<%BANNER%>

Characterizing, Identifying, and Imaging Oxidized Phospholipids by Matrix-Assisted Laser Desorption/Ionization Tandem Ma...

MISSING IMAGE

Material Information

Title:
Characterizing, Identifying, and Imaging Oxidized Phospholipids by Matrix-Assisted Laser Desorption/Ionization Tandem Mass Spectrometry
Physical Description:
1 online resource (134 p.)
Language:
english
Creator:
Stutts, Whitney L
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Chemistry
Committee Chair:
Yost, Richard A
Committee Members:
Fanucci, Gail E
Powell, David Hinton
Bova, Frank J
Garrett, Timothy J

Subjects

Subjects / Keywords:
lipids -- maldi -- ms -- oxidation
Chemistry -- Dissertations, Academic -- UF
Genre:
Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Understanding the involvement of lipid oxidation in both inflammatory processes and the pathogenesis of various diseases is an important endeavor. Despite growing interest in the field of lipidomics, lipid oxidation products remain poorly characterized. However, mass spectrometry, coupled with soft-ionization techniques, has become a key tool for investigating lipids and elucidating structural changes resulting from oxidative modifications. In this research, matrix-assisted laser desorption/ionization tandem mass spectrometry (MALDI MSn) techniques were developed for characterizing, identifying, and imaging oxidized phospholipids. Specifically, the research presented in this work focused on three distinct aspects of oxidized phospholipid analysis. Initially, the MALDI MSn (where n = 2, 3, or 4) fragmentation pathways of short-chain phosphatidylcholine oxidation products (OxPCs) were characterized. This study identified several fragmentation pathways that allowed for unambiguous detection of low abundant OxPC species, and the developed method was applied for the identification of these species in rat spinal cord tissue. Next, a similar study was conducted for phosphatidylethanolamine oxidation products (OxPEs). In this second study, novel fragmentation pathways were identified for short-chain aldehydic OxPEs. This proposed pathway resulted in a macrocyclic gas-phase rearrangement that can be exploited for MSn detection of these OxPEs in biological samples. The final study presented in this work describes a methodology utilizing principal component analysis (PCA) for untargeted detection of oxidized phospholipids following mass spectrometric imaging (MSI) experiments. In summary, this work demonstrates that MALDI MSn provides a selective method for investigating individual molecular species resulting from phospholipid oxidation. Utilizing the MSn capabilities of the linear ion trap, OxPLs were distinguished from isobaric, and even isomeric, phospholipids. Furthermore, the utility of MALDI MSn in conjunction with multivariate data analysis techniques, namely PCA, for discovering, identifying, and imaging individual phospholipid oxidation products in spinal cord tissue is exemplified.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Whitney L Stutts.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Yost, Richard A.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045345:00001

MISSING IMAGE

Material Information

Title:
Characterizing, Identifying, and Imaging Oxidized Phospholipids by Matrix-Assisted Laser Desorption/Ionization Tandem Mass Spectrometry
Physical Description:
1 online resource (134 p.)
Language:
english
Creator:
Stutts, Whitney L
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Chemistry
Committee Chair:
Yost, Richard A
Committee Members:
Fanucci, Gail E
Powell, David Hinton
Bova, Frank J
Garrett, Timothy J

Subjects

Subjects / Keywords:
lipids -- maldi -- ms -- oxidation
Chemistry -- Dissertations, Academic -- UF
Genre:
Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Understanding the involvement of lipid oxidation in both inflammatory processes and the pathogenesis of various diseases is an important endeavor. Despite growing interest in the field of lipidomics, lipid oxidation products remain poorly characterized. However, mass spectrometry, coupled with soft-ionization techniques, has become a key tool for investigating lipids and elucidating structural changes resulting from oxidative modifications. In this research, matrix-assisted laser desorption/ionization tandem mass spectrometry (MALDI MSn) techniques were developed for characterizing, identifying, and imaging oxidized phospholipids. Specifically, the research presented in this work focused on three distinct aspects of oxidized phospholipid analysis. Initially, the MALDI MSn (where n = 2, 3, or 4) fragmentation pathways of short-chain phosphatidylcholine oxidation products (OxPCs) were characterized. This study identified several fragmentation pathways that allowed for unambiguous detection of low abundant OxPC species, and the developed method was applied for the identification of these species in rat spinal cord tissue. Next, a similar study was conducted for phosphatidylethanolamine oxidation products (OxPEs). In this second study, novel fragmentation pathways were identified for short-chain aldehydic OxPEs. This proposed pathway resulted in a macrocyclic gas-phase rearrangement that can be exploited for MSn detection of these OxPEs in biological samples. The final study presented in this work describes a methodology utilizing principal component analysis (PCA) for untargeted detection of oxidized phospholipids following mass spectrometric imaging (MSI) experiments. In summary, this work demonstrates that MALDI MSn provides a selective method for investigating individual molecular species resulting from phospholipid oxidation. Utilizing the MSn capabilities of the linear ion trap, OxPLs were distinguished from isobaric, and even isomeric, phospholipids. Furthermore, the utility of MALDI MSn in conjunction with multivariate data analysis techniques, namely PCA, for discovering, identifying, and imaging individual phospholipid oxidation products in spinal cord tissue is exemplified.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Whitney L Stutts.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Yost, Richard A.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045345:00001


This item has the following downloads:


Full Text

PAGE 1

CHARACTERIZING, IDENTIFYING, AND IMAGING OXIDIZED PHOSPHOLIPIDS BY MATRIX ASSISTED LASER DESORPTION/ IONIZATION TANDEM MASS SPECTROMETRY By WHITNEY LEIGH STUTTS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Whitney Leigh Stutts

PAGE 3

3 T o my loving family and cherished friends

PAGE 4

4 ACKNOWLEDGMENTS First, I thank my advisor, Dr. Richard Yost for his guidance throughout the last five years and for facilit ating a learning environment where ideas are always shared and independent thinking is always encouraged. I am also grateful to my committee members Dr. David Powell, Dr. Gail Fanucc i, Dr. F rank Bova, and Dr. Tim Garrett who have offered their advice and time in support of this research. Additionally, I thank all my fellow Yost group members, past and present, for their insightful discussions and review of this work I would also like to especially thank Dr. Jodie Johnson, Dr. Mike Napolitano and Dr. Rob Menger who have been excellent colleagues and friend s. I also express my sincere appreciation for two undergraduat e research assi stants, Kerolyn de Souza Valente and Katelyn Bobek, who se enthusiasm and dedication have aided the success of this work. I must also acknowledge several people who i nspired and motivated me along this journey to my Ph.D. My love for chemistry was sparked by Dr. Roger Bacon, a professor at Western Carolina University and his wife Lisa during a summer chemistry program for high school students. Their passion for chemistry was obvi ous and apparently contagious Once at North Carolina State University, Dr. Phil Brown inspired me with his enthusiastic advanced organic lectures, stories of graduate school, and hobby of backyard biodiesel synthesis. M y undergraduate research advisor, Dr Damian Shea gave me the opportunity to test my aptitude for research and to participate in a study abroad program for water management and environmental forensics in China and Vietnam which taught me the value of interdisciplinary collaborations for solving global problems

PAGE 5

5 During my graduate career, I have also had the pleasure of collaborating with many great scientists around the world. I thank Dr. Paolo Di Ma scio at the University of S o Paulo, Brazil for allowing me the opportunity to visit his lab and learn more about lipid oxidation and the detection of these products by mass spectrometry. I would also like to thank Dr. Ron Heeren and his group at FOM Institute AMOLF, especially Andrs Kiss for a rewarding collaboration and valuable research experience. Additionally much of this work would not have been possible without Dr. Nigel Calcutt at the University of California San Diego. Dr. Calcutt provided all of the rat tissue for these studies and valuable insight into the involvement of lipid oxidation in diabetic peripheral neuropathy. I would also like to thank my family and friends for all their love and support. Most of all, I thank my mom ; w ithout her, I would not be the person I am today. Furthermore, I thank both my dad for instilling in me the importance of higher education and my sister, Lauren, for her unconditional love and for sharing with me so many of lifes finest moments. I would als o like to thank m y grandparents and my A unt Delita, who have been a solid support system throughout my life, even when life took me to Florida In addition to my loving family, I have also been blessed with great friends who have made Gainesville feel like home. I thank Jess for a truly incredible friendship and many memories that will last a life time Lastly, I am grateful to my b oyfriend Mike who has loved, encouraged, and supported me throughout the last five years.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 8 LIST OF FIGURES .......................................................................................................... 9 ABSTRACT ................................................................................................................... 12 CHAPTER 1 ANALYSIS OF OXIDIZED PHOSPHOLIPIDS BY MATRIX ASSISTED LASER DESORPTION/IONIZATION TANDEM MASS SPECTROMETRY ......................... 14 Research Motivation ............................................................................................... 14 Lipids and Lipid Oxidation ....................................................................................... 15 Phospholipid Oxidation ..................................................................................... 16 Techniques for Measuring Phospholipid Oxidation .......................................... 18 Matrix Assisted Laser Desorption/Ionization ........................................................... 19 Matrix Selection ................................................................................................ 20 MALDI Process ................................................................................................. 21 Instrumentation ....................................................................................................... 22 Linear Ion Trap ................................................................................................. 23 Tandem Mass Spectrometry ............................................................................ 25 Mass Spectrometric Imaging .................................................................................. 26 Ov erview of Dissertation ......................................................................................... 27 2 CHARACTERIZATION OF PHOSPHATIDYLCHOLINE OXIDATION PRODUCTS BY MALDI MSn .................................................................................. 38 Introduction ............................................................................................................. 38 Experimental Methods ............................................................................................ 40 Chemical s ......................................................................................................... 40 Preparation of OxPC Standards ....................................................................... 40 Preparation of Spinal Cord Tissue .................................................................... 41 MALDI MS Instrumentation .............................................................................. 41 MSn Instrumental Parameters .......................................................................... 42 Accurate Mass Measurements ......................................................................... 43 Results and Discussion ........................................................................................... 43 MS2 Characterization of the [M+H]+ Ions of Short Chain OxPCs ...................... 44 MSn Characterization of the [M+Na]+ Ions of Short Chain OXPCs ................... 45 Identifying and Imaging OxPCs in Spinal Cord Tissue ..................................... 50 Conclusions ............................................................................................................ 51

PAGE 7

7 3 MALDI MSn CHARACTERIZATION OF PHOSPHATIDYLETHANOLAMINE OXIDATION PRODUCTS ....................................................................................... 67 Introduction ............................................................................................................. 67 Experimental Section .............................................................................................. 69 Chemical s ......................................................................................................... 69 Oxidation Procedure ......................................................................................... 69 MSn Experimental Parameters ......................................................................... 70 Results and Discussion ........................................................................................... 70 Autoxidation of POPE, PLPE, PDPE, and SDPE ............................................. 70 MSn of [M+Na]+ Ions of OxPEs ......................................................................... 71 MSn of [M H+2Na]+ Ions of OxPEs .................................................................. 76 Conclusions ............................................................................................................ 78 4 IDENTIFYING AN D IMAGING PHOSPHOLIP ID OXIDATION PRODUCT S BY MALDI MSn ............................................................................................................. 91 Introduction ............................................................................................................. 91 Experimental Methods ............................................................................................ 92 Chemical s ......................................................................................................... 92 Tissue Preparation ........................................................................................... 93 In Vitro Oxidation .............................................................................................. 93 Mass Spectrometry .......................................................................................... 94 Statistical Analysis ............................................................................................ 94 Results and Discussion ........................................................................................... 95 MS Analysis of UV exposed and Unexposed Spinal Cord ............................... 95 PCA of UV exposed and Unexposed Spinal Cord ............................................ 96 MSn Analysis of Oxidized Phospholipids in the Spinal Cord ............................. 98 Unsaturated phosphatidylcholines ............................................................. 98 Long chain oxidation products ................................................................. 100 Short chain oxidation products ................................................................. 103 Conclusions .......................................................................................................... 105 5 CONCLUSIONS AND FUTURE WORK ............................................................... 121 Conclusions .......................................................................................................... 121 Future Work .......................................................................................................... 124 LIST OF REFERENCES ............................................................................................. 126 BIOGRAPHICAL SKETCH .......................................................................................... 134

PAGE 8

8 LIST OF TABLES Table page 2 1 Nomenclature of the OxPC standards analyzed. ................................................ 53 2 2 Product ions observed in the MS2 spectra of [M+H]+ ions of PAzPC, PONPC, PGPC, POVPC, and KOdiAPC. .......................................................................... 54 2 3 Product ions observed in MS2 spectra of [M+Na]+ ions of PAzPC, PONPC, PGPC POVPC, and KOdiAPC. .......................................................................... 56 2 4 Product ions observed in MS3 3)3]+ ions of PAzPC, PONPC, PGPC, POVPC, and KOdiAPC. ........................................................... 58 3 1 PE standards autoxidized and analyzed. ............................................................ 79 3 2 Nomenclature of the observed PE oxidation products. ....................................... 80

PAGE 9

9 LIST OF FIGURES Figure page 1 1 Glycerophospholipid structure and select head groups. ..................................... 29 1 2 Two types of phospholipid oxidation products investigated in this work. ............ 30 1 3 Structures of the three MALDI matrices evaluated for the analysis of oxidized phospholipids. ..................................................................................................... 31 1 4 Schematic of the MALDI process ...................................................................... 32 1 5 Schematic of the Thermo Scientific MALDI LTQ XL used in this work. .............. 33 1 6 Schematic of the twodimensional LIT adapted from Schwartz et al. Ions transmitted through the front section along the z axis and are trapped in the center section by DC axial trapping and RF radial trapping. ............................... 34 1 7 Mathieu stability diagram of the twodimensional LIT. Since LITs are typically operated in RF only mode (ax = 0), ions are successfully trapped along the qx axis with a low mass cut off (LMCO) qx = 0.908. ................................................ 35 1 8 Sample mass spectra illustrating an MS2 experiment ........................................ 36 1 9 Mass spectrometric imaging (MSI) workflow. ..................................................... 37 2 1 MS2 product ion spectrum of m/z 666, the [M+H]+ ion of PAzPC. The structure and proposed fragmentation of the ion at m/z 666 are also shown. .... 55 2 2 MS2 product ion spectrum of m/z 688, the [M+Na]+ ion of PAzPC. The structure and proposed fragmentation of the ion at m/z 688 are also s hown. .... 57 2 3 MS3 product ion spectrum of m/z 3)3]+ ion of PAzPC. The structure and proposed fragmentation of the ion at m/z 629 are also shown. ......................................................................................................... 59 2 4 MS4 product ion spectrum of m/z 4(CH2)2N(CH3)3]+ ion of PAzPC. The proposed structure and fragmentation of the ion at m/z 505 are also shown. .......................................... 60 2 5 MS4 product ion spectrum of m/z 4(CH2)2N(CH3)3]+ ion of PONPC. The proposed structure and fragmentation of the ion at m/z 489 ar e also shown. .......................................... 61 2 6 MS4 product ion spectrum of m/z 4(CH2)2N(CH3)3]+ ion of PGPC. The proposed structure and fragmentation of the ion at m/z 449 is also shown. ............................................. 62

PAGE 10

10 2 7 MS4 product ion spectrum of m/z 4(CH2)2N(CH3)3]+ ion of POVPC. The proposed structure and fragmentation of the ion at m/z 433 is also shown. ............................................. 63 2 8 MS4 product ion spectrum of m/z 4(CH2)2N(CH3)3]+ ion of KOdiAPC. The proposed structure and fragmentation of the ion at m/z 503 are also shown. .......................................... 64 2 9 Schemes for the proposed intramolecular retroene mechanism. ...................... 65 2 10 Identification and imaging of two proposed OxPCs at m/z 716. The MS spectrum illustrates the complex mixture of biomolecules observed in tissue ... 66 3 1 MS spectrum of POPE A) before and B) after autoxidation. ............................... 81 3 2 Fragmentation pattern of [PAzPE+Na]+ from autoxidized POPE. ....................... 82 3 3 Fragmentation pattern of [PSuPE+Na]+ from autoxidized PDPE. ....................... 83 3 4 Fragmentation pattern of [SSuPE+Na]+ from autoxidized SDPE. ....................... 84 3 5 Fragmentation pattern of [PONPE+Na]+ from autoxidized POPE. ...................... 85 3 6 Fragmentation pattern of [POBPE+Na]+ ( m/z 560) from autoxidized PDPE ...... 86 3 7 Fragmentation pattern of [SOBPE+Na]+ ( m/z 588) from autoxidized SDPE ...... 87 3 8 MS2 + ( m/z 668) from autoxidized POPE. The proposed structure and sites of fragmentation are also shown. .................. 88 3 9 MS2 + ( m/z 652) from autoxidized POPE. The proposed structure and sites of fragmentation are also shown. .................. 88 3 10 MS2 + ( m/z 598) from autoxidized PDPE. The proposed structure and sites of fragmentation are also shown. .................. 89 3 11 MS2 + ( m/z 582) from autoxidized PDPE. The proposed structure and sites of fragmentation are also shown. .................. 89 3 12 MS2 + ( m/z 626) from autoxidized SDPE. The proposed structure and sites of fragmentation are also shown. .................. 90 3 13 MS2 + ( m/z 610) from autoxidized SDPE. The proposed structure and sites of fragmentation are also shown. .................. 90 4 1 MALDI MS spectrum from the l eft half of the tissue, which was exposed to UV for 4 hours (top) and from the right half of the tissue, which was not exposed to UV (bottom). ................................................................................... 106

PAGE 11

11 4 2 Principal component 1 scaled loadings plot (left) and scores images (right) from the MSI data set in Figure 4 1. Principal component 1 dict ates separation of the oxidized and unoxidized sides of the spinal cord.. ................ 107 4 3 MSn of PC(16:0/18:1), an abundant unsaturated phospholipid in spinal cord tissue ............................................................................................................... 108 4 4 MS2 product ion spectrum of m/z 814, a proposed long chain oxidation product o f phosphatidylcholine. ........................................................................ 109 4 5 MS3 product ion spectrum of m/z mapping the intensity of the product ions corresponding to the NLs of cyclophosphane (124 u), sn 1 moiety (256 u), and sn 2 moiety (314 u) .......... 110 4 6 MS4 product ion spectrum of m/z corresponding to the NLs of the sn 1 and oxidized sn 2 fatty acid substituents confirm the addition of two oxygens to the sn 2 moiety. ................................... 111 4 7 MS2 product ion spectrum of m/z 842, a proposed long chain oxidation product of phosphatidylcholine. ........................................................................ 112 4 8 MS3 product ion spectrum of m/z .................................................... 113 4 9 MS4 product ion spectrum of m/z corresponding to the NLs of the sn 1 and oxidized sn 2 fatty acid substituents confirm the addition of two oxygens to the sn 2 moiety. ................................... 114 4 10 MS2 product ion spectrum of m/z 688, a proposed short chain oxidation product of phosphatidylcholine. ........................................................................ 115 4 11 MS3 product ion spectrum of m/z 29 .................................................... 116 4 12 MS4 product ion spectrum of m/z ........................................... 117 4 13 MS2 product ion spectrum of m/z 716. Two possible isomeric short chain oxidation products of phosphatidylcholine occur at m/z 716. ............................ 118 4 14 MS3 product ion spectrum of m/z PC(16:0/11:0 COOH) and PC(18:0/9:0 COOH) are marked with stars and diamonds, respectively. .................................................................................... 119 4 15 MS4 product ion spectrum of m/z fragmentation, the presence of two isomeric oxidized PCs, PC(16:0/11:0 COOH) and PC(18:0/9:0 COOH), was confirmed. ............................................ 120

PAGE 12

12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CHARACTERIZING, IDENTIFYING, AND IMAGING OXIDIZED PHOSPHOLIPIDS BY MATRIX ASSIS TED LASER DESORPTION/IONIZATION TANDEM MASS SPEC TROMETRY By Whitney Leigh Stutts August 2013 Chair: Richard A. Yost Major: Chemistry Understanding the involvement of lipid oxidation in both inflammatory processes and the pathogenesis of various diseases is an important endeavor Despite growing inter est in the field of lipidomics, lipid oxidation products remain poorly characterized. However, mass spectrometry, coupled with soft ionization techniques, has become a key tool for investigating lipids and elucidating structural changes resulting from oxid ative modifications. In this research, matrix assisted laser desorption/ionization tandem mass spectrometry (MALDI MSn) techniques were developed for characterizing, identifying and imaging oxidized phospholipids Specifically, the research presented in t his work focused on three distinct aspects of oxidized phospholipid analysis. Initially, the MALDI MSn (where n = 2, 3, or 4) fragmentation pathways of short chain phosphatidylcholine oxidation products (OxPCs) were characterized. This study identified several fragmentation pathways that allowed for unambiguous detection of low abundant OxPC species, and t he developed method was applied for the identification of these species in rat spinal cord tissue. Next, a similar study was conducted for phosphatidylethanolamine oxidation products (OxPEs).

PAGE 13

13 In this second study, novel fragmentation pathways were identifi ed for short chain aldehydic OxPEs. This proposed pathway resulted in a macrocyclic gas phase rearrangement that can be exploited for MSn detection of these OxPEs in biological samples. The final study presented in this work describes a methodology utilizi ng principal component analysis (PCA) for untargeted detection of oxidized phospholipids following mass spectrometric imaging (MSI) experiments. In summary, t his work demonstrates that MALDI MSn provides a selective method for investigating individual molecular species resulting from phospholipid oxidation. Utilizing the MSn capabilities of the linear ion trap, OxPLs were distinguished from isobaric, and even isomeric, phospholipids. Furthermore, the utility of MALDI MSn in conjunction with multivariate dat a analysis techniques, namely PCA, for discovering, identifying and imaging individual phospholipid oxidation products in spinal cord tissue is exemplified.

PAGE 14

14 CHAPTER 1 ANALYSIS OF OXIDIZED PHOSPHOLIPIDS BY MATRIX ASSISTED LASER DESORPTION/IONIZATION TANDEM MASS SPECTROMETRY Research Motivation Increasing evidence suggests that oxidized lipids are markers and pathogenic factors in a variety of disease states.1 Lipid oxidation has been implicated in numerous diseases including Alzheimers,2 4 age related macular degeneration,5 atherosclerosis,5 6 cataractogenesis,5 multiple sclerosis,7 and rheumatoid arthritis.8 Furthermore, several studies have demonstrated the involvement of oxidative stress and lipid oxidation in central nervous system injuries, including stroke, traumatic brain injury, and spinal cord injury.9 Thus, there is growing interest in developing methods to identify potential biomarkers of oxidative damage or disease, which may lead to a b etter understanding of the involvement of lipid oxidation produc ts in these various diseases With the advent of soft ionization techniques such as electrospray ionization (ESI)10 and matrix assisted laser desorption/ionization ( MALDI ) ,11, 12 mass spectrometry (MS) has become an indispensable tool in the emerging field of oxidative lipidomics.13 Although still in a state of infancy, several review articles have provided the framework for and outline d the goals and challenges associated with this field.13 15 Thus far, oxidative lipidomics studies have primarily focused on ESI of samples in a liquid state such as brain lipid extracts; however, the potential of MALDI for in situ identification and imaging has also been proposed.13 16 This work makes significant contributions to the field of oxidative lipidomics as the first studies are described reporting MALDI tandem mass spectrometry (MSn) characterization of both short and long chain phospholipid oxidation products (OxPLs) and MS imaging of these products in intact tissue sections.

PAGE 15

15 Lipids and Lipid Oxidation Lipids are vital structur al components of cell membranes and are known to h ave various biochemical roles including energy storage, maintenance of electrochemical gradients, and nerve cell insulation.17, 18 D efined as water insoluble biomolecules that are highly soluble in organic solvents ,19 l ipids can be classified into storage and membrane lipids .20 Membrane lipids can be sub classified as phospholipids and g lycolipids and then further classified according to the chemical structure of the lipid backbone (i.e., glycerol or sphingosine) .20 Glyc erophospholipids (GPLs), ofte n referred to as phospholipids although not all phospholipids contain a glycerol backbone, are abundant in cell membranes and lipoproteins.21 These lipids are amphipathic (i.e., contain both hydrophobic and hydrophilic regions) and can, therefore, aggregate into bilayers.22 GPLs consist of a glycerol backbone, a phosphatecontaining polar headgroup at the sn 3 position of the glycerol backbone, and two nonpolar fatty acid tails esterified to the sn 1 and sn 2 positions (Figure 11).1 Depending on the nature of the polar head group at the sn 3 position, GPLs can be subdivided into distinct classes including phosphatidylcholines (PCs), phosphatidylethanolamines (PEs) and phosphatidylserines (PSs) as shown in Figure 11 .23 In eukaryotic cells, phospholipid fatty acid moieties vary in length, typically having an even number of carbons between 14 and 24.24 These fatty acids can be saturated (i.e., contain no double bonds) or unsaturated (i.e., contain one or more double bonds). Saturated fatty acids are typically bound to the sn 1 position, whereas unsaturated fatty acids, which have between one and six double bonds are preferentially bound to the sn 2 position.25 As saturated fatty acids are not as prone to oxidative modification as

PAGE 16

16 mono or poly unsaturated fatty acids, most oxidized phospholipids are modified at the sn 2 position.26 Phospholipid Oxidation The term oxidized phospholipids was coined in 1939 by Fredrick and Mary Bernheim, who were investigating the catalytic effects of vanadium salts on phospholipid oxidation in heart and brain tissue.27 Since the 1940s, studies on the mechanisms of lipid oxidation have been ongoing. Although there are various mechanisms for lipid oxidation, the most commonly referenced mechanism is the classic free radical chain reaction that proceeds through three steps: initiation, propagation, and termination.28 The first step, initiation, occurs w hen an initiator abstracts an electron from the unsaturated fatty acid moiety of a phospholipid. Initiators of phospholipid oxidation, summarized in a recent review article by Bochkov et al. and commonly referred to as reactive oxygen species (ROS), can be non radical species (e.g., ozone (O3) or singlet oxygen (O2 1)) or radical species (e.g., superoxide anion radical (O2 ) or hydroxyl radical (OH)).1 Furthermore, several catalysts of phospholipid oxidation have been reported including redox active metals and ultraviolet (UV) light.29 Once a hydrogen is abstracted from the unsaturated phospholipid, the carboncentered radical (L) reacts with molecular oxygen to produce a lipid peroxyl radi cal (LOO). The propagation step involves the subsequent reaction of the lipid peroxyl radical with other unsaturated phospholipids in the cell membrane; thus, the reaction is self propagating.29 Although the overall lipid oxidation chain reaction is not expected to fully stop, termination is used to describe the event that occurs when radicals recombine or otherwise terminate to form nonradical products. Greater detail on the theoretical

PAGE 17

17 aspects of lipid oxidation and generation of oxidized phospholi pids in vivo is provided by Schaich29 and Bochkov et al.,1 respectively. In vitro oxidation of phospholipids may be initiated through three pathways: enzymatic; nonenzymatic, freeradical; and nonenzymatic, nonradical. However, it is difficu lt to determine which method most accurately mimics the naturally occurring oxidation in living cells.30 Enzymatic oxidation of phospholipids has been investigated using low density lipoproteins and lipoxygenases.3133 Non enzymatic, free radical oxidation has been studied using H2O2/Fe2+, H2O2/ Cu2+, or t butylhydroperoxide to generate hydroxyl radicals.14 Additionally, t he n onenzymatic, free radical pathway has also been promoted by autoxidation, which involves the reaction of phospholipids wi th trace amounts of metals (e.g., iron) or phospholipid hydroperoxides.34 Previous studies reporting nonenzymatic, nonradical oxidation have exposed phospholipids to singlet oxygen or ozone to produce lipid oxidation products.1 35 For the experiments detailed in this work, autoxidation involving the exposure of unsaturated phospholipids to ambient air and light was used to promote oxidation of phospholipid standards. Furthermore, in vitro oxidation in tissue was catalyzed by UV light. Follo wing oxidative stress, many phospholipid oxidation products with varying structures and biological activities are formed.1 14 These products are classified according to the nature of the modification: 1) long chain products, formed by insertion of oxygen atoms; 2) short chain products, formed by oxidative cleavage of the uns aturated fatty acid tails; and 3) phospholipid adducts, formed by reaction between oxidation product s a nd/or molecules containing nucleophilic groups.14 Additionally,

PAGE 18

18 highly reactive degradation products of lipid peroxidation such as low molecular weight aldehydes (e.g., 4hydroxynonenal and malondialdehyde) may be formed.26 In this work, the first two types of oxidation products, long chain and short chain, were investigated; examples of these products are shown in Figure 12. Due to the relatively high abundance in animal tissue,22 PCs and PEs were the primary phospholipids investigated. Moreover, previous studies reported that most oxidized phospholipids detected in mammalian tissue contained a choline head group or, in the case of retina, an ethanolamine head group.26 Techniques for M easuring P hospholipid O xidation Numerous techniques for the detection of phospholipid oxidation have been repor ted.36, 37 Biochemical assays, such as the thiobarbituric acid (TBA ) assay ,38 have been widely used t o study lipid oxidation, but these assays lack the selectivity needed for structural elucidation. 36, 39 Another commonly used method for determining the extent of phospholipid oxidation is the measurement of diene conjugation by UV absorbance at 234 nm ; however, this method also suffers from poor selectivity.36 MS which characterizes biomolecules based on mass to charge ratios ( m/z ) has become increasingly important in phospholipid oxidation research.14 Several studies have investigated phospholipid oxidation by gas chromatography (GC ) MS and liquid chromatog raphy (LC) MS; however, these methods require extraction and often deriv ati zation.14, 36 Most of the recent MS studies of oxidized phospholipi ds use electrospray ionization.14 Although ESI has several advantages, such as the ability to be coupled to high performance LC (HPLC) there are als o many drawbacks. One limitation is that conventional ESI requires analytes to be in solution;40, 41 thus, direct analysis of intact

PAGE 19

19 tissue is not possible. Furthermore, salts and other impurities found in the analyte solution may cause ion suppression in ESI.42 These limitations can be overcome by matrix assisted laser desorption/ionization, as it is not limited to analytes in solution and is also less sensitive to salts and other impurities found in complex sample matric es;39 43 thus, MALDI is amenable to the analysis of intact biological tissue.44 In terms of lipid oxidation studies, an other important advantage of using MALDI is the minimal sample handling and preparation required, which reduces the likelihood of artifactual oxidation products.40, 43 The application of MALDI for lipid oxidation studies has been described previously; yet, in these experiments, MALDI is most often coupled to a timeof fligh t (To F) mass analyzer.14 45, 46 For the experiments described in this work, a linear ion trap (LIT) will be used as the mass analyzer. The primary advantage of the LIT is the MSn capabilities, which provide enhanced selectivity resulting in more reliable compound identification.43 47 Furthermore, MALDI LIT MSn is a powerful technique for mass spectrometric imaging of lipids4850 and, as shown in this work, lipid oxidation products. Matrix Assisted Laser Desorption/I onization MALDI was first introduced by Karas and Hillenkamp in 1985.51 Over the past 28 years, MALDI developed into an indispens a ble t echnique for MS analysis of labile and nonvolatile molecules .43 Since its introduction, MALDI has been widely used for the investigation of large biomolecules including peptides and proteins.43 Recent developments in sample prep aration, matrix selection, and applications of tandem mass spectrometry have revealed the utility of MALDI for the analysis of small molecules such as drugs, metabolites, and lipids.52, 53

PAGE 20

20 MALDI is a soft ionization technique (i.e., gas phase ions are produced without extensive fragmentation) that uses a pulsed laser beam and a chemical matrix to generate gas phase ions from both small and large molecules. The MALDI matrix typically consists of a small organic acid that has a strong molar absorptivity at the laser 337 nm.43 However, there is no single MALDI matrix or sample preparation protocol for MALDI MS analyses. Thus, the MALDI matrix solution and sample preparation methods must be developed for the analytes o f interest. Matrix S election The view that selection of a proper mat r ix is critical in MALDI MS analysis is widely held and commonly cited.40, 43, 54 In addition to strong absorptivity at the wavelength of the MALDI laser, an ideal matrix typically has a low vapor pressure, appropriate crystallization properties, and sufficient gas phase acidity/basicity.54 One of the most commonly used matrices for phospholipid analysis is 2, 5 dihydroxybenzoic acid (DHB) .43, 55 In addition, 2,4,6t rihydroxyacetophenone (THAP) is commonly used for the analysis of phospholipids and offers the advantage of fewer matrix background ions.45 A lesser known MALDI matrix, 6aza 2 thiothymine (ATT), was recently evaluated for the analysis of oxidized phospholipids and was found to offer higher sensitivity and less fragmentation than either DHB or THAP.45 The structures of these three matrices are illustrated in Fi gure 1 3. Once a MALDI matrix is selected, parameters such as solvent systems and additives may then be optimized. When choosing a solvent system, the solubility of the MALDI matrix and the analyte should be considered. A proper solvent system should be ca pable of dissolving the analyte and also allow for a saturated solution of MALDI

PAGE 21

21 matrix.56 Additionally, solvent systems should have sufficiently high vapor pressure to facilitate rapid evaporation and the formation of a microcrystal line film.56 Furthermore, s everal studies have illustrated the utility of adding salts (e.g., lithium acetate, potassium acetate, or sodium acetate) to MALDI matrix solutions for improving sensitivity, reducing spectral complexity, and enhancing the structural analysis of lipids by MALDI MSn.57, 58 One recent investigation by Griffiths and Bunch demonstrated that not only cation choice, but also salt type (e.g., acetates, chlorides, or nitrates) and concentration, impacted s pectral results.58 Thus, the selection of an additive should also be an important consideration for optimizing MALDI experiments. For MSn and imaging of the OxPLs investigated in this work, no significant differenc es in performance of DHB, THAP, and ATT were observed. Given that DHB was far less expensive than either ATT or THAP DHB was the primary MALDI matrix used in these studies. However, for analysis of low abundant, endogenous lipid oxidation products, ATT may be a more suitable matrix. Regarding solvent selection, this work used solvents including methanol, water, and chloroform. Lastly, sodium acetate was added to the matrix solutions used in this work because enhanced structural information was obtained fro m the sodium adducts of OxPLs. MALDI P rocess For MALDI MS analysis, the analyte solution and an excess of matrix are typically pipetted onto a stainless steel MALDI target plate.43 As the solvent evaporates, the matrix and analytes co crystallize. T he sample plate is then inserted into the mass spectrometer and rastered beneath a pulsed laser beam. When the laser strikes the sample surface, matrix analyte cocrystals a re desorbed into the gas phase plume consisting of positive ions, negative ions, and neutrals (Figure 1 4).

PAGE 22

22 The ionization process of MALDI is still under investigation; h owever, two theories for ion formation in MALDI have been proposed: the lucky survivor model59, 60 and the gas phase protonation model.61, 62 The lucky survivor model proposes that analytes retain their respective charge states from solution when they are incorporated into the matrix analyte cocrystals. Following desorption, ions that are not neutralized in the plume are detected and are socalled lucky survivors.59 In contrast, the gas phase protonation model postulates that neutral analytes undergo gas phase collisions with protonated [ma+H]+ or deprotonated [ma H] matrix leading to proton transfer and the detection of protonated [A+H]+ or deprotonated [A H] analyte.61 Recently, evidence sup porting both of these ionization pathways led to a unified model that combines the lucky survivor and gas phase protonation theories.63 Instrumentation In this work, the primary instrument used was the Thermo Scientific MALDI LTQ XL ( San Jose, CA ). A schematic of this instrument is displayed in Figure 15. The MALDI source is composed of an optics module and sample module. The optics module contains a stationary 60Hz N2 = 337 nm) a nd optics, which include lenses, mirrors, and neutral density filters that focus, direct and attenuate the laser beam as needed prior to reaching the sample target. This module is also equipped with a chargecoupled device (CCD) camera that cap tures optical images of the sample inside the instrument and provides a real time view of the sample during analysis. The sample module transfers the sample plate from atmospheric pressure to intermediate pressure (75 mTorr) and moves the sample target in the xy plane relative to the laser. Ion optics for transferring ions to the MS detector are also contained with in the sample module.64

PAGE 23

23 Follow ing Ionization in the MALDI source, ions are transmitted to the mass analyzer via a series of ion optics. The ions first pass through a square array of squareprofile rods, the Q00 quadrupole. Next, the ions are transmi tted through the Q0 quadrupole and the Q1 octo pole and into the mass analyzer which on this instrument is a linear ion trap (LIT) Linear Ion Trap In the Thermo LTQ XL, t he LIT mass analyzer is the site of storag e, isolation, dissociation, and ejection of ions.64 As depicted in the schematic in Figure 16, the two dimensional LIT is composed of a rectangular array of hyperbolic profile rods, which are divided into three sections: front, center, and back.65 The front and back sections are 12 mm in length, whereas the center section is 37 mm in length with 30mm 0.25mm ejection slits in the two x axis rods. To account for field imperfections due to these slits, the x axis rods are spaced farther apart (9.5 mm) than the y axis rods that are 8 mm apart.66 Ions are trapped in stable orbits within the center section by a timevarying radiofrequency (RF) electric field that is applied to opposite pairs of rods .67 The stability of ions in a LIT is governed by the solutions (Equations 1 1 and 12) to a set of nonlinear equation known as the Mathieu equations.67 = 8 ( 1 1 ) = 4 ( 1 2 )

PAGE 24

24 In these equations, e is the electronic charge of the ion, U is the applied DC potential, V is the applied AC potential, m is the mass of the ion, r0 is the radius of an inscribed circle of the rod array, and is the angular frequency.67 Ions have stable trajectories ins ide the trap when they are at ax, qx values that fall within the shaded portion of the Mathieu stability diagram (Figure 17).65 However, mass analysis in the LIT is performed by mass selective instability scanning, whereby the LIT is operated in RF only mode (i.e., U = 0 and, therefore, ax = 0); thus, allowing for the expression of ion stability in terms of qx only. After trapping, mass analysis is performed by linearly increasing the amplitude of the RF potential and simultaneously applying a resonance excitation voltage across the two x electrodes. As the resonance excitation frequency reaches the secular frequency of a specific m/z the ions at this m/z value are radially ej ected through the slits in the two center section rods.66 Although not shown in Figure 15, ions radially ejected from the LIT strike one of two conversion dynodes on either side of the LIT creating secondary charged particles, that are then detected by electron multipliers.64 The current on the multipliers is amplified and then the onboard data acquisition computer synchronizes the time of the ions ejection to the current to produce a mass spectrum. In si ngle stage full scan mode (MS), ions are stored in the LIT and sequentially scanned out o f the ion trap by mass selective instability to produce a mass spectrum.64 MS spectra yield profiles of ions within a given sample and are ideal for initial comparison of two data sets (e.g., control vs diseased). Additionally, one of the primary advantages of the LIT mass analyzer is the MSn capabilities, w hich allow for enhanced selectivity and structural information. In contrast to many other tandem mass analyzers

PAGE 25

25 (e.g., ToF ToF or QqToF) that can only perform up to two stages of mass analysis, the LIT can perform up to 10 stages of mass analysis. Tandem M ass Spectrometry Tandem mass spectrometry (MSn) is valuable technique f or characterizing phospholipids, allowing for the elucidation of phospholipid structure and differentiation of isobaric species .68 Although MSn can be performed on a number of mass analyzers, this work will focus on MSn in ion traps. Once in the ion trap, ions within a certain m/z range can be isolated and then fragmented by ion activation methods, the most common of which is collision induced dissociation ( CID) .69 CID occurs by a twostep mechanism involving excit ation and fragmentation of precursor ions.69 First, the mass selected precursor ions are excited by an applied resonant excitation voltage. As these ions gain translational energy inelastic collisions occur between the ions and the helium damping gas resulting in the transfer of part of the ions kinetic e nergy into internal energy Second, f ragmentation occurs when the precursor ion s gain enough internal energy to dissociate into one or more product ions and neutrals.69 These product ions are stored and then ejected from the trap in order of increasing m/z by linearly increasing the main RF voltage and applying a supplementary resonant ejection voltage, which is used to improve mass resolution in the LIT.66 The result of this technique is a product ion (MS2) spectrum consisting of fragment ion s from the massselected p recursor ion as shown in Figure 18 During storage of the MS2 product ions, but prior to scanout, a product ion can be isolated and further fragmented by CID (MS3).43 Assuming sufficient isolation and fragmentation efficiency, with each additional stage of MS selectivity is gained and enhanced structural information is often attained. For the identification of phospholipid

PAGE 26

26 oxidation products in this dissertation, multiple stages of MS (up to MS4) were used. MSn is also applicable to mass spectrometric imaging (MSI) experiments; thus images of product ions can be generated allowing the spatial distributions of isobaric ions to be differentiated .70 Mass Spectrometric Imaging MSI has emerged as a powerful technique for investigating regional distributions of specific lipids in biological tissues71 including brain,68, 72 heart ,50 lung,73 and spinal cord .48 This technique, which has been the subject of several recent review articles,44, 74, 75 combines the multichannel ( m/z ) measurement capabilities of mass spectrometers with surface sampling methods (e.g., MALDI) that allow for retention of spatial ly resolved chemical distributions.7 6 Furthermore, MSI data can be correlated to anatomical structures in tissues, which may offer insight into a number of complex biochemical processes that occur within living organisms.76 In MSI experiments, thin tissue sections are directly interrogated in a distinct pattern so that the spatial distribution s of target compounds within the tissue section may be observed.43, 70 Figure 1 9 illustrates the typical MSI workflow. MSI experiments are performed by first sectioning the frozen tissue on a cryostat (temperature of ca. 25 C) into thin (10 20 m) sections. For applications outside of MS, such as histology, the tissue sample is usually affixed to the sample stage using an optimal cutting temperat ure medium (OCT); however, OCT is formulated from w ater soluble glycols and resins and typically contains a number of compounds, such as benzalkonium chloride, that can interfere with mass spectrometric analysis.77, 78 Th erefore, for MSI analysis, HPLC grade water is often used to freeze the tissue to the sample stage.

PAGE 27

2 7 Once cut, the frozen tissue sections are transferred onto a cold MALDI target surface that is usually a glass or conductive glass microscope slide or, in some cases, a metal surface such as stainless steel.79 Sections are then thaw mounted by localized warming of the tissue by plac ement of a gloved finger on the back side of the MALDI target, causing the tissue to adhere to the target surface. Once collected, the tissue sections on the MALDI target are dried ( ca. 30 min) in a vacuum desiccator and coated with MALDI matrix V arious tec hniques for depositing matrix atop tissue samples for MSI have been reported including dry powder coating,80 electrospray deposition,77 81 inkjet printing,82 pneumatic spraying,77, 83 and sublimation.84 Next, the tissue is rastered with respect to the laser in a predefined, equal stepsize pattern. At each spot where the tissue is i rradiated, a mass spectrum is recorded along with its rel ative position. Once the data are acquired, imaging software is employed to extract and display the intensity of individual ions as a function of position.85 Overview of Dissertation To better understand phospholipid oxidation and the involvement of oxidation products in the development and progression of disease, methods for identifying these products in biological samples must be developed. The purpose of this research was to utilize MALDI MSn to develop such methods. In the following chapters, MSn is used as a selective method for characterizing, identifying, and imaging phospholipid oxidation products. Chapter 2 details the MS2, MS3, and MS4 fragmentation pathways of various PC oxidation products and identifies structurally informative product ion s. This chapter illustrates the utility of the MSn capabilities on the LIT for providing enhanced selectivity, which will be important for detection of these low abundance oxidation products in

PAGE 28

28 complex biological samples. Chapter 3 describes the use of MALDI MSn for identifying PE oxidation products following autoxidation of unsaturated PE standards. Various structurally informative product ions were discovered that allowed for confirmation of oxidative modification to the sn 2 fatty acid substituent. Furth ermore, key differences in the fragmentation of OxPE aldehyde and carboxylic acid derivatives are discussed. In Chapter 4, principal component analysis and MALDI MSn imaging were combined to identify and localize lipid oxidation products in spinal cord tis sue following in vitro oxidation. This chapter illustrates the potential of the MALDI MSn methodologies developed in this work for identifying and imaging lipid oxidation products in tissue from animal models of oxidative stress, injury, or disease. Lastly Chapter 5 provides conclusions to the work presented within this dissertation and insight into future experiments that could expand upon this work to further the understanding of the biological roles of phospholipid oxidation products in human health and disease.

PAGE 29

29 Figure 11. Glycerophospholipid structure and select head groups.

PAGE 30

30 Figure 1 2. Two types of phospholipid oxidation products investigated in this work. O OO P O N+O O-O O O O H OOOPON+OO-OOO Long Chain Oxidation Product Short Chain Oxidation Product

PAGE 31

31 Figure 13. Structures of the three MALDI matrices evaluated for the analysis of oxidized phospholipids.45 OHOHOHO O H O H O H O NNHNHOS DHB THAP ATT MW=154.12 MW=168.15 MW=143.16

PAGE 32

32 Figure 14. Schematic of the MALDI process. Adapted from Chughtai et al.44 Matrix Analyte MS ++ + + N 2 Laser

PAGE 33

33 Figure 15. Schematic of the Thermo Scientific MALDI LTQ XL used in this work. Adapted from Strupat et al.86 and G arrett et al.68

PAGE 34

34 Figure 16. Schematic of the twodimensional LIT adapted from Schwartz et al.66 Ions transmitted through the front section along the z axis and are trapped in the center section by DC axial trapping and RF radial trapping. Front Section Center Section Back Section

PAGE 35

35 Figure 17. Ma thieu stability diagram of the twodimensional LIT. Since LITs are typically operated in RF only mode (ax = 0), ions are successfully trapped along the qx axis with a low mass cut off (LMCO) qx = 0.908. Adapted from Krishnaveni et al.65 axqx

PAGE 36

36 Figure 18. Sample mass spectra illustrating an MS2 experiment. First, precursor ions of a chosen m/z value are selected for MS2 analysis (top). Next, the precursor ions at this m/z are isolated (middle) and CID is applied to produce the MS2 product ion spectrum (bottom). 790 795 800 805 810 815 820 825 830 m/z 0 20 40 60 80 100Relative Abundance 810.8 798.8 826.8 822.9 814.8 792.8 804.8 818.829 MS Isolation MS2 550 600 650 700 750 800 850 m/z 0 20 40 60 80 100Relative Abundance 755.5 814.6 631.5 771.5 690.5 737.5 550 600 650 700 750 800 850 m/z 0 20 40 60 80 100Relative Abundance 814.6CID=30

PAGE 37

37 Figure 19 Mass spectrometric imaging (MSI) workflow. 770 775 780 785 790 795 800 805 810 815 820 825 830 835 m/z 0 10 20 30 40 50 60 70 80 90 100 Relative Abundance 782.75 810.75 798.67 832.75 822.75 778.83 794.75 772.75 788.75 776.83 828.67 1 mm 1 mm1. Section tissue (~10 m) 2 Thaw mount tissue on glass slide 3. Coat tissue with matrix solution 4. Raster tissue beneath the laser 5. Obtain position specific MS spectra 6. Extract to generate MS images

PAGE 38

38 CHAPTER 2 CHARACTERIZATION OF PHOSPHATIDYLCHOLINE OXIDATION PRODUCTS B Y MALDI MSn Introduction Lipid oxidation has been implicated in the pathogenesis and/or progression of various human disorders and diseases including Alzheimers,2 4 age re lated macular degeneration,5 a therosclerosis,5 6 cataractogenesis,5 multiple sclerosis,7 and rheumatoid arthritis.8 Probing lipid oxidation has proved challenging for a number of reasons, including the diversity of oxidation products, the i nstability of select oxidation products, and the sensitivity necessary to detect these oxidation products.87 Furthermore, several studies have illustrated the various important biological activities of lipid oxidation products and have shown that many of these activities (e.g., inflammatory vs anti inflammatory) are dependent on the specific chemical structure of the oxidized species;88 the type of modification to the sn 2 substituent (terminal carboxylic acid vs terminal aldehyde), the chemical bond linking the sn 1 substituent (ether vs ester), the fatty acid chain length, and the charge on the head group have all been shown to affect the biological activities of lipid oxidation products.8892 Thus, there is growing interest in developing methods for characterizi ng and identifying lipid oxidation products in biological samples.28 Biochemical assays, such as the thiobarbituric acid (TBA) assay,38 have been widely used to detect total lipid oxidation; however, these assays lack the selectivity needed for structural elucidation.39 Diene conjugation, measured by UV absorbance at 234 nm has also been utilized to determine the extent of lipid oxidation, but this method also suffers from poor selectivity.39 Mass spectrometry (MS) has become increasingly utilized in lipid oxidation research due to the ability to characterize biomolecules based

PAGE 39

39 on their mass to charge ( m/z ) ratios yielding superior selectivity relative to the aforementioned techniques.1 14, 15 Despite the various challenges, several MS methods have been developed for the analysis of the oxidatively modified phospholipids (OxPLs). Gas chromatography mass spectrometry (GC MS) and liquid chromatography mass spectrometry (LC MS) have been used for identification and quantitation of phospholipid oxidation products;9398 however, many of these methods require hydrolysis of the oxidized free fatty acid moieties from the phospholipid head group and extensive derivatization prior to analysis. More recently, soft ionization methods have been employed for the analysis of OxPLs without derivatization.30, 99108 Electrospray ionization mass spectrometry (ESI MS) combined with reversedphase high performance liquid chromatography (RP HPLC) has become the most widely used method for the analys is of OxPLs.15 This method offers several advantages. First, OxPLs readily ionize by ESI and RP HPLC offers chromatographic separation based on fatty acid composition. Second, these methods are amenable to tandem mass spectrometry (MSn), though in most cases only two stages of MS (i.e., MS2) have been perf ormed.14, 15 Matrix assisted laser desorption/ionization (MALDI) has also been used to investigate OxPLs,16 but few published studies have used this ionization method in spite of the many inherent advantages.45 In comparison to LC coupled with ESI MS, MALDI MS requires less sample and offers more rapid analysis. Furthermore, MALDI is less sensitive to salts and other impurities found in complex biological samples;39 in fact, addition of salts is often advantageous for the structural elucidation of various phospholipids by MALDI

PAGE 40

40 MSn.49, 50, 68, 106 Additionally, MALDI can be coupled to mass spectrometric imaging (MSI) techniques for in situ detection and localization of PLs in biological tissues.13, 4850 This investigation characterized oxidation products of phosphatidylcholines, the most abundant phospholipids in cell membranes ,109 by MALDI MSn using a linear ion trap (LIT) mass analyzer. The MSn (where n = 2, 3, or 4) capabilities of the LIT were exploited for enhanced selectivity resulting in more reliable characterization of the collision induced dissociation (CID) fragmentation pathways of OxPCs. Additionally, preliminary results illustrate the utility of the developed MALDI MSn method f or identifying and imaging these OxPCs in biological tissues. Experimental Methods Chemicals Avanti Polar Lipids ( Alabaster, AL) was the source for all lipid standards except 1 (palmitoyl 2 (5 keto 6 octenedioyl) sn glycero3 phosphatidylcholine ( KOdiA PC), which was purchased from Cayman Chemical (Ann Arbor, MI). MALDI matrix, 2,5 dihydroxybenzoic acid (DHB), was purchased from Acros Organics (Geel, Belgium). Sodium acetate trihydrate (NaOAc) and HPLC grade water and methanol were purchased from Fisher Scientific ( Fair Lawn, NJ ). Ethanol (200 proof) was purchased from Decon Labs (King of Prussia, PA). P reparation of OxPC Standards The short chain oxidation product (OxPC) standards listed in Table 21 were dissolved in cold, degassed ethanol to a concentr ation of 100 ppm. A MALDI matrix solution consisting of 40 mg/mL DHB in 70:30 methanol : water (v/v) and 10 mM NaOAc (final concentration) was prepared. Sodium acetate was added to enhance the relative intensity of the [M+Na]+ ions, which yielded more struct urally informative fragmentation.

PAGE 41

41 OxPC standards were deposited atop a 384well stainless steel MALDI plate using a modified drieddroplet method.12 For this metho d, 1 L of the 100 ppm OxPC standards followed by 1 L of MALDI matrix were pipetted into the sample wells. The solvent was allowed to evaporate in ambient air, resulting in matrix/analyte cocrystals. Preparation of Spinal Cord Tissue To illustrate the potential for identification and MSI of OxPCs, spinal cord tissue from the lumbar region of adult, female SpragueDawley rats (Harlan, San Diego, CA) was utilized. Animal studies were approved by the local IACUC at the University of California, San Diego, and were performed in an AAALAC approved vivarium. Following euthanasia, excised tissue was immediately flashfrozen in liquid nitrogen and stored at cryostat (Walldorf, Germany). To avoid analyte ion suppression resulting from tissue mounti ng media (e.g., optimal cutting temperature polymer),77 spinal cord tissue was fused to the cutting stage using HPLC grade water. Cross sections (10 m thick) were thaw mounted atop cold glass microscope slides and subsequently dried in a vacuum desiccator for approximately 30 min to remove excess water. The s lides were then spray coated with MALDI matrix using a glass Type A Meinhard nebulizer (Golden, CO) and nitrogen (30 PSI) as the nebulizing gas. Matrix was applied until a homogenous layer of matrix analyte cocrystals was observed over the entire tissue. MALDI MS Instrumentation Mass spectra were acquired in positive ion mode using an intermediatepressure (70 mTorr) MALDI LIT mass spectrometer, a Thermo Scientific MALDI LTQ XL (San Jose, CA). This ins trument was equipped with a 337nm nitrogen laser with a 60Hz repetition rate and 100 m laser spot diameter. Laser energies of 2.0 8 .0 J per laser

PAGE 42

42 shot and 3 4 laser shots per raster spot were used in these experiments; automatic gain control (AGC) was toggled off. For MSI experiments, images were generated using Thermo ImageQuest 1.0.1 software (San Jose, Ca). Mass spectral images generated following one stage of mass analysis (MS) were normalized to the total ion current (TIC). For images generated from higher stages of mass analysis, the intensity of the most abundant product ion was plotted at each pixel and was not normalized to TIC. For all other data processing, Thermo QualBrowser 2.0.7 (San Jose, Ca) was used. MSn Instrumental Parameters For MS2 experiments, a precursor ion isolation width of 1.2 u was utilized. For MS3, the precursor ion isolation widths for the first and second isolation events were adjusted to 1.5 and 2.0 u, respectively. Lastly, for MS4, isolation widths of 1.5, 2.0, and 2.5 u were utilized for the first, second, and third isolation events, respectively. Each mass spectrum presented represents an average of 100 analytical scans unless otherwise noted. MS2, MS3, and MS4 experiments were performed using CID. Collision energies were optimized to achieve maximum product ion intensity. In general, the precursor ion was depleted to <20% of the most intense fragment ion; however, without exception, the precursor ion was detected above the baseline. For precursor ions with higher m/z values (greater than m/z 599), the q of activation ( qact) was reduced to 0.22 (typically qact = 0.25 on the LTQ XL), when necessary, to decrease the low mass cutoff (LMCO). Specifically, the LMCO was decreased to observe the product ion at m/z 184 for protonated OxPCs.

PAGE 43

43 Accurate Mass Measurements To validate several proposed fragmentation pathways and product ion identifications, accurate mass measurements were performed using a 7T hybrid LIT Fourier transform ion cyclotron resonance (FTICR) mass spectrometer, the Finnigan LTQ FT (Thermo Fisher Scientific, Bremen, Germany). This instrument was equipped with a Thermo Scientific nanospray ionization (NSI) source. For these experiments, OxPC standards diluted to 1000 nM in ethanol were directly infused at a flow rate of 0.5 to +2 kV. The heated capillary on the n isolation and excitation were performed in the LIT, whereas mass analysis of the product ions was conducted in the ICR cell. AGC was toggled on for all experiments with a maximum i njection time of 100 ms. For MSn experiments on the LTQ FT, the parameters for precursor ion isolation widths were the same as those outlined above, with the exception of the third isolation event in MS4, which was lowered from 2.5 to 2.0 u. Additionally, collision energies were optimized as described above. The FT portion of the instrument was operated in wide scan mode at a resolving power of 100,000 with 50 analytical scans averaged to obtain accurate mass values. Results and Discussion Ionization of unm odified PCs by MALDI produces both protonated species [M+H]+, and alkali metal adducts such as [M+Na]+.14 Depending on the precursor ion selected, the fragmentation pathways observed by MS2 vary greatly.47 In contrast to the [M+H]+ ions of unmodified PCs, which yield one predominant MS2 product ion ( m/z 184) and little structural information, CID of the [M+Na]+ ions yields many structural ly informative product ions.68, 110 In this work, MALDI MS utilizing DHB as a positive mode

PAGE 44

44 matr ix produced both protonated and sodiated ions of OxPCs. Subsequently, the fragmentation pathways of the [M+H]+ and [M+Na]+ ions were explored for each of the OxPCs listed in Table 21. MS2 C haracterization of the [M+H]+ I ons of S hort Chain OxPCs The observed MS2 product ions resulting from CID of the [M+H]+ ions of PAzPC, PONPC, PGPC, POVPC, and KOdiAPC are listed in Table 2 2. Many of these product ions are in accordance with the ESI MS2 fragmentation of [M+H]+ ions of 1 palmitoyl 2 linoleoyl sn glyc ero 3 phosphocholine (PLPC) and 1 palmitoyl 2 arachidonoyl sn glycero 3 phosphocholine (PAPC) oxidation products reported by Reis et al. 104 Similar to unmodif ied PCs, following CID of the [M+H]+ ion, short chain OxPCs yielded a dominant product ion at m/z 184, corresponding to the protonated phosphocholine head group ( [H2PO4(CH2)2N(CH3)3]+).111, 112 Although much lower in abundance (0.5 4% relative abundance), the product ions corresponding to the neutral losses (NLs) of the fatty acid moieties, as free acids or as ketenes, were also observed. These product ions were reported previously in ESI MS2 studies of unmodified PCs and allowed for differentiation of positional isomers based on the relative abundances of the sn 1 and sn 2 ketene losses; the product ion corresponding to the sn 2=C=O) demonstrated greater abundance than the product ion due to the sn 1 ketene NL 1=C=O).112 This trend was also observed for all short chain OxPCs studied in this work; that is, the loss of the saturated sn 1 ketene was, in all cases, less abundant than the loss o f the oxidized sn 2 ketene. Furthermore, the modified sn 2 fatty acid chain was 2=C=O) while the sn 1 fatty acid chain was 1COOH).

PAGE 45

45 As a specific example, Figure 21 illustrates t hese fragmentation pathways for MS2 of the [M+H]+ ion of PAzPC ( m/z 666). The product ion corresponding to the phosphocholine head group, m/z 184, was the base peak; all other product ions were observed below 2% relative abundance. Upon 20 magnification o f the upper mass region of the spectrum (greater than m/z 184), product ions related to the fatty acid tails were observed. The losses of the sn 2 tail as the ketene (NL of 170 u ) and as the free acid (NL of 188 u ), although lower in abundance, were observ ed at m/z 496 and 478, respectively. Likewise, losses of the sn 1 tail as the ketene and as the free acid were observed at m/z 428 and 410, respectively; however, the sn 1 tail was preferentially lost as the free acid instead of the ketene. The trend menti oned above for the relative abundance of the sn 1 ketene loss and the sn 2 ketene loss is also illustrated in the MS2 product ion spectrum of m/z 666; that is, the ion corresponding to the sn 2 ketene loss ( m/z 496) was greater in abundance than that of the sn 1 ketene loss ( m/z 428). Although structural information may be obtained from CID of the [M+H]+ ions, the majority of the product ion intensity falls at m/z 184, a structurally uninformative fragment ion with regards to the substituents bound to the sn 1 and sn 2 positions of the glycerol backbone. Thus, fragmentation of the [M+H]+ ion presents a relatively inefficient method for determining fatty acid composition and position. Accordingly, the fragmentation pathways related to the [M+Na]+ ions of OxP Cs were explored. M Sn Characterization of t he [M+Na]+ I ons of Short C hain OXPC s The MSn product ion spectra of the short chain products studied illustrated distinct fragmentation patterns for the [M+Na]+ ions. For each of the OxPCs, the MS2 product ions resulting from CID of the [M+Na]+ precursor ions are listed in Table 23

PAGE 46

46 Many of these product ions are in agreement with ESI MS2 product ions previously observed following CID of PLPC and PAPC oxidation products.104 Characteristic MS2 fragmentation resulting in product ions produced from NLs of 59 u (trimethylamine) and 183 u (phosphocholine) were observed.47, 104, 113 Although these were the two most abundant product ions of each of the short chain OxPCs investigated in this work, product ions resulting from the NLs of the sn 1 and sn 2 fatty acid tails were also observed following CID. In MS2, the sn 1 tail was lost as the free acid or as the concurrent loss of the sn 1 free fatty acid and trimethylamine. Additionally, various losses of the sn 2 tail were observed for each of the OxPCs investigated. For all of the short chain OxPCs except PGPC the sn 2 tail was preferentially lost as the neutral, sodiated free acid; PGPC fragmentation also yielded a loss of the sn 2 tail as the neutral, sodiated free acid, but the sn 2 ketene loss (NL of 114 u ) was more abundant. Product ions resulting from the concurrent loss of the sn 2 tail and the trimethylamine were also observed at approximately equal or greater relative abundance than that of the product ions resulting from concurrent loss of the sn 1 tail and the trimethylamine. This fragmentation pathw ay has been previously reported for lithium adducts of unmodified PCs.112 However, the opposite trend was previously observed (i.e., the [M+Li R13)3]+ ion was greater in abundance than the [M+Li R23)3]+ ion).112 Therefore, t he oxidative modification to the sn 2 tail and/or the sodium likely influences this fragmentation pathway. Figure 22 illustrates the MS2 fragmentation pathways of the PAzPC [M+Na]+ ion ( m/z 688). CID of m/z 688 resulted in a predominant product ion at m/z 629, resulting

PAGE 47

47 from the NL of trimethylamine (NL of 59 u ). Loss of the phosphocholine head group (NL of 183 u ) was observed at m/z 505. Additionally, NLs of the sn 1 free fatty acid (NL of 256 u ) and the sodiated sn 2 free fatty acid (NL of 210 u ) were observed at m/z 432 and 478, respectively. At even lower relative abundances (<0.1%), but still detectable above baseline, product ions corresponding to the loss of the sn 2 free acid ( m/z 500), sn 1 free acid combined with trimethylamine ( m/z 373), and sn 2 free acid combined with trimethylamine ( m/z 441) were observed. Although MS2 provides useful structural information, in situ identification of these OxPCs would likely be confounded by isobaric species. Thus, further stages of MS were utilized for enhanced selectivity and increased structural information. For these studies, MS3 was performed on the product ions resulting from the loss of trimethylamine (NL of 59 u ). Table 24 summarizes the product ions and relative abundances observed in the MS3 spectra f 3)3]+ ions of the short chain OxPCs. As was shown previously for unmodified PCs,47 the predominant MS3 product ion observed resulted from the loss of the remaining portion of the phosphocholine head group (cyclophosphane; NL of 124 u ). Although much lower in abundance, MS3 of these OxPCs yielded product ions corresponding to the NLs of the sn 1 and sn 2 tails (both as the free acid). In contrast to MS2 fragmentation, the sodium was not lost with the sn 2 tail. Instead, a minor product ion corresponding to the NL of cyclophosphane with sodium (NL of 146 u ) was present in the M S3 spectra from each of the oxidation products studied. Furthermore, an MS3 product ion at m/z 415, proposed to be the loss of the sn 2 tail combined with the loss of acetylene (possibly from the remaining portion of the PC head group), was observed in the MS3 spectra from each of the short chain

PAGE 48

48 OxPCs. Lastly, structurally informative product ions, which were only observed for OxPCs containing a terminal carboxylic acid on the sn 2 fatty acid tail, resulted from the loss of CO2 (NL of 44 u ). Thus, the NL o f CO2 results in a diagnostic product ion that can be used to differentiate short chain OxPCs containing a terminal carboxyl group from those containing a terminal aldehyde. Figure 23 illustrates the MS3 product ion spectrum and fragmentation pathways of 3)3]+ ion ( m/z 629) of PAzPC. In addition to the product ions observed for all OxPCs investigated, MS3 product ions of PAzPC resulting from the concurrent loss of the sn 1 free acid and cyclophosphane or the sodiated cyclophosphane were obser ved at m/z 249 and 227, respectively. Interestingly, this MS3 fragmentation pathway was not observed for any of the other short chain products studied in this work. Furthermore, since PAzPC contains a terminal carboxylic acid group on the sn 2 tail, a loss of CO2 (NL of 44 u ) was observed, indicating cleavage of bond relative to the terminal functional group. MS4 HPO4(CH2)2N(CH3)3]+ ions varied drastically depending on the sn 2 fatty acid length and composition. Despite the variation, the observed product ions yield valuable structural information. Figure 24 illustrates the characteristic fragmentation observed in the MS4 product ion spectrum of m/z HPO4(CH2)2N(CH3)3]+ ion o f PAzPC. Although recent evidence has suggested that either a 5or 6 member cyclic structure forms for unmodified PCs prior to MS4, 112 at present, the exact structure of this ion is unclear due to the presence of the modified sn 2 fatty acid. For simplicity, this ion is displayed without the ring structure. Regardless of the structure, the predominant

PAGE 49

49 product ions at m/z 249 and 227 resulted from the NL of the sn 1 fatty acid tail and the sn 1 tail with sodium, respectively. Thes e product ions, in addition to the product ions at m/z 279, 211, 193, 171, and 153, which were attributed to [R1COOH+Na]+, [R2 COOH+Na]+, [R22O]+, [R22O]+, and [R22O]+, respectively, confirmed the identification of both fatty acid substituents of PAzPC. Furthermore, the NL of 56 u ( m/z 449) was observed in the MS4 product ion spectrum of PAzPC and likely corresponds to the loss of C3H4O as previously reported for MSn on the [M+Na]+ ions of triacylglycerols.114 However, this loss of 56 was observed in the MS4 product ion spectrum from each of the OxPCs investigated in this work and is therefore less structurally informative than the product ions related to the fat ty acid tails. The MS4 4(CH2)2N(CH3)3]+ ion of PONPC, PGPG, POVPC and KOdiAPC are depicted in Figures 2 5, 2 6, 2 7, and 28 respectively. Many product ions analogous to those described above for PAzPC were also observed in the MS4 spectra of the other OxPCs; however, MS4 product ions specific to the type of oxidative modification were also detected. One diagnostic fragmentation pathway was the loss of the sn 1 fatty acid tail; in contrast to the carboxylic acid derivatives, the OxPCs containing a terminal aldehyde lost the sn 1 tail (palmitic acid; 256 u) as the NL of 254 u rather than 256 u Based on previously reported MS2 fragmentation of [M+Li]+ ions of unmodified PCs115 and triacylglycerols,114 the NL of 254 u unsaturated fatty acid from the sn unsaturat ed fatty acid from the sn 2 substituent, not from the sn 1 substituent;115 thus the terminal aldehyde

PAGE 50

50 on the sn 2 substituent of the OxPCs is likely playing some role in the formation of this product ion. Furthermore, MS4 of the keto containing OxPC, KOdiAPC, yielded multiple fragmentation pathways not observed for other short chain OxPCs; specifically, cleavages between C2 C3 ( m/z 377) and C3 C4 ( m/z 367) of the sn 2 fatty acid tail were observed (Figure 2 8) To confirm these fragmentation pathways, accurate mass measurements were performed on the hybrid LTQ FT mass spectrometer. The FTICR mass spectra obtained from MS4 4(CH2)2N(CH3)3]+ ion from KOdiAPC exhibited product ions at m/z 377.2676 (error of 2.18 ppm) and m/z 367.2856 (error of 2.10 ppm), thereby confirming the proposed molecular formulae. Although the mechanism was not confirmed, one possible rationale for the C2 C3 cleavage is an intramolecular retroene reaction involving the C5 carbonyl group ( Figure 29 Scheme 1 cleavage of oxofatty acids.116 Although cleavage was observed relative to the C5 carbonyl group, the C3 C4 cleavage is thought to result from an intramolecular retro ene reaction involving the C1 carbonyl group ( Figure 29 Scheme 2), rather than 1,4hydrogen elimination as previously reported.116 1,4 hydr ogen elimination was considered; however, the six membered cyclic intermediate required for 1,4hydrogen elimination is not favorable, as the sn 2 tail does not contain four consecutive carbons with accessible hydrogens. Identifying and Imaging OxPCs in Sp inal Cord Tissue Following MSn characterization of OxPC standards, [M+Na]+ ions of potential OxPCs were targeted for in situ identification and MSn imaging. Figure 210 illustrates the identification and localization of two isobaric OxPCs in spinal cord ti ssue. The

PAGE 51

51 MALDI MS spectrum obtained by averaging the mass spectra across the entire tissue section (approximately 700 scans) depicts the complex mixture of biomolecules obtained from tissue. One potential PC oxidation product, PC(18:0/9:0 COOH), was targe ted for MSn analysis. Upon CID of the expected [M+Na]+ ion ( m/z 716), MSn product ions (indicated with purple diamonds) were analogous to those predicted based on the fragmentation pathways outlined in Tables 22, 2 3, and 24. Thus, product ions observed in the MS2, MS3, and MS4 spectra allowed for in situ identification of PC(18:0/9:0 COOH). However, additional MS2, MS3, and MS4 product ions (indicated with orange stars) indicated the presence of an isobaric ion at m/z 716. Based on these product ions, another OxPC was identified at m/z 716, [PC(16:0/11:0 COOH)+Na]+. After collecting the MSn spectra, extracted ion images were generated to determine the localization of the OxPCs at m/z 716 in tissue. The MS image of [PC(1 6:0/16:0)+Na]+ at m/z 756 ( normalized to the TIC) is provided to distinguish the gray matter (outlined in red) from the w hite matter of the spinal cord; this same outline is superimpos ed atop the other images, demonstrating the localization of these OxPCs in the gray matter. The MSn spectra and images depicted in Figure 210 illustrate the enhanced selectivity afforded by the LIT, allowing for identification and localization of OxPCs in situ. Conclusions This study has established that MALDI MSn is a power ful technique for characterizing various short chain oxidation products of phosphatidylcholines. Based on the characteristic MALDI MSn fragmentation of the [M+H]+ and [M+Na]+ ions of various short chain OxPCs, valuable structural information is obtained. I n MS2, CID of the [M+H]+ and [M+Na]+ ions of PAzPC, PONPC, PGPC, POVPC, and KOdiAPC led to

PAGE 52

52 various product ions, which were analogous to those previously reported for unmodified PCs. For additional structural information, the ion corresponding to the NL of 59 u from the [M+Na]+ was fragmented in MS3 and yielded multiple product ions, one of which (the NL of CO2) was diagnostic of a terminal carboxyl group on the sn 2 tail. Furthermore, MS4 of the [M+Na]+ resulted in drastically different fragmentation depending on the sn 2 f atty acid length and type of functional group(s) added. An MS4 fragmentation pathway characteristic of OxPCs containing a terminal aldehyde rather than a carboxylic acid was the NL of 254 u unsaturated sn 1 fatty acid substituent (16:1). Additionally, the presence of the keto group in KOdiAPC led to mid chain cleavages of the sn 2 fatty acid moiety. This work also illustrates the feasibil ity of this MALDI MSn methodology for the analysis and identification of individual PC oxidation products in complex mixtures including intact biological tissues. The diagnostic ions detailed in this study were utilized for targeted MALDI MSn imaging studi es to determine the distribution of OxPCs in tissue sections. Through the development of selective methods for identifying these OxPCs and determining their in situ localization, a greater understanding of the biological and physiopathological activities of these phospholipid oxidation products may be achieved.

PAGE 53

53 Table 2 1. Nomenclature of the OxPC standards analyzed .a Chemical Name [shorthand] R1 R2 Monoisotopic Mass 1 palmitoyl 2 (5 oxo valeroyl) sn glycero 3 phosphocholine [POVPC] 16:0 5:0 (CHO) 593.37 1 palmitoyl 2 glutaryl sn glycero 3 phosphocholine [PGPC] 16:0 5:0 (COOH) 609.36 1 palmitoyl 2 (9 oxo nonanoyl) sn glycero3 phosphocholine [PONPC] 16:0 9:0 (CHO) 649.43 1 palmitoyl 2 azelaoyl sn glycero 3 phosphocholine [PAzPC] 16:0 9:0 (COOH) 665.43 1 pal mitoyl 2 (5 keto 6 octenedioyl) sn glycero 3 phosphocholine [KOdiAPC] 16:0 8:1 (COOH)b 663.37 a Nomenclature according to the LIPID MAPS systematic naming is used (www.lipidmaps.org). R1 represents the number of carbons and double bonds (#C:#DB) in the fatty acid substituents esterified to the sn 1 position of the glycerol backbone. R2 represents the number of carbons and double bonds in the oxidatively truncated fatty acid substituent. These sn 2 acyl groups contain 5 9 carbon atoms and a terminal aldehyde (CHO) or carboxyl group (COOH). b In unsaturated carboxyl group, the sn 2 tail of KOdiAPC also contains a ketone at C5.

PAGE 54

54 Table 2 2. Product ions o bserved in the MS 2 s pectra of [M+H] + i ons of PAzPC, PONPC, PGPC, POVPC, and KOdiAPC a PAzPC PONPC PGPC POVPC KOdiAPC m/z 666 m/z 650 m/z 610 m/z 594 m/z 664 [H 2 PO 4 (CH 2 ) 2 N(CH 3 ) 3 ] + b 184 (100) 184 (100) 184 (100) 184 (100) 184 (100) 1 =C=O 428 (0.4) 412(1.6) 372(0.6) 356 (0.5) 426 (1.0) 2 =C=O 496 (1.1) 496 (3.4) 496 (1.1) 496 (0.7) 496 (2.5) 1 COOH 410 (1.2) 394 (3.5) 354 (1.2) 338 (0.8) 408 (1.4) 2 COOH 478 (0.5) 478 (2.1) 478(0.7) 478 (0.7) 478 (1.1) 4 (CH 2 ) 2 N(CH 3 ) 3 483 (0.3) 467 (1.2) 427 (0.3) 411 (0.5) 481 (1.2) 2 O 648 (1.5) NA 592 (1.1) NA 646 (9.6) 3 ) 3 NA 591 (0.4) NA NA 605 (8.9) 2 NA NA 566 (0.1) NA 620 (1.9) a Relative abundances (%) for each product ion are given in parentheses. The abundance was rounded to the nearest b Indicates a product ion rather than a NL.

PAGE 55

55 Figure 2 1. MS2 product ion spectrum of m/z 666, the [M+H]+ ion of PAzPC. The structure and proposed fragmentation of the ion at m/z 666 are also shown.

PAGE 56

56 Table 2 3. Product ions observed in MS2 spectra of [M+Na]+ i ons of PAzPC, PONPC, PGPC, POVPC, and KOdiAPC .a PAzPC PONPC PGPC POVPC KOdiAPC m/z 688 m/z 672 m/z 632 m/z 616 m/z 686 3 ) 3 629 (100) 613 (100) 573 (100) 557 (100) 627 (100) 4 (CH 2 ) 2 N(CH 3 ) 3 505 (12.2) 489 (7.5) 449 (15.1) 433 (10.5) 503 (7.2) 2 COONa 478 (0.3) 478 (0.1) 478 (0.2) 478( 0.3) 478 (0.2) 1 COOH 432 (0.2) 416 (0.1) 376 (0.1) 360 (0.1) 430 (0.1) 2 COOH 500 (0.1) 500 (0.1) 500 (0.3) 500 (0.2) 500 (0.2) 1 COOH & N(CH 3 ) 3 373 (0.1) 357 (0.1) 317 (0.1) 301 (0.1) 371 (0.1) 2 COOH & N(CH 3 ) 3 441 (0.1) NA 441 (0.4) 441 (0.3) 441 (0.2) 2 =C=O NA NA 518 (0.6) NA 518 (0.1) a Relative abundances (%) for each product ion are given in parentheses. The abundance was rounded to the nearest

PAGE 57

57 Figure 2 2. MS2 product ion spectrum of m/z 688, the [M+Na]+ ion of PAzPC. The structure and proposed fragmentation of the ion at m/z 688 are also shown.

PAGE 58

58 Table 2 4. Product ions observed in MS3 s3)3]+ i ons of PAzPC, PONPC, PGPC, POVPC, and KOdiAPC .a PAzPC PONPC PGPC POVPC KOdiAPC m/z 629 m/z 613 m/z 573 m/z 557 m/z 627 4 (CH 2 ) 2 505 (100) 489 (100) 449 (100) 433 (100) 503 (100) 4 (CH 2 ) 2 & Na 483 (0.3) 467 (0.4) 427 (0.4) 411 (1) 481 (0.5) 2 COOH 441 (0.5) 441 (0.7) 441 (2.5) 441 (2.5) 441 (2.6) 2 415 (0.1) 415 (0.1) 415 (0.1) 415 (0.2) 415 (0.2) 1 COOH 373 (0.4) 357 (0.5) 317 (0.5) 301 (0.5) 371 (0.4) 1 COOH & HPO 4 (CH 2 ) 2 249 (0.8) NA NA NA NA 1 COOH & HPO 4 (CH 2 ) 2 & Na 227 (0.3) NA NA NA NA 2 585 (0.1) NA 529 (0.1) NA 583 (0.1) 6 H 6 O 3 & HPO 4 (CH 2 ) 2 b NA NA NA NA 377 (1.1) a Relative abundances (%) for each product ion are given in parentheses. The abundance was rounded to the nearest b Fragmentation resulting from proposed intramolecular retroene reaction discussed later in the text.

PAGE 59

59 Figure 2 3. MS3 product ion spectrum of m/z 3)3]+ ion of PAzPC. The structure and proposed fragmentation of the ion at m/z 629 are also shown. Fragmentation pathways leading to the product ions marked with an asterisk are discussed in the text.

PAGE 60

60 Figure 24. MS4 product ion spectrum of m/z HPO4(CH2)2N(CH3)3]+ ion of PAzPC. The proposed structure and fragmentation of the ion at m/z 505 are also shown. Product ions marked with an asterisk are discussed in the text.

PAGE 61

61 Figure 25 MS4 product ion spectrum of m/z 4(CH2)2N(CH3)3]+ ion of PONPC. The proposed structure and fragmentation of the ion at m/z 489 are also shown. Product ions marked with an asterisk were identified as follows. The ion at m/z 433 results from the loss of C3H4O from either the terminal portion of the sn 2 tail or from the glycerol backbone as previously reported.13 The ion at m/z 379 likely results from the loss of C7H16, potentially from the sn 1 fatty acid tail. The product ions resulting from the NL of R2COOH and R1=C=O were also observed at m/z 319 and m/z 251, respectively. Furthermore, the product ions at m/z 177, 165, and 149 are believed to correspond to [R22O]+, [R22O]+, and [R2+, respectively.

PAGE 62

62 Figure 26 MS4 product ion spectrum of m/z 4(CH2)2N(CH3)3]+ ion of PGPC. The proposed structure and fragmentation of the ion at m/z 449 is also shown. Product ions marked with an asterisk were identified as follows. The ion at m/z 393 results from the loss of C3H4O most likely from the glycerol backbone as previously reported.13 The ion at m/z 313 likely results from the combined NL of R2=C=O and Na and the ion at m/z 211 results from the NL of R1=C=O. Lastly, m/ z 137 corresponds to [R22O]+.

PAGE 63

63 Figure 27 MS4 product ion spectrum of m/z 4(CH2)2N(CH3)3]+ ion of POVPC. The proposed structure and fragmentation of the ion at m/z 433 is also shown. Product ions marked with an asterisk were identified as follows. The ion at m/z 377 results from the loss of C3H4O from either the terminal portion of the sn 2 tail or from the glycerol backbone as previously reported.13 The product ions at m/z 319 and 195 result from the NL of R2COOH and the NL of R1=C=O, respectively.

PAGE 64

64 Figure 28 MS4 product ion spectrum of m/z 4(CH2)2N(CH3)3]+ ion of KOdiAPC. The proposed structure and fragmentation of the ion at m/z 503 are also shown. Product ions marked with an asterisk were identified as follows. The NL of 28 u ( m/z 475) likely results from the NL of C2H4; however, the exact site of fragmentation is unknown. Although low in abundance, m/z 447 is observed and is likely produced by the NL of C3H4O from the glycerol backbone as previously reported.13 The ion at m/z 313 is proposed to result from the combined NL of R2=C=O and Na and the ion at m/z 265 from the NL of R1=C=O. Lastly, m/z 191 corresponds to [R22O]+.

PAGE 65

65 Figure 29. Schemes for the proposed intramolecular retroene mechanism.

PAGE 66

66 Figure 210. Identification and imaging of two proposed OxPCs at m/z 716. The MS spe ctrum illustrates the complex mixture of biomolecules observed in tissue. MS2, MS3, and MS4 spectra and images demonstrate the feasibility of identifying and localizing OxPCs in biological tissues using the MALDI MSn methods developed in this work. Based on the product ions observed, two isobaric OxPCs were identified, PC(16:0/11:0 COOH) and PC(18:0/9:0 COOH). The proposed structures are displayed in the orange and purple boxes and the product ions specific to each of these OxPCs are indicated by orange stars and purple diamonds for PC(16:0/11:0 COOH) and PC(18:0/9:0 COOH), respectively. Furthermore, upon comparison of the MSn images of m/z 716 ([PC(16:0/11:0 COOH)+Na]+ and [PC(18:0/9:0 COOH)+Na]+) to the MS image of m /z 756 ( [PC(16:0/16:0) +Na]+, a PL known to localize in the gray matter of rat spinal cord), the ions related to the OxPCs appear mostly in the gray matter (outlined in red).

PAGE 67

67 CHAPTER 3 MALDI MSn CHARACTERIZATION OF PHOS PHATIDYLETHANOLAMINE OXIDATION PRODUCTS Introduction Under oxidative conditions, unsaturated phospholipids undergo structural modifications that result in alterations to the fluidity, permeability, and integrity of cellular membranes.14 In addition to phosphatidylcholines (PCs), phosphatidylethanolamines (PEs) are a major class of phospholipids found within the bilayer of cellular membranes.109 PEs are extremely abundant in the retina, where photoreceptor cell membranes are expected to be highly susceptible to oxidative damage as these membranes contain the most unsaturated fatty acids found in vertebrate tissues ; greater than 50% of the total retinal fatty acids are unsaturated.35 Furthermore, photorecep tor membrane disks are also regularly exposed to light and oxygen making them prone to photogenerated free radical induced oxidative modification.35 Although the high susceptibility of retinal tissue to oxidative damage has led to the implication of phospholipid oxidation in diseases such as agerelated macular degeneration (AMD),5 35 117 there is only limited information regarding the lipid composition of normal and diseased retinal ti ssue.118 However, one specific PE oxidation product, 4hydroxy 7 oxohept 5 enoic acid phosphatidylethanolamine (HOHA PE) is thought to play a central role in promoting the most devastating form of AMD, choroidal neovascularization (CNV) which causes irreversible loss of vision in elderly humans.35 Therefore, there is growing interest in investigating PE oxidation products (OxPEs) and developing methods to identify these products in biological tissues such as the retina.

PAGE 68

68 Recently, there have been several published studies reporting the direct analysis of OxPEs by MS using soft ionization techniques, such as ESI and MALDI.14, 15 In 1998, Khaselev and Murphy reported the use of ESI MS2 for the investigation of plasmenyl OxPEs resulting from in vitro oxidation of a bovine brain PE mixture.119 Gugiu et al. also used ESI MS and MS2 in combination with RP HPLC to investigate short chain OxPEs produced by in vitro oxidation of unsaturated PEs and to identify these OxPEs in retinal extracts.35 Another recent study used ESI MS in combination with thinlayer chromatography (T LC) to monitor autoxidation of various PLs including 1palmitoyl 2 arachidonoyl sn glycero 3 phosphoethanolamine (PAPE) prior to investigating lipopolysaccharide inhibition by these OxPLs.120 Furthermore, oxidative modifications to one specific unsaturated PE, 1palmitoyl2 linoleoyl sn glycero 3 phosphoethanolamine (PLPE), were studied by RP HPLC ESI MS2.103 Although most studies have employed ESI MS as the method of choice for analysis of OxPLs, Stbiger et al. published in 2010 the first study investigating sample preparation techniques for MALDI MS analysis of OxPCs, OxPEs, and OxPSs.45 Additionally, TLC MALDI MS methods for monitoring oxidation products of PC s and PE s have also been recently reported.108 To build upon these works, this study investigates the MALDI MSn (where n = 2, 3, and 4) fragmentation of several short chain OxPEs. Four different unsaturated PE standards were exposed to 48 h of autoxidation in ambient air and light. The observed oxidation products were then identified by MALDI MSn utilizing a linear ion trap (LIT) mass analyzer. Up to four stages of mass analysis (MS4) were used to fully characterize these OxPEs. This work illustrates the feasibility of MSn experiments for obtaining valuable structural information on OxPEs

PAGE 69

69 Experimental Section Chemicals Fisher Scientific ( Fair Lawn, NJ) was the supplier of chloroform (CHCl3), methanol (MeOH), and sodium acetate trihydrate (NaOAc). All lipid standards, which are summarized in Table 31, were purchase d in CHCl3 (10 mg/mL) from Avanti Polar Lipids ( Alabaster, AL). These standards were further diluted in CHCl3 to a final concentration of 100 ppm. MALDI matrix, 2,5 dihydroxybenzoic acid (DHB), was purchased from Acros Organics (Geel, Belgium). A matrix solution of 40 mg/mL was prepared in a solvent system of 50:50 CHCl3:MeOH with a final concentration of 10 mM NaOAc. Oxidation P rocedure The unsaturated PEs listed in Table 31 were autoxidized to produce OxPEs. The reported method for autoxidation of lipids3 was modified by performing the reaction directly on the stainless steel MALDI plate rather than in glass vials, thereby simplifying sample preparation. For the 0 h time point, 1 L aliquots of each of the 100 ppm PE standards were pipetted atop individual sample wells on the MALDI target plate, immediately followed by 1 L of the MALDI matrix solution. For the 48 h autoxidation, 1 L aliquots of each of the100 ppm PE standards were pipetted atop individual sample wells on the MALDI target plate, the solv ent was evaporated under a gentle stream of nitrogen, and the lipid residue was exposed to ambient air and light for 48 h. Immediately following autoxidation, 1 L of the MALDI matrix solution was deposited atop the autoxidized PE standards.

PAGE 70

70 MSn Experiment al P arameters All mass spectra were acquired in positive ion mode using a Thermo Scientific MALDI LTQ XL (San Jose, CA) with an intermediatepressure MALDI source (75 mTorr) and a LIT mass analyzer. The MALDI source was equipped with a nitrogen laser ( = 337 nm) with a 60 Hz repetition rate and 100m laser spot diameter. For these experiments, laser energies between 3 and 10 J per laser shot and 3 laser shots per raster spot were used. Furthermore, all mass spectra were collected in profile mode. MS2 ex periments were performed using a precursor ion isolation width of 1.2 u. For MS3 analysis, the precursor ion isolation width for the first and second isolation events were adjusted to 1.5 and 2.0 u, respectively. For MS4, isolation widths for the first, second, and third isolation events were 1.5, 2.0, and 2.5 u, respectively. Each mass spectrum presented in this work represents an average of 50 analytical scans. Collision induced dissociation (CID) was used to perform MS2, MS3, and MS4 experiments. Norma lized c ollision energies were optimized to achieve maximum product ion intensity and ranged from 29 45% In general, the precursor ion intensity was decreased to ca. 1 10% of the most intense fragment ion. Results and Discussion Autoxidation of POPE, PLPE, PDPE, and SDPE Table 32 summarizes the oxidation products observed following 48 h of autoxidation of the unsaturated PEs listed in Table 31. Due to prolonged storage, SDPE was already oxidized prior to the 48 h autoxidation method. Thus, all other lipid standards were autoxidized and analyzed within one month of purchasing. The most abundant oxidation products observed for each PE resulted from oxidative modification at the first double bond position in the sn 2 substituent as previously reported.35 For

PAGE 71

71 POPE and PLPE, the most abundant OxPEs resulted from cleavage between C9 and C1 0 of the unsaturated sn 2 substituent (PONPE and PAzPE). PDPE and SDPE on the contrary, yielded OxPEs resulting primarily from cleavage between C4 and C5 of the unsaturated sn 2 substituent; PDPE autoxidation produced POBPE and PSuPE and SDPE autoxidation produced SOBPE and SSuPE The mass spectrum obtained before and after 48 h autoxidation of POPE is depicted in Figure 31. Although two ions corresponding to PONPE ( m/z 630 and 652) are observed prior to oxidation (Figure 31A), the relative abundances of ions corresponding to the proposed oxidation products, including PAzPE at m/z 646 and 668 are higher followi ng autoxidation (Figure 31B). Additionally, the relative abundances of the [M+H]+, [M+Na]++ ions of intact POPE, observed at m/z 718, 740 and 762, respectively, decreased following autoxidation. Although low in intensity due to the addition of sodium acetate to the matrix, the [M+H]+ ions of each of the short chain OxPEs were observed; however, MSn of most of these ions did not yield informative fragmentation. Thus, fragmentation of the monosodiated ([M+Na]+) and disodiated +) ions were investigated in this work. MSn of [M+Na]+ I ons of OxPEs MSn fragmentation of OxPEs was dependent on the type of modification to the sn 2 s ubstituent (terminal aldehyde or carboxylic acid) and the length of the oxidatively modified sn 2 substituent. Following MS2 of the [M+Na]+ ion of each of the OxPEs, the most abundant MS2 product ion was isolated and f ragmented by CID to produce MS3 spectr a. MS4 was acquired by selecting the most abundant MS3 product ion for further fragmentation. By performing multiple stages of MS, characteristic fragmentation pathways were determined for each of the OxPEs investigated.

PAGE 72

72 Figure 32 depicts the MSn fragment ation of the [M+Na]+ ion of PAzPE produced from autoxidation of POPE. Following CID, [PAzPE+Na]+ ( m/z 646) fragmented similarly to the [M+Li]+ ions of unoxidized PEs,121 yielding MS2 product ions resulting from NLs of aziridine (C2H5N; 43 u), dehydrated phosphoethanolamine (C2H6NO3P; 123 u) phosphoethanolamine (C2H8NO4P; 141 u) and the sn 1 moiety (C16H32O2; 256 u) at m/z 603, 523, 505, and 390, respectively (Figure 32A). MS3 2H5N]+ ( m/z ct ion ( m/z 505) resulting from a NL of phosphoric acid (H3PO4; 98 u) (Figure 32B). MS4 2H8NO4P]+ ( m/z m/z 279 ([ sn 1+Na]+), 249 (NL of sn 1 ; 256 u ), 227 (NL of sn 1 a nd Na; 278 u), and 211 ([ sn 2+Na]+) (Figure 3 2C). These MS4 product ions allowed for identification of fatty acid substituents, providing confirmation of the oxidative modification to the sn 2 moiety. Likewise, MSn of the [M+Na]+ ions of PSuPE and SSuPE, which are the carboxylic acidcontaining oxidation products of PDPE and SDPE respectively, also yielded informative product ion spectra (Figure 33 and 34). MS2 of [PSuPE+Na]+ at m/z 576 (Figure 33A) and [SSuPE+Na]+ at m/z 604 (Figure 34A) also yielded product ions analogous to those reported above for PAzPE; a predominant NL of 43 u was observed followed by NLs of 141 u and 123 u. As expected, MS3 of 2H5N]+ ( m/z 3B) and [SSuPE+Na]+ ( m/z Figure 34B) resulted in one predominant product ion corresponding to the NL of 98 u; however, an additional product ion corresponding to the NL of the oxidized sn 2 moiety (118 u) was observed in the MS3 spectrum from PSuPE and SSuPE at m/z 415 and 443, respectively Since this frag mentation pathway was not observed for PAzPE, the

PAGE 73

73 NL of the sn 2 substituent in MS3 appears to be dependent on the sn 2 acyl chain length. MS4 2H8NO4P]+ ( m/z 3C) and 2H8NO4P]+ ( m/z 4C) pr oduced numerous structurally informative product ions related to the sn 1 and sn 2 moieties. In contrast to PAzPC, the predominant MS4 product ions from PSuPE and SSuPE resulted from cleavage of the sn 2 substituent rather than the sn 1. These product ions resulted from the NL of 100 u (C4H4O3) and 122 u (C4H3NaO3) from the sn 2 substituents. Also, both of these OxPEs yielded an abundant MS4 product ion at m/z 141 that was identified as the [ sn 2+Na]+ ion. In addition to the carboxylic acidcontaining OxPEs, aldehydecontaining OxPEs were also interrogated by MSn. MS2, MS3, and MS4 spectra of the [PONPE+Na]+, a product of both POPE and PLPE autoxidation, are shown in Figure 3 5A, 3 5B, and 3 5C, respectively. Unexpectedly, MS2 of the precursor ion at m/ z 630 resulted in an abundant NL of 18 u ( m/z 612) corresponding to the loss of water. Product ions resulting from the NLs of 43 ( m/z 587) and 141 u ( m/z 489) were also observed, but were of much lower abundance than m/z 612 (Figure 3 5A). This loss of wat er is proposed to result from the gas phase reaction of the primary amine with the carbonyl carbon forming a macrocycle as shown in Figure 35B. MS3 2H5N]+ ion ( m/z and resulted in a dominant product ion at m/z 514 (NL of phosphoric acid ; 98 u) and lower abundant product ions at m/z 492 (NL of sodium dihydrogen phosphate; 120 u), 415 (NL of sn 2 and C2H5N ; 197 u), 374 (NL of sn 1 ketene; 238 u), and 356 (NL of sn 1 free

PAGE 74

74 fatty acid ; 256 u) as displayed in Figure 3 5B. The observed MS3 product ion at m/z 514 supported the proposed macrocycle formation as one would not expect the loss of phosphoric acid without the remaining portion of the ethanolamine headgroup from a noncyclic structure. Additionally, the product ion at m/z 415 supported the proposed reaction of the primary amine in the headgroup with the carbonyl carbon of the oxidized sn 2 substituent as the NL of 197 u corresponded to the loss of the remaining sn 2 moiety (less water) with C2H5N from the ethanolamine headgroup. For further structural characterization, MS4 was performed on the 2H6NO4P]+ ion ( m/z 5C). The product ions at m/z 260 (NL of sn 1 moiety as a ketene; 254 u) and m/z 220 ([ sn 2O+C2H5N]+) confirmed that the ion at m/z 514 still contained the intact sn 1 moiety, the sn 2 moiety less H2O, and C2H5N. Additionally, the product ion at m/z 220 ([ sn 2O+C2H5N]+) further corroborated the proposed reaction of the primary amine with the sn 2 carbonyl carbon expelling water. Similar MSn fragmentation pathways were observed for the [M+Na]+ ions of POBPE and SOBPE; however, both the NL of aziridine and water were observed as abundant MS2 product ions. Thus, both fragmentation pathways were followed as depicted in Figures 36 and 3 7. MS2 of the [ POBPE+Na]+ ion ( m/z 560; Figure 36A) and [S OBPE+Na]+ ion ( m/z 588; Figure 37A) yielded abundant product ions corresponding to the NL of water ( m/z 542 and 570, respectively), C2H5N ( m/z 517 and 545, respectively), dehydrated phosphoethanolamine ( m/z 437 and 465, respectively), and phosphoethanolamine ( m/z 419 and 447, respectively)

PAGE 75

75 MS3 of [POBPE+Na C2H5N]+ ( m/z ; Figure 36B) and [SOBPE+Na C2H5N]+ ( m/z ; Figure 37B) resulted mainly in a NL of 98 u, as was observed for PONPE (Figure 35B). Additionally, MS3 product ions resulting from a NL of sodium dihydrogen phosphate (NL of 120 u) and a NL of the sn 2 fatty acid substitue nt (NL of 102 u) were observed at m/z 397 and 415, respectively, for POBPE and m/z 425 and 443, respectively, for SOBPE. MS4 of m/z 6C) and m/z 7C) resulted in similar fragmentation pathways and product ions tha t confirmed the identification of these precursor ions. Several NLs related to the sn 1 and sn 2 substituents were observed. The most abundant MS4 product ion in these two spectra resulted from the NL of 84 u from the sn 2 substituent at m/z 335 and 363 for POBPE and SOBPE, respectively. Additionally, the loss of the entire sn 2 moiety was observed as a NL of 100 u at m/z 319 and 347 for POBPE and SOBPE, respectively. Following the fragmentation pathway initiated by the loss of water in MS2, MS3 of 2O]+ ( m/z 2O]+ ( m/z 7D) revealed multiple product ions that resulted from various crossring cleavages of the proposed macrocycle, similar to the proposed fragmentation pathways for P ONPE. The most abundant MS3 product ion, proposed to res ult from concurrent loss of part of the sn 2 moiety and C2H5N ( m/z 433 for POBPE and m/z 461 for SOBPE), was further interrogated by MS4. MS4 of m/z 3 6E) and m/z re 3 7E) yielded various product ions that confirmed the assignment of the ions at m/z 433 and 461, respectively.

PAGE 76

76 MSn of [M H+2Na]+ Ions of OxPEs + ions of OxPEs resulted in numerous structurally informative product ions. As more comprehensive fragmentation coverage was observed for these disodiated ions, only two stages of MS (MS2) were required to identify the fatty acid composition. Ions correspon ding to the NL of each of the fatty acid substituents were observed in addition to product ions corresponding to the disodiated oxidized sn 2 substituent. Figures 38 and 39 illustrate the MS2 spectrum from + ( m/z + ( m/z 652), respectively. As evident in these two figures, MS2 fragmentation pathways were dependent on the type of modification to the sn 2 moiety. CID of the carboxylic acidcontaining OxPE, PAzPE ( m/z 668) yielded dominant product ions resulting from the NL of aziridine (43 u), dehydrated phosphoethanolamine (123 u), and the sn 1 substituent concurrently with C2H5N (299 u). Conversely, CID of the aldehydecontaining OxPE (PONPE; m/z 652) yielded one dominant product ion resulting from the loss of the sn 1 fa tty acid + and + resulted in a NL of dehydrated phosphoethanolamine and Na (C2H6N Na O3P; 145 u ) NL of the sn 2 substituent (NL 188 u for PAzPE and NL of 172 u for PONPE), and a product ion corresponding to the disodiated sn 2 substituent ( m/z 233 and 217 for PAzPE and PONPE, respectively). T he product ions marked with asterisks in Figures 3 8 and 39 indicate the concurrent NL of one of the fatty acid substituent and 43 u from the PE headgroup. Furthermore, product ions at lower relative abundances suggested more than one potential structure for the ion at m/z 652. Again, the presence of a terminal aldehyde led to product ions corresponding to the loss of phosphori c acid (NL of 98 u)

PAGE 77

77 (without the remainder of the PE headgroup) and the loss of the sodiated sn 2 substituent with C2H5N (NL of 237 u). Thus, the PONPE ion at m/z 652 likely has two conformations : one cyclic and one noncyclic. Figures 310 and 311 illust rate the MS2 + ( m/z + ( m/z 582), respectively. These spectra again illustrate the effect of the terminal sn 2 functional group on the fragmentation pathways observed and, upon comparison with Figures 38 and 3 9 indicate that the sn 2 acyl chain length has a direct impact upon fragmentation. As described above for the [M+Na]+ ions, CID of the shorter chain OxPEs (PSuPE and POBPE) resulted in a greater number of cleavages related to the sn 2 substituent rather than the sn 1 substituent as observed for the 9 carboncontaining OxPEs (PAzPE and PONPE). The primary difference in the MS2 fragmentation pathways of carboxylic acidcontaining OxPE (PSuPE) and the aldehydecontaining OxPE (POBPE) was the site of cleavage of the oxidized sn 2 + resulted in a primary product ion at m/z 480, corresponding to the NL of the sn 2 as the free fatty acid. In contrast, CID of [ + resulted in a primary production ion at m/z 498, corresponding to the NL of the sn 2 from the other side of the glycerol oxygen. These same trends were observed for the analogous precursor ions of SSuPE ( m/z 626; Figure 3 12) and SOBPE ( m/z 610; Figure 3 13). Namely, the most abundant product ions resulted from cleavage of the sn 2 moiety rather than sn 1 Furthermore, the aldehydecontaining OxPE (SOBPE) preferentially lost the sn 2 as the free fatty acid whereas the carboxylic acidcontaining OxPE (SOBPE) lost the sn 2 moiety without the glycerol oxygen.

PAGE 78

78 Conclusions MALDI MS of autoxidized PE standards revealed two major oxidation products for each of the unsaturated PEs studied, and MSn fragmentation patterns allowed for structural characterization of thes e oxidation products. For the [M+Na]+ a NL of water in MS2 and NL of the sn 2 moiety concurrently with C2H5N in MS3 were only observed for PE oxidation products containing a terminal aldehyde and not those containing a terminal carboxylic acid. Thus, a nov el fragmentation pathway resulting in the formation of a macrocycle is proposed for oxidized PEs containing a terminal aldehyde on the sn 2 fatty acid substituent. Additionally, the relative abundance of the macrocyclic structure increased as a function of sn 2 chain length, as demonstrated by the MS2 fragmentation of PONPE and POBPE. This trend correlates with previously reported observations for bn peptide ions, in which the relative abundance of the macrocyclic structure increased as a function of fragme nt size.122 Lastly, MS2 of the disodiated oxidized PE ions provided better fragmentation coverage than MS2 of singly sodiated oxidized PE ions. The MALDI MSn technique developed in this work for identifying OxPEs resulting from autoxidation is expected to be directly appli cable to detecting and identifying in situ OxPEs in biological tissues such as the retina. Future studies will aim to identify and quantify OxPEs in retina tissue by MALDI MSn and mass spectrometric imaging. These studies may potentially provide insight into the involvement of lipid oxidation, specifically OxPEs, in aging of the retina and diseases such as AMD.

PAGE 79

79 Table 31. PE standards autoxidized and analyzed.a Chemical Name [Shorthand] Monoisotopic Mass R1 R2 1 palmitoyl 2 oleoyl sn glycero 3 phosphoethanolamine [POPE] 717.53 16:0 18:1 1 palmitoyl 2 linoleoyl sn glycero 3 phosphoethanolamine [PLPE] 715.52 16:0 18:2 1 palmitoyl 2 docosahexaenoyl sn glycero 3 phosphoethanolamine [PDPE] 763.52 16:0 22:6 1 stearoyl 2 docosahexaenoyl sn glycero 3 phosphoethanolamine [SDPE] 791.55 18:0 22:6 a Nomenclature according to the LIPID MAPS systematic naming is used (www.lipidmaps.org). R1 and R2 represent the number of carbons and double bonds (#C:#DB) in the fatty acid substituents esterified to the sn 1 and sn 2 position of the glycerol backbone, respectively.

PAGE 80

80 Table 3 2 N omenclature of the observed PE oxidati on products .a Observed Oxidation Products [Shorthand] Monoisotopic Mass R1 R2 Products of:b 1 palmitoyl 2 (9 oxo nonanoyl) sn glycero 3 phosphoethanolamine [PONPE] 607.38 16:0 9:0 (CHO) POPE, PLPE 1 palmitoyl 2 azelaoyl sn glycero 3 phosphoethanolamine [PAzPE] 623.38 16:0 9:0 (COOH) POPE, PLPE 1 palmitoyl 2 (4 oxo butyroyl) sn glycero 3 phosphoethanolamine [POBPE] 537.31 16:0 4:0 (CHO) PDPE 1 palmitoyl 2 succinoyl sn glycero 3 phosphoethanolamine [PSuPE] 553.30 16:0 4:0 (COOH) PDPE 1 stearoyl 2 (4 oxo butyroyl) sn glycero 3 phosphoethanolamine [SOBPE] 565.34 18:0 4:0 (CHO) SDPE 1 stearoyl 2 succinoyl sn glycero 3 phosphoethanolamine [SSuPE] 581.33 18:0 4:0 (COOH) SDPE a R1 represents the number of carbons and double bonds (#C:#DB) in the fatty acid substituents esterified to the sn 1 position of the glycerol backbone. R2 represents the number of carbons and double bonds in the oxidatively truncated sn 2 fatty acid substitu ent. b The se PE oxidation products are produced from oxidative cleavage along the unsaturated sn 2 fatty acid of POPE, PLPE, PDPE, or SDPE and result in addition of a terminal aldehyde (CHO ) or carboxyl group(COOH) .

PAGE 81

81 Figure 3 1. MS spectrum of POPE A) before and B) after autoxidation. A decrease in the relative abundances of ions corresponding to intact POPE and an increase in the relative abundances of POPE oxidation products were observed. m/z 560 580 600 620 640 660 680 700 720 740 760 780 0 20 40 60 80 100 0 20 40 60 80 100Relative Abundance 762.5 740.5 577.6 697.5 718.4 599.6 652.4 762.5 652.4 740.5 577.7 630.4 668.4 646.4 697.5 612.4 718.4 630.4 [M H+2Na]+[ M+Na ]+[M+H]+ [M+H141]+ ++ [ PONPE+Na ]+ [ PAzPE+Na ]+ 624.4NL: 4.28E5 NL: 2.38E5 A B

PAGE 82

82 Figure 32 Fragmentati on pattern of [ P A z PE+Na]+ from autoxidized POPE. A) MS2 of m/z 646 results in a NL of 43 u ( m/z 603) corresponding to t he loss of aziridine, which is typically the most abundant loss observed from sodiated PEs B) Fragment ions observed in MS3 of m/z 646603, and C) MS4 of m/z 646 603 505 support the identification of m/z 646 as [ P A z PE+Na]+. The proposed structures of the ions undergoing fragmentation and the sites o f fragmentation are also shown. OOOOHPOHOONH2OOHO Na+ NL 43 NL 256 NL 141 NL 123 NL 98 OOOOHPOHOHOOOHO Na+ OOOOOHO Na+ NL 256 NL 278 NL 56 NL 294 NL 226 200 250 300 350 400 450 500 550 600 650 700 0 20 40 60 80 100Relative Abundance 603.3 505.3 523.3 646.3 390.0 200 250 300 350 400 450 500 550 600 650 700 0 20 40 60 80 100Relative Abundance 505.3 559.1 603.3 150 200 250 300 350 400 450 500 550 600 650 700 m/z 0 20 40 60 80 100Relative Abundance 249.1 227.0 449.3 211.1 171.1 279.3 505.3A B C NL 334 NL 44

PAGE 83

83 Figure 33 Fragmentation pattern of [ PSu PE+Na]+ from autoxidized PD PE. A) MS2 of m/z 576 results in a NL of 43 u ( m/z 533) corresponding to t he loss of aziridine, which is typically the most abundant loss observed from sodiated PEs B) Fragment ions observed in MS3 of m/z 576533 and C) MS4 of m/z 576 53 3 435 support the identification of m/z 576 as [ PSu PE+Na]+. The proposed structures of the precursor ions and the sites o f fragmentation are also shown. OOOCH2OHOO Na+ Na+OOOHPOHOHOOOHOO Na+O O O H P O H O O N H2O O H O O NL 43 NL 141 NL 123 NL 98 NL 118 600 650 150 200 250 300 350 400 450 500 550 m/z 0 20 40 60 80 100Relative Abundance141.0 335.3 407.3 313.4 367.4 435.5 391.3 279.2 180.8 NL 294 NL 100 NL 156 NL 254 NL 44 NL 122 NL 68 NL 28A B C 200 250 300 350 400 450 500 550 600 650 0 20 40 60 80 100Relative Abundance 533.3 435.3 453.3 576.1 415.3 200 250 300 350 400 450 500 550 600 650 0 20 40 60 80 100Relative Abundance 435.3 415.3 295.0 533.3 NL 238

PAGE 84

84 Figure 34 Fragmentation pattern of [ SSu PE+Na]+ from autoxidized SD PE. A) MS2 of m/z 604 r esults in a NL of 43 u ( m/z 561) corresponding to t he loss of aziridine, which is typically the most abundant loss observed from sodiated PEs B) Fragment ions observed in MS3 of m/z 604561 and C) MS4 of m/z 604 561 463 support the identification of m/z 604 as [ SSu PE+Na]+. The proposed structures of the precursor ions and the sites o f fragmentation are also shown. 200 250 300 350 400 450 500 550 600 650 0 20 40 60 80 100Relative Abundance 561.3 463.4 586.3 481.4 443.3 533.3 604.4 506.4 200 250 300 350 400 450 500 550 600 650 0 20 40 60 80 100Relative Abundance 463.3 443.3 295.1 150 200 250 300 350 400 450 500 550 600 650 m/z 0 20 40 60 80 100Relative Abundance 341.4 363.2 141.1 435.4 407.3 222.2 395.2 463.3 285.3 307.1 419.4 196.9 156.9 Na+OOOHPOHOONH2OOHOO Na+OOOHPOHOHOOOHOO Na+O O O C H2O H O O NL 43 NL 141 NL 123 NL 98 NL 118 NL 322 NL 100 NL 156 NL 241 NL 44 NL 122 NL 56 NL 28 NL 178 561.2 NL 266A B C

PAGE 85

85 Figure 3 5 Fragmentation pattern of [PON PE+Na]+ from autoxidized POPE. A) MS2 of m/z 630 results in an abundant NL of 18 u ( m/z 612) corresponding to a loss of water, which suggests the reaction of the primary amine with the carbonyl carbon resulting in the formation of a macrocycle. B) Fragment ions observed in MS3 m/z 630612 and C) MS4 m/z 630612 514 support this hypothesis. T he proposed structures of the ions undergoing fragmentation and the sites of fragmentation are also shown. 612.3 586.8 629.8 532.3 463.2 390.9 489.3 OOOOHPOHOONH2OO Na+ NL 18 NL 43 NL 141 514.3 356.1 415.2 612.3 492.3 374.2 514.3 220.1 458.3 202.1 260.1 319.0 OO O P O O O O O H N Na+ Na+OOCH2OOCH2N NL 256 NL 197 NL 98 NL 256 NL 294 NL 56 A B C 200 250 300 350 400 450 500 550 600 650 700 0 20 40 60 80 100Relative Abundance 200 250 300 350 400 450 500 550 600 650 700 0 20 40 60 80 100Relative Abundance 200 250 300 350 400 450 500 550 600 650 700 m/z 0 20 40 60 80 100Relative Abundance NL 120 NL 238 NL 312 NL 195

PAGE 86

86 Figure 36 Fragmentation pattern of [ P OBPE+Na]+ ( m/z 560) from autoxidized P DPE. A) MS2 of m/z 560 yields fr agment ions typically observed for sodiated PEs (NLs of 43, 123, 141, and 256 u); however, the ion resulting from a loss of water (NL of 18 u) is also an abundant fragment suggesting POBPE may form a macrocycle similar to PONPE. B) MS3 of m/z 560 517 results mainly in a NL of 98 u, typical for a sodiated PE. C) MS4 of m/z 560 517 419 results in NLs from the sn 1 and sn 2 tails confirming the identification of this ion. D) MS3 of m/z 560542 yields fragmentation pathways similar to those proposed for PONPE, many of which are crossring cleavages of the macrocycle. E) MS4 m/z 560542433 confirms the assignment of m/z 433; NLs corresponding to the sn 2 tail are not observed. The proposed structures of the precursor ions and the sites of fragmentation are shown. O O O O O C H2 Na+ 0 20 40 60 80 100Relative Abundance 517.3 542.3 419.3 437.3 560.3 Na+O O O O O H P O H O O N H2O NL 43 NL 18 NL 141 NL 123 150 200 250 300 350 400 450 500 550 600 m/z 0 20 40 60 80 100Relative Abundance 335.3 419.3 363.2 319.2 125.0 307.2 401.2 164.7 NL 294 NL 84 NL 100 NL 56 NL 18A CNL 254 150 200 250 300 350 400 450 500 550 600 m/z 0 20 40 60 80 100Relative Abundance 415.3 313.3 353.3 335.3 195.0 176.9 433.2 Na+O O O H P O H O H O O NL 18 NL 98 NL 80 NL 238 NL 256 NL 120 E 150 200 250 300 350 400 450 500 550 600 m/z O O O O O H P O H O H O O Na+ NL 98 NL 102 B 150 200 250 300 350 400 450 500 550 600 m/z 0 20 40 60 80 100Relative Abundance 419.3 397.3 517.3 279.0 261.1 NL 238 NL 254 415.3 Na+OOOOPOHOONO NL 256 NL 127 NL 98 NL 312 NL 18 NL 80 NL 238 0 20 40 60 80 100Relative Abundance433.3 415.3 230.0 444.3 286.1 462.3 304.1 NL 109D 150 200 250 300 350 400 450 500 550 600 m/z 524.3 542.2

PAGE 87

87 Figure 37 Fragmentation pattern of [SOBPE+Na]+ ( m/z 588) from autoxidized SDPE. A) MS2 of m/z 588 yields fr agment ions typically observed for sodiated PEs (NL of 43, 123, 141, 256 u); however, the ion resulting from a loss of water (NL of 18 u) is also an abundant f ragment suggesting SOBPE may form a macrocycle similar to PONPE. B) MS3 of m/z 4 of m/z 588545 447 results in neutral losses from the sn 1 and sn 2 tails confirming the identification of this ion. D) MS3 of m/z crossring cleavages of the macrocycle. E) MS4 m/z 588570461 confirms the assignment of m/z 461; NLs corresponding to the sn 2 tail a re not observed. The proposed structures of the ions undergoing fragmentation and the s ites of fragmentation are shown. 150 200 250 300 350 400 450 500 550 600 650 m/z 0 20 40 60 80 100 Relative Abundance Na+OOOOOHPOHOONH2O 150 200 250 300 350 400 450 500 550 600 650 m/z 0 20 40 60 80 100Relative Abundance 447.3 363.2 391.3 428.9 165.1 305.1 347.2 0 20 40 60 80 100Relative Abundance545.3 570.3 447.3 465.3 341.3 588.3 NL 256 NL 141 NL 123 NL 18 NL 43 NL 98 Na+OOOOOHPOHOHOO 447.3 545.3 425.3 Na+OOOOOCH2 NL 84 NL 56 NL 100 NL 142 NL 18 NL 282A B C443.3 NL 102 0 20 40 60 80 100Relative Abundance461.3 552.3 443.3 230.1 472.4 286.2 490.4 304.2 341.4 570.2 Na+OOOOPOHOONO NL 340 NL 98 NL 127 NL 109 NL 284 NL 80 NL 18 150 200 250 300 350 400 450 500 550 600 650 m/z 0 20 40 60 80 100Relative Abundance 443.3 341.3 195.1 381.4 363.4 461.5 Na+O O O H P O H O H O O NL 120 NL 18 NL 266 NL 98 D ENL 80 200 250 300 350 400 450 500 550 600 650 m/z 200 250 300 350 400 450 500 550 600 650 m/z

PAGE 88

88 Figure 38. MS2 fragmentation of [ PAz + ( m/z 668) from autoxidized POPE. The proposed structure and sit es of fragmentation are also shown. Product ions marked with an asterisk indicate the concurrent NL of one of the fatty acid substituent s and part of the PE headgroup ( C2H5N ; 43 u). Figure 39. MS2 fragmentati on of [PON + ( m/z 652) from autoxidized POPE. The proposed structure and sites o f fragmentation are also shown. Product ions marked with an asterisk indicate the concurrent NL of one of the fatty acid substituent s and part of the PE headgroup ( C2H5N ; 43 u). 200 250 300 350 400 450 500 550 600 650 700 m/z 0 20 40 60 80 100Relative Abundance 369.0 545.3 625.3 412.1 480.2 523.3 668.3 233.0 OOOOHPOOON H2O O H O Na Na+ NL 43 NL 299 NL 123 NL 256 NL 188 NL 435 NL 145 347.1 NL 321 437.2 NL 231* * 200 250 300 350 400 450 500 550 600 650 700 m/z 0 20 40 60 80 100Relative Abundance OOOOHPOOONH2OONa Na+ 396.1 480.2 331.1 609.3 507.3 415.2 353.1 217.1 652.3 NL 43 NL 256 NL 299 NL 321 NL 435 NL 172 NL 237 NL 145 *

PAGE 89

89 Figure 310. MS2 fragmentati on of [PSu + ( m/z 598) from autoxidized PD PE. The proposed structure and sites o f fragmentation are also shown. Product ions marked with an asterisk indicate the concurrent NL of one of the fatty acid substituent s and part of the PE headgroup ( C2H5N ; 43 u). Figure 311. MS2 fragmentati on of [POB + ( m/z 582) from autoxidized PD PE. The proposed structure and sites o f fragmentation are also shown. Product ions marked with an asterisk indicate the concurrent NL of one of the fatty acid substituent s and part of the PE headgroup ( C2H5N ; 43 u). Na+OOOHPOOONH2ONaOOHO 200 250 300 350 400 450 500 550 600 650 m/z 0 20 40 60 80 100Relative Abundance 498.3 475.3 437.3 375.3 555.3 508.3 415.3 598.3 580.3 NL 43 NL 123 NL 100 NL 18 NL 161 NL 183 NL 223 342.1 NL 256 * Na+OOOOOHPOOONH2ONa NL 18 NL 43NL 124 NL 102 NL 256 NL 84 NL 145 NL 167 NL 414 200 250 300 350 400 450 500 550 600 m/z 0 20 40 60 80 100Relative Abundance 480.3 415.3 437.3 539.3 326.1 168.0 498.3 582.3 458.3 428.3 261.0 564.3 NL 321*

PAGE 90

90 Figure 312. MS2 fragmentati on of [SSu + ( m/z 626) from autoxidized SD PE. The proposed structure and sites o f fragmentation are also shown. Product ions marked with an asterisk indicate the concurrent NL of one of the fatty acid substituent s and part of the PE headgroup ( C2H5N ; 43 u). Figure 313. MS2 fragmentati on of [SOB + ( m/z 610) from autoxidized SD PE. The proposed structure and sites o f fragmentation are also shown. Product ions marked with an asterisk indicate the concurrent NL of one of the fatty acid substituent s and part of the PE headgroup ( C2H5N ; 43 u). Na+OOOHPOOONH2ONaOOHO NL 123 NL 100 NL 266 NL 18 NL 43 NL 161 NL 141 NL 223 NL 183 * 200 250 300 350 400 450 500 550 600 650 700 m/z 0 20 40 60 80 100Relative Abundance 526.3 503.3 608.3 583.3 465.3 403.3 485.3 626.4 286.2 443.4 360.2 206.2 Na+OOOOOHPOOONH2ONa 200 250 300 350 400 450 500 550 600 650 m/z 0 20 40 60 80 100Relative Abundance 508.3 443.3 465.3 567.3 326.0 526.3 486.3 261.0 592.3 610.0 NL 102 NL 18 NL 43 NL 124 NL 284 NL 84 NL 145 NL 167 NL 349 *

PAGE 91

91 CHAPTER 4 IDENTIFYING AND IMAG ING PHOSPHOLIPID OXIDATION PRODUCTS BY M ALDI MSn Introduction Oxidative stress has been implicated in the pathogenesis of various human diseases.2 6 In addition to DNA and proteins, unsaturated lipids are known targets of oxidative modification.123 Following oxidative stress, numerous lipid oxidation products, with diverse biological activities, are formed.1 124 These activities (e.g., inflammatory vs anti inflammatory) are dependent on the chemical structure of the oxidized species.88 Furthermore, freeradical production and lipid oxidation have been shown to occur in localized regions of biological tissues such as ischemic rat brain.125 Thus, there is growing interest in developing methods to identify and determine the spatial distributions of individual lipid oxidation products in biological tissues.13, 15 Although techniques for measuring lipid oxidation have been reported,37 many of these techniques (e.g., the thiobarbituric acid assay38) suffer from poor selectivity, and are therefore ill suited for analysis of complex biological samples.39 For greater selectivity, mass spectrometric techniques have become i ncreasingly utilized for the analysis of intact lipid oxidation products in complex sample matrices.14, 37, 101, 126 Of these techniques, reversedphase highperformance liquid chromatography (RP HPLC) coupled to electrospray ionization mass spectrometry (ESI MS) has become the m ost widely utilized, permitting both separation and selective detection of intact lipid oxidation products.15 Despite these advantages, conventional ES I MS requires extraction from tissue, precluding the analysis of these oxidation products in situ. Conversely, this work demonstrates that matrix assisted laser desorption/ionization mass spectrometry

PAGE 92

92 (MALDI MS) offers the ability to perform in situ analys is and imaging of oxidized lipids in biological tissues. Although mass spectrometric imaging (MSI) is a powerful technique for obtaining spatially resolved chemical distributions of biomolecules in tissue, MSI typically generates large datasets that are ti me consuming to process. Therefore, multivariate data analysis techniques such as principal component analysis have been employed to reduce the dimensionality of MSI datasets, allowing for unsupervised data mining.50, 127, 128 In many cases, various regions of tissue can be differentiated using a single principal component, thereby reducing analysis time and allowing for rapid extraction of relevant information.50 In this work, in vitro oxidation was performed on thin tissue sections. Although several methods of in vitro oxidation were tested (e.g., the addition of the phot osensitizer methylene blue), UV light exposure in ambient air was found to produce lipid oxidation without requiring the addition of any chemicals (excluding the MALDI matrix) to the tissue. Following oxidati on, principal component analysis (PCA) in combination with MALDI MSn and mass spectrometric imaging (MSI) was utilized to discover, identify, and localize lipid oxidation products in thin tissue sections following in vitro oxidation. As illustrated in this work, several phospholipid oxidation products are nominally isobaric, some even isomeric, with both oxidized and unoxidized phospholipids; thus, the need for the enhanced selectivity of MSn is demonstrated. Experimental Methods Chemicals The MALDI matrix 2,5 dihydroxybenzoic acid ( DHB) was purchased from Acros Organics (Geel, Belgium). Sodium acetate trihydrate (NaOAc) and HPLC grade

PAGE 93

93 methanol and water were obtained from Fisher Scientific (Fair Lawn, NJ). The MALDI matrix solution was prepared by dissolvi ng DHB in 70:30 methanol:water to a final concentration of 40 mg/mL. To promote sodiation, NaOAc (10 mM final concentration) was added to the matrix solution. Tissue Preparation Spinal cords from adult female SpragueDawley rats (Harlan, San Diego, CA) we re u sed in these experiments. Animal studies were performed in an AAALAC approved vivarium and were approved by the local IACUC at the University of California, San Diego. Following euthanasia, excised tissue was immediately flash frozen in liquid nitrogen Microm HM 505E cryostat (Walldorf, Germany). Cross obtained from the cervical region of the spinal cord and were thaw mounted atop glass microscope slides. To remove excess water from the tissue prior to in vitro oxidation, these slides were dried in a vacuum desiccator for approximately 10 min. Although tissue is typically dried for ca. 30 min to 1 h, these longer drying times were not need ed since the tissues were not coat ed with MALDI matrix directly after vacuum desiccation. In Vitro Oxidation Following vacuum desiccation, the right half of the tissue was covered with an aluminum foil wrapped microscope slide to prevent light exposure and oxidation. The glass microscope slide containing the spinal cord sections was then placed ca. 1 cm = 254 nm) mineralight lamp (model UVG 1; 1 UVP, Inc., San Gabriel, CA) and the left half was UV irradiated in ambient air for 4 h. Immediately following this in vitro ox idation procedure, the aluminum foil was removed from the right

PAGE 94

94 side of the microscope slide and MALDI matrix solution was deposited atop the tissue with a Meinhard Type A3 glass nebulizer (Golden, CO). Mass Spectrometry All MS and MSn spectra were acquir ed using a Thermo Scientific MALDI LTQ XL (San Jose, Ca). This instrument was eq uipped with a 337nm nitrogen laser with a 60Hz repetition rate and a 100energies of 2.8 shots per raster spot were used. Additionally, MS and MSn spectra were collected in profile mode using a raster step size of 100 m. For MSn experiments, which were performed by collision induced dissociation (CID) with helium as the collision gas, instru mental parameters were adjusted as follows: 1.2 u isolation width for MS2, 1.5 and 2.0 u isolation widths for the first and second isolation events of MS3, 1.5, 2.0, and 2.5 u isolation widths for the first, second, and third isolation events of MS4, and normalized collision energies of 29 45%. All mass spectra were processed using Thermo Scientific QualBrowser. Following MS and MSn data collection, images were generated using Thermo Scientific ImageQuest 1.0.1 software. To account for the inherent signal v ariability in MALDI MS analysis of tissue, MS images were normalized to the total ion current (TIC); however, MSn images were not normalized. Statistical Analysis Prior to MSn analysis, PCA was performed to determine m/z values that distinguished the UV ex posed and unexposed sides of the tissue. PCA was conducted on the MALDI MSI datasets using the ChemomeTricks toolbox developed at FOM Institute AMOLF (Amsterdam, Netherlands). MS imaging files collected on the Thermo LTQ XL were first converted to .NetCDF using the Roadmap file converter provided in

PAGE 95

95 Xcalibur. Software developed at FOM Institute AMOLF was then employed to extract the data from the .NetCDF files and create Matlab compatible files (.MAT). To reduce the size of the dataset, the collected profil e data were converted to centroid data using PEAPI software, which also performed peaking picking and spectral alignment.129 Prior to performing PCA, the data were normalized to the TIC, meancentered, and divided by the standard deviation of each respective variables intensity Additionally, m/z values know to be related to DHB were removed from the dataset. Following meance ntering and standardization, each pixel in the MS image was treated as a separate sample, and PCA was performed. Scores images were plotted for the first 20 principal components and 1dimensional loadings plots were generated for principal components that demonstrated discrimination between the oxidized and control regions of the tissue. Results and Discussion MS Analysis of UV exposed and Unexposed Spinal Cord Presented in Figure 41 are representative mass spectra from the left half of the tissue, exposed to UV light for 4 h (top) and from the right half of the tissue, which was not exposed to UV light (bottom). Each spectrum was produced by averaging 20 analytical scans acquired from the respective sides of the tissue. The optical image, photographed aft er matrix deposition but prior to MS analysis, depicts the division in the tissue, as the UV exp osed side of the tissue appears slightly darker (Figure 41). Furthermore, the gray matter (butterfly shape) and white matter (area surrounding the gray matter) of the spinal cord are evident in the optical image. Although these studies were aimed at targeting phospholipid oxidation, one overt difference in the spectra relates to the oxidation of cholesterol. Specifically cholesterol, which is observed as the [M + ion ( m/z 369) as previously reported,130 was lower

PAGE 96

96 in relative abundance on the UV exposed side of the tissue. Moreover, the reported oxidation products of cholesterol, such as the oxysterols that were observed at m/z 367 2O]+) and m/z 385 ([M 2O]+) and 7ketocholesterol that was observed at m/z 401 ([M+H]+),130, 131 were all higher in relative abundance on the UV exposed side of the tissue. This difference in localization is further illustrate d by the MS images in Figure 4 1 ; after extracting the intensity of m/z 369 and m/z 367 and normalizing each image to the TIC, opposite spatial distributions were observed for these two ions (Figure 4 1). Additionally, several ions corresponding to ceramides ( m/z 630 and 648) were lower in intensity on the UV exposed portion of the tissue. As hypothesized, a decrease in the relative abundance for a number of ions in the phospholipid region (ca. m/z 700 900) was observed following in vitro oxidation. Despite t his decrease in signal, the hypothesized oxidation products of these phospholipids (as discussed in Chapters 2 and 3) were not readily apparent above the baseline. This lack of abundant oxidation products is not unexpected, as previous studies have suggest ed that each unsaturated phospholipid precursor may produce at least 50 oxidation products ,6 effec tively diluting the MS signal from each individual oxidation product. Furthermore, the inherent spectral complexity associated with tissue analysis makes identifying targets for MSn analysis challenging. Thus, PCA was employed to determine potential phosph olipid oxidation products PCA of UV exposed and Unexposed Spinal Cord PCA was performed on the MSI dataset using the ChemomeTricks toolbox. As illustrated in Figure 42, scores images from principal component 1, accounting for ca. 33% of the variance, distinguished the UV exposed (oxidized) and unexposed (control) halves of the tissue. Accordingly, a 1dimensional loadings plot for this principal

PAGE 97

97 component was generated ( Figure 42 ) to determine which ions exhibited appreciable loadings coefficients on principal component 1the loadings coefficients provided information concerning the relative contribution of each m/z value to the separation along a principal component. Thus, ions loading positively on principal component 1 demonstrated a positive correl ation with this principal component as displayed in the scores images (Figure 42), and are therefore expected to be higher in intensity on the UV exposed half of the tissue. Conversely, ions loading negatively on principal component 1 demonstrate a negati ve correlation with principal component 1 and are therefore expected to be higher in intensity on the unexposed side of the tissue. Several ions dictating separation between the UV exposed and unexposed portions of the tissue were also ions that were ident ified as potential oxidation products based on the extracted MS spectra and MS images detailed in Figure 41 For example the proposed oxidation products of cholesterol at m/z 367, 385, and 401 were positively loaded on principal component 1 and demonstrated elevated intensity on the left side of the scores image ( Figures 42 2O]+ ion of cholesterol at m/z 369 loaded negatively on principal component 1 and appeared on the right side of the scores image. Furthermore, unsaturated phos pholipids, including [PC(16:0/18:1)+Na]+ ( m/z 782) and [PC(18:0/18:1)+Na]+ ( m/z 810), were negatively loaded on principal component 1 and ions 32 u higher at m/z 814 and 842, respectively, were positively loaded on principal component 1. The latter two ion s, proposed to be the respective peroxidation products of PC (16:0/18:1) and PC (18:0/18:1), were targeted for MSn analysis. In addition to these long chain products, ions corresponding to known short chain oxidation products of phosphatidylcholines (e.g., m/z 688 and 716) also

PAGE 98

98 loaded positively on principal component 1 and were therefore targeted for i nterrogation by MSn as well MSn Analysis of Oxidized Phospholipids in the Spinal Cord Based on princi pal component 1 loadings values, ions were selected for MSn analysis. M ultiple stages of mass analysis permitted identification and localization of many of these ions within the spinal cord. Furthermore, several of these ions were detected at the same nomi nal mass as other endogenous species in the tissue; in these instances MSn provided a selective method for differentiating isobaric and even isomeric species. In the figures detailed below, the fragmentation of each ion is indicated by the colored boxes on the spectra and color matched arrows on the structures. The direction of the arrow indicates the portion of the species that retains the charge after dissociation. In some cases, two or more product ions a re outlined in the same colored box; typically, t hese product ions differ by 22 u, and therefore represent a NL with (less a proton) or without sodium. Also, for spectra containing product ions characteristic of more than one phospholipid, stars and diamonds were used to denote the respective product ions for each precursor ion. In these spectra, product ions boxed, but not marked with a star or diamond, were common to both precursor ions. Unsaturated phosphatidylcholine s One of the most abundant phospholipids observed by MALDI MS of spinal cord is PC(16:0/18:1) detected as the [M+Na]+ ion at m/z 782. As m/z 782 exhibited an appreciable negative loadings coefficient on principal component 1, MSn was performed to co nfirm the identity of this ion ( Figure 43 ) MS2 of m/z 782 produced NLs of 59 u ( m/z 7 23) and 183 u ( m/z 599), which are characteristic NLs corresponding to alkali

PAGE 99

99 metal adducts of PC s. As evident in the MS2 image mapping the product ion at m/z 723, this ion is localized on the control side of the tissue. To confirm the fatty acid assignments, MS3 of m/z Although the predominant product ions corresponded to NLs of cyclophosphane ( m/z 599) and sodiated cyclophosphane ( m/z 577) from the remaining portion of the PC headgroup, a NL of the sn 1 substituent ( m/z 467) was also observed. In MS4, product ions related to both fatty acid substituents are observed; the product ions at m/z 305 and 279 corresponded to the [ sn 2+Na]+ and [ sn 1+Na]+ ions, respectively. Although the MS image quality significantly declines between MS3 and MS4 due to the loss in ion signal, the characteristic ion at m/z 305 does appear to be higher in intensity on the control side of the sample. The MS4 product ions at m/z 543 (NL of 56 u) and 568 (NL of 18 u) most likely correspond to the NLs of C3H4O and water, respectively. These fragment ions are commonly observed in the MS4 analysis of both unmodified phosphatidylcholines. Furthermore, the fragmentation pathways yielding these ions were generally uninformative regarding fatty acid and headgroup assignment; thus, the fragmentation pathways leading to these product ions were not indicated in Figure 43. In addition to m/z 782, the ion at m/z 810, proposed to be PC(18:0/18:1), was also interrogated by MSn. MS2 and MS3 spectra not only exhibited product ions corresponding to the 18:0 and 18:1 fatty acid substituents, but also product ions suggesting the presence of 16:0, 20:1, and 20:4 substituents. The product ions related to each of these PC species were mapped in tissue and were similarly loc alized on the control side of the tissue.

PAGE 100

100 Longchain oxidation products One of the ions exhibiting a positive loadings coefficient on principal component 1 was m/z 814, which is 32 u higher than m/z 782; thus this ion was targeted for MSn analysis. Based o n the product ions observed and their spatial distributions, two potential phospholipids were proposed (Figure 44). The product ion at m/z 755, which resulted from the NL of 59 u, indicated the presence of PC. The localization of this ion on the UV expose d side of the sample suggested that the PC containing species was potentially an oxidation product. In addition to the product ion at m/z 755, an ion resulting from the loss of phosphocholine at m/z 631 (NL of 183 u) was also localized on the UV exposed side of the tissue. The product ion boxed in red ( m/z 690) may result cleavage relative to the hydroperoxide in the sn 2 substituent, although this claim is not yet confirmed. Also, the unmarked product ion at m/z 796 resulted from the NL of water, most likely from the hydroperoxyl group, and was localized on the UV exposed portion of the tissue. In contrast to the aforementioned ions, product ions related to the isobaric PE were localized in the gray matter of the control side of the tissue; m/z 771 (NL of 43 u) and 673 (NL of 141 u) corresponded to NLs of the PE headgroup. A lower intensity fragment ion at m/z 530 (NL of 284) was hypothesized to arise from the NL of the sn 1 substituent (steari c acid; 18:0). This product ion also demonstrated colocalization with the more intense PE headgroup fragments in the gray matter of the unoxidized half of the spinal cord, thereby confirming this identification as the [M+Na]+ ion of PE(18:0/22:6). To confi rm the identification and spatial localization of the PC species at m/z 814, the most abundant MS2 product ion related to the PC, m/z 755, was targeted for MS3

PAGE 101

101 analysis (Figure 45). Upon CID, the most abundant MS3 product ion ( m/ z 631) resulted from the N L of the remaining portion of the PC headgroup (NL of 124 u; cyclophosphane). In addition to this dominant product ion, two other characteristic product ions are observed at much lower relative abundances: the first at m/z 441, corresponding to the NL of t he oxidatively modified sn 2 substituent, and the second at m/z 499, corresponding to the NL of the sn 1 substituent. The MS3 images demonstrating the localization of these three product ions on the UV exposed side of the tissue are also displayed in Figur e 4 5. Although not yet identified, the product ion at m/z 291 appears to be characteristic of the 18:1 moiety containing a hydroperoxide. To obtain further structural information, MS4 was performed on the most abundant MS3 product ion, m/z 631 (Figure 46 ). The resulting MS4 spectrum of m/z from the unmodified PC described above ( m/z Figure 4 cleavage relative to the hydroperoxide group with concomitant loss of water (Hock fragmentation132). Based on the observed product ions, this fragmentation pathway was responsible for the ions observed at m/z 475, 489, 503, and 517, which corresponded to the C8, C9, C10, and C1 1 isomers, respectively, of the hydroperoxidecontaining PC. In addition to these ions, product ions corresponding to the NL of the sn 1 moiety ( m/z 377) and the [M+Na]+ ions of the sn 2 free acid ( m/z 337) and sn 2 ketene ( m/z 319) confirmed the assignmen t of the PC fatty acid substituents as 16:0 and18:1 hydroperoxide. Other product ions in the MS4 spectrum, which will be the subject of future studies, may result from gas phase rearrangement or from other isomeric ions at m/z 631.

PAGE 102

102 Another potential long chain oxidation product observed at m/z 842 was discovered by PCA and targeted for MSn identification and imaging (Figure 47). Similar to m/z 814, MS2 product ions were indicative of at least two isobaric phospholipids that had opposite spatial distributions, one related to a PC (product ions at m/z 783 and 659, corresponding to the NL of trimethylamine and phosphocholine, respectively) and the other to PE (product ions m/z 799 and 560, corresponding to the NL of aziridine and an 18:1 fatty acid moiety). However, the exact structure of the PE species could not be determined from the MS3 spectrum of m/z PE are outlined in black without corresponding fragmentation arrows. In addition to the isobaric PE, the MS2 spect rum of m/z 842 also suggests the presence of a phosphatidylserine ( m/z 755, corresponding to a NL of 87 u from the PS headgroup). To determine the fatty acid composition of the proposed PC oxidation product, MS3 was performed on the most abundant MS2 product ion related to the PC ( m/z 783) as displayed in Figure 48. Structurally informative product ions corresponding to the NLs of 124 u (cyclophospane), 284 u (the sn 1 moiety), and 314 u (the oxidized sn 2 moiety) were observed. Moreover, the MS3 ima ges of these product ions confirm the localization of this long chain PC oxidation product on the UV exposed side of the tissue. For added confirmation, MS4 was conducted on the MS3 product ion at m/z 659. The fragmentation of this ion, as depicted in Figure 4 9, was analogous to that reported above for MS4 of m/z cleavage relative to the hydroperoxides were detected at m/z 503, 517, 531, and 545 and are proposed to result from Hock fragmentation. Additionally, the product ion resulting from the NL of the sn 1 moiety ( m/z 375) and ions corresponding to the [M+Na]+ ions of the

PAGE 103

103 oxidized sn 2 free fatty acid ( m/z 337) and ketene ( m/z 319) were observed. As was mentioned above for the MS4 spectrum from m/z product ions were observed in the MS4 spectrum of m/z these ions will be the subject of future experiments. Short chain oxidation products In addition to the long chain oxidation products formed by the addition of oxygen(s) to the unsaturated phospholipids, several potential short chain oxidation products (e.g., m/z 688 and 716) were discovered by PCA (Figure 4 2) These ions, which loaded positively on principal component 1 and were observed at lower m/z values than typical phospholipids were subjected to MSn for identification. MS2 of m/z 688, proposed to be the [M+Na]+ ion of the PC oxidation product 1 palmitoyl 2 azelaoyl sn glycero 3 phosphocholine (PAZPC), resulted in product ions indicative of an ion containing a PC headgroup (Figure 410); NLs of 59 u ( m/z 629) and 183 u ( m/z 505) were observed. As expected, the MS2 image of m/z localization of this ion on the UV exposed port ion of the tissue. Furthermore, the MS3 spectrum of m/z 11) contained product ions at m/z 505 (NL of cyclophosphane; 124 u) and 249 (NL of the sn 1 moiety and cyclophosphane; 380 u). To obtain structural information related to the sn 2 m oiety, MS4 was performed on m/z 12. The product ion at m/z 211 corresponded to the [M+Na]+ of the oxidized sn 2 substituent, thus confirming that m/z 688 contained a 9:0 COOH moiety in addition to the 16:0 sn 1 fatty ac id, which was also confirmed by the MS4 product ions at m/z 249 (NL of the sn 1 fatty acid; 256 u) and 227 (NL of the sn 1 fatty acid and sodium; 278 u). Figure 412 also illustrates the

PAGE 104

104 potential of MS3 imaging for enhanced selectivity, which is often req uired for identification of individual molecular species in complex samples such as tissue. Another potential short chain oxidation product that was discovered using PCA was m/z 716, an ion expected to correspond to two isobaric OxPCs based on the work pre sented in Chapter 2. The MS2 spectrum of m/z 716, along with the proposed structures and fragmentation of these ions, are shown in Figure 413. As expected, the primary MS2 product ions resulted from the NLs of trimethylamine (NL of 59 u) and phosphocholine (NL of 183 u). As these two species both contained a PC headgroup, these oxidation products were not differentiated by MS2; however, by performing multiple stages of mass analysis, product ions specific to each of these isomeric ions were observed. Figures 414 and 415 demonstrated the MS3 and MS4 characterization of these isomers. Product ions specific to PC(16:0/11:0 COOH) and PC(18:0/9:0 COOH) are denoted by stars and diamonds, respectively; product ions common to both of these oxidized PCs are unmarked. In addition to product ions corresponding to the NL of cyclophosphane, MS3 of m/z m/z 373 and 469 resulting from the NLs of the sn 1 and sn 2 substituents of PC(18:0/9:0 OOH) and product ions at m/z 401 and 441 cor responding to the analogous losses from PC(16:0/11:0 COOH). Also observed in the MS3 spectrum were two product ions at m/z 249 and 277, which corresponded to the concomitant loss of the sn 1 moiety and cyclophosphane from PC(18:0/9:0 COOH) and PC(16:0/11:0 COOH), respectively. Thus, MS3 provides significant evidence that m/z 716 corresponds to both of these oxidation products. Although potentially unnecessary for the identification of these ions

PAGE 105

105 at m/z 716, as fatty acid composition was determined in MS3, M S4 was performed on m/z The MS4 product ions support the identification of these two isobaric oxidized PCs at m/z 716, as product ions related to both of the fatty acid substituents for both of these products ( m/z 171, 193, 19 9, 211, 221, 227, 239, 249, 255, and 277) were assigned ( Figure 415) Conclusions This work demonstrates, for the first time, that PCA is a powerful tool for rapidly discovering potential biomarkers of oxi dative stress in MALDI MSI data sets and that MSn p rovides a selective method for identifying and localizing these oxidation products in situ. The MSn fragmentation pathways for several representative oxidation products, including long chain and short chain products of PCs were presented. Thus, this work p rovides the foundation for future studies that will investigate these and other oxidation products as potential biomarkers of disease or oxidative stress. Furthermore, the utility of MSn imaging for differentiating isobaric and even isomeric molecular spec ies was illustrated. Through this work the potential of MALDI MSn for identifying individual molecular species, such as oxidized phospholipids, and localizing these individual species in situ, was realized.

PAGE 106

106 Figure 41. MALDI MS spectrum from the left half of the tissue, which was exposed to UV for 4 hours (top) and from the right half of the tissue, which was not exposed to UV (bottom). Each spectrum was produced by averaging 20 analytical scans. On the right, an optical image of tissue and MS images of an oxidation product of cholesterol ([ oxysterol 2O]+ at m/z 2O]+ at m/z 369 normalized to the TIC; bottom) are displayed. 40 40 80 60 100 20 20 60 80 100 200 400 600 800 1000 0 0 NL: 4.50E4 NL: 5.01E4 273.3 385.6 401.5 177.2 782.7 848.8 367.6 518.5 630.8 782.8 810.8 760.8 369.6 958.5 648.8 850.8 Left Half of Tissue Exposed to 4 h UV Right Half of Tissue Not Exposed to UV177.2m/z MS Image of m/z 367/TIC MS Image of m/z 369/TIC 0 0 No UV 4 h UVOptical ImageRelative Abundance 1 mm

PAGE 107

107 Figure 4 2. Principal component 1 scaled loadings plot (left) and scores images (right) f rom the MSI data set in Figure 4 1 Principal component 1 dictates separation of the oxidized (positive) and unoxidized (negative) sides of the spinal cord. Various m/z valu es with positive loadings on principal component 1, such as m/z 814 and 716, were targeted for MSn studies. The loadings were scaled by the standard deviation for each variable. 30 25 20 15 10 5 0 5 10 15 Left Half of Tissue Exposed to 4 h UV Right Half of Tissue Not Exposed to UV385 367 401 184 264 369 630 842 814 648 782 810 850 688 716 880 644 518 768 866Principal Component 1 Scaled Loadings Principal Component 1 (Positive) Principal Component 1 (Negative)

PAGE 108

108 Figure 43. MSn of PC(16:0/18:1) an abundant unsaturated phospholipid in spinal cord tissue. A) MS2 of m/z 782 ( [PC(16:0/18:1)+Na]+) and MS image of the product ion at m/z 723. B) MS3 of m/z 723 ( [PC(16:0/18:1)+Na N(CH3)3]+) and MS image of the product ion at m/z 599. C) MS4 of m/z 599, ( [PC(16:0/18:1)+Na HPO4(CH2)2N(CH3)3]+) and MS image of the product ion at m/z 305. OOOOHPOOO-ON+ Na+ 200 300 400 500 600 700 800 m/z 0 20 40 60 80 100 Relative Abundance 200 300 400 500 600 700 800 m/z 0 20 40 60 80 100 Relative Abundance 599.6 723.6 577.7 467.3 599.6 568.8 543.6 305.3 O O O O Na+ OOOOHPOOOO Na+ 279.3 300 400 500 600 700 800 m/z 0 20 40 60 80 100 Relative Abundance 599.6 782.7 723.6MS2m/z 782 MS3m/z MS4m/z 5.8E5 0 1.3E6 0 1.6E3 0 A B C

PAGE 109

109 Figure 44. MS2 product ion spectrum of m/z 814, a proposed long chain oxidation product of phosphatidylcholine. Based on the observed fragmentation, two dif ferent phospholipids with opposite localization were present at m/z 814: one containing a PC headgroup and one containing a PE headgroup. The proposed structure s and fragmentation pathways of these two phospholipids are shown. Additionally, MS2 images of t he characteristic product ions of the PC at m/z 755 (NL of 59 u) and the PE at m/z 771 (NL of 43 u) are depicted. 250 300 350 400 450 500 550 600 650 700 750 800 850 0 10 20 30 40 50 60 70 80 90 100 Na+O O O O H P O O O-O N+O O H [PC(16:0/18:1)+2O+Na]+ 755.5 814.8 771.5 796.5 690.6 673.6 631.6 530.3 0 0 3.26E4 2.66E4m/zMS2m/z 814Relative Abundance Na+OOOOHPOOOHONH2 [ ]+

PAGE 110

110 Figure 45. MS3 product ion spectrum of m/z ions corresponding to the NLs of cyclophosphane ( 124 u), sn 1 moiety ( 256 u) and sn 2 moiety ( 314 u). The MS3 images illustrate the localization of the proposed PC oxidation product on the UV exposed side of the tissue. 250 300 350 400 450 500 550 600 650 700 750 800 0 10 20 30 40 50 60 70 80 90 100 x20 Na+O O O O H P O O O O O O H 631.6 755.6 441.3 291.3 0 8.77E4 0 1.40E3 0 9.18E2m/zRelative AbundanceMS3m/z 499.3

PAGE 111

111 Figure 46. MS4 product ion spectrum of m/z sn 1 and oxidized sn 2 fatty acid substituents confirm the addition of two oxygens to the sn 2 moiety Although the exact location of oxidative modification has not been confirmed, product ions indicated in red are proposed to correspond to cleavages relative to the hydroperoxide; thus, several isomers are probable. 200 250 300 350 400 450 500 550 600 650 0 10 20 30 40 50 60 70 80 90 100 Relative Abundance Na+O O O O O O H 631.6 613.7 575.5 531.5 517.5 503.5 489.5 475.4 377.3 337.3 319.3 307.3 291.3 587.6 H2O C9 C10 C8 C11m/zMS4m/z

PAGE 112

112 Figure 47. MS2 product ion spectrum of m/z 8 4 2 a proposed long chain oxidation product of phosphatidylcholine. Based on the observed fragmentation and MS2 images two different phospholipids with opposite localization were present at m/z 842, one containing a PC headgr oup and the other containing a PE headgroup. The proposed structure and fragmentat ion pathways of the long chain PC oxidation product are shown. 250 300 350 400 450 500 550 600 650 700 750 800 850 900 0 10 20 30 40 50 60 70 80 90 100 PE Na+O O O O H P O O O-O N+O O H [ PC(18:0/18:1)+2O+Na]+ 783.6 842.6 799.5 718.6 560.3 755.5 659.6 0 0 2.50E 4 5.50 E3824.5 0 2.87 E3 0 2.92E3 MS2m/z 842m/zRelative Abundance

PAGE 113

113 Figure 48. MS3 product ion spectrum of m/z MS i mages were generated by mapping the intensity of the product ions corresponding to the NLs of cyclophosphane ( 124 u), sn 1 moiety ( 284 u) and sn 2 moiety ( 314 u ). The MS3 images illustrate the localization of the proposed PC oxidation product on the UV exposed side of the tissue 250 300 350 400 450 500 550 600 650 700 750 800 850 900 / 0 10 20 30 40 50 60 70 80 90 100 x20 Na+OOOOHPOOOOOOH 659.6 783.6 469.4 499.4 291.3 389.4 319.3 0 4.02 E4 0 1.10E3 0 7.78E2m/zRelative AbundanceMS3m/z

PAGE 114

114 Figure 49. MS4 product ion spectrum of m/z Product ions corresponding to the NLs of the sn 1 and oxidized sn 2 fatty acid substituents confirm the addition of two oxygens to the sn 2 moiety. Although the exact location of oxidative modification has not been confirmed, product ions indicated in red are proposed to correspond to cleavages relative to the hydroperoxide; thus, several isomers are probable. 200 250 300 350 400 450 500 550 600 650 700 0 10 20 30 40 50 60 70 80 90 100 517.5 531.6 503.5 641.6 659.6 319.3 613.6 337.3 347.3 545.6 403.3 365.3 615.6 603.5 307.4 559.5 405.3 627.5 375.3 489.6 599.6 435.5 475.5 291.3 Na+O O O O O O H H2O C9 C10 C8 C11MS4m/z m/zRelative Abundance

PAGE 115

115 F igure 410. MS2 product ion spectrum of m/z 688, a proposed short chain oxidation product of phosphatidylcholine. The MS2 images demonstrate localization on the UV exposed side of the tissue, and the obser ved product ions confirm the presence of a phosphocholine headgroup. However, product ions related to the fatty acid substituents are not observed; therefore, further stages of mass analysis were performed. 200 250 300 350 400 450 500 550 600 650 700 750m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 629.4 688.7 505.5 670.4 564.5 644.4 1.47E5 0 Na+O O O P O-O O N+O OHO O H 1.45E5 0MS2m/z 688

PAGE 116

116 Figure 411. MS3 product ion spectrum of m/z Based on the observed product ions, identi fied as indicated by the color matched boxes and arrow s, this ion contained a 16:0 fatty acid substituent and a 9:0 COOH fatty acid substituent, confirming the identity of the ion at m/z 688. Furthermore, t he MS3 image s illustrate the localization of this oxidation product on the UV exposed side of the tissue. 200 250 300 350 400 450 500 550 600 650 700m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 505.5 461.4 629.5 249.2 373.3MS3m/z 1.46E5 0 Na+OOOPOOOOOHOOH 441.3 1.77E3 0

PAGE 117

117 Figure 412. MS4 product ion spectrum of m/z The observed product ions further substantiate the identification of this PC oxidation product, as multiple ions related to each of the fatty acid substituents were observed as indicated by the color matched boxes and arrows 150 200 250 300 350 400 450 500 550m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 249.2 227.2 505.5 449.4 293.3 211.2 461.4 487.4 193.1 335.3 377.3 171.2 279.3 Na+O O O O O O H MS4m/z

PAGE 118

118 Figure 413. MS2 product ion spectrum of m/z 716. Two possible isomeric short chain oxidation product s of phosphatidylcholine occur at m/z 716. The MS2 images demonstrate localization on the UV exposed side of the tissue, and the obser ved product ions confirm the presence of a phosphocholine headgroup. However, product ions related to the fatty acid substituents are not observed; therefore, further stages of mass analysis were performed. 200 250 300 350 400 450 500 550 600 650 700 750 0 10 20 30 40 50 60 70 80 90 100 716.5 657.3 698.3 672.4 592.4 533.3 Na+OOOPO-OON+OOOOHH 1.99E4 0PC(16:0/11:0 COOH) m/zRelative AbundanceMS2m/z 716 Na+O O O P O-O O N+O OHO O H PC(18:0/9:0 COOH) 4.89E 3 0

PAGE 119

119 Figure 414. MS3 product ion spectrum of m/z P roduct ions related to PC(16:0/11:0 COOH) and PC(18:0/9:0 COOH) are marked with stars and diamonds, respectively. Based on these observed product ions, the fatty acid composition of the two isomeric PCs was confirmed; the product ions at m/z 401 and 441 corresponded to the NL of 16:0 and 11:0 COOH, respectively, whereas the product ions at m/z 373 and 469 corresponded to the NL of 18:0 and 9:0 COOH respectively Furthermore, t he MS3 image s illustrate the localization of both of these oxidation products on the UV exposed side of the tissue. 200 250 300 350 400 450 500 550 600 650 700 0 10 20 30 40 50 60 70 80 90 100 x20 Na+OOOPOOOOOOOHH 657.3 533.4 511.3 469.3 373.2 277.2 249.2 4.17E4 0 401.2 441.3 m/zRelative AbundanceMS3m/z Na+O O O P O O O O O H O O H 6.60E2 0 1 .27E3 0

PAGE 120

120 Figure 415. MS4 product ion spectrum of m/z he presence of two isomeric oxidized PCs, PC(16:0/11:0 COOH) and PC(18:0/9:0 COOH), was confirmed. Analogous product ions related to PC(16:0/11:0 COOH) and PC(18:0/9:0 COOH) are indicated by the color matched boxes and arrows; the product ions of PC(16:0/11:0 COOH) and PC(18:0/9:0 COOH) are marked with stars and diamonds, respectively. 150 200 250 300 350 400 450 500 550 0 10 20 30 40 50 60 70 80 90 100 533.5 477.3 277.2 321.3 255.2 249.1 227.1 515.3 489.3 OOOOOOH Na+ m/zRelative Abundance171.0 193.0 199.1 211.1 221.1 239.1 MS4m/z Na+O O O O O O H

PAGE 121

121 CHAPTER 5 CONCLUSIONS AND FUTURE WORK Conclusions To gain a comprehensive understanding of the role of oxidized phospholipids (OxPLs) in disease, methods for detecting individual OxPLs in biological and clinical samples must first be developed. In this work, matrix assisted laser desorption/ionization (MALDI) tandem mass spectrometry (MSn) techniques for characterizing, identifying and imaging phospholipid oxidation pro ducts are described. The MSn capabilities of the linear ion trap (LIT) were exploited for selective detection of individual phosphatidylcholine oxidation products (OxPCs) and phosphatidylethanolamine oxidation products (OxPEs). Up to four stages of mass analysis were used to obtain enhanced structural information and, hence, improve confidence in ion identifica tion First, five different short chain OxPC standards were characterized by MALDI MSn. Ions of both [M+H]+ and [M+Na]+ for each of the OxPC s were interrogated. Although several product ions corresponding to the fatty acid substituent were observed following CID of the [M+H]+ ions, the [M+Na]+ ion s yielded more structurally informative product ions that were targeted for further stages of mass analysis MS3 of the 3)3]+ ions yielded fragmentation indicative of the OxPC modification; specifically, a product ion corresponding to the neutral loss of CO2 (NL of 44) was observed for OxPCs containing a terminal carboxylic acid rather than an aldehyde. Furthermore, MS4 HPO4(CH2)2N(CH3)3]+ ions for each OxPCs resulted in fragmentation pathways dependent on the sn 2 fatty acid chain length and type of functional group(s). In MS4, OxPCs with a terminal aldehyde yielded

PAGE 122

122 fragmentation of the sn 1 fatty acid distinct from terminal carboxylic acidcontaining OxPCs. Specifically, aldehydecontaining OxPCs with palmitic acid esterified to the sn 1 position of the glycerol backbone yielded a NL of 254 u 2 u less than the nominal mass of palmitic a cid, whereas the analogous terminal carboxylic acidcontaining OxPCs demonstrated a NL of 256 u ketone relative to the terminal carboxyl group resulted in C C bond cleavages along the sn 2 substituent, providing diagnostic MS4 product ions for ketocontaining OxPCs; two intramolecular retroene reactions were proposed to explain these observed product ions. The applicability of this developed MALDI MSn method for i dentifying short chain OxPCs in biological tissue samples was de monstrated using control rat spinal cord tissue. The utility of MSn for distinguishing individual OxPLs in complex biological samples was also illustrated as two isomeric OxPCs were discerned based on the observed MS3 and MS4 product ions. Furthermore, MSI performed on these tissues illustrated localization of two isomeric OxPCs in the gray matter of spinal cord tissue from the cervical region. The above MALDI MSn techniques were also employed to identify and characterize [M+Na]+ + ions of shor t chain PE oxidation products formed by autoxidation of various unsaturated PE standards. Six different OxPEs were investigated and the MSn (where n= 2, 3, or 4) fragmentation was shown to depend on the nature of the oxidative modification, as was the case for OxPCs. Specifically, OxPEs containing a terminal aldehyde at the sn 2 position yielded an abundant MS2 product ion corresponding to the NL of water. This fragmentation pathway was proposed to result from a reaction between the primary amine of the eth anolamine head group and the

PAGE 123

123 terminal carbonyl carbon of the sn 2 moiety, which resulted in the NL of water and the formation of a macrocyclic structure. Additionally, the relative abundance of the ion corresponding to the loss of water increased as a func tion of chain length of the sn 2 moiety containing the terminal aldehyde. Thus, the macrocyclic structure was preferentially formed during CID of PE(16:0/ 9:0 CHO), but was not as favorable for PE(16:0/4:0 COOH) or PE(18:0/4:0 CHO), the shorter chain aldehy de deriv atives. These results correlate well with previous studies reporting the formation of macrocyclic structures in the gas phase following MS2 of bn peptide ions ; moreover, an analogous increase in the relative abundance of the macrocyclic structure as a function of fragment size was also reported.122 Following characterization of OxPL standards, in vitro oxidation was performed on spinal cord tissue to identify and image these oxidation products in situ. Select areas of t hin tissue sections from the cervical region of contr ol rat spinal cord were exposed to UV light and ambient air for 4 h. MSI was then performed on t he tissue and principal component analysis (PCA) was used for data mining to determine m/z values to be targeted for MSn analysis. Several long chain and short chain OxPCs were discovered by PCA and identified based on characteristic MSn fragmentation. Furthermore, the MSn images illustrated localization of these OxPCs on the UV exposed side of the tissue as expected. These studies demonstrate the potential for using PCA in combination with MALDI MSn imaging to identify and localize oxidized lipids in thin tissue sections. Thus, this work provides the foundation for future studies that will utilize MALDI MSn and MSI to gain a better understanding to OxPLs and their role in human disease.

PAGE 124

124 Future Work Future studies will investigate the effects of in vitro oxidation on other lipid classes. Phosphatidylserine and cardiolipin have been shown to undergo extensive oxidation, which could be involved in the pathology of traumatic brain injury.13, 133 These oxidation products would be interesting to study by MSn as they are preferentially ionized in negative mode; thus, a more basic MALDI matrix such as 9aminoacridine would need to be employed. MSn of oxidized cardiolipin would also likely yield extensive fragmentation as each cardiolipin has four fatty acid substituents, that are typically unsaturated;134 therefore, one could potentially determine by MSn whether one or more of these fatty acids become oxidatively modified following in vitro oxidation. Lastly, cholesterol is another target of oxidation that could be studied using the MALDI MSn methods developed in this work. As presented in Chapter 4, chol esterol appears to be a primary target of oxidation in tissue exposed to UV light. For in vitro oxidation experiments, this work primarily focused on autoxidation and UV light induced oxidation; however, many other in vitro oxidation methods have been repo rted and could be utilized to investigate whether the mechanism of oxidation (e.g., free radical vs singlet oxygen) has any effect on what oxidation products are formed. Additionally, oxidation products caused by enzymatic oxidation by either lipoxygenase or myeloperoxidase are expected to selectively oxidize unsaturated fatty acid s at specific positions. All of these aforementioned in vitro methods are applicable to oxidation of phospholipid standards; however, in vitro oxidation of tissue would be more ch allenging from a method development standpoint Issues involving the application of oxidant to the tissue without causing analyte migration would need to be considered.

PAGE 125

125 In addition to studying oxidation of other phospholipid classes and investigating diffe rent in vitro oxidation methods, future studies will aim to identify OxPLs in biological samples from various disease models. Due to the high abundance of polyunsaturated p hospholipids and exposure to UV light, the retina is an ideal model for studying in vivo lipid oxidation.35 However, preliminary experiments with retina tissue from adult Rhesus monkeys suggested that better methods for preparation of retinal flat mounts must be developed before informative MS images can be obtained ; thus, this work was not included in the previous chapters Most notably, cracking of the tissue was observed which resulted in varied topography causing artifacts in the MS images Furthermore brain, spinal cord, and sciatic nerve tissues from SpragueDawley rat models of two similar disease states, diabetic neuropathy and dichloroacetate (DCA) induced neuropathy, hypothesized to involve oxidative damage to nervous tissue, will be studied.135 136, 137 Nervous tissues fr om healthy rats will also be analyzed as control samples for comparison. Since spatial information can be retained in MALDI, the distributions of oxidation products in various tissues will be investigated by MSI which may lead to a n improved understanding of the biological and physiopathological activities of OxPLs.

PAGE 126

126 LIST OF REFERENCES (1) Bochkov, V. N.; Oskolkova, O. V.; Birukov, K. G.; Levonen, A. L.; Binder, C. J.; Stockl, J. Antioxid. Redox Signaling 2010, 12 10091059. (2) Ma rkesbery, W. R.; Lovell, M. A. Neurobiol. Aging 1998, 19 33 36. (3) Koppaka, V.; Axelsen, P. H. Biochemistry 2000, 39 10011 10016. (4) Koppaka, V.; Paul, C.; Murray, I. V. J.; Axelsen, P. H. J. Biol. Chem. 2003, 278 3627736284. (5) Huang, L.; Estrada, R.; Yappert, M. C.; Borchman, D. Free Radical Biol. Med. 2006, 41 14251432. (6) Berliner, J. A.; Leitinger, N.; Tsimikas, S. J. Lipid Res. 2009, 50, S207 S212. (7) Qin, J.; Goswami, R.; Balabanov, R.; Dawson, G. J. Neurosci. Res. 2007, 85, 977984. (8) Baskol, G.; Demir, H.; Baskol, M.; Kilic, E.; Ates, F.; Karakukcu, C.; Ustdal, M. Cell Biochem. Funct. 2006, 24 307 311. (9) Adibhatla, R. M.; Hatcher, J. F. Antioxid. Redox Signaling 2010, 12, 125169. (10) Yamashita, M.; Fenn, J. B. J. Phys. Chem. 1984, 88 44514459. (11) Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.; Yoshida, Y.; Yoshida, T.; Matsuo, T. Rapid Commun. Mass Spectrom. 1988, 2 151 153. (12) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 22992301. (13) Sparvero, L. J.; Amos cato, A. A.; Kochanek, P. M.; Pitt, B. R.; Kagan, V. E.; J. Neurochem. 2010, 115 13221336. (14) Domingues, M. R. M.; Reis, A.; Domingues, P. Chem. Phys. Lipids 2008, 156 1 12. (15) O'Donnell, V. B. Biochim. Biophys. Acta, Mol. Cell Biol. Lipids 2011, 1811, 818 826. (16) Fuchs, B.; Bresler, K.; Schiller, J. Chem. Phys. Lipids 2011, 164 782 795. (17) Watson, A. D. J. Lipid Res. 2006, 47 21012111. (18) Alberts, B.; Johnson, A.; Lewis, J.; Raff, M.; Roberts, K.; Walter, P. Molecular Biology of the Cell 4 ed.; Garland Science: New York, 2002.

PAGE 127

127 (19) Berg, J. M.; Tymoczko, J. L.; Stryer, L. Biochemistry 5th ed.; W H Freeman: New York, 2002. (20) Lehninger, A. L.; Nelson, D. D. L.; Cox, M. M. Lehninger Principles of Biochemistry 4th ed.; W.H. Freeman: New York, 2005. (21) Vance, J. E.; Vance, D. E. Biochemistry of Lipids, Lipoproteins and Membranes 5th ed.; Elsevier Science: Oxford, 2008. (22) Farooqui, A. A.; Horrocks, L. A.; Farooqui, T. Chem. Phys. Lipids 2000, 106 1 29. (23) Prasad, R. Ma nual on membrane lipids ; Springer Verlag: New York, 1996. (24) Beckerman, M. In Molecular and Cellular Signaling; Springer: New York, 2005, pp 161186. (25) Yeagle, P. The membranes of cells ; Academic Press: Orlando, 1987. (26) Fruhwirth, G. O.; Loidl, A.; Hermetter, A. Biochim. Biophys. Acta 2007 1772 718736. (27) Bernheim, F.; Bernheim, M. L. C. J. Biol. Chem. 1939, 127 353 360. (28) Reis, A.; Spickett, C. M. Biochim. Biophys. Acta 2012 1818 23742387. (29) Schaich, K. M. In Bailey's Industrial Oil and Fat Products ; John Wiley & Sons: New York, 2005. (30) Megli, F. A.; Russo, L. Biochim. Biophys. Acta 2008, 1778, 143 152. (31) Sparrow, C. P.; Parthasarathy, S.; Steinberg, D. J. Lipid Res. 1988, 29, 745753. (32) Khn, H.; Belkner, J.; Suzuki, H.; Yam amoto, S. J. Lipid Res. 1994 35, 17491759. (33) Khn, H.; Heydeck, D.; Hugou, I.; Gniwotta, C. J. Clin. Invest. 1997, 99, 888. (34) Gunstone, F. D., Ed. Fatty Acid and Lipid Chemistry ; Blackie Acandemic & Professional: Bishopbriggs, Glasgow, 1996. (35) G ugiu, B. G.; Mesaros, C. A.; Sun, M. J.; Gu, X. R.; Crabb, J. W.; Salomon, R. G. Chem. Res. Toxicol. 2006, 19 262 271. (36) Niki, E. Free Radical Biol. Med. 2009, 47, 469484. (37) Halliwell, B.; Chirico, S. Am. J. Clin. Nutr. 1993 57 S715S725. (38) Oh kawa, H.; Ohishi, N.; Yagi, K. Anal. Biochem. 1979, 95 351358.

PAGE 128

128 (39) Spalteholz, H.; Zschrnig, O.; Arnold, K. Prog. Lipid Res. 2004, 43, 449488. (40) Watson, J. T.; Sparkman, O. D., Eds. Introduction to Mass Spectrometry: Instrumentation, Applications and Strategies for Data Interpretation, 4th ed.; John Wiley & Sons, Ltd: West Sussex, 2007. (41) Benabdellah, F.; Seyer, A.; Quinton, L.; Touboul, D.; Brunelle, A.; Laprvote, O. Anal. Bioanal. Chem. 2010, 396 151162. (42) Schmidt, A.; Karas, M.; Dulcks, T. J. Am. Soc. Mass Spectrom. 2003 14, 492 500. (43) Hillenkamp, F.; Katalinic, J. P., Eds. MALDI MS a Practical Guide to Instrumentation, Methods and A pplications ; Wiley VCH: Weinheim, 2007. (44) Chughtai, K.; Heeren, R. M. A. Chem. Rev. (Washington, DC, U. S.) 2010 110 32373277. (45) Stubiger, G.; Belgacem, O.; Rehulka, P.; Bicker, W.; Binder, B. R.; Bochkov, V. Anal. Chem. 2010 82 55025510. (46) Cohen, S. L. Anal. Chem. 2006, 78 43524362. (47) Garrett, T. J.; Yost, R. A. Anal. Chem. 2006, 78, 24652469. (48) Landgraf, R. R.; Prieto Conaway, M. C.; Garrett, T. J.; Stacpoole, P. W.; Yost, R. A. Anal. Chem. 2009, 81, 84888495. (49) Garrett, T. J.; Yost, R. A. In Mass Spectrometry Imaging: Principles and Protocols ; Rubakhin, S. S., Sweedler, J. V., Eds.; Humana Press, 2010; Vol. 656, pp 209230. (50) Menger, R. F.; Stutts, W. L.; Anbukumar, D. S.; Bowden, J. A.; Ford, D. A.; Yost, R. A. Anal. Chem. 2011, 84 1117 1125. (51) Karas, M.; Bachmann, D.; Hillenkamp, F. Anal. Chem. 1985, 57 29352939. (52) Guo, Z.; He, L. Anal. Bioanal. Chem. 2007, 387 19391944. (53) Cohen, L.; Gusev, A. Anal. Bioanal. Chem. 2002, 373 571 586. (54) Gross, J. In Mass Spectrometry ; Springer Verlag, 2011, pp 507559. (55) Fuchs, B.; Schiller, J. In Lipids in Health and Disease; Quinn, P., Wang, X., Eds.; Springer Netherlands, 2008; Vol. 49, pp 541565. (56) Dreisewerd, K. Chem. Rev. (Washington, DC, U. S.) 2003 103 395 425.

PAGE 129

129 (57) Stubiger, G.; Belgacem, O. Anal. Chem. 2007, 79 3206 3213. (58) Griffiths, R. L.; Bunch, J. Rapid Commun. Mass Spectrom. 2012, 26 15571566. (59) Karas, M.; Glckmann, M.; Schfer, J. J. Mass Spectrom. 2000, 35 1 12. (60) Karas, M.; Krger, R Chem. Rev. (Washington, DC, U. S.) 2003 103 427440. (61) Zenobi, R.; Knochenmuss, R. Mass Spectrom. Rev. 1998, 17 337 366. (62) Knochenmuss, R.; Zenobi, R. Chem. Rev. (Washington, DC, U. S.) 2003, 103 441452. (63) Jaskolla, T.; Karas, M. J. Am. Soc Mass Spectrom. 2011, 22, 976 988. (64) Qual Browser, v. 2.0.7; LTQ XL Hardware Manual; Thermo Scientific: San Jose, Ca, 2006. (65) Krishnaveni, A.; Kumar Verma, N.; Menon, A. G.; Mohanty, A. K. Int. J. Mass Spectrom. 2008, 275 11 20. (66) Schwartz, J. C.; Senko, M. W.; Syka, J. E. P. J. Am. Soc. Mass Spectrom. 2002, 13, 659669. (67) March, R. E.; Todd, J. F. J. Quadrupole Ion Trap Mass Spectrometry 2nd ed.; Wiley: Hoboken, 2005. (68) Garrett, T. J.; Prieto Conaway, M. C.; Kovtoun, V.; Bui H.; Izgarian, N.; Stafford, G.; Yost, R. A. Int. J. Mass Spectrom. 2007, 260 166176. (69) Sleno, L.; Volmer, D. A. J. Mass Spectrom. 2004, 39, 10911112. (70) Rubakhin, S. S.; Sweedler, J. V., Eds. Mass Spectrometry Imaging: Principles and Protocols. ; Springer: New York, 2010. (71) Murphy, R. C.; Hankin, J. A.; Barkley, R. M. J. Lipid Res. 2009, 50 Suppl S317 322. (72) Sjvall, P.; Lausmaa, J.; Johansson, B. Anal. Chem. 2004, 76 42714278. (73) Berry, K. A.; Li, B.; Reynolds, S. D.; Barkley, R. M.; Gi jon, M. A.; Hankin, J. A.; Henson, P. M.; Murphy, R. C. J. Lipid Res. 2011 52, 15511560. (74) McDonnell, L. A.; Heeren, R. M. A. Mass Spectrom. Rev. 2007, 26 606643. (75) Amstalden van Hove, E. R.; Smith, D. F.; Heeren, R. M. A. J. Chromatogr., A 2010, 1217 39463954.

PAGE 130

130 (76) Cornett, D. S.; Reyzer, M. L.; Chaurand, P.; Caprioli, R. M. Nat. Methods 2007 4 828 833. (77) Schwartz, S. A.; Reyzer, M. L.; Caprioli, R. M. J. Mass Spectrom. 2003, 38, 699 708. (78) Rmpp, A.; Spengler, B. Histochem. Cell Biol. 2013, 139 759 783. (79) Landgraf, R. R. Analysis of lipids in nerve tissue by MALDI tandem mass spectrometric imaging. Ph.D. Thesis, University of Florida, Gainesville, FL, 2009. (80) Puolitaival, S.; Burnum, K.; Cornett, D. S.; Caprioli, R. J. Am. Soc. Mass Spectrom. 2008, 19, 882 886. (81) Axelsson, J.; Hoberg, A. M.; Waterson, C.; Myatt, P.; Shield, G. L.; Varney, J.; Haddleton, D. M.; Derrick, P. J. Rapid Commun. Mass Spectrom. 1997, 11, 209 213. (82) Baluya, D. L.; Garrett, T. J.; Yost, R. A. Anal. C hem. 2007, 79 68626867. (83) Chen, Y.; Allegood, J.; Liu, Y.; Wang, E.; CachonGonzalez, B.; Cox, T. M.; Merrill, A. H.; Sullards, M. C. Anal. Chem. 2008, 80 2780 2788. (84) Hankin, J. A.; Barkley, R. M.; Murphy, R. C. J. Am. Soc. Mass Spectrom. 2007, 1 8 16461652. (85) Garrett, T. J. Imaging small molecules in tissue by matrix assisted laser desorption/ionization tandem mass spectrometry. Ph.D. Thesis, University of Florida, Gainesville, FL, 2006. (86) Strupat, K.; Kovtoun, V.; Bui, H.; Viner, R.; Stafford, G.; Horning, S. J. Am. Soc. Mass Spectrom. 2009, 20, 14511463. (87) Yin, H.; Cox, B. E.; Liu, W.; Porter, N. A.; Morrow, J. D.; Milne, G. L. J. Mass Spectrom. 2009, 44, 672 680. (88) Salomon, R. G.; Gu, X. Chem. Res. Toxicol. 2011, 24 17911802. (8 9) Sun, M. J.; Finnemann, S. C.; Febbraio, M.; Shan, L.; Annangudi, S. P.; Podrez, E. A.; Hoppe, G.; Darrow, R.; Organisciak, D. T.; Salomon, R. G.; Silverstein, R. L.; Hazen, S. L. J. Biol. Chem. 2006 281 4222 4230. (90) Leitinger, N. Curr. Opin. Lipidol. 2003 14 421 430. (91) Lee, S.; Birukov, K. G.; Romanoski, C. E.; Springstead, J. R.; Lusis, A. J.; Berliner, J. A. Circ. Res. 2012, 111 778 799. (92) Gao, D.; Ashraf, M. Z.; Kar, N. S.; Lin, D.; Sayre, L. M.; Podrez, E. A. J. Biol. Chem. 2010, 285 44474454.

PAGE 131

131 (93) Tokumura, A.; Sumida, T.; Toujima, M.; Kogure, K.; Fukuzawa, K.; Takahashi, Y.; Yamamoto, S. J. Lipid Res. 2000, 41, 953962. (94) KayganichHarrison, K. A.; Murphy, R. C. Anal. Biochem. 1994 221, 16 24. (95) Watson, A. D.; Leitinger, N.; Navab, M.; Faull, K. F.; Hrkk, S.; Witztum, J. L.; Palinski, W.; Schwenke, D.; Salomon, R. G.; Sha, W.; Subbanagounder, G.; Fogelman, A. M.; Berliner, J. A. J. Biol. Chem. 1997, 272 13597 13607. (96) Frey, B.; Haupt, R.; Alms, S.; Holzmann, G.; Kn ig, T.; Kern, H.; Kox, W.; Rstow, B.; Schlame, M. J. Lipid Res. 2000 41, 11451153. (97) Nakamura, T.; Henson, P. M.; Murphy, R. C. Anal. Biochem. 1998, 262, 23 32. (98) Hall, L. M.; Murphy, R. C. Anal. Biochem. 1998, 258 184 194. (99) Reis, A.; Domingu es, P.; Ferrer Correia, A. J. V.; Domingues, M. R. M. J. Mass Spectrom. 2004, 39, 15131522. (100) Nakanishi, H.; Iida, Y.; Shimizu, T.; Taguchi, R. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2009 877 1366 1374. (101) Spickett, C. M.; Rennie, N.; Winter, H.; Zambonin, L.; Landi, L.; Jerlich, A.; Schaur, R. J.; Pitt, A. R. Biochem. J. 2001 355 449457. (102) Oskolkova, O. V.; Afonyushkin, T.; Preinerstorfer, B.; Bicker, W.; von Schlieffen, E.; Hainzl, E.; Demyanets, S.; Schabbauer, G.; Lindner, W.; Tselepis, A. D.; Wojta, J.; Binder, B. R.; Bochkov, V. N. J. Immunol. 2010, 185 7706 7712. (103) Domingues, M. R. M.; Simes, C.; da Costa, J. P.; Reis, A.; Domingues, P. Biomed. Chromatogr. 2009, 23 588601. (104) Reis, A.; Domingues, P.; Ferrer Correia, A. J. V.; Domingues, M. R. M. Rapid Commun. Mass Spectrom. 2004, 18, 28492858. (105) Reis, A.; Domingues, M. R. M.; Amado, F. M. L.; Ferrer Correia, A. J. V.; Domingues, P. Biomed. Chromatogr. 2005 19 129 137. (1 06) Stbiger, G.; Belgacem, O. Anal. Chem. 2007, 79 3206 3213. (107) van den Brink, O. F.; Boon, J. J.; O'Connor, P. B.; Duursma, M. C.; Heeren, R. M. A. J. Mass Spectrom. 2001, 36 479492. (108) Bischoff, A.; Eibisch, M.; Fuchs, B.; Suss, R.; Schurenber g, M.; Suckau, D.; Schiller, J. Acta Chromatogr. 2011, 23, 365 375. (109) Vance, J. E. J. Lipid Res. 2008 49 13771387.

PAGE 132

132 (110) Domingues, P.; Domingues, M. R. M.; Amado, F. M. L.; Ferrer Correia, A. J. Rapid Commun. Mass Spectrom. 2001, 15, 799 804. (111) Pulfer, M.; Murphy, R. C. Mass Spectrom. Rev. 2003, 22, 332 364. (112) Hsu, F. F.; Turk, J. J. Am. Soc. Mass Spectrom. 2003, 14, 352 363. (113) Han, X.; Gross, R. W. J. Am. Soc. Mass Spectrom. 1995, 6 12021210. (114) Hsu, F. F.; Turk, J. J. Am. Soc. Mass Spectrom. 2010, 21, 657 669. (115) Hsu, F. F.; Turk, J. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2009, 877, 26732695. (116) Cheng, C.; Gross, M. J. Am. Soc. Mass Spectrom. 1998, 9 620 627. (117) Suzuki, M.; Kamei, M.; Itabe, H.; Yoneda, K .; Bando, H.; Kume, N.; Tano, Y. Mol. Vis. 2007, 13, 772 778. (118) Busik, J. V.; Reid, G. E.; Lydic, T. A. Methods Mol. Biol. 2009 579 33 70. (119) Khaselev, N.; Murphy, R. C. Free Radical Biol. Med. 1999, 26, 275 284. (120) von Schlieffen, E.; Oskolkov a, O. V.; Schabbauer, G.; Gruber, F.; Bluml, S.; Genest, M.; Kadl, A.; Marsik, C.; Knapp, S.; Chow, J.; Leitinger, N.; Binder, B. R.; Bochkov, V. N. Arterioscler., Thromb., Vasc. Biol. 2009, 29 356 362. (121) Hsu, F. F.; Turk, J. J. Mass Spectrom. 2000, 3 5 595 606. (122) Chen, X.; Steill, J. D.; Oomens, J.; Polfer, N. C. J. Am. Soc. Mass Spectrom. 2010, 21 13131321. (123) Radak, Z.; Zhao, Z.; Goto, S.; Koltai, E. Mol. Aspects Med. 2011, 32 305 315. (124) Subbanagounder, G.; Watson, A. D.; Berliner, J. A. Free Radical Biol. Med. 2000, 28 17511761. (125) Watson, B. D.; Busto, R.; Goldberg, W. J.; Santiso, M.; Yoshida, S.; Ginsberg, M. D. J. Neurochem. 1984, 42 268 274. (126) Marathe, G. K.; Harrison, K. A .; Murphy, R. C.; Prescott, S. M.; Zimmerman, G. A.; McIntyre, T. M. Free Radical Biol. Med. 2000, 28 17621770. (127) Amstalden van Hove, E. R.; Blackwell, T. R.; Klinkert, I.; Eijkel, G. B.; Heeren, R. M. A.; Glunde, K. Cancer Res. 2010 70 90129021. (128) Alexandrov, T. BMC Bioinf. 2012, 13 Suppl 16, S11.

PAGE 133

133 (129) Heeren, R. M. A. Surf. Interface Anal. 2009, 41 675 685. (130) Hasegawa, M.; Hakamata, H.; Matsunaga, I.; Kusu, F. Eur. J. Lipid Sci. Technol. 2011, 113 423 429. (131) Lyons, M. A.; Brown, A. J. Int. J. Biochem. Cell Biol. 1999, 31, 369 375. (132) Milne, G. L.; Porter, N. A. Lipids 2001, 36 1265 1275. (133) Bayir, H.; Tyurin, V. A.; Tyurina, Y. Y.; Viner, R.; Ritov, V.; Amoscato, A. A.; Zhao, Q.; Zhang, X. J.; JaneskoFeldman, K. L.; Alexander, H.; Basova, L. V.; Clark, R. S. B.; Kochanek, P. M.; Kagan, V. E. Ann. Neurol. 2007 62, 154169. (134) Tyurina, Y. Y.; Kini, V.; Tyurin, V. A.; Vlasova, I. I.; Jiang, J.; Kapralov, A. A.; Belikova, N. A.; Yalowich, J. C.; Kurnikov, I. V.; Kagan, V. E. Mol. Pharmacol. 2006, 70 706717. (135) Calcutt, N. A.; Lopez, V. L.; Bautista, A. D.; Mizisin, L. M.; Tor res, B. R.; Shroads, A. L.; Mizisin, A. P.; Stacpoole, P. W. J. Neuropathol. Exp. Neurol. 2009, 68 985 993. (136) Feldman, E. L. J. Clin. Invest. 2003, 111 431433. (137) Cunha, J. M.; Jolivalt, C. G.; Ramos, K. M.; Gregory, J. A.; Calcutt, N. A.; Mizisi n, A. P. Metab., Clin. Exp. 2008, 57, 873 881.

PAGE 134

134 BIOGRAPHICAL SKETCH Whitney Leigh Stutts, the daughter of Debbie and Terry Stutts and older sister to Lauren Stutts was born in WinstonSalem, North Carolina She grew up in the beautiful foothills of NC where she enjoyed the outdoors and local music. After graduating from East Wilkes High School, Whitney began a degree in c hemistry at N orth Carolina State University. While at NC State, Whitney was involved in the University Scholars Program and Alpha Chi Sigma, the professional chemistry fraternity. She also conducted undergraduate research under the direction of Dr. Damian Shea, studying the fate and concentrations of steroid hormones in North Carolinas surface waters. During the summer s, Whitney worked as a clinical laboratory technician at Suncare Research Laboratory assessing the photostability, phototoxicity, and photoallergenicity of sunscreens and dermatological products through in vitro and in vivo studies In December of 2007, Whitney graduated m agna cum laude with a Bachelor of Science degr ee in chemistry and a minor in environmental toxicology. In 2008 Whi t ney began her graduate career at the Uni versity of F lorida, pursuing a doctorate of philosophy in chemistry. She joined Dr. Richard A. Yosts group and, for the past five years, has invest igated oxidized lipids by MALDI MSn. Over the course of her graduate career, Whitney had several collaborations, domesti c and international, and greatly enjoyed merging two of her passions, science and traveling. Whitney completed her doctorate in August 2013. She hopes to pursue a career in mass spectrometry and lead others in their scientific endeavors.