

m/s

FIGURE 22.8 Mass spectra of soil samples exposed to H. (A) MS-1 spectrum, 0.5 molecular layer. (B) MS/MS isolation of the ions in the vicinity of m/z 123, [H–Cl]+. (C) MS/MS fragmentation of m/z 123.

for use as a military agent, and interestingly, these compounds have been evaluated for use as anticancer compounds. In contrast to the sulfur mustard compounds, intact, protonated molecular ions were generated in the IT-SIMS analysis for all three of the nitrogen mustards. The minimum detection limit (in terms of molecular layers) for these compounds was lower than for the sulfur mustards, which is indicative of the fact that the HN compounds are basic and form ionic species much more eficiently. SCHEME 22.4 Ion fragmentation observed in the analyses of [H–Cl]+ formed in the IT-SIMS from H adsorbed to soil particles.

that the majority of H will partition onto surfaces and, when trace quantities are present, is best detected from surfaces. The IT-SIMS was also applied for the characterization of soil samples contaminated with nitrogen mustard compounds [16]. These are much less common than sulfur mustards and are comprised by three compounds: tris(2-chloroethyl)amine (HN-3), bis(2-chloroethyl)methylamine (HN-2), and bis(2-chloroethyl)ethylamine (HN1). Of these, only HN-3 received serious consideration

22.5 ANALYSIS OF NERVE AGENTS 22.5.1

Background for the Analysis of Nerve Agents

The common nerve agents that are encountered in chemical warfare agent-illed munitions are isopropyl methylphosphonoluoridate (GB or sarin), pinacolyl methylphosphonoluoridate (GD or soman), and O-ethyl, S-(N,N-diisopropyl)aminoethyl methylphosphonothiolate (VX). These compounds are acutely toxic, with median lethal dose (LD50) values on the order of tens of micrograms per kilogram for humans, which make detection of trace quantities extremely important, and also render them dificult to handle in conducting analytical experiments. Given this limitation, the small size and ease of operation are favorable attributes of the

ANALYSIS OF NERVE AGENTS

IT-SIMS that have enabled application to the analysis of nerve agent-contaminated surfaces. Speciically, the bread box-sized IT-SIMS shown in Figure 22.2 is small enough to be situated in a surety laboratory, which is exactly what was done to enable the characterization of VX on soil surfaces. VX is very readily detected using the IT-SIMS and is in sharp contrast to GB and GD, which are not easily detected. This is principally because the G-agents were too volatile to effectively adsorb to the surfaces of sample materials, and hence, were lost in the evacuation of the sample in the direct insertion probe forelock. VX, on the other hand, aggressively adsorbed to sample surfaces; this, together with its high proton afinity, resulted in extremely sensitive detection. 22.5.2

Protocol for the Analysis of Nerve Agents

22.5.2.1 Instrumental Setup The IT-SIMS based on either Teledyne or Varian designs was situated in a surety hood for use in analysis. This setup obviated the need for moving contaminated samples outside of the hood for analysis, and therefore, sample loading and analysis procedures were conducted within the hood. Electronic cables were run from the instrument to the power supplies, which were located on a cart that was adjacent to the hood. The pumping hose between the turbo pump on the IT-SIMS and the rotary pump also ran outside the hood; however, the rotary pump was vented into the hood. This arrangement contained all vapors generated during the nerve agent experiments. 22.5.2.2 Sample Generation Analytical standards of VX and GB (approximately 0.1% in isopropanol) were diluted to the desired concentration and then applied to soil samples that had previously been characterized in terms of speciic surface area. The volume of standard was roughly equivalent to the volume of solid sample in most cases. Residual isopropanol was allowed to evaporate, at which time particles were afixed to a sample holder (again the #18 inishing nail head) using double-sided tape. The sample holder was itted into the end of the direct insertion probe and locked into place using a set screw. All these manipulations were performed inside the surety hood. 22.5.2.3 IT-SIMS Analysis The loaded sample probe was evacuated in the insertion probe forelock to about 100 mTorr, then inserted into the IT-SIMS vacuum chamber. This procedure caused a temporary increase in pressure to about 10−4 Torr, which is followed by a decrease to ∼10−5 Torr, which is where the instrument is operated during analysis. Background pressure for the IT-SIMS used in these experiments was on the order of

501

5 × 10−7 Torr, and for analysis, He buffer gas was added such that the total pressure was 1 × 10−5 Torr. Ion generation was initiated by bombarding the sample for typically tens to a few hundreds of milliseconds, during which time the ReO4− primary ion beam was directed through the IT and onto the sample surface. Typical primary ion currents were 50 pA, and the total primary ion dose was kept 9.5) [24,28,29] or standard diethyl-p-phenylenediamine (DPD) titration [30]. Standards containing monochloramine and dichloramine can be prepared by slowly pouring a free chlorine solution over an ammonium chloride solution at a chlorine-to-ammonia molar ratio of 1.31:1.00 under rapid stirring. One should wait for 1 h for the reactions to complete in darkness [3]. Standards of trichloramine can be prepared similarly but at a chlorine-to-ammonia molar ratio of 3.15:1.00. The pH of the standard solution can be adjusted to that of the sample solution by phosphate buffers. The standard solutions can be standardized by DPD titration [30]. 27.2.2

Identiication and Quantiication of DBPs

The application of MIMS to the measurement of DBPs in swimming pools has recently attracted considerable attention. Because of the typical high organic load found in swimming pools from sunscreens, perspiration, urine, and other sources, the chlorination of pool water often generates high concentrations of DBPs [31]. The volatile DBPs are particularly troublesome because they can be taken up by swimmers through respiration and skin absorption [32]. The high DBP concentrations in swimming pools are likely to be well above MIMS detection limits, unlike drinking water applications where much lower concentrations are typical (cf. Table 27.1).

600

DISINFECTANT AND BY-PRODUCT ANALYSIS IN WATER TREATMENT (A)

Abundance

52

44 CO2

30,000

HO35CI

24,000 18,000 HO37CI

12,000

35

6000 m/z

0

36 30

45 40

50 60 51 NH235CI

(B) 25,000 Abundance

Cl2 70 37 Cl2 66

119

70

80

90

100

110

120

130

70

80

90

100

110

120

130

110

120

130

20,000 15,000 10,000 5000

NH237CI

44

36 38

0 m/z

30

40

50 60 49 N35CI

(C) 30,000 Abundance

24,000 18,000 12,000

44

m/z

Abundance

(D)

87

36

6000 0

NH35Cl2 85

N37CI 51

30

40

50

70 70

60

NH35Cl37CI

NH37Cl2 80

90

100

84 60,000

N35Cl2 N35Cl37CI

45,000 30,000 49

N37Cl2 88

15,000 36 m/z

0

30

44 40

51 66 50

60

70 70

80

90

100

109 110

N35Cl3 119 123 120 130

FIGURE 27.5 Representative EI mass spectra (35 ≤ m/z ≤ 125) of (A) free chlorine (2000 mg/L as Cl2), (B) monochloramine (40 mg/L as Cl2), (C) dichloramine (20 mg/L as Cl2), and (D) trichloramine (20 mg/L as Cl2). Reprinted from Shang and Blatchley [3] with the permission of the American Chemical Society.

The identiication of DBPs can be challenging because the volatile compounds form a mixture after they pass the semipermeable silicone membrane in MIMS. The challenge is somewhat alleviated for halogenated DBPs by the isotopic signatures of chlorine and bromine. The halogenated compounds that have been identiied as DBPs in the literature are summarized in Table 27.1. They include several trihalomethanes (THMs) such as chloroform (CHCl3), dichlorobromomethane (CHBrCl2), dibromochloromethane (CHBr2Cl), and bromoform

(CHBr3). Other DBPs are dichloroacetonitrile (Cl2CHC≡N), N,N-dichloroaminoacetonitrile (Cl2N– CH2C≡N), cyanogen chloride (CNCl), cyanogen bromide (CNBr), and dichloromethylamine (CH3NCl2). Like chloramines, their signature m/z peaks also obey the isotopic ratios of [35Cl] : [37Cl] ≈ 3:1 and [79Br] : [81Br] ≈ 1:1. The distinctive MIMS peaks of volatile disinfection products are summarized in Table 27.1. Even for chlorinated compounds, their m/z signatures can still overlap when more than one of them is

APPLICATIONS

present in a sample. This is particularly problematic in quantiication. For example, CHBr2Cl has the most abundant response at m/z 127. However, CHBrCl2 can also contribute to this peak. When both of them are present, the contribution from CHBrCl2 has to be subtracted from the total abundance at m/z 127: 79

35

I m / z 127 (CH Br Cl

•+

from CHBrCl 2 ) =

I m / z 127 (total) − I m / z 127 (CH 79 Br 35Cl •+

from CHBr2Cl), (27.19)

I m / z 127 (CH 79 Br 35Cl •+ 35

35

χI m / z 83 (CH Cl Cl

from CHBr2Cl) = •+

from CHBrCl 2 ),

(27.20)

where χ is the abundance ratio of the peaks at m/z 127 and 83 measured with CHBr2Cl standards. According to Table 27.1, CHCl3 can also contribute to m/z 83. If CHCl3 is in solution, then its contribution must also be considered: I m / z 83 (CH 35Cl 35Cl •+

from CHBrCl 2 ) =

I m / z 83 (total) − I m / z 83 (CH 35Cl 35Cl •+

from CHCl 3 ). (27.21)

With the addition of even more volatile components in a water sample, the deconvolution of their contributions to speciic MIMS peaks can quickly become unmanageable. The application of tandem mass spectrometry (MS/MS) can be helpful for cases where two different fragments from two different sources give the same m/z value within uncertainty. Like the case mentioned above, MS/MS cannot distinguish the sources by further fragmenting the ions because different sources give the same ion at the speciied m/z. One possible solution to such a problem may be to place a short GC column between the semipermeable membrane and the MS ionization chamber, which can be readily envisioned for the external MIMS coniguration. We are, however, not aware of any study that has incorporated chromatography into MIMS. Nevertheless, the complexity of the volatile composition of a sample should be checked on before MIMS is considered. Another advantage of MIMS is that it does not require sample preparation and can be operated in a low-through mode; therefore, MIMS can be employed for automatic, on-line monitoring. A recent study that continuously monitored THMs in a Danish pool has shown that the evaporation of volatile DBPs was greatly enhanced by the activities of the swimmers when the pool was open [2]. Such long-term monitoring data are expected to provide valuable insights for improving the

601

management of water treatment in swimming pools so that the adverse health effects of swimming in chlorinated pools can be minimized. 27.2.3

Formation and Stability of DBPs

MIMS has been shown to be suitable for studying the kinetics of DBP formation and decay reactions that have half-lives greater than the mass-transfer delay (typically a few minutes) [16,17]. A traditional technique for studying disinfection reactions is UV-vis spectrophotometry [33]; however, not all reactants and products have signiicant light absorption. Another option to study kinetics besides direct measurement is quenching the reaction; however, this method only works when the reaction can be stopped by either a rapid change in the experimental condition or the addition of chemicals to exhaust a reactant. Compared with spectrophotometry and quenching, MIMS has the advantage of being applicable to most volatile compounds and requiring minimum sample preparation. To perform kinetic studies in the laboratory, the MIMS should be connected with a reactor and set up in a continuous low-through mode (i.e., in-line MIMS). An example of such a coniguration is shown in Figure 27.6 [16]. Because the analytes are volatile, void space should be minimized in the setup. One strategy, as illustrated in Figure 27.6, is to direct the solution back to the reactor after MIMS measurement. If this strategy is used, the reactor volume should be large enough to ensure that the loss of volatile components in the MIMS will not affect reaction progress.

0RWRU

3XPS 0HPEUDQH ,QWURGXFWLRQ 3UREH

0DVV 6SHFWURPHWHU

,VRWHPS &LUFXODWRU 5HDFWRU

FIGURE 27.6 An example of in-line MIMS system used in kinetic studies. This system consists of a constant temperature and well-mixed batch reactor, a high performance liquid chromatography (HPLC) piston pump (Aculow Series I, Lab Alliance Inc., State College, PA), a direct membrane inlet probe (MIMS Technology), and a Saturn 4D ion trap (Varian) or Hewlett-Packet 5972 quadrupole mass spectrometer (MS) (Agilent Technologies, Santa Clara, CA). The MS operated in the electron impact (EI) mode. Reprinted from Na and Olson [16] with the permission of the American Chemical Society.

602

DISINFECTANT AND BY-PRODUCT ANALYSIS IN WATER TREATMENT

Depending on the half-life of the reaction, one may decide whether the mass-transfer delay should be included in data analysis. For the MIMS system that has generated the CNCl spectrum in Figure 27.4, the total delay is approximately 2 min. If a 5% measurement discrepancy is assumed to occur in this period, then a pseudo-irst-order reaction with a half-life greater than 27 min would not require consideration of the delay (i.e., t1/2 ≥ 2 min × [ln(0.5)/ln(0.95)]). This is the case for CNCl decay in the presence of free chlorine, which has a half-life of ca. 60 min [16]. The kinetic data must be analyzed following the procedure outlined in Section 27.2.4 for reactions with a half-life close to or shorter than ca. 30 min. This is the case for the formation of CNCl from the chlorination of glycine, a prevalent nitrogenous compound in natural water [17]. Extensive calibration is required to obtain important parameters such as Dm, td, and σ. Because a mass spectrometer is a universal detector, MIMS offers the advantage that unexpected, short-lived reaction intermediates can often be spectrally identiied during kinetic studies. N-chloromethylimine (ClN=CH2) is such an example intermediate that was found to play an important role in the formation of the DBP CNCl due to the reaction of chlorine and glycine [17] (cf. Table 27.1). The direct observation and identiication of ClN=CH2 by MIMS greatly simpliied the process of elucidating the reaction pathway for CNCl formation.

REFERENCES 1. Johnson, R.C., Cooks, R.G., Allen, T.M., Cisper, M.E., Hemberger, P.H. (2000) Membrane introduction mass spectrometry: trends and applications. Mass Spectrometry Reviews, 19, 1–37. 2. Kristensen, G.H., Klausen, M.M., Hansen, V.A., Lauritsen, F.R. (2010) On-line monitoring of the dynamics of trihalomethane concentrations in a warm public swimming pool using an unsupervised membrane inlet mass spectrometry system with off-site real-time surveillance. Rapid Communications in Mass Spectrometry, 24, 30–34. 3. Shang, C., Blatchley, E.R. (1999) Differentiation and quatiication of free chlorine and inorganic chloramines in aqueous solution by MIMS. Environmental Science & Technology, 33, 2218–2223. 4. Lee, W., Westerhoff, P., Yang, X., Shang, C. (2007) Comparison of colorimetric and membrane introduction mass spectrometry techniques for chloramine analysis. Water Research, 41, 3097–3102. 5. Hoch, G., Kok, B. (1963) A mass spectrometer inlet system for sampling gases dissolved in liquid phases. Archives of Biochemistry and Biophysics, 101, 160–170. 6. Haas, C.N. (1999) Disinfection. In Water Quality and Treatment, edited by Letterman, R.D. New York: McGraw Hill.

7. USEPA. (2003) National primary drinking water standards. EPA 816-F-03-016. 8. Richardson, S.D. (2008) Environmental mass spectrometry: emerging contaminants and current issues. Analytical Chemistry, 80, 4373–4402. 9. Riter, L.S., Peng, Y.A., Noll, R.J., Patterson, G.E., Aggerholm, T., Cooks, R.G. (2002) Analytical performance of a miniature cylindrical ion trap mass spectrometer. Analytical Chemistry, 74, 6154–6162. 10. Ketola, R.A., Kotiaho, T., Cisper, M.E., Allen, T.M. (2002) Environmental applications of membrane introduction mass spectrometry. Journal of Mass Spectrometry, 37, 457–476. 11. Bier, M.E., Cooks, R.G. (1987) Membrane interface for selective introduction of volatile compounds directly into the ionization chamber of a mass spectrometer. Analytical Chemistry, 59, 597–601. 12. Kotiaho, T., Lauritsen, F.R., Choudhury, T.K., Cooks, R.G., Tsao, G.T. (1991) Membrane introduction massspectrometry. Analytical Chemistry, 63, 1794–1801. 13. Lopez-Avila, V., Benedicto, J., Prest, H., Bauer, S. (1999) Automated MIMS for direct analysis of organic compounds in water. American Laboratory, 31, 32. 14. Geinoz, S., Rey, S., Boss, G., Bunge, A.L., Guy, R.H., Carrupt, P.A., Reist, M., Testa, B. (2002) Quantitative structure-permeation relationships for solute transport across silicone membranes. Pharmaceutical Research, 19, 1622–1629. 15. Overney, F.L., Enke, C.G. (1996) A mathematical study of sample modulation at a membrane inlet mass spectrometer—potential application in analysis of mixtures. Journal of the American Society for Mass Spectrometry, 7, 93–100. 16. Na, C., Olson, T.M. (2004) Stability of cyanogen chloride in the presence of free chlorine and monochloramine. Environmental Science & Technology, 38, 6037–6043. 17. Na, C., Olson, T.M. (2006) Formation of cyanogen chloride from glycine in chlorination. Environmental Science & Technology, 40, 1469–1477. 18. Sysoev, A.A., Ketola, R.A., Mattila, I., Tarkiainen, V., Kotiaho, T. (2001) Application of the numerical model describing analyte permeation through hollow iber membranes into vacuum for determination of permeation parameters of organic compounds in a silicone membrane. International Journal of Mass Spectrometry, 212, 205–217. 19. Jaeger, J.C. (1940) Radial heat low in circular cylinders with a general boundary condition. Journal and Proceedings of Royal Society of New South Wales, 74, 342–352. 20. Westerhoff, P., Mash, H. (2002) Dissolved organic nitrogen in drinking water supplies: a review. Journal of Water Supply Research and Technology-Aqua, 51, 415–448. 21. Connell, G.F., Routt, J.C., Macler, B., Andrews, R.C., Chen, J.M., Chowdhury, Z.K., Crozes, G.F., Finch, G.B., Hoehn, R.C., Jacangelo, J.G., Penkal, A., Schaeffer, G.R., Schulz, C.R., Uza, M.P. (2000) Committee report: disinfection at

REFERENCES

22.

23.

24.

25.

26.

27.

28.

29.

30.

large and medium-size systems. Journal American Water Works Association, 92, 32–43. Connell, G.F., Routt, J.C., Macler, B., Andrews, R.C., Chen, J.M., Chowdhury, Z.K., Crozes, G.F., Finch, G.B., Hoehn, R.C., Jacangelo, J.G., Penkal, A., Schaeffer, G.R., Schulz, C.R., Uza, M.P. (2000) Committee report: disinfection at small systems. Journal American Water Works Association, 92, 24–31. Haas, C.N., Jacangelo, J.G., Bishop, M.M., Cameron, C.D., Chowdhury, Z.K., Connell, G.F., Doty, G.A., Finch, G.R., Gates, D.J., Greenberg, A.E., Hoehn, R.C., Huebner, W.B., Jensen, J.N., Lange, A.L., Long, B.W., Moyer, N.P., Nagel, W.H., Noran, P.F., Palin, A.T., Regli, S.E., Routt, J.C., Symons, J.M., Thompson, C.K., Voyles, C.F. (1992) Survey of water utility disinfection practices. Journal American Water Works Association, 84, 121–128. Morris, J.C. (1966) Acid ionization constant of HoCl from 5 to 35 degrees. Journal of Physical Chemistry, 70, 3798–3805. Vikesland, P.J., Ozekin, K., Valentine, R.L. (2001) Monochloramine decay in model and distribution system waters. Water Research, 35, 1766–1776. Donnermair, M.M., Blatchley, E.R. (2003) Disinfection eficacy of organic chloramines. Water Research, 37, 1557–1570. McLafferty, F.W., Turecek, F. (1993) Interpretation of Mass Spectra, 4th ed. Sausalito, CA: University Science Books. Galal-Gorchev, H., Morris, J.C. (1965) Formation and stability of bromamide bromimide and nitrogen tribromide in aqueous solution. Inorganic Chemistry, 4, 899–905. Friedman, H.L. (1953) On the ultraviolet absorption spectra of uninegative ions. Journal of Chemical Physics, 21, 318–322. American Public Health Association, American Water Works Association, Water Pollution Control Federation, and Water Environment Federation. (1995) Standard Methods for the Examination of Water and Wastewater, 19th ed., Vol. (4500-Cl F and G for free chlorine, 4500-CND for KCN, and 4500-CN-J for CNCl). New York [etc.]: American Public Health Association.

603

31. Zwiener, C., Richardson, S.D., De Marini, D.M., Grummt, T., Glauner, T., Frimmel, F.H. (2007) Drowning in disinfection byproducts? Assessing swimming pool water. Environmental Science & Technology, 41, 363–372. 32. Weisel, C.P., Richardson, S.D., Nemery, B., Aggazzotti, G., Baraldi, E., Blatchley, E.R., Blount, B.C., Carlsen, K.H., Eggleston, P.A., Frimmel, F.H., Goodman, M., Gordon, G., Grinshpun, S.A., Heederik, D., Kogevinas, M., LaKind, J.S., Nieuwenhuijsen, M.J., Piper, F.C., Sattar, S.A. (2009) Childhood asthma and environmental exposures at swimming pools: state of the science and research recommendations. Environmental Health Perspectives, 117, 500–507. 33. Margerum, D.W., Gray, E.T., Huffman, R.P. (1978) Chlorination and the formation of N-Chloro compounds in water treatment. In Organometals and Organometalloids. ACS Symposium Series, 82, 278–291. 34. Li, J., Blatchley, E.R. (2007). Volatile disinfection byproduct formation resulting from chlorination of organicnitrogen precursors in swimming pools. Environmental Science & Technology, 41, 6732–6739. 35. Yang, X., Shang, C., Westerhoff, P. (2007) Factors affecting formation of haloacetonitriles, haloketones, chloropicrin and cyanogen halides during chloramination. Water Research, 41, 1193–1200. 36. Weaver, W.A., Li, J., Wen, Y.L., Johnston, J., Blatchley, M.R., Blatchley, E.R. (2009) Volatile disinfection byproduct analysis from chlorinated indoor swimming pools. Water Research, 43(13), 3308–3318. 37. Shang, C., Gong, W.L., Blatchley, E.R. (2000) Breakpoint chemistry and volatile byproduct formation resulting from chlorination of model organic-N compounds. Environmental Science & Technology, 34(9), 1721–1728. 38. Kristensen, G.H., Klausen, M.M., Hansen, V.A., Lauritsen, F.R. (2010) On-line monitoring of the dynamics of trihalomethane concentrations in a warm public swimming pool using an unsupervised membrane inlet mass spectrometry system with off-site real-time surveillance. Rapid Communications in Mass Spectrometry, 24(1), 30–34. 39. Yang, X., Shang, C. (2005) Quantiication of aqueous cyanogen chloride and cyanogen bromide in environmental samples by MIMS. Water Research, 39(9), 1709–1718.

28 PROTON TRANSFER REACTION MASS SPECTROMETRY (PTR-MS) Yujie Wang, Chengyin Shen, Jianquan Li, Haihe Jiang, and Yannan Chu

28.1

INTRODUCTION

Proton transfer reaction mass spectrometry (PTR-MS) was irst developed at the Institute of Ion Physics of Innsbruck University in the 1990s. Nowadays, PTR-MS is a well-developed and commercially available technique for the on-line monitoring of trace volatile organic compounds (VOCs) down to parts per trillion by volume (ppt) level. PTR-MS has some advantages such as rapid response, soft chemical ionization (CI), absolute quantiication, and high sensitivity. In general, a standard PTR-MS instrument consists of external ion source, drift tube, and mass analysis detection system. Figure 28.1 illustrates the basic composition of the PTR-MS instrument constructed in our laboratory using a quadrupole mass spectrometer as the detection system.

is measured. If a chromatographic separation method is not used prior to MS, then the resulting mass spectra from EI may be so complicated that identiication and quantiication of the compounds can be very dificult. In PTR-MS instrument, the hollow cathode discharge is served as a typical ion source [1], although plane electrode direct current discharge [2] and radioactive ionization sources [3] recently have been reported. All of the ion sources are used to generate clean and intense primary reagent ions like H3O+. Water vapor is a regular gas in the hollow cathode discharge where H2O molecule can be ionized according to the following ways [4]: e + H 2O → H 2 + + O + 2e,

(28.1)

e + H 2O → H + + OH + 2e,

(28.2)

e + H 2O → O + H 2 + 2e,

(28.3)

e + H 2O → H 2O+ + 2e.

(28.4)

+

28.1.1

Ion Source

Perhaps the most remarkable feature of PTR-MS is the special CI mode through well-controlled proton transfer reaction, in which the neutral molecule M may be converted to a nearly unique protonated molecular ion MH+. This ionization mode is completely different from traditional mass spectrometry (MS) where electron impact (EI) with energy of 70 eV is often used to ionize chemicals like VOCs. Although the EI source has been widely used with the commercial MS instruments most coupled with a variety of chromatography techniques, these MS platforms have a major deiciency: in the course of ionization, the molecule will be dissociated to many fragment ions. This extensive fragmentation may result in complex mass spectra especially when a mixture

The above ions are injected into a short source drift region and further react with H2O ultimately leading to the formation of H3O+ via ion–molecule reactions: H 2 + + H 2O → H 2O + + H 2 → H3O+ + H

(28.5a) (28.5b)

H + + H 2O → H 2O + + H

(28.6)

O + H 2O → H 2O + O

(28.7)

+

+

OH + + H 2O → H3O+ + O → H 2O+ + OH

(28.8a) (28.8b)

H 2O+ + H 2O → H3O+ + OH

(28.9)

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

605

606

PROTON TRANSFER REACTION MASS SPECTROMETRY (PTR-MS)

SD

HC

quadrupole mass filter

SEM Water out

Water in

IC sample inlet turbo pump

ion source

drift tube

turbo pump ion detection system

FIGURE 28.1 Schematic diagram of the PTR-MS instrument that contains a hollow cathode (HC), a source drift (SD) region, an intermediate chamber (IC), and a secondary electron multiplier (SEM).

Unfortunately, the water vapor in the source drift region inevitably can form a few of cluster ions H3O+(H2O)n via the three-body combination process H3O+ (H 2O)n−1 + H 2O + A → H3O+ (H 2O)n + A (n ≥ 1), (28.10) where A is a third body. In addition, small amounts of NO+ and O2+ ions occurred due to sample air diffusion into the source region from the downstream drift tube. Thus, an inlet of venturi type has been employed on some PTR-MS systems to prevent air from entering the source drift region [5,6]. At last, the H3O+ ions produced in the ion source can have the purity up to >99.5%. Thus, unlike the selected ion low tube mass spectrometry (SIFT-MS) technique [7], the mass ilter for the primary ionic selection is not needed and the H3O+ ions can be directly injected into the drift tube. In some PTR-MS, the ion intensity of H3O+ is available at 106∼107 counts per second on a mass spectrometer installed in the vacuum chamber at the end of the drift tube. Eventually, the limitation of detection of PTR-MS can reach low ppt level. Instead of H3O+, other primary reagent ions, such as NH4+, NO+, and O2+, have been investigated in PTR-MS instrument [8–10]. Because the ion chemistry for these ions is not only proton transfer reaction, the technique is sometimes called CI reaction MS. However, the potential beneits of using these alternative reagents are usually minimal, and to our knowledge, H3O+ is still the dominant reagent ion employed in PTR-MS research [1,6,11,12].

28.1.2

Drift Tube

The drift tube consists of a number of metal rings that are equally separated from each other by insulated rings. Between the adjacent metal rings, a series of resistors is connected. A high voltage power supplier produces a voltage gradient and establishes a homogeneous electric ield along the axis of the ion reaction drift tube. The primary H3O+ ions are extracted into the ion reaction region and can react with analyte M present in the sample air, which through the inlet is added to the upstream of the ion reaction drift tube. According to the values of proton afinity (PA) (see Table 28.1), the reagent ion H3O+ does not react with the main components in air like N2, O2, and CO2. In contrast, the reagent ion can undergo proton transfer reaction with M as long as the PA of M exceeds that of H2O [6]: M + H3O+ → MH + + H 2O.

(28.11)

Thus, the ambient air can be directly introduced to achieve an on-line measurement in the PTR-MS operation. Due to the presence of electric ield, in the reaction region, the ion energy is closely related to the reducedield E/N, where E is the electric ield and N is the number density of gas in the drift tube. In a typical PTR-MS measurement, E/N is required to set to an appropriate value normally in the range of 120∼160 Td (1 Td = 10−17 Vcm2/molecule), which may restrain the formation of the water cluster ions H3O+(H2O)n (n = 1–3) to avoid the ligand switch reaction with analyte M [6]: H3O+ (H 2O)n + M → H3O+ (H 2O)n−1 M + H 2O.

(28.12)

INTRODUCTION

TABLE 28.1

Proton Afinities of Some Compounds

Compound

Helium Neon Argon Oxygen Nitrogen Carbon dioxide Methane Carbon monoxide Ethane Ethylene Water Hydrogen sulide Hydrogen cyanide Formic acid Benzene Propene Methanol Acetaldehyde Ethanol Acetonitrile Acetic acid Toluene Propanal O-xylene Acetone Isoprene Ammonia Aniline

Molecular Formula He Ne Ar O2 N2 CO2 CH4 CO C2H6 C2H4 H2O H2S HCN HCOOH C6H6 C3H6 CH3OH CH3COH C2H5OH CH3CN CH3COOH C7H8 CH3CH2COH C8H10 CH3COCH3 CH2C(CH3) CHCH2 NH3 C6H7N

Molecular Proton Weight Afinity [13] (kJ/mol) 4 20 40 32 28 44 16 28 30 28 18 34 27 46 78 42 32 44 46 41 60 92 58 106 58 68

177.8 198.8 369.2 421 493.8 540.5 543.5 594 596.3 680.5 691 705 712.9 742 750.4 751.6 754.3 768.5 776.4 779.2 783.7 784 786 796 812 826.4

17 93

853.6 882.5

However, a higher reduced-ield E/N can cause the collision-induced dissociation (CID) of the protonated products, thereby complicating the identiication of detected analytes. 28.1.3

Mass Analyzer

At the end of the drift tube, there is an intermediate chamber in which most of the air from the drift tube through a small oriice is pumped away. The ions in the drift tube are extracted and focused by the ion optical lens and inally, in a high vacuum chamber, are detected by a quadrupole mass spectrometer with an ion pulse counting system. The ionic count rates I(H3O+) and I(MH+) are measured in counts per second, which are proportional to the respective densities of these ions. Although quadrupole mass ilter is a traditional analyzer in the current PTR-MS instrument, other MS analyzers have been investigated including time-of-light

607

mass spectrometer (TOF-MS) [14–16], ion trap mass spectrometer (IT-MS) [17], and linear ion trap mass spectrometer (LIT-MS) [18]. These MS techniques have been used to distinguish isomeric/isobaric compounds as discussed in the later section. 28.1.4 Absolute Quantiication Normally, PTR-MS can determine the absolute concentrations of trace VOCs according to well-established ion–molecular reaction kinetics. If trace analyte M reacts with H3O+, then the H3O+ signal does not decline signiicantly and can be deemed to be a constant. Thus, the density of product ions [MH+] at the end of the drift tube is given in Equation 28.13 [6]: [ MH + ] = [H3O+ ]0 (1 − e − k[ M ]t ),

(28.13)

where [H3O+]0 is the density of reagent ions at the end of the drift tube in the absence of analyte M, k is the rate constant of reaction 28.11, and t is the average reaction time the ions spend in the drift tube. In the trace analysis case k[M]t 1014 W/cm2) that induces nonlinear (multiphoton) absorption of the radiation, a process that is virtually independent of the central wavelength of the laser. While this technique could unequivocally detect petroleum from single inclu-

sions by GC-MS, biomarkers that are usually the most telling molecular structures for pinpointing their source could not be detected [102]. Coupling this method of luid inclusion extraction with TOF-SIMS for the detection of biomarkers [103] may be another way of signiicantly enhancing the sensitivity of the laser decrepitation system [102]. 30.4.3 TOF-SIMS TOF-SIMS is a high-resolution (m/Δm ∼5000–10,000) mass spectrometric surface-sensitive technique based on the secondary ions that are emitted from a surface when bombarded with energetic primary ions [104]. This technique has been applied to molecular imaging of Archaea-derived lipids in a microbial mat [105,106], and has now been shown to unambiguously detect sterane and hopane biomarkers in a complex matrix of a crude oil from the Siljan impact crater without the need for gas chromatographic separation [103]. Detection limit tests suggest that hopanes and steranes could be detected in an oil with a similar biomarker concentration from a single luid inclusion covering an area of 10 µm2 [103]. It has recently been demonstrated that single luid inclusions can be opened by ion etching with a C60+ gun inside the TOF-SIMS instrument, while recording in real time the opening of the inclusion and the instrument response to the exposed inclusion contents [107]. Using this technique, it is possible to map oil inclusions in three dimensions using MS response, for example, from the m/z 55 ion (Figure 30.8). It is likely

FIGURE 30.8 Three-dimensional representation of the m/z 55.06 signal strength obtained during high mass resolution depth proiling using time-of-light secondary ion mass spectrometry of a sample region (100 × 100 µm2) containing several inclusions. The z-direction (3.1 µm) of the cube is the depth axis; hence, the top side of the cube is the surface of the thin section, whereas the lower spheres are oil-bearing luid inclusions. Pale gray at the top of the index indicates the strongest signal intensity while dark gray at the bottom of the index is the weakest (from Reference [107]). Reproduced from Geobiology with permission of Wiley-Blackwell.

GEOLOGICAL APPLICATIONS OF OIL INCLUSION ANALYSIS BY GC-MS

that future applications of this technique will yield new evidence on the evolution of life on early Earth when applied to Precambrian oil inclusions, and combinations of this technique with laser ablation of single inclusions followed by chromatographic separation prior to SIMS detection may also be achievable.

30.5 GEOLOGICAL APPLICATIONS OF OIL INCLUSION ANALYSIS BY GC-MS The molecular composition of an oil is a ingerprint of its source, alteration, and thermal history, and this is also true of oil in luid inclusions. Early work demonstrated that oil inclusions contained hydrocarbons [108,109], and by 1988, biomarkers were being detected in inclusions by MS and used to constrain the physiochemical aspects of mineral diagenesis [110,111]. Several studies up to 1998 examined the molecular composition of oil inclusions [33,34,45–48,58,59,62,75,81,85,108–117], and many more have been published subsequently. Space does not permit an exhaustive review of all the post1998 papers. Some highlights from key early papers are presented in the sections below, together with more recent work that particularly demonstrates the breadth of use of MS for analysis. 30.5.1 Composition of Early Oil Charge Compared with Current Charge Reservoir illing histories can be reined by understanding the composition of luid inclusion oils, because these histories often relect earlier oil charge than is now present in a reservoir. Quartz cementation is a slow process that is strongly temperature controlled [118,119], and it may take several million years for oil inclusions within quartz overgrowths to be formed. However, oil inclusions in healed fractures are likely to be sealed and encapsulated much quicker than in overgrowths. Previous geochemical analyses of oil inclusions from current oil columns have provided data that suggest that most oil inclusions are trapped during the early stages of trap illing [33,34,46,62,75,120–124]. This is shown by the usually lower thermal maturities of included oil compared with co-occurring crude oil in the reservoir. This observation is consistent with diagenetic studies showing that the rate of quartz diagenesis, and hence, the rate of inclusion entrapment slows down markedly once oil saturation increases in an oil reservoir [125]. Thus, the oil inclusion record is biased toward the composition of the earlier charge into a reservoir, so geochemical analysis of the oil trapped in inclusions provides a powerful technique for understanding the evolution of petroleum within a reservoir.

661

For example, oil trapped in calcite cements in the Fateh Field, Dubai, was generated from an early-mature source rock prior to the time when the pore system was becoming dominated by oil [48]. Similarly, oil trapped in quartz grains in Jabiru Field, NW Australia, was generated from source rocks of slightly lower maturity than the oil in the reservoir, suggesting charge of progressively more mature oil [59,75]. A more complex oil charge history is recorded by oil inclusions in the mineral cements from Ula Field reservoir sandstones, North Sea [46,85]. Oil trapped in inclusions in potassium feldspar overgrowths, the earliest diagenetic phase to host oil inclusions, has a low maturity and different source characteristics than oil in the Ula Formation reservoirs, which was generated from the Mandal Formation [46,85]. Oil inclusions in later authigenic albite and quartz have a composition intermediate between the potassium feldspar inclusion oil and the accumulated oil in the reservoir, suggesting that they record progressive dilution of the early-migrated oil with oil from the Mandal Formation [46,85]. Another example where inclusion oil has different source characteristics from the oil in the reservoir is at South Pepper, Carnarvon Basin, NW Australia [34,62]. Fluid inclusions in carbonate cements trapped early oil derived from a more calcareous and anoxic source rock than the oil currently in the reservoir, which was derived from the Jurassic Dingo Claystone and contains high amounts of C29 25-norhopane (Figure 30.9). This biomarker is present in lower abundance in the inclusion oil. It was deduced that after trapping of oil in inclusions, the early charge was heavily biodegraded during the Miocene and then the structure was recharged with fresh oil from the Dingo Claystone, which mixed with the biodegraded residue in the reservoir [34,62]. The relatively common inding that paleo-oils represented by luid inclusion oils were generated from geochemically different source rock facies compared with oil currently in the reservoir implies that (1) generation and expulsion, trap illing, leakage, and continued recharge are common in petroleum systems over geological time, such that many crude oils represent “mixtures”; and (2) fresh and recently charged crude oil in a reservoir is not always trapped in large amounts in luid inclusion oils. One criticism sometimes made of luid inclusion oil geochemistry is regarding the validity of interpretation of data from numerous luid inclusions in a rock sample. The crushing methods, especially the offline crushing method (Section 30.3.2), use signiicant amounts (typically 3–15 g) of cleaned mineral grains, such that very large numbers of luid inclusions contribute to the overall inclusion oil signature. If more than one generation of luid inclusion oils are present [49,126], then the geochemical signature is of the mixture. In

662

MASS SPECTROMETRY TECHNIQUES FOR ANALYSIS OF OIL AND GAS TRAPPED IN FLUID INCLUSIONS

FIGURE 30.9 Partial m/z 191 and 205 mass chromatograms for luid inclusions oils and crude oil, South Pepper-1, showing the distribution of hopanes and methylhopanes. Peak assignments deine the stereochemistry at C-22 (S and R); αβ and βα denote 17α(H),21β(H) and 17β(H),21α(H)-hopanes, respectively. Ts = C27 18α(H)-22,29,30-trisnorneohopane; Tm = C27 17α(H)-22,29,30trisnorhopane; * = diahopane and 2α(Me) = methylhopane. Numbers in italics refer to the relative abundance of chromatograms (modiied from Reference [34]).

answer to this criticism, one needs to look no further than typical crude oils in reservoirs, which are often complex mixtures or “incrementers” relecting progressive oil charge over signiicant lengths of time, partial loss, and in-reservoir oil alteration events [94,127]. Despite this, crude oils are routinely analyzed and interpreted. Experience indicates that luid inclusion oils may be simpler than crude oils, probably because they

are trapped over more limited geological time when inclusion trapping processes are operational. 30.5.2

Isotopic Measurements for Oil Correlation

Section 30.3.5 provides details on methodologies for isotopic measurements on luid inclusion oils. In this section, two examples are provided to show the application. At

GEOLOGICAL APPLICATIONS OF OIL INCLUSION ANALYSIS BY GC-MS

663

This insight would have been precluded when looking at molecular evidence alone. 30.5.3 Composition of Oil Charge Prior to Seal or Fault Breach, or Other Loss of Charge

FIGURE 30.10 Carbon isotopic compositions of n-alkanes in the Jabiru luid inclusion oil, production oil, and urea nonadducted fraction of the production oil (modiied from Reference [75]).

Jabiru Field, NW Australia, the good correlation between luid inclusion oil and production oil established by biomarkers was checked using compound-speciic carbon isotopic compositions of n-alkanes [75]. The δ13C results of n-alkanes from the unfractionated luid inclusion oil and the production oil appear to be similar and are all within analytical error (Figure 30.10), indicating that the n-alkanes in both oils are derived from a similar source. This is the best means of comparison between luid inclusion oils and oils in reservoirs where small sample size precludes isolation steps. Improved accuracy can be obtained by fractionation of oil samples using column chromatography and molecular sieving techniques. With luid inclusions, this is rarely possible, but one example where it was achieved was in the Perth Basin, Western Australia [76]. δ13C proiles of n-alkanes were obtained from puriied fractions of two lean luid inclusion oils from Leander Reef-1 and Houtman-1. The Leander Reef-1 luid inclusion oil shares many biomarker similarities with oils from the Dongara and Yardarino oil ields, which have been correlated with the Early Triassic Kockatea Shale, but the heavier isotopic values for the C15–C25 n-alkanes in the Leander Reef-1 luid inclusion oil indicated that it is a mixture, and in fact, the main part of this oil (∼90%) was probably sourced from more terrestrial and isotopically heavier Early Permian rocks [76].

A common observation concerning oil inclusion abundance is that paleo-oil columns are common not only below current oil columns, but also in wells that are completely dry at present. For example, in the Timor Sea region of northern Australia, there was an extensive period of fault-seal breach of oil reservoirs during Late Miocene/Early Pliocene fault reactivation, and this left many paleo-oil columns in the presently water-illed reservoir sections [28,128]. In dry wells, the analysis of luid inclusion oils offers the possibility of understanding a petroleum system without having access to current luids in the reservoir [49,55,129], and potentially then being able to predict where the oil may have leaked to, or where it may be trapped in nonbreached structures. A good example of this application is the dry Octavius-2 well in Vulcan Sub-basin, northern Australia [58]. Octavius-2 luid inclusion oil is pristine and rich in aliphatic hydrocarbons, and has a high thermal maturity. A full range of C5–C37 hydrocarbons, including n-alkanes, isoprenoids, alkylcyclohexanes, hopanes, and aromatic hydrocarbons were detected by the off-line and on-line crushing techniques. Alkylnaphthalene distributions of Octavius-2 luid inclusion oil (Figure 30.11) contain relatively large amounts of the more thermally stable β-substituted isomers (e.g., 2-MN; 2,6- and 2,7-DMN; 2,3,6-TMN), indicating high thermal maturity [58]. Approximate calibrations of alkylnaphthalene ratios suggest that the Octavius-2 luid inclusion oil has a thermal maturity in the range of 1.0%–1.3% vitrinite relectance equivalent, and this is corroborated by the methylphenanthrene ratio and by the alkylbiphenyl maturity ratios. Biomarkers are in very low abundance owing to thermal cracking or dilution with hydrocarbons with nonbiological hydrocarbons skeletons such as n-alkanes. 30.5.4 Composition of Oil Charge Prior to Drilling Contamination A common problem with many oil exploration wells is that they are drilled with oil-based drilling muds, or are contaminated with other anthropogenic substances. These can make evaluation of oils recovered from drill stem tests or by solvent extraction of reservoir lithologies very dificult, due to overprinting of the crude oil. Analysis of oil inclusions can avoid these problems because the luid inclusions are trapped before drilling, so extraneous or drilling contaminants can be removed

664

MASS SPECTROMETRY TECHNIQUES FOR ANALYSIS OF OIL AND GAS TRAPPED IN FLUID INCLUSIONS

FIGURE 30.11 Partial added m/z 128.1 + 142.1 + 156.1 + 170.1 + 184.1 mass chromatogram of the luid inclusion oil from Octavius-2 (3200 m) in the Vulcan Sub-basin, northern Australia, showing the distribution of alkylnaphthalenes. The enlargements show details of the partial m/z 170.1 and 184.1 mass chromatograms over the indicated retention times. EN, ethylnaphthalene (modiied from Reference [58]).

by careful cleanup of the mineral surfaces of the reservoir samples prior to inclusion oil analysis. For example, geochemical evidence for a previously unrecognized light oil or gas-condensate source rock in the Nancar Trough area of the Timor Sea was found by GC-MS analysis of luid inclusion oil from the Ludmilla-1 well [130]. This oil has an unusually high abundance of midchain substituted monomethylalkanes relative to nalkanes (Figure 30.12), and was derived from an unknown source rock deposited in a marine environment under sub-oxic to oxic conditions with limited sulfur content, a low contribution of terrestrial organic matter, and a high contribution of organic matter from bacterial activity, particularly cyanobacteria [130]. This information was only revealed following luid inclusion

analysis, because recovered oils and core extracts from this and nearby wells were heavily contaminated by alkene-based drilling mud additives. 30.5.5

Composition of Precambrian Oils

One of the irst studies involving molecular compositional analysis of oil-bearing luid inclusions was conducted on organic-rich inclusions in quartz crystals from calcite veins in Precambrian metasedimentary rocks from South-West Africa [109]. GC-MS analysis revealed the presence of a range of low- and high molecular weight hydrocarbons, including CH4, C2H6, C3H8, and possibly C4H10, n-alkanes from n-C10 to n-C33 as well as fairly high concentrations of isoprenoids, particularly

GEOLOGICAL APPLICATIONS OF OIL INCLUSION ANALYSIS BY GC-MS

665

FIGURE 30.12 Partial m/z 85 mass chromatogram of the Ludmilla-1 luid inclusion oil, showing the distribution of n-alkanes, monomethylalkanes, and isoprenoids. Compound abbreviations: n-Cxx, n-alkane; i-Cxx, isoprenoid; MUD, methylundecanes; MDD, methyldodecanes; MTD, methyltridecanes; MTeD, methyltetradecanes; MPD, methylpentadecanes; MHD, methylhexadecanes; Pr, pristane (modiied from Reference [130]).

pristane [109]. Although the age of the inclusions was not constrained, the study demonstrated the potential of the GC-MS technique for studying biomarker distributions in very small volumes of oil extracted from rocks of Precambrian age. Case studies on Precambrian inclusion oils [6,57], whose age is constrained by petrographic relationships and metamorphic events, show that their biomarker attributes are consistent with Proterozoic sources [131] and that they are similar in composition to older Paleoproterozioc oils [6] and late Archean rock extracts [132,133]. Oil inclusions from the 2.45 Ga Matinenda Formation, Canada [57], from the 2.1 Ga FA Formation from Franceville Basin, Gabon [66] and from the 1.4 Ga Bessie Creek Formation from the Roper Superbasin, Australia [134,135], show many broadly similar characteristics. The oils are mature to highly mature, nonbiodegraded or only slightly biodegraded, and all contain abundant monomethylalkanes and hopanes, including 2α-methyl hopanes derived from cyanobacteria. These are common characteristics of many, but not all [136], organic-rich Precambrian sedimentary rocks [131,137]. Steranes have also been detected in inclusion oils from the Matinenda Formation [6,57] and the FA Formation [66], and indicate that some Proterozoic hydrocarbons have been derived from a eukaryotic source. Although sterane biomarkers from Archean shale extracts have been described as “probably syngenetic” [132], the steranes reported from luid inclusions are from an oil of normal thermal maturity that was trapped early in the burial history of the host rocks and has remained uncontaminated within sealed inclusion cavities. Notably, land plant biomarkers such as oleanane are absent from Precambrian inclusion oils, indicating that following a thorough and careful cleanup procedure of the sample

prior to crushing Precambrian inclusion oils can provide a genuine record of Precambrian life. To date, these studies have concerned only Proterozoic inclusion oils, the oldest thus far published being the 2.45 Ga Matinenda Formation [6,57]. GC-MS analysis of the oils have provided insights into the composition of the Proterozoic biosphere, particularly before and after the Great Oxidation Event [6] as well as insights into the thermal stability of oil [57]. Currently, there are no published studies on the molecular compositions of Archean oil inclusions and very few on source rock hydrocarbon extracts from rocks of that age [132,133,138]. Bitumen extracts from the late Archean Transvaal Supergroup sedimentary rocks have biomarker distributions that were strikingly similar to those in Proterozoic rocks, suggesting that the biomass responsible for generating younger oils was already well established by about 2.7 Ga [133]. This leaves open the possibility of inding and analyzing intact oil-bearing luid inclusions (“biogeochemical time capsules”) from Archean formations so as to help constrain the earliest development of life on Earth.

30.5.6 Fluid Inclusion Oils in Igneous Rocks and Ore Deposits Hydrocarbons are sometimes trapped in inclusions within igneous intrusions [16,115,117,135] or crystalline basement [139,140], or are associated with ore deposits [110]. For example, hydrocarbons trapped in ore veins from the Aberfoyle tin–tungsten deposit in Tasmania, Australia, include biomarkers that suggest that the mineralizing luid evolved through interaction with sedimentary country rocks [110]. Sphalerite- and

666

MASS SPECTROMETRY TECHNIQUES FOR ANALYSIS OF OIL AND GAS TRAPPED IN FLUID INCLUSIONS

dolomite-hosted oil inclusions in lead–zinc deposits in Canning Basin, Western Australia, are more mature than the host rocks for the ore. The oil inclusions show different source parameters than Ordovician oils and source rocks in this basin, suggesting that deeper, more mature organic-rich strata contributed to the oreforming luids [111]. These data were obtained by a mixture of thermal desorption MS and bulk analysis of luid inclusions using GC-MS. Crystalline basement in southern Norway contains oil inclusions with hydrocarbons and dark carbonaceous solids [139]. Analysis shows C1–C5 gaseous hydrocarbons, and complex hydrocarbon in two quartz veins, and are thought to have been derived from overlying sedimentary rocks by the deep percolation of basinal luids [139]. A ∼1280 Ma with Myr-old dolerite sill within the Mesoproterozoic Roper Group in Roper Superbasin, Australia, contains evidence for at least two episodes of hydrocarbon migration represented, respectively, by solid bitumen with a ketone-rich extract [115], and a mixture of a high-maturity gas condensate and a lower-maturity oil within oil-bearing luid inclusions [16]. The ketone isomers were formed by lash pyrolysis of kerogen during the intrusion of the dolerite sill [115] and represent the irst and oldest phase of hydrocarbon migration. The gas condensate and oil were subsequently trapped as a mixture within the luid inclusions at diagenetic temperatures and pressures, signiicantly after cooling of the sill and likely during the Neoproterozoic reactivation of the Roper Superbasin [16]. The hydrocarbons are biogenic and were likely derived from Proterozoic source rocks such as the directly overlying Velkerri Formation and/or the underlying Barney Creek Formation from the McArthur Group [16].

30.6 RELIABILITY AND CONSTRAINTS ON FLUID INCLUSION OIL ANALYSIS BY GC-MS All the applications discussed above are based on the assumption that the oil extracted from luid inclusions represents a sample of oil that was present in the reservoir at some time in the geological past or, in the case of progressive trapping and/or multiple generations of oil inclusions, represents a mixture of oils generated from different source rocks or from the same source rock but produced over a relatively long period of time. Criticisms are likely on the use of luid inclusion oil geochemistry, especially with regard to the validity of interpretation of data from numerous luid inclusions in a rock sample. Therefore, it is pertinent to make a few comments about the reliability of oil inclusion analysis [61].

First, it is essential that the distribution and attributes of oil inclusions in the reservoir samples of interest are fully understood prior to analysis, and this is achieved by using a variety of petrographic and screening techniques, as described in Section 30.2. Multiple generations of oil inclusions may be discernible using textural, luorescence, and thermometric parameters, and would generally be analyzed together using the off-line method, as there is no practical way of separating very large numbers of grains with oil inclusions having different attributes. This does not always apply to the on-line method, where only ∼50 mg of grains are typically crushed, potentially enabling hand-picking of grains if needed to separate multiple generations. However, quite often different luid inclusion assemblages can be found within the same grain, rendering hand-picking of no use for separating assemblages. Second, it is known that luid inclusion oils are sometimes enriched in polar compounds compared with crude oils from the same reservoir [46,141], although this is not always the case [142]. This is likely due to fractionation of the oil in the reservoir, with preferential adsorption of polar compounds onto mineral surfaces [143–146]. The possibility of carryover of polar compounds highly adsorbed onto grain surfaces into inclusion oils is limited by the extensive cleanup steps advocated and can be checked by outside rinse blanks (see Section 30.3.1) and/or by the use of surrogate extraction standards [50]. Third, no signiicant molecular fractionation caused by preferential trapping or experimental artifacts (e.g., oxidation products) is evident in luid inclusion oils. For example, the C30 pentacyclic hydrocarbons oleanane, lupane, C30 αβ hopane, and C30 βα hopane are physically very similar compounds, so are unlikely to have very different afinities to rock surfaces, yet some inclusion oils contain oleanane and others do not contain oleanane, irrespective of whether associated crude oils contain this angiosperm marker [33,67,129,147]. Such differences are thus not the result of preferential trapping. Lack of any signiicant molecular fractionation caused by trapping is further demonstrated by oil ields with simple charge histories, where the luid inclusion oils look very similar to that of associated crude oils (e.g., Jabiru oil ield, Vulcan Sub-basin [75], McKee oil ield, Taranaki Basin [123,124], Gidgealpa oil ield, Cooper Eromanga basins [148], and Cliff Head oil ield [149]). These examples provide high conidence in the interpretation of samples where the inclusion oils are dissimilar to associated crude oils, leading to deduced multiple source rocks and complex charge histories [33,34,62,129,147]. Contaminants of unknown origin sometimes appear in luid inclusion oils (e.g., C27 cholestane ααα 20R is

REFERENCES

a persistent contributor to low-level luid inclusion oils at present in the CSIRO labs), but these are monitored by the system blanks and can thus be excluded from interpretation if necessary [35,57,66]. Anthracene and methylanthracenes are sometimes reported to occur in unusually high abundances in inclusion oils compared with associated crude oils [33,50], but this only affects a small proportion of inclusion oils ( 2 billion years. Geochimica et Cosmochimica Acta, 72, 844–870. George, S.C., Lisk, M., Eadington, P.J., Quezada, R.A. (1998) Geochemistry of a palaeo-oil column, Octavius 2, Vulcan Sub-basin. In The Sedimentary Basins of Western Australia: Proceedings of Petroleum Exploration Society of Australia Symposium, 2, edited by Purcell, P.G., Purcell, R.R., Perth: PESA, pp. 195–210. George, S.C., Lisk, M., Eadington, P.J., Quezada, R.A., Krieger, F.W., Greenwood, P.F., Wilson, M.A. (1996) Comparison of palaeo oil charges with currently reservoired hydrocarbons using the geochemistry of oilbearing luid inclusions. In Society of Petroleum Engineers, Asia Paciic Oil and Gas Conference, SPE paper 36980, 159–171. Society of Petroleum Engineers, Adelaide, Australia. George, S.C., Volk, H., Ahmed, M. (2004) Oil-bearing luid inclusions: geochemical analysis and geological applications. Acta Petrologica Sinica, 20, 1319–1332. George, S.C., Volk, H., Ahmed, M. (2007) Geochemical analysis techniques and geological applications of oilbearing luid inclusions, with some Australian case studies. Journal of Petroleum Science and Engineering, 57, 119–138. Lisk, M., George, S.C., Summons, R.E., Quezada, R.A., O’Brien, G.W. (1996) Mapping hydrocarbon charge histories: detailed characterisation of the South Pepper oil ield, Carnarvon Basin. Australian Petroleum Production and Exploration Association Journal, 36(1), 445–464. Volk, H., Horsield, B., Mann, U., Suchý, V. (2002) Variability of petroleum inclusions in vein, fossil and vug cements: geochemical study in the Barrandian Basin (Lower Palaeozoic, Czech Republic). Organic Geochemistry, 33, 1319–1341. Volk, H., George, S.C., Kempton, R.H., Liu, K., Ahmed, M., Ambrose, G.J. (2007) Petroleum migration in the Georgina Basin: evidence from the geochemistry of oil inclusions and bitumens. In Proceedings of the Central Australian Basins Symposium (CABS), Special Publication 2, edited by Munson, T.J., Ambrose, G.J. Northern Territory Geological Survey, Alice Springs, Northern Territory, August 16–18, 2005, pp. 282–303. Gardner, P.M., Whitehead, E.V. (1972) Isolation of squalane from a Nigerian petroleum. Geochimica et Cosmochimica Acta, 36, 259–263. Dutkiewicz, A., George, S.C., Mossman, D., Ridley, J., Volk, H. (2007) Oil and its biomarkers associated with the Palaeoproterozoic Oklo natural ission reactors, Gabon. Chemical Geology, 244, 130–154.

67. George, S.C., Lisk, M., Eadington, P.J. (2004) Fluid inclusion evidence for an early, marine-sourced oil charge prior to gas-condensate migration, Bayu-1, Timor Sea, Australia. Marine and Petroleum Geology, 21, 1107–1128. 68. George, S.C., Dutkiewicz, A., Volk, H., Ridley, J., Mossman, D.J., Buick, R. (2009) Oil-bearing luid inclusions from the Palaeoproterozoic: a review of biogeochemical results from time-capsules > 2.0 Ga old. Science in China Series D-Earth Sciences, 52, 1–11. 69. Ahmed, M., George, S.C. (2004) Changes in the molecular composition of crude oils during their preparation for GC and GC-MS analyses. Organic Geochemistry, 35, 137–155. 70. Merritt, D.A., Brand, W.A., Hayes, J.M. (1994) Isotoperatio-monitoring gas chromatography mass spectrometry—methods for isotopic calibration. Organic Geochemistry, 21, 573–583. 71. Hayes, J.M., Freeman, K.H., Popp, B.N., Hoham, C.H. (1990) Compound speciic isotope analysis for reconstruction of ancient biogeochemical processes. Organic Geochemistry, 16, 1115–1128. 72. Boreham, C.J., Summons, R.E., Roksandic, Z., Dowling, L.M., Hutton, A.C. (1994) Chemical, molecular and isotopic differentiation of organic facies in the Tertiary lacustrine Duaringa oil shale deposit, Queensland, Australia. Organic Geochemistry, 21, 685–712. 73. Murray, A.P., Summons, R.E., Boreham, C.J., Dowling, L.M. (1994) Biomarker and n-alkane isotope proiles for Tertiary oils—relationship to source rock depositional setting. Organic Geochemistry, 22, 521–542. 74. Summons, R.E., Freeman, K.H., Grice, K., Pancost, R.D. (2008) Where would we be without the isotopes? Organic Geochemistry, 39, 483–484. 75. George, S.C., Greenwood, P.F., Logan, G.A., Quezada, R.A., Pang, L.S.K., Lisk, M., Krieger, F.W., Eadington, P.J. (1997) Comparison of palaeo oil charges with currently reservoired hydrocarbons using molecular and isotopic analyses of oil bearing luid inclusions: Jabiru oil ield, Timor Sea. Australian Petroleum Production and Exploration Association Journal, 37(1), 490–504. 76. Volk, H., Boreham, C., Kempton, R.H., George, S.C. (2004) Geochemical and compound speciic carbon isotopic characterisation of luid inclusion oils from the offshore Perth Basin (Western Australia): implications for recognising effective oil source rocks. The Australian Petroleum Production and Exploration Association Journal, 44(1), 223–239. 77. Hoering, T.C., Freeman, D.H. (1984) Shape-selective sorption of monomethylalkanes by silicalite, a zeolitic form of silica. Journal of Chromatography, 316, 333–341. 78. West, N., Alexander, R., Kagi, R.I. (1990) The use of silicalite for rapid isolation of branched and cyclic alkane fractions of petroleum. Organic Geochemistry, 15, 499–501.

REFERENCES

79. Jochum, J. (2000) Variscan and post-Variscan lead-zinc mineralization, Rhenish Massif, Germany: evidence for sulide precipitation via thermochemical sulfate reduction. Mineralium Deposita, 35, 451–464. 80. Volk, H., Mann, U., Burde, O., Horsield, B., Suchý, V. (2000) Petroleum inclusions and residual oils: constraints for deciphering petroleum migration. Journal of Geochemical Exploration, 69, 595–599. 81. Ruble, T.E., George, S.C., Lisk, M., Quezada, R.A. (1998) Organic compounds trapped in aqueous luid inclusions. Organic Geochemistry, 29, 195–205. 82. Pironon, J., Barrés, O. (1990) Semi-quantitative FT-IR microanalysis limits: evidence from synthetic hydrocarbon luid inclusions in sylvite. Geochimica et Cosmochimica Acta, 54, 509–518. 83. Dubessy, J., Buschaert, S., Lamb, W., Pironon, J., Thiéry, R. (2001) Methane-bearing aqueous luid inclusions: Raman analysis, thermodynamic modelling and application to petroleum basins. Chemical Geology, 173, 193–205. 84. Salvi, S., Williams-Jones, A.E. (2003) Bulk analysis of volatiles in luid inclusions. In Fluid Inclusions––Analysis and Interpretation, Short Course Series, edited by Samson, I., Anderson, A., Marshall, D. Vancouver: Mineralogical Association of Canada, pp. 247–278. 85. Nedkvitne, T., Karlsen, D.A., Bjorlykke, K., Larter, S.R. (1993) Relationship between reservoir diagenetic evolution and petroleum emplacement in the Ula Field, North Sea. Marine and Petroleum Geology, 10, 255–270. 86. Ohm, S.E., Karlsen, D.A. (2007) Biogenic gas in luid inclusions from sandstones in contact with oil-mature coals. American Association of Petroleum Geologists Bulletin, 91, 715–739.

92.

93.

94.

95.

96.

97.

98.

87. Shi, J.Y., Liu, D.H., Shi, J.X., Wang, B.S., Xiang, M.J. (1994) Evaluation of the petroleum source rock potential of samples from the Shichuan Basin, P.R. China. Organic Geochemistry, 21, 343–350.

99.

88. Gong, S., Peng, P.A., Shuai, Y.H., Dai, J.X., Zhang, W.Z. (2008) Primary migration and secondary alteration of the Upper Paleozoic gas reservoir in Ordos Basin, China— application of luid inclusion gases. Science in China Series D-Earth Sciences, 51, 165–173.

100.

89. Potter, J., Longstaffe, F.J. (2007) A gas-chromatograph, continuous low-isotope ratio mass-spectrometry method for δ13C and δD measurement of complex luid inclusion volatiles: examples from the Khibina alkaline igneous complex, northwest Russia and the south Wales coalields. Chemical Geology, 244, 186–201.

101.

102.

90. Bray, C.J., Spooner, E.T.C., Thomas, A.V. (1991) Fluid inclusion volatile analysis by heated crushing, online gaschromatography—applications to Archean luids. Journal of Geochemical Exploration, 42, 167–193.

103.

91. Jochum, J., Friedrich, G.H., Leythaeuser, D., Littke, R., Ropertz, B. (1995) Hydrocarbon-bearing luid inclusions in calcite-illed horizontal fractures from mature

104.

671

Posidonia Shale (Hils Syncline, NW Germany). Ore Geology Reviews, 9, 363–370. Bennett, B., Larter, S.R. (1997) Partition behaviour of alkylphenols in crude oil/brine systems under subsurface conditions. Geochimica et Cosmochimica Acta, 61, 4393–4402. Lafargue, E., Barker, C. (1988) Effect of water washing on crude oil composition. American Association of Petroleum Geologists Bulletin, 72, 263–276. Larter, S.R., Aplin, A.C. (1995) Reservoir geochemistry: methods, applications and opportunities. In The Geochemistry of Reservoirs, edited by Cubitt, J.M., England, W.A. Bath: Geological Society of London Special Publication, pp. 5–32. George, S.C., Volk, H., Dutkiewicz, A., Ridley, J., Mossman, D. (2006) Oil-bearing luid inclusions: biogeochemical time-capsules for >2.0 billion years. Geochimica et Cosmochimica Acta, 70, A198. Thiéry, R., Pironon, J., Walgenwitz, F., Montel, F. (2000) PIT (Petroleum Inclusion Thermodynamic): a new modeling tool for the characterization of hydrocarbon luid inclusions from volumetric and microthermometric measurements. Journal of Geochemical Exploration, 69, 701–704. Heinrich, C.A., Pettke, T., Halter, W.E., Aigner-Torres, M., Audetat, A., Gunther, D., Hattendorf, B., Bleiner, D., Guillong, M., Horn, I. (2003) Quantitative multi-element analysis of minerals, luid and melt inclusions by laserablation inductively-coupled-plasma mass-spectrometry. Geochimica et Cosmochimica Acta, 67, 3473–3497. Horn, I., von Blanckenburg, F., Schoenberg, R., Steinhoefel, G., Markl, G. (2006) In situ iron isotope ratio determination using UV-femtosecond laser ablation with application to hydrothermal ore formation processes. Geochimica et Cosmochimica Acta, 70, 3677–3688. Greenwood, P.F., George, S.C., Wilson, M.A., Hall, K.J. (1996) A new apparatus for laser micropyrolysis-gas chromatography/mass spectrometry. Journal of Analytical and Applied Pyrolysis, 38, 101–118. Greenwood, P.F., George, S.C., Hall, K. (1998) Applications of laser micropyrolysis gas chromatography mass spectrometry. Organic Geochemistry, 29, 1075–1089. Hode, T., Zebühr, Y., Broman, C. (2006) Towards biomarker analysis of hydrocarbons trapped in individual luid inclusions: irst extraction by ErYAG laser. Planetary and Space Science, 54, 1575–1583. Volk, H., Fuentes, D., Fuerbach, A., Miese, C., Koehler, W., Bärsch, N., Barcikowski, S. (2010) First on-line analysis of petroleum from single inclusion using ultrafast laser ablation. Organic Geochemistry, 41, 74–77. Siljeström, S., Hode, T., Lausmaa, J., Sjovall, P., Toporski, J., Thiel, V. (2009) Detection of organic biomarkers in crude oils using ToF-SIMS. Organic Geochemistry, 40, 135–143. Benninghoven, A. (1994) Chemical analysis of inorganic and organic surfaces and thin ilms by static time of lightsecondary ion mass spectrometry (ToF-SIMS).

672

105.

106.

107.

108.

109.

110.

111.

112.

113.

114.

115. 116.

117.

MASS SPECTROMETRY TECHNIQUES FOR ANALYSIS OF OIL AND GAS TRAPPED IN FLUID INCLUSIONS

Angewandte Chemie International Edition, England, 33, 1023–1042. Thiel, V., Heim, C., Arp, G., Hahmann, U., Sjövall, P., Lausmaa, J. (2007) Biomarkers at the microscopic range: ToF-SIMS molecular imaging of Archaea-derived lipids in a microbial mat. Geobiology, 5, 413–421. Thiel, V., Toporski, J., Schumann, G., Sjövall, P., Lausmaa, J. (2007) Analysis of archaeal core ether lipids using time of light-secondary ion mass spectrometry (ToF-SIMS): exploring a new prospect for the study of biomarkers in geobiology. Geobiology, 5, 75–83. Siljeström, S., Lausmaa, J., Sjövall, P., Broman, C., Thiel, V., Hode, T. (2010) Analysis of hopanes and steranes in single oil-bearing luid inclusions using time-of-light secondary ion mass spectrometry (ToF-SIMS). Geobiology, 8, 37–44. Murray, R.C. (1957) Hydrocarbon luid inclusions in quartz. American Association of Petroleum Geologists Bulletin, 41, 950–956. Kvenvolden, K.A., Roedder, E. (1971) Fluid inclusions in quartz crystals from South-West Africa. Geochimica et Cosmochimica Acta, 35, 1209–1229. Hoffmann, C.F., Henley, R.W., Higgins, N.C., Solomon, M., Summons, R.E. (1988) Biogenic hydrocarbons in luid inclusions from the Aberfoyle tin-tungsten deposit, Tasmania, Australia. Chemical Geology, 70, 287–299. Etminan, H., Hoffmann, C.F. (1989) Biomarkers in luid inclusions: a new tool in constraining source regimes and its implications for the genesis of Mississippi Valley-type deposits. Geology, 17, 19–22. Andrawes, F., Holzer, G., Roedder, E., Gibson, E.K.J., Oro, J. (1984) Gas chromatographic analysis of volatiles in luid and gas inclusions. Journal of Chromatography, 302, 181–193. Jixi, S., Benchao, L., Jiamo, F., Dehan, L., Pingan, P. (1988) Organic inclusions and their relations to oil and gas occurrences. Organic Geochemistry, 13, 1101–1107. Kihle, J., Andresen, B., Munz, I.A. (1995) Quantitative C1-C15+ extraction of hydrocarbons trapped as luid inclusions; improved possibilities for isotope determinations and PVT modelling. In Organic Geochemistry: Developments and Applications to Energy, Climate, Environment and Human History, edited by Grimalt, J.O., Dorronsoro, C. Donostia-San Sebastian: A.I.G.O.A., pp. 736–738. George, S.C., Jardine, D.R. (1994) Ketones in a Proterozoic dolerite sill. Organic Geochemistry, 21, 829–839. George, S.C., Eadington, P.J., Hamilton, P.J. (1995) Organic geochemistry of petroleum-bearing luid inclusions in quartz grains from a Cretaceous sandstone. In Organic Geochemistry: Developments and Applications to Energy, Climate, Environment and Human History, edited by Grimalt, J.O., Dorronsoro, C. Donostia-San Sebastian: A.I.G.O.A., pp. 282–283. Kihle, J., Johansen, H., Karlsen, D. (1995) A luid inclusion and geochemical study of liquid hydrocarbon-

118.

119.

120.

121.

122.

123.

124.

125.

126.

127.

128.

129.

bearing dyke at Tvedestrand, on-shore Norway. Geonyrr, Norwegian Geological Survey, 22, 40. Walderhaug, O. (1994) Precipitation rates for quartz cement in sandstones determined by luid-inclusion microthermometry and temperature-history modeling. Journal of Sedimentary Research Section A—Sedimentary Petrology and Processes, 64, 324–333. Walderhaug, O. (1994) Temperatures of quartz cementation in Jurassic sandstones from the Norwegian continental shelf—evidence from luid inclusions. Journal of Sedimentary Research Section A—Sedimentary Petrology and Processes, 64, 311–323. Isaksen, G.H., Pottorf, R.J., Jenssen, A.I. (1998) Correlation of luid inclusions and reservoired oils to infer trap ill history in the South Viking Graben, North Sea. Petroleum Geoscience, 4, 41–55. Bhullar, A.G., Karlsen, D.A., Backer-owe, K., Seland, R.T., Letran, K. (1999) Dating reservoir illing—a case history from the North Sea. Marine and Petroleum Geology, 16, 581–603. Pan, C.C., Yang, J.Q. (2000) Geochemical characteristics and implications of hydrocarbons in reservoir rocks of Junggar Basin, China. Chemical Geology, 167, 321–335. Volk, H., George, S.C., Killops, S.D., Lisk, M., Ahmed, M., Quezada, R.A. (2002) The use of luid inclusion oils to reconstruct the charge history of petroleum reservoirs— an example from the Taranaki Basin. In Proceedings of the 2002 New Zealand Petroleum Conference, Crown Minerals, Auckland, pp. 221–233. Volk, H., George, S.C., Lisk, M., Killops, S.D., Ahmed, M., Quezada, R.A. (2001) Charge histories of petroleum reservoirs in the Gippsland and Taranaki Basins—Evidence from the analysis of oil inclusions and crude oils. In Eastern Australasian Basins Symposium, A Refocused Energy Perspective for the Future, edited by Hill, K.C., Bernecker, T. Melbourne: Petroleum Exploration Society of Australia, Special Publication, pp. 413–422. Worden, R.H., Oxtoby, N.H., Smalley, P.C. (1998) Can oil emplacement prevent quartz cementation in sandstones? Petroleum Geoscience, 4, 129–137. Teinturier, S., Pironon, J., Walgenwitz, F. (2002) Fluid inclusions and PVTX modelling: examples from the Garn Formation in well 6507/2-2, Haltenbanken, Mid-Norway. Marine and Petroleum Geology, 19, 755– 765. van Koeverden, J.H., Nakrem, H.A., Karlsen, D.A. (2010) Migrated oil on Novaya Zemlya, Russian Arctic: evidence for a novel petroleum system in the eastern Barents Sea and the Kara Sea. American Association of Petroleum Geologists Bulletin, 94, 791–817. O’Brien, G.W., Lisk, M., Duddy, I., Eadington, P.J., Cadman, S., Fellows, M. (1996) Late Tertiary luid migration in the Timor Sea: a key control on thermal and diagenetic histories? Australian Petroleum Production and Exploration Association Journal, 36(1), 399–427. Volk, H., George, S.C., Middleton, H., Schoield, S. (2005) Geochemical comparison of luid inclusion and present-

REFERENCES

130.

131.

132.

133.

134.

135.

136.

137.

138.

139.

140.

141.

day oil accumulations in the Papuan Foreland—evidence for previously unrecognised petroleum source rocks. Organic Geochemistry, 36, 29–51. George, S.C., Volk, H., Ruble, T.E., Brincat, M.P. (2002) Evidence for an unusual oil family in the Nancar Trough area, Timor Sea. Australian Petroleum Production and Exploration Association Journal, 42(1), 387–404. Summons, R.E., Jahnke, L.L., Hope, J.M., Logan, G.A. (1999) 2-Methylhopanoids as biomarkers for cyanobacterial oxygenic photosynthesis. Nature, 400, 554–557. Brocks, J.J., Logan, G.A., Buick, R., Summons, R.E. (1999) Archean molecular fossils and the early rise of eukaryotes. Science, 285, 1033–1036. Waldbauer, J.R., Sherman, L.S., Sumner, D.Y., Summons, R.E. (2009) Late Archean molecular fossils from the Transvaal Supergroup record the antiquity of microbial diversity and aerobiosis. Precambrian Research, 169, 28–47. Dutkiewicz, A., Volk, H., Ridley, J., George, S. (2003) Biomarkers, brines, and oil in the Mesoproterozoic, Roper Superbasin, Australia. Geology, 31, 981–984. Volk, H., George, S.C., Dutkiewicz, A., Ridley, J. (2005) Characterisation of luid inclusion oil in a Mid-Proterozoic sandstone and dolerite (Roper Superbasin, Australia). Chemical Geology, 223, 109–135. Brocks, J.J., Love, G.D., Summons, R.E., Knoll, A.H., Logan, G.A., Bowden, S.A. (2005) Biomarker evidence for green and purple sulphur bacteria in a stratiied Palaeoproterozoic sea. Nature, 437, 866–870. Summons, R.E., Jahnke, L.L. (1990) Identiication of the methylhopanes in sediments and petroleum. Geochimica et Cosmochimica Acta, 54, 247–251. Eigenbrode, J.L., Freeman, K.H., Summons, R.E. (2008) Methylhopane biomarker hydrocarbons in Hamersley Province sediments provide evidence for Neoarchean aerobiosis. Earth and Planetary Science Letters, 273, 323–331. Munz, I.A., Yardley, B.W.D., Banks, D.A., Wayne, D. (1995) Deep penetration of sedimentary luids in basement rocks from southern Norway: evidence from hydrocarbon and brine inclusions in quartz veins. Geochimica et Cosmochimica Acta, 59, 239–254. Schubert, F., Diamond, L.W., Toth, T.M. (2007) Fluidinclusion evidence of petroleum migration through a buried metamorphic dome in the Pannonian Basin, Hungary. Chemical Geology, 244, 357–381. Macleod, G., Petch, S., Larter, S.R., Aplin, A.C. (1993) Investigations of the composition of hydrocarbon luid inclusions. In Abstracts of the 205th ACS National

142.

143.

144.

145.

146.

147.

148.

149.

150.

151.

673

Meeting, San Diego, CA, Division of Geochemistry. Abstract 86. Pang, L.S.K., George, S.C., Quezada, R.A. (1998) A study of the gross compositions of oil-bearing luid inclusions using high performance liquid chromatography. Organic Geochemistry, 29, 1149–1161. Clementz, D.M. (1976) Interaction of petroleum heavy ends with montmorillonite. Clays and Clay Minerals, 24, 312–319. Crocker, L., Wang, T.B., Kebarle, P. (1993) Electron afinities of some polycyclic aromatic hydrocarbons, obtained from electron-transfer equilibria. Journal of the American Chemical Society, 115, 7818–7822. Horstad, I., Larter, S.R., Dypvik, H., Aagaard, P., Bjørnvik, A.M., Johansen, P.E., Eriksen, S. (1990) Degradation and maturity controls on oil ield petroleum column heterogeneity in the Gullfaks ield, Norwegian North Sea. Organic Geochemistry, 16, 497–510. Karlsen, D.A., Larter, S.R. (1991) Analysis of petroleum fractions by TLC-FID: applications to petroleum reservoir description. Organic Geochemistry, 17, 603–617. George, S.C., Eadington, P.J., Lisk, M., Quezada, R.A. (1998) Geochemical comparison of oil trapped in luid inclusions and reservoired oil in Blackback oilield, Gippsland Basin, Australia. Petroleum Exploration Society of Australia Journal, 26, 64–81. Brünsing, J., Volk, H., George, S.C. (2005) Charge history of the Gidgealpa Oilield, Cooper Eromanga Basins— evidence from luid inclusion oil analysis. In System Earth–Biosphere Coupling, Vol. 39, edited by Freiwald, A., Röhling, H.G., Löfler, S.B. Hannover: Deutschen Gesellschaft für Geowissenschaften. Volk, H., Kempton, R.H., Ahmed, M., Gong, S., George, S.C., Boreham, C.J. (2009) Tracking petroleum systems in the Perth Basin by integrating microscopic, molecular and isotopic information on petroleum luid inclusions. Journal of Geochemical Exploration, 101, 109. Dutkiewicz, A., Volk, H., Ridley, J., George, S.C. (2007) Precambrian inclusion oils in the Roper Group. In Proceedings of the Central Australian Basins Symposium (CABS), Special Publication 2, edited by T.J. Munson, G.J. Ambrose, Northern Territory Geological Survey, Alice Springs, Northern Territory, August 16–18, 2005, pp. 326–348. Mossman, D., Gauthier-Lafaye, F., Dutkiewicz, A., Brüning, R. (2008) Carbonaceous substances in Oklo reactors––analogue for permanent deep geologic storage of anthropogenic nuclear waste. Reviews in Engineering Geology, XIX, 1–13.

31 LA-MC-ICP-MS APPLIED TO U-PB ZIRCON GEOCHRONOLOGY Alain Cocherie and Michèle Robert

31.1

INTRODUCTION

During the mid-1980s, elemental and isotopic studies of the solid earth started by coupling an inductively coupled plasma (ICP) source with a quadrupole mass spectrometer (MS) [1]. Some years later, during the early 1990s, geoscientists linked a laser ablation (LA) system to such an ICP-MS instrument [2] to investigate minerals at the 30- to 50-µm-diameter crater scale. In parallel with improvements to ICP-MS technology, Nd :YAG light ampliication by stimulated emission of radiation (LASER) technology operating in the near infrared at 1064 nm was replaced by frequencyquadrupled UV lasers operating at lower wavelength (266 nm) generating lower elemental fractionation. The pioneering studies applied to geochronology provided the irst in situ dating [3] limited to 207Pb/206Pb ages on zircon, because elemental fractionation led to poor reproducibility for U/Pb ratios and unreliable 206Pb/238U ages. On the one hand, Tiepolo proposed using a singlecollector magnetic sector ICP-MS to obtain higher abundance sensitivity and a resulting lat-topped peak shape, which allows more precise isotope ratios to be recorded [4]. On the other hand, laser improvements led to reduced elemental fractionation (U/Pb) for new lasers operating at 213 nm [5] and slightly later in the case of excimer gas lasers operating at a shorter wavelength of 193 nm. Despite these improvements, laser systems still provided transient signals recorded in dynamic mode (peak jumping) using either quadrupole ICP-MS or magnetic sector ICP-MS. One of the main

improvements has been simultaneously collecting all of the peaks of interest at the same time on different collectors. This principle was applied for about 20 years to all isotope systematics in the geosciences ield after dissolution of rock samples using thermal ionization mass spectrometer (TIMS) measurements. Nowadays, a Faraday cup system, including about nine cups and usually a central ion counter, is available. At the beginning of the twenty-irst century, this type of coniguration was also installed on magnetic sector ICP-MS systems equipped with a Faraday cup multicollector (MC) system, creating MC-ICP-MS instruments. However, some minor Pb peaks (207Pb and 204Pb) obtained after vaporization of zircon by a laser cannot be properly recorded by a Faraday cup collector. Consequently, when the irst MC-ICP-MS was equipped with a multi-ion counting (MIC) system, this new technology was applied for the irst time to U-Pb zircon geochronology [6]. A set of reference zircons with a wide age range was dated using this method [7,8]. The results were successfully compared with the in situ reference method using a secondary ion mass spectrometer (SIMS) dedicated to U-Pb zircon geochronology: the sensitive high-resolution ion microprobe (SHRIMP). The protocol applied to collect raw data and calculate the associated mean age using LA-MC-ICP-MS and an MIC system will be developed below. Before using the laser system for geochronological purposes, the sensitivity and the precision that can be achieved using such MC-ICP-MS (MIC) instrument was evaluated for determining Pb/Pb ratios on natural water [9].

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

675

676

LA-MC-ICP-MS APPLIED TO U-PB ZIRCON GEOCHRONOLOGY

FIGURE 31.1 Microphotograph (transmitted light) of small zircon grains exhibiting cracks and inclusions, which can be avoided using in situ dating (scale bar 50 µm).

These measurements on solutions will ix the best precision that cannot at present be exceeded using a laser for sample introduction in the MC-ICP-MS. After preliminary remarks about the U-Pb system in a zircon mineral phase, the irst section describes the method to achieve convenient preparation of the samples. In the second section, the laser conditions to achieve isotope measurements on zircon are listed, and some reference zircon samples, already dated by using other techniques, are dated to check the precision and accuracy of the LAMC-ICP-MS protocol. Finally, an example of an investigation of very young zircons, which are the most dificult samples because they contain very little amounts of radiogenic Pb*, is discussed.

31.2

PRELIMINARY REMARKS

Some questions may need to be answered in this preliminary section for scientists not directly involved in geosciences: (1) What is the speciic interest of in situ dating? (2) Why is the U-Pb system so frequently used to date minerals? (3) Why is zircon one of the best candidates for dating geological events? 31.2.1

Interest of In Situ Dating

Despite considerable improvements to the isotope dilution (ID)-TIMS technique to reduce the amount and size of the analyzed zircon grains, the minimum volume investigated by this method is about the size of a rather large single mineral grain (∼100–200 µm in length).

Consequently, one obvious interest of in situ dating at 20–30 µm size (in diameter) is to enable dating of grains signiicantly smaller than 100 µm, which is quite common. In addition, most grains contain solid inclusions belonging to other mineral species and therefore must not be analyzed simultaneously with the zircon material itself (Figure 31.1). Since zircon mineral can contain signiicant amounts of U, α-particles may be emitted during the natural radioactive disintegration of this element, which can damage the mineral structure, especially when luid circulation is also involved (Figure 31.2). Tectonic events also produce cracks that can contain impurities. Only an in situ approach allows damaged and cracked areas to be avoided. The inal critical point is that despite zircon being a robust mineral, a “recent” rim can recrystallize during some high-temperature geological events. Therefore, inherited cores occur very commonly in natural zircons (Figure 31.3). Thus, after a careful observation of the scanning electron microscopic images (mainly cathodoluminescence [CL]), suitable domains and age heterogeneities can be investigated only by in situ analyses. 31.2.2

U-Pb Isotope System

The natural radioactivity of uranium can be summarized as the decay of 238U to stable 206Pb and as the decay of 235U to stable 207Pb. The half-lives of 238U and 235U (4.468 × 109 and 0.7038 × 109 years, respectively) are much longer than those of their daughters. Therefore, these decay series satisfy the condition of secular equilibrium in a U-bearing mineral, which remains as a

MATERIALS AND METHODS

577 ± 12 Ma

677

351 ± 7 Ma

1889 ± 29 Ma 337 ± 8 Ma

200µm

FIGURE 31.3 Scanning electron microscope image (cathodoluminescence) of two zircon grains from the same granite (V. Thiery, pers. comm.). Note that the elongated simple grain gives the same age (∼340–350 Ma) as the edge of the complex grain. In addition, the inner inherited domain is constituted of two domains very different in age (577 and 1889 Ma).

31.2.3 Why Use Zircon for U-Pb Geochronology? FIGURE 31.2 Backscattered electron image of a zircon grain showing botryoidal alteration (dark area) and fractures. The lighter domains are suitable for LA-MC-ICP-MS (scale bar 100 µm).

closed system. Subsequently, the decay of uranium in minerals can be directly treated as two simple equations: 206

Pb = ( 206 Pb )0 + 238 U ( eλ 238 t − 1)

and 207

206

Pb = ( 207 Pb )0 + 235 U ( eλ 235t − 1) , 207

Zircon (ZrSiO4) is a heavy accessory mineral (density 4.6), which is never abundant in rocks but occurs commonly as rare grains in many crustal rocks. Uranium is present in most zircon grains in reasonable amounts to allow Pb isotopic daughters to be determined under good analytical conditions. In addition, “zircon is a remarkably robust reservoir for radiogenic Pb, even when subject to intense and prolonged heating” [10]. Experimental and empirical studies show that zircon remains closed under most crustal processes. Thus, a U-Pb zircon geochronometer is one of the most powerful tools for deciphering magma processes over a wide age range: from the irst witness of the crustal growth (about 4000 Ma) to quaternary times (about 1 Ma or less).

31.3

MATERIALS AND METHODS

206

where ( Pb)0 and ( Pb)0 are the initial amount of Pb and 207Pb isotopes at the time of the mineral formation, λ238 and λ238 are the decay constants of 238U and 235U isotopes (1.55125 × 10−10 year−1 and 9.8485 × 10−10 year−1, respectively), t is the time elapsed since closure of the mineral to U and Pb. Theoretically, initial Pb (also called common-Pb) is easily subtracted by measuring 206Pb/204Pb in the mineral. Then, the corrected data plotted in the (206Pb/238U) = f(206Pb/238U) should be on the Concordia curve if the system has remained closed for U and Pb from the time of the formation of the mineral. Potentially, a meaningful age can be calculated for each spot analysis.

31.3.1

Zircon Extraction from the Rock

All the analytical methods used for U-Pb geochronology on zircon require the separation of grains from the rock sample. Usually, 3–5 kg of the rock sample is progressively crushed. More than 10 kg of basic rocks (low SiO2) can be required because zircon is sometimes very rare in such materials. Then, a riddling stage is applied at 400 µm and 1 mm. The fraction below 400 µm is poured into a pan to reject the very ine grains 55 ka as dem-

131.

132.

133.

134.

135.

136.

137.

138.

139.

140.

141.

142.

onstrated by protein and DNA sequences. Geology, 30(12), 1099–1102. Nielsen-Marsh, C.M., Richards, M.P., Hauschka, P.V., Thomas-Oates, J.E., Trinkaus, E., Pettitt, P.B., Karavanic, I., Poinar, H., and Collins, M.J. (2005) Osteocalcin protein sequences of Neanderthals and modern primates. Proceedings of the National Academy of Sciences of the United States of America, 102(12), 4409–4413. Ostrom, P.H., Schall, M., Gandhi, H., Shen, T.L., Hauschka, P.V., Strahler, J.R., Gage, D.A. (2000) New strategies for characterizing ancient proteins using matrix-assisted laser desorption ionization mass spectrometry. Geochimica et Cosmochimica Acta, 64(6), 1043–1050. Ostrom, P.H., Gandhi, H., Strahler, J.R., Walker, A.K., Andrews, P.C., Leykam, J., Stafford, T.W., Kelly, R.L., Walker, D.N., Buckley, M., Humpula, J. (2006) Unraveling the sequence and structure of the protein osteocalcin from a 42 ka fossil horse. Geochimica et Cosmochimica Acta, 70(8), 2034–2044. Buckley, M. (2008) Species identiication in ancient and heated bone fragments using protein mass spectrometry. Unpublished PhD thesis, University of York, York. Buckley, M., Collins, M.J., Thomas-Oates, J. (2008) A method of isolating the collagen (I) a2 chain carboxytelopeptide for species identiication in bone fragments. Analytical Biochemistry, 374, 325–334. Buckley, M., Whitcher-Kansa, S., Howard, S., Campbell, S., Thomas-Oates, J., Collins, M.J. (2010) Distinguishing between archaeological sheep and goat bones using a single collagen peptide. Journal of Archaeological Science, 37, 13–20. Buckley, M., Collins, M.J., Thomas-Oates, J., Wilson, J.C. (2009) Species identiication by analysis of bone collagen using matrix-assisted laser desorbtion/ionisation time-oflight mass spectrometry. Rapid Communications in Mass Spectrometry, 23, 3843–3854. Hauschka, P.V. (1980) Osteocalcin: a speciic protein of bone with potential for fossil dating. In Biogeochemistry of Amino Acids, edited by Hare, P.E., Hoering, T.C., King, K.J. New York: Wiley, pp. 75–82. Hauschka, P.V., Wians, F.H. (1989) Osteocalcinhydroxyapatite interaction in the extracellular organic matrix of bone. Anatomical Record, 224, 180–188. Collins, M.J., Gernaey, A.M., Nielsen-Marsh, C.M., Vermeer, C., Westbroek, P. (2000) Slow rates of degradation of osteocalcin: green light for fossil bone protein? Geology, 28(12), 1139–1142. Muyzer, G., Sandberg, P., Knapen, M.H.J., Vermeer, C., Collins, M., Westbroek, P. (1992) Preservation of the bone protein osteocalcin in dinosaurs. Geology, 20(10), 871–874. Termine, J.D., Kleinman, H.K., Whitson, S.W., Conn, K.M., McGarvey, M.L., Martin, G.R. (1981) Osteonectin, a bone-speciic protein linking mineral to collagen. Cell, 26, 99–105.

REFERENCES

143. Fisher, L.W., Termine, J.D., Dejter, J.S.W., Whitson, S.W., Yanagishita, M., Kimura, J.H., Hascall, V.C., Kleinman, H.K., Hassel, J.R., Nilsson, B. (1983) Proteoglycans of developing bone. The Journal of Biological Chemistry, 258(10), 6588–6594. 144. Franzen, A., Heinegard, D. (1985) Isolation and characterization of two sialoproteins present only in bone calciied matrix. Biochemical Journal, 232(15), 715–724. 145. Heinegard, D., Hultenby, K., Oldberg, A., Reinholt, F., Wendel, M. (1989) Macromolecules in bone matrix. Connective Tissue Research, 21(1–4), 3–11. 146. Price, P.A., Urist, M.R., Otawara, Y. (1983) Matrix Gla protein, a new gamma-carboxyglutamic acid-containing protein which is associated with the organic matrix of bone. Biochemical and Biophysical Research Communications, 117, 765–771. 147. Cattaneo, C., Gelsthorpe, K., Phillips, P., Sokol, R.J. (1990) Blood in ancient human bone. Nature, 347, 339.

795

148. Shoshani, J., Lowenstein, J.M., Walz, D.A., Goodman, M. (1985) Proboscidean origins of mastodont and woolly mammoth demonstrated immunologically. Paleobiology, 11, 429–437. 149. Kadler, K.E., Holmes, D.F., Trotter, J.A., Chapman, J. (1996) Collagen ibril formation. Biochemistry Journal, 316, 1–11. 150. Poser, J.W., Price, P.A. (1979) A method for the decarboxylation of γ-carboxyglutamic acid in proteins. Journal of Biological Chemistry, 254(2), 431–436. 151. Asara, J.M., Schweitzer, M.H., Freimark, L.M., Phillips, M., Cantley, L.C. (2007) Protein sequences from mastodon and tyrannosaurus tex revealed by mass spectrometry. Science, 316(5822), 280–285. 152. McBride, D.J. Jr., Choe, V., Shapiro, J.R., Brodsky, B. (1997) Altered collagen structure in mouse tail tendon lacking the alpha 2(I) chain. Journal of Molecular Biology, 270(2), 275–284.

36 ARCHAEOMETRIC DATA FROM MASS SPECTROMETRIC ANALYSIS OF ORGANIC MATERIALS: PROTEINS, LIPIDS, TERPENOID RESINS, LIGNOCELLULOSIC POLYMERS, AND DYESTUFF Maria Perla Colombini, Francesca Modugno, and Erika Ribechini

36.1

Since ancient times, natural organic materials have been used as paint binders, adhesives, waterprooing materials, ointments, balms, and pharmaceutical preparations. A chemical analysis of the residues from organic materials associated with human activity in the past, which can be found in a variety of locations and deposits at archaeological sites, can contribute to a better understanding of the everyday life, crafts, and technologies of past societies. The information obtained from chemical analysis can also help to map ancient trade routes, assess the origin of materials, and devise conservation and restoration procedures [1–5]. However, the identiication of organic materials recovered from archaeological indings is particularly arduous since organic substances: •





course of time, and is deeply inluenced by environmental conditions.

INTRODUCTION

are biosynthesized by living organisms (plants or animals), and are thus complex mixtures of many chemical species, which are quite often very similar to each other; have a chemical composition that has been altered by humans before and during use (e.g., since the Neolithic period, treatments based on the distillation or pyrolysis (Py) of resinous wood have been carried out to produce pitches and tars); are sensitive to chemical, physical, and biological factors, so that their composition changes in the

Therefore, the fate of the components in the complex organic materials from an archaeological residue is dificult to predict. Furthermore, the chemical transformations that the residues are likely to have undergone need to be considered. Speciic studies have been carried out over the last 20 years using artiicial aging and experimental archaeology in order to reconstruct the decay patterns of proteinaceous, lipid, polysaccharide, and resinous materials in artworks and archaeological materials. The main categories of organic materials that are encountered in archaeological organic residues are summarized in Table 36.1. Mass spectrometry (MS) has proved to be one of the most powerful analytical tools in archaeometric research on organic materials for the following reasons: •



the range of available MS-based methods allows for the molecular analysis of a wide variety of organic species, from low molecular weight volatile monoterpenes in plant extracts to macromolecules such as proteins, or polysaccharides, or lignocellulosic polymers; the coupling of MS with separation techniques, such as gas chromatography (GC) or high performance

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

797

798

ARCHAEOMETRIC DATA FROM MASS SPECTROMETRIC ANALYSIS OF ORGANIC MATERIALS

TABLE 36.1 Types of Organic Materials Encountered in Archaeology and Examples of Their Uses Category Proteins Glycerolipids

Natural waxes Natural resins

Polysaccharide materials Fossil materials Lignocellulosic materials Organic dyes





Organic Materials

Uses

Egg, milk, and casein; animal glue; silk; wool; vegetable proteins; human and animal tissues Animal fats, vegetable oils including drying oils

Beeswax, spermaceti wax, Chinese wax, lanolin, carnauba, candelilla, Japan wax, esparto wax Conifer resins, sandarac, mastic, dammar, frankincense, benzoe, styrax, myrrh, shellac (animal resin), tar and pitch (from thermal treatment of plant resins or wood) Starch, cellulose, plant gums (arabic gum, tragacanth, karaya, ghatti, guar, locust bean, fruit tree gum) Bitumen, asphalt, fossil resins

Food, commodities, adhesives, textiles, paint binders, parchment Food, commodities, paint binders, varnishes, illuminants, ingredients in cosmetic and pharmaceutical preparations Coatings, sealants, molding materials, writing tablets, ingredients in cosmetic, pharmaceutical preparations Varnishes, coatings, waterprooing materials, paint binders, mordants, ingredients in cosmetic and pharmaceutical preparations Paper, paint binders, adhesives, coatings

Wood, ibers

Molding materials, adhesives, pigments, ingredients in cosmetic and pharmaceutical preparations Furniture, tools, architectural structures, ships, textiles

Cochineal, madder, kermes, saffron, purple, indigo, etc.

Colorants for dying textiles, organic pigments (lakes), ingredients in cosmetics

liquid chromatography (HPLC) and with analytical Py, provides optimal solutions for the resolution of complex mixtures of organic species represented by naturally aged materials; the ability to obtain quantitative and qualitative information via standards, reference materials, and calibration curves, and the identiication of unexpected species such as unforeseen components or degradation products, through the interpretation of mass spectra; the low detection limits and the negligible amount of sample required for the analysis.

This review provides an up-to-date overview of the application of analytical procedures based on MS for the characterization of organic natural materials in archaeological and historical objects. Applications that feature the use of gas chromatography/mass spectrometry (GC/MS), Py-GC/MS, high performance liquid chromatography combined with mass spectrometry (HPLC/MS), and direct MS analysis such as matrixassisted laser desorption/ionization mass spectrometry (MALDI-MS), electrospray ionization mass spectrometry (ESI-MS), and direct exposure mass spectrometry (DE-MS) are summarized to highlight the different information provided by each of the various analytical approaches. Case studies and examples are also presented and include a description of the molecular markers and of the molecular proiles that are used to identify the original materials.

36.2

GC/MS

GC/MS is the most frequently used MS-based analytical technique in archaeometry. The versatility of GC for the investigation of organic materials in this ield was pioneered by Mills and White [1] and conirmed by a number of successful applications and case studies [2,4,6] relative to a variety of materials and matrices. The relatively low cost and widespread availability of GC/MS instruments have contributed to make this analytical platform the irst choice for solving many analytical problems that involve the organic chemistry of archaeological and historical objects. The natural organic substances found in archaeological samples are complex mixtures of many chemical species, some of which were originally present, while others are degradation products. Therefore, the resolution and determination of the molecular proile is essential to identify the original materials and their aging pathways. Consequently, the coupling of GC with MS is essential due to the high number of compounds with similar retention times and the possible presence of unexpected or unknown species. Furthermore, reliable identiication cannot be based only on retention time since many compounds are not available as commercial standards. Thus, the conirmation from mass spectra is required. GC/MS chromatograms allow the qualitative and quantitative molecular ingerprint of the injected samples to be inferred. The use of internal standards and calibration curves allow for accurate, absolute, and

GC/MS

routine quantitation. In addition, GC/MS can determine different chemical families in a single analytical run, which is particularly useful when samples are limited and precious, as is the case with artistic and archaeological objects. However, only in a few cases is GC/MS used to directly analyze analytes in the sample, due to the fact that many of the natural materials that humans have exploited since ancient times are macromolecular. Often, these materials correspond to natural polymers, such as proteins, plant gums, lignocellulosic materials, or fossil resins. Other materials undergo polymerization or cross-linking as an effect of exposure to light and oxygen, such as terpenic resins and drying oils. Some materials contain polar functional groups and are not volatile for GC separation. For these reasons, it is not possible to directly analyze these materials. In the majority of cases, GC/MS determination is preceded by sample pretreatments aimed at chemically transforming the macromolecules or the low volatile analytes in the sample into a volatile form, suitable for GC. Fragmentation of macromolecules can be achieved by coupling thermochemolysis (Py) with GC/MS (Py-GC/ MS), or, more commonly, performing a wet chemical treatment of the sample. Although the analytical procedures are relatively laborious and time-consuming, GC/ MS is still unsurpassed in its capacity to unravel the composition of the archaeological organic materials at a molecular level. Table 36.2 shows a representative list of applications reported in the literature. The wet chemical sample pretreatments for the GC/ MS analysis often entail [2,6] (1) extraction and puriication of the compounds, (2) chemolysis of macromolecules and high molecular weight species into smaller molecules, and (3) derivatization of polar functional groups. Extraction is necessary to obtain the analytes in solution and, in some cases, to separate analytes from interfering species that may be present in the sample matrix. For example, when organic residues in porous ceramic vessels need investigating, potsherds are generally ground into a ine powder, prior to adding a suitable solvent (dichloromethane, methanol, acetone, or mixtures). This maximizes the contact between the residue and the solvent, thus enabling the lipid materials entrapped in ceramic pores to be characterized [7]. The organic extracts can then be separated into fractions for analysis to simplify data interpretation and to allow for the optimization of the chromatographic conditions for each class of compounds [8–15]. In some cases, puriication steps are included in the procedure. For example, paint samples that contain interfering pigments need to be puriied in order to identify proteinaceous binders on the basis of the quan-

799

titation of amino acids [16]. Metal cations and anions give rise to analytical interferences in the hydrolysis of proteins and the derivatization of amino acids. To overcome these problems, the extract or the hydrolysate can be puriied using a miniaturized sorbent tip (C18 or C4 stationary phase) [17,18]. Figure 36.1 illustrates the result of the amino acid analysis of a wall painting sample from the Etruscan tomb “Tomba della Quadriga Infernale” (Sarteano, Siena, Italy) that dates back to the second half of the fourth century b.c. GC/MS analysis was performed after acidic hydrolysis and puriication using a C18 pipette tip, followed by silylation of the amino acids. The results revealed that egg had been used as an organic binder to disperse the pigments [19]. The same procedure was recently used to demonstrate that egg was the proteinaceous material used as a paint medium for the “Terracotta Army,” in the burial complex of the emperor Qin Shihuangdi (210 b.c., Lintong, China) [20]. Amino acid analysis is commonly used for the identiication of proteins in paint layers, where contamination is low, the conservation state is generally good, and the range of possible proteins is very limited (mainly egg, collagen, or casein). However, the identiication of archaeological proteinaceous residues is by no means an easy task and requires peptide mapping using HPLC/ MS or MALDI techniques (see Section 36.5). With the increased availability of advanced instrumentation, proteomics may become the favored approach for protein determination also in paint layers [21]. Chemolysis is needed in order to free small molecules from macromolecules or polymers, such as amino acids from proteins, fatty acids from triglycerides, or sugars from polysaccharides. Speciically: •



proteins and polysaccharides are submitted to acidic hydrolysis in order to free the amino acids [16] and sugars [22,23], respectively; to ensure the completeness of the reaction and to minimize any loss of labile components, the hydrolysis conditions must be optimized appropriately (i.e., milder for polysaccharides, harsher for proteins); the chemolysis of ester bonds of triglycerides and wax esters is performed by alkaline hydrolysis or methanolysis; selective transesteriication to obtain fatty acid methyl esters (FAMEs) is one of the most common approaches [24,25], together with alkaline saponiication in hydroalcoholic KOH [8,26–28], which also enables saponiiable fraction (acids) and neutral fraction (alcohols, sterols, alkanes) to be separated.

Derivatization reactions are required in order to transform molecules that contain polar functional

800 TABLE 36.2

Examples of GC/MS Investigations of Samples Collected from Archaeological Findings

Identiied Material

Sample Matrix

Egg protein [19,20]

Paint binder samples

Polysaccharide gums [22]

Mural paintings from Macedonian tombs (third to fourth century b.c.) and from the Mycenaean Nestor palace (thirteenth century b.c.), Greece Residues of illuminants in Roman and Coptic lamps from Egypt

Oil from the seeds of Brassicaceae [25,48,49] Diterpenoid resin, beeswax, plant waxes [27]

Balms in Roman glass unguentaria

Birch bark tar [47,84,94]

Hafting material from prehistoric stone tool

Pitch obtained from Pinaceae wood [91]

Internal coating in Roman amphorae

Benzoe resin [8]

Residue in a censer from Roman Egypt

Sample Pretreatment Acidic hydrolysis, puriication, and silylation Acidic hydrolysis, conversion of sugars into diethyl-dithioacetals and silylation

Alkaline saponiication and trimethylsilylation [48,49]; transesteriication [25] Alkaline saponiication and trimethylsilylation, and HTGC/MS of the total lipid extract Alkaline saponiication, solvent extraction, and trimethylsilylation [84]; solvent extraction and trimethylsilylation [47,94] Alkaline saponiication and trimethylsilylation

Alkaline saponiication and trimethylsilylation

Characteristic Proile and Markers Amino acids Sugars and uronic acids

Characteristic fatty acids (11,12-dihydroxyeicosanoic acid and 13,14-dihydroxydocosanoic acid) Characteristic proile of fatty acids, longchain alcohols, alkanes, cerides, diterpenoids (dehydroabietic acid) Betulin, betulone, lupenone, lupeol, lupa-2,20(29)-dien-28-ol

Diterpenoids with abietane structure (dehydroabietic acid, 7-oxodehydroabietic acid, 15-hydroxy-7-oxodehydroabietic acid, 15-hydroxy-dehydroabietic acid, retene, tetrahydroretene, norabietatriene, norabietatetraene, and methyldehydroabietate) 4-Hydroxy-benzaldehyde, vanillin, 3-hydroxy-benzoic acid, 4-hydroxybenzoic acid, and vanillic acid

Identiied Material

Sample Matrix

Sample Pretreatment

Characteristic Proile and Markers

Frankincense [83]

Resinous sample from Dahshour site (Egypt, XIIth Dynasty)

Trimethylsilylation of the methanolic extract

Beeswax [35,46]

Residues from ceramic vessels and adhesives from Egyptian opus sectile

Fats, oils, beeswax, bitumen, coniferous resin, mastic resin [54,56]

Egyptian human and animal mummies dating from the mid-dynastic period (1900 b.c.) to the late Roman period (395 a.d.)

Solvent extraction and trimethylsilylation [35], alkaline saponiication and trimethylsilylation [46] Alkaline saponiication and trimethylsilylation, and HTGC/MS of the total lipid extract

3-epi-β-amyrin, 3-epi-α-amyrin, 3-epi-lupeol, 3α-hydroxy-lupen-24-oic acid, α- and β-boswellic acids, their O-acetates and their products of degradation Characteristic proile of fatty acids, longchain alcohols, alkanes, cerides

Fats, beeswax, diterpenic resins, vegetable tannin, castor oil [24] Diterpenoid and triterpenoid resins [88] Bitumen [9]

Ptolemaic mummy (100 b.c.) from the Guimet Museum in Lyon

Methanolysis

Archaeological samples from ancient Egypt Coating inside ceramic buff ware recovered in Anuradhapura (Sri Lanka, third to ninth century)

Headspace solid-phase microextraction (SPME)-GC/MS Direct analysis of total lipid extracts

Diterpenoids (7-oxo-dehydroabietic and 15-hydroxy-7-oxo-dehydroabietic acid); triterpenoids (isomasticadienonic, masticadienonic, moronic and oleanonic acids); alkanes (C25-C33), wax esters (C40-C50), hydroxy wax esters (C42-C54); hydroxyaromatic acids; fatty acids; steranes and hopanes Fatty acids (including 12-hydroxy-(Z)-9octadecenoic acid), diterpenoids, gallic acid, inositols Mono- and sequiterpenoids Sterane (m/z 217) and hopane (m/z 191) patterns

801

802

ARCHAEOMETRIC DATA FROM MASS SPECTROMETRIC ANALYSIS OF ORGANIC MATERIALS Abundance gly 60,000 55,000 50,000 45,000 ala 40,000 35,000 30,000 25,000 20,000 15,000 10,000 5000

s.i.2

s.i.1

leu val ile pro

ser phe

asp

glu lys

tyr

16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00 32.00 34.00 36.00 38.00 Time-->

FIGURE 36.1 GC/MS chromatogram, acquired in single ion monitoring (SIM) mode, of the proteinaceous fraction of a paint sample from the Etruscan tomb “Tomba della Quadriga Infernale” (Sarteano, Siena, Italy) [19], containing egg protein. The analysis entailed hydrolysis, puriication using a C18 pipette tip and silylation of the amino acids with N-methyl-N-(t-butyldimethylsilyl) triluoroacetamide (MTBSTFA) in pyridine.

groups, such as carboxylic, hydroxy, and amminic functions, into the corresponding esters, ethers, and amides. This derivatization procedure increases the volatility of the analytes and improves the chromatographic performance. For example, carboxylic functional groups can be transformed into the corresponding methyl esters, acetyl derivatives, ethylchloroformate derivatives, or trimethylsilyl (TMS) ester derivatives [2]. Many derivatizing reagents are commercially available and include methanolic HCl and diazomethane for methylation, and N,O-bis(trimethylsilyl)triluoroacetamide (BSTFA) for silylation. Hydroxy moieties can be silylated in the same step with BSTFA, along with carboxylic moieties to form the corresponding TMS ethers. The derivatization of monosaccharides can be particularly tricky, as sugars undergo intramolecular reactions to form ive- and six-membered cyclic hemiacetals. Each hexose can be present in ive forms (two pyranoside, two furanoside, and an open-ring one) and thus produce highly complex chromatograms and result in irreproducible quantiication and a loss of sensitivity. To obtain a single peak for each sugar, aldoses and uronic acids can be converted into the corresponding diethyldithioacetals and diethyl-dithioacetal lactones before silylation [22]. The mass spectra of the derivatives obtained are shown in Figure 36.2 for xylose and galacturonic acid. This approach has been used for the characterization of plant gums used as paint materials in Macedonian tombs (third to fourth century b.c.) [22]. Lipid materials and terpenic resins have always been widely exploited in crafts, art, and everyday life, and are the most commonly found organic materials in relation to archaeological artifacts [1,2,4,5,8,29–32]. The preservation of these materials is due to their hydrophobic

behavior and to their higher resistance to biological agents, compared with more easily degradable substances such as proteins and polysaccharides. There is a huge number of GC/MS applications used for lipid analysis in archaeological samples reported in the literature, and lipid characterization, together with the study of lipid degradation, is an important research area in archaeometry [1,2,4,5,33–35]. Lipid analysis has meant that not only can the source of food residues be identiied [34,36–42], but also information can be gained on the use of vessels [43] and on many others commodities and goods, such as ointments and personal care products [27,44,45], adhesives [46,47], paint binders [6,13], lighting fuels [25,48–51], and ritual preparations such as incense and mummiication balms [52–57]. Lipid analysis has also been used to study the preservation of body tissues [58–62]. All the approaches based on chemolysis and GC analysis give detailed information on fatty acid and alcohol proiles, but do not reveal anything about the extent of the hydrolysis of triglycerides or wax esters in the sample. The determination of the degree of lipid hydrolysis is particularly interesting in terms of degradation, since the hydrolysis of the ester bond is one of the main decay paths for triglycerides and wax esters. In order to measure the degree of hydrolysis, the lipid extracts need to be analyzed directly by high-temperature gas chromatography/mass spectrometry (HTGC/MS) [35,53,63]. Using HTGC/MS, it is possible to study a large range of molecular weights, by adopting short nonpolar capillary columns (10–15 m) with a thin stationary phase and a GC temperature up to 350°C. Beeswax, the most widely exploited natural wax since ancient times, can be identiied by analytical

GC/MS

803

(A) SEt

73 EtS

H OTMS

H TMSO

H OTMS

H

249

319

205

CH2OTMS

147 103 176 m/z--> 50 (B)

100

150

277 200

250

300

349 393421 454482 350

400

450

73

550

601 643 600

650

217

100

161 150

189 200

305

245 250

300

SEt

EtS O

OTMS

TMSO

103 50

500

135

O

m/z-->

568

530

OTMS

498 596 335 408 526 564 437 469 363 625 350

400

450

500

550

600

650

FIGURE 36.2 Mass spectra of (A) trimethylsilylated xylose-dithioacetal (MW 544) and (B) trimethylsilylated galacturonic aciddithioacetal (MW 498).

methods based on conventional GC/MS after hydrolysis of the wax esters, in terms of the occurrence of [1,2,64]: long-chain fatty acids with an even number of carbons (from C16 to C34, with a very high abundance of palmitic acid); long-chain linear alcohols with an even number of carbons (from C24 to C34, peaking at C30); long-chain linear saturated hydrocarbons with the prevalence of an odd number of carbons (from C21 to C33), as shown in Figure 36.3 [46]. (ω-1)-Hydroxy acids with an even number of carbons and long-chain (α,ω-1)-diols with an even number of carbons are also present. On the other hand, when the lipid extract is directly analyzed using HTGC/MS, the proile of beeswax is characterized by long-chain palmitate esters (C40–C52 peaking at C46) and long-chain hydroxypalmitate esters (C40–C52), in addition to odd-numbered linear hydrocarbons (C21–C33), and even-numbered long chain fatty acids (C22–C34 peaking at C24), as shown in the case study in Figure 36.4 [35,65]. Figure 36.5 shows the mass spectra of triacontanoylhexadecanoate ester, TMS ester of tetracosanoic acid, and TMS ether of triacontanol. The identiication of the speciic source of lipids after the GC/MS analysis of an archaeological sample is by no means straightforward. The analysis is complicated

due to the occurrence of mixtures, the unspeciic fatty acid composition of many lipid materials, and the changes induced by aging. Thus, the exact origin of lipids can only be assessed by GC/MS in favorable situations. Such favorable situations are represented by natural waxes, with a quite speciic molecular pattern, and by a few oils and lipids that contain uncommon fatty acids that can be used as biomarkers that have been exploited for their recognition in organic archaeological residues. For example, castor oil contains large amounts of 12-hydroxy-(Z)-9-octadecenoic acid (ricinoleic acid), which produces a very characteristic oxidation product: 9,12-dihydroxy-octadecanoic acid [24,49]. Oils obtained from the seeds of Brassicaceae, such as rapeseed oil and radish oil, are characterized by abundant amounts of uncommon fatty acids such as Z-11-eicosenoic (gondoic) acid and Z-13-docosenoic (erucic) acid and, after aging or heating, by their oxidation products 11,12dihydroxyeicosanoic acid and 13,14-dihydroxydocosanoic [48,49,52]. The electron ionization (EI) mass spectra of the TMS derivatives of the latter compounds are shown in Figure 36.6. The α-cleavage of the bond between the two vicinal trimethylsiloxy groups leads to the formation of fragments at m/z 215 and 345 for

 



























7LPH PLQ







2+

$

2+



2+

UHWHQH



Q&



Q& 2+



2+

,6



Q&



Q&



Q&



5HODWLYHLQWHQVLW\

2+

 

,6

%

,6



$ $







2+



$



2+



2+



$



$



,6

5HODWLYHLQWHQVLW\



2+$ 2+$ GLGHK\GURDELHWLF DF $ GHK\GURDELHWLF DF 2+$2+$ R[RGHK\GURDELHWLFDF $

$

$

2+

ARCHAEOMETRIC DATA FROM MASS SPECTROMETRIC ANALYSIS OF ORGANIC MATERIALS

PHWK\OGHK\GURDELHWDWH

804











27

C24:0

Relative intensity (%)

100

C16:0

FIGURE 36.3 GC/MS/total ion chromatogram (TIC) of the trimethylsilylated (A) acidic fraction and (B) neutral fraction of a sample of adhesive from Egyptian opus sectile, containing beeswax and pine pitch. AX:Y is the fatty acids of chain length x and degree of unsaturation y; OHx is the n-alkanols of chain length x; XOHAX is the hydroxy fatty acids of chain length x and with hydroxy group at position X; n-Cx is the n-alkanes of chain length x. IS1 and IS2 are the n-hexadecane and tridecanoic acid internal standards [46].

Wax esters 46

10

15

C34:0

50 52

A34

20

48

IS A32 C30:0

33

44 42

C32:0

C20:0

25

A28 C28:0

C22:0 A24 A26 C26:0

C18:0

29

A30

40 31

25

30

35

Retention time (min)

FIGURE 36.4 HTGC/MS/total ion chromatogram (TIC) of the trimethylsilylated charred residue OD3006C1 from Bercy, showing the presence of altered beeswax. The numbered peaks between 25 and 33 correspond to n-alkanes; Ax corresponds to linear long-chain alcohols containing x carbon atoms, and Cy:0 corresponds to saturated long-chain fatty acids with y carbon atoms. IS corresponds to the internal standard [35].

11,12-dihydroxyeicosanoic acid, and at m/z 215 and 373 for 13,14-dihydroxydocosanoic acid. Other abundant fragment ions and ion radicals at m/z 204, 217, [M-105]+, and [M-90]+ are present in the mass spectra of both compounds. The peak at [M-90]+ is due to the loss of a trimethysilanol molecule from the molecular ion,

while the peak at [M-105]+ is related to the consecutive loss of trimethysilanol and of a methyl radical. The peak corresponding to [M-15]+, related to the loss of a methyl radical from the TMS group, is also evident. The loss of nonanal ([M-CH3(CH2)7CHO]+) via a rearrangement process leads to the formation of a cation

GC/MS

805

(A)

Relative Abundance

57

Relative Abundance

(B)

Relative Abundance

(C)

100 257 95 90 85 80 75 70 65 71 60 8397 55 50 45 40 35 111 30 25 20 125 15 139 10 153 185 213 239 285 5 0 50 100 150 200 250 300 100 95 90 85 80 75 70 65 60 55 50 75 45 40 35 30 57 25 73 103 83 20 97 15 111 10 125 153 5 139 0 50 100 150

100 73 95 90 85 80 75 70 65 75 60 55 50 45 40 35 57 30 25 20 69 15 83 97 10 5 0 50 100

676

466

350

400 m/z

450

500

550

600

650

700

500

550

600

650

700

500

550

600

650

495

496

185 207

200

241 269

283

250

325 354

300

350

396

424 460 480

400 m/z

450

117

425

129 145

426 201 185 171 227 257269

150

200

250

382 398 325 341 355

300

350

400 m/z

440

450

700

FIGURE 36.5 Mass spectra of (A) triacontanoylpalmitate (molecular weight [MW] = 676), (B) TMS ether of triacontanol (MW = 510), and (C) TMS ester of tetracosanoic acid (MW = 440).

radical corresponding to the observed peak at m/z [M-142]+. Lipids from marine products have been identiied in archaeological residues in pottery on the basis of the presence of ω-(o-alkylphenyl)alkanoic acids with 16, 18, and 20 carbon atoms, together with isoprenoid fatty acids (4,8,12-trimethyltetradecanoic acid and phytanic acid). These products are presumed to have been formed during the heating of tri-unsaturated fatty acids (C16:3, C18:3, and C20:3), fatty acyl components of marine lipids [36,37]. However, the majority of lipid sources present an unspeciic proile of fatty acids. Thus, the lipid sources

are rarely identiied in archaeological residues by GC/ MS. Over the last few years, gas chromatography/ combustion/isotope ratio mass spectrometry (GC/C/ IRMS) has become an increasingly important tool to glean information on the use and consumption of lipid materials in ancient times [66,67]. The evaluation of δ13C values of fatty acids in lipids trapped in archaeological potsherds has increased the potential of GC/MS in palaeodiet studies. The online coupling of IRMS with GC in GC/C/IRMS, which has been commercially available for the last 10 years, has led to the determination of stable isotope values for speciic molecular components. Bulk δ13C and δ15N analysis has been used to correlate

806

ARCHAEOMETRIC DATA FROM MASS SPECTROMETRIC ANALYSIS OF ORGANIC MATERIALS (A)

345

Relative Abundance

Relative Abundance

100 95 90 85 m/z = 215 O-TMS 80 75 COO-TMS 70 65 TMS-O 147 m/z = 345 60 55 73 11,12-dihydroxyeicosanoate, MW = 560 50 204 45 40 69 81 35 [M-142]+· 95 217 30 [M-90-15]+ 25 20 129 165 [M-90]+· 418 119 15 [M-15]+ 273 +· 291 10 317 455 183 545 M 329 373 439 470 5 560 0 50 100 150 200 250 300 350 400 450 500 550 600 m/z (B) 373 100 95 90 m/z = 215 85 O-TMS 80 COO-TMS 75 70 TMS-O m/z = 373 65 60 13,14-dihydroxydocosanoate, MW = 588 55 50 73 45 204 40 147 35 69 95 [M-142]+· 83 30 [M-90-15]+ 217 25 109 175 20 133 [M-90]+· 446 15 [M-15]+ 283 189 10 +· 467 301 319 357 573[M] 273 483 5 498 588 0 50 100 150 200 250 300 350 400 450 500 550 600 m/z

FIGURE 36.6 Mass spectra of TMS derivatives of (A) 11,12-dihydroxyeicosanoic acid and (B) 13,14-dihydroxydocosanoic acid.

the composition of the bone collagen of fossils/skeletal remains with diet [68], while the strong speciicity of the determination of compound-speciic stable isotope values after GC separation makes this method suitable for interpreting the isotope values of food residues in archaeological pottery. Figure 36.7 illustrates the GC/C/ IRMS instrumentation. This platform has proven to be a suitable approach to distinguish between plant and animal lipids, and between ruminant and porcine adipose fats, on the basis of the differences in the δ13C values of fatty acids [8,66,67,69,70]. In addition, differences in the δ13C values of fatty acids of ruminant adipose and dairy fats have been exploited to address issues related to prehistoric subsistence practices, animal husbandry, and dairying activities [71–75]. Recent studies have attempted to clarify how the δ13C values of individual fatty acids are

affected by environmental variables, geographical origin, climatic inluences, and cooking. The aim is to obtain globally applicable parameters for the classiication of fats [41,42,74,76]. A similar approach has been applied to palaeodiet studies to determine the δ13C values of cholesterol [77], fatty acids, and individual amino acids in skeletal and body remains [78–80], interpreted together with bulk δ13C and δ15N of collagen and δ13C of apatite. The assessment of the origin of terpenic resins by GC/MS is generally based on the qualitative recognition of speciic molecular biomarkers that are normally speciic for the genus [1,2,8,27,81–86]. Figures 36.8–36.10 illustrate the mass spectra of several biomarkers of mastic resin, frankincence resin, and birch bark tar, respectively. The GC/MS analysis of di- and triterpenoids requires the extraction and derivatization of

GC/MS

Al2 O3 furnace housing Autosampler

Heater lead Alumina tube

(a)

Alumina tube

807

Catalyst

(1.6-mm i.d., 260-mm length)

Cu/Pt (850°C) Cu/Pt/Ni wire (940°C)

movable

COMBUSTION REACTOR

Tpiece

REFERENCE GAS

Combustion reactor

Fused-silica capillary

5 mm of stationary phase removed

Ion source

Column Reduction reactor (optional) Vent Vent O2

Water separator He

Vent Open-split interface He

Areas prone to blockages, leaks, and adsorption effects

Computer data system

Triple collector for m/z 44, 45, and 46

Amplifier array

FIGURE 36.7 Schematic diagram of a GC/C/IRMS instrument conigured for the determination of δ13C values of organic compounds. GC-separated compounds are individually combusted on-line over a catalyst generating CO2 and H2O. H2O is removed by a water separator, and the remaining CO2 is introduced into a mass spectrometer equipped with a triple collector comprising three Faraday cups monitoring simultaneously m/z 44, 45, and 46, which corresponds to the ions of the three isotopomers 12C16O2, 13 16 C O2, and 12C18O16O, respectively [66].

acidic functions [2]. The derivatization of alcoholic functions is also often performed. Mono- and sesquiterpenoids are normally not encountered in aged samples, due to their volatility, but in some cases, they are retained in bulk materials, such as mummiication balms, and their selective sampling and recognition can be performed using solid-phase microextraction (SPME)-GC/ MS [87–89]. The botanical species exploited in the past for the collection of resins vary according to the geographical area, and thus, the availability of adequate reference materials of a known botanical origin is crucial for resin identiication. Pine resin and resin-derived pitch, mastic, frankincense, myrrh, benzoe, and birch bark tar are the most common terpenoid materials in archaeological artifacts from the Mediterranean area. Terpenic resin composition depends on aging and exposure to air and light, so the biomarkers used for their identiication are

rarely the original components of the native materials. More commonly, the biomarkers are stable aging products or compounds formed during the collection, puriication, and transformation of the terpenic materials before use, such as the retorting process of resinous wood to obtain tar and pitches used as waterprooing material and adhesives. For example, tar and pitch are produced from Pinaceae resin and woods, and retene is considered to be a stable end product of the reaction pathways that occur during the production of pitches. Norabietatrienes, simonellite, and tetrahydroretene represent the intermediates of these reactions [4,90– 92]. Lupa-2,20(29)-diene, lupa-2,20(29)-dien-28-ol, and lupa-2,20(29)-diene are indicative of a thermal treatment of birch bark and, with betulin and lupeol together with low amounts of lupenone, betulone, and betulinic acid, have been used as biomarkers to identify birch bark tar in prehistoric hafting materials [47,84,93,94].

408 100 95 90 85 COOTMS COOTMS 80 75 70 O O 65 189 60 Oleanonic acid TMS Moronic acid TMS 203 55 409 derivative, MW 526 derivative, MW 526 50 45 187 409 40 35 391 526 30 306 133 511 25 471 393 173 20 119 381 526 320 15 215 145163 472 511 105 281 10 363 73 95 235 436453 353 5 67 436 486 255269 295 365380 469485 0 350 400 450 500 550 600 650 50 100 150 200 250 300 350 400 450 500 550 600 650

Relative Abundance

Relative Abundance

189

100 95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0

187 306

203

307

119

73

50

235 217

175 147 105 91 132 161

100

150

281 295 267

200

250

300

m/z

m/z

Relative Abundance

421

100 95 COOTMS 90 85 80 O Masticadienonic acid TMS 75 derivative, MW 526 70 65 60 COOTMS 55 50 45 40 O 393 35 Isomasticadienonic acid TMS 403 30 derivative, MW 526 311 365 25 257 20 15 169 297 10 157 357 207227251 271 5 6775 97 123143 181 0 50 100 150 200 250 300 350 400

511

422 493

526

462

450

500

550

600

650

m/z

FIGURE 36.8 Mass spectra of the characteristic compounds of mastic resin. MW, molecular weight.

292

292

TMSO COOTMS a-boswellic acid TMS derivative, MW 600

381392 364

350

400

585 600

471 495 510

450

500

100 95 90 85 80 75 70 65 60 202 55 50 45 40 35 218 30 175 293 147 25 189 20 161 119 15 73 107 133 276 95 10 231 257 5 5567 0 650 50 100 150 200 250 300

Relative Abundance

Relative Abundance

100 95 90 85 203 80 75 70 65 202 60 55 218 50 45 40 175 35 293 30 189 25 73 119 20 161 95107 15 276 257 10 67 79 231 306 5 59 0 50 100 150 200 250 300

550

600

381 392 364

350

495 510

585

600

600

650

100 95 90 85 80 75 70 65 218 174 60 AcO 55 COOTMS 203 50 45 O-acetyl-b-boswellic acid TMS 189 147 40 derivative, MW 570 119 161 35 107 133 30 293 25 91 20 15 73 495 351 377393 10 67 510 229 257270 364 316 5 55 421 535 570 0 650 50 100 150 200 250 300 350 400 450 500 550 600

650

400

450

500

550

AcO

COOTMS O-acetyl-a-boswellic acid TMS derivative, MW 570

351

377 393 405 421

350

m/z

400

495 510

450

500

555 570

550

600

Relative Abundance

292

203

Relative Abundance

COOTMS b-boswellic acid TMS derivative, MW 600

m/z

m/z 100 95 90 85 80 75 70 292 65 218 60 55 50 175 189 45 174 40 35 30 119 25 105 133147 293 20 161 95 79 15 10 67 229 257276 5 55 305 0 50 100 150 200 250 300

TMSO

m/z

FIGURE 36.9 Mass spectra of the characteristic compounds of frankincense resin. MW, molecular weight.

PY-GC/MS 189

100 95 90 85 80 75 70 65 TMSO 60 203 Lupeol TMS derivative, 55 MW 498 50 175 45 107 40 218 95 121 147 161 35 135 393 30 81 498 25 73 297 229 20 67 408 369 483 257 279 325 15 385 10 415 455 243 306 339 5 55 442 0 650 50 100 150 200 250 300 350 400 450 500 550 600

Relative Abundance

Relative Abundance

203

100 95 90 189 85 CH2 OTMS 80 75 70 496 TMSO 393 65 60 Betulin TMS derivative, 55 50 MW 586 45 483 147 40 119 95 35 133 161 216 73 30 279 25 229 20 391 15 67 255 293 453 337 363 406 10 427 349 5 55 0 50 100 150 200 250 300 350 400 450 500

571 586

550

600

m/z 100 406 95 90 85 CH2 OTMS 80 187 75 70 65 159 60 Lup-2,20(29)-dien-28-ol 55 TMS derivative, 189 50 203 119 163 45 MW 496 40 145 105 133 95 35 173 30 91 229 25 81 363 20 73 311 324 15 67 255 281297 337 377 414 10 496 243 271 350 427 450 481 5 55 0 50 100 150 200 250 300 350 400 450 500 550 600 650

Relative Abundance

Relative Abundance

650

m/z 393

409

100 422 95 90 85 80 CH2OTMS 75 203 70 O 65 163 189 60 BetuloneTMS derivative, 55 50 MW 512 45 95 147 40 107 119 35 91 30 379 81 25 245 327 353 231 20 15 67 271 483 298311 339 367 10 497 391 438 5 55 458 512 0 50 100 150 200 250 300 350 400 450 500 550 600 650

m/z

m/z

189

189

100 100 95 95 205 90 90 85 85 204 80 80 119 75 75 95 70 70 297 65 65 409 107 147 O 60 60 313 161 119 133 55 175 55 133 91 91 Lupenone, MW 424 393 218 50 50 79 147 424 159 45 45 40 40 79 218 395 408 35 67 35 67 173 381 245 30 30 298 229 25 25 368 255 243 20 283 20 229 257 271 357 365 15 15 323 341 55 10 55 10 311 269 381 5 5 0 0 100 150 200 250 300 350 400 50 100 150 200 250 300 350 400 450 500 550 600 650 50

Relative Abundance

Relative Abundance

809

m/z

Lup-2,20(29)-diene, MW 408

450

500

550

600

650

m/z

FIGURE 36.10 Mass spectra of the characteristic compounds of birch bark tar. MW, molecular weight.

The composition of tar and pitches and the relative proportions of their constituents can change profoundly and depend on the duration and the intensity of the hard-heating treatment.

36.3

PY-GC/MS

The need for time-consuming sample preparation, which always carries the risk of both sample loss and contamination, is the main drawback of the GCbased approach for the analysis of archaeological samples. In addition, GC/MS can only provide informa-

tion on polymeric materials when they are eficiently chemolyzed using wet chemical pretreatments. This is not the case with lignocellulosic polymers and wood, fossil resins such as amber, terpenic resins containing a polymeric fraction, and synthetic polymers added during restoration. Online coupling of analytical Py with the GC/MS system (Py-GC/MS) achieves the thermal cleavage of macromolecules, as a rapid alternative to chemical cleavage, and offers the possibility of investigating high molecular weight/polymeric materials that cannot be made volatile for GC/MS analysis. Online coupling also avoids time-consuming sample pretreatments. Evidence

810

ARCHAEOMETRIC DATA FROM MASS SPECTROMETRIC ANALYSIS OF ORGANIC MATERIALS

of many natural materials such as lipids, resinous materials, wood, polysaccharides, and proteins, and of synthetic polymers that may have been introduced during restoration, can be simultaneously provided with Py-GC/ MS. Thus, Py-GC/MS is of particular interest for handling retreatment problems. Py-GC/MS is often used to perform a fast initial screening of samples in order to choose the appropriate samples and procedures for a subsequent, more detailed, GC/MS analysis. The potential of analytical Py in cultural heritage studies was irst explored in the late 1980s [95,96], and one of the irst archaeometric applications was the recognition of rosin, beeswax, and polysaccharide gums in samples from Egyptian mummy cartonnages [97]. Since then, Py-GC/ MS has been used to study terpenic materials [46, 98–105], glycerolipids [53,106–110], natural waxes [64,111,112], proteins [113–115], organic dyes [116,117], tannins [118], iron-gallic inks [119], polysaccharides [44,120], and lignocellulosic materials [121–123] in a large number of studies related to historical, archaeological, and artistic objects [2,124–127]. A list of case studies is shown in Table 36.3. When Py-GC/MS is used, the chemical composition of the sample is reconstructed on the basis of an interpretation of the molecular proile of the thermal degradation products of the original components, and on the recognition of speciic molecular markers or of characteristic molecular patterns, which act as ingerprints of the pyrolyzed material. Py proiles are strongly dependent on the instrument type and the experimental parameters, particularly the Py temperature, types of pyrolyzer (microfurnace, Curie point, resistively heated

TABLE 36.3

ilament probe), Py-GC interface, and sample composition and morphology. One example is the use of Py-GC/MS to examine archaeological wooden artifacts and to study the degradation patterns of wood. Archaeological wooden objects are relatively rare, as they are very sensitive to attack from insects, fungi, and microorganisms. Consequently, they are recovered only in very particular sites such as waterlogged or aqueous environments where, due to low temperatures and oxygen concentrations, fungi and insects are not active. Waterlogged wooden artifacts, such as shipwrecks, are sometimes recovered in a seemingly very good condition. However, even under water in near anoxic conditions, some species of anaerobic bacteria slowly attack the wood, mainly by eroding cellulose and hemicellulose as sources of nutrients [128]. This condition leads to long-term degradation phenomena, which seriously threaten the stability of the artifacts, especially during/after the recovery and drying, and thus need speciic conservation treatments [129]. Although the macromolecular complexity of wood limits the possibility of obtaining complete chemical information on its alteration, Py-GC/MS has proven to be a valuable tool for investigating wood, lignin, cellulose, and hemicellulose in plant remains and archaeological artifacts [121,122,130,131]. The use of Py-GC/MS avoids the long procedures classically used in wood analysis and only requires a minimal sample size (a few tenths of a microgram). Thus, Py-GC/MS is suitable for screening many samples in large artifacts such as shipwrecks. When wood is pyrolyzed, a complex mixture of products is formed including low molecular weight com-

Examples of Py-GC/MS Applications in Archaeometry

Identiied/Investigated Material

Sample

Natural gums, di- and triterpenic resins, waxes [97] Fatty acids and characteristic markers for proteins and polysaccharides [110] Beeswax, animal glue [111]

Egyptian cartonnages

Lipids [53]

Egyptian mummy balms

Pine pitch, beeswax, plant oil [46]

Adhesive from Egyptian opus sectile

Wood components [121]

Archaeological waterlogged wood (oak) from various shipwrecks Archaeological waterlogged wood (pine, elm, beech) of the Roman period Roman cosmetic (second century a.d.) Archaeological wine of the Roman period

Wood components [122] Starch [44] Wine polyphenols [118]

Pyrolysis Experimental Conditions

Charred food remains from early Roman pottery

Curie-point pyrolyzer; Py temperature 610°C

Paint layers from Egyptian sarcophagi

Curie-point pyrolyzer; Py temperature 610°C Curie-point pyrolyzer; TMAH; Py temperature 610° Pt-heated ilament pyrolyzer; HMDS; Py temperature 550°C Curie point; Py temperature 610°C Double-shot micro furnace pyrolyzer; Py temperature 500°C Not reported Pt-heated ilament pyrolyzer; TMAH; Py temperature 450°C

PY-GC/MS

811

(A) Abundance 6e+07 S

4e+07 C

2e+07

C C

0 Time-->

GS

G

C C C C

C

C C C

CC

G G S G G S G G S

G G G H

H

C

G G

G

S

G

G

C

5.00

S

S

G

10.00

C

G S

S S S G S

S G S G S

G

S S S

S G

15.00

20.00

S

S

25.00

(B) Abundance

G

S

S

S S G

6e+07

S

G

G

G S

4e+07 G

G G G

2e+07

G

G

S G S

C

C

H

H

G G S

G G

S

C C C

G S S S S S

G

S S

S S S

G S G

C

0 Time-->

5.00

10.00

15.00

20.00

25.00

FIGURE 36.11 Py-GC/MS proiles of (A) sound beech wood (Fagus sylvatica) and (B) archaeological beech wood from the Pisa site. C, carbohydrates pyrolysis products; G, guaiacyl-lignin pyrolysis products; S, siringyl-lignin pyrolysis products. Pyrolysis was performed with a Py-2020iD double-shot micro furnace pyrolyzer at T 500°C [122].

pounds and levoglucosan derived from polysaccharides, and relatively simple phenols that result from the cleavage of ether and C–C lignin bonds. The phenols produced retain their substitution patterns from the lignin polymer; thus, it is possible to identify the components of the wood from the p-hydroxyphenylpropanoid (H), guaiacylpropanoid (G), and syringylpropanoid (S) lignin units [132]. A detailed molecular analysis of Py products achieves semiquantitative results on the extent of cellulosic loss and highlights the chemical modiications undergone by lignin, such as the demethylation of guaiacyl and syringyl units [121,122]. Figure 36.11 shows a Py-GC/MS chromatogram obtained from the analysis of a sample of waterlogged beech from the excavation of the site of San Rossore in Pisa (Italy), where more than 30 Roman shipwrecks have been found in a relatively good condition, dating to a period between the fourth century b.c. and the eighth century a.d. [122,133]. The irst section of the pyrograms (3–7 min) is dominated by products that arise from the Py of carbohydrates (C): furfural 2,3-dihydro-5-methylfuran-2-one; 2(5H)-furanone; 2hydroxy-3-methyl-2-cyclopenten-1-one; and levoglucosan. Lignin thermal cleavage leads to a complex mixture of phenolic products (7–27 min), mainly guaiacyl (G) and syringyl (S) moieties with different alkyl substituents: guaiacol, 4-vinylguaiacol, syringol, E-isoeugenol, 4-methylsyringol, 4-vinylsyringol, 4-allylsyringol, E-4propenylsyringol, and syringaldehyde are the major peaks. Small but not negligible amounts of catechols

and methoxy catechols that are derived from the demethylation of lignin units are also observed [122]. The differences between archaeological waterlogged wood and sound wood lie in the relative amounts of the products observed, which highlight how lignin was well preserved in archaeological wood and, at the same time, clearly shows polysaccharide degradation. This difference provides evidence of a substantial degradation in the wood due to the selectivity of the anaerobic degradation toward polysaccharide components in an aqueous environment. These components were preferentially removed and a residue enriched in lignin remained. The relative loss of polysaccharide Py products, in terms of peak areas, from sound to archaeological beech wood, is up to 90%. Many organic natural substances that can be found in association with archaeological artifacts under Py produce low volatile molecules that contain polar functionalities, which are not eficiently separated by GC. This is the case with fatty acids and di- and triterpenoid acids produced in the Py of lipids and natural resins. Over the last few years, the most common approach to overcome problems related to GC analysis of low volatile polar Py products is to use thermally assisted reactions. Py of the sample in the presence of a suitable derivatization reagent, which transforms the polar functionalities of the Py products into less polar moieties, improves the analytical performance and detection limits. Thermally assisted hydrolysis and methylation (THM) based on quaternary alkylammonium

812

ARCHAEOMETRIC DATA FROM MASS SPECTROMETRIC ANALYSIS OF ORGANIC MATERIALS DAB

Phenolic markers

Ret Isomers

N-Sim of PIM

ABT OAB

Retention time

FIGURE 36.12 Pyrogram obtained by THM-GC/MS from Roman Caecuban wine from an amphorae recovered in the shipwreck Madrague de Giens. N-Sim, norsimonellite; Ret, retene; PIM, pimaric acid; DAB, dehydroabietic acid; ABT, abietic acid; OAB, 7-oxodehydroabietic acid [118].

hydroxides, and thermally assisted silylation are the most widely adopted in situ thermally assisted reactions [2,124]. Experimental conditions, such as the amount of sample, the concentration of the reagent, the temperature of the Py chamber, and the Py-GC interface strongly inluence the mechanisms and the yields of the reactions involved. Tetramethylammonium hydroxide (TMAH) is the most commonly used reagent for THM, and TMAH thermochemolysis has been extensively applied to the characterization of organic natural materials [98,103,104,107,108,124,127,134–138]. Py with TMAH involves the deprotonation of carboxylic acids and the hydrolysis of ester and ether bonds, followed by the formation of tetramethylammonium salts, which are subsequently subjected to thermal dissociation and leads to the formation of the corresponding methyl derivatives. Figure 36.12 shows the chromatogram obtained with TMAH thermally assisted hydrolysis and methylation gas chromatography/mass spectrometry (THM-GC/ MS) on a wine residue sample (Caecuban) from a sealed Roman amphorae recovered from the shipwreck Madrague de Giens [118]. Thermochemolysis revealed the preservation of tannins during past millennia via the identiication of a large number of pyrolytic phenolic markers from fruit derivatives. Di- and trimethoxylated benzenoid compounds (trimethoxybenzene, dimethoxybenzaldeide, trimethoxystyrene, trimethoxybenzoic acid-methyl ester, and dimethoxycinnamic acid-methyl ester, among the most abundant) formed in the Py of proanthocyanidins gave an indication of the preservation of the polyphenolic molecular structure in the sample. The diterpenic pitch used to waterproof the ceramic vessel was also evidenced. Despite the utility of the THM process, an interpretation of the obtained pyrograms must take into account the fact that the alkalinity of TMAH causes side reac-

tions such as decarboxylation, dehydroxylation, recombination, isomerization of unsaturated acylic chains, degradation of alkaline-sensitive compounds, keto-enol tautomerization, and α-methylation of acidic moieties [98,136,137,139,140]. For example, in the analysis of terpenic resins, Py reactions lead to a chromatographic proile that is generally more complex than that obtained with GC/MS analysis after extraction and derivatization. The effectiveness of the methylation reaction of the hydroxy groups varies according to the chemical structure of the analytes, the solvent chosen for the TMAH solution, and the Py temperature applied. As an alternative to TMAH, other less alkaline reagents have been proposed. These include trimethylsulfonium hydroxide (TMSH) [141], tetramethylammonium acetate (TMAAc) [142], (m-triluoromethylphenyl) trimethylammonium hydroxide (TFTMTH) [134,143], or silylating reagents such as BSTFA [144] and hexamethyldisilazane (HMDS) [45,46,98,101,102,105,108,114– 116,120,123,145]. HMDS and BSTFA have the advantage of silylating carboxylic, alcoholic, and phenolic functions. Incomplete derivatization and extensive fragmentation can, in some cases, increase the number of peaks and make an interpretation of the pyrograms rather complex. For example, in the analysis of triglycerides, a shortening of the acylic chains and side Py reactions produce a fatty acid proile in the pyrograms that does not relect the original lipids in the sample. HMDS provide excellent results for the analysis of natural and fossil resins and allow for complete silylation of the diterpenoid species. The result of the Py(HMDS)GC/MS analysis of succinite, also known as Baltic amber, is shown in Figure 36.13. This analysis revealed the presence of mono-, sesqui-, and diterpenoids together with succinic acid. In addition, several bicyclic Py degradation products were observed. These compounds are related to the polylabdanic structure of the polymeric fraction of succinite.

36.4

HPLC/MS

Over the past decade, the use of HPLC/MS has grown enormously in the cultural heritage ield. Table 36.4 shows a summary of the HPLC/MS applications of archaeometric interest. HPLC/MS is essential for the analysis of organic dyes that originate from antraquinoid, indigoid, and lavonoid species and is also essential for the study of lipids. HPLC/MS has also been successfully used in the study of archaeological proteins [146], wine [147], and cacao [148] residues. The primary advantage of HPLC/MS is its ability to analyze a wide range of compounds. Analytes that are thermally labile, exhibit a high polarity, or have a high molecular weight and are

Succinic acid-TMS

Relative abundance

HPLC/MS

Pyrolysis degradation products

COOTMS

Ploymer

COOTMS Ploymer

COOTMS R



R

COOTMS

Mono-and sesquiterpenes

Time 10.00

15.00

20.00

813

25.00

30.00

35.00

40.00

45.00

FIGURE 36.13 Pyrogram of a succinite sample obtained by in situ thermally assisted silylation with HMDS.

not suitable for GC analysis, can all be analyzed by HPLC/MS. The possibility of using both atmospheric pressure chemical ionization (APCI) and ESI gives information both on polar and apolar compounds: ESI is more suitable for polar compounds, and APCI for apolar compounds. In addition, the technique has the advantage of short analytical run times, a straightforward sample preparation, and a very high analytical speciicity especially when multidimensional mass spectrometry (MSn) or single ion monitoring (SIM) I acquisition modes are used. The most frequent application of HPLC/MS to archaeological materials currently involve the study of natural organic dyes [149–154]. HPLC/MS methods for the analysis of natural organic dyestuffs have been reviewed [155–157]. HPLC/MS is often used in conjunction with HPLC-UV/VIS and HPLC-diode array detector (DAD) technique to conirm the presence of speciic compounds. In fact, MS detection, both with ESI and APCI, can provide structural information on the basis of fragmentation patterns and thus give insight into the nature of unknown compounds. This aspect is particularly important for the analysis of yellow dyes, which are mainly composed of a mixture of O-glycoside lavonoid isomers. The distribution of O-glycoside lavonoids and the position of the glycoside bond in the molecular structure are highly speciic of the origin of the plant [158]. When analyzing organic dyes, ESI is normally preferred to APCI because it is more suitable for polar compounds, such as anthraquinones, lavonoids, and tannins. APCI generally provides better results for the analyses of less polar compounds such as indigoids [155,156].

Good results for the identiication of triacylglycerols (TAGs) of a vegetable and animal origin, and of longchain esters from beeswax were achieved with HPLC combined with APCI [25,34,50,159]. When these compounds survive in archaeological indings, their distribution and composition can be put in relation with the original lipid material. In fact, HPLC/MS with APCI reveals the positional distribution of fatty acids in TAGs in the corresponding mass spectra via the protonated molecular ion ([M+H]+), diacylglycerol (DAG) ions ([M-RCO2]+), and acylium ions (RCO+) [160]. The positional distributions can be identiied by the abundance of DAG ions: the least abundant DAG ion is formed by a cleavage of the sn-2 position fatty acid, which, due to the steric hindrance, is less favorable than the cleavage at sn-1 and sn-3 positions (Figure 36.14) [34,160]. An evaluation of the ratio of palmitic acid to stearic acid (P/S) at position 2 provided information on the origin of archaeological animal lipids: ruminant lipids show a P/S of 60/40, while nonruminant lipids show a ratio of 95/5 [34]. In terms of the beeswax long-chain esters, monoesters (C40–C50), diesters (C40–C50), hydroxymonoesters (C56–C66), and hydroxydiesters (C56–C64) have been identiied by HPLC/MS [159]. Table 36.5 shows a list of the most important ion fragments of beeswax components revealed using APCI source. Normally, in the HPLC analyses of lipids, the sample is submitted to solvent extraction using a mixture of chloroform/methanol (2:1) and is directly injected into the HPLC/MS system. The extraction can be assisted with sonication, Soxtec, or Soxhlet apparatus. Interestingly, Romanus et al. [25] submitted the solvent

814

TABLE 36.4

Examples of the HPLC/MS Investigations of Samples Collected from Archaeological Findings

Analyte Class/Compound Anthraquinoids (alizarin, purpurin, carminic acid, kermesic acid) [149] Indigoids (indogotin, indirubin), anthraquinoids (alizarin, purpurin), lavonoids (apigenin, luteolin, chrysoeriol, apigenin 7-glucoside, luteolin 7-glucoside, chrysoeriol 7-glucoside, luteolin 7-glucuronide) [150] Flavonoids (quercetin 3-sulfate, quercetin 3-glucoside, isorhamnetin 3-sulfate, isorhamnetin 3-glucoside, kaempferol 3-glucoside, kaempferol, quercetin) [151] Flavonoids (apigenin glucoside, isomer of luteolin glucoside, chrysoeriol glucoside, luteolin glucuronide) [152] Indigoids (indigotin, 6-monobromoindigotin, 6,6′-dibromoindigotin, 6,6′-dibromoindirubin) [153] Anthraquinoids (purpurin, munjistin) [154]

Triacylglycerols [34] Long-chain monoesters (C40–C50) and diesters (C56–C64), di- and triacylglycerols (mainly saturated) [159] Di- and triacylglycerols (mainly unsaturated) [50] Di- and triacylglycerols (mainly unsaturated), oxidized triacylglycerols [25] Myoglobin peptides [146]

Identiied Material

Sample Matrix Pazyryk textiles, 500–200 b.c.

MS System APCI operating in positive-ion mode, single quadrupole analyzer APCI operating in positive- and negative-ion mode, single quadrupole analyzer

Dyes from Rubia spp. and cochineal Indigo, dye from Rubia spp., yellow dyes (unknown plant source)

Chinese textiles, 2000 b.c.

Dye from Flaveria haumanii

Pre-Columbian Andean textiles, 1050–1200 a.d.

APCI operating in positive- and negative-ion mode, single quadrupole analyzer

Yellow dye (unknown plant source)

Chinese textiles, 3000 b.c.

APCI operating in positive- and negative-ion mode, single quadrupole analyzer

Purple

Greek painting, seventeenth century b.c.

Madder

Greek painting materials, Hellenistic period (second to third century b.c.) and Roman period (second century a.d.) Ceramic vessels, Medieval period Roman cooking pots, ifth to sixth century a.d.

APCI operating in negative-ion mode, single quadrupole analyzer with SIM acquisition mode ESI operating in negative-ion mode, single quadrupole analyzer with SIM acquisition mode

Ruminant and nonruminant animal fat Beeswax, ruminant animal fat Olive oil, animal fat Olive oil

Tartaric and syringic acids [147]

Marine muscle tissue protein, probably of seal species Wine

Theobromine, caffeine [148]

Cacao

Roman oil lamps, ifth to sixth century a.d. Coptic oil lamp shells, fourth to tenth century a.d. Alaskan pottery fragment, 1200–1400 a.d. Egyptian ceramic jar, Dynasty I (2920–2770 b.c.), Dynasty XVIII (1550–1307 b.c.), and Dynasty XIX (1307–1196 b.c.) Mayan ceramic vessels, 600 b.c.–a.d. 250

APCI operating in positive-ion mode, triple quadrupole operated in single analyzer mode APCI operating in positive-ion mode, ion trap

APCI operating in positive-ion mode, ion trap APCI operating in positive-ion mode, ion trap NanoESI operating in positive-ion mode, hybrid quadrupole FT-ICR operating in single and tandem mass spectrometry ESI (turbo ion spray) operating in negativeion mode, triple quadrupole MS/MS system with multiple reaction monitoring (MRM) acquisition mode APCI operating in positive-ion mode, single quadrupole analyzer with SIM acquisition mode

HPLC/MS

extracts to a derivatization reaction with N-methyl-Ntrimethylsilyl-triluoroacetamide (MSTFA) prior to the HPLC/MS analyses. This derivatization allows di- and trihydroxylated TAGs, and highly oxidized compounds to be determined in archaeological samples and demonstrates that the oxidation of TAGs may also take place

before their hydrolysis. These compounds contain two or three (CH3)3SiO groups that are derived from OH moieties silylated during the derivatization reaction. As evidenced in the mass spectrum of glycerol-lauratemyristate-(13,14-di-trimethylsilyloxy)docosanoate, shown in Figure 36.15, these compounds show very particular fragmentation patterns dominated by two or three consecutive losses of 90 mass units such as trimethylsylanol [–(CH3)3SiOH]. More recently, peptide sequencing methods have been introduced for the study of protein remains in archaeological ceramics [146]. An analytical procedure

577 [OP]+

Relative abundance

100

815

549 [MyO]+

80 60

523 [MyP]+

TABLE 36.5 Ion Fragments of Beeswax Components Obtained Using APCI Source

40 805 [M+H]+

20

200

400

600 m/z

800

Compounds

Ion Fragments

Monoesters

[M–H]+, [RCOOH–H]+, [RCOO–H2O]+ [M+H]+, [M+H–RCOOH]+, [RCOOH–H]+ [M-H]+, [ROHCOOH–H]+ [M+H]+, [M+H–RCOOH]+

Diesters

1000

Hydroxymonoesters Hydroxydiesters

FIGURE 36.14 APCI mass spectrum of 1-myristoyl-2oleoyl-3-paimitoyl glycerol (MyOP) [160]. O O O

100

O

tms

O O O

O

tms

1000.5 spectrum 27.14 min

Relative Abundance (%)

80

60

910.7

40

20

Relative Intensity (%)

391.1 30 25

0 200

400

820.7 600 m/z 800

1000

1200

20 15 10 5 0 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 Time (min)

FIGURE 36.15 APCI mass spectrum of glycerol-laurate-myristate-(13,14)-dihydroxydocosanoate after derivatization with MSTFA and HPLC chromatogram [25].

816

ARCHAEOMETRIC DATA FROM MASS SPECTROMETRIC ANALYSIS OF ORGANIC MATERIALS

FIGURE 36.16 FT-ICR-MS/MS spectrum obtained from an archaeological protein extract that illustrates the y and b fragments of the myoglobin peptide (VETDLAGHGQEVLIR) via the doubly charged ion at m/z 818.936 [146].

was developed and involves protein extraction, hydrolysis, and peptide analysis with a nano-high performance liquid chromatography (nanoHPLC) nanoESI/MS/MS. The protein extraction was carried out using a solution of triluoroacetic acid (1%) on a inely crushed fragment of archaeological potsherd (roughly 250 mg), and an enzymatic hydrolysis with trypsin was performed on the whole protein extract solution in order to obtain peptides. Lastly, the peptide sequence was obtained using a nanoHPLC coupled with a Fourier transform ion cyclotron resonance mass spectrometer (nanoLC-FT-ICR/ MS) equipped with a nanoelectrospray source [146]. Both MS and MS/MS experiments were carried out. The FT-ICR/MS is particularly suited for the analysis of ultratrace protein remains as in the case of archaeological indings, due to its high mass accuracy, its high resolution, and sensitivity. This was the irst study to identify proteins by sequencing from the archaeological remains of potsherd. It highlighted the presence of a muscle tissue protein, myoglobin (Figure 36.16) with four peptides including one peptide sequence speciic to pinniped and cetacean species and one speciic to seal species. Peptide mass mapping by HPLC/MS has also been used in the study of proteinaceous binders in paintings. Peptide mapping can provide an identiication of the binder used (egg, collagen, or casein), leading to distinguish between egg yolk and egg glair, and determine the animal species from which collagen or casein was obtained [21].

36.5 DIRECT MS TECHNIQUES: DE-MS, DTMS, DI-MS, ESI-MS AND ESI-MS/MS, MALDI-MS, AND LDI-MS Analytical techniques based on direct MS such as DE-MS, direct temperature-resolved mass spectrometry (DTMS), direct inlet mass spectrometry (DI-MS), ESI-MS, laser desorption/ionization mass spectrometry (LDI-MS), and MALDI-MS have become popular as alternative approaches to chromatography for the analysis of organic residues from archaeological indings. In some cases, these techniques, which involve minimal sample manipulation and no sample pretreatment, can reduce contamination and sample loss problems that are associated with wet chemical procedures. In addition, these analytical instruments require very small amounts of sample (a few nano/micrograms). The techniques based on direct MS can be performed rapidly (a few minutes for the analysis and data acquisition), and their sensitivity has been shown to be suitable for the identiication of many organic substances in archaeological objects. In some cases, direct MS approaches can be used as preliminary screening techniques; in others, they can give a complete compositional and structural identiication as discussed in the examples reported below. Table 36.6 shows a summary of the direct mass spectrometric applications for the analysis of archaeological samples. Techniques based on the direct introduction of the sample into the mass spectrometer, such as DTMS,

TABLE 36.6 Technique DE-MS

Examples of the Investigations of Samples Collected from Archaeological Findings Using Direct Mass Spectrometric Techniques Analyte Class/ Compound

Identiied Materials [Reference]

Sample Matrix

Diterpenoids

Pine pitch [161]

Roman amphora

Diterpenoids, triterpenoids

Pine resin, mastic resin, birch bark tar [52,84,91]

Acyl-lipids, cholesterol Long-chain esters Lignins

Fat of animal origin [39]

Roman and Egyptian ceramic vessels, Paleolithic stone tools Egyptian ceramic jar

Waxy material [27] Hard- and softwood [122]

Indigoids

Purple [163]

DTMS

Proteins, lipids, polysaccharides

DI-MS

Diterpenoids, triterpenoids, longchain esters

Charred animal products; starch, charred vegetable, and grain mixtures; protein-rich, lipid-free nonfood products (possibly used to decorate vessels); bone or skin glue [164] Pine resin, birch bark tar, beeswax [165]

Roman glass unguentaria Roman waterlogged wood Roman burial Roman ceramic vessels

Bronze Age ceramic sherd, Iron Age bronze sword, Iron Age ceramic vessel

Ionization Mode Chemical ionization (CI) with isobutane at 70 eV in positive-ion mode EI at 70 eV in positiveion mode

EI at 70 eV in negativeion mode EI at 16 eV in positiveion mode

EI at 70 eV in positiveion mode

MS System Ion trap analyzer

Double-focusing mass spectrometer with B/E geometry

Ion trap analyzer

(Continued)

817

818 TABLE 36.6 Technique ESI-MS and ESI-MS/MS

(Continued) Analyte Class/ Compound

MALDI-MS and MS/MS

Sample Matrix

Ionization Mode

Long-chain esters

Beeswax [168]

Etruscan ceramic vessels

ESI in positive-ion mode

Triacylglycerols

Cow milk, goat milk, and sheep adipose fat [170] Animal fats [169]

Neolithic ceramic vessels

Anthraquinoids (purpurin, pseudopurpurin)

Madder [171]

Roman cosmetics

Anthraquinoids (purpurin, alizarin), triacylglycerols Proteins

Madder, acylglycerolipids [45]

Roman cosmetics

Goat and sheep collagen peptides [172]

Neolithic goat and sheep bones

Mycolic acids

Mycobacterial cell [173]

Human skeletal remains

Proteins

Keratin peptides [174]

Animal hair

NanoESI in positive-ion mode NanoESI in positive-ion mode 337-nm pulsed nitrogen laser (20 Hz) in positive- and negativeion modes 337-nm pulsed nitrogen laser (20 Hz) in positive- and negativeion modes α-Cyano-4hydroxycinnamic acid as the matrix 337-nm pulsed nitrogen laser (20 Hz) in positive-ion mode, fullerene as the matrix 355-nm pulsed Nd :Yag laser (20 Hz) in positive-ion mode

Triacylglycerols LDI-MS

Identiied Materials [Reference]

Ceramic oil lamps

MS System Triple quadrupole mass spectrometer Q-q-TOF analyzer Hybrid quadrupoleFT-ICR/MS TOF analyzer

TOF analyzer

TOF/TOF analyzer

DIRECT MS TECHNIQUES: DE-MS, DTMS, DI-MS, ESI-MS AND ESI-MS/MS, MALDI-MS, AND LDI-MS

819

Wood DE—mass spectra PC1/PC2 score plot

PC2 17.31% total variance

2.0 1.5

Pinus Uimus Fagus Archaeological Pinus Archaeological Uimus Archaeological Fagus

1.0 0.5 –2.5

–1.5

0.0 –0.5

0.5

1.5

2.5

–0.5 –1.0 –1.5 PC1 61.30% total variance

FIGURE 36.17 PCA score plot of PC1 and PC2 of mass spectral data, accounting for 78.61% of total variance [122].

DE-MS, and DI-MS, have been used for the analysis of a variety of chemical substances and materials such as organic colorants, di- and triterpenoid resinous materials, lipids, lignin, polysaccharides and proteins, found in archaeological samples [27,39,52,90,121,161–165]. Wideranging reviews on these techniques used for the analysis of lipid and resinous materials in the cultural heritage ield have been reported [166,167]. Such techniques are based on directly introducing the sample into the mass spectrometer. The sample is placed onto a ilament (as a solid, as a solution, or as a dispersion in a suitable solvent) or in a small glass cup. After the ilament or cup has been introduced into the ion source, the sample is desorbed or pyrolyzed by controlled heating of the ilament. The main advantage is the minimal or no pretreatment of the sample and the quick response. Due to the dificulty in interpreting the mass spectra, multivariate analysis using principal component analysis (PCA) [84,121,162], discriminant analysis (DA), and completelinkage cluster analysis (CLCA) [164] of mass spectral data (relative abundance of peaks) were used to compare, differentiate, and classify the samples. For instance, DE-MS together with PCA was successfully used in the study of degradation processes undergone by archaeological waterlogged wood and lignin. The results revealed the syringyl/guaiacyl ratio and the loss of polysaccharides as the effect of degradation in a waterlogged environment. Figure 36.17 illustrates the score plot for the irst two principal components obtained by PCA of the mass spectral data corresponding to sound and archaeological Roman woods [122]. The score plot, accounting for 79% of the total variance, highlights that the irst principal component discriminates between softwoods (PC1 negative values) and hardwoods (PC1 positive values), while the second principal component discriminates between sound wood

(PC2 positive values) and archaeological wood (PC2 negative values) on the basis of the polysaccharide content. ESI is often referred to as a “soft ionization” method as very little residual energy is retained by the analytes upon ionization. Very little fragmentation is produced, and the mass spectrum often contains a molecular ion indicative of molecular weight. Therefore, MS/MS is often necessary for structural elucidation studies. Sample pretreatments are often required in order to purify the sample and to increase the yield of ionization, and improve the selectivity and sensitivity of the method. High molecular weight and nonvolatile compounds are particularly suitable for an analysis by direct infusion ESI-MS. ESI has been used to study the long-chain esters of beeswax [168] and TAGs of animal fats [169,170]. ESI-MS and MS/MS were successfully used to investigate the origin of the organic residues found in an Etruscan ceramic vessel for the study of beeswax [168]. In the ESI mass spectrum, the mass range between m/z 500 and m/z 900 was the most informative, since the protonated fragments of three homologous series of biomarkers were present: monoesters, hydroxymonoesters, and diesters. NanoESI mass spectra of archaeological samples collected from pottery that contain animal fats are dominated by peaks related to the molecular cation of lithiated adducts of TAGs and DAGs [169,170]. The structure of such compounds was investigated more thoroughly by nanoESI/MS/MS. In the mass spectra of TAGs, a series of peaks corresponding to the loss of fatty acid fragments [M+Li-RCOOH]+ and lithiated fatty acid moieties [M+Li-RCOOLi]+ from the lithiated adduct cation, and to acylium ions [RCO]+ are evident (Figure 36.18). Laser-based ionization techniques, which include LDI-MS and MALDI-MS, are useful tools in the

820

ARCHAEOMETRIC DATA FROM MASS SPECTROMETRIC ANALYSIS OF ORGANIC MATERIALS 18,000

–C16:0

Cow milk: T44:0, m/z 757.69 [M + Li-Rcooh]+ / [M + Li-RcooLi]+ Acylium ions RCO+

495.44

C8:0 to C20:0

12,000 10,000

–C14:0 239.24

Intensity (counts)

14,000

8000 6000

–C18:0 –C12:0

C C14:0 C16:0 C10:0 12:0

4000

–C10:0 –C8:0

–C20:0

2000 0 50

501.45

16,000

100

150

200

250

300

350

400

450

500

550

600

650

m/z

FIGURE 36.18 Mass spectrum obtained by nanoESI/MS/MS of triacylglycerol with 44 carbon atoms [170].

cultural heritage ield. These techniques involve irradiating a sample that contains an analyte substance with a short pulse of radiation emitted by a laser. The radiation is absorbed by the sample and results in the desorption and ionization of molecules from the sample. The sample is mixed with a matrix material that is highly absorbent at the irradiation wavelength and assists the desorption and ionization of the analytes. LDI-MS has been used to analyze Roman and Greek pink-reddish cosmetics [45,171]. The protocol featured the use of the positive- and negative-ion modes and revealed the presence of very small amounts of madder characteristic compounds in all the cosmetics. The negative-ion LDI mass spectra of one of the samples [45], shown in Figure 36.19A, indicated the presence of the characteristic ions at m/z 240, 239, 256, and 255, which correspond to the radical alizarin ion [A]− and its deprotonated molecule [A–H]−, and to the purpurin ion [P]− and its deprotonated molecule [P–H]−. The positiveion LDI mass spectra (Figure 36.19B) conirmed the presence of madder compounds. Potassium adduct ions of alizarin [A+K]+ (m/z 279) and purpurin [P+K]+ (m/z 295) were observed. In addition, in one of the samples (Figure 36.19B), further information on the presence of a lipid component was provided by the positive-ion LDI mass spectra: the presence of TAGs, from the presence of fragments at around m/z 900, and DAGs, from the presence of fragments at around m/z 600 [45]. The most intense peaks were observed at m/z 664, 636, 608, and 619, which could correspond to potassium DAG adducts or could be the result of the TAG fragmentation (corresponding losses of fatty acid residues RCOO−K+). MALDI-MS/MS has been used to distinguish between degraded archaeological sheep and goat bones, thus

overcoming the problems associated with morphological criteria [172]. The method is based on collagen peptide sequencing. It consisted in the isolation and identiication of a single peptide capable of distinguishing between sheep and goat collagen. Other promising results have also been obtained by MALDI in the study of archaeological human bones in order to detect and identify mycobacterial infections such as tuberculosis [173]. Mycolic acids cover a diverse group of fatty acids with molecular weights ranging mainly between m/z 1000 and m/z 2000. They are present in the bones of people affected by tuberculosis as components of the mycobacterial cell envelope. The 1400-year-old mycolic acids are unique tuberculosis biomarkers and were extracted and identiied for the irst time using MALDI-time-of-light (TOF)MS. Figure 36.20 shows the mass spectrum of bone remains found in the cemetery of Felgyő-Ürmös Tanya (Hungary). The fragment ions at m/z 1290, 1318, and 1346 were ascribed to an oxygenated methoxy-mycolic acid methyl ester typical of Mycobacterium tuberculosis. In addition, the mass spectrum conirmed the presence of oxygenated ketomycolates at m/z 1246 and 1274, corresponding to mycolic acids with 82 and 84 carbon atoms, respectively. MALDI has also been used to identify the origins of four animal hair samples from the accoutrement of the 5300-year-old Tyrolean mummy, called Oetzi [174]. The analytical method involved the determination of proteins by analyzing peptides directly derived from tryptic hair digests without any separation or enrichment. Using this method, two samples obtained from Oetzi’s coat and a sample from his leggings were found to be made from sheep hide, while the upper leather of his moccasins was from cattle hide.

DIRECT MS TECHNIQUES: DE-MS, DTMS, DI-MS, ESI-MS AND ESI-MS/MS, MALDI-MS, AND LDI-MS

821

(A) x104

225

Intens. [a.u.]

2.0

1.5

239

1.0 211

0.5

255

195

263 285 265 295

0.0

200

(B)

250

482 375

300

350

445 461

400

m/z

450

235

5000

264

4000

335 397 369

248

Intens. [a.u.]

279

3000

219 353

2000

323 295

479

1000

522 636 608 664

0 200

300

400

500

600

935

700

800

900 m/z

FIGURE 36.19 LDI mass spectra of the sample collected from a Roman bronze tool. (A) Negative-ion mass spectrum. (B) Positive-ion mass spectrum [45]. Intens. [a.u.]

1318.3

600 1290.0 1274.3

400

200 1226.9

1346.4 1379.2 1424.7

1470.0 1515.1 1560.0 1604.8 1649.2

0 1200

1250

1300

1350

1400

1450 m/z

1500

1550

1600

1650

FIGURE 36.20 MALDI-TOF mass spectrum of a tuberculosis-infected archaeological bone sample (vertebra, Felgyő-Ürmös Tanya grave 205) using C60 fullerene as the matrix in the range of m/z 1200–1700 [173].

822

ARCHAEOMETRIC DATA FROM MASS SPECTROMETRIC ANALYSIS OF ORGANIC MATERIALS

REFERENCES 1. Mills, J.S., White, R. (1994) The Organic Chemistry of Museum Objects. Oxford: Butterworths-Heinemann, pp. 31–44. 2. Colombini, M.P., Modugno, F., eds. (2009) Organic Mass Spectrometry in Art and Archaeology. Chichester: John Wiley & Sons. 3. Stuart, B.H. (2007) Analytical Techniques in Material Conservation. Chichester: John Wiley & Sons. 4. Pollard, A.M., Heron, C. (1996) Archaeological Chemistry. Cambridge, UK: Royal Society of Chemistry. 5. Nicholson, P., Shaw, I., eds. (2000) Ancient Egyptian Materials and Technology. Cambridge, UK: Cambridge University Press. 6. Doménech-Carbò, M.T. (2008) Novel analytical methods for characterising binding media and protective coatings in artworks. Analytica Chimica Acta, 621, 109–139. 7. Evershed, R.P. (2008) Experimental approaches to the interpretation of absorbed organic residues in archaeological ceramics. World Archaeology, 40, 26–47. 8. Modugno, F., Ribechini, E., Colombini, M.P. (2006) Aromatic resin characterisation by gas chromatography– mass spectrometry: raw and archaeological materials. Journal of Chromatography A, 85, 164–173. 9. Stern, B., Connan, J., Blakelock, E., Jackman, R., Coningham, R.A.E., Heron, C. (2008) From Susa to Anuradhapura: reconstructing aspects of trade and exchange in bitumen-coated ceramic vessels between Iran and Sri Lanka from the third to the ninth centuries AD. Archaeometry, 50, 409–428. 10. Evershed, R.P., Dudd, S.N., Charters, S., Mottram, H., Stott, A.W., Raven, A., Van Bergen, P.F., Bland, H.A. (1999) Lipids as carriers of anthropogenic signals from prehistory. Philosophical Transactions of the Royal Society of London Series B, 354, 19–31. 11. Koller, J., Baumer, U. (2000) An investigation of the “Lacquers of the West”. A methodical survey. In Japanese and European Lacquerware, edited by Kuhlenthal, M. Munich: Karl M. Lipp Verlag, pp. 339–348. 12. Burger, P., Charrié-Duhaut, A., Connan, J., Flecker, M., Albrecht, P. (2009) Archaeological resinous samples from Asian wrecks: taxonomic characterization by GC–MS. Analytica Chimica Acta, 648, 85–97. 13. Andreotti, A., Bonaduce, I., Colombini, M.P., Gautier, G., Modugno, F., Ribechini, E. (2006) Combined GC/MS analytical procedure for the characterization of glycerolipid, waxy, resinous, and proteinaceous materials in a unique paint microsample. Analytical Chemistry, 78, 4490–4500. 14. Erhardt, D., Tumosa, C.S., Mecklenburg, M.F. (2005) Long-term chemical and physical processes in oil paint ilms. Studies in Conservation, 50, 143–150. 15. Charrié-Duhaut, A., Connan, J., Rouquette, N., Adam, D., Barbotin, C., de Rozières, M.F., Tchapla, A., Albrecht, P. (2007) The canopic jars of Rameses II: real use revealed

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

by molecular study of organic residues. Journal of Archaeological Science, 34, 957–967. Colombini, M.P., Modugno, F. (2004) Characterisation of proteinaceous binders in artistic paintings by chromatographic techniques. Journal of Separation Science, 27, 147–160. Gautier, G., Colombini, M.P. (2007) GC-MS identiication of proteins in wall painting samples: a fast clean-up procedure to remove copper-based pigment interferences. Talanta, 73, 95–102. Bonaduce, I., Cito, M., Colombini, M.P. (2009) The development of a gas chromatographic–mass spectrometric analytical procedure for the determination of lipids, proteins and resins in the same paint micro-sample avoiding interferences from inorganic media. Journal of Chromatography A, 1216, 5931–5939. Pallecchi, P., Giachi, G., Colombini, M.P., Modugno, F., Ribechini, R. (2009) The painting of the Etruscan “Tomba della Quadriga Infernale” (4th century BC), in Sarteano (Siena, Italy): technical features. Journal of Archaeological Science, 36, 2635–2642. Bonaduce, I., Blaensdorf, C., Dietemann, P., Colombini, M.P. (2008) The binding media of the polychromy of Qin Shihuang’s Terracotta Army. Journal of Cultural Heritage, 9, 103–108. Kuckova, S., Hynek, R., Kodicek, M. (2007) Identiication of proteinaceous binders used in art-works by MALDITOF mass spectrometry. Analytical and Bioanalytical Chemistry, 388, 201–206. Bonaduce, I., Brecoulaki, H., Colombini, M.P., Lluveras, A., Restivo,V., Ribechini, E. (2007) Gas chromatographic– mass spectrometric characterisation of plant gums in samples from painted works of art. Journal of Chromatography A, 1175, 275–282. Pitthard, V., Finch, P. (2001) GC–MS analysis of monosaccharides mixtures as their diethyldithioacetal derivatives: application to plant gums used in art works. Chromatographia, 53, 317–321. Tchapla, A., Mejanelle, P., Bleton, J., Goursaud, S. (2004) Characterisation of embalming materials of a mummy of the Ptolemaic era. Comparison with balms from mummies of different eras. Journal of Separation Science, 27, 217–234. Romanus, K., Van Neer, W., Marinova, E., Verbeke, K., Luypaerts, A., Accardo, S., Hermans, I., Jacobs, P., Vos, D.D., Waelkens, M. (2008) Brassicaceae seed oil identiied as illuminant in Nilotic shells from a irst millennium AD Coptic church in Bawit, Egypt. Analytical and Bioanalytical Chemistry, 390, 783–793. Colombini, M.P., Giachi, G., Modugno, F., Pallecchi, P., Ribechini, E. (2003) Characterisation of paints and waterprooing materials of the shipwrecks found in the archaeological site of the Etruscan and Roman Harbour of Pisa (Italy). Archaeometry, 45, 649–664. Ribechini, E., Modugno, F., Colombini, M.P., Evershed, R.P. (2008) Gas chromatographic and mass spectrometric investigations of organic residues from Roman glass

REFERENCES

28.

29. 30. 31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

unguentaria. Journal of Chromatography A, 1183, 158–169. Regert, M., Bland, H.A., Dudd, S.N., van Bergen, P.F., Evershed, R.P. (1998) Free and bound fatty acid oxidation products in archaeological ceramic vessels. Proceedings of the Royal Society of London B, 265, 2027–2032. Evershed, R.P. (1993) Biomolecular archaeology and lipids. World Archaeology, 25, 74–93. Mills, J.S., White, R. (1997) Natural resins in art and archaeology. Studies in Conservation, 22, 12–31. Heron, C., Nemeck, N., Bonield, K.M., Dixon, D., Ottaway, B.S. (1994) The chemistry of neolithic beeswax. Naturwissenschaften, 81, 266–269. Peris-Vicente, J., Gimeno Adelantado, J.V., Doménech Carbó, M.T., Mateo Castro, R., Bosch Reig, F. (2006) Characterization of waxes used in pictorial artworks according to their relative amount of fatty acids and hydrocarbons by gas chromatography. Journal of Chromatography A, 1101, 254–260. Evershed, R.P. (2008) Organic residues analysis in archaeology: the archaeological biomarker revolution. Archaeometry, 50, 895–924. Evershed, R.P., Dudd, S.N., Copley, M.S., Berstan, R., Stott, A.W., Mottram, H.R., Buckley, S.A., Crossman, Z. (2002) Chemistry of archaeological animal fats. Accounts of Chemical Research, 35, 660–668. Regert, M., Colinart, S., Degrand, L., Decavallas, O. (2001) Chemical alteration and use of beeswax through time: accelerated ageing tests and analysis of archaeological samples from various environmental contexts. Archaeometry, 43, 549–569. Hansel, F.A., Copley, M.S., Madureira, L.A., Evershed, R.P. (2004) Thermally produced ω-(o-alkylphenyl)alkanoic acids provide evidence for the processing of marine products in archaeological pottery vessels. Tetrahedron Letters, 45, 2999–3002. Craig, O.E., Forster, M., Andersen, S.H., Koch, E., Crombe, P., Milner, N.J., Stern, B., Bailey, G.N., Heron, C.P. (2007) Molecular and isotopic demonstration of the processing of aquatic products in Northern European prehistoric pottery. Archaeometry, 49, 135–152. Ribechini, E., Modugno, F., Baraldi, C., Baraldi, P., Colombini, M.P. (2008) An integrated analytical approach for characterizing an organic residues from an archaeological glass bottle recovered in Pompeii (Naples, Italy). Talanta, 74, 555–561. Ribechini, E., Colombini, M.P., Giachi, G., Modugno, F., Pallecchi, P. (2009) A multi-analytical approach for the characterization of commodities in a ceramic jar from Antinoe (Egypt). Archaeometry, 51, 480–494. Evershed, R.P., Heron, C., Goad, L.J. (1991) Epicuticular wax components preserved in potsherds as chemical indicators of leafy vegetables in ancient diets. Antiquity, 65, 540–544. Evershed, R.P., Payne, S., Sherratt, A.G., Copley, M.S., Coolidge, J., Urem-Kotsu, D., Kotsakis, K., Ozdogan, M.,

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

823

Ozdogan, A., Nieuwenhuyse, O., Akkermans, P.M., Bailey, D., Andeescu, R.-R., Campbell, S., Farid, S., Hodder, I., Yalman, N.O., Ozbaşaran, M., Bicakci, E., Garinkel, Y., Levy, T., Burton, M.M. (2008) Earliest date for milk use in the Near East and Southeastern Europe linked to cattle herding. Nature, 455, 528–531. Outram, A.K., Stear, N.A., Bendrey, R., Olsen, S., Kasparov, A., Zaibert, V., Thorpe, N., Evershed, R.P. (2009) The earliest horse harnessing and milking. Science, 323, 1333–1335. Evershed, R.P., Dudd, S.N., Anderson-Stojanovic, V.A., Gebhard, E.R. (2003) New chemical evidence for the use of combed ware pottery vessel as beehives in ancient Greece. Journal of Archaeological Science, 30, 1–12. Evershed, R.P., Berstan, R., Grew, F., Copley, M.S., Charmant, A.J.H., Barham, E., Mottram, H.R., Brown, G. (2004) Archaeology: formulation of a Roman cosmetic. Nature, 432, 35–36. Pérez-Arantegui, J., Ribechini, E., Cepriá, G., Degano, I., Colombini, M.P., Paz-Peralta, J., Ortiz-Palomar, E. (2009) Colorants and oils in Roman make-ups. Trends in Analytical Chemistry, 28, 1019–1028. Ribechini, E., Orsini, S., Silvano, F., Colombini, M.P. (2009) Py-GC/MS, GC/MS and FTIR investigations on late Roman-Egyptian adhesives from opus sectile: new insights into ancient recipes and technologies. Analytica Chimica Acta, 638, 79–87. Regert, M. (2004) Investigating the history of prehistoric glues by gas chromatography-mass spectrometry. Journal of Separation Science, 27, 244–254. Colombini, M.P., Modugno, F., Ribechini, E. (2005) Organic mass spectrometry in archaeology: evidence for Brassicaceae seed oil in Egyptian ceramic lamps. Journal of Mass Spectrometry, 40, 890–898. Copley, M.S., Bland, H.A., Rose, P., Horton, M., Evershed, R.P. (2005) Gas chromatographic, mass spectrometric and stable carbon isotopic investigations of organic residues of plant oils and animal fats employed as illuminants in archaeological lamps from Egypt. Analyst, 130, 860–871. Kimpe, K., Jacobs, P.A., Waelkens, M. (2001) Analysis of oil used in late Roman oil lamps with different mass spectrometric techniques revealed the presence of predominantly olive oil together with traces of animal fat. Journal of Chromatography A, 937, 87–95. Evershed, R.P., Vaughan, S.J., Dudd, S.N., Soles, J.S. (1997) Fuel for thought? Beeswax in lamps and conical cups from late Minoan Crete. Antiquity, 71, 979–985. Colombini, M.P., Giachi, G., Modugno, F., Ribechini, E. (2005) Characterisation of organic residues in pottery vessels of the Roman age from Antinoe (Egypt). Microchemical Journal, 79, 83–90. Buckley, S.A., Stott, A.W., Evershed, R.P. (1999) Studies of organic residues from ancient Egyptian mummies using high temperature-gas chromatography-mass spectrometry and sequential thermal desorption-gas chromatography-mass spectrometry and pyrolysis-gas

824

54.

55.

56.

57.

58.

59.

60. 61. 62.

63.

64.

65.

66.

67.

ARCHAEOMETRIC DATA FROM MASS SPECTROMETRIC ANALYSIS OF ORGANIC MATERIALS

chromatography-mass spectrometry. The Analyst, 124, 443–452. Buckley, S.A., Evershed, R.P. (2001) Organic chemistry of embalming agents in Pharaonic and Greco-Roman mummies. Nature, 413, 837–841. Koller, J., Baumer, U., Kaup, Y., Schmid, M., Weser, U. (2003) Analysis of a pharaonic embalming tar. Nature, 424, 784. Buckley, S.A., Clark, K.A., Evershed, R.P. (2004) Complex organic chemical balms of Pharaonic animal mummies. Nature, 431, 294–299. Papageorgopoulou, C., Xirotiris, N.I., Iten, P.X., Baumgartner, M.R., Schmid, M., Rühli, F. (2009) Indications of embalming in Roman Greece by physical, chemical and histological analysis. Journal of Archaeological Science, 36, 35–42. Gülaçar, F.O., Buchs, A., Susini, A. (1989) Capillary gas chromatography-mass spectrometry and identiication of substituted carboxylic acids in lipids extracted from a 4000-year–old Nubian burial. Journal of Chromatography, 479, 61–72. Makristathis, A., Schwarzmeier, J., Mader, R.M., Varmuza, K., Simonitsch, I., Chavez, J.C., Platzer, W., Unterdorfer, H., Scheithauer, R., Derevianko, A., Seidler, H. (2002) Fatty acid composition and preservation of the Tyrolean Iceman and other mummies. Journal of Lipid Residues, 43, 2056–2056. Evershed, R.P., Connolly, R.C. (1987) Lipid preservation in Lindow Man. Naturwissenschaften, 75, 143–145. Evershed, R.P. (1992) Chemical composition of bog body adipocere. Archaeometry, 34, 253–265. Varmuza, K., Makristathis, A., Schwarzmeier, J., Seidler, H., Mader, R.M. (2005) Exploration of anthropological specimens by GC-MS and chemometrics. Mass Spectrometry Reviews, 24, 427–452. Evershed, R.P., Heron, C., Goad, J. (1990) Analysis of organic residues of archaeological origin by high-temperature gas chromatography and gas chromatography-mass spectrometry. Analyst, 115, 1339–1342. Bonaduce, I., Colombini, M.P. (2004) Characterisation of beeswax in works of art by gas chromatography-mass spectrometry and pyrolysis-gas chromatography-mass spectrometry procedures. Journal of Chromatography A, 1028, 297–306. Regert, M., Langlois, J., Colinart, S. (2005) Characterisation of wax works of art by gas chromatographic procedures. Journal of Chromatography A, 1091, 124–136. Evershed, R.P. (2009) Compound-speciic stable isotopes in organic residue analysis in archaeology. In Organic Mass Spectrometry in Art and Archaeology, edited by Colombini, M.P., Modugno, F. Chichester: John Wiley & Sons, pp. 392–432. Evershed, R.P., Arnot, K.I., Collister, J., Eglinton, G., Charters, S. (1994) Application of isotope ratio monitoring gas chromatography-mass spectrometry to the analy-

sis of organic residues of archaeological origin. Analyst, 119, 909–914. 68. Tauber, H. (1981) 13C evidence for dietary habits of prehistoric man in Denmark. Nature, 292, 332–333. 69. Evershed, R.P., Mottram, H.R., Dudd, S.N., Charters, S., Stott, A.W., Gibson, A.M., Conner, A., Blinkhorn, P.W., Reeves, V. (1997) New criteria for the identiication of animal fats preserved in archaeological pottery. Naturwissenschaften, 84, 402–406. 70. Mottram, H.R., Dudd, S.N., Lawrence, G.J., Stott, A.W., Evershed, R.P. (1999) New chromatographic, mass spectrometric and stable isotope approaches to the classiication of degraded animal fats preserved in archaeological pottery. Journal of Chromatography A, 833, 209–221. 71. Dudd, S.N., Evershed, R.P. (1998) Direct demonstration of milk as an element of archaeological economies. Science, 282, 1478–1481. 72. Copley, M.S., Berstan, R., Dudd, S.N., Straker, V., Payne, S., Evershed, R.P. (2005) Dairying in antiquity. I. Evidence from absorbed lipid residues dating to the British Iron Age. Journal of Archaeological Science, 32, 485–503. 73. Mukherjee, A.J., Gibson, A.M., Evershed, R.P. (2008) Trends in pig product processing at British Neolithic Grooved Ware sites traced through organic residues in potsherds. Journal of Archaeological Science, 35, 2059– 2073. 74. Spangenberg, J.E., Jacomet, J., Schibler, J. (2006) Chemical analyses of organic residues in archaeological pottery from Arbon Bleiche 3, Switzerland—evidence for dairying in the late Neolithic. Journal of Archaeological Science, 33, 1–13. 75. Craig, O.E., Taylor, G., Collins, M., Parker Pearson, M. (2005) The identiication of prehistoric dairying activities in the Western Isles of Scotland: an integrated biomolecular approach. Journal of Archaeological Science, 32, 91–103. 76. Gregg, M.W., Banning, E.B., Gibbs, K., Slater, G.F. (2009) Subsistence practices and pottery use in Neolithic Jordan: molecular and isotopic evidence. Journal of Archaeological Science, 36, 937–946. 77. Jim, S., Ambrose, S.H., Evershed, R.P. (2004) Stable carbon isotopic evidence for differences in the dietary origin of bone cholesterol, collagen and apatite: implications for their use in palaeodietary reconstruction. Geochimica et Cosmochimica Acta, 68, 61–72. 78. Fogel, M.L., Tuross, N. (2003) Extending the limits of palaeodietary studies of humans with compound speciic carbon isotope analysis of amino acids. Journal of Archaeological Science, 30, 535–545. 79. Corr, L.T., Richards, M.P., Jim, S., Ambrose, S.H., Mackie, A., Beattie, O., Evershed, R.P. (2008) Probing dietary change of the Kwädąy Dän Ts’ìnchį individual, an ancient glacier body from British Columbia: I. Complementary use of marine lipid biomarker and carbon isotope signatures as novel indicators of a marine diet. Journal of Archaeological Science, 35, 2102–2110.

REFERENCES

80. Corr, L.T., Richards, M.P., Grier, C., Mackie, A., Beattie, O., Evershed, R.P. (2009) Probing dietary change of the Kwädąy Dän Ts’ìnchį individual, an ancient glacier body from British Columbia: II. Deconvoluting whole skin and bone collagen δ13C values via carbon isotope analysis of individual amino acids. Journal of Archaeological Science, 36, 12–18. 81. Evershed, R.P., van Bergen, P.F., Peakman, T.M., LeighFirbank, E.C., Horton, M.C., Edwards, D., Biddle, M., Kjolbye-Biddle, B., Rowley-Conwy, P.A. (1997) Archaeological frankincense. Nature, 390, 667–668. 82. Stern, B., Heron, C., Corr, L., Serpico, M., Bourriau, J. (2003) Compositional variations in aged and heated Pistacia resin found in late Bronze Age Canaanite amphorae and bowls from Amarna, Egypt. Archaeometry, 45, 457–469. 83. Mathe Culioli, C.G., Archier, P., Vieillescazes, C. (2004) Characterization of archaeological frankincense by gas chromatography-mass spectrometry. Journal of Chromatography A, 1023, 277–285. 84. Colombini, M.P., Modugno, F., Ribechini, E. (2006) Chemical study of triterpenoid resinous materials in archaeological indings by means of direct exposure electron ionisation mass spectrometry and gas chromatography/mass spectrometry. Rapid Communications in Mass Spectrometry, 20, 1787–1800. 85. Regert, M., Vacher, S., Moulherat, C., Decavallas, O. (2003) Adhesive production and pottery function during iron age at the site of Grand Aunay (Sarthe, France). Archaeometry, 45, 101–120. 86. Baumer, U., Dietemann, P., Koller, J. (2009) Identiication of resinous materials on 16th and 17th century reverseglass objects by gas chromatography/mass spectrometry. International Journal of Mass Spectrometry, 284, 131–141. 87. Regert, M., Alexandre, V., Thomas, N., LattuatiDerieux, A. (2006) Molecular characterisation of birch bark tar by headspace solid-phase microextraction gas chromatography–mass spectrometry: a new way for identifying archaeological glues. Journal of Chromatography A, 1101, 245–253. 88. Hamm, S., Bleton, J., Tchapla, A. (2004) Headspace solid phase microextraction for screening for the presence of resins in Egyptian archaeological samples. Journal of Separation Science, 27, 235–243. 89. Hamm, S., Bleton, J., Connan, J., Tchapla, A. (2005) A chemical investigation by headspace SPME and GC–MS of volatile and semi-volatile terpenes in various olibanum samples. Phytochemistry, 66, 1499–1514. 90. Robinson, N., Evershed, R.P., Higgs, W.J., Jerman, K., Eglinton, G. (1987) Proof of a pine wood origin for pitch from Tudor (Mary Rose) and Etruscan shipwrecks: application of analytical organic chemistry in archaeology. Analyst, 112, 637–643. 91. Colombini, M.P., Modugno, F., Ribechini, E. (2005) Direct exposure electron ionization mass spectrometry and gas chromatography/mass spectrometry techniques to study

92.

93.

94. 95.

96.

97.

98.

99.

100.

101.

102.

103.

825

organic coatings on archaeological amphorae. Journal of Mass Spectrometry, 40, 675–687. Charrié-Duhaut, A., Connan, J., Darnell, M., Spangenberg, J., Szymczyk, E., Bissada, A., Albrecht, P. (2009) Molecular and isotopic characterization of organic samples from the wreck of the Saint-Etienne merchant ship (XVIIIth century): identiication of pitch, fat, hair and sulphur. Organic Geochemistry, 40, 647–665. Hayek, E.W.H., Krenmayr, P., Lohninger, H., Jordis, U., Moche, W., Sauter, F. (1990) Identiication of archaeological and recent wood tar pitches using gas chromatography/ mass spectrometry and pattern recognition. Analytical Chemistry, 62, 2038–2043. Grunberg, J.M. (2002) Middle palaeolithic birch-bark pitch. Antiquity, 76, 15–16. Shedrinsky, A.M., Wampler, T.P., Indictor, N., Baer, N.S. (1989) Application of analytical pyrolysis to problems in art and archaeology: a review. Journal of Analytical and Applied Pyrolysis, 15, 393–412. Chiavari, G., Ferretti, S., Galletti, G.C., Mazzeo, R. (1991) Analytical pyrolysis as a tool for the characterization of organic substances in artistic and archaeological objects. Journal of Analytical and Applied Pyrolysis, 20, 253–261. Wright, M.M., Wheals, B.B. (1987) Pyrolysis mass spectrometry of natural gums, resins, and waxes and its use for detecting such materials in ancient Egyptian mummy cases (cartonnages). Journal of Analytical and Applied Pyrolysis, 11, 195–211. Van den Berg, K.J., Boon, J.J., Pastorova, I., Spetter, L.F.M. (2000) Mass spectrometric methodology for the analysis of highly oxidized diterpenoid acids in Old Master paintings. Journal of Mass Spectrometry, 35, 512–533. Chiavari, G., Fabbri, D., Prati, S. (2002) Characterisation of natural resins by pyrolysis-silylation. Chromatographia, 55, 611–616. Shedrinsky, A.M., Grimaldi, D.A., Boon, J.J., Baer, N.S. (1993) Application of pyrolysis-gas chromatography and pyrolysis-gas chromatography/mass spectrometry to the unmasking of amber forgeries. Journal of Analytical and Applied Pyrolysis, 25, 77–95. de la Cruz-Canizares, J., Domenech-Carbo, M.T., Gimeno-Adelantado, J.V., Mateo-Castro, R., BoschReig, F. (2005) Study of Burseraceae resins used in binding media and varnishes from artworks by gas chromatography-mass spectrometry and pyrolysisgas chromatography-mass spectrometry. Journal of Chromatography A, 1093, 177–194. Osete-Cortina, L., Domenech-Carbo, M.T. (2005) Analytical characterization of diterpenoid resins present in pictorial varnishes using pyrolysis-gas chromatographymass spectrometry with on line trimethylsilylation. Journal of Chromatography A, 1065, 265–278. Scalarone, D., Lazzari, M., Chiantore, O. (2002) Pyrolytic characterisation and ageing behaviour of diterpenic resins used as art materials: colophony and Venice

826

104.

105.

106.

107.

108.

109.

110.

111.

112.

113.

114.

115.

116.

ARCHAEOMETRIC DATA FROM MASS SPECTROMETRIC ANALYSIS OF ORGANIC MATERIALS

turpentine. Journal of Analytical and Applied Pyrolysis, 64, 345–361. Scalarone, D., Lazzari, M., Chiantore, O. (2003) Ageing behaviour and analytical pyrolysis characterisation of diterpenic resins used as art materials: Manila copal and sandarac. Journal of Analytical and Applied Pyrolysis, 68–69, 115–136. Colombini, M.P., Bonaduce, I., Gautier, G. (2003) Molecular pattern recognition of fresh and aged, shellac. Chromatographia, 58, 357–364. Raven, A.M., van Bergen, P.F., Stott, A.W., Dudd, S.N., Evershed, R.P. (1997) Formation of long-chain ketones in archaeological pottery vessels by pyrolysis of acyl lipids. Journal of Analytical and Applied Pyrolysis, 40–41, 267–285. Cappitelli, F., Learner, T., Chiantore, O. (2002) An initial assessment of thermally assisted hydrolysis and methylation-gas chromatography/mass spectrometry for the identiication of oils from dried paint ilms. Journal of Analytical and Applied Pyrolysis, 63, 339–348. Mazzeo, R., Cam, D., Chiavari, G., Fabbri, D., Ling, H., Prati, S. (2004) Analytical study of traditional decorative materials and techniques used in Ming Dynasty wooden architecture. The case of the Drum Tower in Xi’an, P.R. of China. Journal of Cultural Heritage, 5, 273–283. Chiavari, G., Fabbri, D., Prati, S. (2001) In-situ pyrolysis and sylilation for analysis of lipid materials used in paint layers. Chromatographia, 53, 311–314. Oudemans, T.F.M., Boon, J.J. (1991) Molecular archaeology: analysis of charred (food) remains from prehistoric pottery by pyrolysis-gas chromatography/mass spectrometry. Journal of Analytical and Applied Pyrolysis, 20, 197–227. Chiavari, G., Fabbri, D., Galletti, G.C., Mazzeo, R. (1995) Use of analytical pyrolysis to characterize Egyptian painting layers. Chromatographia, 40, 594–600. Asperger, A., Engewald, W., Fabian, G. (2001) Thermally assisted hydrolysis and methylation—a simple and rapid on line derivatization method for gas chromatographic analysis of natural waxes. Journal of Analytical and Applied Pyrolysis, 61, 91–109. Chiavari, G., Galletti, G.C., Lanterna, G., Mazzeo, R. (1993) The potential of pyrolysis gas chromatography mass spectrometry in the recognition of ancient painting media. Journal of Analytical and Applied Pyrolysis, 24, 227–242. Chiavari, G., Lanterna, G., Luca, C., Matteini, M., Prati, S., Sandu, I.C.A. (2003) Analysis of proteinaceous binders by in-situ pyrolysis and silylation. Chromatographia, 57, 645–648. Bonaduce, I., Colombini, M.P. (2003) Gas chromatography/ mass spectrometry for the characterization of organic materials in frescoes of the Monumental Cemetery of Pisa (Italy). Rapid Communications in Mass Spectrometry, 17, 2523–2527. Casas-Catalán, M.J., Doménech-Carbó, M.T. (2005) Identiication of natural dyes used in works of art by

117.

118.

119.

120.

121.

122.

123.

124.

125.

126.

127.

128.

129.

pyrolysis–gas chromatography/mass spectrometry combined with in situ trimethylsilylation. Analytical and Bioanalytical Chemistry, 382, 259–268. Fabbri, D., Chiavari, G., Ling, H. (2000) Analysis of anthraquinoid and indigoid dyes used in ancient artistic works by thermally assisted hydrolysis and methylation in the presence of tetramethylammonium hydroxide. Journal of Analytical and Applied Pyrolysis, 56, 167–178. Garnier, N., Richardin, P., Cheynier, V., Regert, M. (2003) Characterization of thermally assisted hydrolysis and methylation products of polyphenols from modern and archaeological vine derivatives using gas chromatography– mass spectrometry. Analytica Chimica Acta, 493, 137–157. Chiavari, G., Montalbani, S., Prati, S., Keheyan, Y., Baroni, S. (2007) Application of analytical pyrolysis for the characterisation of old inks. Journal of Analytical and Applied Pyrolysis, 80, 400–405. Fabbri, D., Chiavari, G., Prati, S., Vassura, I., Vangelista, M. (2002) Gas chromatography/mass spectrometric characterisation of pyrolysis/silylation products of glucose and cellulose. Rapid Communications in Mass Spectrometry, 16, 2349–2355. van Bergen, P.F., Poole, I., Ogilvie, T.M., Caple, C., Evershed, R.P. (2000) Evidence for demethylation of syringyl moieties in archaeological wood using pyrolysisgas chromatography/mass spectrometry. Rapid Communications in Mass Spectrometry, 14, 71–79. Łucejko, J.J., Modugno, F., Ribechini, E., del Río, J.C. (2009) Characterisation of archaeological waterlogged wood by pyrolytic and mass spectrometric techniques. Analytica Chimica Acta, 654, 26–34. Colombini, M.P., Lucejko, J.J., Modugno, F., Orlandi, M., E.- L.Tolppa, L.Z. (2009) A multi-analytical study of degradation of lignin in archaeological waterlogged wood. Talanta, 80, 61–70. Sobeih, K.L., Baron, M., Gonzalez-Rodriguez, J. (2008) Recent trends and developments in pyrolysis–gas chromatography. Journal of Chromatography A, 1186, 51–66. Bocchini, P., Traldi, P. (1998) Organic mass spectrometry in our cultural heritage. Journal of Mass Spectrometry, 33, 1053–1062. Chiavari, G., Prati, S. (2003) Analytical pyrolysis as diagnostic tool in the investigation of works of art. Chromatographia, 58, 543–554. Prati, S., Smith, S., Chiavari, G. (2004) Characterization of siccative oils, resins and pigments in art works by thermochemolysis coupled to thermal desorption and pyrolysis GC and GC/MS. Chromatographia, 59, 227–231. Blanchette, R.A. (2000) A review of microbial deterioration found in archaeological wood from different environments. International Biodeterioration & Biodegradation, 46, 189–204. Rowell, R.M., Barbour, R.J. (1990) Archaeological Wood Properties, Chemistry, and Preservation. Washington, DC: American Chemical Society.

REFERENCES

130. Saiz-Jimenez, C., Boon, J.J., Hedges, J.I., Hessels, J.K.C., De Leeuw, J.W. (1987) Chemical characterization of recent and buried woods by analytical pyrolysis: comparison of pyrolysis data with 13C NMR and wet chemical data. Journal of Analytical and Applied Pyrolysis, 11, 437–450. 131. Yang, H., Huang, Y., Leng, Q., Le Page, B.A., Williams, C.J. (2005) Biomolecular preservation of Tertiary Metasequoia Fossil Lagerstätten revealed by comparative pyrolysis analysis. Review of Palaeobotany and Palynology, 134, 237–256. 132. Saiz-Jimenez, C., de Leeuw, J.W. (1986) Lignin pyrolysis products: their structures and their signiicance as biomarkers. Organic Geochemistry, 10, 869–876. 133. Giachi, G., Bettazzi, F., Chimichi, S., Staccioli, G. (2003) Chemical characterisation of degraded wood in ships discovered in a recent excavation of the Etruscan and Roman harbour of Pisa. Journal of Cultural Heritage, 4, 75–83. 134. Rie, E.R., Watts, S. (2002) GCMS analysis of terpenoid resins: in situ derivatization procedures using quaternary ammonium hydroxides. Studies in Conservation, 47, 257–272. 135. Challinor, J.M. (2001) Review: the development and application of thermally assisted hydrolysis and methylation reactions. Journal of Analytical and Applied Pyrolysis, 61, 3–34. 136. Pastorova, I., van den Berg, K.J., Boon, J.J., Verhoeven, J.W. (1997) Analysis of oxidized diterpenoid acids using thermally assisted methylation with TMAH. Journal of Analytical and Applied Pyrolysis, 43, 41–57. 137. Challinor, J.M. (1996) A rapid simple pyrolysis derivatisation gas chromatography-mass spectrometry method for proiling of fatty acids in trace quantities of lipids. Journal of Analytical and Applied Pyrolysis, 37, 185–197. 138. Riedo, C., Scalarone, D., Chiantore, O. (2010) Advances in identiication of plant gums in cultural heritage by thermally assisted hydrolysis and methylation. Analytical and Bioanalytical Chemistry, 396, 1559–1569. 139. Joll, A., Huynh, T., Heitz, A. (2003) Off-line tetramethylammonium hydroxide thermochemolysis of model compound aliphatic and aromatic carboxylic acids: decarboxylation of some ortho- and/or para-substituted aromatic carboxylic acids. Journal of Analytical and Applied Pyrolysis, 70, 151–167. 140. van den Berg, J.D.J., Boon, J.J. (2001) Unwanted alkylation during direct methylation of fatty (di)acids using tetramethylammonium hydroxide reagent in a Curiepoint pyrolysis unit. Journal of Analytical and Applied Pyrolysis, 61, 45–63. 141. Ishida, Y., Wakamatsu, S., Yokoi, H., Ontani, H., Tsurge, S. (1999) Compositional analysis of polyunsaturated fatty acid oil by one-step thermally assisted hydrolysis and methylation in the presence of trimethylsulfonium hydroxide. Journal of Analytical and Applied Pyrolysis, 49, 267–276.

827

142. Hardell, H.L., Nilvebrant, N.O. (1999) A rapid method to discriminate between free and esteriied fatty acids by pyrolytic methylation using tetramethylammonium acetate or hydroxide. Journal of Analytical and Applied Pyrolysis, 32, 1–14. 143. Piccirillo, A., Scalarone, D., Chiantore, O. Comparison between off-line and on-line derivatisation methods in the characterisation of siccative oils in paint media. Journal of Analytical and Applied Pyrolysis, 74, 2005, 33–38. 144. Kuroda, K.I. (2000) Pyrolysis-trimethylsilylation analysis of lignin preferential formation of cinnamyl alcohol derivatives. Journal of Analytical and Applied Pyrolysis, 56, 79–87. 145. Chiantore, O., Riedo, C., Scalarone, D. (2009) Gas chromatography–mass spectrometric analysis of products from on-line pyrolysis/silylation of plant gums used as binding media. International Journal of Mass Spectrometry, 284, 35–41. 146. Solazzo, C., Fitzhugh, W.W., Rolando, C., Tokarski, C. (2008) Identiication of protein remains in archaeological potsherds by proteomics. Analytical Chemistry, 80, 4590–4597. 147. Guasch-Janes, M.R., Ibern-Gomez, M., Andres-Lacueva, C., Jauregui, O., Lamuela-Raventos, R.M. (2004) Liquid chromatography with mass spectrometry in tandem mode applied for the identiication of wine markers in residues from ancient Egyptian vessels. Analytical Chemistry, 76, 1672–1677. 148. W.J. Hurst, S.M. Tarka, T.G. Powis, F. Valdez, T.R. Hester, (2002) Archaeology: cacao usage by the earliest Maya civilization. Nature, 418, 289–290. 149. Balakina, G.G., Vasiliev, V.G., Karpova, E.V., Mamatyuk, V.I. (2006) HPLC and molecular spectroscopic investigations of the red dye obtained from an ancient Pazyryk textile. Dyes and Pigments, 71, 54–60. 150. Zhang, X., Good, I., Laursen R. (2008) Characterization of dyestuffs in ancient textiles from Xinjiang. Journal of Archaeological Science, 35, 1095–1103. 151. Zhang, X., Boytner, R., Cabrera, J.L., Laursen, R. (2007) Identiication of yellow dye types in pre-Columbian Andean textiles. Analytical Chemistry, 79, 1575–1582. 152. Zhang, X., Laursen, R. (2009) Application of LC–MS to the analysis of dyes in objects of historical interest. International Journal of Mass Spectrometry, 284, 108–114. 153. Karapanagiotis, I. (2006) identiication of indigoid natural dyestuffs used in art objects by HPLC coupled to APCIMS. American Laboratory, 38, 36–40. 154. Karapanagiotis, I., Chryssoulakis, Y. (2006) Investigation of red natural dyes used in historical objects by HPLCDAD-MS. Annali di Chimica, 96, 75–84. 155. Lech, K., Polec′-Pawlak, K., Jarosz, M. (2009) Characterization of organic natural dyes by electrospray mass spectrometry coupled with HPLC and/or capillary electrophoresis. In Organic Mass Spectrometry in Art and

828

156.

157.

158.

159.

160.

161.

162.

163.

164.

165.

ARCHAEOMETRIC DATA FROM MASS SPECTROMETRIC ANALYSIS OF ORGANIC MATERIALS

Archaeology, edited by Colombini, M.P., Modugno, F. Chichester, UK: John Wiley & Sons, pp. 365–388. Degano, I., Ribechini, E., Modugno, F., Colombini, M.P. (2009) Analytical methods for the characterisation of organic dyes in artworks and in historical textiles. Applied Spectroscopy Reviews, 44, 363–410. Rosenberg, E. (2008) Characterisation of historical organic dyestuffs by liquid chromatography–mass spectrometry. Analytical and Bioanalytical Chemistry, 391, 33–57. Zhang, X., Laursen, R.A. (2005) Development of mild extraction methods for the analysis of natural dyes in textiles of historical interest using LC-diode array detector-MS. Analytical Chemistry, 77, 2022–2025. Kimpe, K., Jacobs, P.A., Waelkens, M. (2002) Mass spectrometric methods prove the use of beeswax and ruminant fat in late Roman cooking pots. Journal of Chromatography A, 968, 151–160. Mottram, H.R., Evershed, R.P. (1996) Structure analysis of triacylglycerol positional isomers using atmospheric pressure chemical ionisation mass spectrometry. Tetrahedron Letters, 37, 8593–8596. Ribechini, E., Modugno, F., Colombini, M.P. (2008) Direct exposure-(chemical ionization)-mass spectrometry for a rapid characterization of raw and archaeological diterpenoid resinous substances. Microchimica Acta, 162, 405–413. Modugno, F., Ribechini, E., Calderisi, M., Giachi, G., Colombini, M.P. (2008) Analysis of lignin from archaeological waterlogged wood by direct exposure mass spectrometry (DE-MS) and PCA evaluation of mass spectral data. Microchemical Journal, 88, 186–193. Deviese, T., Ribechini, E., Baraldi, P., Farago-Szekeres, B., Duday, H., Regert, M., Colombini, M.P. (2011) First chemical evidence of royal purple as a material used for funeral treatment discovered in a Gallo–Roman burial (Naintré, France, third century AD). Analytical and Bioanalytical Chemistry, 40, 1739–1748. Oudemans, T.F.M., Eijkel, G.B., Boon, J.J. (2007) Identifying biomolecular origins of solid organic residues preserved in Iron Age pottery using DTMS and MVA. Journal of Archaeological Science, 34, 173–193. Regert, M., Rolando, C. (2002) Identiication of archaeological adhesives using direct inlet electron ionization mass spectrometry. Analytical Chemistry, 74, 965– 975.

166. Ribechini, E. (2009) Direct mass spectrometric techniques: versatile tools to characterise resinous materials. In Organic Mass Spectrometry in Art and Archaeology, edited by Colombini, M.P., Modugno, F. Chichester, UK: John Wiley & Sons, pp. 77–95. 167. Regert, M. (2009) Direct mass spectrometry to characterize wax and lipid. In Organic Mass Spectrometry in Art and Archaeology, edited by Colombini, M.P., Modugno, F. Chichester, UK: John Wiley & Sons, pp. 97–129. 168. Garnier, N., Cren-Olivé, C., Rolando, C., Regert, M. (2002) characterization of archaeological beeswax by electron ionization and electrospray ionization mass spectrometry. Analytical Chemistry, 74, 4868–4877. 169. Garnier, N., Rolando, C., Høtje, J.M., Tokarski, C. (2009) Analysis of archaeological triacylglycerols by high resolution nanoESI, FT-ICR MS and IRMPD MS/MS: application to 5th century BC–4th century AD oil lamps from Olbia (Ukraine). International Journal of Mass Spectrometry, 284, 47–56. 170. Mirabaud, S., Rolando, C., Regert, M. (2007) Molecular criteria for discriminating adipose fat and milk from different species by nanoESI MS and MS/MS of their triacylglycerols: application to archaeological remains. Analytical Chemistry, 79, 6182–6192. 171. Van Elslande, E., Guérineau, V., Thirioux, V., Richard, G., Richardin, P., Laprévote, O., Hussler, G., Walter, P. (2008) Analysis of ancient Greco–Roman cosmetic materials using laser desorption ionization and electrospray ionization mass spectrometry. Analytical and Bioanalytical Chemistry, 390, 1873–1879. 172. Buckley, M., Whitcher Kansa, S., Howard, S., Campbell, S., Thomas-Oates, J., Collins, M. (2010) Distinguishing between archaeological sheep and goat bones using a single collagen peptide. Journal of Archaeological Science, 37, 13–20. 173. Mark, L., Patonai, Z., Vaczy, A., Lorand, T., Marcsik, A. (2010) High-throughput mass spectrometric analysis of 1400-year-old mycolic acids as biomarkers for ancient tuberculosis infection. Journal of Archaeological Science, 37, 302–305. 174. Hollemeyer, K., Altmeyer, W., Heinzle, E., Pitra, C. (2008) Species identiication of Oetzi’s clothing with matrixassisted laser desorption/ionization time-of-light mass spectrometry based on peptide pattern similarities of hair digests. Rapid Communications in Mass Spectrometry, 22, 2751–2767.

37 LASER ABLATION ICP-MS IN ARCHAEOLOGY Hector Neff

With few exceptions, the most abundant components of the archaeological record are inorganic solids, which typically include ceramics, lithics, metals, glass, and mineralized animal tissue (bone and teeth). Archaeologists, whose job is to describe and account for this record in terms of past events and processes, have long sought to develop and/or adapt techniques for characterizing such materials. Questions that can be addressed via inorganic material characterization range from diet to dating (i.e., U-series dating and dosimetry in luminescence dating). Perhaps the best-known kinds of questions relate to provenance. By linking artifacts to geographically restricted source materials, provenance determination leads directly to inference of mobility patterns and economic interaction patterns of prehistoric populations [1–6]. Elemental characterization also addresses technological evolution, serving to elucidate how metalalloying, glaze recipes, or other technological practices changed in response to changing environmental and social conditions [7]. To the extent that climate-change processes may leave an elemental signature in sediments [8–11], speleothem growth rings [12,13], or tree rings [14], solid-sample analysis also reveals the environmental changes that affect human adaptation. Finally, analysis of mineralized animal tissue can be used to infer human diet via trace element analysis [15,16], human population movements via strontium isotope ratio analysis [17–20], or trace element analysis [21,22], or to shed light on human subsistence via the isotopic composition of nonhuman animal bone [23,24].

Three basic approaches to characterizing inorganic solids may be discerned. First, techniques based on inductively coupled plasma (ICP) including emission spectroscopy (ICP-AES) and mass spectrometry (ICP-MS) generally require digestion of solid samples with acids and heat, so that they can be introduced to the ICP torch as liquids. Second, techniques that depend on emissions of electromagnetic radiation of higher energy, including X-ray luorescence (XRF), electron beam-induced XRF in scanning electron microscopes (SEMs) or microprobes, proton-induced X-ray emission (PIXE) spectroscopy, and instrumental neutron activation analysis (INAA), are inherently solid-sample techniques: the solid is exposed to electrons, X-rays, or neutrons, and the induced emission of electromagnetic radiation at energies in the X-ray or gamma-ray region is used to characterize the sample. Lastly, laser ablation (LA) of a solid sample may be used to vaporize material from the solid that can be analyzed by one of the ICP-based techniques. This chapter focuses on the last of the above-listed techniques, laser ablation-ICP-MS (LA-ICP-MS). A virtual explosion of recent applications (199 papers mentioning LA-ICP-MS and archaeology since 2008 according to a Google Scholar search in December 2009) would seem to indicate that the time is right for an overview. Rather than attempt a comprehensive summary, a sample of the range of archaeological applications is presented, and some examples of recent successes and future potential are described. For a more thorough summary of LA-ICP-MS applications in archaeology, art history, and other cultural heritage ields, the reader should consult a recent review [25].

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

829

830

37.1

LASER ABLATION ICP-MS IN ARCHAEOLOGY

BACKGROUND AND CONTEXT

Since LA-ICP-MS is a relatively recent innovation in solid-sample characterization, archaeological consumers will want to make comparisons to older, betterestablished techniques. All techniques are strong in some respects and weaker in others, but a case can be made that INAA has become the “technique of choice” in many archaeological applications [26–28]. INAA is a true bulk technique that does not require digestion. INAA is highly sensitive and precise, data collection and reduction are automated, and many elements can be measured simultaneously [1,29]. These qualities have helped establish a strong track record for INAA over the past 40 years, since the early pioneering applications of “standard comparator” INAA at Lawrence Berkeley [30,31] and Brookhaven National Laboratory [32,33] [34]. Data generated by INAA are extremely robust in the face of changes in gamma-spectroscopy instruments, varying multielement standards, and even changing irradiation and decay conditions, so that data generated in different laboratories are fully compatible [35]. Literally hundreds of articles, monographs, and edited volumes have been published, especially during the past two decades that report INAA-based provenance studies of obsidian [36], chert [37], ceramics [38], metal [39], and other artifact classes. This track record makes INAA the benchmark against which other elemental characterization techniques must be evaluated. While INAA has an excellent track record in archaeology, bulk techniques like INAA have inherent limitations. In bulk analysis, powdered, homogenized whole samples are characterized, so contributions from individual components of a composite material, such as a ceramic, cannot be separated. One reason to characterize individual components in ceramics is that patterned elemental variation may arise not only from provenance differences, but also from paste preparation and diagenesis [5]. Microprobe techniques, either electron microprobes [5] or LA-ICP-MS [1,40], offer a means to identify where within the ceramic fabric the important elements are concentrated or diluted. Microprobes can also be targeted at ceramic slips, glazes, and pigments, as discussed in greater detail below. The broad range of elements to which ICP-MS is sensitive is another advantage for archaeology-based studies. Some important elements found in metal alloys, glass colorants, ceramic paints, and/or glazes cannot be measured by INAA (i.e., Pb) or require nonstandard analytical procedures (i.e., Cu, Ag, Sn) due to nuclear properties or matrix effects. ICP-MS also permits isotope ratio characterization (albeit with limitations, as discussed below), which can be as or more useful than elemental abundances in provenance investigations of

metals, glass, and glazes [41], in population movement studies based on strontium isotopes in tooth enamel [42,43] and other applications.

37.2

LA-ICP-MS ELEMENTAL ANALYSIS

Archaeologists began to use ICP-MS in the early or mid-1990s. In 1996, for instance, a textbook on archaeological chemistry [44] contained a description of quadrupole mass spectrometers along with a discussion of the potential—unrealized in archaeology as of 1996—of LA-ICP-MS. ICP mass spectrometers exploit the fact that hightemperature argon plasmas, which were irst used in light-emission spectrometry, very eficiently ionize atoms in a sample, so that different atomic masses (more precisely, different mass-to-charge ratios) can be measured. Generally, samples are introduced to the plasma as liquids. Therefore, solid samples must be digested with heat and strong acids, a step that is time-consuming, unpleasant, and potentially dangerous. Archaeological ceramics and most lithic materials, for example, require hydroluoric acid digestion [45,46]. In addition to being inconvenient, the digestion step also enhances the potential for subtle random and systematic effects to attenuate the stability of analytical results, especially over the long term. LA is an alternative to digesting samples for ICP-MS analysis. Not only does LA eliminate the need for sample digestion, it also turns ICP-MS instruments into extremely sensitive microprobes capable of solid-sample determination of the vast majority of elements in the periodic table at concentrations below 1 ppm. Currently, available LA systems such as the New Wave UP-213 can ablate spots as small as 10-µm diameter. A typical LA-ICP-MS setup (Figure 37.1) has a 213nm Nd-YAG laser either with argon or helium lowing through the sample chamber, coupled to a quadrupole ICP-MS. The sample chamber on commercially available lasers admits samples up to about 5-cm diameter; usually, 10–50 smaller samples are mounted on a slide together with reference standards, and the slide is then placed in the chamber. Software controls x-y-z movement of the ablation chamber and projects a video image of the inside of the chamber. Ablation patterns (lines, rasters, spots, or lines of spots) are drawn on the samples, and characteristics of the ablation (spot size, scan speed, laser power, repetition rate, number of passes or dwell time) are speciied. When the ablation starts, the material vaporized from the sample and ejected from the site of ablation is entrained in the helium stream lowing through the chamber. Outside of the sample chamber, the ablated aerosol joins the argon

LA-ICP-MS ELEMENTAL ANALYSIS

831

Ar

He

Argon Sample Gas Mass Spectrometer

Laser ICP torch He + Sample Aerosol

Computer

Interface

Computer

FIGURE 37.1 Diagram of a typical laser ablation-ICP-MS setup.

gas stream and eventually lows into the argon plasma of the ICP torch. The sample aerosol is sometimes passed through a spray chamber prior to analysis [47], but most often, it is injected directly into the plasma. Ions from the plasma then pass through the interface into the mass spectrometer. Characterization of a sample usually starts after a pre-ablation pass, which removes surface contamination, and following a short period to allow the sample to reach the ICP torch. Signal intensities for the analytes of interest are monitored, as in normal ICP-MS applications, and the resulting data saved to a ile on the ICP-MS computer. Because calibration, discussed further below, requires ablation of a series of solid, multielement reference standards, it is generally impossible to have the ICP-MS software calculate standard curves and elemental concentrations. Thus, raw signal intensities are usually downloaded and processed off-line. Quantiication of elemental concentrations in ICPMS requires prior knowledge of one or more internal standards in the samples to be characterized. In solution, ICP-MS internal standards that are known not to be present in the samples are added in known concentrations. This approach can be used in LA-ICP-MS as well, if bulk characterization is the goal. Heterogeneous samples, such as pottery pastes or sediments, are irst ground and homogenized to a ine powder, which is then spiked with an indium internal standard solution. When the standard solution dries, the sample is homogenized again and then pressed into a wafer either using an XRF binder or a high-purity graphite powder. If samples are to be analyzed without homogenization, then the concentration of an internal standard can

sometimes be assumed. Obsidian, for instance, is rhyolitic volcanic glass, in which silicon concentrations remain very close to 35%. Thus, silicon, monitored at the minor isotope mass-30 because of the high concentrations, can be used as the internal standard, with 35% as the assumed concentration. Another approach to internal standardization involves independent determination of the concentrations of an appropriate analyte, for example, by electron microprobe or XRF. In the case of the sediment-sample pressed wafers discussed above, for instance, calcium, titanium, iron, strontium, and other major and minor elements with moderate abundance in sediments can be determined easily by XRF, and then used as internal standards, providing the reference standards that also have those elements. If internal standard analytes are known by one of the above means (i.e., because they have been spiked, can be assumed, or have been independently determined), then signal intensities can be calibrated to elemental concentrations as follows. After subtracting the background counts for each mass, the standardized signal for element y can be deined as SSy =

Signal y Signalis

(subscript “is” indicates internal standard). For the reference standards, in which all elements are known, standardized concentrations for all elements are deined as SC y =

Conc y . Concis

832

LASER ABLATION ICP-MS IN ARCHAEOLOGY

Using the reference standards, linear regression of SCy on SSy yields parameters (slope, K, and intercept, b) for calculating SCy in the unknowns. Parameters K and b from the regression can then be used to calculate SCy in the unknowns by SC y = K (SSy ) + b. And, since SCy is deined as the actual concentration divided by the internal standard concentration, the concentration of each analyte can be determined from the internal standard concentration as follows: Conc y = SC y (Concis ). Since the intercept can be assumed to be 0, the concentration of element y in the unknowns can be found by Conc y = K (SSy )Concis. In many applications of LA-ICP-MS, an internal standard concentration is unknown and cannot be assumed. Ceramic studies, for instance, sometimes call for characterization of individual components within a heterogeneous matrix [40]. As another example, several pigment colors on a single polychrome sherd may be targeted [48]. While an electron microprobe or microXRF could be used to obtain point-speciic analysis of the individual components, the extra sample preparation and analysis would dramatically increases total analysis time, and the necessary ancillary instrumentation may not always be available. Gratuze [49,50] proposed a way around the problem of unknown internal standard concentrations that applies to silicate materials and has been widely adapted in archaeological LAICP-MS studies. Gratuze’s approach relies on measuring most major, minor, and trace elements and then accounting for all oxygen atoms, so that very close to 100% of the whole rock is represented in the list of monitored signal intensities. In this case, the concentration of some element, y, is known to be some proportion of the total concentrations measured. Thus, the slope parameter, Ky, relating standardized concentration to standardized signal, can be used to write Conc y

=

K y (SSy )Concis

∑ Conc ∑ K (SS )Conc m

.

i

i

i

is

i =1

If all concentrations are converted to oxides, then the denominator on both sides of the equation becomes 100%, and can be deined as

Conc yox Oy K y (SSy ) = m , 100% Oi Ki (SSi )i

∑ i =1

where Oi is the element-to-oxide conversion factor and Concyox is the oxide concentration of element y. The individual oxide concentrations can then be expressed as

Conc yox

   O (SS )K  y y y  100%. = m   Oi (SSi )Ki    i =1



If elemental concentration rather than oxide concentrations are required, then the oxide conversion can be removed at this point. For obsidian (rhyolitic volcanic glass), using silicon as an internal standard yields results in very close agreement with the Gratuze approach, as shown in Table 37.1. Gratuze [49] recommended dividing the elements into several groups to reduce analysis time and minimize problems of fractionation that can occur as the laser ablates deeper into the sample. In this approach, major elements are measured in the irst group of elements, and then one of these elements is selected to serve as the internal standard for subsequent groups. If line scans rather than spot ablations are used, then all elements can be collected simultaneously. Also, with time-of-light (TOF) instruments (see below), analysis time does not increase with the number of elements. Therefore, the total time can be as short as a fraction of a second. In archaeology, LA-ICP-MS has recently been used in a wide range of material characterization studies that were considered to be either impractical or impossible. Some examples of these studies included the elemental analysis of ceramic slips and pigments [3,48,51–53]; analysis of individual nonplastic particles within ceramic pastes [40,54]; analysis of clay matrix in ceramic pastes [54–59]; analysis of historic glass trade beads from various world regions [60,61]; analysis of copper and copper alloys [62,63]; analysis of Bronze Age glass from Egypt, Greece, and Mesopotamia [64]; analysis of gold [65]; analysis of Maya Blue pigment [66]; and analysis of tooth enamel for population movement studies [21,22]. Minimal sample preparation, speed of sample throughput, and sensitivity to a wide range of elements are distinct advantages of LA-ICP-MS, even with materials that can be easily analyzed by other techniques, such as chert and lint [67,68], obsidian [69,70], jade [71], turquoise [72], and quartzite [73].

TABLE 37.1 Comparison of Concentrations in Obsidian Based on Assumed Value of 35% for the Internal Standard (Si) and Based on the Gratuze Approach to Calibration, in Which Oxide Concentrations Are Summed to 100% (See Text for Full Explanation) Assumed 35% Si Internal Standard Concentrations

833

Mg Al Si K Ca Sc Ti V Cr Mn Fe Ni Co Cu Zn As Rb Sr Y Zr Nb Sn Sb Cs Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Pb Th U

Gratuze Approach

COD125

COD126

COD127

COD88

COD55

UK1

UK4

UK6

AR14

63.04 53,139 350,000 38,908 2327 7.00 254 0.00 0.00 183 6853 3.23 0.59 3.16 53.62 6.55 242 2.27 22.27 61.57 33.39 7.71 0.69 4.92 6.26 8.25 24.06 2.37 14.05 2.76 0.04 2.81 0.44 3.42 0.63 2.13 0.29 2.60 0.30 2.56 2.26 22.69 13.54 5.22

65.27 53,828 350,000 48,711 2348 8.11 262 0.00 0.00 184 6836 2.59 0.60 2.82 52.38 6.36 234 2.46 23.40 64.17 34.25 7.73 0.71 4.80 6.50 8.61 24.67 2.53 14.90 2.85 0.04 2.81 0.45 3.83 0.65 2.34 0.30 2.82 0.31 2.62 2.40 22.08 14.37 5.18

61.77 59,487 350,000 25,360 2259 6.71 246 0.00 0.00 173 6529 1.89 0.56 2.65 52.08 6.20 231 2.21 22.44 60.48 33.22 7.80 0.67 4.84 6.26 8.27 23.60 2.41 14.53 2.90 0.05 2.72 0.46 3.46 0.64 2.28 0.30 2.67 0.30 2.49 2.27 22.51 13.76 5.18

54.24 50,696 350,000 41,861 2127 7.99 230 0.00 0.00 164 6090 1.81 0.50 2.76 52.47 6.66 239 1.66 20.39 51.68 31.66 8.00 0.74 5.06 4.87 7.16 21.51 2.11 12.96 2.58 0.03 2.41 0.41 3.19 0.57 2.07 0.25 2.38 0.27 2.27 2.08 23.02 12.51 5.52

81.37 54,602 350,000 50,544 2387 8.72 330 0.00 0.00 156 7321 1.72 0.72 3.46 55.80 6.55 248 4.25 23.90 78.47 33.91 8.51 0.65 5.20 22.09 15.53 39.37 3.57 21.68 3.16 0.05 3.05 0.52 3.75 0.71 2.45 0.31 2.80 0.33 2.80 2.31 22.46 14.86 5.69

1301.46 72,776 350,000 32,391 6501 12.05 1634 10.35 0.00 194 13,240 3.17 2.73 10.37 40.57 3.27 159 74.49 12.27 146.29 7.09 3.34 0.55 8.67 848.61 15.93 34.10 3.30 19.79 2.48 0.41 1.97 0.32 2.18 0.36 1.24 0.16 1.61 0.17 3.21 0.46 17.54 7.71 3.94

403.03 55,397 350,000 46,158 3962 9.91 561 0.57 0.00 219 5364 0.87 0.80 2.90 26.76 0.40 96 35.39 9.99 47.58 5.47 1.28 0.20 2.88 1197.84 14.90 31.78 2.75 17.10 1.85 0.41 1.38 0.19 1.61 0.30 0.96 0.14 1.30 0.14 1.54 0.30 13.31 3.81 2.44

894.07 61,658 350,000 54,210 4417 10.94 1210 5.84 0.00 189 9693 0.78 1.78 8.72 34.27 7.69 140 43.91 13.25 108.71 7.72 2.51 1.23 6.38 711.36 15.62 35.53 3.32 20.00 2.64 0.29 2.09 0.32 2.18 0.40 1.40 0.18 1.75 0.19 2.73 0.42 16.90 6.98 3.51

558.36 60,843 350,000 59,624 3999 9.90 853 3.15 0.00 215 9961 0.00 1.20 3.26 43.42 7.73 155 54.83 7.60 105.69 10.31 2.56 1.06 4.45 1041.81 23.03 44.82 3.61 22.70 1.82 0.27 1.35 0.18 1.30 0.24 0.79 0.10 0.91 0.12 2.50 0.60 25.72 7.78 4.00

COD125

COD126

COD127

COD88

COD55

UK1

UK4

UK6

AR14

69.18 58,318 384,115 42,700 2554 7.69 278 0.00 0.00 201 7521 3.54 0.65 3.47 58.85 7.19 266 2.49 24.45 67.57 36.65 8.46 0.75 5.40 6.87 9.05 26.41 2.60 15.42 3.03 0.05 3.08 0.49 3.75 0.69 2.34 0.32 2.85 0.33 2.81 2.48 24.90 14.86 5.73

70.61 58,235 378,655 52,700 2540 8.78 283 0.00 0.00 199 7395 2.80 0.64 3.05 56.67 6.88 253 2.66 25.31 69.42 37.05 8.36 0.77 5.19 7.04 9.32 26.69 2.74 16.12 3.08 0.05 3.04 0.49 4.14 0.70 2.53 0.33 3.05 0.34 2.84 2.59 23.88 15.55 5.61

68.16 65,640 386,206 27,983 2493 7.41 271 0.00 0.00 191 7204 2.08 0.62 2.92 57.46 6.84 255 2.44 24.76 66.73 36.66 8.61 0.74 5.34 6.91 9.13 26.05 2.66 16.03 3.20 0.05 3.01 0.51 3.81 0.71 2.52 0.34 2.95 0.33 2.75 2.51 24.84 15.18 5.72

59.69 55,793 385,190 46,070 2341 8.79 253 0.00 0.00 181 6703 1.99 0.55 3.04 57.75 7.33 264 1.83 22.45 56.87 34.85 8.80 0.81 5.57 5.36 7.88 23.67 2.33 14.26 2.83 0.04 2.65 0.46 3.51 0.63 2.28 0.28 2.62 0.30 2.50 2.29 25.33 13.77 6.08

87.59 58,778 376,773 54,410 2570 9.39 355 0.00 0.00 168 7881 1.85 0.77 3.73 60.07 7.05 267 4.57 25.73 84.47 36.50 9.16 0.70 5.59 23.78 16.72 42.38 3.84 23.34 3.40 0.06 3.28 0.56 4.04 0.76 2.64 0.33 3.01 0.35 3.01 2.49 24.17 15.99 6.12

1354.52 75,743 364,270 33,712 6766 12.54 1701 10.77 0.00 202 13,780 3.30 2.84 10.80 42.23 3.41 166 77.53 12.77 152.26 7.38 3.47 0.57 9.02 883.21 16.58 35.50 3.44 20.60 2.58 0.43 2.05 0.33 2.26 0.38 1.29 0.17 1.68 0.18 3.34 0.48 18.25 8.03 4.10

434.98 59,789 377,746 49,817 4276 10.69 605 0.62 0.00 236 5790 0.93 0.86 3.13 28.88 0.44 104 38.20 10.78 51.35 5.90 1.38 0.21 3.11 1292.80 16.08 34.30 2.97 18.45 2.00 0.44 1.49 0.21 1.74 0.32 1.04 0.15 1.40 0.15 1.66 0.32 14.37 4.12 2.63

934.79 64,466 365,939 56,678 4618 11.44 1266 6.10 0.00 198 10,135 0.82 1.86 9.11 35.83 8.04 146 45.91 13.85 113.67 8.07 2.63 1.29 6.67 743.76 16.34 37.15 3.47 20.91 2.76 0.30 2.18 0.33 2.27 0.42 1.47 0.19 1.83 0.20 2.85 0.43 17.67 7.30 3.67

581.30 63,343 364,384 62,074 4164 10.31 888 3.28 0.00 224 10,370 0.00 1.25 3.39 45.21 8.05 162 57.08 7.91 110.03 10.74 2.66 1.11 4.63 1084.63 23.98 46.66 3.76 23.63 1.89 0.28 1.40 0.19 1.35 0.25 0.82 0.11 0.95 0.12 2.61 0.62 26.77 8.10 4.16

834

LASER ABLATION ICP-MS IN ARCHAEOLOGY

The microspatial capability of LA-ICP-MS can be its greatest advantage. Several studies of pigments on pottery have demonstrated that paint and glaze recipes can carry a signal of interaction at a spatial scale different from the signal carried by the ceramic paste [74]. Duwe and Neff [51], for instance, demonstrated that, while ceramics from Bailey Ruin, eastern Arizona, were derived from a single (probably local) source, pigments on those vessels formed multiple groups that showed clear spatial patterning within the Pueblo, patterning that appears to document networks of social interaction within the Pueblo. Although the microprobe capability of LA-ICP-MS is an advantage in many cases, such as the pigment studies just mentioned, some provenance investigations are better served by bulk characterization of heterogeneous matrices. For instance, bulk characterization of ceramic pastes by INAA has proved incomparably effective for assigning provenance to ceramics throughout the world [35,38,75,76]. Yet, because the number of research reactors is shrinking, there is latent demand for additional reliable approaches to bulk characterization. LA-ICP-MS could serve this demand, but additional work is needed to develop protocols for sample homogenization and reliable standardization. One approach, described above, is to infuse pressed-powder wafers with an indium standard [10]. Another possible approach to the analysis of homogenized and pressed powders would be to optimize XRF analysis for select major and minor elements, and then to use those elements as internal standards in LA-ICP-MS analysis of the same wafers.

37.3 ELEMENTAL ANALYSIS WITH LA-TOF-ICP-MS While LA systems are often coupled with scanning mass spectrometers (quadrupoles and magnetic sector instruments), TOF-ICP-MS is better suited for many LA applications. Rather than scanning the mass range, ions generated from the plasma in a TOF-ICP-MS format are sampled at a single instant in time. Different masses are detected by monitoring how long it takes them to reach a detector: the heavier ions arrive at the detector later than lighter ions. Many “snapshots” of the entire mass spectrum can thus be obtained in a fraction of a second. By sampling ions produced by the plasma at a single instant in time, TOF-ICP-MS eliminates variation in ion production over time due to plasma licker and variation in the sample stream reaching the plasma. The speed and reduction of variation due to ion production make TOF-ICP-MS ideal for the analysis of transient signals, such as the signals produced by LA sample

introduction. The studies that would be problematic with the relatively slow analysis time of scanning mass spectrometers are highlighted in this section. Fast analysis time with TOF-ICP-MS has proven to be useful for the analysis of obsidian microdebitage. Fraser Shapiro and Gilstrap [77] demonstrated proofof-concept, generating experimental samples by crushing obsidian from several California sources, then sieving the powder into size classes, the smallest being 0.09- to 0.125-mm diameter. Time-scan data were collected by spot ablation of individual grains and standardized using well-characterized obsidian source samples. Microdebitage samples from different sources were easily discriminated, even in the smallest size fraction. The importance of including small obsidian lakes and microdebitage in provenance investigations was recently demonstrated by Eerkens et al. [78], who concluded that failure to include microdebitage will bias provenance studies in favor of local sources and can underestimate obsidian assemblage diversity. Fraser Shapiro and Gilstrap [77] speciied additional uses and advantages of microdebitage sourcing, such as the potential to obtain source representation data even at heavily looted archaeological sites. LA-TOF-ICPMS analysis of obsidian microdebitage from archaeological contexts has recently been reported by Morgan et al. [79]. As noted above, a number of LA-ICP-MS studies of ceramic slips and pigments have been undertaken that have featured the use of scanning mass spectrometers [3,48,52,53]. Sometimes, pigments are very thin or poorly afixed to the underlying paste, either because of weathering or because they were applied after iring. An example of the latter is the hematite pigment applied to carved and incised pottery produced by the San Lorenzo Olmec of Mesoamerica. Therefore, Backes et al. [80] used LA-TOF-ICP-MS in time-scan mode to detect highly transient signals produced during the ablation of hematite patches on Olmec gray pottery from San Lorenzo and Canton Corralito. Also included in the study were raw pigments found in archaeological contexts at both San Lorenzo and Canto Corralito and geological hematite from the vicinity of San Lorenzo. Because the iron signal from hematite overwhelmed other components, Backes utilized ratios of other analytes to iron in pattern recognition analysis. The results showed that pigments from the pottery group together with the raw pigments. Raw pigments were exported from San Lorenzo, and carved gray pots at Canton Corralito; both imports from San Lorenzo and local copies were decorated with San Lorenzo-derived hematite. This study provided a remarkable complement to the results of an earlier INAA study of Olmec pottery, which detected high-volume export of Olmec pots from

ISOTOPE RATIO ANALYSIS WITH LA-ICP-MS

the Gulf Coast but no other Early Formative ceramic exchange [75,81,82]. LA-TOF-ICP-MS can also be used for elementalmapping applications that would be dificult or impossible to perform with quadrupole LA-ICP-MS. In one study [83], ceramics from the Sepik Coast of New Guinea at the Field Museum revealed some clear associations between ind spot and chemical group. This study raised the possibility that diagenetic enrichment of barium may contribute to some of the apparent discrimination. In order to test this possibility, microscale mapping of elemental concentrations in Sepik Coast sherds was undertaken with LA-TOF-ICP-MS. The microprobe capabilities of the TOF allowed quantitative data to be obtained from spots less than 100-µm diameter. Elements other than barium showed no trends across sherd cross sections. Similarly, in low barium, there were no trends in barium concentrations. Highbarium sherds, however, showed enrichment near one or both surfaces. These observations are consistent with diagenetic enrichment of barium [83].

37.4 ISOTOPE RATIO ANALYSIS WITH LA-ICP-MS Possibly the greatest potential for future development of LA-ICP-MS in archaeology is in the area of rapid, precise isotope ratio measurement. Lead isotope ratios have been used for some time in provenance studies of copper and bronze [41,84], and lead, together with strontium isotopes, has been used successfully in sourcing turquoise [85]. Lead isotope ratios in human tooth enamel can also be used in studies of population movements [86,87] and in studies of ancient human health status [88]. Strontium isotopes have been used with tremendous success over the last decade for the analysis of human bone and tooth enamel [43,86,89,90]. The resulting inferences of prehistoric population movements have had a major impact on understanding ancient societies in Europe [42,43,91–94], Mesoamerica [95,96], and the Andes [97]. Isotope ratio determination is also useful in dating applications that could potentially have major impacts in archaeology, as discussed further below. Thermal ionization mass spectrometry (TIMS) has been the technique of choice in most lead and strontium isotope ratio studies. TIMS is a bulk technique, where samples are powdered and concentrated prior to analysis. This bulk technique has two negative features. First, processing of samples is time-consuming and requires additional laboratory equipment. Second, the samples must be homogeneous. This aspect of TIMS is problematic in samples where small-scale compositional vari-

835

ability is of interest, or potentially contributes error to the result. In teeth, for example, strontium isotopes may vary in a sample from interior to exterior, and may relect diet, time and growth, remodeling in the dentin, or postdepositional alteration [98]. Although the potential of ICP-MS isotope ratio analysis has been recognized for some time [99–101], few convincing studies have appeared, and the technique was considered experimental until relatively recently [102]. The main drawback with scanning instruments such as quadrupoles or single-collector magnetic sector instruments is the limited precision that results from plasma licker. LA further degrades precision, as ablation yield inevitably varies over time, although Jochum et al. [103] reported substantial gains in precision in 208 Pb/206Pb and 207Pb/206Pb ratios over standard instrumentation when a 193-nm wavelength solid-state laser was coupled with a single-collector magnetic sector ICP-MS. In principle, TOF-ICP-MS, with or without LA, should afford superior performance for the determination of lead isotope ratios. Since individual ion packets are monitored, the effects of variation in ionization in the plasma or variation in yield from the laser should be reduced. A major limitation with available instruments is that both precision and accuracy are highly dependent on count rate. Nonetheless, Dudgeon et al. [104] reported partial success determining lead isotopes. 207 Pb/206Pb ratios are precise and accurate without calibration, while very good accuracy and precision on 208 Pb/206Pb can be achieved by monitoring ratios across a range of count rates and calibrating against a standard. Walton [105] reported a more elaborate approach to calibration of TOF lead isotope ratios, with apparent improvements in precision and accuracy. Unfortunately, since TOF isotope ratios depend so crucially on count rate, and 204Pb is so much lower than the radiogenic isotopes, useful precisions have not yet been achieved for ratios of the radiogenic isotopes to 204Pb. Some effort has also been focused on the determination of other isotope ratios, particularly 87Sr/86Sr with TOF-ICP-MS and LA-TOF-ICP-MS [104]. These efforts have not, however, yielded precision and accuracy useful for the characterization of human tooth enamel for population movement studies. Part of the problem probably arises from interferences produced by the calcium phosphate matrix [106], but there also appears to be TOF detector-related fractionation akin to but more extreme than that observed for 208Pb/206Pb. Considering the limitations of quadrupole, magnetic sector, and TOF isotope ratio studies, there seems little doubt that the future of LA-ICP-MS isotope ratio work lies with multicollector (MC) instruments. MC-ICP-MS instruments are sector devices that permit sampling of

836

LASER ABLATION ICP-MS IN ARCHAEOLOGY

ions from the plasma simultaneously through a series of Faraday cups and discrete dynode electron multipliers (typically 12–15 Faraday cups and one to three electron multipliers). The detector array is conigured so that a range of up to about 16 atomic masses can be monitored simultaneously. This coniguration is essentially the same as that used in thermal ionization mass spectrometer (TIMS) instruments mentioned above. MC-ICPMSs are capable of determining isotope ratios with a precision that far surpass those achievable with a singlecollector ICP-MS [107]. Furthermore, MC-ICP-MS technology can determine all single-element isotope ratios of interest simultaneously with no degradation in precision. MC-ICP-MS also offers high sensitivity for the individual analytes, so that isotope ratios can be measured at high precision with analyte concentrations D/(Δh)2.

η = 0: α bulk = α s . Here, αg, αs, and αbulk are the atomic fractions of 18O in the gas-phase oxygen, on oxide surface and in the oxide bulk, respectively; f34 is the fraction of 16O18O molecule in the gas phase; CO2 is the gas-phase oxygen concentration (mol/mol); τ is the residence time (s); b is the total number of surface sites (mol) per mole of gas molecules present in the catalyst section; R0, R1, and R2 are the rates of different types of exchange as calculated per active site of the surface (s−1); D is the diffusion coeficient of 18O in the oxide bulk (m2/s); h is the characteristic size of oxide particle (m); NS and Nbulk are the quantities of oxygen atoms on the surface and in the oxide bulk, respectively; ξ is the dimensionless reactor length; η is the dimensionless depth of the oxide layer. This model is analogous to the model proposed by Klier (Equations 51.11, 51.13, and 51.14) for batch systems, but applied to a plug-low reactor and supplemented by Equation 51.12 that describes the variation of the 16O18O fraction. The rate of heteroexchange on the oxide surface and the diffusion coeficient of labeled atoms in the bulk can be determined by the analysis of the isotope fraction αg(t) variation. The shape of calculated responses αg(t)|ξ=1 and f34(t)|ξ=1 as a function of the ratio between the rate of interphase exchange V and that of diffusion is shown in Figure 51.14. Three cases can be considered:

1. If the rate of isotope exchange is strongly determined by the heteroexchange on the oxide surface (V > D/(Δh)2), the αg(t) response curve is smoothed (Figure 51.14B) and cannot be approximated by an exponential curve, then the kinetic distinctions between the different types of mechanisms tend to decrease. 3. In the extreme case where the rate of isotope exchange is determined only by 18O diffusion in the oxide bulk (V >> D/(Δh)2), all differences between the various mechanisms of exchange vanish (Figure 51.14C).

1248

MASS SPECTROMETRY IN THE SSITKA STUDIES A

B Ne

Ne 1.0

1.0 αg

isotope fraction

αg 0.8

0.8

0.6

0.6

0.4

0.4

- Pt-CeO2-ZrO2-La2O3

f34

0.2

0.0

0.2

f34

0.0 0

120 240

300 400 500

time (s)

FIGURE 51.15 and 850°C (B).

- CeO2-ZrO2-La2O3

0

120

240

300 400 500

time (s)

Experimental αg(t) and f34(t) responses observed for CeO2-ZrO2-La2O3 and Pt/CeO2-ZrO2-La2O3 at 650°C (A)

Thus, the dynamics of isotope response in the pluglow reactor is highly sensitive to the relative rates of interphase exchange and 18O diffusion in the oxide bulk. In the case of exchange of type III, which is most typical of oxides, the sensitivity of f34(t) is even higher than that of αg(t). In particular, the maximal value of f34(t) depends prominently on the V/D ratio (Figure 51.14), and preliminary conclusions on the mechanism of exchange and type of rate-determining step can be done without modeling. 51.6.1 Effect of Pt on the Oxygen Mobility in Fluorite-Like CeO2-ZrO2-Based Mixed Oxides Platinum supported on ceria-zirconia-based oxides shows high activity and selectivity in selective oxidation of hydrocarbons to syngas. The oxygen transport from the support to the metal phase plays a key role in the mechanism of this reaction and therefore deserves to be investigated carefully for advanced kinetics and mechanistic pathways. To elucidate the mechanism and kinetics of oxygen transfer in such materials, both CeO2-ZrO2-La2O3 and Pt/CeO2-ZrO2-La2O3 samples were studied [25]. First, the amount of exchangeable oxygen both for the pure oxide and for the Pt-modiied oxide one exceeds the overall amount of oxygen in the sample by several monolayers. The oxygen exchange over the support obeys a type III mechanism (i.e., determined by the rate

of interphase exchange) over the entire temperature interval as determined by comparison of αg(t) and f34(t) experimental curves that were observed over both samples (Figure 51.15) with the above-discussed models (Figure.51.14). The increase of f34(t) maximal value with temperature rise is due to the fact that the diffusion processes are less dependent on T than surface processes. The close amount of exchanged oxygen during the initial period after the switch observed for Pt/CeO2ZrO2-La2O3 and CeO2-ZrO2-La2O3 despite a more than twofold decrease in surface area means that the rate of isotope exchange over Pt-containing sample is substantially higher than that over the mixed oxide alone. Since the total quantity of Pt atoms in the sample was substantially lower than the amount of fast substituted oxygen, the promoting effect of Pt on the observed rate of isotopic exchange was related to the oxygen spillover between Pt and the support. The numerical analysis of transient isotope responses conirmed the conclusions based on the qualitative analysis of αg(t) and f34(t) transients within the model described by Equations 51.11– 51.14. Table 51.3 lists the obtained values of R2 as calculated per active site of the CeO2-ZrO2-La2O3 material, diffusion coeficients, and the intrinsic amount of exchangeable oxygen deduced from the composition of the mixed oxide. The value of oxygen diffusion coeficient within the CeO2-ZrO2-La2O3 mixed oxide calculated in this work (D = 1.6 × 10−18 m2/s) is three orders of magnitude

SSITKA APPLICATION FOR THE STUDY OF OXYGEN TRANSPORT IN SOLIDS

1249

TABLE 51.3 Calculated Values of the Rates of Surface Oxygen Exchange (R2), Diffusion Coeficients (D), and Total Amount of Exchanged Oxygen (NOexchange) as Compared with the Intrinsic Amount of Oxygen in CeO2-ZrO2-La2O3 Lattice (NOoxide) R2 (s−1)

R1

R0

D (m2/s)

NOexchange (at/g)

NOoxide (at/g)

0.12 (±0.01) 0.30 (±0.02)

0 0

0 0

1.5 ÷ 2 × 10−18 1.6 (±0.2) × 10−18

9.0 (±0.5) × 1021 9.0 (±0.5) × 1021

7.2 × 1021 7.2 × 1021

T (°C) 650 850

higher than the value obtained after extrapolation of that reported by other authors at a substantially lower temperature (D = 1.5 × 10−21 m2/s). Such difference can be related to a much higher concentration of defects and domain boundaries in the mixed oxides considered in our study. Such disordered boundaries, which are formed by the stacking of nanodomains that result from the use of Pechini route for sample preparation, promote oxygen mobility and ability to accommodate surstoichiometric oxygen and/or hydroxyl groups. The increase in the concentration of defects can be caused by the generation of additional oxygen vacancies in the lattice at a higher measurement temperature. In the case of dense and regular crystalline lattice, the isotope transfer implies the exchange between neighboring oxygen atoms included into the crystalline lattice. In the defect-rich structure, dissociatively adsorbed oxygen can diffuse easily into the bulk via the defects and domain boundaries. The rate of oxygen diffusion is thus determined by the length and surface of boundaries and by the concentration of defects. During diffusion, atomic oxygen can exchange with the oxide domains lattice; thus, the overall rate of oxygen exchange increases. In accordance with the above reasoning, a scheme of the isotope transfer within a defectrich CeO2-ZrO2-La2O3 mixed oxide was proposed: 18

O

O Ce

O

O Ce

O

O D = 1.5·10

–21

2

ms

Oads

Fast diffusion

Oads

Domain boundary, defects

Slow exchange

18

18

Omobile

(Table 51.3), thus amounting to ca. 20% of NOexchange, the diffusion coeficient of “mobile” oxygen DOmobil = 8 × 10−18 m2/s (850°C) was calculated as follows: DOmobil ≈ D

(51.15)

with an activation energy less than 7 kJ/mol that is about ive times smaller than the activation energy for the overall diffusion coeficient. However, in the case of Pt/CeO2-ZrO2-La2O3, the numerical analysis has shown that even in the presence of fast oxygen spillover between Pt and support, it was impossible to describe the experimental results by increasing only the rate of oxygen exchange between the gas phase and the catalyst surface (R2). Both the rate of labeled oxygen transfer to the support bulk (near subsurface layers mainly) and the overall quantity of exchangeable support oxygen increased in the presence of Pt (Table 51.4). The increase in the rate of gas-phase oxygen exchange with support surface (R2sup) can be related to the increased concentration of surface/near subsurface oxygen species accumulated on the oxygen vacancies (defects) of the support that result from an electronic transfer with metal atoms, especially due to the incorporation of Pt atoms into the mixed oxide lattice through the domain boundaries. This increase in the content of surface/near subsurface oxygen species would account also for a higher amount of available oxygen from the support and the enhanced lux of oxygen along these extended defects to the surface. A twofold decrease of SBET value in the presence of Pt evidences the restructuring of the nanodomains with partial plugging the nanoporosity and increases the concentration of the so-called internal boundaries, which favors the mobility of bulk oxygen in the immediate vicinity of Pt clusters (i.e., in the near subsurface layers). The following averaged estimates of diffusion coeficient of the “mobile” oxygen in the Pt-containing sample were obtained at 850°C according to Equation 51.15:

D = 1.7·10–18 m2s •

Supposing that the concentration of “mobile” oxygen is equal to the difference between the quantities of exchanged oxygen (NOexchange) and lattice oxygen (NOoxide)

N Oexhange Obulk =D , Omobil N Oexhange − N Ooxide



averaged by overall support volume (η = 0 ÷ 1) − DOmobil = 45 × 10−18 m2/s, averaged by surface/near subsurface layers (η = 0 ÷ 0.3) − DOmobil = 70 × 10−18 m2/s.

1250

MASS SPECTROMETRY IN THE SSITKA STUDIES

TABLE 51.4 Calculated Values of the Rates of Surface Oxygen Exchange with Pt (R2Pt) and Support (R2sup), Coeficients of Exchange between Pt and Support (βPt-sup), Diffusion Coeficients (D), and Total Amount of Exchanged Oxygen (NOexchange) as Compared with the Intrinsic Amount of Oxygen in the CeO2-ZrO2-La2O3 Lattice (NOoxide) T (°C)

R2Pt (s−1)

βPt-sup (s−1)

R2sup (s−1)

DSURF/hSURF (s−1)

D0 (η = 0) (m2/s)

650 850

0.3 {10} ≥3 {≥100} R2Pt + R2sup = 1.2 (±0.1)

0.3(±0.02)

0.4(±0.1)

40 (±10) × 10 40 (±10) × 10−18

a

a

a

−18

NOexchange (at/g) 21

10.5 (±0.5) × 10 10.5 (±0.5) × 1021

NOoxide (at/g) 7.1 × 1021 7.1 × 1021

In braces, values of R2Pt and βPt-sup as calculated per Pt atom.

High sensitivity of the maximal f34(t) value of the ratio between the rate heteroexchange on the oxide surface and the diffusion coeficient helped to determine the kinetic characteristics of the spillover step between Pt and the support. Although the characteristic time of the oxygen spillover from Pt to the support is less than 10−2 s at and below 650°C, the rate of spillover is determined by the surface diffusion with a character2 istic time (hSURF / DSURF) of ca. 2 s (at 650°C). The rate of surface diffusion prominently increases with temperature. As a consequence, at 850°C, the rate of spillover is determined by the diffusion into the bulk, which depends weakly on temperature. All these intrinsic kinetic parameters related to oxygen activation and diffusion for ceria-zirconia-based mixed oxide in the absence or in the presence of Pt were used in a further transient kinetic description of the partial oxidation of methane over these reference materials. 51.6.2 Inluence of Oxygen Mobility on Catalytic Activity of La-Sr-Mn-O Composites in the Reaction of High-Temperature N2O Decomposition Similar studies of the mechanism and kinetics of oxygen exchange were performed for three La1-xSrxMnO3 (x = 0, 0.3, and 0.5) samples considerably differing in catalytic activity toward high-temperature N2O decomposition [26]. A combination of X-ray diffractometry (XRD) (including high-temperature measurements), X-ray photoelectron spectroscopy (XPS), and differential dissolution phase analysis (DDPA) indicates that the replacement La for Sr (x = 0.3) results in the transformation of orthorhombic cell (Pnma space group) characterizing the perovskite lattice of LaMnO3 to pseudocubic structure (Pm3m space group). The pseudocubic structure corresponds to the solid solution of perovskite La1-xSrxMnO3 with a variable Sr content that gradually decreases depthward in the bulk. The Srsupersaturated solid solution of perovskite (x = 0.5) transforms into the phase of (La1-ySry)2MnO4 with tetragonal K2NiF4-type structure (P4/nmm space group)

and consists of alternate layers of La1-xSrxMnO3 and SrO that are stable in the conditions of activity measurements, and highly dispersed MnOx. All the samples are characterized by the developed network of grain boundaries and their density changes in the order LaMnO3 > La0.7Sr0.3MnO3 > La0.5Sr0.5MnO3. Both Srcontaining phases are characterized by the presence of oxygen vacancies: isolated at x = 0.3, and highly concentrated—due to the removal of oxygen atoms from the lanthanide planes of layer-structured phase (La1-ySry)2MnO4 with a decreased oxygen bonding strength (x = 0.5). Therefore, oxygen-deicient systems should possess both high oxygen mobility in the bulk of the solid and increased ratio between the rates of surface-oxygen desorption and adsorption. N2O decomposition was chosen as the reaction that can be sensitive to these parameters. This reaction can be envisioned to occur via the following mechanism [27]:

2 → N 2 +O*, 1. N 2 O * k k1 → N 2 O*, 2. N 2 O + * ← k−1 k3 O → 3. 2O* ← 2 + * (* is active site for N 2 O k−3

adsorption), where step 3—desorption of oxygen—is most likely the rate-determining step, as was already observed for many catalysts. In this case, at low and medium temperatures ( La0.7Sr0.3MnO3 > La0.5 Sr0.5MnO3, which indicates that the substitution of Sr for La in manganites increases the lattice oxygen mobility. Since for all samples, the distribution of isotope molecules 16O18O and 18O2 at the outlet from the catalyst bed is approaching the equilibrium (it is envisaged as f34(t)max ∼ 0.5), the conclusion about the realization of R1 or R2 mechanisms of exchange cannot be done without numerical analysis of the isotope responses. The isotope exchange between the gas-phase oxygen and oxygen in the oxide was modeled for a plug-low reactor using the mass balance Equations 51.11–51.14 that describe changes in the 18O fraction. To describe oxygen transfer in all the systems studied, a general scheme that includes two pathways of oxygen exchange in perovskites with different rates should be proposed,

1252

MASS SPECTROMETRY IN THE SSITKA STUDIES

TABLE 51.6 Calculated Values of Oxygen Exchange Parameters for Different La-Sr-Mn-O Samples Catalyst

Exchange on the Surface R (s−1)a

Exchange in the Bulk Model I—one type of bulk oxygen

τs (s)

D (m2/s)

Nbulk (at/g) LaMnO3

8

20

73 × 10

≈0.13

τbulk (s)

0.6·10

−14

≈60

Model II—two types of bulk oxygen

La0.7Sr0.3MnO3 La0.5Sr0.5MnO3 a

15 15

≈0.07 ≈0.07

2

fast (at/g) N bulk

Dfast (m /s)

τfast bulk (s)

slow (at/g) N bulk

−1 βslow bulk (s )

τslow bulk (s)

35 × 1020 66.0 × 1020

2.3·10−14 6.4·10−14

≈10 ≈4

35 × 1020 17 × 1020

0.02 0.02

≈50 ≈50

Supposing a concentration of active sites is 1 × 1019 at/m2.

which is principally very close to that considered for luorite-based systems: O2 gas

O2 gas

R

R

Regular lattice

ª DLaMnO3

Dfast

Slow exchange Obulk

βslow bulk

Fast diffusion through oxygen vacancies

OSOS

OSOS

Obulk

Within this scheme, the slow pathway is due to exchange with participation of the regular lattice oxygen (LaMnO3 case). The fast pathway can be realized by oxygen diffusion through vacancies and lattice defects formed after Sr substitution for La. As follows from numerical analysis of αg(t) and f34(t) responses observed for LaMnO3, in the initial period of time after the switch oxygen exchange on the surface is the rate-determining step, while diffusion in the bulk becomes more signiicant with time. A complete heteromolecular exchange between the O2 molecule and two oxygen atoms of the solid (R2 mechanism) provides a more accurate description of f34(t) response, which is in agreement with the literature [30]. The best it of experimental αg(t) and f34(t) data for the La0.7Sr0.3MnO3 sample can be achieved with the assumption that the bulk of the solid contains two different types of oxygen characterized by different mobility: Ofast and Oslow in the ratio 1:1. Finally, for La0.5Sr0.5MnO3 sample, in addition to Dfast, the ratio between fast and slow exchangeable oxygen should be increased to 4:1.

Table 51.6 shows a list of the oxygen exchange rate values on the surface R and the diffusion coeficient D as well as the characteristic time of oxygen exchange on the surface and in the bulk (τs, τbulk). Since La0.5Sr0.5MnO3 possessed the lowest density of network boundaries, the highest rate of oxygen diffusion obtained for the multiphase sample La0.5Sr0.5MnO3 was related to the presence of the layer-structured phase LaSrMnO4 on the surface of the particles. However, based on the quantitative assessment of LaSrMnO4 phase made by DDPA, the layer-structured phase can supply only 55% of fast exchangeable oxygen. Therefore, the surface MnOx and/or the La1-xSrxMnO3 phase localized in the bulk should make an additional contribution to the fast oxygen exchange. As follows from the DDPA data, single-phase Sr-doped perovskite has the stoichiometric composition La0.78Sr0.22MnO3 that is quite close to La0.7Sr0.3MnO3 that contains ca. 50% of fast exchangeable oxygen. Thus, formation of oxygen vacancies in La0.78Sr0.22MnO3 can also make a signiicant contribution to the fast oxygen mobility in the bulk. Study of the tested samples in high-temperature nitrous oxide decomposition showed a direct correlation between activity and oxygen mobility in the samples (Figure 51.17).

51.7 BRØNSTED ACIDITY STUDY OF FIBERGLASS MATERIALS BY H/D EXCHANGE The catalysts based on silica iberglass materials can be eficiently used in many gas-phase and liquid-phase catalytic processes such as acid-catalyzed isopropanol dehydration reactions. To date, the acid–base properties of iberglass materials, including the Brønsted acidity, were not studied in detail. The lack of detailed studies could be due to the low speciic surface area of iber-

BRøNSTED ACIDITY STUDY OF FIBERGLASS MATERIALS BY H/D EXCHANGE

1253

sponds to the simple heteroexchange between hydrogen molecule and one hydrogen atom of the catalyst. This type of isotope exchange is the most typical of the isolated sites.

FIGURE 51.17 A direct correlation between catalytic activity of the La1–xSrxMnO3 (x = 0, 0.3, 0.5) samples in N2O decomposition in the absence () or presence ( ) of oxygen (900°C, contact time 5·10−4 s) and the diffusion coeficient ().



glass materials and the presence of active sites on both the external surface and in subsurface layers of ibers. Frequently used techniques based on chemical adsorption of probe molecules are ineffective for the characterization of iber acidity due to inaccessibility of acid sites located in the bulk of glass matrix (at low temperatures at least). The SSITKA study of H/D exchange is the most appropriate technique to study the catalyst acidity under reaction conditions. This method allows for the determination of acidic site concentration and the rate of deuterium exchange with molecular hydrogen, which depends on the strength of acidic sites. The results of H/D exchange study over the iberglass materials were compared with those for zeolite HZSM-5, SiO2, and the H3PO4/SiO2 reference standard. Thus, the acidity of iberglass materials in the traditional scale of proton afinity was estimated for the irst time [31]. 51.7.1 A Correlation between the Rate of H/D Exchange and Brønsted Acidity Expressed in Terms of Proton Afinity (PA) Figure 51.18 illustrates the time dependences of the atomic fraction of labeled hydrogen atoms (αD) and the molecular fraction of HD (fHD) in the gas phase observed over HZSM-5, SiO2, and H3PO4/SiO2 following H2 + N2 − D2 + Ar + N2 step change. The numerical analysis of the response curves (αD(t)) obtained with HZSM-5 zeolite showed the following: 1. There are three types of OH groups on the surface of zeolite distinguished through the H/D exchange rate with molecular hydrogen. 2. The H/D exchange proceeds via type II mechanism (R1) for all three types of sites and corre-

The calculated concentrations of different OH groups of HZSM-5 and their rates of H/D exchange (calculated per one site) with molecular hydrogen are listed in Table 51.7. The estimated concentrations of OH groups and their rates of H/D exchange for SiO2 and H3PO4/SiO2 are presented as well. The values of concentrations of different OH groups of zeolite obtained by the numerical modeling of the response curves are in good agreement with the results of acidity measurement by infrared (IR) spectroscopy of adsorbed CO (Table 51.8). The exchange rates of (OH)1 and (OH)2 zeolite sites are comparable, whereas the exchange rate of (OH)3 sites is slightly higher, which could be ascribed to the participation of the Lewis acid site (Al cations) in the exchange process. The exchange rate of H3PO4/SiO2 sample with weaker acid sites (PA = 1270 kJ/mol) is signiicantly lower than that of zeolite. In turn, the rate of H/D exchange on SiO2 (PA = 1390 kJ/mol) is at least two orders of magnitude lower than that for zeolite. Therefore, there is an obvious correlation between the strength of acid site and the rate of H/D exchange. Figure 51.19 illustrates a clear correlation between H/D exchange rates and traditional scale of proton afinity. This dependence is used further to characterize the acidity of catalysts in PA scale based on measured rates of H/D exchange. The proposed technique for the kinetic analysis of H/D exchange allows quantitative evaluation of the acid sites with different strength. 51.7.2

H/D Exchange over Fiberglass Catalysts

The study was performed with two types of iberglass materials: (1) nonmodiied leached iberglass material (FG) and (2) leached iberglass material modiied by alumina compounds (FG-M). Modifying was carried out via impregnation by Al2(SO4)3 solution with a concentration of 5 g Al/L (pH = 2.2). Figure 51.20 illustrates the time dependences of isotope fractions αD and fHD obtained for both samples of iberglass material in comparison with the normalized response curve for Ar. The observed exchange rate over the nonmodiied sample remains almost constant (i.e., the extent of isotope substitution of the OH groups that directly participate in the exchange with D2 is negligible) through a period of D2 + Ar + N2 low. This exchange rate is due to fast diffusion of the labeled atoms into the catalyst bulk. A direct isotope exchange between the molecular deuterium and the OH groups of iberglass on the

1254

MASS SPECTROMETRY IN THE SSITKA STUDIES

FIGURE 51.18 Experimental (points) and calculated (lines) time dependences of atomic fraction αD and the molecular fraction fHD obtained over HZSM-5, SiO2, and H3PO4/SiO2 after the switching of H2 + N2 low to D2 + Ar + N2 one, and vice versa.

TABLE 51.7 Calculated Values of the Acid Sites Concentration (Ni) and Their H/D Exchange Rates (Ri) for Zeolite HZSM-5, SiO2, and H3PO4/SiO2 Sample HZSM-5

SiO2 H3PO4/SiO2

Site

Ni·10−20 (sites/g)

R1i·103 (s−1)

(OH)1 (OH)2 (OH)3 (OH)1 (OH)1

1.5 0.3 0.2 0.6 0.10

1.7 1.3 2.6 ≤2·10−2 0.3

TABLE 51.8 Acidic Properties of HZSM-5, SiO2, and H3PO4/SiO2 According to IR Spectroscopy Data Sample HZSM-5

SiO2 H3PO4/SiO2

Site

NOH·10−20 (sites/g)

PA (kJ/mol)

Si-OH-Al (channels) Si-OH-Al (external surface) Al-OH Si-OH P-OH

1.50 0.30

1180 1190

0.24 0.60 0.10

1260 1390 1270

surface and in the subsurface layers of glass iber (OHSS) can be assumed. The labeled hydrogen atoms diffuse throughout the glass iber bulk (OHBULK) by isotope exchange between neighboring OH groups. Unlike H/D exchange over the FG sample, the observed H/D

FIGURE 51.19 Correlation between the H/D exchange rate (calculated per 1 Brønsted acid site) and PA. The circle shows the data referring to iberglass materials.

exchange rate over the modiied sample (FG-M) signiicantly decreases with the isotope substitution of the catalyst OH groups. High exchange rate at the initial period could be ascribed to the presence of another type of acidic sites. Most probably, the new hydroxyl groups arose from the modiication of iberglass materials with aluminum sulfate. The numerical analysis showed that the OH groups of the second type (denoted as OHS) do not participate in exchange with OH groups located in the catalyst bulk (OHBULK). Hence, OHSS groups can be assumed to be located in the subsurface layers of iber,

REFERENCES

1255

FIGURE 51.20 Experimental (points) and calculated (lines) time dependences of atomic fraction αD and the molecular fraction fHD obtained over FG and FG-M samples after the switching of H2 + N2 low to D2 + Ar + N2 one and vice versa.

TABLE 51.9 Calculated Values of Exchange Parameters for FG-M and FG Samples Sample

OHSS·10−20 (sites/g)

R11 (s−1)

OHS·10−20 (sites/g)

R21 (s−1)

OHBULK·10−20 (sites/g)

D/h2·105 (s−1)

FG-M FG

0.06 0.02

5·10−3 5·10−3

0.045 0

4.4·10−3 –

6 5 (±2)

2.5 >1

while OHS groups are at their external surface. The calculated values of parameters of H/D exchange over FG-M and FG samples are presented in Table 51.9. The exchange rates of the sites located in the subsurface layers of ibers (5·10−3/s) and at the external surface (4.4·10−3/s) are close (distinctions within the estimation accuracy range). Acidity in PA scale can be estimated with the correlation curve (Figure 51.19) as 1130– 1140 kJ/mol. Therefore, the strength of Brønsted acid sites of iberglass materials can be concluded to be comparable with zeolite. The increase in the rate of H/D exchange over iberglass materials as compared with zeolite could be related to the fact that iberglass materials represent viscous liquids (i.e., silica glasses); therefore, the H/D exchange proceeds in a liquid medium where the energy of polarized states depends on dielectric constant.

51.8

CONCLUSION

MS is the most eficient method for fast analysis of differently labeled molecules, in particular, when their con-

centration change is ine. The dynamics of their transfer both in the chemical/catalytic reaction and under adsorption–desorption equilibrium as studied by SSITKA allows to: •



reveal the mechanisms and kinetics of both fast (characteristic time up to ∼1 µs) chemical reactions and those complicated by mass transfer processes including evaluation of concentrations of intermediates and reaction rate coeficients (by the example of ethylene epoxidation, NO reduction with methane, Fischer–Tropsch synthesis); characterize both the surface (Brønsted acidity) and the bulk (oxygen mobility) of the solid.

REFERENCES 1. Happel, J. (1986) Isotopic Assessment of Heterogeneous Catalysis. New York: Academic Press. 2. Shannon, S.L., Goodwin, J.G. (1995) Characterization of catalytic surfaces by isotopic-transient kinetics during steady-state reaction. Chem. Rev., 95(3), 677–695.

1256

MASS SPECTROMETRY IN THE SSITKA STUDIES

3. Bennet, C.O. (1976) The transient method and elementary steps in heterogeneous catalysis. Catal. Rev. Sci. Eng., 13(2), 121–148. 4. Bal’zhinimaev, B.S., Sadovskaya, E.M., Suknev, A.P. (2009) Transient isotopic kinetics study to investigate reaction mechanisms. Chem. Eng. J., 154, 2–8. 5. Dawson, P.H., ed. (1976) Quadrupole Mass Spectrometry and Its Applications. Amsterdam: Elsevier. 6. Sadovskaya, E.M., Bulushev, D.A., Bal’zhinimaev, B.S. (1999) Dynamics of isotopic label transfer in catalytic reactions. Kinet. Catal., 40, 54–61. 7. Sadovskaya, E.M., Suknev, A.P., Pinaeva, L.G., Goncharov, V.B., Bal’zhinimaev, B.S., Chupin, C., Mirodatos, C. (2001) Mechanism and kinetics of the selective NO reduction over Co-ZSM-5 studied by the SSITKA. J. Catal., 201, 159–168. 8. Sadovskaya, E.M., Suknev, A.P., Pinaeva, L.G., Goncharov, V.B., Bal’zhinimaev, B.S., Chupin, C., Perez-Ramirez, J., Mirodatos, C. (2004) Mechanism and kinetics of the selective NO reduction over Co-ZSM-5 studied by the SSITKA technique: reactivity of NOx-adsorbed species with methane. J. Catal., 225, 179–189. 9. Sadovskaya, E.M., Suknev, A.P., Goncharov, V.B., Bal’zhinimaev, B.S., Mirodatos, C. (2004) Reaction kinetics and mechanism of selective NO reduction on a CoZSM-5 catalyst as studied by SSITKA. Kinet. Catal., 45, 436. 10. Hindermann, J.P., Hutchings, G.J., Kiennemann, A. (1993) Mechanistic aspects of the formation of hydrocarbons and alcohols from CO hydrogenation. Catal. Rev. Sci. Eng., 35, 1. 11. Anderson, R.B. (1984) The Fischer-Tropsch Synthesis. Orlando, FL: Academic Press. 12. Vannice, M.A. (1976) The catalytic synthesis of hydrocarbons from carbon monoxide and hydrogen. Catal. Rev. Sci. Eng., 14, 153–193. 13. Winslow, P., Bell, A.T. (1984) Application of transient response techniques for quantitative determination of adsorbed carbon monoxide and carbon present on the surface of a ruthenium catalyst during Fischer-Tropsch synthesis. J. Catal., 86, 158. 14. Brady, R.C., Petit, R. (1980) Reactions of diazomethane on transition-metal surfaces and their relationship to the mechanism of the Fischer-Tropsch reaction. J. Am. Chem. Soc., 102, 6181. 15. Iglesia, E., Reyes, S.C., Madon, R.J. (1991) Transportenhanced [alpha]-olein readsorption pathways in Rucatalyzed hydrocarbon synthesis. J. Catal., 129, 238. 16. Kummer, T., Emmet, P.H. (1953) Fisher Tropsch synthesis mechanism studies: the addition of radioactive alcohols to the synthesis gas. J. Am. Chem. Soc., 75, 5177. 17. Gafoor, M.A., Hutton, A.T., Moss, J.R. (1996) Synthesis, characterization and structure of the ethylene-bridged dimer [Cp(CO)2RuCH2CH2RU(CO)2Cp] and comparison of its reactivity with that of [Cp(CO)2RuCH2CH3] and

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

[Cp(CO)2Ru(CH2)5Ru(CO)2Cp]: models for FischerTropsch surface intermediates. J. Org. Chem., 510, 233. van Dijk, H.A.J., Hoebnik, J.H.B.J., Schouten, J.C. (2001) Steady-state isotopic transient kinetic analysis of the Fischer-Tropsch synthesis reaction over cobalt-based catalysts. Chem. Eng. Sci., 56, 1211. van Dijk, H.A.J., Hoebnik, J.H.B.J., Schouten, J.C. (2003) A mechanistic study of the Fischer-Tropsch synthesis using transient isotopic tracing. Part 1: model identiication and discrimination. Topics Catal., 26, 111. van Dijk, H.A.J., Hoebnik, J.H.B.J., Schouten, J.C. (2003) A mechanistic study of the Fischer-Tropsch synthesis using transient isotopic tracing. Part2: model quantiication. Topics Catal., 26, 163. van Dijk, H.A.J. (2001) The Fischer-Tropsch synthesis: a mechanistic investigation using transient isotopic tracing. PhD dissertation, Eindhoven University Press, Eindhoven. Sadovskaya, E.M., Suknev, A.P., Toktarev, A.V., Simonova, L.G., Paukshtis, E.A., Bal’zhinimaev, B.S. (2006) Isotope exchange between NO and H2O on a platinum-containing catalyst based on iberglass. Kinet. Catal., 47, 131. Klier, K., Novakova, J., Jiru, P. (1963) Exchange reactions of oxygen between oxygen molecules and solid oxides. J. Catal., 2, 479. Happel, J., Walter, E., Lecourtier, Y. (1990) Modeling transient tracer studies in plug-low reactors. J. Catal., 123, 12–20. Sadovskaya, E.M., Ivanova, Y.A., Pinaeva, L.G., Grasso, G., Kuznetsova, T.G., van Veen, A., Sadykov, V.A., Mirodatos, C. (2007) Kinetics of oxygen exchange over CeO2-ZrO2 luorite-based catalysts. J. Phys. Chem. A, 111, 4498. Ivanov, D.V., Sadovskaya, E.M., Pinaeva, L.G., Isupova, L.A. (2009) Inluence of oxygen mobility on catalytic activity of La-Sr-Mn-O composites in the reaction of high temperature N2O decomposition. J. Catal., 267, 5–13. Kapteijn, F., Rodriguez-Mirasol, J., Moulijn, J. (1996) Heterogeneous catalytic decomposition of nitrous oxide. Appl. Catal. B Environ., 9, 25–64. Muzykantov, V.S., Popovsky, V.V., Boreskov, G.K. (1969) Reactivity of oxygen adsorbed on oxide surface: 2. Isotopic exchange of oxygen in Fe-Sb-O catalysts. Kinet. Catal., 10, 1270–1277 (in Russian). Muzykantov, V.S., Popovsky, V.V., Boreskov, G.K. (1973) Determination of types of homomolecular oxygen exchange at oxides. Kinet. Catal., 14, 948. Doornkamp, C., Clement, M., Ponec, V. (1999) The isotopic exchange reaction of oxygen on metal oxides. J. Catal., 182, 390–399. Glazneva, T.S., Sadovskaya, E.M., Suknev, A.P., Goncharov, V.B., Simonova, L.G., Paukshtis, E.A., Bal’zhinimaev, B.S. (2009) Bronsted acidity study of iberglass materials by H/D-exchange. Appl. Catal. A Gen., 366, 262–268.

52 PROTON TRANSFER REACTION MASS SPECTROMETRY: APPLICATIONS IN THE LIFE SCIENCES Elena Crespo, Marco M.L. Steeghs, Simona M. Cristescu, and Frans J.M. Harren

52.1

INTRODUCTION

The air in the atmosphere is a mixture that consists of nitrogen, oxygen, argon, carbon dioxide, and many other compounds at (sub)-parts per million by volume (ppmv) level. Additional trace compounds can be found in air from speciic sources, such as air pollution from industry or cars, and on a smaller scale, cigarette smoke and emission by plants, animals, or humans. All these biological, atmospheric, and industrial sources inluence the composition of the air even when present in minute quantities. Nowadays, trace gas detection techniques can routinely detect quantities below 1 part per billion by volume (ppbv; 1: 109). These techniques allow the temporal/spatial distribution and quantity of the emitted gas to be studied. The measurement of those trace gases is important in many ields. Atmospheric chemists are interested in the origin of air pollution and its consequences, such as the greenhouse effect or the depletion of the ozone layer. Factory owners want to monitor the gas composition inside a production site to monitor the quality of the production process or the potential exposure to dangerous gases of employees. Biologists might want to know how a plant reacts on certain stress situations, such as drought, looding, high/low temperatures, or attack by insects. Food dealers might want to monitor the stage of ripening of the fruits, or deterioration of the meat in their storage rooms by the air around the food (headspace). Also, physicians can use gas analysis for human health research; the analysis of the content of

human breath can tell something about the status of the (human) body, with the advantage that there is no discomfort to the patient. Highly sensitive detectors are required to study the emissions of these trace molecules due to the sometimes very low release rate by the sources. The sources of gaseous compounds vary, just as the amounts and types of molecules emitted. Molecules from various chemical groups can be emitted such as hydrocarbons, alcohols, aldehydes, ketones, acids, esters, aromatics, halogenated hydrocarbons, and sulfur compounds. The only condition for the emission of a compound is that at a speciic temperature, the vapor pressure should be suficient. Many different detection methods that have been developed range from electronic noses, infrared spectrometers, laser-based detectors, and mass spectrometers, whether or not combined with gas chromatography (GC). Some of the detectors are developed for a speciic molecule resulting in highly dedicated sensors; others are capable of measuring many different compounds simultaneously. Within mass spectrometry, electron ionization (EI) typically results in the extensive fragmentation of the molecule under investigation. Therefore, the interpretation of the mass spectra (i.e., conirmation of molecular weight) can be dificult. A way to avoid fragmentation is via chemical ionization (CI), which is done in our case using proton transfer reaction (PTR) ionization. Similar to CI, molecules that are ionized using PTR techniques typically form a protonated molecular ion (M+H)+, in

Mass Spectrometry Handbook, First Edition. Edited by Mike S. Lee. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

1257

1258

PROTON TRANSFER REACTION MASS SPECTROMETRY: APPLICATIONS IN THE LIFE SCIENCES

which M is the molecular weight of the parent molecule. In addition, PTR has the advantage to be very eficient (the reaction rate constant is close to the collision rate constant), resulting in a very good performance with trace gas detection. The strength of proton transfer reaction mass spectrometry (PTR-MS) is its capability to perform detection of trace gases from various chemical groups in the order of seconds at (sub) part per billion levels. This strength is in contrast to GC methods in which the analytic capabilities are strongly enhanced but the time response is enlarged to 20–30 min. Therefore, PTR-MS is best used with on-line experiments when a fast time response is expected or when a number of experiments are monitored in parallel. This chapter describes the application and development of trace gas detection based on PTR-MS within life sciences. The chapter begins with a short overview about the ion chemistry that is used in these mass spectrometer systems to sensitively measure trace gases. The overview is followed by the experimental description of the system, including practical aspects such as how to perform a calibration or the use of natural isotopic ratios to gain some information about the identity of the detected compounds. The main part of the chapter deals with applications and measurements performed with PTR-MS to study: processes inside plants, fruit, bacteria, and insects; interactions between plants and pathogens; and also as a tool for human health research.

52.2 52.2.1

PTR-MS Ion Chemistry

PTR-MS is a form of CI mass spectrometry, which was developed in the mid-1990s [1]. In PTR-MS, a trace gas neutral molecule is ionized via a chemical reaction with H3O+, and the products are selected and detected according to their mass-to-charge ratio (m/z). As a general rule, only ions with a proton afinity (PA) higher than the one of water will react with H3O+. The advantages of PTR-MS are that it is fast, sensitive, and versatile; many different trace gas species can be measured in parallel and on-line, at the (sub) part per billion levels. Due to its inherent speed, measurements can be obtained every few minutes, or even every second, if necessary. One reason H3O+ is used as the reactant ion is that is does not react with compounds such as oxygen, nitrogen, CO2, or methane, which are present in large quantities in atmospheric air. The PA of water is larger than the above-mentioned molecules but lower than most of the volatile organic compounds (VOCs) that are present as trace gas in the air. The reac-

tion is often referred to as “soft” because of the abundant production of protonated molecular ions (M+H)+ without any dissociation. Reactions between ions and molecules are among the fastest chemical reactions known, because of the long-distance electrostatic interaction between the charge of the ion and the polar or polarizable molecule; the resulting interaction energy at short range is often suficient to overcome intrinsic energy barriers. This ion–molecule reaction is in contrast to molecule–molecule reactions that normally involve activation energy. To determine whether a certain compound can be studied and quantiied, three parameters are important: 1. The PA of the neutral molecule determines if the PTR can occur. Hunter and Lias listed the PA values of over 1700 compounds [2]; the National Institute of Standards and Technology (NIST) chemistry webbook is based on these values [3]. In practice, the structure of the molecule is a good estimator for whether the PA of a trace gas compound is higher than that of water. The PA of oxygenated, aromatic hydrocarbons, and hydrocarbons with an N, P, S, or Cl atom incorporated are generally higher than that of water and can be measured by PTR-MS. Saturated hydrocarbons (alkanes) and the “normal constituents” of air (N2, O2, NO, NO2, CO2, Ar, etc.) have PAs lower than water. 2. The collision rate constant determines the speed of the reaction. Eficient reactions proceed at or close to the collision rate (∼1). The reaction eficiency Φ is given by the ratio between the reaction and collision rate coeficients Φ = kr/kc. If such a PTR is exo-energetic, then the eficiency is ∼1, so the reaction rate constant can be considered equal to the collision rate constant. In a drift tube reaction between H3O+ and R, the reaction rate constant k appears as −

d [ H 3O + ] dt

=

d [ RH+ ] dt

= k ⋅ [ H 3O+ ] ⋅ [ R ].

(52.1)

Many collision rate constants of neutral gaseous molecules with H3O+ are listed in literature [4]. The uncertainty in calculated and measured values is typically 10%–20%. The values of these rate constants are compound speciic and vary around a value of about 2·10−9 cm3/s; therefore, when the collision rate of a compound is unknown, this value is often used in the calculation of its concentration. 3. Additionally, the reaction time determines the number of reactions that can take place. The reac-

PTR-MS

tion time in the PTR-MS instrument is the time it takes for an H3O+ ion to cross the drift tube. As a result of the electric ield E over the drift tube, the so-called drift velocity is given by vd = μ · E where μ is the ion mobility, which has been determined for numerous ions in different buffer gases, including for H3O+ ions in a nitrogen buffer gas (2.76 cm2/V/s) [5]. The reduced mobility is given as  p µ0 =   p0

  T0   N  T  ⋅µ =  N  0  

  ⋅ µ, 

E vd = µ 0 N 0   . N

(52.3)

The drift velocity is a function of the parameter E/N, which is a frequently used parameter in ion mobility studies and is expressed in Townsend (1 Td = 10−17 Vcm2). De Gouw et al. showed that this calculation is in excellent agreement with experiment [6]. The reaction time t = L/υd is around 110 μs, at 120 Td for a drift tube length of L = 10 cm. In addition to the normal PTR, the H3O+ and RH+ ions can cluster with water molecules: H 3O+ + n ( H 2O ) ↔ H 3O+ ⋅ ( H 2O )n RH+ + n ( H 2O ) ↔ RH+ ⋅ ( H 2O )n

Higher drift tube voltages also increase the kinetic energy of the product ions, which will fragment to an increasing extent with increasing drift tube voltage. Proton transfer results in little or no fragmentation as compared with other ionization techniques such as EI. It is known, however, that several compounds do fragment upon proton transfer and that the degree of fragmentation increases with increasing kinetic energy [7]. For instance, alcohols are known to break down easily, losing a water molecule via the dehydration channel:

(52.2)

where p is the pressure, T is the temperature, and N is the gas number density in the drift tube. N0 denotes the gas number density at standard temperature and pressure (STP): pressure p0 (1 atm) and temperature T0 (273.15 K). Substituting this in the drift velocity, we get

.

(52.4)

These clusters complicate the interpretation of the mass spectra. Depending on the pressure and the E/N value, the H3O+ (H2O)n clusters can be present in the drift tube and react with the trace gas compounds. Since the PA of the clusters is higher than the PA of water, the PTR with a water cluster is more selective. This reaction can be equally eficient as the PTR, depending on the dipole moment of the neutral R. For nonpolar molecules like benzene, cluster reactions will not take place. Therefore, the sensitivity or detection eficiency of a molecule like benzene can be humidity dependent, since the amount of water clusters depends on humidity. The formation of these clusters with PTR-MS techniques can be limited and controlled by increasing the electric ield applied over the reaction region or lowering the pressure.

1259

RH+ → ( R −OH ) + H 2O. +

(52.5)

This fragmentation may be dependent not only on the structure of the molecule itself, but also on the drift tube E/N value that controls the kinetic energy of the molecules. Therefore, it is important to know, either via direct measurements of pure compounds or literature values, the behavior of the trace gas compounds under study. Even though there are libraries available with fragmentation patterns obtained with EI, these spectra cannot be used as reference for PTR-MS because of the completely different nature of ionization. In a PTR-MS experiment, where speciicity and sensitivity are both important, there is usually a trade-off between high kinetic energies for low mass spectral complexity and low kinetic energies for high sensitivity and low degree of fragmentation. 52.2.2

Quantiication of Trace Gas Concentrations

The reaction equation of a PTR is given in Equation 52.1. Rearranging and solving the differential equation an exponential dependency on the trace gas concentration R can be determined:

[ RH + ] = [ H 3O+ ]0 ( 1 − e −k[R ]t ) ,

(52.6)

where [H3O+]0 is the concentration of the water ions at t = 0. For trace gas experiments, [R] is small and the exponent can be approximated by a Taylor expansion (see Equation 52.7). Usually, the assumption is made that the ratio between the detected signals is proportional to the ratio of concentrations in the drift tube, meaning that the detection eficiencies for both ions is the same. Thus, the ratio between the two concentrations can be replaced by the ratio of the detected signals i(RH+) and i(H3O+):

[RH + ] ≈ [ H 3O+ ]0 ( 1 − 1 + k [R ] t ) + 1 [ RH ] i (RH + ) 1 i (RH + ) ≈ ⋅ ⋅ FT = Ccal ⋅ . [R ] = ⋅ + + kt [ H 3O ] 0 kt i (H 3O ) i (H 3O+ ) (52.7)

1260

PROTON TRANSFER REACTION MASS SPECTROMETRY: APPLICATIONS IN THE LIFE SCIENCES

To take the difference in detection eficiency into account, the transmission factor FT is introduced, which is the ratio between the transmission eficiencies of both ions. It should be noted that the transmission factor FT is a factor speciic for an instrument and depends on the mass. The amount of ionized trace gas molecules in the drift tube, and therefore its calculated concentration, is linearly proportional to the amount of H3O+ ions. Consequently, every variation in production of these primary ions will result in a luctuation in the measured and calculated concentration of the trace gas component. To avoid this luctuation, the measured number of counts on a speciic mass is normalized to the ixed amount of 1·106 primary ions: i ( RH + )norm =

106 × i (RH + ) . i ( H 3O + )

vent this is to keep the concentration of the water cluster H3O+·H2O in the drift tube low (lower than 10% of the primary ion count rate). Calculation and experimental determination of the calibration factor can be in reasonable agreement. Since instrument performance and drift tube humidity can vary over time and fragmentation will be different for every compound for every E/N value, calibration measurements are preferred. However, several errors can also be introduced in calibration measurements by an inaccurate determination of the concentrations in the certiied mixture or inaccuracies in the mixing of lows with mass low controllers (MFCs), for example. If calculation and calibration are in reasonable agreement, then determined concentrations can be regarded as reliable. With proper calibration measurements, the uncertainty in the measured concentration can be decreased to ∼5%–10%. From the calibration factors obtained for the ixed set of compounds in the certiied mixture, the calibration factors of other compounds on a speciic m/z can then be calculated by taking into account the difference in collision rate constant, transmission eficiency factors, and fragmentation ratios.

(52.8)

The measured concentration is obtained from the count rate by division by the calibration constant Ccal (see Equation 52.7), which is expressed as the normalized number of counts per part per billion volume (ncps/ ppbv). As noted above, the water cluster H3O+·H2O reacts differently with the trace molecules as the water ion does. A correction factor is introduced to take this difference into account [8]: i (RH + )norm =

106 × i (RH+ ) . i (H 3O ) + X × i (H 3O+ ⋅ H 2O) +

52.2.3 Experimental Description PTR Mass Spectrometer A PTR-MS (see Figure 52.1) normally consists of an ion source (1) in which H3O+ ions are produced by a discharge in water vapor; a drift tube (2) where the trace gases from the gas sample are ionized by PTRs with H3O+ ions; a quadrupole mass ilter (3) where the ions are mass iltered based on its m/z; and a secondary electron multiplier (4) that counts the ions. The primary ions (H3O+) are produced in the source by a discharge in water vapor. These ions are directed

(52.9)

The factor X should be experimentally determined for every single compound, which implies a positive identiication of the compound behind the observed ion intensity; this is not always possible. One way to circum-

Water inlet

3 1

2

Trace gas inlet

4

TMP TMP

FIGURE 52.1 PTR-MS scheme with (1) an ion source, (2) the drift tube, (3) a quadrupole mass selector, and (4) the secondary electron multiplier. TMP, turbo molecular pump.

PTR-MS

52.2.3.1 Gas Handling Setup The trace gas sampling system should be comprised of inert materials to minimize adsorption in the inlet system and decrease the potential for memory effects. Materials such as Telon or perluoroalcoxy (PFA) plastics or Silcosteel (stainless steel with an inert coating) are used, while materials such as nylon and stainless steel should be avoided. Also, the use of MFCs with stainless steel inner surfaces should be avoided between the sampling port and the PTR-MS. When performing a calibration using stainless steel MFCs, the low of calibration gas mixture should be established through a low controller for at least a day before performing the calibration. In this way, a constant and reproducible trace gas mixture can be ensured. The addition of a catalytic converter or a carbohydrates ilter in the gas handling system is also recommended, to be able to measure hydrocarbon free (zero air) background values. 52.2.3.2 Calibration A calibration of the system is needed before starting a set of experiments in order to convert detector units (counts per second, cps) into concentration units (ppbv). Standard mixtures of compounds in known concentrations (preferably of interest for our application) are required that cover the mass spectral range of detection of the instrument. A calibration will be done mixing a ixed low of calibration mixture with a variable low of zero (free of hydrocarbons) air. As previously mentioned, when performing a calibration using stainless steel MFCs, the low of calibration gas mixture should be established through the low controller at least a day before performing the calibration, to ensure stable values.

Methanol, m/z 33 Acetaldehyde, m/z 45 Acetone, m/z 59 Isoprene, m/z 59 Benzene, m/z 59 Styrene, m/z 59

200

Ion Intensity (ncps)

toward the drift tube by applying an electric ield. The drift tube consists of a number of electrically isolated stainless steel rings; the thickness and the inner diameter of those rings can vary per design. The trace gas low enters the drift tube and the molecules are protonated by collisions with the primary ions. A homogeneous electric ield is established over the length of the drift tube by applying a voltage difference between beginning and end of the drift tube; this electric ield guides the protonated trace gas molecules toward the exit skimmer of the drift tube and into the buffer chamber. This chamber (at a pressure of about 10–4 mbar) separates the high pressure in the drift tube (2 mbar) from the high vacuum pressure (10–6 mbar) chamber with the quadrupole mass spectrometer. The advantage of having a low pressure in the quadrupole chamber is to decrease the number of collisions (i.e., ions with the residual particles and residual particles with the detector), and therefore, decrease the background signal.

1261

150

100

50

0 0

20 40 Concentration (ppbv)

60

80

FIGURE 52.2 Typical calibration curves for various mixtures of methanol, acetaldehyde, acetone, isoprene, benzene, and styrene between 0 and 80 ppbv.

As an example, the steps followed to make the calibration shown in Figure 52.2 are described. A gas cylinder containing a mixture of methanol, acetaldehyde, acetone, isoprene, benzene, and styrene was used for this calibration. In addition to the ions that correspond to the calibration mixture, other ion masses should also be monitored. The ion at m/z 19 corresponds to the primary ion mass, which amount is used to normalize the secondary ion counts. Because the value of the ion at m/z 19 is a factor 103 higher in intensity than most of the other masses (value 106 cps), it can age the electron multiplier very fast. Therefore, the ion at m/z 21 is used as an indicator for the primary ion signal, taking into account that this is the 18O isotope of H3O+ with a natural isotopic ratio H316O+/H318O+ of 500. The ion at m/z 25 is an indicator for the noise of the detector (since no compound is detected at this mass); the ion at m/z 30 is mostly representing NO+, a product from the primary ion discharge region, but also able to ionize trace gas compounds; the ion at m/z 32 represents O2+ with similar origin. NO+ and O2+ will both ionize the trace gas molecules without transferring a proton. To avoid confusion how the trace gas molecules are ionized, these primary ions should be kept as low as possible (