A Dissertation

entitled

The Investigation of Potential Salivary Biomarkers of Acute Stress Using Proteomics

and Mass Spectrometry

by

Rachel K. Marvin

Submitted to the Graduate Faculty as partial fulfillment of the requirements for the

Doctor of Philosophy Degree in Chemistry

______Dr. Dragan Isailovic, Committee Chair

______Dr. Mark R. Mason, Committee Member

______Dr. John J. Bellizzi, Committee Member

______Dr. Kenneth Hensley, Committee Member

______Dr. Amanda Bryant-Friedrich, Dean College of Graduate Studies

The University of Toledo

August 2016

Copyright 2016, Rachel K. Marvin

This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author.

An Abstract of

The Investigation of Potential Salivary Protein Biomarkers of Acute Stress Using

Proteomics and Mass Spectrometry

by

Rachel K. Marvin

Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Doctor of Philosophy Degree in Chemistry

The University of Toledo

August 2016

The goal of the present research was to elucidate salivary biomarkers of acute stress using proteomic approaches. Acute stress is marked by an increased activity of the sympathetic branch of the autonomic nervous system. Prolonged exposure to acute stress may result in fatigue and chronic stress. Therefore, acute stress has potential social and economic implications in which the performance of workers in high stress occupations

(e.g., health care and law enforcement professionals) may be negatively affected. To aide in the monitoring a stress an objective assessment using biomarkers is desirable.

Saliva is an optimal body fluid for the discovery and investigation of biomarkers of this physiological state because its secretion is controlled by the autonomic nervous system, and its collection is easy and noninvasive. As people secrete 1-2 L of saliva per day, ample quantities of saliva are typically obtained for qualitative and quantitative analyses of potential biomarkers. In addition, saliva is primarily comprised of water, electrolytes, and hormones whose abundance may be monitored. Consequently, iii

the present analyses involved the examination of the human salivary proteomes at different time points corresponding to non-stressed and acutely stressed states to discover novel protein markers of acute stress.

The first project utilized a model system employing the cold pressor test (CPT) to reproducibly induce stress. Participants immersed their non-dominant hand into cold water for a maximum of five minutes, which increases the activity of sympathetic nervous system for the majority of people. Whole saliva was collected before, immediately after and 20 minutes after the CPT. A variety of gel electrophoresis techniques (i.e., SDS-PAGE, 2D-PAGE and 2D-DIGE) was used to assess alterations in the salivary proteome at these three time points. When equal volumes of saliva were analyzed by SDS-PAGE, minor differences in the abundance of some salivary proteins were observed after the CPT. For instance, the amount of cystatins and alpha-amylase is increased immediately after and 20 minutes after the CPT in comparison to the sample obtained prior to the CPT. These differences were also noticed after HPLC was used to separate equal amounts of proteins from the whole saliva samples. Hypothesizing that protein phosphorylation may be influenced by the CPT, tryptic phosphopeptides were isolated from the salivary proteome using immobilized metal affinity chromatography

(IMAC). After enrichment, the phosphopeptides were analyzed by nanoHPLC-MALDI-

MS/MS, and some differences in the phosphorylation of salivary proteins were observed.

Comparing the nanoHPLC chromatograms the peptides eluting at retention times of ~25 and 29.7 minutes had the highest absorbance for the samples collected immediately and

20 minutes after the CPT, indicating that these peptides are present in higher amounts after the stressor. In contrast, the peptide(s) at retention time of ~26.7 minutes had the

iv

highest absorbance, and the samples collected after the CPT stressor had a relatively low absorbance. The peptides at these retention times were identified as originating from salivary acidic proline-rich phosphopeptide 1/2. However, the changes in salivary protein abundances and phosphorylation require further characterization before they can be verified as biomarkers of acute stress induced by the CPT.

The second project investigated proteins from saliva of medical residents who were placed into a stressed state by performing emergency medicine simulations. Whole saliva was collected from eight medical residents prior to performing the simulation, after the simulation, and the next morning upon waking. Salivary proteins were identified with

SDS-PAGE and peptide mass fingerprinting, as well as by nanoHPLC-ESI-MS/MS.

Relative quantification of the alpha-amylase, cystatin-type S family, a ~26 kDa band and a low-molecular weight (<10 kDa) band was investigated in relation to the stressful emergency medicine scenarios. Densitometry and statistical analyses of Coomassie stained SDS-PAGE gels indicated an increase in the relative amount of salivary alpha- amylase and cystatin type-S bands after simulated emergency interventions in comparison to the pre-simulation and waking time points. Some differences in the ~26 kda and the low-molecular weight bands at different time points were also observed. The results indicate that proteins, such as alpha-amylase, cystatins and -3, may be considered as putative salivary biomarkers of acute stress, but they need further validation in a larger sample population.

v

Acknowledgements

First and foremost, I would like to thank my advisor, Dr. Dragan Isailovic. He has been an invaluable mentor that has played an instrumental role in my success in the lab. I am extremely grateful for all the opportunities and knowledge he has provided me. I would also like to thank my committee members, Drs. Mason, Bellizzi and Hensley.

Additionally, I thank our collaborators Dr. Surya Nauli, Maki Takahashi, Dr.

Giovannucci, Dr. Hensley, Muncharie Saepoo models, Dr. Don White and Simao Ye.

Also thanked are my labmates who have come and gone during my time in the lab. This includes Zhen, Yang, Suraj, Raymond, Ravi, Siddhita, Krishani, Sangee, Thilini and all the undergraduates who have worked in the lab, especially Jonathan. They have been very helpful and made the lab enjoyable. I also need to thank Charlene Hansen and the rest of the graduate students in the department for their support. In particular, I need to acknowledge Jen, Steve and Jared for their constant motivation and friendship.

Most importantly, I need to thank my family for their support, encouragement and love. My mom has been a source of never-ending strength and inspiration always pushing me to reach my maximum potential for which I will forever be grateful. Both my mom and my dad provided many meals after long days in the lab. My sisters, Andrea and

Jessica, have also been a source of great advice and fun throughout the years. Without them, none of this would have been possible.

vi

Table of Contents

An Abstract of ...... iii

Acknowledgements ...... vi

Table of Contents ...... vii

List of Tables ...... xiii

Table of Figures ...... xv

List of Abbreviations ...... xx

Chapter 1 ...... 1

Introduction to Mass Spectrometry for Biomolecule Analysis ...... 1

1.1 Mass Spectrometry Background ...... 1

1.2 Mass Spectrometer Components ...... 2

1.2.1 Ion Sources...... 3

1.2.1.1 MALDI ...... 3

1.2.1.2 ESI ...... 6

1.2.1.3 NanoESI ...... 7

1.2.2 Mass Analyzers ...... 8

1.2.2.1 TOF ...... 10

1.2.2.2 Quadrupole ...... 13

vii

1.2.2.3 Ion Trap ...... 14

1.2.2.4 Orbitrap ...... 16

1.3 Biomolecule Analysis by Mass Spectrometry ...... 17

1.3.1 Proteins and Peptides ...... 18

1.3.1.1 Peptide Mass Fingerprinting ...... 18

1.3.1.2 Tandem Mass Spectrometry ...... 21

1.3.1.2.1 MS/MS using Bruker’s MALDI-TOF/TOF ...... 23

1.3.1.2.2 MS/MS using Thermo Orbitrap Fusion ...... 24

Chapter 2 ...... 28

Sample Preparation and Separation Techniques Used for Proteomic Analyses ...... 28

2.1 Sample Preparation for Mass Spectrometric Analyses ...... 28

2.2 Desalting of Proteins and Peptides ...... 29

2.3 Gel Based Separations ...... 31

2.3.1 SDS-PAGE ...... 31

2.3.2 2D-PAGE ...... 33

2.3.3 Difference Gel Electrophoresis ...... 34

2.4 HPLC Separations ...... 35

2.5 Fe-NTA columns ...... 37

Chapter 3 ...... 39

Mass Spectrometric Investigation of Potential Salivary Protein Biomarkers of Stress

Induced by the Cold Pressor Test ...... 39

3.1 Introduction ...... 39

3.1.1 Diagnostic utility of saliva and methods of collection ...... 39 viii

3.1.2 Acute stress and the cold pressor test ...... 42

3.2 Experimental ...... 43

3.2.1 Chemicals and Materials ...... 43

3.2.2 Cold Pressor Test and Saliva Collection ...... 45

3.2.3 Salivary protein separation and identification ...... 46

3.2.3.1 SDS-PAGE ...... 47

3.2.3.2 2D-PAGE ...... 48

3.2.3.3 2D-Difference gel electrophoresis (DIGE) ...... 51

3.2.3.4 In-gel digestion and peptide mass fingerprinting ...... 53

3.2.3.5 HPLC of salivary proteins in whole saliva ...... 54

3.2.3.6 nanoHPLC-MALDI-MS/MS and ESI-MS/MS of trypsin digested whole

saliva ...... 56

3.2.4 Depletion of alpha-amylase from whole saliva ...... 58

3.2.5 Investigation of salivary phosphoproteins and phosphopeptides...... 59

3.2.5.1 Gel-based analysis of phosphoproteins ...... 59

3.2.5.2 Gel-free analysis of phosphopeptides ...... 60

3.3 Results and Discussion ...... 63

3.3.1 Protein Identification by Gel-Based Methods ...... 63

3.3.2 Protein identification using HPLC separation of saliva ...... 68

3.3.2.1 HPLC of salivary proteins in whole saliva ...... 68

3.3.2.2 Protein identification using nanoHPLC-MALDI-MS/MS and ESI-

MS/MS of trypsin digested whole saliva ...... 71

3.3.3 Amylase depletion for lower abundance protein identification ...... 75

ix

3.3.4 Investigation of salivary phosphoproteins...... 78

3.3.4.1 Gel-based methods of protein identification ...... 79

3.3.4.2 Gel-free methods of salivary phosphoprotein identification ...... 81

3.4 Conclusions and Future Directions ...... 88

Chapter 4 ...... 92

A Preliminary Investigation of Changes in the Salivary Proteome in Response to Acute

Stress in Medical Residents Performing Advanced Clinical Simulations ...... 92

4.1 Introduction ...... 92

4.2 Experimental ...... 93

4.2.1 Chemicals and Materials ...... 93

4.2.2 Ethics Approval ...... 94

4.2.3 Emergency Medicine Simulations and Saliva Collection ...... 94

4.2.4 SDS-PAGE ...... 96

4.2.5 In-gel digestion and peptide mass fingerprinting ...... 97

4.2.6 Statistical Analysis ...... 97

4.2.7 Whole saliva digestion ...... 98

4.2.8 nanoHPLC-ESI-MS/MS ...... 99

4.2.9 Analysis of commercial hisatin-3 ...... 100

4.3 Results and Discussion ...... 101

4.3.1 SDS-PAGE and peptide mass fingerprinting ...... 101

4.3.2 Densitometry and statistical analyses ...... 106

4.3.2.1 Results of the t-test ...... 110

4.3.2.2 Results of the Wilcoxon signed-rank test ...... 112

x

4.3.3 Further analyses of the low-molecular weight band ...... 114

4.3.4 nanoHPLC-ESI-MS/MS ...... 116

4.3.5 Additional Discussion ...... 118

4.4 Conclusions and Future Directions ...... 120

Chapter 5 ...... 122

Summary ...... 122

References ...... 124

Appendix A: Supplemental Cold Pressor Test Salivary Proteome Figures and

Identifications ...... 136

Appendix B: Tables of the raw and normalized band areas ...... 146

Appendix C: List of proteins identified by nanoHPLC-ESI-MS/MS from medical resident saliva ...... 151

Appendix D: Imaging Mass Spectrometry of Normal and Polycystic Kidney Disease

Mouse Kidney Tissues ...... 155

D.1 Introduction to Polycystic Kidney Disease ...... 155

D.2 Principle of MALDI-MSI ...... 158

D.2.1 Sample Preparation...... 160

D.2.2 Matrix Application ...... 163

D.2.3 Data Acquisition ...... 165

D.2.4 Data Normalization ...... 166

D.2.5 Statistical Analysis ...... 167

D.2.5.1 Hierarchical clustering ...... 167

D.2.5.2 Principal Component Analysis ...... 168

xi

D.3 Experimental ...... 169

D.3.1 Chemicals and Materials ...... 169

D.3.2 Animal Samples ...... 170

D.3.3 Tissue Preparation ...... 170

D.3.4 MALDI-MS Imaging ...... 172

D.3.5 nanoHPLC-MALDI-MS/MS ...... 173

D.4 Results and Discussion ...... 175

D.4.1 Normal and PKD Kidney Protein Distribution Imaging ...... 175

D.4.2 Normal and PKD Kidney Peptide Distribution Imaging...... 177

D.4.3 PCA ...... 178

D.4.4 nanoHPLC-MALDI-MS/MS of Extracted Peptides ...... 180

D.5 Conclusions and Future Directions ...... 184

Appendix References ...... 186

xii

List of Tables

Table 1.1 Comparison of common mass analyzers adapted from de Hoffmann and

Stroobant2 ...... 9

Table 2.1 Comparison of solvents and volumes used for ZipTip and StageTip desalting

of peptides ...... 30

Table 3.1 Isoelectric focusing program used for pH 3-10, 7 cm IPG strips...... 50

Table 3.2 Protein identified by peptide mass fingerprinting from 2D-PAGE ...... 66

Table 3.3 List of Mascot search identified CPT salivary phosphopeptides originating

from salivary acidic proline-rich phosphopeptide ½ (Uniprot # P02810)...... 88

Table 4.1 Demographics of the first-year medical residents participating in the study ... 95

Table 4.2 Peptide mass fingerprinting identified salivary proteins ...... 104

Table 4.3 Average of the ratios of the relative amounts of the four protein bands

subjected to densitometric analysis ...... 107

Table 4.4 Summary of the relative quantification of the protein bands at the three time

points as determined by the t-test ...... 110

Table 4.5 Summary of the relative quantification of the protein bands determined by

Wilcoxon signed-rank test ...... 113

Table 4.6 NanoHPLC-ESI-MS/MS identifications of salivary proteins co-separating in

the low-molecular weight SDS-PAGE protein band ...... 117

xiii

Table 4.7 NanoHPLC-ESI-MS/MS identifications of salivary proteins co-separating in

the ~26 kDa SDS-PAGE protein band ...... 118

Table A.1 List of salivary proteins identified from cold pressor test saliva ...... 139

xiv

List of Figures

Figure 1-1 Generalized schematic of a mass spectrometer ...... 2

Figure 1-2 Scheme of the MALDI principle adapted from de Hoffmann and Stroobant2 . 4

Figure 1-3 Overall schematic of the ion formation and desolvation process in ESI

(adapted from Ho et al.)3 ...... 7

Figure 1-4 General scheme of linear time-of-flight mass analyzer. The molecular weight

of the ion is indicated by the size of the circle ...... 11

Figure 1-5 Illustration of a reflectron time-of-flight mass analyzer. The spheres represent

two ions with the same mass and charge where the outlined one has slightly

higher kinetic energy than the shaded in sphere...... 12

Figure 1-6 General diagram of ion trajectories through a quadrupole mass analyzer ..... 13

Figure 1-7 Basic layout of a 3D ion trap ...... 15

Figure 1-8 Overview of peptide mass fingerprinting for protein identification ...... 19

Figure 1-9 Biemann nomenclature for distinguishing common peptide fragmentation

sites on a generic pentapeptide (adapted from Hjernø and Jensen)13 ...... 22

Figure 1-10 General schematic of MS/MS using the LIFT on the Bruker MALDI-

TOF/TOF (adapted from Suckau et al.)16 ...... 24

Figure 1-11 Overall reaction of fluoranthene radical anions with a peptide for

fragmentation (adapted from Udeshi et al.)20 ...... 27

xv

Figure 2-1 Simplified phosphopeptide enrichment workflow using Fe-NTA columns

(adapted from Dunn et al.)43 ...... 38

Figure 3-1 Chemical structures of (A) Cyanine3 NHS ester and (B) Cyanine5 NHS ester

...... 44

Figure 3-2 Participant immersing their hand in cold water as part of the cold pressor test

...... 46

Figure 3-3 Simplified workflow of the methods used for the separation and identification

of salivary proteins ...... 47

Figure 3-4 Overall 2D-PAGE workflow ...... 49

Figure 3-5 2D-DIGE workflow...... 53

Figure 3-6 Overview of the commercial Fe-NTA phosphopeptide enrichment kit

workflow ...... 61

Figure 3-7 Image of SDS-PAGE separation of equal volumes and equal total protein

loaded on a 12% polyacrylamide gel visualized with (A) Coomassie blue G-250

and (B) SYPRO Ruby...... 64

Figure 3-8 Image of 2D-PAGE separation of proteins originating from saliva collected

prior to the CPT labeled with some of the identified salivary proteins ...... 65

Figure 3-9 HPLC chromatogram showing the separation of salivary proteins. Pooled

saliva A was collected prior to the CPT. Pooled saliva B was collected

immediately after the CPT. Pooled saliva C was collected 20 min after CPT.

Proteins were identified by nano-HPLC-MALDI-MS/MS...... 70

Figure 3-10 Base peak chromatogram of the separation of 1.33 µg of salivary tryptic

peptides from a sample collected before the CPT by nanoHPLC-ESI-MS/MS ... 74

xvi

Figure 3-11 Image of SDS-PAGE of whole saliva and saliva subjected to alpha-amylase

depletion using a potato starch column ...... 78

Figure 3-12 Image of SDS-PAGE of whole saliva samples prior to, immediately after

and 20 minutes after the CPT. Samples were ran in duplicate on the gel with 35

μL and 45 μL of saliva loaded, respectively. (A) Total protein visualization using

Coomassie blue G-250 (B) Phosphoprotein visualization using Pro-Q Diamond.80

Figure 3-13 MALDI-MS spectra of phosphopeptides isolated using the Fe-NTA

Phosphopeptide Enrichment Kit. (A) Pooled saliva collected prior to the CPT (B)

pooled saliva collected immediately after the CPT (C) pooled saliva collected 20

min after the CPT ...... 84

Figure 3-14 nanoHPLC chromatogram overlay of the salivary phosphopeptides isolated

using the Fe-NTA Phosphopeptide Enrichment Kit from saliva collected prior to,

immediately after and 20 min after the CPT ...... 85

Figure 3-15 nanoHPLC chromatogram of saliva collected immediately after the CPT.

Peptides were identified by MALDI-MS/MS ...... 87

Figure 4-1 Overview of the timeline of saliva collection and the emergency medicine

simulation ...... 96

Figure 4-2 Images of the SDS-PAGE separation of the salivary proteins obtained from

(A) resident 1, (B) resident 5, (C) resident 7 and (D) resident 8, the four male

medical residents, collected upon waking, prior to the simulation (pre) and after

the simulation (post)...... 102

Figure 4-3 A representative SDS-PAGE gel labeled with the most prevalent proteins

identified by PMF ...... 103

xvii

Figure 4-4 Boxplots of the relative ratios of alpha-amylase, the ~26 kDa, cystatins, and

the low-molecular weight protein bands for the ratio of the protein band areas (A)

after the simulation compared to waking the morning after the simulation, (B)

after the simulation compared to before the simulation and (C) waking the

morning after the simulation compared to before the simulation...... 109

Figure 4-5 MALDI-MS of the tryptic digest of (A) the 10 kDa band in-gel digest from

medical resident saliva, (B) the solution digest of the synthetic histatin-3 and (C)

the in-gel digest of the synthetic histatin-3. Common histatin-3 peaks are marked

by a solid outline box, whereas histatin-3 peaks unique to each sample are marked

by a dashed outline box...... 115

Figure 4-6 Separation of 300 ng of synthetic histatin-3 and medical resident saliva

collected upon waking, prior to the emergency medicine simulation (pre) and after

the emergency medicine simulation (post) ...... 116

Figure A-1 nanoHPLC chromatogram of the trypsin digested HPLC fraction collected

from 24-26 min from pooled saliva B...... 136

Figure A-2 nanoHPLC-MALDI-MS/MS of trypsin digested whole saliva ...... 137

Figure A-3 Venn diagram of the functions of the salivary proteins identified using

nanoHPLC-ESI-MS/MS ...... 137

Figure A-4 MALDI-MS of (A) phosphopeptide enriched casein digest and (B) casein

digest prior to enrichment ...... 138

Figure D-1 Overview of the MALDI mass spectrometry imaging workflow for the

analysis of proteins and peptides ...... 171

xviii

Figure D-2 Normalized average spectra (A-F) of six normal mouse kidney tissues from

the same region of the kideny mounted on the same ITO slide ...... 177

Figure D-3 PCA of MALDI-MSI data. All spectra from a single healthy kidney section

and from a late-stage PKD kidney section were loaded for PCA. (A) PCs in 3D

plot, (B) PCs in 2D plot of the first two principal components, (C) 3D loading

plot, (D) 2D loading plot ...... 180

Figure D-4 nanoHPLC of peptides extracted from six imaged sections of normal mouse

kidney tissue...... 182

Figure D-5 nanoHPLC-MALDI-MS/MS of the fraction containing the peptide ion with

m/z 1529.8 which was selected for MS/MS analysis for protein identification.. 183

Figure D-6 nanoHPLC-MALDI-MS/MS of the peptide ion with m/z 1529.8 identified as

hemoglobin subunit alpha by Mascot ...... 183

Figure D-7 nanoHPLC-MALDI-MS/MS of peptide ion with m/z 1168 identified as

cytochrome c by Mascot ...... 184

xix

List of Abbreviations

ACN ...... Acetonitrile ADPKD ...... Autosomal dominant polycystic kidney disease ARPKD ...... Autosomal recessive polycystic kidney disease

CHCA ...... α-Cyano-4-hydroxycinnamic acid CPT ...... Cold pressor test CT ...... Computed tomography Cy3 ...... Cyanine3 NHS ester Cy5 ...... Cyanine5 NHS ester

Da ...... Dalton DHB ...... 2,5-dihydroxybenzoic acid DI ...... Deionized DIGE ...... Difference gel electrophoresis DMSO ...... Dimethyl sulfoxide DTT ...... Dithiothreitol

FA ...... Formic Acid FFPE ...... Formalin-fixed and paraffin embedded

H&E ...... Hematoxylin and eosin HPLC ...... High-performance liquid chromatography IAM ...... Iodoacetamide IPG ...... Immobilized pH gradient ITO ...... Indium-tin oxide

LDI ...... Laser desorption/ionization MALDI ...... Matrix-assisted laser desorption/ionization MRI ...... Magnetic resonance imaging MS ...... Mass spectrometry MSI ...... Mass spectrometry imaging MS/MS ...... Tandem mass spectrometry m/z ...... Mass-to-charge ratio xx

NTA ...... Nitrilotriacetate OCT...... Optimal cutting temperature media

PC1 ...... First principal component PCA ...... Principal component analysis pI ...... Isoelectric point PKD...... Polycystic kidney disease ppm ...... Parts per million rpm ...... Revolutions per minute SA ...... Sinapinic acid SDS-PAGE ...... Sodium dodecyl sulfate polyacrylamide gel electrophoresis SPE ...... Solid-phase extraction

TEMED ...... N,N,N,N’-tetramethylenediamine TFA ...... Trifluoroacetic acid TIC ...... Total ion current TOF ...... Time-of-flight

UV ...... Ultraviolet v:v ...... volume-to-volume

xxi

Chapter 1

Introduction to Mass Spectrometry for Biomolecule

Analysis

1.1 Mass Spectrometry Background

Mass spectrometry is a common analytical technique that is used to provide qualitative and quantitative information. A mass spectrometer can be broken down into four main components: the inlet, the ion source, the mass analyzer and the detector.1, 2

The inlet is where the sample to be analyzed is introduced. The sample can be directly infused as a liquid or introduced as a solid or a gas. Other techniques can also be directly coupled to a mass spectrometer, including gas chromatography and liquid chromatography.2, 3 The ion source is where gaseous ions are formed. After ion formation, the ions travel to the mass analyzer where they are separated according to their mass-to-charge (m/z) ratios.4 Then, they pass to the detector which measures their abundance and can be transformed into an electronic signal that can be read by a computer. Further data analysis can also be performed on the computer. A simplified scheme of these components is depicted in Figure 1-1. The main components of a mass spectrometer are under vacuum. This vacuum prevents unintended collisions of the ions 1

with other molecules from occurring, which could complicate or hinder spectral analysis.2

Figure 1-1 Generalized schematic of a mass spectrometer

1.2 Mass Spectrometer Components

As previously mentioned, a mass spectrometer is composed of an ion source, mass analyzer and detector. In most cases, ionization is performed under vacuum. The subsequent separation of ions in the mass analyzer and their detection are also performed under vacuum. A variety of ion sources and mass analyzers exist. Additionally, multiple mass analyzers can be coupled together. Furthermore, mass analyzers can be used as collision cells to induce fragmentation of ions of interest. This gives rise to several different instrument configurations with a variety of analysis methods that are suitable for different classes of molecules.

2

1.2.1 Ion Sources

The purpose of the ion source is to ionize the samples that are to be analyzed.

There are two main types of ionization sources: hard-ionization and soft-ionization. Hard- ionization is very energetic, resulting in fragmentation of the analytes.2 Consequently, it is typically used with small molecules. A common hard-ionization source is electron ionization.2 On the other hand, soft-ionization is less energetic and tends to produce mainly intact molecular ions.2 Typically, soft-ionization sources are used for ionization of biomolecules and other large compounds. Common soft-ionization sources include matrix-assisted laser desorption/ionization (MALDI), electrospray ionization (ESI), atmospheric pressure chemical ionization and secondary ion mass spectrometry.2 Two related sources are laser desorption ionization (LDI) and nanoESI.

1.2.1.1 MALDI

MALDI was introduced as a technique by Karas and Hillenkamp in 1988.2

Common applications of the ionization source include the analysis of proteins, synthetic polymers, oligonucleotides and large inorganic compounds. In fact, proteins with a mass of up to 300,000 Da have been detected in the femtomole range.2 Compared to other ionization techniques, it is very tolerant of contaminants such as buffers, detergents and salts.2 In addition, MALDI tends to produce predominately singly charged ions, which make the spectra less complicated to interpret compared to other ionization sources.

In MALDI, a solid sample is introduced into the ion source. To form the solid, the analyte of interest is co-spotted with a solution containing matrix which will absorb energy from a laser. Many matrices exist; the most common of which include sinapinic acid (SA), 2,5-dihydroxybenzoic acid (DHB) and α-cyano-4-hydroxycinnamic acid 3

(CHCA). In the absence of matrix, the technique is commonly termed laser desorption ionization (LDI). The mechanism for MALDI ionization has not been fully elucidated.

Two prevailing theories exist at the moment, and it is yet unclear if one model predominates or if they contribute to ionization.5, 6 The older model is photoionization

(also termed Coupled Physical and Chemical Dynamics model), and the more recent model is the ―lucky survivor,‖ each will be described in more detail. A generalized scheme of the MALDI principle is provided in Figure 1-2. The basic steps involve irradiation of the sample with a laser, desorption of the matrix analyte and the ionization of the analyte.

Figure 1-2 Scheme of the MALDI principle adapted from de Hoffmann and Stroobant2

The Coupled Physical and Chemical Dynamics model assumes that analyte molecules are neutral in the matrix crystals, and they obtain their charge in the gaseous desorption plume by means of a charge transfer from the photoionized matrix molecules.

It is believed that the matrix crystal has a slightly lower ionization potential. In the gas

4

phase, the neutral analytes collide with either protonated or deprotonated matrix ions.

These collisions result in proton transfer reactions, forming either protonated or deprotonated analyte ions.5 The matrix ions occur from the reaction of the neutral matrix molecules with photoelectrons which will form both positive and negative ions.7 Then, an efficient charge transfer results from the higher proton affinities of basic amino acids compared to the protonated matrix molecules. Consequently, a positively charged analyte ion is favorable to be produced.

In contrast, the ―lucky survivor‖ model has two main premises. The first is that analytes have a charge in solution, which they retain when incorporated into the matrix.

The second is that these charged ions happen to survive neutralization, retaining their charge.7 It is also possible that the ablated clusters containing matrix and analyte have precharged analytes with a corresponding amount of counterions generating no net charge, and the cluster is not detected.5 After desorption, the ablated plume expands, and the matrix-analyte cluster loses neutral matrix molecules and solvent. In addition, the counterions undergo proton-transfer neutralization via interaction with analyte sites.

Consequently, many of the charges have been neutralized, except for a remaining excess charge. Thus, the charged analyte is a ―lucky survivor‖ of neutralization.7 The ―lucky survivor‖ model also works for obtaining negative ions under acidic conditions and for positively charged polyanionic compounds, such as nucleic acids.7 This model also accounts for the low intensity negative ion analytes that can be seen using acidic matrices. Their low intensity can be explained by the fact that many matrix anions have carboxylates, whose acidity is roughly the same as the analyte carboxylic acid groups.

5

Therefore, deprotonation of the analyte would not be favored.5 In addition, this model explains the process for both UV- and IR-MALDI.

1.2.1.2 ESI

ESI was introduced as a technique by Fenn.2 Unlike MALDI which produces mainly singly charged ions, ESI commonly produces multiply charged ions. However, smaller molecules (< 1000 Da) may be detected as singly charged ions. Furthermore, ESI is not as tolerant to contamination as MALDI. For ESI, up to 10-3 M of electrolytes can be in solution before the sensitivity suffers significantly.2 Common applications of ESI include the analysis of proteins, biopolymers, polymers and small polar molecules.2 For protein analysis, it has been shown that there is approximately one charge for every 1000

Da of the protein.2, 8 One of its major advantages is its ability to be coupled to separation techniques, such as high-performance liquid chromatography (HPLC). It is also important to note that the sensitivity of ESI is related to the concentration of the analytes and not the total amount injected. For instance, when coupled to HPLC if the same amount of sample is loaded on a column with a smaller diameter and flow rate, its concentration will be increased, resulting in increased sensitivity of ESI.2

In ESI, a liquid sample is introduced to the mass spectrometer for ionization. This process is comprised of three main steps. The first step is the formation of a spray of charged droplets. In the second step, these charged droplets undergo solvent evaporation forming highly charged droplets, and ions are ejected from these droplets in the third step.3 An overview of this process is depicted in Figure 1-3. Specifically, a liquid solution is infused through a stainless steel capillary (the electrospray tip). A high voltage is applied to the capillary resulting in a ―Taylor cone‖ and a mist of charged droplets.2 6

These droplets have the same polarity as the applied voltage and carry a high charge.3

After exiting the electrospray tip, the droplets pass through a pressure and potential gradient as they enter the mass analyzer region. Throughout this gradient, the charged droplets are reduced in size due to an elevated temperature in the source and a stream of nitrogen gas for desolvation resulting in evaporation of the solvent. As the droplet size

(radius) decreases, it experiences an increase in surface charge density. A critical point is reached, and the droplets are ejected into the gas phase, generating ions.2, 3 This critical point is referred to as the ―Rayleigh limit.‖ At this point Coulomb repulsion occurs to overcome the surface tension resulting from the large surface charge density.8 The final diameter of the charged particles is often around 10 nm.8

Figure 1-3 Overall schematic of the ion formation and desolvation process in ESI (adapted from Ho et al.)3

1.2.1.3 NanoESI

A primary advantage of nanoESI compared to capillary scale ESI is that it uses significantly smaller volumes of sample. For ESI, the flow rates are typically 1-100

µL/min. Additionally, nanoESI has higher sensitivity which allows samples down to the low attomol range to be analyzed.7 However, in order to use lower flow rates certain 7

modifications are employed for a nanoESI source. For example, the stainless steel capillary is typically replaced with a drawn glass capillary that has a diameter of 1-10

µm. Furthermore, the distance between the capillary and the entrance to the mass analyzer is decreased to 0.5-2 mm.8 NanoESI can also be on-line coupled to separation techniques, especially as HPLC.

Although, nanoESI appears to be a miniaturized version of ESI, it utilizes an altered mechanism which leads to important differences in its use. For instance, the droplets formed by ESI have a diameter of approximately 150 nm, whereas in nanoESI their diameter is approximately 1.5 µm. This results in one less step being performed in the reduction of the droplet size during nanoESI. By omitting this additional step, the salt contamination becomes less concentrated, effectively increasing the technique’s tolerance to the overall concentration of salt in the sample.7

1.2.2 Mass Analyzers

The purpose of a mass analyzer is to separate the generated ions according to their m/z. Mass spectrometers can have one mass analyzer or can have multiple mass analyzers coupled together. Furthermore, ion fragmentation can be performed in a mass analyzer.

In general, mass analyzers use electric and magnetic fields to separate ions based on their m/z. As with ionization sources, mass analyzers can be classified into two main categories: static or dynamic.2 In general, static mass analyzers work by transmitting ions of particular mass-to-charge ratios as a function of time. Common static mass analyzers include magnetic sectors and quadrupoles. Dynamic mass analyzers work by transmitting

8

ions simultaneously. Common dynamic mass analyzers include time of flight (TOF), ion traps, ion cyclotron resonance and orbitraps.2

Overall, mass analyzers have five main characteristics that need to be considered.

One characteristic is the mass range. Another mass analyzer property is the scan speed.

The third characteristic is the transmission of the ions. This is generally defined as the ratio of the ions reaching the detector to the ions that entered the mass spectrometer.2 The fourth characteristic is the mass accuracy. The mass accuracy is defined by the difference between the measured m/z of an analyte and its theoretical m/z. It is commonly expressed in terms of parts per million (ppm).7 The mass accuracy is also related to the last property, resolution. Resolution is defined as the ability to differentiate two ions with a small m/z difference as two distinct signals.2 A comparison of the major characteristics of the most common mass analyzers (two configurations of TOF, quadrupole, ion trap and orbitrap) is provided in Table 1.1. The resolution in the table is defined as full width at half maximum at m/z 1000 as discussed by de Hoffmann and Stroobant.2

Table 1.1 Comparison of common mass analyzers adapted from de Hoffmann and Stroobant2

Reflectron TOF TOF Quadrupole Ion Trap Orbitrap Mass limit > 1,000,000 10,000 4,000 6,000 50,000

Resolution 5,000 20,000 2,000 4,000 100,000

Accuracy 200 ppm 10 ppm 100 ppm 100 ppm < 5 ppm

Ion sampling Pulsed Pulsed Continuous Pulsed Pulsed

Pressure 10-6 torr 10-6 torr 10-5 torr 10-3 torr 10-10 torr

9

1.2.2.1 TOF

Time-of-flight (TOF) mass analyzers utilize pulsed ion sampling. This makes them ideal to be coupled with pulsed ionization sources, especially MALDI.2 In general, they consist of a flight tube that separates the packet of ions by their molecular weight.

Common types include linear and reflectron TOFs. These two main configurations will be discussed in greater detail below.

In a linear TOF, generated ions are accelerated towards the mass analyzer and gain equal kinetic energy. These ions, which have the same kinetic energy and the same charge, enter the flight tube of the TOF where they encounter a field-free region in which they are separated by their velocities.2 A general schematic of linear TOF is shown in

Figure 1-4. After detection, the m/z of the ion can be determined through the rearrangement of several basic equations. An ion can be described as having a mass of m and a total charge of q which is the number of charges, z, times the charge of an electron, e. The ion was accelerated by a potential Vs. Consequently, its kinetic energy, Ek, is given by equation 1. As all the ions are imparted with the same kinetic energy, their travel in a straight line through the flight tube is dictated by their velocity. Therefore, we can rearrange equation 1 to equation 2 to isolate the ion’s velocity as a function of its mass and charge. The flight tube is a set distance, L. The time, t, it takes an ion to traverse the length of the flight tube and reach the detector can be given by equations 3 and 4 where v is the velocity of the ion.2 Equation 4 can be rearranged to isolate the m/z in equation 5.

Consequently, the m/z of an ion is proportional to t2 (the time it takes an ion to travel through the drift tube to the detector, squared).

10

(1)

(2)

(3)

(4)

(5)

Figure 1-4 General scheme of linear time-of-flight mass analyzer. The molecular weight of the ion is indicated by the size of the circle

There are several advantages of linear TOF mass analyzers. In theory, there is no mass limit for a TOF. In fact, samples with a mass greater than 300,000 Da have been analyzed by MALDI-TOF.2 Additionally, TOFs generally have a high ion transmission efficiency, which allows them to have a high sensitivity. Previous studies have observed down to 100 attomole of protein.2

Linear TOFs also have some disadvantages. Its primary disadvantage is that compared to other mass analyzers it has lower resolution.2 One main way to overcome this issue is by employing a reflectron TOF. In this method, the ions traverse the field-

11

free region in the same manner as for a linear TOF. At the opposite end of the ion source, they encounter a reflectron which is a series of electrodes that reflect the ions back through the flight tube before detection.2 The detector can be placed in two positions, but the most common is an ―off-axis‖ detector in which the reflector deflects the ions at a slight angle away from the ion source. This allows the detector to be placed next to the ion source.2 This general layout of a reflectron TOF is illustrated in Figure 1-5. In practicality, ions with the same m/z may enter the TOF with a slight kinetic energy dispersion. The purpose of the reflectron is to correct for these differences in energy.2

Ions of the same m/z with higher kinetic energy will penetrate further into the reflectron compared to those with lower kinetic energy. Therefore, those with higher kinetic energy that travel to the reflectron faster will spend more time in the reflectron before being deflected back into the flight tube. Consequently, they will reach the detector at the same time as the ions with lower kinetic energy that traveled slower and did not penetrate as far into the reflectron.2 Overall, this process increases the mass resolution of the TOF but imposes a limitation to its mass range.

Figure 1-5 Illustration of a reflectron time-of-flight mass analyzer. The spheres represent two ions with the same mass and charge where the outlined one has slightly higher kinetic energy than the shaded in sphere.

12

1.2.2.2 Quadrupole

Quadrupoles are scanning mass analyzers. Their development was described in

1953 by Paul and Steinwegen.2 Quadrupoles consist of four perfectly parallel circular rods to which a potential and an electric field are applied.2 Specifically, all the rods have an applied RF voltage. The polarity of adjacent rods is 180 degrees out of phase, such that opposite rods has the same polarity.2 In order for an ion to be detected, its trajectory must be stable through the alternating fields of the quadrupole as depicted in Figure 1-6.

If an ion does not have a stable trajectory it will hit a rod and be discharged.2

Figure 1-6 General diagram of ion trajectories through a quadrupole mass analyzer

A limitation of quadrupoles, it that their mass range is often limited to under

4,000 m/z.2 Additionally, they are considered low resolution instruments. However, this makes them useful by serving as ion transmission guides when only a RF voltage is applied.2 Multiple quadrupoles can also be used in tandem, where one quadrupole serves as a collision cell for fragmentation of ions. Additionally, quadrupoles can be coupled to other mass analyzers, especially TOF.

13

1.2.2.3 Ion Trap

Ion traps were first described by Paul and Steinwedel in 1960.2 Generally speaking, ion traps store ions by using an oscillating electric field. Ion traps can be divided broadly into 2D or 3D ion traps, which will be described separately. Of these, the

3D ion trap was the first type of ion trap developed.2 The 2D ion trap is commonly referred to as a linear ion trap, and the 3D ion trap is also commonly referred to as a quadrupole ion trap or as a Paul ion trap. A major advantage of ion traps is that MSn can be performed to obtain multiple fragmentations of different precursor ion originating from the same molecule.

A 3D ion trap is comprised of a circular electrode and two ellipsoid caps, one on the top and the other on the bottom. This configuration is depicted in Figure 1-7 and can be thought of as a quadrupole that has been bent in on itself which creates a closed loop.2

Multiple masses of ions are stored in the trap at the same time, and they are ejected from the trap using a resonant frequency in order to produce the mass spectrum.2 In order for the ions to enter the 3D ion trap, a potential is adjusted. Specifically, to inject positively charged ions into the ion trap, a negative potential is applied to inject a pre-determined number of ions.2 The number of ions is essential because too few ions results in a loss of sensitivity, but too many ions will result in a loss of resolution. Within the confines of the ion trap, the ions will repel each other causing their trajectories to expand. To counteract this expansion, helium gas is used in the ion trap as it will collide with the ions and reduce their energy. Furthermore, a RF voltage is applied to the central ring electrode with a constant frequency and varying amplitudes. To the end caps, RF voltages are also applied with selected frequencies and amplitudes. To expel an ion, the voltage will be 14

increased resulting in an unstable trajectory for the specified ion. It is also important to note that 3D ion traps are subject to the space charge effect. This occurs when too many ions are contained in the ion trap, and those on the outside act as a shield which changes the field acting on the ions in the middle.2

Figure 1-7 Basic layout of a 3D ion trap

A 2D ion trap is comprised of a quadrupole with lenses at the ends to repel the ions within the rods of the quadrupole. Similar to a quadrupole, the ions will oscillate in the xy plane of the 2D ion trap due to an applied RF potential on the rods. The ions are trapped by the application of a DC voltage to the ends.2 To eject ions, an AC voltage is applied either between the rods and the exit lens for axial ejection or to two opposite rods in which slots have been hollowed for radial ejection. Compared to 3D ion traps, 2D ion traps have a higher trapping efficiency and 10 times greater trapping capacity which makes them less prone to space charging effects.2

15

1.2.2.4 Orbitrap

The design of orbitraps is based from a Kingdon trap which was introduced in

1923. Advances in technology have allowed many of the deficiencies of the Kingdon trap to be overcome since that time.9-11 The orbitrap mass analyzer was introduced commercially in 2005.2, 10 A smaller design was released in 2011 known as a high-field compact trap version.10 In most commercial instruments, the orbitrap is part of a hybrid mass spectrometer where it is coupled to at least one other mass analyzer. For instance, in

Thermo’s LTQ-Orbitrap it is coupled to a linear trap quadrupole.9 Two of the major benefits of the orbitrap are its high resolving power and its mass accuracy, which have made it a powerful mass analyzer in the field of bottom-up proteomics as it drastically reduces false positive peptide identifications.9 Specifically, Thermo’s LTQ-Orbitrap has a resolving power greater than 150,00 and a mass accuracy of 2-5 ppm.9

An orbitrap consists of three electrodes. Two outer electrodes are cup-shaped and face each other creating a barrel-like shape.9, 10 However, they are electrically isolated by a hair-thin gap. The third electrode is a spindle-like central electrode. For analysis, a voltage is applied between the outer and central electrodes. In order to have the necessary long mean free path (many kilometers), the orbitrap utilizes an ultrahigh vacuum.10

Additionally, the central electrode serves as a shield for the ions increasing the charge capacity of the mass analyzer.10

In many commercial instruments, ions are collected in a C-trap before injection into the orbitrap. The C-trap is used as an external storage device for the Orbitrap mass analyzer that accumulates ions before injection into the Orbitrap making it compatible

16

with a continuous ESI source.11 The C-trap is a curved quadrupole ion trap that only utilizes RF. 9 To inject the ion packet, the RF is decreased on the C-trap followed by DC pulses which direct the ions out of the C-trap through a slot for radial ejection. As the ions are injected into the orbitrap, they experience ―electrodynamic squeezing‖ which results in contraction of the ion cloud and its attraction to the z-axis (along the central electrode). Furthermore, larger m/z ions are injected into the orbitrap later than smaller m/z ions. These larger m/z ions experience a larger mean orbital radius and amplitude of axial oscillation due to electrodynamic squeezing.9 Once ions are injected, the central electrode is held at a constant voltage. This results in a linear electric field along the axis of the central electrode with purely harmonic oscillations.10 In order to have a stable trajectory in the orbitrap, ions will oscillate around the central electrode along its z-axis while orbiting around the central electrode. The m/z ratio of the ion and the electrode potential determines the frequency of the ion oscillations.9 Ions with the same m/z will oscillate together in a thin ring. For detection, the outer electrodes are used to detect the image current of the ions which are Fourier-transformed to produce the mass spectrum.

1.3 Biomolecule Analysis by Mass Spectrometry

Many different classes of molecules are analyzed by mass spectrometry. Common applications include the analysis of biomolecules (e.g., proteins, peptides, carbohydrates, lipids, nucleic acids and metabolites). As MALDI and ESI are soft-ionization sources, they are amenable to biomolecule analysis. Different applications exist to study biomolecules. This section will focus on the primary techniques to analyze proteins and peptides.

17

1.3.1 Proteins and Peptides

Mass spectrometry is commonly used to study proteins and peptides. One common application is the use of mass spectrometry for molecular mass determination.2

Additionally, protein and peptide sequences can be derived from fragmentation. It is also useful for the identification, localization and quantitation of post-translational modifications.2 Proteins and peptides are commonly identified by two different techniques, either peptide mass fingerprinting or tandem mass spectrometry (MS/MS).

Both of these methods will be described in the following sections.

1.3.1.1 Peptide Mass Fingerprinting

Proteins separated by gel electrophoresis (e.g., SDS-PAGE or 2D-gel electrophoresis) are generally subjected to peptide mass fingerprinting (PMF) for protein identification. An overview of the workflow is provided in Figure 1-8. In this technique, protein bands of interest are isolated and enzymatically digested. The most common enzyme used is trypsin. The resulting peptides are analyzed by MALDI-MS.

Alternatively, the enzymatic digest can be performed in solution for purified proteins.12

After obtaining the mass spectrum, the obtained masses of the tryptic peptides are searched against an in silico digest of all the proteins in a given database.13 Different programs utilize slightly different parameters, but generally speaking the closer the experimental data match the in silico digest of a protein, the more likely a positive match or identification is made.

18

Figure 1-8 Overview of peptide mass fingerprinting for protein identification

One of the most common programs to perform PMF is Mascot, which is a probability-based scoring search. This probability-based scoring is used to calculate the likelihood that the observed match of the experimental data is a chance event. The common threshold of p < 0.05 is utilized to determine a positive identification. However, for a large database the probability of a match becomes a very small number.

Consequently, the probability is converted to a score, where the score is defined as -

14 10Log10 (probability).

When performing peptide mass fingerprinting, several parameters must be determined by the user. The most important parameters of PMF, in descending order, are the number of peptides matched, the mass error threshold, the percent coverage of the sequence and the number of missed cleavage sites.15 The parameter of mass accuracy is extremely important. Specifically, while a large mass error window will increase the

19

number of peptides matched, more random identifications will occur. Conversely, if the mass error window is too small, the user risks missing legitimate matches.14 The user must also determine how many missed cleavages were likely to occur during the digestion process for the peptides. The two main reasons that missed cleavages arise is that either too short a digestion time was employed or that the ratio of enzyme to substrate was too low.14 Another factor for PMF is that the digest may contain modified peptides which must be accounted for in the database search. However, as you increase the number of possible modifications in a search, the number of possible random matches increase, increasing the number of false positives.13 These modifications can either be quantitative (global or fixed) or non-quantitative (variable). For a quantitative modification, one would expect to always see that modification on the given amino acid residue. An example is the chemical derivation via methylation of cysteine. For a non- quantitative modification, one would expect to occasionally see that modification. An example is the in vivo phosphorylation of serine.14 In this case some, but not all, the serines in a given protein are phosphorylated. However, it is important to note that

Mascot does not use the known post-translational modifications present in protein databases.

Several other intrinsic factors may influence the results of PMF. For instance, a difference in ionization of the peptides can influence the results of PMT. Specifically, tryptic peptides containing arginine tend to be more intense than peptides containing lysine.12, 13 Furthermore, PMF is often challenging for proteins with a molecular weight less than 15 kDa as they have a low number of peptides produced from enzymatic digestion.16 The presence of multiple proteins in a single sample can also hinder 20

identification by PMF.16 However, these factors cannot be accounted for in the search parameters.

1.3.1.2 Tandem Mass Spectrometry

Shotgun proteomics is another common method utilized for performing protein identification. For shotgun proteomics, a sample is subjected to enzymatic digestion, again typically using trypsin, and the resulting peptides are separated by LC (or multidimensional LC). The separated peptides are then subjected to tandem mass spectrometry (MS/MS), and the spectra are subjected to automated database searching. 16,

17 Compared to PMF, shotgun proteomics typically has a higher throughput and better sensitivity.17 Fragmentation of the peptides produces different possible fragments depending on which bond is cleaved. A standard nomenclature proposed by Biemann is used to distinguish these fragments.2 There are six common fragmentation sites for peptides. Overall, the charge of the fragment is either retained on the N-terminus or the

C-terminus of the peptide. If the charge is retained on the N-terminus fragment, a, b and c ions are observed. Conversely if the charge is retained on the C-terminus fragment, x, y and z ions are observed. The peptide fragmentation sites and their nomenclature are provided for a pentapeptide in Figure 1-9. For example, b and y ions are formed if the bond between the carboxyl group and the amine is cleaved. These fragments are also numbered from the respective ends to indicate the position along the peptide backbone.

For low-energy collision-induced dissociation (CID), b and y ions are the most commonly observed peptide fragment ions, but higher energy fragmentation can produce a, c, x, and z ions.

21

Figure 1-9 Biemann nomenclature for distinguishing common peptide fragmentation sites on a generic pentapeptide (adapted from Hjernø and Jensen)13

Many of the same parameters used for PMF searching of a database are also used for searching MS/MS spectra. Instead of a search against tryptic peptide mass lists, the search engine compares the fragmentation of peptides. The most common search engines are SEQUEST and MASCOT. SEQUEST uses a cross correlation score, Xcorr, for scoring purposes. First, low-intensity peaks are removed from the experimental spectrum, and the peak intensities are normalized. This processed spectrum in compared to a theoretical spectrum to see how many peaks match.17 In its simplest sense, the Xcorr counts the number of matching fragment ions between the experimental spectrum and the theoretical spectrum. SEQUEST also uses a derivative score, ΔCn, which is the difference between the best Xcorr and the second best Xcorr. In general, the higher the Xcorr and the ΔCn, the more likely the match is correctly assigned. However, these parameters are not independent of the peptide length.17

In contrast, Mascot uses the probability that the number of matched peaks would be due to chance alone to determine the score. As it uses probabilities, Mascot has a score

22

threshold, while SEQUEST does not.18 In both Mascot and SEQUEST, the user must specify the expected ion series based on the type of instrument and fragmentation used.

1.3.1.2.1 MS/MS using Bruker’s MALDI-TOF/TOF

Bruker’s MALDI UltrafleXtreme is one of the primary instruments used in these studies. It has the capability of performing laser-induced dissociation (LID) and CID for fragmentation. For protein and peptide analysis, LID is commonly used for protein identification, whereas CID is used for de novo sequencing due to its higher energy.16

Specifically, this instrument uses LIFT-TOF-TOF for MS/MS analyses (Figure 1-9). For

LID, the laser fluence is increased to obtain more precursor ions per shot which are accelerated at a low voltage of 8 kV. During their long flight time the precursor ions can undergo fragmentation, generating an ―ion family‖ with the same velocity as the precursor ion. Different precursor ions will generate different ion families that will reach the timed ion selector at different times. The gate voltage of the timed ion selector will be turned off to allow the selected ion family to pass through, while it will be turned on to deflect the other ―ion families.‖ At this point, the ―ion family‖ passes into the ―LIFT‖ device which consists of three stages. In the first stage, the grid potential is increased from ground to 19 kV, and the ions pass into the second stage where they are held at 19 kV to ensure all the ions are traveling with the same speed. The grid potential at the end of the second stage is reduced by 2-3 kV to accelerate the ions into the third stage. In the third stage, the ions are accelerated and enter the second TOF for detection allowing both the selected precursor and the fragment ions to be analyzed in a single spectrum.16 The ions are also focused using the reflector before reaching the detector. Between the LIFT device and the reflector is a ―post lift metastable suppressor‖ that deflects any remaining 23

precursor ions that could produce undesirable fragments after acceleration into the second

TOF.

Figure 1-10 General schematic of MS/MS using the LIFT on the Bruker MALDI- TOF/TOF (adapted from Suckau et al.)16

1.3.1.2.2 MS/MS using Thermo Orbitrap Fusion

Thermo’s ESI-Orbitrap Fusion Tribrid mass spectrometer was used for high- throughput proteomics. As a tribrid mass spectrometer it combines three mass analyzers, a quadrupole mass filter with an Orbitrap and a linear ion trap. It also has a maximum resolving power of 500,000 at m/z 200 making it a high resolution mass spectrometer. As it has three separate mass analyzers, analyses can be fully parallelized to allow simultaneous detection in both the Orbitrap and the linear ion trap.11 It is also capable of several different modes of fragmentation, namely collision-induced dissociation (CID), high-energy collisional dissociation (HCD) and electron transfer dissociation (ETD).11

24

CID is commonly used in instruments with ion trap mass analyzers. Overall, it is a low energy process that activates and deactivates ions at a rate of 1 to 100 Hz (activation time ~30 ms).19 While CID is commonly used for peptide characterization, there are a few drawbacks to the method. For instance, CID in ion traps suffers from the one-third effect in which fragment ions with a mass of less than one-third of the parent ion are lost.19 Furthermore, while CID is useful for peptide identification, it has limited utility for analyzing post-translational modifications, particularly phosphorylation, as the modifications are typically labile and are lost by CID. Additionally, for large, highly charged peptides it provides limited sequence information.11, 20

HCD occurs in the octapole collision cell of Orbitrap instruments. In comparison to CID, it utilizes higher energy and shorter times to activate the ions (0.1 ms).19 Unlike

CID, HCD does not suffer from the one-third effect. However, HCD fragmentation spectra tend to be more complex than CID spectra as 53% of the total ion intensity is from a, b and y ions. In comparison, the same ion families account for 72% of the total ion intensity of CID fragmentation spectra.19 As it does not suffer from the one-third effect, HCD also produces more fragments in the lower mass range.11 For HCD fragmentation, a quadrupole filled with gas is used. This quadrupole is placed between the C-trap and the linear ion trap to diminish the amount of gas that would enter the linear ion trap. In order to fragment ions, the DC offset on the rod electrodes is adjusted.11

ETD was first developed in 2004 to overcome several of CID’s limitations.11, 20 In addition, it can easily be used with online chromatographic separations as it has a short duty cycle, requiring tens of milliseconds for reaction times.20 Both CID and ETD result

25

in cleavage of the peptide backbone. However, unlike CID, ETD leaves post-translational modifications intact which is especially important for labile modifications such as phosphorylation and glycosylation. 11 ETD generates fragmentation by a nonergodic pathway (i.e., it does not involve the redistribution of intramolecular vibrational energy).21 For ETD fragmentation, the Orbitrap Fusion ionizes the reagent molecules

(fluoranthene) using a Townsend discharge whereas traditional methods use a filament- based ionization.11 Radical anions of fluoranthene are generated, and they react with multiply charged peptide cations by transferring an electron as shown in Figure 1-11.20

From odd-electron components, c and z-type of fragment ions are generated. From even- electron components, b- and y-type ions are produced.21 Specifically, this results in a highly selective fragmentation of the N-Cα bonds of the peptide backbone generating c and z-radical ions. However, charge reduction of the multiply charged peptides may also occur. During charge reduction, the fragments are unable to dissociate due to multiple hydrogen bonds or salt bridge interactions.20

26

Figure 1-11 Overall reaction of fluoranthene radical anions with a peptide for fragmentation (adapted from Udeshi et al.)20

27

Chapter 2

Sample Preparation and Separation Techniques Used for Proteomic Analyses

2.1 Sample Preparation for Mass Spectrometric Analyses

Proper sample preparation is an essential step in the proteomics workflow. A variety of techniques exists to perform sample preparation. A primary category includes sample cleanup steps, such as desalting, to minimize or eliminate unwanted molecules from the analytes of interest. For instance, many salts and buffers are nonvolatile resulting in noisy mass spectra and contaminated mass spectrometer sources. Sample preparation may also include the separation of complex mixtures. This enables only a few proteins or peptides to be analyzed at one time, allowing for efficient isolation of each analyte for characterization. Furthermore, differences in ionization efficiencies may result in ion suppression effects that may make it difficult to detect one analyte in the presence of another. Overall, these purification and separation steps are essential to generate the high-quality mass spectra necessary for protein identification. The main methods used in the conducted proteomic studies will be discussed in further detail below.

28

2.2 Desalting of Proteins and Peptides

Desalting of proteins and peptides is commonly performed using solid-phase extraction (SPE). The basic principle of SPE involves the retention of analytes on a solid phase to isolate them from the rest of the sample prior to their elution using the necessary solvent.22 Several commercial devices are available including cartridges, ZipTip pipette tips (Millipore) and StageTips (Thermo). While commercially available, StageTips can also be constructed from materials in the laboratory.23 Their primary benefit is that they can clean and concentrate peptides in a single step. This is achieved by binding the peptides to a small amount of reversed-phase material (e.g., C18) packed in a pipette tip and eluting them using a small volume of organic solution.23 Consequently, they work in a similar manner to reversed-phase HPLC. In both methods, the analytes partition between the non-polar stationary phase and the mobile phase. As the stationary phase media is non-polar, non-polar (or hydrophobic) peptides will have a greater affinity (i.e., retention to the column or media. Polar analytes, such as salts, will have a smaller affinity to the column and are easily washed off the column with water. To elute the peptides, a more polar solvent (such as acetonitrile) is typically passed over the column.

These commercial tips have several advantages. One of the advantages of using

ZipTips compared to traditional methods is that they allow a bi-directional flow over their 15 µm particles enabling fast loading, washing and elution.24 They also allow quick adsorption kinetics.24 Another advantage is that small volumes are loaded (2-4 µL).

Furthermore, as there is essentially no dead volume, small volumes can be eluted from the ZipTips.24 Therefore, ZipTips and similar media allow for the quick purification and concentration of small volumes of sample that may be obtained from biological studies. 29

ZipTips and StageTips follow the same basic principles as they contain the same type of media in the pipette tip format. Both commercial tips follow a similar protocol that involves wetting the dry media followed by an equilibration step. The wetting and equilibration steps are important as the solid-phase material is C18 silica. Also, the tips should not be allowed to dry prior to loading the sample.25 Next, the sample is loaded and washed in slightly acidic solvent prior to elution using organic solvent. However, the exact composition of the solvents and volumes used for each brand of tips differs slightly in their standard protocols. These differences are summarized in Table 2.1. Overall, the

ZipTips use a combination of acetonitrile (ACN), water and trifluoroacetic acid (TFA).

Whereas, the StageTips use a combination of ACN, water and formic acid (FA).

However, these are the recommended protocols and modifications can be made to the protocols.

Table 2.1 Comparison of solvents and volumes used for ZipTip and StageTip desalting of peptides

ZipTips StageTips

Step Solvent Volume Solvent Volume

Wetting 100% ACN 10 µL 80% ACN, 5% FA 20 µL

Equilibration 0.1% TFA in dI H2O 10 µL 5% FA 20 µL

Bind sample 0.1% TFA in dI H2O 10 µL 5% FA 20 µL

Wash 0.1% TFA in dI H2O 10 µL 5% FA 20 µL Elute A 50% ACN in 0.1% TFA 1-4 µL 80% ACN, 5% FA 20 µL

Elute B 75% ACN in 0.1% TFA 1-4 µL n/a n/a n/a indicates step not used

30

2.3 Gel Based Separations

Gel electrophoresis is commonly used to separate proteins based on their size and charge. In general, an electric field is applied causing the charged proteins to move toward the oppositely charged electrode. Gel electrophoresis can be split into different classes. For instance, proteins can either be separated in their native states or under denaturing conditions. One-dimensional denaturing gel electrophoresis is commonly referred to as sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE).

In addition, one-dimensional or two-dimensional separations can be used. For two- dimensional separations proteins are separated by their isoelectric point (pI) prior to molecular weight separation. The different methods will be discussed in further detail below.

2.3.1 SDS-PAGE

For SDS-PAGE, proteins are denatured prior to separation using a polyacrylamide gel. In order to denature the proteins, SDS is added to the protein sample, and the mixture is heated. The SDS binds to the proteins in a ratio of one molecule of SDS to two amino acids, and it will also give the proteins a net negative charge.26 Reducing agents, such as

2-mercaptoethanol, are added to the protein solution to further denature the proteins by cleaving any disulfide bonds.26, 27 Overall, the proteins will migrate in the same direction on the polyacrylamide gel towards the anode which has a positive charge to attract the negatively charged proteins. In general, smaller proteins will travel through the gel matrix faster. Additionally, the molecular weight of a protein can be estimated by its migration in the gel as its migration rate is inversely proportional to the logarithm of its weight.28

31

However, some proteins will run differently than expected including very basic proteins, very acidic proteins, glycoproteins and lipoproteins.28

The SDS-PAGE gels are formed by polymerizing acrylamide monomers with a bisacrylamide crosslinker.28 As a safety precaution, care should be taken while working with the acrylamide monomer as it is classified as a neurotoxin.28 Ammonium persulfate is used to initiate the polymerization by a free radical mechanism, and the addition of

N,N,N,N’-tetramethylenediamine (TEMED) accelerates the reaction.26, 28 Different concentrations of acrylamide will determine the pore size of the gel which influences the separation of the proteins. Specifically, gels with lower concentrations of acrylamide

(e.g., 8%) will have larger pore sizes that allow large proteins to migrate faster, resulting in better resolution of high molecular weight proteins, whereas gels with higher concentrations of acrylamide (e.g., 15 or 18%) will have smaller pore sizes that will hinder the migration of large proteins. This gives better separation to the lower molecular weight proteins.26, 27 The common method for casting gels follows the Laemmli method which is a discontinuous system using two distinct gels (e.g., a resolving and stacking gel) in a continuous manner.28 A stacking gel is often used which contains a lower acrylamide concentration and lower pH than the main resolving gel. This allows the proteins to concentrate into a narrow band at interface between the stacking and resolving gels, resulting in better resolution of the proteins.26, 28

Different protein stains exist to visualize the separated proteins. Common stains used are Coomassie blue, silver staining and fluorescent stains. Additional specific stains exist, such as those that will visualize only phosphoproteins. Of these, Coomassie blue is

32

one of the most popular choices. Its primary mechanism of staining is through interaction with lysine, tyrosine and arginine residues in the proteins.29 It is capable of easily detecting approximately 1 μg of protein in a band.28 Silver staining is another common method for visualizing proteins. An advantage of silver staining is that it is up to 100 times more sensitive than traditional Coomassie blue staining.28, 29 Thus, 10-100 ng of protein can be detected.28 Furthermore, fluorescent total protein stains are an alternative to silver staining as they are also more sensitive than Coomassie.30 Of the various fluorescent stains, SYPRO Ruby protein stain is a common choice. It exhibits a larger dynamic range than silver staining. Specifically, silver staining has been shown to have a linear dynamic range from 8-60 ng, whereas SYPRO Ruby has a linear dynamic range from 1-1000 ng.30

2.3.2 2D-PAGE

Another common separation technique used in proteomics is 2D-PAGE. In the first dimension, proteins are separated by their isoelectric point (pI) during isoelectric focusing. In the second dimension, proteins are separated by their molecular weight.31

These two orthogonal separations are capable of resolving more proteins than the one- dimensional SDS-PAGE separations. It is capable of separating hundreds to thousands of proteins, depending on the pI range for isoelectric focusing, the size of the second- dimension gel and staining choice.32 Instead of bands, proteins are typically resolved into spots. One major drawback to 2D-PAGE is that it is difficult to compare spots from different gels.33

33

Overall, the procedure can be divided into five major steps: sample preparation, isoelectric focusing, interface to the second-dimension separation, SDS electrophoresis and protein detection.34 For sample preparation, proteins must be solubilized without altering the protein’s pI. Isoelectric focusing is performed using immobilized pH gradients (IPG strips).34 A high field strength is used to focus proteins after a lower field strength step which allows salts and carrier ampholytes to reach their position without causing a significant amount of heating of the IPG strip.34 Once the proteins have undergone isoelectric focusing, an equilibration step is necessary prior to SDS electrophoresis. In this step, SDS is introduced for protein mobility in the second- dimension gel. Proteins are then separated in the same manner as in SDS-PAGE. Also in a similar accord, separated proteins can be visualized using Coomassie blue, silver staining or fluorescence.

2.3.3 Difference Gel Electrophoresis

Difference gel electrophoresis (DIGE) is commonly used for comparing multiple samples by 2D-PAGE using a single gel, overcoming reproducibility issues. This allows for quantification of proteins present in various states. In this method, the lysine residues of proteins are labeled with one of three fluorescent dyes prior to electrophoresis.31 The three dyes used are Cyanine-2, Cyanine-3 and Cyanine-5. A pooled internal standard is typically labeled with the Cyanine-2 dye.35 These dyes are spectrally distinct, compatible for mass spectrometry, have of limit of detection around 100-200 pg of protein and exhibit a linear range of at least five orders of magnitude.31 Additionally, they have similar masses and have a positive charge which prevents any change in the overall charge of the proteins as it replaces the positive charge on lysine.36 34

The overall workflow for DIGE is very similar to 2D-PAGE with the addition of a labeling step. For instance, extracted proteins must first be denatured and solubilized.

However, the buffer used must be the proper pH (pH 8.5) for labeling with the cyanine dyes. Afterwards, minimal labeling is performed in which 1-2% of the lysines are labeled with the N-hydroxy succinimidyl ester cyanine dyes. This prevents issues with increased hydrophobicity of the proteins after labeling, leading to possible precipitation issues.36 It also prevents multiple dye additions to a single protein that would result in multiple detected spots.35 For minimal labeling, a ratio of 100-300 pmol of CyDye to 50 µg of sample is typically used.32 Labeling of the lysine occurs via a nucleophilic substitution reaction in which the N-hydroxy succinimidyl ester group reacts with the epsilon amino group of lysine.35, 36 The reaction can be quenched by adding in a solution of lysine. After labeling, the samples are mixed together prior to isoelectric focusing of the proteins.

Once the proteins have been fully separated, the proteins are visualized and spot matching from superimposing the images of each CyDye is performed prior to statistical analysis.

2.4 HPLC Separations

Proteins and peptides can also be separated using reversed-phase HPLC. In general, reversed-phase HPLC is marked by using a nonpolar stationary phase with a polar mobile phase.37 For the stationary phase, the column is comprised of a silica gel.38

This silica gel contains alkyl chain ligands, typically C4, C8 or C18 alkyl chains. The mobile phase usually contains water and an organic solvent that is miscible with water.

Common organic solvents used include acetonitrile, methanol and isopropanol.38 Acids, such as formic acid or trifluoroacetic acid, may also be added to the mobile phase. This 35

addition has two main consequences. First, it minimizes undesired interactions of the proteins or peptides with the stationary phase. Second, it gives the analytes a net positive charge.38 This is important as HPLC can also be coupled with ESI-MS with on-line coupling or MALDI-MS with off-line coupling for analyte identification.39 Additionally,

HPLC is a common separations technique due to its high resolving power, its reproducibility and its high recovery of the analytes.40

A mixture of proteins or peptides is applied to the column using an aqueous solvent. Overall, their retention depends on their hydrophobicity.40 Specifically, polar

(i.e., hydrophilic) proteins or peptides are the least retained on the column, and nonpolar

(i.e., hydrophobic) analytes are eluted later. To elute the retained proteins and peptides, the concentration of the organic mobile phase is increased.38 This increase in organic solvent can be performed in either an isocratic manner or using a gradient. In an isocratic separation, the concentration of the organic mobile phase is held constant throughout the run, whereas a gradient separation uses increasing concentrations of the organic mobile phase.40 For optimal separations, linear gradients are typically used instead of isocratic elutions.

For the analysis of peptides, nanoHPLC can also be used. This method differs from traditional HPLC in that its flow rate is in the nanoliter per minute range.39 It has several advantages including that is has a small dead volume, reduced analyte dilution due to the low flow rate and a high recovery of low abundant peptides.39 This leads to higher sensitivity than traditional HPLC. It also leads to the detection of low amounts

(e.g., femtomoles) of peptides.41 In order to minimize the dead volumes of the system,

36

capillary tubing with inner diameters of 10-50 µm are typically used instead of the traditional HPLC tubing.41 As the tubing is smaller than traditional HPLC, the columns used for peptide separation in nanoHPLC are also smaller. The inner diameter of the column is typically 50-100 µm, and the column length is 50-150 mm.41

2.5 Fe-NTA columns

While gel electrophoresis and HPLC are common approaches to globally separate proteins and peptides, additional separation techniques exist. Several of these techniques are used for specific applications to separate a specific class of proteins or peptides. For instance, characterization of phosphopeptides can be challenging for two main reasons.

First, they are present in relatively low amounts. Second, they typically exhibit lower ionization efficiency than non-phosphorylated peptides.42 Both of these contribute to difficulty in analysis, necessitating enrichment of the phosphopeptides for their characterization. One common method for phosphopeptide enrichment is the use of immobilized metal affinity chromatography, commonly using Fe-NTA columns.42

Fe-NTA columns use a stationary phase in which Fe3+ metal ions are chelated to beads coated with nitrilotriacetate (NTA). This stationary phase will selectively bind the negatively charged phosphate group of the phosphopeptides.42, 43 The overall procedure involves sample application to the Fe-NTA column followed by a rinsing step to remove unbound peptides and an elution step (Figure 2-1). The final solution contains the phosphopeptides in a concentrated amount.43 However, it has been reported that the technique may exhibit a preference for enrichment of multiply phosphorylated peptides than monophosphorylated peptides due to their stronger interaction with the resin.43

37

Figure 2-1 Simplified phosphopeptide enrichment workflow using Fe-NTA columns (adapted from Dunn et al.)43

Certain considerations should be taken into account. For instance, the Fe3+ metal ions are not covalently bound to the stationary phase. Consequently, they could leach from the column resulting in a loss of enriched phosphopeptides.43 Additionally, nonspecific binding can occur, particularly from acidic residues (i.e., glutamic acid and aspartic acid). This can be mitigated by two main approaches. One method is by acidifying the sample prior to enrichment. This results in protonation of the carboxyl groups of the amino acids, which hinders their interaction with the stationary phase.42 It has also been noted that by increasing the ionic strength of either the loading solution or the rinsing solution, electrostatic interactions between these residues and the metal ion are minimized.43

38

Chapter 3

Mass Spectrometric Investigation of Potential Salivary

Protein Biomarkers of Stress Induced by the Cold

Pressor Test

3.1 Introduction

The use of biomarkers has become common for the diagnosis of diseases and psychological states or to monitor the progression and treatment of a disease. Several body fluids can be used for investigation of biomarkers including plasma, urine, cerebrospinal fluid, saliva, bronchoalveolar lavage fluid, synovial fluid and amniotic fluid.44 This study aims to use saliva to determine potential salivary biomarkers of stress using the cold pressor test (CPT) as a model system to reproducibly induce stress in participants.

3.1.1 Diagnostic utility of saliva and methods of collection

Saliva has gained popularity for biomarker elucidation and monitoring as it has several advantageous attributes compared to other body fluids, particularly blood and urine. For instance, collection of saliva is considered to be easy and noninvasive, facilitating multiple collections.45, 46 Additionally, saliva does not have the same

39

collection risks as blood for contracting infectious agents.47, 48 The ease of collection and reduced risk means that it does not necessitate the need of health care workers for collection.49 Compared to obtaining blood samples, saliva collection is associated with reduced anxiety as it is a painless procedure.46, 50 Additionally, sample collection is straightforward and sufficient quantities can be obtained.51 Specifically, the average person secretes 1-2 L of saliva each day.46, 48 Compared to serum and urine, saliva has lower costs of storage and shipping.44, 51 Saliva, in comparison to urine, has a lower salt concentration.50 Furthermore, saliva is capable of reflecting real-time biomarker levels.52

Whole saliva is secreted by three major glands and various minor salivary glands.

The major glands include the parotid, submandibular and sublingual glands, and they are responsible for approximately 90% of fluid production.51, 53 The minor salivary glands are primarily located on the tongue, palate, buccal mucosa and labial mucosa.46 Overall, whole saliva is composed mainly of water with electrolytes, enzymes, hormones, lipids, mucins, proteins and small organic compounds.46, 54, 55 Moreover, many of the biomolecules found in blood and urine are also found in saliva. In fact, 50% of salivary proteins can also be detected in serum.56 However, their levels are one-tenth to one- thousandth of those in blood.48 The primary purpose of saliva is to lubricate oral tissue for processes including eating and speaking, to protect teeth and mucosal surfaces and to aide in digestion.51, 57

Saliva can be collected as either whole saliva or from an individual gland in either a resting (nonstimulated) or stimulated state. Whole saliva is deemed simpler than glandular saliva to obtain as minimal equipment is required.45 However, it is influenced

40

to a larger extent by the oral environment.45 In comparison, collection from an individual gland is controlled and provides more specific information for salivary gland diseases.45

Collection of nonstimulated whole saliva is typically performed by draining, aspiration or using an absorbent material.58 Typical flow rates range from 0.1 to 0.5 mL per minute.49,

54 Several absorbent materials are fabricated of cotton and range from simple dental rolls to specialized devices which require centrifugation to gather the saliva. For instance, the

Salivette is a commercial device consisting of a cotton roll that is sucked or chewed on.52

Saliva production can be stimulated using sour candies, lemon juice or citric acid.

Alternatively, stimulation by chewing paraffin wax is also used.46 Stimulation is oftentimes beneficial in order to study conditions that have a low salivary flow.45

As of a report in 2012, more than 3000 salivary proteins have been identified.50

Several major classes of proteins can be readily detected in saliva. The most abundant salivary proteins include amylase, carbonic anhydrase, cystatins, , mucins and proline-rich proteins.50, 59 In fact, 50% of the total salivary proteins are derived of amylase, immunoglobulins, mucins and proline-rich proteins.46 Although alpha-amylase is the most abundant protein in whole saliva, it only accounts for 20% of the salivary proteome by weight.55 The class of proline-rich proteins can be further divided into acidic, basic or glycosylated proline-rich proteins, and as a class, they account for 60% of the saliva proteome by weight.53, 55 To complicate analysis, many post-translational modifications of the proteins are present, especially glycosylation and phosphorylation.57

Furthermore, multiple variables may affect the composition of saliva. Its secretion is controlled by the autonomic nervous system.46, 54 Specifically, secretion is modulated

41

by both the sympathetic and parasympathetic nervous system. Activation of the sympathetic nervous system results in smaller volumes of saliva produced with high protein concentrations. Conversely, activation of the parasympathetic nervous system results in large volumes of saliva secreted with low protein concentrations.46 In addition, the composition and flow can be altered by age, gender, food ingestion, health status, medication, circadian rhythms and physical exercise.46, 51 For example, while the average salivary protein concentration is 0.5-2 mg/mL, the concentration is higher in the afternoon than in the morning.48, 53, 54 These rhythms can also be observed for specific proteins, such as alpha-amylase which has a maximum secretion in the late afternoon.52

The flow rate also exhibits a circadian rhythm in which the flow rate of whole saliva has a minimum in the morning around 3-4 am and has a maximum in late afternoon around 3-

4 pm.49, 60

3.1.2 Acute stress and the cold pressor test

Acute stress results in the activation of the sympathetic nervous system.61

Multiple laboratory stressors exist to induce acute stress. One of the common is the cold pressor test (CPT). It has been used since the 1930s to study pain.62 The cold pressor test is considered a robust, reliable and noninvasive autonomic challenge in which cold is applied to tissue. Typically, the non-dominant hand is immersed into cold water (0-

10 °C) and is removed when the sensation becomes intolerable or when a pre-determined time ceiling has been reached.62, 63 This serves to activate dermal thermoreceptors resulting in an increase in the sympathetic nervous system in the majority of the general population.64 65 Common indicators of an increased sympathetic system include vasoconstriction, increased heart rate, increased blood pressure and increased respiration 42

frequency.63, 64, 66 Effects of the CPT are time dependent. It is well tolerated during the initial 30 seconds, but a large increase in blood pressure can be observed after two minutes.61 Use of the CPT also results in increased skin conductance.66 Moderate increases in cortisol, which is indicative of activation of the hypothalamus-pituitary- adrenal axis, are also observed.66 An increase in salivary alpha-amylase in response to the cold pressor test has also been observed.67 While increases in salivary alpha-amylase have been correlated with stress, it lacks specificity as increases have been observed in other disease states (e.g., gingivitis, Sjögren’s syndrome) necessitating the discovery of novel, objective salivary protein biomarkers of acute stress.50, 68

3.2 Experimental

3.2.1 Chemicals and Materials

A water purification unit from Thermo was used to obtain DI water. SDS-PAGE gels were cast using 30% acrylamide/bis-acrylamide (29:1) solution, 1.5 M Tris-HCl (pH

8.8), 0.5 M Tris-HCl (pH 6.8), 10% (w/v) sodium dodecyl sulfate solution and TEMED that were purchased from Bio-Rad (Hercules, CA). Bio-Safe Coomassie stain, 2x

Laemmli sample buffer (65.8 mM Tris-HCl (pH 6.8); 26.3% (w/v) glycerol; 2.1% SDS and 0.01% bromophenol blue), ReadyStrip (7 cm, pH 3-10) IPG strips, mineral oil and

ReadyPrep Overlay Agarose (0.5% low melting point agarose in 1x Tris/glycine/SDS and

0.003% bromophenol blue) were also acquired from Bio-Rad. Additional materials from

Bio-Rad include ReadyPrep 2-D Starter Kit Rehydration/Sample Buffer containing 8 M urea, 2% CHAPS, 50 mM DTT, 0.2% Bio-Lyte 3/10 ampholyte and 0.001% bromophenol blue. For 2D gel electrophoresis, Bio-Rad’s 12% Mini-PROTEAN TGX

Precast gels were used. 43

Ammonium peroxydisulfate, HPLC grade methanol, HPLC grade acetonitrile, acetic acid, glycerol, dimethyl sulfoxide (DMSO), urea, sodium acetate and trifluoroacetic acid (TFA) were obtained from Fisher Scientific (Pittsburgh, PA).

Ammonium bicarbonate, β-mercaptoethanol, trypsin, formic acid (FA), dithiothreitol

(DTT) and iodoacetamide (IAM) were purchased from Sigma (St. Louis, MO). SYPRO

Ruby Protein gel stain, Pro-Q Diamond Phosphoprotein gel stain and PeppermintStick

Phosphoprotein molecular weight standards were obtained from Molecular Probes, Inc.

(Eugene, OR). Cyanine3 NHS ester (Cy3) and Cyanine5 NHS ester (Cy5) whose structures are shown in Figure 3-1 were purchased from Lumiprobe (Hallandale Beach,

Florida). CHCA matrix was obtained from Bruker Daltonics (Bremen, Germany). A commercial Fe-NTA phosphopeptide enrichment kit was purchased from Pierce

(Rockford, IL). Bovine casein and sodium dodecyl sulfate from JT Baker were acquired.

Ethanol was obtained from Pharmco-AAPER.

Figure 3-1 Chemical structures of (A) Cyanine3 NHS ester and (B) Cyanine5 NHS ester

For MALDI-MS, Bruker’s MALDI-TOF/TOF (equipped with a pulsed

Smartbeam II Nd:YAG 355 nm laser) was used to analyze samples. FlexControl software was used to acquire the spectra in positive ion reflectron mode. Prior to data acquisition,

44

calibration was performed using a peptide standard obtained from Bruker Daltonics. The peptide standard contained angiotensin II, angiotensin I, substance P, bombesin, ACTH fragment 1-17, ACTH fragment 18-39 and somatosatin. All identified proteins were further verified as known salivary proteins by searching UCLA Dental Research

Institute’s Salivary Proteome Knowledge Base.69, 70

3.2.2 Cold Pressor Test and Saliva Collection

Whole saliva from 25 individuals (16 males and 9 females) was collected and pooled. Participants were asked to refrain from eating, drinking or smoking for at least two hours prior to collection. Samples were collected around the same time in the afternoon (2:30 pm) from each individual. Additionally, participants were asked to rinse their mouth with water prior to saliva collection. For the CPT, participants immersed their non-dominant hand in a container filled with ice-water for as long as they could tolerate, up to five minutes as demonstrated in Figure 3-2. Samples were collected prior to, immediately after and 20 minutes after the CPT. The passive drool collection method was used to collect the saliva into cryo vials. All samples were centrifuged at 3500 rpm for 15 minutes to remove cell debris and were stored at -20 °C until analyzed.

45

Figure 3-2 Participant immersing their hand in cold water as part of the cold pressor test

3.2.3 Salivary protein separation and identification

Both gel-based and gel-free separation techniques were used to analyze the salivary proteins. Figure 3-3 provides an overview of the workflow used for saliva analyses. Gel-based separations included the use of SDS-PAGE to compare saliva samples obtained prior to the cold pressor test, immediately after the cold pressor test and

20 minutes after the cold pressor test. In addition, 2D-PAGE was used in order to obtain better resolution facilitating the identification of additional salivary proteins.

Furthermore, 2D-DIGE compared the protein expression between collection time points overcoming 2D-PAGE reproducibility issues. Bottom-up proteomics was also utilized where the pooled saliva samples were enzymatically digested with trypsin, separated by nanoHPLC and analyzed by MALDI-MS/MS and ESI-MS/MS for identification.

Conversely, salivary proteins in whole saliva were initially separated by HPLC, and fractions of the eluting proteins were collected. These fractions were digested, separated by nanoHPLC and analyzed by MALDI-MS and MS/MS.

46

Figure 3-3 Simplified workflow of the methods used for the separation and identification of salivary proteins

3.2.3.1 SDS-PAGE

Saliva samples were prepared by mixing an aliquot of saliva with sample buffer containing 19:1 (v:v) Laemmli sample buffer: β-mercaptoethanol. The samples were heated at ~90 °C for five minutes to facilitate protein denaturing and to coat the proteins with SDS. Then, the salivary proteins were loaded onto 12% polyacrylamide gels in either equal mass loadings or equal volumes from each of the collection time points. The gels were run at a constant 200 V for approximately 45 minutes.

The salivary proteins were visualized by staining with either Bio-Safe Coomassie

Blue G-250 or SYPRO Ruby protein gel stain. Staining with Bio-Safe Coomassie Blue involved rinsing the gel with DI water, incubation with the stain for at least one hour and

47

destaining with DI water until a sufficient background was obtained. The Coomassie stained gels were scanned using a desktop scanner at its maximum resolution for further analyses.

Staining with SYPRO Ruby utilized an initial fixation step where the gel was placed in a container with 100 mL of a solution containing 50% methanol and 7% acetic acid, and it was gently rocked for 30 minutes. Afterwards, the solution was poured off, and this step was repeated once more with fresh solution. Then, 60 mL of the SYPRO

Ruby stain was added, and the gel was incubated in the solution overnight with gentle agitation. Finally, the gel was washed for 30 minutes using 100 mL of a solution containing 10% methanol and 7% acetic acid. The gel was further rinsed two times, for five minutes each, using ultrapure water before it was transferred to a fluorescence imager. As a fluorescent stain, the staining and destaining steps were performed excluding light from the gel container. SYPRO Ruby has an excitation maximum at ~450 nm and an emission maximum near 610 nm. The gels were visualized using a Typhoon

TRIO Variable Mode Imager (GE Healthcare Life Sciences, Pittsburgh, PA) using excitation at 488 nm and emission at 610 nm.

3.2.3.2 2D-PAGE

The overall 2D-PAGE workflow is depicted in Figure 3-4 along with the approximate time required for each step. For IPG strip rehydration, 200 µL of a saliva sample was dried down using a vacuum concentrator and redissolved in 125 µL of the

Bio-Rad’s starter kit rehydration/sample buffer. This solution containing buffer and salivary proteins was dispensed into a channel of a rehydration tray such that it covered

48

the entire length of the channel. The plastic cover from a commercial IPG strip was removed, and the strip was placed gel-side down over the solution containing the salivary proteins being careful not to trap any air bubbles under the strip. The IPG strips used in these experiments were 7 cm long and have a linear pH range of 3-10. Then, 4 mL of mineral oil was overlaid to completely cover the strip preventing the precipitation of urea form the rehydration buffer. It was left covered on the bench top at room temperature overnight (~14 hours).

Figure 3-4 Overall 2D-PAGE workflow

After rehydration of the IPG strip, salivary proteins were subjected to isoelectric focusing. The mineral oil was completely removed from the strip. Then, the IPG strip was transferred from the rehydration tray and placed gel-side up in the focusing tray.

Prewetted electrode wicks were placed on the ends of the strip. Again, the strip was overlaid with 4 mL of mineral oil. A multistep program, provided in Table 3.1, was used for isoelectric focusing. The first step was used to facilitate desalting of the sample without overheating the IPG strip and proteins as the ionic salts migrate towards either the anode or the cathode by quickly increasing the voltage to 250 V. The second step utilized a gradual increase in the voltage from 250 V to 4000 V in one hour in order to mobilize the proteins. The third step in the program held the voltage at 4000 V and was 49

used to complete the isoelectric focusing of the proteins. The elapsed time in this step was based on volt hours instead of a straightforward time interval for increased reproducibility between samples. Once isoelectric focusing was complete, the strip was held at 500 V until removal for the next step.

Table 3.1 Isoelectric focusing program used for pH 3-10, 7 cm IPG strips

Step Voltage (V) Ramp Time Time Units 1 250 Rapid 0:15 HH:MM 2 4000 Gradual 1:00 HH:MM 3 4000 Rapid 15,000 VoltHr 4 500 Hold

Once the salivary proteins have undergone isoelectric focusing, the IPG strip must be equilibrated prior to second dimension separation (i.e., by their molecular weight).

This step involves the denaturing of the focused proteins. The IPG strip was removed from the focusing tray and placed gel-side up into a clean channel of the rehydrating tray;

2 mL of equilibration buffer 1 (containing DTT) is added, and it is placed on a shaker for

10-15 minutes. Equilibration buffer 1 has 50 mg of DTT in 5 mL of an equilibration stock buffer which contains 6 M urea, 30% (w/v) glycerol, 2% (w/v) SDS in 0.05 M Tris-

HCl buffer (pH 8.8). This step ensures the denaturing of the proteins and the cleavage of cysteine-cysteine disulfide bonds. Afterwards, equilibration buffer 1 was poured off, and

2 mL of equilibration buffer 2 was added with gentle rocking for 10-15 minutes.

Equilibration buffer 2 contains 200 mg of iodoacetamide in 5 mL of the aforementioned equilibration stock buffer. This step is used to alkylate the cysteine residues, preventing the reformation of the disulfide bonds.

50

Once equilibrated, the strip can be placed on the second dimension gel. The plastic backing of the strip was placed against the back of the glass plate containing the gel (i.e., gel-side up). The IPG strip was gently pushed until it touched the bottom of the well. A small piece (few millimeters) of a wick was soaked with MW ladder standard and placed at one end of the well with the IPG strip. Then, the strip was overlaid with molten agarose. The agarose solution should be liquid, but it should not be boiling. Care should be taken to not trap any air bubbles, especially between the IPG strip and the second dimension gel. The gel should be left for 5-10 minutes to allow the agarose to solidify.

Initially, a low voltage (~50 V) can be used to allow the proteins from the IPG strip to enter the SDS-PAGE gel prior to separating the sample at the desired voltage.

Afterwards, the separated proteins were visualized using Bio-Safe Coomassie Blue G-

250.

3.2.3.3 2D-Difference gel electrophoresis (DIGE)

The protocol for 2D-DIGE was adapted from work by Kasamatsu et al. and

Nakashima et al.33, 71 A simplified workflow is depicted in Figure 3-5. Minimal labeling of salivary proteins from two different time points could be compared as Cy3 and Cy5 were used. From each time point, 200 µL of pooled saliva was dried down. The samples were reconstituted in 20 µL of 10 mM Tris-HCl (pH 8.8) to ensure the samples were at the proper pH for the labeling reaction. Then, 5 µL of 0.2 mM of Cy3 or Cy5 dye in

DMSO was added to the respective time points. The mixture was vortexed and incubated on ice for 30 minutes in the dark. Afterwards, the reaction was stopped by adding 7.35

µL of 1 mg/mL lysine. The labeled extracts were combined together, and the total volume was adjusted to 125 µL with rehydration buffer. 51

At this point, the proteins were subjected to isoelectric focusing and separation as described in the previous section ensuring that the samples were kept in the dark throughout the procedure. After isoelectric focusing and the second dimension separation, the gel was stored in DI water overnight before imaging using the Typhoon TRIO

Variable Mode Imager (GE Healthcare Life Sciences). Proteins labeled with Cy3 were visualized using an excitation wavelength of 555 nm and an emission wavelength of 570 nm. Proteins labeled with Cy5 were visualized using an excitation wavelength of 646 nm and an emission wavelength of 662 nm. The images were overlaid in ImageJ (NIH) for differential analysis.

52

Figure 3-5 2D-DIGE workflow

3.2.3.4 In-gel digestion and peptide mass fingerprinting

Proteins bands from SDS-PAGE and spots from 2D-PAGE were subjected to in- gel digestion for identification following a protocol described by Shevchenko et al. 72

Briefly, the bands or spots were excised from the gel, cut into ~ 1 mm3 cubes and placed into microcentrifuge tubes. The gel pieces were destained by adding 1:1 (v:v) 100 mM ammonium bicarbonate: acetonitrile with occasional vortexing for 30 minutes.

Dehydration and further destaining was achieved by adding neat acetonitrile to the microcentrifuge tube for ten minutes. At this point, all the liquid was removed from the 53

tube, and 50 µL of a 13 ng/µL trypsin solution in 10 mM ammonium bicarbonate (pH 8-

8.5) was added. If necessary, 10 mM ammonium bicarbonate was added to completely cover the pieces. Then, the samples were placed in the fridge to allow the trypsin to fully saturate the gel pieces. After two hours, 15 µL of 100 mM ammonium bicarbonate was added, and the microcentrifuge tubes were transferred to a 37 °C water bath for incubation overnight (~16 hours).

The trypsin digested samples were analyzed by MALDI-MS for peptide mass fingerprinting. A 1 µL aliquot of the digest was co-spotted with 1 µL of 10 mg/mL

CHCA matrix in 60:40 (v:v) acetonitrile: 0.1% TFA onto a MTP 384 ground steel plate.

Spectra were acquired from m/z 600 to m/z 3500. A MASCOT (Matrix Science) search was performed using the SwissProt database of tryptic peptides. The taxonomy was set to

Homo sapiens, and one missed cleavage as allowed. Methionine oxidation was set as a variable modification. A 0.2 Da mass tolerance was permitted. The threshold for confident protein identification was determined by the software to be a score equal to or greater than 56, which corresponds to p < 0.05.

3.2.3.5 HPLC of salivary proteins in whole saliva

Salivary proteins in whole saliva from each time point (i.e., prior to CPT, immediately after CPT and 20 minutes after CPT) were separated by HPLC. A

Prominence HPLC (Shimadzu) equipped with a Phenomenex Jupiter C4 column (5 µm,

300 Å, 2.00 mm I.D. x 250 mm) was used. Mobile phase A was H2O + 0.1% FA, and mobile phase B was acetonitrile + 0.1% FA. The total flow rate was 0.2 mL/min. A linear gradient of 15-75% B in 50 minutes and 75-90% B in the next ten minutes was used. A

54

PDA detector was used to determine the absorbance of the eluting compounds from 190-

600 nm. For all samples, 20 µL of saliva was manually injected onto the column.

Fractions containing the eluted proteins were collected every two minutes and dried down.

These collected fractions were digested with trypsin. First, they were reconstituted in 25 µL of 100 mM ammonium bicarbonate, and 2.5 µL of 200 mM dithiothreitol was added. The solution was incubated for one hour at 60 °C. Then, 10 µL of 200 mM iodoacetamide was added, and the solution was incubated for one hour at room temperature in the dark. An additional 2.5 µL of 200 mM dithiothreitol was added followed by incubation for one hour at room temperature in the dark. Finally, 300 µL of water, 100 µL of 100 mM ammonium bicarbonate and 20 µL of a trypsin solution containing 2.5 mg/mL of trypsin in 50 mM acetic acid were added. The samples were placed in a 37 °C water bath overnight (~16 hours), and the digestion was halted by adding 2 µL of TFA.

The tryptic peptides from the digested fractions were separated and identified by nanoHPLC-MALDI-MS/MS. An Ultimate 3000 nanoHPLC (Dionex) equipped with an

Acclaim PepMap 100 C18 column (3 µm, 100 Å, 75 µm I.D. x 15 cm) was used. The column was held at 25 °C, and the flow rate was 300 nL/min. Mobile phase A was H2O +

0.05% TFA, and mobile phase B was 0.05% TFA in 80:20 (v:v) acetonitrile: water. A linear gradient of 3-60% B in 80 minutes and 60-100% B in the next 15 minutes was used. Eluting peptides were monitored using a UV detector at 214 nm. As described in

Section 3.3.6, the nanoHPLC system was connected to the Proteineer fc spotting device

55

where the eluting peptides were mixed with CHCA matrix and spotted on a MALDI

Anchorchip plate.

As with the in-gel digests, MALDI-MS spectra were acquired from m/z 600 to m/z

3500. Selection of precursor ions was performed automatically by the software. The top nine most intense peaks with intensity greater than 1000 arbitrary units were selected for

MS/MS. A MASCOT (Matrix Science) search of the MS/MS spectra was performed using the SwissProt database of tryptic peptides. The taxonomy was set to Homo sapiens, and one missed cleavage as allowed. Methionine oxidation and carbamidomethylation of cysteine were set as a variable modifications. The parent ion mass tolerance was 0.2 Da, and the fragment ion mass tolerance was 0.7 Da.

3.2.3.6 nanoHPLC-MALDI-MS/MS and ESI-MS/MS of trypsin digested whole saliva

Alternatively, the whole saliva was initially digested using with trypsin prior to any separation. From each time point (i.e., prior to CPT, immediately after CPT and 20 minutes after CPT), 200 µL of pooled saliva was dried down. The samples were reconstituted in 25 µL of 100 mM ammonium bicarbonate. To this solution, 2.5 µL of

200 mM dithiothreitol was added, and it was incubated for one hour at 60 °C. Afterwards,

10 µL of 200 mM iodoacetamide was added. The sample was incubated at room temperature for one hour in the dark. Then, another 2.5 µL of the dithiothreitol solution was added. It was incubated for an additional hour in the dark before addition of 300 µL of water, 100 µL of 100 mM ammonium bicarbonate and 20 µL of a 0.045 mg/mL trypsin solution in 50 mM acetic acid. The samples were incubated overnight (~16 hours) in a 37 °C water bath. The digestion was halted by adding 2 µL of TFA.

56

The tryptic peptides from the whole saliva digestion were separated and identified using nanoHPLC-MALDI-MS/MS in the same manner as for the digested fractions

(Section 4.2.3.5), except the gradient was extended as this mixture is more complex. A linear gradient of 3-55% B in 120 minutes and 55-100% B in the next 10 minutes was used. The eluting peptides also co-spotted with CHCA matrix. These spots were subjected to automated MALDI-MS/MS whose spectra were also subjected to a

MASCOT search as previously described.

Alternatively, the tryptic peptides were analyzed by nanoHPLC-ESI-MS/MS using a Thermo ESI-Orbitrap Fusion. The peptides were separated using an Acclaim®

PepMap RSLC C18 column (75 µm x 15 cm, 2 µm particles, 100 Å pore size) on a

Dionex Ultimate 3000 HPLC RSLCnano system. The column oven was 35 °C, and the flow rate was 300 nL/min. Mobile phase A was 0.1% FA in water, and mobile phase B consisted of 0.08% FA in 80:20 (v:v) acetonitrile: water. Injections were made using

0.05% TFA in 98:2 (v:v) water: acetonitrile. The peptides were loaded on a precolumn for five minutes for desalting while the C18 column was held at the initial composition of the mobile phase. A gradient consisted of a linear increase from 2.5-25% B in 100 min and linear ramp from 25-40%B in 20 min followed by column washing and re- equilibration was used.

The eluting peptides were directly coupled to the ESI-Orbitrap Fusion. Mass spectra were acquired with a resolution of 120,000 and a scan range of m/z 400 to m/z

1600 in positive ion mode. MS/MS spectra were acquired in a data dependent manner in the ion trap using CID fragmentation. The parameters for MS/MS selection were a charge

57

state of 2-4 with a minimum intensity of 5.0e3. Once an ion was selected, it was excluded for 40 seconds before it could be selected for fragmentation again. The MS/MS spectra were subjected to SEQUEST HT searching using Proteome Discoverer (version 1.4,

Thermo). The taxonomy was set to Homo sapiens, and two missed cleavages were allowed. Methionine oxidation and phosphorylation of serine, threonine or tyrosine were set as variable modifications. Carbamidomethylation of cysteine was set as a fixed modification. The parent ion mass tolerance was set to 10 ppm, and the fragment ion mass tolerance was 0.6 Da. The results were run through Percolator with a relaxed false discovery rate of 0.05 and a strict false discovery rate of 0.01.73

3.2.4 Depletion of alpha-amylase from whole saliva

Alpha-amylase was removed from whole saliva using a starch column as previously described.74 A 0.45 µm filter was placed at the tip of a 1 mL syringe. The syringe was filled with approximately 350 mg of potato starch. First, the column was wetted by pressing 600 µL of water through the potato starch filled syringe. Then, 400

µL of whole saliva was slowly pressed through the potato starch. The eluting proteins from this step should be free of alpha-amylase. To elute the alpha-amylase, 500 µL of

10% (w/v) SDS, 3% (v/v) β-mercaptoethanol was applied to the starch column and incubated for 10 minutes at room temperature at which point it was pressed through. This step was repeated once more.

The whole saliva, alpha-amylase free fraction and alpha-amylase fractions were analyzed by SDS-PAGE. For the purified fractions, half of the collected sample was dried down and reconstituted in 100 µL of water and mixed with an appropriate amount

58

of Laemmli sample buffer and β-mercaptoethanol. The other half was desalted using

SepPack C18 SPE cartridges (Waters, Millford, MA). For SPE, the column was wetted with 1 mL of ethanol, three times. Then, the column was equilibrated with 1 mL of water, three times. Next, the sample was applied five times to the cartridge. A washing step using 1 mL of water was used, and the proteins were eluted with 1 mL of acetonitrile.

The eluted proteins were also dried down, reconstituted in 100 µL of water and mixed with the appropriate amount of Laemmli sample buffer and β-mercaptoethanol.

3.2.5 Investigation of salivary phosphoproteins and phosphopeptides

Phosphorylation of salivary proteins and their relationship to the cold pressor test was also explored. Their analyses can be divided into two main procedures, gel based and gel-free.

3.2.5.1 Gel-based analysis of phosphoproteins

SDS-PAGE of whole saliva at the three different time points was performed.

Different amounts of saliva were studied. For increased resolution of the proteins smaller than 50 kDa, 18% polyacrylamide gels were run at 175 V. PeppermintStick phosphoprotein molecular weight standards were loaded into one lane of the gel. From the stock solution, 1 µL of the standard was combined with 6 µL of the loading buffer and heated for four minutes at ~90 °C. This standard contains four nonphosphorylated proteins and two phosphorylated proteins. The nonphosphorylated proteins are β- galactosidase (MW 116,250), bovine serum albumin (MW 66,200), avidin (MW 18,000) and lysozyme (MW 14,400). The two phosphorylated proteins are ovalbumin (MW

45,000) and β-casein (MW 23,600).

59

Following separation, the gel was stained using Pro-Q Diamond phosphoprotein gel stain which preferentially interacts with the phosphate group allowing the detection of phosphoproteins. The first step involved fixing the gel in 100 mL of 50% methanol, 10% acetic acid for 30 min, twice. Next, the gel was washed using 100 mL of ultrapure water for 10 minutes, three times. The gel was incubated with 60 mL of the stain for 60-90 minutes. Afterwards, the gel was destained in 100 mL of 20% acetonitrile, 50 mM sodium acetate (pH 4) three times for 30 minutes each. Finally, the gel was washed with

100 mL of ultrapure water for five minutes, twice before imaging. As a fluorescent stain, the gel container was kept in the dark during the staining and destaining steps. Pro-Q

Diamond has an excitation maximum at ~555 nm and an emission maximum near 580 nm. The gels were visualized using the Typhoon TRIO Variable Mode Imager using excitation at 532 nm and emission at 580 nm. Once imaged, the gel was stained using

Bio-Safe Coomassie G-250 for total protein visualization as previously described.

3.2.5.2 Gel-free analysis of phosphopeptides

Salivary phosphopeptides were also enriched using a commercial Fe-NTA phosphopeptide enrichment kit, as summarized in Figure 3-6. It is capable of enriching up to 150 µg of phosphopeptides. The sample was first dried down and resuspended in

200 µL of their binding buffer. The resuspended sample was added to the Fe-NTA spin column and incubated at room temperature for 20 minutes with end-over-end rotation.

Afterwards, the column was centrifuged at 1,000 x g for one minute. This flow-through

(i.e., sample prep fraction) was collected for analysis. Next, 100 µL of wash buffer A

(150 µL of their wash buffer 2X stock + 150 µL of ultrapure water) was added to the column, gently mixed and centrifuged at 1,000 x g for one minute. This flow-through 60

(i.e., wash A fraction) was collected, and the step was repeated combining the aliquots.

Then, 100 µL of wash buffer B (150 µL of their wash buffer 2X stock, 30 µL of acetonitrile and 120 µL of ultrapure water) was added to the column, gently mixed and centrifuged at 1,000 x g for one minute. Again, the flow-through (i.e., wash B fraction) was collected, and the step was repeated. At this point, the column was prepared for phosphopeptide elution by adding 50 µL of their elution buffer to the resin. It was incubated at room temperature for three to five minutes. Afterwards, the column was centrifuged at 1,000 x g for one minute. This fraction contained the isolated phosphopeptides. The elution buffer was applied two more times. The three phosphopeptide fractions were combined together and acidified by adding 200 µL of

2.5% TFA.

Figure 3-6 Overview of the commercial Fe-NTA phosphopeptide enrichment kit workflow For initial studies, casein was used as a phosphoprotein standard. A trypsin digest was performed by dissolving 1.5 mg of casein in 50 µL of 100 mM ammonium

61

bicarbonate. Then, 5 µL of 200 mM dithiothreitol was added, and the sample was incubated at 60 °C for one hour. Afterwards, 20 µL of 200 mM iodoacetamide was added and incubated at room temperature for one hour in the dark. An additional 5 µL of 200 mM dithiothreitol was added with incubation in the dark for one hour at room temperature. Finally, 600 µL of water, 200 µL of 100 mM ammonium bicarbonate and 40

µL of 2.5 mg/mL trypsin in 50 mM acetic acid were added. The sample was incubated overnight (~16 hours) in a 37 °C water bath. The digestion was halted by adding 2 µL of

TFA. From this mixture 250 µg of casein was dried down for phosphopeptide enrichment. Salivary phosphopeptides from the trypsin digestion of whole saliva digestion described in Section 3.2.3.6 were also enriched using the kit. The tryptic digests and collected fractions were co-spotted with either 10 mg/mL CHCA in 60:40 (v:v) acetonitrile: water with 0.1% TFA or 20 mg/mL DHB in 50:50 (v:v) methanol: water with 0.3% TFA for MALDI-MS. Mass spectra were acquired from m/z 500 to m/z 3500.

Isolated phosphopeptides were subjected to nanoHPLC-MALDI-MS/MS as in

Section 3.2.3.5, with the exception of the gradient. A linear gradient of 3-55% B in 30 minutes followed by 55-100% B in the next 5 minutes was used. The eluting peptides were also co-spotted with CHCA matrix and were subjected to automated MALDI-

MS/MS. A MASCOT search was performed. Variable modifications included methionine oxidation and phosphorylation of serine, threonine or tyrosine. The enzyme was set to either trypsin or semi-trypsin. The semi-trypsin setting only requires one trypsin cleavage site, whereas the other cleavage site could result from endogenous enzymes.

62

3.3 Results and Discussion

3.3.1 Protein Identification by Gel-Based Methods

For protein identification, salivary proteins were separated using SDS-PAGE. A representative image of the separated proteins is shown in Figure 3-7. In general, less than ten bands could be routinely visualized. This relatively low number of bands is in good agreement with previous reports.75 In addition, a comparison of total protein visualization using Coomassie blue and SYPRO Ruby was performed. SYPRO Ruby is more sensitive than Coomassie, so it was expected to stain additional proteins. Both equal volumes and equal total amount of protein from each collection time point were analyzed by SDS-PAGE. A similar number of proteins were visualized with SYPRO Ruby and

Coomassie blue. However, one additional faint band below the cystatin band could be visualized exclusively after SYPRO Ruby staining. A few other bands were stained more intensely using SYPRO Ruby in comparison to Coomassie blue. Overall, no major benefit was obtained from using the more sensitive fluorescent protein stain for these samples.

63

Figure 3-7 Image of SDS-PAGE separation of equal volumes and equal total protein loaded on a 12% polyacrylamide gel visualized with (A) Coomassie blue G-250 and (B) SYPRO Ruby.

The bands were excised and digested with trypsin for peptide mass fingerprinting.

Overall, 31 salivary proteins were identified from these bands by peptide mass fingerprinting (Appendix Table A.1). As was expected from previous studies, alpha- amylase was the most prevalent band.51, 74 The cystatin type S family (cystatin-S, cystatin-SN and cystatin-SA) was also a prominent group of proteins on the gels with a molecular weight of slightly less than 15 kDa. Other common proteins include serum albumin, keratins, carbonic anhydrase 6 and lysozyme C.

For increased resolution of the salivary proteins, 2D-PAGE was also performed.

After focusing was complete, the cysteine disulfide bonds were cleaved using DTT and alkylated using iodoacetamide to prevent their reformation. At this point, the focused proteins were separated using SDS-PAGE. A representative gel is shown in Figure 3-8.

As expected, a greater number of protein spots were resolved using 2D-PAGE. Using one

64

dimension SDS-PAGE, less than 10 salivary protein bands were routinely observed.

Comparatively, at least 26 discrete regions were visualized from 2D-PAGE. Each of these regions contained multiple spots. However, few proteins were observed in the region above 70 kDa. Consequently, they may have not efficiently entered the IPG strip.

Not all the proteins were completely resolved though. In particular, a relatively large protein spot was typically observed in the low molecular weight (< 10 kDa) and basic pI region of the gel.

Figure 3-8 Image of 2D-PAGE separation of proteins originating from saliva collected prior to the CPT labeled with some of the identified salivary proteins

The salivary protein spots were also excised and subjected to in-gel trypsin digestion for peptide mass fingerprinting. Overall, fewer proteins were positively identified by peptide mass fingerprinting from 2D-PAGE compared to the one-dimension gels. In total, 15 proteins were identified and confirmed as salivary proteins which are summarized in Table 3.2. A few of the identified proteins spots are labeled on the 2D-

65

PAGE in Figure 3-8. As before, the most common proteins identified are alpha-amylase, cystatin-SN, cystatin-S and keratins. Periplakin was another protein identified. It is a member of the plakin class, and periplakin may aid in wound healing.76 Vinculin is a multiprotein linker involved in cellular adhesion that was also identified. Kim et al. have identified increased plasma levels of vinculin as a potential biomarker for the diagnosis of age-related macular degeneration.77 Hardt et al. observed zinc-alpha-2-glycoprotein in saliva as three separate spots due to post-translational modifications using 2D-PAGE.57

This protein is involved in the metabolism of lipids, particularly for decreasing fatty acids from adipose tissue.57, 78 Furthermore, its increase has been associated with several types of carcinoma making it a potential biomarker of cancer.78 Differences in their abundances have been shown to play a role in diseases, which may enable them to be used as biomarkers.

Table 3.2 Protein identified by peptide mass fingerprinting from 2D-PAGE

Protein Code Protein Name Mass Score Matches AMY1 Alpha-amylase 1 55,909 175 22 AMY2B Alpha-amylase 2B 55,891 165 21 AMYP Pancreatic alpha-amylase 55,889 128 19 CYTN Cystatin-SN 14,316 115 11 CYTS Cystatin-S 14,189 78 8 IGHA1 Ig alpha-1 chain C region 37,631 63 9 IGHA2 Ig alpha-2 chain C region 36,503 64 9 K1C10 Keratin, type I cytoskeletal 10 58,827 73 13 K1C9 Keratin, type I cytoskeletal 9 62,064 59 11 K22E Keratin, type II cytoskeletal 2 epidermal 65,433 96 15 K2C1 Keratin, type II cytoskeletal 1 65,908 153 22 PEPL Periplakin 204,747 68 13 TMEM9 Transmembrane protein 9 18,568 56 7 VINC Vinculin 123,668 65 16 ZA2G Zinc-alpha-2-glycoprotein 32,145 67 12

66

Overall, these results had some similarities with those in the literature. Vitorino et al. were able to identify 43 different spots from their 2D gels which had better resolution.51 Their increased resolution is a likely consequence of utilizing larger format gels that employed a 13 cm IPG strip instead of a 7 cm strip. Hu et al. also identified a series of alpha-amylase in the same region as was detected in these experiments.79

Additionally, Ghafouri et al. were able to identify 101 proteins spots corresponding to 21 different proteins using a pH 4-7 IPG strip indicating that narrower ranges may also improve protein separation.59 Common proteins identified from their gels and the gels ran in these experiments include alpha-amylase, cystatin-S and cystatin-SN. This indicates that the most abundant proteins were identified in our experiments, but additional studies need to be performed in order to increase the resolution of the separations and the number of proteins identified.

For relative quantification, 2D-DIGE was used to compare two collection time points. Equal volumes of each saliva sample were used. The salivary proteins were minimally labeled with either Cy3 or Cy5. Figure 3-5 shows the image of the proteins from saliva collected prior to the CPT labeled with Cy3 dye, the image of the Cy5 labeled proteins from saliva collected immediately after the cold pressor test and an overlay of these images. Overall, the gels images were similar to the 2D-PAGE of a single sample that were visualized using Coomassie blue G-250. However, some of the lower molecular weight proteins are not as apparent after labeling with the cyanine dye. On the other hand, several regions in the high molecular weight region were visualized with using the cyanine dyes that were not detected using Coomassie blue. Furthermore, the proteins

67

were not typically detected as discrete spots. Instead, two vertical streaks were present.

One vertical streak was towards pI 3 and the other closer to pI 10.

In the overlay image, differential color may indicate a difference in the relative abundance of the salivary proteins. Specifically, proteins that are green have a greater abundance in the saliva sample labeled with Cy3 (i.e., collected prior to the CPT), and proteins that are red are present in greater abundance in the saliva sample labeled with

Cy5 (i.e., collected immediately after the CPT). Proteins whose relative abundance does not appreciably differ between the two samples are present in yellow. Overall, most of the proteins have similar abundances between the two collection time points. However, some spots appear to have a slightly redder hue which may indicate an increase in their relative abundances after the cold stress. Due to a lack of software to appropriately align the images and define the spot areas, further relative quantitation was not performed at this time.

3.3.2 Protein identification using HPLC separation of saliva

3.3.2.1 HPLC of salivary proteins in whole saliva

Proteins from equal volumes of saliva collected prior to, immediately after and 20 minutes the cold pressor test were separated by HPLC. The absorbance at 280 nm was monitored to determine the elution of the salivary proteins. Figure 3-9 depicts an overlay of the HPLC chromatograms of the separated salivary proteins from each time point.

Overall, a relatively low number peaks were observed, especially considering the complex nature of saliva. Furthermore, these results were not in good agreement with other reported separations. For instance, Hu et al. obtained at least ten resolved peaks in 68

their HPLC chromatograms.79 Millea et al. also observed a complex LC-MS chromatogram from their separation of whole saliva with numerous proteins separated.47

As similar columns, flow rates and solvents were used, the reasons for these differences are not entirely clear. One potential explanation is that larger salivary protein amounts were loaded in the other studies.

For a relative comparison of the time points based on protein amount, all the absorbances were normalized to 25 µg of loaded protein. This normalization was performed by multiplying the recorded absorbance as a function of time by the number obtained from the amount of protein loaded on the column divided by 25 µg. For example, 20 µL of saliva collected prior to the CPT contains 30.6 µg of protein.

Therefore, taking the 30.6 µg of protein loaded divided by 25 µg yields 0.816, and the absorbance for this sample was multiplied by 0.816. Some slight differences between the samples were observed. Comparing the absorbances, the largest absorbance was typically observed for the saliva sample collected 20 minutes after the cold pressor test (e.g., retention times of 4 and 24 minutes). This likely indicates that some salivary proteins increase in abundance in response to the CPT.

69

Figure 3-9 HPLC chromatogram showing the separation of salivary proteins. Pooled saliva A was collected prior to the CPT. Pooled saliva B was collected immediately after the CPT. Pooled saliva C was collected 20 min after CPT. Proteins were identified by nano-HPLC-MALDI-MS/MS.

In order to identify the eluting salivary proteins, fractions were collected at two minute intervals. These fractions were subjected to trypsin digestion. The peptides from each digested fraction were further separated by nanoHPLC which was off-line coupled to MALDI-MS/MS. This allowed for greater separation of co-eluting salivary proteins, facilitating their identification.

Many of the proteins identified by this strategy were also identified by peptide mass fingerprinting from gel electrophoresis. Some of the most prevalent salivary

70

proteins identified by HPLC fractionation of the whole saliva include alpha-amylase, basic salivary proline rich protein 3, cystatin, lysozyme c, prolactin-inducible protein and protein S100-A8. Several forms of keratin were also observed. An example of nanoHPLC chromatogram of the tryptic peptides labeled with their identifications is depicted in

Figure A-1.

3.3.2.2 Protein identification using nanoHPLC-MALDI-MS/MS and ESI-MS/MS of trypsin digested whole saliva

Alternatively, the proteins in whole saliva were reduced, alkylated and digested with trypsin. This complex mixture was separated using nanoHPLC. Appendix A Figure

A-2 depicts the UV-detected nanoHPLC chromatogram of 500 ng of trypsin digested whole saliva collected immediately after the CPT labeled with proteins identified by

MALDI-MS/MS. As expected, more peptides were present in this mixture compared to the digested fractions from HPLC-separated salivary proteins. A few of the peaks were well resolved from adjacent peaks, but many of the peaks were not fully resolved. This indicates that the peptides are not fully separated, and that it may be necessary to extend the gradient further to obtain an increased separation. The eluting peptides were co- spotted with CHCA matrix on a MALDI Anchorchip plate. In total, 39 salivary proteins were identified by MALDI-MS/MS. Common protein families include actins, alpha- amylase and cystatins. Some of the other prevalent proteins include carbonic anhydrase,

Ig alpha-1 chain C region, lipocalin-1, mucin-5B, polymeric-inducible protein, serum albumin and zinc-alpha-2-glycoprotein.

71

Thus far three main techniques were used for separation and identification of salivary proteins and peptides: gel electrophoresis, HPLC fractionation and whole saliva digestion. Of these methods, digestion of whole saliva with trypsin followed by nanoHPLC-MALDI-MS/MS yielded the greatest number of proteins identified. In comparison to SDS-PAGE, this was not surprising as the limit of detection for SDS-

PAGE is higher than for nanoHPLC-MALDI-MS/MS. Consequently, only the most abundant proteins in saliva would be visualized and subjected to peptide mass fingerprinting. However, it would have been expected that HPLC fractionation will yield more identifications than the whole saliva digestion. By collecting fractions, the complex mixture would have been simplified before trypsin digestion and nanoHPLC. However, the lower number of identifications may also be a result of only the most abundant proteins were efficiently detected.

To overcome the challenge of identifying a limited number of salivary proteins, the higher-throughput method of nanoHPLC-ESI-MS/MS was utilized to improve the number of identified proteins. This method was expected to increase the number of identifications for multiple reasons. Specifically, it is an on-line technique in which the eluting peptides are directly coupled to the mass spectrometer. The primary advantage of the on-line coupling is the increased speed of analysis. In addition, the scan rate of the

Thermo ESI-Orbitrap is higher than Bruker’s MALDI-TOF/TOF. This allows more

MS/MS spectra to be collected and subjected to database searching using ESI-MS/MS.

Both instruments utilize data-dependent acquisition to determine the precursor ion selection. However, a dynamic exclusion could be used for the Thermo ESI-Orbitrap.

Once a precursor ion has been subjected to MS/MS fragmentation, dynamic exclusion 72

permits that ion to be excluded from MS/MS for a specified time interval that corresponds to the average nanoLC peak width. This enables more peptides to be subjected to MS/MS and database searching. Another difference between the techniques is that nanoHPLC-ESI-MS/MS does not necessitate the use of matrix. Consequently, peptides with a small mass, particularly those below 700 Da, may be better detected by

ESI as they may have mass interferences or ionization suppression with the matrix in

MALDI-MS.

Likewise, tryptic peptides from the whole saliva digests were separated and identified by nanoHPLC-ESI-MS/MS. Different gradients were used to obtain an optimal separation of the peptides. The base peak from one of the separations is shown in Figure

3-10. Numerous peaks were observed in the base peak chromatogram. Many of the peaks were relatively well resolved with peak widths less than one minute. While separation could be seen in the peaks eluting between 80 and 128 minutes, their peak widths increased. As two to four minute wide peaks were observed, it indicates that the peptides in this region were not as efficiently separated. However, a compromise must be made between extending the length of the gradient against increasing the analysis time and possible band broadening.

73

Figure 3-10 Base peak chromatogram of the separation of 1.33 µg of salivary tryptic peptides from a sample collected before the CPT by nanoHPLC-ESI-MS/MS

This methodology greatly enhanced the number of salivary proteins identified.

For the single run shown, 100 salivary proteins were identified. In comparison, less than

80 salivary proteins were identified by combining all the different methods using

MALDI-MS/MS. At this point, 172 salivary proteins have been identified from the nanoLC-ESI-MS/MS of the whole saliva tryptic peptides. As in the previous methods, alpha-amylase, cystatin-S, cystatin-SA, cystatin-SN, serum albumin, carbonic anhydrase

6 and lysozyme C were identified. However, additional forms of the cystatin family (e.g., cystatin-B, cystatin-C and cystatin-D) were also identified. Many proline-rich proteins were identified including small proline-rich protein 3, basic salivary proline-rich protein

3 and basic salivary proline-rich protein 2. Three different mucin family members were also present. Numerous immunoglobulin and keratin proteins were identified.

74

Furthermore, identification of the low-molecular weight proteins histatin-1 and histatin-3 was also possible. Additional identifications may be facilitated by using multidimensional separations, including strong cation exchange chromatography prior to reversed-phase HPLC analysis of the salivary peptides.70

Overall four different methods were used for the separation and identification of salivary proteins: gel electrophoresis and peptide mass fingerprinting; HPLC fractionation; nanoHPLC-MALDI-MS/MS of trypsin digested whole saliva; and nanoHPLC-ESI-MS/MS of trypsin digested whole saliva. Multiple proteins could be identified using at least two of the methods, and the most abundant salivary proteins were identified using all four methods. These proteins include alpha-amylase, cystatin-S, cystatin-SA, cystatin-SN, Ig alpha-1 chain C region, Ig alpha-2 chain C region, prolactin- inducible protein and serum albumin. However, nanoHPLC-ESI-MS/MS of the tryptic peptides identified the greatest number of proteins. It increased the number of identified proteins by an order of magnitude. Furthermore, 138 proteins were only identified using the nanoHPLC-ESI-MS/MS workflow. A complete listing of identified salivary proteins is included in Appendix Table A.1. In addition, a Venn diagram of the functions of the salivary proteins identified using nanoHPLC-ESI-MS/MS is depicted in Appendix A-3.

The most common functions include catalytic activity and protein binding.

3.3.3 Amylase depletion for lower abundance protein identification

Most of the aforementioned methods were only capable of identifying the most abundant proteins in saliva. In order to aide in the identification of lower abundant proteins, it was necessary to remove alpha-amylase as it is the most abundant salivary

75

protein. A relatively simple starch column was fabricated. The alpha-amylase will interact with the starch and be retained, while the other salivary proteins will flow through the column.74 Compared to other techniques, such as immunoprecipitation, this method was relatively inexpensive, simple and quick to perform. It is also less expensive as antibodies are not utilized.

A 0.4 mL aliquot of whole saliva was subjected to alpha-amylase depletion. The initial flow-through was collected as the alpha-amylase free fraction. The two fractions containing the eluted alpha-amylase were combined together. One half of each fraction

(~200 µL of amylase depleted saliva and ~500 µL of the amylase fraction) was dried down and reconstituted in 100 µL of water to concentrate the samples. The other half of each fraction was subjected to desalting using SepPack C18 SPE cartridges. This desalting step was performed because the potato starch was obtained from a grocery store and likely contained many salts that were introduced to the saliva samples. After desalting, the fractions were also dried down and reconstituted in 100 µL of water.

The effectiveness of this procedure for removing alpha-amylase was determined using SDS-PAGE. A 12% polyacrylamide gel was loaded with the whole saliva prior to depletion, the amylase depleted fraction (with and without desalting) and the amylase fraction (with and without desalting). The separated proteins were visualized with

Coomassie blue. An image of the gel is shown in Figure 3-11. The first lane contains 50

µL of the saliva sample prior to loading on the starch column. A large band corresponding to alpha-amylase is located slightly above the 47 kDa molecular weight marker band. The next sample lane contains the alpha-amylase depleted saliva fractions

76

without desalting. Many of the same salivary protein bands were observed in this lane.

However, the band corresponding to alpha-amylase is no longer present, indicating that the potato starch column removed a majority of the protein. Comparing these two lanes, no additional bands were present after amylase depletion. The two lanes were not loaded with the same amount of total protein though, so it is possible that the lower abundance proteins could be more concentrated in the depleted sample compared to the whole saliva and too little of the fraction was analyzed to detect them. The lane corresponding to the alpha-amylase fraction contains one major band, identified as alpha-amylase. No additional bands were present which indicates that the potato starch column specifically bound only the alpha-amylase in saliva. No bands were present from the samples that were desalted. A few explanations are possible. The first possibility is that the protein concentration was too low. Alternatively, the proteins may not have interacted with the

SPE cartridge optimally resulting in a lack of retention and their loss from the column.

Conversely, the proteins may have been retained on the SPE cartridge and were not eluted.

77

Figure 3-11 Image of SDS-PAGE of whole saliva and saliva subjected to alpha-amylase depletion using a potato starch column

These results were verified using nanoHPLC-MALDI-MS/MS of the trypsin digest of the alpha-amylase depleted fractions. Many of the same proteins were identified

(e.g., cystatins, serum albumin, mucin-5B and prolactin-inducible protein). However, alpha-amylase was not identified in the fraction. In total, 24 proteins were identified from the preliminary separations of the alpha-amylase depleted saliva. Further concentration of the sample to enable loading more protein on the column should result in more proteins, especially lower abundance proteins, to be identified.

3.3.4 Investigation of salivary phosphoproteins

Salivary proteins undergo many post-translational modifications. One of the most common post-translational modifications is phosphorylation. Phosphorylation is involved in many cellular processes such as cell growth, proliferation, apoptosis and metabolism.80

78

Consequently, salivary phosphoproteins may be indicative of various physiological conditions, including stress. Their presence and relative abundances were investigated using differential staining of SDS-PAGE gels and their enrichment in saliva using a commercial Fe-NTA kit.

3.3.4.1 Gel-based methods of protein identification

SDS-PAGE using a phosphoprotein specific stain was used to examine global changes in the salivary phosphoproteins in response to the cold pressor test. Proteins from whole saliva collected prior to the CPT, immediately after the CPT and 20 minutes after the CPT were separated using an 18% polyacrylamide gel. Higher acrylamide content was used to increase the separation of proteins smaller than 50 kDa which were targeted as being differentially phosphorylated in response to the cold pressor test. After separation, the gels were visualized using the fluorescent stain Pro-Q Diamond which is specific for phosphoproteins. It is capable of detecting phosphoserine, phosphothreonine and phosphotyrosine with comparable sensitivity as the dye binds directly with the phosphate. This binding is a noncovalent interaction which makes the stain mass spectrometry compatible enabling peptide mass fingerprinting or MS/MS for protein identification and characterization.81, 82 The total protein stain Coomassie blue G-250 was also utilized on the same gel (Figure 3-12).

79

Figure 3-12 Image of SDS-PAGE of whole saliva samples prior to, immediately after and 20 minutes after the CPT. Samples were ran in duplicate on the gel with 35 μL and 45 μL of saliva loaded, respectively. (A) Total protein visualization using Coomassie blue G-250 (B) Phosphoprotein visualization using Pro-Q Diamond.

For each collection time point, the samples were run in duplicate using two different volumes of saliva loaded. As expected, the lanes with 45 µL of saliva loaded showed more intense staining for all of the proteins compared to the lanes with 30 µL of saliva loaded (Figure 3-12A). Several phosphorylated proteins were visible using Pro-Q

Diamond (Figure 3-12B). All of these proteins have a smaller molecular weight, appearing below the 23.6 kDa marker. Overall, many of the bands exhibit a low intensity of staining indicating that they are present in lower concentrations in saliva. As in previous experiments, the bands were subjected to peptide mass fingerprinting for identification.

The two different staining techniques, Coomassie blue and Pro-Q Diamond, can be compared in order to determine the proteins’ relative amount of phosphorylation in regards to its amount in whole saliva. For instance, alpha-amylase is a highly abundant protein in saliva as observed by Coomassie staining, but it is not phosphorylated as

80

indicated by a lack of staining with Pro-Q Diamond. Conversely, there are a series of unidentified phosphorylated protein bands between the 23.6 kDa molecular marker band and the cystatin band that are low in abundance in saliva as they cannot be visualized with Coomassie. These proteins have a relatively high amount of phosphorylation (in regards to the amount of number of phosphate groups detected). A similar trend is seen with a low-molecular weight protein (<14.4 kDa) in which it may be highly phosphorylated but its total abundance in saliva is low.

A series of three bands for cystatin can be observed with Coomassie. However, only one of these cystatin bands appears to be phosphorylated. It is postulated that the upper band is cystatin-S because it is known to be either monophosphorylated at Ser23 or diphosphorylated at Ser21 and Ser23.83, 84 Neither cystatin-SN nor cystatin-SA have been shown to be phosphorylated. The results of peptide mass fingerprinting and MALDI-

MS/MS were inconclusive for differentiating which form of cystatin was present in the band or for determining the site(s) of phosphorylation.

3.3.4.2 Gel-free methods of salivary phosphoprotein identification

Gel-free approaches were also utilized for investigating salivary proteins and peptides in saliva in response to the cold pressor test. In this approach, phosphopeptides were enriched using Fe-NTA spin columns. Then, the phosphopeptides were analyzed by

MALDI-MS and MS/MS. Initial studies used a phosphoprotein standard, bovine casein.

The casein standard was digested with trypsin, and 250 µg of the digest was loaded onto the Fe-NTA spin column. A total of four fractions were collected. The first three fractions

81

were proteins that did not strongly interact with the column, predominately peptides that are not phosphorylated. The fourth fraction primarily contains phosphorylated peptides.

Appendix Figure A-4 is the MALDI-MS of the trypsin-digested casein sample and the enriched phosphopeptide fraction. Many peptides corresponding to alpha-S1 of bovine casein were identified from the digest. The most intense peaks (m/z 1267.6,

1383.7, 1759.9 and 2316.0) correspond to nonphosphorylated peptides. However, the phosphorylated peptides were not readily identified. Very low intensity peaks could be seen at m/z 1660.8 (intensity ~250 a.u.) and m/z 1951.9 (intensity ~ 320 a.u.) corresponding to known phosphopeptides. Their intensity was three orders of magnitude lower than the nonphosphorylated peptides in the digest. Conversely, after enrichment, they were the most intense peaks seen in the mass spectrum. This highlights the common challenge in the analysis of phosphopeptides by MALDI-MS. Specifically, phosphopeptides are typically present in lower abundance than non-phosphorylated peptides. In addition, they experience ionization suppression.42

After enrichment, the phosphopeptides could be readily detected by MALDI-MS.

Two alpha-S1 casein phosphopeptides and their non-phosphorylated analogs were identified. The peak at m/z 1661.051 corresponds to the VPQLEIVPNS(Phospho)AEER tryptic peptide. The peak at m/z 1952.245 corresponds to the

YKVPQLEIVPNS(Phospho)AEER tryptic peptide of alpha-S1 bovine casein. The peak at m/z 1871.927 was identified as having the same sequence as the ion at m/z 1952.2 without phosphorylation of the serine residue. However, a few very low intensity peaks

82

corresponding to nonphosphorylated peptides were present with very low intensities, such as m/z 1267.8.

At this point, the workflow for enrichment and identification of phosphopeptides using the Fe-NTA spin columns had been verified to work appropriately. The trypsin digested saliva samples collected before the cold pressor test, immediately after and 20 minutes after the CPT were loaded onto the Fe-NTA spin columns. Figure 3-13 compares the three saliva samples after phosphopeptide enrichment. The pooled saliva collected prior to the CPT (Figure 3-13A) was spotted with 10 mg/mL CHCA matrix. Numerous ions were present from m/z 900 to m/z 2000 after phosphopeptide enrichment. The most intense peak was located at m/z 1897.8. The pooled saliva collected immediately after the

CPT (Figure 3-13B) was spotted with 20 mg/mL DHB matrix. There were some similarities between the two samples. For instance, the peak at m/z 1897 was observed in both samples. However, their relative intensities were different. In addition, a few phosphopeptides could also be observed in the region greater than m/z 3300 in this sample. The pooled saliva collected 20 minutes after the CPT (Figure 3-13C) was also spotted with 20 mg/mL DHB matrix. The spectrum was similar to that of the saliva immediately after the CPT. However, the overall intensity of the ions was lower and fewer signals were observed. These differences could be the result of differences in the salivary phosphopeptides as a result of the cold pressor test. Care must be taken though when comparing the spectra as two different matrices were used which may influence which phosphopeptides are better ionized.

83

Figure 3-13 MALDI-MS spectra of phosphopeptides isolated using the Fe-NTA Phosphopeptide Enrichment Kit. (A) Pooled saliva collected prior to the CPT (B) pooled saliva collected immediately after the CPT (C) pooled saliva collected 20 min after the CPT For comparison, the enriched phosphopeptides from each time point were separated and identified by nanoHPLC-MALDI-MS/MS. An overlay of the nanoHPLC chromatograms is depicted in Figure 3-14. There are several similarities and differences between the samples. For instance, all three samples have peaks at approximately 25,

26.7 and 29.7 minutes. However, the absorbances of these peaks are different between the samples. The highest absorbance for the peaks at ~25 and 29.7 minutes was observed for the fraction from saliva immediately after the CPT. Conversely, the highest absorbance for the peak at ~26.7 minutes is from the fraction of saliva collected prior to the CPT, and the other two saliva samples had a relatively low absorbance. This may indicate that the peptide is present in all the samples but may change in abundance in response to the cold stressor. 84

Figure 3-14 nanoHPLC chromatogram overlay of the salivary phosphopeptides isolated using the Fe-NTA Phosphopeptide Enrichment Kit from saliva collected prior to, immediately after and 20 min after the CPT

The eluting peptides were co-spotted with CHCA matrix on a MALDI

Anchorchip plate for MALDI-MS/MS analysis. A database search was performed using

MASCOT for peptide identification. Figure 3-15 is the nanoHPLC chromatogram of the phosphopeptide enriched saliva collected immediately after the CPT with the major peaks identified. Overall, two proteins were readily identified, alpha-amylase and salivary acidic proline-rich phosphopeptide 1/2. Salivary acidic proline-rich phosphopeptide ½ is involved in the inhibition of hydroxyapatite formation and form part of the dental pellicle.85 The peptides originating from alpha-amylase were not shown to be phosphorylated. They were likely present due to the large abundance of alpha-amylase in saliva. Several related peptides originating from salivary acidic proline-rich 85

phosphopeptide 1/2 were observed. For example, the peak with the largest absorbance at

~27 minutes has the sequence VISDGGDS(Phospho)EQFIDEER. Most of the other identified peptides result from truncations from the N-terminal end of salivary acidic proline-rich phosphopeptide 1/2. A complete list of identified phosphopeptides from salivary acidic proline-rich phosphopeptide ½ is provided in Table 3.3. MASCOT scores greater than 46 indicate identity or extensive homology, and scores greater than 31 indicate significant homology. Some of the peptides were also identified using de novo sequencing. Most of the peptides were monophosphorylated, but a few of them were diphosphorylated.

86

Figure 3-15 nanoHPLC chromatogram of saliva collected immediately after the CPT. Peptides were identified by MALDI-MS/MS

87

Table 3.3 List of Mascot search identified CPT salivary phosphopeptides originating from salivary acidic proline-rich phosphopeptide ½ (Uniprot # P02810).

m/z Sequence Mascot Score Saliva Sample 1232.5 pSEQFIDEER dn B, C 1347.5 DpSEQFIDEER 46 B, C 1404.5 GDpSEQFIDEER 67 B, C 1461.5 GGDpSEQFIDEER 83 B, C 1576.6 DGGDpSEQFIDEER 98 B,C 1746.6 pSDGGDpSEQFIDEER dn C 1856.7 IpSDGGDpSEQFIDEER dn B,C 1875.7 VISDGGDpSEQFIDEER 89 B,C 1955.7 VIpSDGGDpSEQFIDEER dn B,C 2184.9 VPLVISDGGDpSEQFIDEER 43 B 2300.1 DVPLVIpSDGGDSEQFIDEER 37 B dn denotes peptides determined by de novo sequencing B: saliva collected immediately after the CPT C: saliva collected 20 minutes after the CPT

3.4 Conclusions and Future Directions

A variety of methods were used to examine whole saliva with respect to the cold pressor test. Initial studies looked at global changes using gel-based techniques (i.e.,

SDS-PAGE, 2D-PAGE and DIGE) to separate the proteins. Less than ten bands corresponding to the most abundant proteins were observed using SDS-PAGE. To name a few, these bands contained alpha-amylase, carbonic anhydrase 6, cystatin type S family, keratins, lysozyme C and serum albumin. A greater number of protein spots were observed using 2D-PAGE. Consequently, 2D-PAGE provided better separation of the salivary proteins. However, further optimization is necessary to further increase the number of distinct spots. Major parameters that could be further optimized include the amount of protein applied to the IPG strips, the use active loading of the salivary proteins prior to isoelectric focusing and the use of narrower range IPG strips. In total, 37 salivary

88

proteins were identified by peptide mass fingerprinting and MALDI-MS/MS from the

SDS-PAGE and 2D-PAGE experiments.

Preliminary relative quantification was also performed using 2D-DIGE for the comparison of saliva collected prior to the CPT and immediately after the CPT. Some differences in abundance are possible but further studies need to be performed. Further experiments include using 2D-DIGE for the comparison of saliva collected prior to the

CPT to 20 minutes after the CPT and of saliva collected immediately after the CPT to 20 min after the CPT. Proper spot alignment between the images is necessary before densitometry can be performed for relative quantification.

Salivary proteins were also separated by HPLC. Overall, the fractionation of the salivary proteins using a C4 column did not separate the proteins as well as expected. It also resulted in relatively few proteins identified. Less than 20 proteins were identified by this method. The HPLC chromatogram indicated that some proteins may increase in abundance in response to the cold stressor, which is in agreement with results obtained by

SDS-PAGE. Alternatively, whole saliva samples that were digested with trypsin prior to separation yielded better separations and more identifications. Using MALDI-MS/MS for identification more than 35 proteins were identified. However, the method that yielded the most identifications was nanoHPLC-ESI-MS/MS of the whole saliva digest in which at least 100 salivary proteins were identified from a single separation.

As a method to remove alpha-amylase from the saliva samples, a potato starch column was used. Analysis of the samples using both SDS-PAGE and nanoHPLC-

MALDI-MS/MS confirmed that the majority of amylase was depleted from the sample.

89

The method proved to be quick, easy and inexpensive. However, the implications of depleting alpha-amylase were not fully studied here. Additional studies should be performed using these samples in order to ascertain if any low abundance salivary proteins may be changing in response to the stress.

Salivary phosphoproteins and phosphopeptides were also examined in this study.

Some phosphoproteins, such as cystatin, were selectively visualized after SDS-PAGE separation using Pro-Q Diamond phosphoprotein gel stain. Visually, no major differences in phosphorylation between the three samples were observed. Furthermore, tryptic phosphopeptides were isolated using the Fe-NTA Phosphopeptide Enrichment kit and were further analyzed using nanoHPLC-MALDI-MS/MS. This enrichment of the salivary phosphopeptides resulted in the identification of eleven phosphopeptides matching two phosphorylated serine residues of the salivary acidic proline-rich phosphopeptide ½.

Examination of the MALDI-MS and the nanoHPLC chromatogram of the enriched fractions from each time point indicate that differences in the abundance of common phosphopeptides between the samples may exist. It also indicates that there may be different phosphopeptides present in the saliva collected before the cold pressor test in comparison to the two saliva samples collected after the cold pressor test.

Future work includes the relative quantification of the salivary proteins and peptides using label-free mass spectrometry. To perform this, equal volumes of the salivary digests will be separated and identified by nanoHPLC-ESI-MS/MS. Thermo’s

SIEVE software will be used to align the peaks in the chromatograms of each sample to correct for any run to run variances. It also performs statistical analysis of their relative

90

abundances from the resulting mass spectra to determine if the peptides are present in differing amounts at the different time points. This should further enable the identification of a putative biomarker of stress that can be validated in future studies.

Additional studies will also be conducted using Thermo’s ESI-Orbitrap Fusion to study the phosphopeptides. It will be coupled to the nanoHPLC to enable high throughput analysis for phosphoprotein identification. The Orbitrap Fusion is equipped with ETD fragmentation which will aide in the identification of phosphorylation sites. In contrast to the MALDI-MS/MS spectra, ETD does not break the bond between the amino acid and the phosphate group. Consequently, the residue remains phosphorylated, and its shift in mass can be used to distinguish it as being phosphorylated. This method may enable the identification of novel sites of phosphorylation in response to the cold stress. It can also be used in conjunction with label-free mass spectrometry for the relative quantification of the phosphopeptides for biomarker discovery.

91

Chapter 4

A Preliminary Investigation of Changes in the Salivary

Proteome in Response to Acute Stress in Medical

Residents Performing Advanced Clinical Simulations

4.1 Introduction

As discussed in the previous chapter, saliva is an optimal body fluid for biomarker discovery and for monitoring diseases or psychological states. For example, salivary cortisol has been shown to be a biomarker of chronic stress.52, 86, 87 Various demanding occupations (i.e., military and medical fields) may experience other types of stress, such as acute stress, which may be beneficial to monitor using biomarkers. Some preliminary studies have begun to address this issue in which they have detected high levels of stress in specific occupational situations. For example, McGraw et al. monitored saliva in Army nurses performing a combat casualty scenario and observed an increase in cortisol and alpha-amylase in response to the stress.88 Strahler and Ziegert used a similar paradigm in which police officers performed a simulation responding to a school shooting. They noted that the officers experienced a decrease in salivary cortisol after completion of the simulation, and salivary alpha-amylase was at its highest level after completing the 92

simulation.89 Groer et al. also utilized simulations with police officers in which they responded to different incident scenarios.90 In response to the simulated incidents, the officers experienced an increase in cortisol, alpha-amylase, interleukin-6 and secretory

IgA in saliva. Another study utilized teams of paramedics and EMS physicians that performed emergency scenarios. They experienced a decrease in salivary cortisol and an increase in salivary alpha-amylase during the scenarios.91 However, all of these studies used a targeted approach in which they used assays to monitor only a few specific analytes.

Instead, this study offers an untargeted proteomics approach to analyze the dynamically changing saliva obtained from medical residents performing emergency medicine simulations to discover potential biomarkers of acute stress. The advantage of using simulations is that they have been shown to create realistic psychological challenges that mimic performance in a real situation with a high level of fidelity.92, 93

These simulations use computerized mannequins in a simulated hospital setting. The mannequins exhibit realistic features and can appropriately respond to the medical residents’ interventions. This type of scenario will place the residents into a brief stressed condition.

4.2 Experimental

4.2.1 Chemicals and Materials

Glacial acetic acid, ammonium persulfate and HPLC-grade acetonitrile were purchased from Fisher Scientific (Pittsburgh, PA). Ammonium bicarbonate, 2- mercaptoethanol, dithiothreitol, formic acid (FA), iodoacetamide, trifluoroacetic acid

93

(TFA), trypsin from bovine pancreas and LC-MS grade water were obtained from Sigma

(St. Louis, MO). Bio-Safe Coomassie Brilliant Blue G-250, 10% SDS, 2x Laemmli sample buffer, 30% acrylamide/bis solution (29:1), 1.5 M Tris-HCl (pH 8.8), 0.5 M Tris-

HCl (pH 6.8) and TEMED were from Bio-Rad (Hercules, CA). Synthetic histatin-3 was from Genemed Synthesis Inc. (San Antonio, TX).

4.2.2 Ethics Approval

The collection of saliva and this study were approved by the University of Toledo

Biomedical Institutional Review Board. Consent was also obtained from all participants of this study.

4.2.3 Emergency Medicine Simulations and Saliva Collection

Eight first-year Emergency Medicine residents were familiarized during their first month of their program with the computerized mannequins and the simulation process using basic, uncomplicated cases that did not require any urgency. Afterwards, the residents provided an assessment of their performance, and a faculty member provided targeted education. Conversely, the simulations used in the studies were more realistic and stressful as they replicated a true emergency. The ―patients‖ experienced a crisis of airway, breathing and/or circulation that required immediate management. The simulations included emergencies such as seizures, traumatic shock, tension pneumothorax and an overdose. Compared to the previous simulations, these were new to the medical residues and required them to prioritize multiple complications that necessitated immediate correction. A ―family member‖ was also involved which added

94

additional complexities and complications to patient management by the medical residents.

Saliva was collected from the eight first-year medical residents (four female and four male, Table 4.1). The passive drool method was used to collect whole saliva, and the residents were instructed to refrain from eating and drinking for at least 20 minutes before sample collection. During the study, one of the residents was fasting from sunrise to sunset. Three saliva samples were collected from each resident over the course of the study. The first sample was collected immediately prior to performing the emergency medicine simulation around 9:00 am at the Immersive Interdisciplinary Simulation

Center at the University of Toledo Medical Center. Shortly after completion of the simulation, the second saliva sample was collected. This was within three hours of collecting the first saliva sample. The third saliva sample was collected the next morning upon waking before eating, drinking or brushing their teeth. An overview of the saliva collection timeline is given in Figure 4-1.

Table 4.1 Demographics of the first-year medical residents participating in the study

Resident Gender Age Post-Simulation a Wake a 1 Male 26 11:18 AM 6:45 AM 2 Female 24 9:51 AM 7:00 AM 3 Female 29 10:57 AM 6:50 AM 4 Female 28 9:51 AM 6:50 AM 5 Male 28 11:19 AM 6:45 AM 6 Female 28 10:56 AM 6:00 AM 7 Male 27 11:48 AM 6:40 AM 8* Male 30 11:47 AM 7:30 AM * Fasting from sunrise to sunset a Refers to time at which saliva was collected

95

Figure 4-1 Overview of the timeline of saliva collection and the emergency medicine simulation

Saliva samples collected prior to the simulation were stored on ice until obtainment of the second saliva sample after completion of the simulation. Afterwards, the saliva was transported back to the lab on ice where it was aliquoted to avoid multiple freeze-thaw cycles prior to being stored at -80 °C. The samples were thawed and centrifuged at 986 x g for 15 minutes before analysis to pellet any debris present.

4.2.4 SDS-PAGE

For SDS-PAGE saliva samples were prepared in two comparable manners using equal volumes of saliva from each time point. In one manner, 30 μL of the saliva sample was combined with 30 μL of sample buffer consisting of 29.25 μL of Laemmli sample buffer and 0.75 μL of 2-mercaptoethanol. From the total sample volume of 60 μL, 25 μL was loaded on 12% polyacrylamide gels. Alternatively, 20 μL of the saliva sample was combined with 10 μL of the sample buffer and the entirety was loaded on the gel for increased sensitivity of the low-molecular weight band. The 12% polyacrylamide gels were run at 120 V for approximately 1.5 hours. Bio-Safe Coomassie Brilliant Blue G-250 was used to visualize the separated proteins, and the gels were scanned at the maximum resolution (1200 dpi) using a desktop scanner. Samples prepared in either manner were

96

analyzed in duplicate to ensure reproducibility. Any gels that showed protein migration into adjacent lanes were omitted for densitometry and statistical analyses.

4.2.5 In-gel digestion and peptide mass fingerprinting

In-gel digestion of the protein bands was performed as described in Section

3.2.3.4 for peptide mass fingerprinting. MALDI-MS was performed using Bruker’s

MALDI-TOF/TOF UltrafleXtreme. The mass spectra were acquired from m/z 600 to m/z

3500 in reflectron and positive ion modes using FlexControl software (version 3.4,

Bruker Daltonics). A standard peptide calibration mixture from Bruker was used to calibrate the instrument prior to data acquisition. Using the dried droplet method, a 1 μL aliquot of the digest was co-spotted with 1 μL of 10 mg/mL CHCA matrix in 60:40 (v:v) acetonitrile: 0.1% TFA onto a MTP 384 ground steel target plate. The mass spectra were analyzed using FlexAnalysis software, and a MASCOT (Matrix Science) search of the

SwissProt Homo sapiens database was performed.14 Typically one missed cleavage was permitted; however, two missed cleavages were allowed for bands with a molecular weight less than 10 kDa. Methionine oxidation was set as a variable modification with a mass tolerance of 0.2 Da. A score equal to or greater than 56 was set as the threshold for confident protein identification which corresponds to p < 0.05. As in the previous chapter, verification of the salivary proteins was performed by searching the UCLA

Dental Research Institute’s Salivary Proteome Knowledge Base.69, 70

4.2.6 Statistical Analysis

The stained protein bands of interest were subjected to densitometric analysis using ImageJ (NIH). The 50 kDa molecular weight ladder band was used for

97

normalization to account for gel to gel differences. Specifically, the area of each protein band was divided by both the volume of the sample (i.e., 25 or 30 μL) and the area of the

50 kDa band. Ratios of the normalized protein band areas were determined for the time points (i.e., post-emergency medicine simulation to wake, post-emergency medicine simulation to pre-emergency medicine simulation and wake to pre-emergency medicine simulation). The average ratio for the multiple runs was calculated. This average ratio was subjected the t-test and the Wilcoxon signed-rank test using Minitab (Minitab Inc.) software in order to determine if significant differences existed between the relative protein abundances at the three time points.

4.2.7 Whole saliva digestion

Without prior separation, the salivary proteins from the three collection time points were digested with trypsin. First, the disulfide bonds were reduced by combining

1.25 μL of 200 mM dithiothreitol with a 10 μL aliquot of saliva and 7.5 μL of 100 mM ammonium bicarbonate. The sample was incubated for one hour at 60 °C. Then, the cysteine residues were alkylated by adding 1.25 μL of 200 mM iodoacetamide. The sample was incubated at room temperature for one hour in the dark. Afterwards, 1.25 μL of 200 mM dithiothreitol was added to the sample with incubation in the dark for one hour at room temperature. Then, 150 μL of water, 50 μL of 100 mM ammonium bicarbonate and 0.9 μg of trypsin in 10 μL of 50 mM acetic acid were added prior to incubation at 37 °C overnight (~16 hours). The digestion was quenched by the addition of

2 μL of TFA.

98

4.2.8 nanoHPLC-ESI-MS/MS

As a secondary method for protein identification, peptides were extracted from the digested gel bands used for PMF. These peptides were analyzed by nanoHPLC-ESI-

MS/MS using Thermo’s ESI-Orbitrap Fusion Tribrid mass spectrometer. The tryptic peptides were extracted as described by Shevchenko et al.72 The tryptic peptides from the whole saliva digestion were also analyzed by nanoHPLC-ESI-MS/MS. A Dionex

Ultimate 3000 HPLC RSLCnano system equipped with an Acclaim PepMap RSLC C18 column (75 μm x 15 cm, 2 μm particles, 100 Å pore size) was used to separate the peptides using a flow rate of 300 nL/min. The column oven was maintained at 35 °C.

Mobile phase A was 0.1% FA in water. Mobile phase B was 0.08% FA in 80:20 (v:v) acetonitrile: water. The samples were loaded onto a precolumn using 0.05% TFA in 98:2

(v:v) water: acetonitrile where they were desalted for five minutes. During this time, the

C18 column was held at 4% B. Depending on the sample, two different gradients were utilized. A linear gradient from 4-55% B in 60 minutes followed by a ramp to 100% B in

0.5 minutes which was maintained for 10 minutes for column washing followed by re- equilibration at 4% B was used to separate the peptides extracted from the SDS-PAGE bands. For the tryptic peptides from the whole saliva digestion, this gradient was extended. A linear gradient of 4-55% B in 120 minutes was used followed for column washing and re-equilibration.

A UV-Vis nanoHPLC detector at 214 nm was used to detect the eluting peptides.

The ESI-Orbitrap Fusion was operated in data-dependent mode to analyze the eluting peptides. A positive spray voltage of 1800 V, and an ion transfer tube temperature of

275 °C were used. Sheath and auxiliary gases were not used. Full MS scans were 99

acquired from m/z 400 to m/z 1600 using Orbitrap detection at a resolution of 120,000.

MS/MS selection was performed for charge states 2+ to 4+, and ions with an intensity of at least 5.0e3. The most intense ions were selected first for fragmentation. CID was performed in the linear ion trap, and the resulting MS/MS were subjected to a SEQUEST

HT search of the SwissProt database of Homo sapiens tryptic peptides using Proteome

Discoverer (version 1.4, Thermo). Two missed cleavages were permitted. The variable modifications for the in-gel digested peptides included methionine oxidation and phosphorylation of serine, threonine or tyrosine. For the whole saliva trypsin digested peptides, carbamidomethylation of cysteine was also set as a fixed modification. The precursor mass tolerance was 10 ppm, and the fragment mass tolerance was 0.6 Da. The results were run through percolator with a relaxed target false discovery rate of 0.05 and a strict false discovery rate of 0.01.73 As with peptide mass fingerprinting, the identifications were confirmed as known salivary proteins using UCLA Dental Research

Institute’s Salivary Proteome Knowledge Base.69, 70

4.2.9 Analysis of commercial hisatin-3

For confirmation studies, a commercially obtained histatin-3 protein was used for comparison with salivary histatin-3. This synthetic protein was digested with trypsin in solution omitting the reduction and alkylation steps as histatin-3 does not contain any cysteine residues. A 10 μL aliquot of 1 μg/μL of histatin-3 in DI water was combined with 50 μL of 100 mM ammonium bicarbonate and 100 ng of trypsin in 20 μL of 50 mM acetic acid. As with the whole saliva digestion, the sample was incubated overnight at

37 °C, and the digestion was halted by the addition of 2 μL of TFA.

100

Alternatively, 300 ng of synthetic histatin-3 was loaded on a 12% polyacrylamide gel with saliva samples collected from a medical resident. After visualization, it was subjected to the aforementioned in-gel digest. Both histatin-3 digests were analyzed by

MALDI-MS using the same conditions described in Section 4.2.5 for peptide mass fingerprinting. In addition, the samples were analyzed by nanoHPLC-ESI-MS/MS as described in the previous section.

4.3 Results and Discussion

4.3.1 SDS-PAGE and peptide mass fingerprinting

Equal volumes of saliva samples for the three time points from each of the medical residents were separated using 12% polyacrylamide SDS-PAGE gels. The decision to use equal volumes of saliva was made in light of future potential clinical applications. Specifically, Millea et al. noted that use of equal volumes of saliva is more practical because many clinical laboratories do not perform protein assays prior to analysis.47 The SDS-PAGE gel images from the four male medical residents are shown in

Figure 4-2. Some inter-individual variability can be observed between the residents.

Similar results for the female medical residents were observed.

101

Figure 4-2 Images of the SDS-PAGE separation of the salivary proteins obtained from (A) resident 1, (B) resident 5, (C) resident 7 and (D) resident 8, the four male medical residents, collected upon waking, prior to the simulation (pre) and after the simulation (post).

The protein bands were excised and subjected to in-gel digestion. Peptide mass fingerprinting was performed for protein identification. A representative gel image labeled with the most prevalent identified proteins is shown in Figure 4-3. After staining with Coomassie blue, a relatively low number of bands were detected. However, this is in agreement with a previous study that detected approximately ten protein bands from whole saliva using SDS-PAGE.75 The most prominent band on the gels was identified as alpha-amylase. This was expected as approximately 60% of saliva is comprised of alpha- amylase.51, 74 In addition, several high-molecular weight proteins were identified. These

102

proteins include mucin-7 and polymeric immunoglobulin receptor. Furthermore, another prominent group of proteins was identified as the cystatin type S family (cystatin-S, cystatin-SN and cystatin-SA). The low-molecular weight (< 10 kDa) band was identified by peptide mass fingerprinting as containing histatin-3. The complete list of proteins identified by peptide mass fingerprinting is provided in Table 4.2.

Figure 4-3 A representative SDS-PAGE gel labeled with the most prevalent proteins identified by PMF

103

Table 4.2 Peptide mass fingerprinting identified salivary proteins

Accession Protein Name Mass Score a No. Seq. Residents (kDa) Peptides b Cov. c P02768 Serum albumin 66.5 186 26 34% 1 to 7 P04745 Alpha-amylase 1 55.9 314 34 62% 1 to 8 P19961 Alpha-amylase 2B 55.9 304 34 59% 1 to 8 P04746 Pancreatic alpha-amylase 55.9 242 30 50% 1 to 8 P01037 Cystatin-SN 14.3 141 14 76% 1 to 8 P01036 Cystatin-S 14.1 155 13 92% 1 to 8 P09228 Cystatin-SA 14.3 86 11 53% 1,3,4,5,8 P15516 Histatin-3 4.1 103 6 84% 1,2,4,7 P01876 Ig alpha-1 chain C region 37.6 66 12 24% 2,5,8 P01877 Ig alpha-2 chain C region 36.5 67 11 22% 2,5,7,8 O43790 Keratin, type II cuticular 53.5 72 13 28% 5 Hb6 P61626 Lysozyme C 14.7 65 9 45% 2,4 Q8TAX7 Mucin-7 36.8 74 10 21% 2,4 P01833 Polymeric immunoglobulin 81.3 94 16 21% 3,5 receptor P12273 Prolactin-inducible protein 13.5 87 7 57% 7 P18206 Vinculin 123.6 73 18 21% 1,4,6 a Protein scores greater than 56 are significant (p<0.05); highest score for digests given b Number of peptides in the mass spectrum matching the protein c Residents in which the protein was identified

Alpha-amylase is responsible for initiating digestion by hydrolyzing α-1,4- glycosidic bonds in starch.52 Salivary alpha-amylase also exists in several glycosylated forms which may affect its detection.44, 51 In addition, salivary alpha-amylase exhibits a circadian rhythm. Its maximum is in the late afternoon, and the minimum is in the morning.52 Therefore, comparisons of the relative amounts of alpha-amylase should be performed during the same part of the day to avoid influences of the circadian rhythm on its abundance.

104

Peptide mass fingerprinting of the ~26 kDa band was unsuccessful for protein identification. However, from previous studies it can be proposed that it may contain proline-rich proteins as it is one of the most abundant classes of salivary proteins. In addition, this group of proteins is challenging to identify by trypsin digestion due to its large number of proline residues.58, 94 If sufficient sample amounts were available the presence of this protein class could be confirmed using Coomassie R-250 instead of

Coomassie G-250 staining. Coomassie R-250 will stain proline-rich proteins a pink-violet color when organic solvent is omitted from the destaining solution, and the other proteins are stained blue.57, 58, 94, 95 In comparison with previous studies, the band may contain a basic proline-rich protein with a mass of 23,462 Da. While this protein is slightly smaller than the calculated mass of the protein band, it has been shown that proline-rich proteins have a tendency to migrate slower than expected.55, 95

Another prominent band detected contained cystatin-S, -SN and –SA. As members of the cystatin family they are cysteine protease inhibitors with a mass of 13 to

14 kDa, and they contain two conserved disulfide bonds.58, 83, 84, 96-98 The cystatin family also regulates salivary calcium levels and has antimicrobial activity.98 Cystatin-S can be further classified as cystatin S1 or S2 depending on its post-translational phosphorylation.

Cystatin S1 is monophosphorylated at Ser23, whereas cystatin S2 is phosphorylated at

Ser21 and Ser23.83, 84

Peptide mass fingerprinting of the low-molecular weight (<10 kDa) band resulted in the identification of histatin-3 which is a small, 32- amino acid protein. Two missed cleavages were permitted in the MASCOT search. This seems reasonable due to the small

105

size of the protein. In addition, histatin-3 has multiple consecutive trypsin cleavage sites in its amino acid sequence (DSHAKRHHGY KRKFHEKHHS HRGYRSNYLY DN).59,

99, 100 Histatin-3 has been shown to have antibacterial and antifungal activity in which it is particularly effective against C. neoformans, C. albicans and P. gingivalis.101-103

4.3.2 Densitometry and statistical analyses

Relative protein quantification of four bands was performed using ImageJ. The bands were alpha-amylase, a ~26 kDa protein band, cystatin type-S family and the low- molecular weight band. The scanned gel images were imported into ImageJ to measure the area of each band. These band areas were normalized to account for gel to gel differences. The raw and normalized band areas are provided in Appendix B. For each medical resident, an average of the ratios of the normalized band areas were determined for the post-simulation to wake, post-simulation to pre-simulation and wake to pre- simulation time points (Table 4.3). The ratios from the 25 μL sample SDS-PAGE runs were averaged for the alpha-amylase, ~26 kDa (likely containing a proline-rich protein) and cystatins bands. However, only the second run using the 30 μL sample was used for the low-molecular weight band as the band was not detected in the other samples.

106

Table 4.3 Average of the ratios of the relative amounts of the four protein bands subjected to densitometric analysis

Resident Amylase Amylase Amylase ~26 kDa ~26 kDa ~26 kDa Post to Wake Post to Pre Wake to Pre Post to Wake Post to Pre Wake to Pre 1 2.1446 2.0135 0.9400 2.2469 1.7804 0.8391 2 1.0268 1.8880 1.9219 0.8680 1.5211 1.7683 3 2.5234 1.5316 0.6071 1.8958 1.5853 0.8355 4 4.8052 1.3706 0.7672 0.8096 1.5369 1.9840 5 1.9183 2.1904 1.1583 0.8620 1.5594 1.8167 6 1.8584 2.1255 1.1531 2.0945 1.7541 0.8390 7 2.3215 2.1428 0.9747 1.3102 1.5487 1.3868 8 0.5630 0.8751 1.7154 0.2732 0.8357 3.6840 Resident Cystatin Cystatin Cystatin Low MW Low MW Low MW Post to Wake Post to Pre Wake to Pre Post to Wake Post to Pre Wake to Pre 1 1.5987 2.4874 1.4570 3.1233 10.8730 3.4813 2 1.4099 1.6461 1.1809 3.8487 2.7068 0.7033 3 2.3810 1.6226 0.6820 n/a n/a n/a 4 2.7556 1.6756 0.6081 3.8756 1.1123 0.2870 5 1.5335 1.7915 1.1699 n/a n/a n/a 6 4.0645 2.4566 0.6044 n/a n/a n/a 7 1.8025 1.7346 0.9878 3.7961 1.8440 0.4858 8 0.9643 1.6922 1.7325 1.2159 0.9492 0.7807

Amylase: alpha-amylase protein band Post: saliva collected after the emergency medicine simulation Wake: saliva collected upon waking the morning after the emergency medicine simulation Pre: saliva collected prior to the emergency medicine simulation n/a: band absent or not sufficiently stained

In order to evaluate the distribution of the ratios of the protein bands at the various time points, boxplots were generated using Minitab (Figure 4-4). The relative protein abundance after the emergency medicine simulation (post-simulation) increased compared to the waking and pre-simulation time points as indicated by ratios greater than one. Whereas, the ratio of the relative protein abundances at the waking and pre-

107

simulation time points was approximately one, indicating that no major difference between the samples exists. To determine if significant differences in the relative protein amount existed in the saliva collected after the simulation, both the parametric t-test and the non-parametric Wilcoxon signed-rank test were used.

108

Figure 4-4 Boxplots of the relative ratios of alpha-amylase, the ~26 kDa, cystatins, and the low-molecular weight protein bands for the ratio of the protein band areas (A) after the simulation compared to waking the morning after the simulation, (B) after the simulation compared to before the simulation and (C) waking the morning after the simulation compared to before the simulation.

109

4.3.2.1 Results of the t-test

The one-sample t-test was used to determine if the mean ratios of the peak areas of alpha-amylase, cystatins, the ~ 26 kDa band and the low-molecular weight band were significantly different from one comparing the stressed and non-stressed time points

(post-simulation vs. wake, post-simulation vs. pre-simulation and wake vs. pre- simulation). A summary of the calculated p-values and the 95% confidence intervals for these ratios of the protein abundances are summarized in Table 4.4.

Table 4.4 Summary of the relative quantification of the protein bands at the three time points as determined by the t-test

Alpha-amylase ~26 kDa

Post to Wake Post to Pre Wake to Pre Post to Wake Post to Pre Wake to Pre p-value 0.0371 0.0024 0.3654 0.2800 0.0016 0.0971

95% confidence 1.09 to 3.20 1.38 to 2.16 0.78 to 1.53 0.70 to 1.89 1.27 to 1.76 0.85 to 2.44 interval

Cystatin Low-molecular weight

Post to Wake Post to Pre Wake to Pre Post to Wake Post to Pre Wake to Pre p-value 0.0184 0.0002 0.7284 0.0129 0.2527 0.8147

95% confidence 1.24 to 2.89 1.58 to 2.19 0.71 to 1.40 1.76 to 4.58 -1.669 to 8.68 -0.49 to 2.79 interval

For the alpha-amylase band, both the ratio of the post-emergency medicine simulation to the waking time points and the ratio of the post-emergency medicine simulation to the pre-emergency medicine simulation were significantly different as the p-values determined by the t-test were less than 0.05 (0.0371 and 0.0024, respectively).

Furthermore, their 95% confidence intervals were further confirmation that the mean ratios for these time points were greater than one. Specifically, the 95% confidence interval of the ratio of the post-simulation to waking time point was from 1.09 to 3.20, and it was 1.38 to 2.16 for the ratio of the post-simulation to the pre-simulation time 110

points. Consequently, the one-sample t-test indicates that there was a significant increase in salivary alpha-amylase after the acute stressor.

For the ~26 kDa band, an insignificant p-value (0.2800) was observed for the average of the ratio of the post-emergency medicine simulation to waking time points. In addition, the 95% confidence interval (0.70-1.89) further indicated that there was not a significant difference between these time points. However, the average ratio of the band areas of the post-simulation to pre-simulation time points was significantly different (p- value of 0.0016). In addition, the 95% confidence interval was greater than one for these time points indicating that the abundance of the ~26 kDa band increased after the acute stressor in comparison to the pre-simulation time point.

For the cystatin band, a significant difference in the average band area ratios for the post-simulation to waking time points and for the post-emergency medicine simulation to the pre-simulation time points was observed (p-values of 0.0184 and

0.0002, respectively). In addition, their 95% confidence intervals indicated that their mean ratios were greater than one (1.24 to 2.89 and 1.58 to 2.19, respectively). Therefore, the t-test indicated that the relative abundance of the salivary cystatins increased after the acute stressor as also observed for alpha-amylase.

For the low-molecular weight band, only the ratio of the band areas for the post- simulation to the waking time points was significant (p-value of 0.0129). Its 95% confidence interval was from 1.76 to 4.58, which is further indication that its abundance increased after the simulation. However, neither the ratio of the band area of the post- simulation to pre-simulation time points nor the waking to pre-simulation time points

111

were significant (0.252 and 0.8147, respectively). In addition, the 95% confidence interval for the post-simulation to pre-simulation time points ranged from -1.69 to 8.68.

This indicated that there was a large amount of variability in the low-molecular weight band, as it is impossible for the ratio of protein abundances to be negative. For the four protein bands analyzed, the ratio of the waking to the pre-emergency medicine simulation time points was not significant based on their p-values and 95% confidence intervals.

Consequently, no major difference in their protein amounts was observed between the two non-stressed time points.

4.3.2.2 Results of the Wilcoxon signed-rank test

In a comparable manner, the non-parametric Wilcoxon singed-rank test was used for determining if the median ratios were significantly different from one comparing the three time points of the alpha-amylase, cystatins, the ~26 kDa and the low-molecular weight bands. A summary of their p-values and 95% confidence intervals is provided in

Table 4.5. In general, the results were in agreement with the results of the t-test. In particular, the median band area ratios for both the post-emergency medicine simulation to the waking time points and the post-emergency medicine simulation to pre-emergency medicine simulation time points for the alpha-amylase band had significant p-values

(0.0234 and 0.0156, respectively). Their 95% confidence intervals also indicated that their medians were greater than one (1.21 to 3.47 and 1.37 to 2.14, respectively).

112

Table 4.5 Summary of the relative quantification of the protein bands determined by Wilcoxon signed-rank test

Alpha-amylase ~26 kDa

Post to Wake Post to Pre Wake to Pre Post to Post to Pre Wake to Pre

Wake p-value 0.0234 0.0156 0.6406 0.3828 0.0156 0.1094

95% confidence 1.21 to 3.47 1.37 to 2.14 0.77 to 1.54 0.57 to 2.07 1.19 to 1.68 0.83 to 2.72 interval

Cystatin Low-molecular weight

Post to Wake Post to Pre Wake to Pre Post to Post to Pre Wake to Pre Wake p-value 0.0156 0.0078 0.8438 0.0625 0.1250 0.6250

95% confidence 1.28 to 2.93 1.65 to 2.14 0.64 to 1.46 1.22 to 3.88 0.95 to 0.28 to 3.48 interval 10.87

In agreement with the t-test, an insignificant p-value (0.3828) was observed for median ratios of the ~26 kDa band areas of the post-simulation to waking time points, and the 95% confidence interval (0.57 to 2.07) further indicated no significant difference existed. Whereas, a significant p-value (0.0156) for the post-emergency medicine simulation to the pre-emergency medicine simulation time points was observed. The 95% confidence interval (1.19 to 1.68) further indicated that the median ratio of the abundance of the ~26 kDa protein band was greater than one, as in the case of the t-test.

The results of the Wilcoxon signed-rank test for the cystatin band were also consistent with the t-test results. The p-values of the median ratio of the post-simulation to waking time points and the post-simulation to pre-simulation time points were significant (0.0156 and 0.0078, respectively). Furthermore, their 95% confidence intervals were greater than one (1.28 to 2.93 and 1.65 to 2.14, respectively). Therefore, the acute stress resulted in an increase in the relative abundance of cystatins.

113

The results for the low-molecular weight band were in partial agreement with the t-test results. Insignificant p-values for the ratio of the band areas for the post-simulation to waking time points and the post-simulation to pre-simulation time points were observed (0.0625 and 0.1250, respectively). Furthermore, their respective 95% confidence intervals ranged from 1.22 to 3.88 and 0.95 to 10.87. The latter is in agreement with the t-test as higher variability exists. This increased variability may be a consequence of the limited number of samples (n=5). Also in agreement with the t-test of all four protein bands, the ratio of the waking to the pre-emergency medicine simulation time points was not significant meaning that no major difference in protein amount existed between the non-stressed time points.

4.3.3 Further analyses of the low-molecular band

As confirmation of the presence of histatin-3 detected in the low-molecular weight band of the medical residents’ saliva, a solution of synthetic histatin-3 was digested and analyzed by peptide mass fingerprinting. A positive identification of the protein was made as expected (score: 80; protein score >56 is significant). The MALDI mass spectra of the in-solution digest of the synthetic histatin-3 protein and the in-gel digest of the low-molecular weight (<10 kDa) protein band identified as histatin-3 by

PMF were similar (Figure 4-5A and B). The spectra had several peptide peaks in common (e.g., m/z 797.4, 953.5, 1214.6, 1342.7) which suggests that the low-molecular weight protein band from the medical resident saliva likely contains histatin-3.

114

Figure 4-5 MALDI-MS of the tryptic digest of (A) the 10 kDa band in-gel digest from medical resident saliva, (B) the solution digest of the synthetic histatin-3 and (C) the in- gel digest of the synthetic histatin-3. Common histatin-3 peaks are marked by a solid outline box, whereas histatin-3 peaks unique to each sample are marked by a dashed outline box.

Furthermore, the migration of the synthetic histatin-3 sample matched the migration of the salivary low-molecular weight band on a 12% polyacrylamide SDS-

PAGE gel as shown in Figure 4-6. This synthetic histatin-3 band was also subjected to peptide mass fingerprinting. A positive protein identification (score: 72; protein score >

115

56 is significant) resulted from the MASCOT database search as expected. Similar to the solution digest of the synthetic histatin-3, there were several common peptide peaks in the MALDI-MS of the in-gel digest of both the synthetic histatin-3 sample and the salivary histatin-3 (e.g., m/z 797, 953, 1214 and 1342) as shown in Figure 4-5A and C.

The matching migration and MALDI-MS spectra of the samples indicate that the low- molecular weight band from the medical resident saliva likely contains histatin-3.

Figure 4-6 Separation of 300 ng of synthetic histatin-3 and medical resident saliva collected upon waking, prior to the emergency medicine simulation (pre) and after the emergency medicine simulation (post)

4.3.4 nanoHPLC-ESI-MS/MS

Additional efforts at confirmation of the presence of histatin-3 in the low- molecular weight protein band were performed using nanoHPLC-ESI-MS/MS on the extracts of the in-gel digests of the salivary low-molecular weight and the synthetic histatin-3 SDS-PAGE bands. Database searching of the MS/MS spectra identified

116

proteins including cystatin, keratin, alpha-amylase and prolactin-inducible protein in the salivary low-molecular band. The complete listing of proteins identified with at least two unique peptides in at least two samples is provided in Table 4.6. Therefore, the low- molecular weight band may contain fragments from several larger salivary proteins.

Salivary histatin-3 was not identified in the band by nanoHPLC-ESI-MS/MS though.

Table 4.6 NanoHPLC-ESI-MS/MS identifications of salivary proteins co-separating in the low-molecular weight SDS-PAGE protein band

Accession Protein Name Score Seq. # Unique # MW Cov. Peptides Peptides (kDa) P01036 Cystatin-S 22.31 44.68% 4 5 14.1 P01037 Cystatin-SN 14.59 32.62% 2 3 14.3 P01833 Polymeric immunoglobulin receptor 5.9 8.51% 2 4 81.3 P04264 Keratin, type II cytoskeletal 1 121.2 52.80% 20 34 65.9 P04745 Alpha-amylase 1 30.69 33.46% 5 9 55.9 P07477 Trypsin-1 9.28 7.29% 2 2 24.1 P12273 Prolactin-inducible protein 16.13 38.36% 3 4 13.5

However, when the extracted peptides from the in-gel digest of the synthetic histatin-3 were analyzed by nanoHPLC-ESI-MS/MS, identification of histatin-3 was successful. Consequently, possible issues with the in-gel digestion, peptide extraction or analysis by nanoHPLC-ESI-MS/MS of the medical resident salivary histatin-3 from the amounts present in the gel may exist. A previous study using LC-MS/MS was capable of detecting histatin-3 and histatin-3 fragments from whole saliva that was not subjected to proteolytic cleavage.100 However, a similar study for additional confirmation of the presence of histatin-3 was not feasible in these analyses due to the limited saliva amounts available.

117

As peptide mass fingerprinting did not result in an identification of the ~26 kDa band, the in-gel digest of the band was analyzed by nanoHPLC-ESI-MS/MS in an attempt to determine its identity. Table 4.7 lists the proteins identified by the database search in at least two samples with at least two unique peptides including keratins, lipocalin-1 and prolactin-inducible protein. Therefore, this ~26 kDa band may contain several different proteins and keratin fragments.

Table 4.7 NanoHPLC-ESI-MS/MS identifications of salivary proteins co-separating in the ~26 kDa SDS-PAGE protein band

Accession Protein Name Score Seq. # Unique # MW Cov. Peptides Peptides (kDa) P04264 Keratin, type II cytoskeletal 1 16.29 23.91% 4 14 65.9 P12273 Prolactin-inducible protein 89.8 65.07% 6 9 13.5 P31025 Lipocalin-1 51.92 52.27% 8 10 17.4 Keratin, type II cytoskeletal 2 P35908 3.3 13.62% 2 6 6.5 epidermal

Tryptic peptides from the in-solution digest of whole saliva at the waking, pre- simulation and post-simulation time points were also subjected to nanoHPLC-ESI-

MS/MS. A total of 57 proteins were identified with at least two unique peptides.

Common proteins identified include alpha-amylase, cystatin-B, cystatin-C, cystatin-D, cystatin-S, cystatin-SA, cystatin-SN, histatin-1, mucin-7 and serum albumin. A complete listing is provided in Appendix C. Many of the identifications confirm the peptide mass fingerprinting identifications.

4.3.5 Additional Discussion

Salivary alpha-amylase has been shown in previous studies as a marker of stress.88-91, 104 The results from the t-test and Wilcoxon signed-rank test of the alpha-

118

amylase band confirm such findings. Furthermore, a previous study observed that the increase in salivary alpha-amylase due to stress is independent of flow rate.105 Therefore, salivary alpha-amylase may be used as a stress biomarker regardless of changes in the salivary flow rate throughout the day. Our results are also in agreement with previous studies that have indicated cystatin-S is a potential marker of acute stress.104, 106 Previous studies have also shown that cystatin-SA is a putative salivary biomarker of oral cancer.84

As the statistical analyses indicated that the low-molecular (<10 kDa) protein band changes in response to acute stress, confirmation of histatin-3 (identified by PMF) and fragments from larger salivary proteins (identified by nanoHPLC-ESI-MS/MS) is necessary. However, it should be considered that salivary histatin-3 also displays a circadian rhythm which parallels the saliva flow rate. Secretion of histatin-3 has a maximum in the late afternoon.99, 107 However, this circadian rhythm was not expected have a significant impact of the relative abundances in this study as a relatively short time period existed between collection of the pre- and post-emergency medicine simulation saliva samples, minimizing circadian rhythm influences.

As a preliminary proteomics study, some limitations were experienced.

Specifically, a small sample population of only eight medical residents was used.

Consequently, the increases in salivary alpha-amylase and cystatin in response to the acute stress need to be confirmed in a larger population. Furthermore, determination regarding which form of cystatin (cystatin-S, -SA or –SN) was changing in response to the acute stress were not feasible due to limited volumes of saliva. In addition, this study was only designed determine how acute stress affects the salivary proteome. Additional

119

insights as to if these changes can also be related to performance in the stressful conditions should be performed.

4.4 Conclusions and Future Directions

In summary, a combination of SDS-PAGE, nanoHPLC and mass spectrometry was used to study the salivary proteome of eight first-year medical residents performing emergency medicine simulations in order to investigate potential biomarkers of acute stress. Whole saliva from the residents was collected at three time points: in the morning prior to the simulation, immediately after the simulation and the next morning upon waking. As expected, there was some variability in the amounts of the salivary proteins between the individuals and at the different time points.

However, there was no significant difference in the alpha-amylase, ~26 kDa, cystatins or low-molecular weight bands from the wake to the pre-simulation time points.

Both the t-test and the Wilcoxon signed-rank test indicated that the relative abundance of salivary alpha-amylase and cystatins changed significantly after the emergency medicine simulation when compared to both the pre-emergency medicine simulation and the waking time points. The results were not as consistent for the ~26 kDa band as it only significantly changed in comparison to the pre-emergency medicine simulation time point. Even more variability was observed for the low-molecular weight protein band. A significant difference in the relative amount of this band was only observed when the post-simulation time point was compared to the waking time point using the t-test. These results indicate that these proteins should be studied further for the identification of a

120

salivary biomarker, or panel of biomarkers, of acute stress. For these studies a larger population should be used to increase the confidence of the findings.

As in the analysis of stress using the cold pressor test, future work includes the use of label-free mass spectrometry for the relative quantification of the salivary proteins at the three time points. Equal volumes of whole saliva from each medical resident will be digested with trypsin. The resulting tryptic peptides will be analyzed in triplicate by nanoHPLC-ESI-MS/MS, and Thermo’s SIEVE software will be used for spectral alignment prior to performing statistical analyses. Afterwards, western blotting can be used as additional confirmation of putative salivary biomarkers of acute stress.

Discovery of novel salivary biomarkers of stress can lead to an objective way to measure the extent and consequences of this physiological state. A quantitative comparison of protein biomarkers of stress with current indicators or stress (e.g., cortisol) will be beneficial both to evaluate the effect of the stress and performance of the individuals affected by the stressful situations.

121

Chapter 5

Summary

Proteomics and mass spectrometry were successfully employed for the analysis of proteins in human saliva to elucidate novel biomarkers that may be used for objective determination of acute stress. SDS-PAGE was useful for identification of the most abundant salivary proteins (e.g., alpha-amylase, serum albumin, cystatin-S, cystatin-SA and cystatin-SN). In order to increase the number of proteins identified, it was necessary to perform a trypsin digest on the whole saliva followed by nanoHPLC-MALDI-MS/MS or nanoHPLC-ESI-MS/MS analyses. While both methods greatly increased the number of salivary proteins identified, nanoHPLC-ESI-MS/MS using the Thermo ESI-Orbitrap

Fusion yielded considerably more identifications in a shorter amount of time. Therefore, future analyses will primarily use nanoHPLC-ESI-MS/MS for the identification, structural characterization, and quantification of peptides and proteins from a complex sample, such as saliva. In addition, phosphopeptide enrichment using Fe-NTA columns vastly improved the detection and identification of phosphopeptides originating from whole saliva. Additional studies using ETD fragmentation in conjunction with nanoHPLC-ESI-MS/MS are needed to fully characterize these enriched salivary phosphopeptides.

122

From the separation techniques utilized (i.e., SDS-PAGE, 2D-DIGE and HPLC), it was possible to detect differences in the relative protein abundances in saliva collected at stressed and non-stressed time points using both the cold pressor test and the emergency medicine simulations. Preliminary results from the emergency medicine simulations are especially encouraging indicating that some salivary proteins (e.g., alpha- amylase and type S cystatins) significantly increased in response to acute stress. Salivary proteins in the ~26 kDa band, which likely contains proline-rich proteins, also increased after the acute stressor. In order to confirm these results, saliva samples from a larger population of people experiencing stressful situations need to be collected and compared to non-stressed controls. Furthermore, label-free quantification should be performed in order to identify lower abundance salivary proteins as potential biomarkers that may have greater specificity for acute stress.

Overall, the discovery of novel salivary protein biomarkers of acute stress has great importance. Specifically, the use of saliva facilitates multiple sample collections during dynamic situations. These novel biomarkers may also be used in conjunction with current stress biomarkers (e.g., cortisol) in order to provide an objective measurement to determine individuals’ physiological state while monitoring their performance during stressful situations. These investigations confirm salivary alpha-amylase as a biomarker of acute stress, although it lacks specificity. Additionally, the type S cystatins and histatin-3 were determined to be putative salivary biomarkers of acute stress.

123

References

1. May, C.; Brosseron, F.; Chartowski, P.; Schumbrutzki, C.; Schoenebeck, B.; Marcus, K. Instruments and Methods in Proteomics. In Data Mining in Proteomics, Hamacher, M.; Eisenacher, M.; Stephan, C., Eds. Humana Press: 2011; Vol. 696, pp 3-26.

2. Edmond de Hoffmann, V. S. Mass Spectrometry: Principles and Application. Third ed.; Jon Wiley & Sons Ltd: West Sussex, England, 2007.

3. Ho, C. S.; Lam, C. W. K.; Chan, M. H. M.; Cheung, R. C. K.; Law, L. K.; Lit, L. C. W.; Ng, K. F.; Suen, M. W. M.; Tai, H. L. Electrospray Ionisation Mass Spectrometry: Principles and Clinical Applications. The Clinical Biochemist Reviews 2003, 24 (1), 3-12.

4. Helsens, K.; Martens, L.; Vandekerckhove, J.; Gevaert, K. Mass Spectrometry- Driven Proteomics: An Introduction. In Gel-Free Proteomics, Gevaert, K.; Vandekerckhove, J., Eds. Humana Press: 2011; Vol. 753, pp 1-27.

5. Jaskolla, T. W.; Karas, M. Compelling Evidence for Lucky Survivor and Gas Phase Protonation: The Unified MALDI Analyte Protonation Mechanism. Journal of the American Society for Mass Spectrometry 2011, 22 (6), 976-988.

6. Karas, M.; Krüger, R. Ion Formation in MALDI: The Cluster Ionization Mechanism. Chemical Reviews 2003, 103 (2), 427-440.

7. Hillenkamp, F.; Jaskolla, T. W.; Karas, M. The MALDI Process and Method. In MALDI MS, Wiley-VCH Verlag GmbH & Co. KGaA: 2013; pp 1-40.

8. Karas, M.; Bahr, U.; Dülcks, T. Nano-electrospray ionization mass spectrometry: addressing analytical problems beyond routine. Fresenius' Journal of Analytical Chemistry 2000, 366 (6), 669-676.

124

9. Perry, R. H.; Cooks, R. G.; Noll, R. J. Orbitrap mass spectrometry: instrumentation, ion motion and applications. Mass Spectrometry Reviews 2008, 27 (6), 661-99.

10. Zubarev, R. A.; Makarov, A. Orbitrap Mass Spectrometry. Analytical Chemistry 2013, 85 (11), 5288-5296.

11. Eliuk, S.; Makarov, A. Evolution of Orbitrap Mass Spectrometry Instrumentation. Annual Review of Analytical Chemistry 2015, 8 (1), 61-80.

12. Hillenkamp, F.; Jaskolla, T. W.; Karas, M. MALDI MS. Wiley-VCH Verlag GmbH & Co. KGaA: 2013.

13. Hjernø, K. Protein Identification by Peptide Mass Fingerprinting. In Mass Spectrometry Data Analysis in Proteomics, Matthiesen, R., Ed. Humana Press: 2007; Vol. 367, pp 61-75.

14. Perkins, D. N.; Pappin, D. J.; Creasy, D. M.; Cottrell, J. S. Probability-based protein identification by searching sequence databases using mass spectrometry data. Electrophoresis 1999, 20 (18), 3551-67.

15. Parker, K. C. Scoring methods in MALDI peptide mass fingerprinting: ChemScore, and the ChemApplex program. Journal of the American Society for Mass Spectrometry 2002, 13 (1), 22-39.

16. Suckau, D.; Resemann, A.; Schuerenberg, M.; Hufnagel, P.; Franzen, J.; Holle, A. A novel MALDI LIFT-TOF/TOF mass spectrometer for proteomics. Analytical and Bioanalytical Chemistry 2003, 376 (7), 952-965.

17. Nesvizhskii, A. Protein Identification by Tandem Mass Spectrometry and Sequence Database Searching. In Mass Spectrometry Data Analysis in Proteomics, Matthiesen, R., Ed. Humana Press: 2007; Vol. 367, pp 87-119.

18. Choi, H.; Nesvizhskii, A. I. False Discovery Rates and Related Statistical Concepts in Mass Spectrometry-Based Proteomics. Journal of Proteome Research 2008, 7 (1), 47-50.

125

19. Shao, C.; Zhang, Y.; Sun, W. Statistical characterization of HCD fragmentation patterns of tryptic peptides on an LTQ Orbitrap Velos mass spectrometer. Journal of Proteomics 2014, 109, 26-37.

20. Udeshi, N. D.; Compton, P. D.; Shabanowitz, J.; Hunt, D. F.; Rose, K. L. Methods for analyzing peptides and proteins on a chromatographic timescale by electron-transfer dissociation mass spectrometry. Nature Protocols 2008, 3 (11), 1709-1717.

21. Syka, J. E. P.; Coon, J. J.; Schroeder, M. J.; Shabanowitz, J.; Hunt, D. F. Peptide and protein sequence analysis by electron transfer dissociation mass spectrometry. Proceedings of the National Academy of Sciences of the United States of America 2004, 101 (26), 9528-9533.

22. Poole, C. F. New trends in solid-phase extraction. TrAC Trends in Analytical Chemistry 2003, 22 (6), 362-373.

23. Rappsilber, J.; Mann, M.; Ishihama, Y. Protocol for micro-purification, enrichment, pre-fractionation and storage of peptides for proteomics using StageTips. Nature Protocols 2007, 2 (8), 1896-906.

24. Pluskal, M. G. Microscale sample preparation. Nature Biotechnology 2000, 18 (1), 104-105.

25. Hennion, M.-C. Solid-phase extraction: method development, sorbents, and coupling with liquid chromatography. Journal of Chromatography A 1999, 856 (1–2), 3-54.

26. Al-Tubuly, A. A. SDS-PAGE and Western Blotting. In Diagnostic and Therapeutic Antibodies, George, A. J. T.; Urch, C. E., Eds. Humana Press: Totowa, NJ, 2000; pp 391-405.

27. Gibbins, J. M. Techniques for Analysis of Proteins by SDS-Polyacrylamide Gel Electrophoresis and Western Blotting. In Platelets and Megakaryocytes: Volume 2: Perspectives and Techniques, Gibbins, J. M.; Mahaut-Smith, M. P., Eds. Humana Press: Totowa, NJ, 2004; pp 139-151.

126

28. Garfin, D. E. Chapter 29 One-Dimensional Gel Electrophoresis1. In Methods in Enzymology, Richard, R. B.; Murray, P. D., Eds. Academic Press: 2009; Vol. Volume 463, pp 497-513.

29. de Moreno, M. R.; Smith, J. F.; Smith, R. V. Mechanism studies of coomassie blue and silver staining of proteins. Journal of Pharmaceutical Sciences 1986, 75 (9), 907-11.

30. Lopez, M. F.; Berggren, K.; Chernokalskaya, E.; Lazarev, A.; Robinson, M.; Patton, W. F. A comparison of silver stain and SYPRO Ruby Protein Gel Stain with respect to protein detection in two-dimensional gels and identification by peptide mass profiling. Electrophoresis 2000, 21 (17), 3673-3683.

31. Lilley, K. S.; Razzaq, A.; Dupree, P. Two-dimensional gel electrophoresis: recent advances in sample preparation, detection and quantitation. Current Opinion in Chemical Biology 2002, 6 (1), 46-50.

32. Tannu, N. S.; Hemby, S. E. Two-dimensional fluorescence difference gel electrophoresis for comparative proteomics profiling. Nature Protocols 2006, 1 (4), 1732-42.

33. Nakashima, D.; Uzawa, K.; Kasamatsu, A.; Koike, H.; Endo, Y.; Saito, K.; Hashitani, S.; Numata, T.; Urade, M.; Tanzawa, H. Protein expression profiling identifies maspin and stathmin as potential biomarkers of adenoid cystic carcinoma of the salivary glands. International Jounal of Cancer 2006, 118 (3), 704-13.

34. Rabilloud, T.; Lelong, C. Two-dimensional gel electrophoresis in proteomics: A tutorial. Journal of Proteomics 2011, 74 (10), 1829-1841.

35. Lilley, K. S.; Friedman, D. B. All about DIGE: quantification technology for differential-display 2D-gel proteomics. Expert Review of Proteomics 2004, 1 (4), 401-9.

36. Arentz, G.; Weiland, F.; Oehler, M. K.; Hoffmann, P. State of the art of 2D DIGE. Proteomics Clinical Applications 2015, 9 (3-4), 277-88.

37. Poole, C. F. Chapter 4 - The Column in Liquid Chromatography. In The Essence of Chromatography, Elsevier Science: Amsterdam, 2003; pp 267-429. 127

38. Josic, D.; Kovac, S. Reversed-Phase High Performance Liquid Chromatography of Proteins. In Current Protocols in Protein Science, John Wiley & Sons, Inc.: 2001.

39. Mirgorodskaya, E.; Braeuer, C.; Fucini, P.; Lehrach, H.; Gobom, J. Nanoflow liquid chromatography coupled to matrix-assisted laser desorption/ionization mass spectrometry: Sample preparation, data analysis, and application to the analysis of complex peptide mixtures. Proteomics 2005, 5 (2), 399-408.

40. Aguilar, M.-I. Reversed-Phase High-Performance Liquid Chromatography. In HPLC of Peptides and Proteins: Methods and Protocols, Aguilar, M.-I., Ed. Springer New York: Totowa, NJ, 2004; pp 9-22.

41. Fröhlich, T.; Arnold, G. A Newcomer’s Guide to Nano-Liquid-Chromatography of Peptides. In Proteomics, Reinders, J.; Sickmann, A., Eds. Humana Press: 2009; Vol. 564, pp 123-141.

42. Tichy, A.; Salovska, B.; Rehulka, P.; Klimentova, J.; Vavrova, J.; Stulik, J.; Hernychova, L. Phosphoproteomics: searching for a needle in a haystack. Journal of Proteomics 2011, 74 (12), 2786-97.

43. Dunn, J. D.; Reid, G. E.; Bruening, M. L. Techniques for phosphopeptide enrichment prior to analysis by mass spectrometry. Mass Spectrometry Reviews 2010, 29 (1), 29-54.

44. Hu, S.; Loo, J. A.; Wong, D. T. Human body fluid proteome analysis. Proteomics 2006, 6 (23), 6326-6353.

45. Al-Tarawneh, S. K.; Border, M. B.; Dibble, C. F.; Bencharit, S. Defining Salivary Biomarkers Using Mass Spectrometry-Based Proteomics: A Systematic Review. OMICS: A Journal of Integrative Biology 2011, 15 (6), 353-361.

46. Amado, F. L.; Vitorino, R. P.; Lobo, M. C.; Domingues, P. D. N. Proteomics of Human Saliva. In Proteomics of Human Body Fluids, Thongboonkerd, V., Ed. Humana Press: 2007; pp 347-376.

47. Millea, K. M.; Krull, I. S.; Chakraborty, A. B.; Gebler, J. C.; Berger, S. J. Comparative profiling of human saliva by intact protein LC/ESI-TOF mass

128

spectrometry. Biochimica et Biophysica Acta) - Proteins and Proteomics 2007, 1774 (7), 897-906.

48. Schulz, B. L.; Cooper-White, J.; Punyadeera, C. K. Saliva proteome research: current status and future outlook. Critical Reviews in Biotechnology 2013, 33 (3), 246-59.

49. Castagnola, M.; Cabras, T.; Vitali, A.; Sanna, M. T.; Messana, I. Biotechnological implications of the salivary proteome. Trends in Biotechnology 2011, 29 (8), 409- 418.

50. Amado, F. M. L.; Ferreira, R. P.; Vitorino, R. One decade of salivary proteomics: Current approaches and outstanding challenges. Clinical Biochemistry 2013, 46 (6), 506-517.

51. Vitorino, R.; Lobo, M. J. C.; Ferrer-Correira, A. J.; Dubin, J. R.; Tomer, K. B.; Domingues, P. M.; Amado, F. M. L. Identification of human whole saliva protein components using proteomics. Proteomics 2004, 4 (4), 1109-1115.

52. Soo-Quee Koh, D.; Choon-Huat Koh, G. The use of salivary biomarkers in occupational and environmental medicine. Occupational and Environmental Medicine 2007, 64 (3), 202-210.

53. Castagnola, M.; Cabras, T.; Iavarone, F.; Fanali, C.; Nemolato, S.; Peluso, G.; Bosello, S. L.; Faa, G.; Ferraccioli, G.; Messana, I. The human salivary proteome: a critical overview of the results obtained by different proteomic platforms. Expert Review of Proteomics 2012, 9 (1), 33-46.

54. Brandão, E.; Soares, S.; Mateus, N.; de Freitas, V. Human saliva protein profile: Influence of food ingestion. Food Research International 2014, 64 (0), 508-513.

55. Messana, I.; Inzitari, R.; Fanali, C.; Cabras, T.; Castagnola, M. Facts and artifacts in proteomics of body fluids. What proteomics of saliva is telling us? Journal of Separation Science 2008, 31 (11), 1948-1963.

56. Zauber, H.; Mosler, S.; Heßberg, A. v.; Schulze, W. X. Dynamics of salivary proteins and metabolites during extreme endurance sports – a case study. Proteomics 2012, 12 (13), 2221-2235.

129

57. Hardt, M.; Thomas, L. R.; Dixon, S. E.; Newport, G.; Agabian, N.; Prakobphol, A.; Hall, S. C.; Witkowska, H. E.; Fisher, S. J. Toward Defining the Human Parotid Gland Salivary Proteome and Peptidome: Identification and Characterization Using 2D SDS−PAGE, Ultrafiltration, HPLC, and Mass Spectrometry. Biochemistry 2005, 44 (8), 2885-2899.

58. Amado, F. M.; Vitorino, R. M.; Domingues, P. M.; Lobo, M. J.; Duarte, J. A. Analysis of the human saliva proteome. Expert Review of Proteomics 2005, 2 (4), 521-39.

59. Ghafouri, B.; Tagesson, C.; Lindahl, M. Mapping of proteins in human saliva using two-dimensional gel electrophoresis and peptide mass fingerprinting. Proteomics 2003, 3 (6), 1003-1015.

60. Dawes, C. Circadian rhythms in human salivary flow rate and composition. The Journal of Physiology 1972, 220 (3), 529-545.

61. Deuter, C. E.; Kuehl, L. K.; Blumenthal, T. D.; Schulz, A.; Oitzl, M. S.; Schachinger, H. Effects of Cold Pressor Stress on the Human Startle Response. Plos One 2012, 7 (11), e49866.

62. Engert, V.; Merla, A.; Grant, J. A.; Cardone, D.; Tusche, A.; Singer, T. Exploring the Use of Thermal Infrared Imaging in Human Stress Research. Plos One 2014, 9 (3), e90782.

63. Smeets, T.; Cornelisse, S.; Quaedflieg, C. W.; Meyer, T.; Jelicic, M.; Merckelbach, H. Introducing the Maastricht Acute Stress Test (MAST): a quick and non-invasive approach to elicit robust autonomic and glucocorticoid stress responses. Psychoneuroendocrinology 2012, 37 (12), 1998-2008.

64. Richardson, H. L.; Macey, P. M.; Kumar, R.; Valladares, E. M.; Woo, M. A.; Harper, R. M. Neural and physiological responses to a cold pressor challenge in healthy adolescents. Journal of Neuroscience Research 2013, 91 (12), 1618-27.

65. Santarcangelo, E. L.; Paoletti, G.; Chiavacci, I.; Palombo, C.; Carli, G.; Varanini, M. Cognitive modulation of psychophysical, respiratory and autonomic responses to cold pressor test. Plos One 2013, 8 (10), e75023.

130

66. Schwabe, L.; Haddad, L.; Schachinger, H. HPA axis activation by a socially evaluated cold-pressor test. Psychoneuroendocrinology 2008, 33 (6), 890-5.

67. van Stegeren, A. H.; Wolf, O. T.; Kindt, M. Salivary alpha amylase and cortisol responses to different stress tasks: impact of sex. International Journal of Psychophysiology 2008, 69 (1), 33-40.

68. Nater, U. M.; Rohleder, N. Salivary alpha-amylase as a non-invasive biomarker for the sympathetic nervous system: Current state of research. Psychoneuroendocrinology 2009, 34 (4), 486-496.

69. Salivary Proteome Knowledge Base. http://hspp.dent.ucla.edu/cgi-bin/hspmscgi- bin/search_pro_c.cgi (accessed 5/10/2016).

70. Yan, W.; Apweiler, R.; Balgley, B. M.; Boontheung, P.; Bundy, J. L.; Cargile, B. J.; Cole, S.; Fang, X.; Gonzalez-Begne, M.; Griffin, T. J.; Hagen, F.; Hu, S.; Wolinsky, L. E.; Lee, C. S.; Malamud, D.; Melvin, J. E.; Menon, R.; Mueller, M.; Qiao, R.; Rhodus, N. L.; Sevinsky, J. R.; States, D.; Stephenson, J. L.; Than, S.; Yates, J. R.; Yu, W.; Xie, H.; Xie, Y.; Omenn, G. S.; Loo, J. A.; Wong, D. T. Systematic comparison of the human saliva and plasma proteomes. PROTEOMICS – Clinical Applications 2009, 3 (1), 116-134.

71. Kasamatsu, A.; Uzawa, K.; Nakashima, D.; Kouzu, Y.; Endo, Y.; Koike, H.; Yokoe, H.; Harada, K.; Sato, M.; Tanzawa, H. A proteomics approach to characterizing human submandibular gland cell lines by fluorescent two- dimensional differential in-gel electrophoresis. International Journal of Molecular Medicine 2006, 17 (2), 253-60.

72. Shevchenko, A.; Tomas, H.; Havlis, J.; Olsen, J. V.; Mann, M. In-gel digestion for mass spectrometric characterization of proteins and proteomes. Nature Protocols 2007, 1 (6), 2856-2860.

73. Käll, L.; Storey, J. D.; MacCoss, M. J.; Noble, W. S. Assigning Significance to Peptides Identified by Tandem Mass Spectrometry Using Decoy Databases. Journal of Proteome Research 2008, 7 (1), 29-34.

74. Deutsch, O.; Fleissig, Y.; Zaks, B.; Krief, G.; Aframian, D. J.; Palmon, A. An approach to remove alpha amylase for proteomic analysis of low abundance biomarkers in human saliva. Electrophoresis 2008, 29 (20), 4150-4157.

131

75. Bandhakavi, S.; Stone, M. D.; Onsongo, G.; Van Riper, S. K.; Griffin, T. J. A Dynamic Range Compression and Three-Dimensional Peptide Fractionation Analysis Platform Expands Proteome Coverage and the Diagnostic Potential of Whole Saliva. Journal of Proteome Research 2009, 8 (12), 5590-5600.

76. Bouameur, J.-E.; Favre, B.; Borradori, L. Plakins, a Versatile Family of Cytolinkers: Roles in Skin Integrity and in Human Diseases. Journal of Investigative Dermatology 2014, 134, 885-894.

77. Kim, H.-J.; Woo, S. J.; Suh, E. J.; Ahn, J.; Park, J. H.; Hong, H. K.; Lee, J. E.; Ahn, S. J.; Hwang, D. J.; Kim, K. W.; Park, K. H.; Lee, C. Identification of Vinculin as a Potential Plasma Marker for Age-Related Macular DegenerationVinculin as a Plasma Marker for AMD. Investigative Ophthalmology & Visual Science 2014, 55 (11), 7166-7176.

78. Hassan, M. I.; Waheed, A.; Yadav, S.; Singh, T. P.; Ahmad, F. Zinc α2- Glycoprotein: A Multidisciplinary Protein. American Association for Cancer Research 2008, 6 (6), 892-906.

79. Hu, S.; Arellano, M.; Boontheung, P.; Wang, J.; Zhou, H.; Jiang, J.; Elashoff, D.; Wei, R.; Loo, J. A.; Wong, D. T. Salivary proteomics for oral cancer biomarker discovery. Clinical Cancer Research 2008, 14 (19), 6246-52.

80. Stone, M. D.; Chen, X.; McGowan, T.; Bandhakavi, S.; Cheng, B.; Rhodus, N. L.; Griffin, T. J. Large-Scale Phosphoproteomics Analysis of Whole Saliva Reveals a Distinct Phosphorylation Pattern. Journal of Proteome Research 2011, 10 (4), 1728-1736.

81. Schulenberg, B.; Goodman, T. N.; Aggeler, R.; Capaldi, R. A.; Patton, W. F. Characterization of dynamic and steady-state protein phosphorylation using a fluorescent phosphoprotein gel stain and mass spectrometry. Electrophoresis 2004, 25 (15), 2526-2532.

82. Steinberg, T. H.; Agnew, B. J.; Gee, K. R.; Leung, W. Y.; Goodman, T.; Schulenberg, B.; Hendrickson, J.; Beechem, J. M.; Haugland, R. P.; Patton, W. F. Global quantitative phosphoprotein analysis using Multiplexed Proteomics technology. Proteomics 2003, 3 (7), 1128-44.

83. Lupi, A.; Messana, I.; Denotti, G.; Schininà, M. E.; Gambarini, G.; Fadda, M. B.; Vitali, A.; Cabras, T.; Piras, V.; Patamia, M.; Cordaro, M.; Giardina, B.; 132

Castagnola, M. Identification of the human salivary cystatin complex by the coupling of high-performance liquid chromatography and ion-trap mass spectrometry. Proteomics 2003, 3 (4), 461-467.

84. Ryan, C. M.; Souda, P.; Halgand, F.; Wong, D. T.; Loo, J. A.; Faull, K. F.; Whitelegge, J. P. Confident assignment of intact mass tags to human salivary cystatins using top-down Fourier-transform ion cyclotron resonance mass spectrometry. Journal of the American Society for Mass Spectrometry 2011, 21 (6), 908-917.

85. Madapallimattam, G.; Bennick, A. Phosphopeptides derived from human salivary acidic proline-rich proteins. Biological activities and concentration in saliva. Biochemical Journal 1990, 270 (2), 297-304.

86. Marrelli, M.; Gentile, S.; Palmieri, F.; Paduano, F.; Tatullo, M. Correlation between Surgeon's Experience, Surgery Complexity and the Alteration of Stress Related Physiological Parameters. Plos One 2014, 9 (11), 1-8.

87. Naumova, E. A.; Sandulescu, T.; Bochnig, C.; Khatib, P. A.; Lee, W.-K.; Zimmer, S.; Arnold, W. H. Dynamic changes in saliva after acute mental stress. Scientific Reports 2014, 4.

88. McGraw, L. K.; Out, D.; Hammermeister, J. J.; Ohlson, C. J.; Pickering, M. A.; Granger, D. A. Nature, correlates, and consequences of stress-related biological reactivity and regulation in Army nurses during combat casualty simulation. Psychoneuroendocrinology 2013, 38 (1), 135-144.

89. Strahler, J.; Ziegert, T. Psychobiological stress response to a simulated school shooting in police officers. Psychoneuroendocrinology 2015, 51, 80-91.

90. Groer, M.; Murphy, R.; Bunnell, W.; Salomon, K.; Van Eepoel, J.; Rankin, B.; White, K.; Bykowski, C. Salivary Measures of Stress and Immunity in Police Officers Engaged in Simulated Critical Incident Scenarios. Journal of Occupational and Environmental Medicine 2010, 52, 595-602.

91. Valentin, B.; Grottke, O.; Skorning, M.; Bergrath, S.; Fischermann, H.; Rortgen, D.; Mennig, M. T.; Fitzner, C.; Muller, M. P.; Kirschbaum, C.; Rossaint, R.; Beckers, S. K. Cortisol and alpha-amylase as stress response indicators during pre-hospital emergency medicine training with repetitive high-fidelity simulation

133

and scenarios with standardized patients. Scandinavian Journal of Trauma, Resuscitation and Emergency Medicine 2015, 23, 31.

92. Müller, M. P.; Hänsel, M.; Fichtner, A.; Hardt, F.; Weber, S.; Kirschbaum, C.; Rüder, S.; Walcher, F.; Koch, T.; Eich, C. Excellence in performance and stress reduction during two different full scale simulator training courses: A pilot study. Resuscitation 2009, 80 (8), 919-924.

93. Clarke, S.; Horeczko, T.; Cotton, D.; Bair, A. Heart rate, anxiety and performance of residents during a simulated critical clinical encounter: a pilot study. BMC Medical Education 2014, 14 (1), 1-8.

94. Beeley, J. A.; Khoo, K. S.; Lamey, P.-J. Two-dimensional electrophoresis of human parotid salivary proteins from normal and connective tissue disorder subjects using immobilised pH gradients. Electrophoresis 1991, 12 (7-8), 493- 499.

95. Beeley, J. A.; Sweeney, D.; Lindsay, J. C. B.; Buchanan, M. L.; Sarna, L.; Khoo, K. S. Sodium dodecyl sulphate-polyacrylamide gel electrophoresis of human parotid salivary proteins. Electrophoresis 1991, 12 (12), 1032-1041.

96. Isemura, S.; Saitoh, E.; Sanada, K.; Minakata, K. Identification of full-sized forms of salivary (S-type) cystatins (cystatin SN, cystatin SA, cystatin S, and two phosphorylated forms of cystatin S) in human whole saliva and determination of phosphorylation sites of cystatin S. Journal of Biochemistry 1991, 110 (4), 648- 54.

97. Hu, S.; Xie, Y.; Ramachandran, P.; Ogorzalek Loo, R. R.; Li, Y.; Loo, J. A.; Wong, D. T. Large-scale identification of proteins in human salivary proteome by liquid chromatography/mass spectrometry and two-dimensional gel electrophoresis-mass spectrometry. Proteomics 2005, 5 (6), 1714-1728.

98. Sousa-Pereira, P. d.; Abrantes, J.; Pinheiro, A.; Colaço, B.; Vitorino, R.; Esteves, P. J. Evolution of C, D and S-Type Cystatins in Mammals: An Extensive Duplication in Primates. Plos One 2014, 9 (10), e109050.

99. Sabatini, L. M.; Azen, E. A. Histatins, a family of salivary histidine-rich proteins, are encoded by at least two loci (HIS1 and HIS2). Biochemical and Biophysical Research Communications 1989, 160 (2), 495-502.

134

100. Castagnola, M.; Inzitari, R.; Rossetti, D. V.; Olmi, C.; Cabras, T.; Piras, V.; Nicolussi, P.; Sanna, M. T.; Pellegrini, M.; Giardina, B.; Messana, I. A Cascade of 24 Histatins ( Fragments) in Human Saliva: Suggestions for a Pre- Secretory Sequential Cleavage Pathway. Journal of Biological Chemistry 2004, 279 (40), 41436-41443.

101. Tsai, H.; Raj, P. A.; Bobek, L. A. Candidacidal activity of recombinant human salivary histatin-5 and variants. Infection and Immunity 1996, 64 (12), 5000-7.

102. Tsai, H.; Bobek, L. A. Human salivary histatin-5 exerts potent fungicidal activity against Cryptococcus neoformans. Biochimica et Biophysica Acta - General Subjects 1997, 1336 (3), 367-369.

103. Gusman, H.; Travis, J.; Helmerhorst, E. J.; Potempa, J.; Troxler, R. F.; Oppenheim, F. G. Salivary Histatin 5 is an Inhibitor of Both Host and Bacterial Enzymes Implicated in Periodontal Disease. Infection and Immunity 2001, 69 (3), 1402-1408.

104. Trueba, A. F.; Mizrachi, D.; Auchus, R. J.; Vogel, P. D.; Ritz, T. Effects of psychosocial stress on the pattern of salivary protein release. Physiology & Behavior 2012, 105 (3), 841-849.

105. Obayashi, K. Salivary mental stress proteins. Clinica Chimica Acta 2013, 425, 196-201.

106. Bosch, J. A.; Ring, C.; de Geus, E. J. C.; Veerman, E. C. I.; Nieuw Amerongen, A. V. Stress and secretory immunity. In International Review of Neurobiology, Academic Press: 2002; Vol. Volume 52, pp 213-253.

107. Castagnola, M.; Cabras, T.; Denotti, G.; Fadda, M. B.; Gambarini, G.; Lupi, A.; Manca, I.; Onnis, G.; Piras, V.; Soro, V.; Tambaro, S.; Messana, I. Circadian Rhythms of , Histatin 3, Histatin 5, Statherin and Uric Acid in Whole Human Saliva Secretion. Biological Rhythm Research 2002, 33 (2), 213-222.

135

Appendix A

Supplemental Cold Pressor Test Salivary Proteome

Figures and Identifications

Figure A-1 nanoHPLC chromatogram of the trypsin digested HPLC fraction collected from 24-26 min from pooled saliva B

136

Figure A-2 nanoHPLC-MALDI-MS/MS of trypsin digested whole saliva

Figure A-3 Venn diagram of the functions of the salivary proteins identified using nanoHPLC-ESI-MS/MS

137

Figure A-4 MALDI-MS of (A) phosphopeptide enriched casein digest and (B) casein digest prior to enrichment

138

Table A.1 List of salivary proteins identified from cold pressor test saliva

Uniprot Average Technique Protein Name Time point b Accession Mass Used a Small proline-rich protein 3 B1AN48 17,059 ESI Pre, Imm (Fragment) Poly(A) binding protein, B1ANR0 cytoplasmic 4 (Inducible form), 67,971 ESI Post isoform CRA_e C9J5S7 Glutathione S-transferase P 12,566 ESI Pre D6RBJ7 Vitamin D-binding protein 37,077 ESI Pre Immunoglobulin J chain D6RD17 15,817 ESI all (Fragment) D6REQ6 Ribonuclease T2 25,309 ESI Pre D6RF35 Vitamin D-binding protein 51,300 ESI Pre D6RF88 Carboxypeptidase E (Fragment) 15,660 ESI Pre E7EV34 Tripeptidyl-peptidase 1 (Fragment) 13,888 ESI Pre, Imm 14-3-3 protein zeta/delta E7EX29 27,790 ESI Pre (Fragment) E9PKE3 Heat shock cognate 71 kDa protein 2,996 ESI Pre DDB1- and CUL4-associated factor E9PKG6 40,369 ESI Pre 5 E9PLJ3 Cofilin-1 (Fragment) 9,091 ESI Pre F2Z393 Transaldolase 35,329 ESI Pre, Imm F5H386 Lactoperoxidase 73,956 ESI Imm F5H7C1 Basic salivary proline-rich protein 3 35,123 ESI Pre F8W6P5 LVV-hemorphin-7 (Fragment) 9,670 ESI Imm F8W9L1 Serpin B3 44,770 ESI Post H0Y300 Haptoglobin 47,250 ESI Pre, Imm H0Y750 Renin receptor (Fragment) 26,633 ESI Pre H0YAS8 Clusterin beta chain (Fragment) 16,177 ESI Imm DDB1- and CUL4-associated factor H0YD18 8,644 ESI Post 5 (Fragment) DDB1- and CUL4-associated factor H0YEG8 23,788 ESI Pre 5 (Fragment) Beta-2-microglobulin form pI 5.3 H0YLF3 7,326 ESI Imm (Fragment) H7C3T4 Peroxiredoxin-4 (Fragment) 12,967 ESI Pre Fatty acid-binding protein, I6L8B7 10,123 ESI Imm epidermal Rho GDP-dissociation inhibitor 1 J3KTF8 21,614 ESI Pre (Fragment) J3QLI0 Beta-2-glycoprotein 1 (Fragment) 22,176 ESI Pre K7EMR1 Paragranulin (Fragment) 1,576 ESI Pre 139

M0QZV0 Kallikrein-11 (Fragment) 5,765 ESI Pre O00391 Sulfhydryl oxidase 1 79,578 ESI Pre O60437 Periplakin 204,747 PMF Imm Ly6/PLAUR domain-containing O95274 31,022 ESI Pre protein 3 P00450 Ceruloplasmin 120,085 ESI Pre P00738-2 Isoform 2 of Haptoglobin 43,349 ESI Imm P01023 Alpha-2-macroglobulin 160,810 ESI Imm P01033 Metalloproteinase inhibitor 1 20,709 ESI Imm ESI and P01034 Cystatin-C 13,347 all MALDI P01036 Cystatin-S 14,189 all all P01037 Cystatin-SN 14,316 all all ESI, P01591 Immunoglobulin J chain 15,594 MALDI all and PMF P01593 Ig kappa chain V-I region AG 11,992 ESI all P01617 Ig kappa chain V-II region TEW 12,316 ESI Pre P01619 Ig kappa chain V-III region B6 11,636 PMF Pre ESI and P01620 Ig kappa chain V-III region SIE 11,775 Pre PMF Ig kappa chain V-III region NG9 P01621 10,389 ESI Post (Fragment) P01623 Ig kappa chain V-III region WOL 11,746 PMF Pre P01714 Ig lambda chain V-III region SH 11,393 ESI all P01763 Ig heavy chain V-III region WEA 12,256 ESI Pre ESI and P01765 Ig heavy chain V-III region TIL 12,356 all MALDI ESI and P01766 Ig heavy chain V-III region BRO 13,227 all MALDI P01767 Ig heavy chain V-III region BUT 12,379 ESI Imm P01772 Ig heavy chain V-III region KOL 13,718 ESI Imm, Post P01776 Ig heavy chain V-III region WAS 13,091 MALDI Pre ESI and P01779 Ig heavy chain V-III region TUR 12,431 Prior MALDI Polymeric immunoglobulin ESI and P01833 81,349 all receptor MALDI ESI, P01834 Ig kappa chain C region 11,609 MALDI all and PMF P01876 Ig alpha-1 chain C region 37,655 all all P01877 Ig alpha-2 chain C region 36,526 all all P02538 Keratin, type II cytoskeletal 6A 59,914 PMF Imm 140

P02647 Apolipoprotein A-I 28,079 ESI Pre, Imm P02750 Leucine-rich alpha-2-glycoprotein 34,346 ESI Pre P02763 Alpha-1-acid glycoprotein 1 21,560 ESI Pre, Imm P02766 Transthyretin 13,761 ESI Pre P02768 Serum albumin 66,472 all all P02787 Serotransferrin 75,195 ESI all ESI and P02788 Lactotransferrin 76,165 all PMF P02790 Hemopexin 49,295 ESI Pre P02812 Basic salivary proline-rich protein 2 39,189 ESI all Submaxillary gland androgen- ESI and P02814 5,810 Pre regulated protein 3B MALDI P03973 Antileukoproteinase 11,726 ESI all P04040 Catalase 59,625 ESI Pre ESI and P04080 Cystatin-B 11,140 all MALDI P04196 Histidine-rich glycoprotein 57,660 ESI Pre P04206 Ig kappa chain V-III region GOL 11,830 PMF Pre P04208 Ig lambda chain V-I region WAH 11,725 ESI Pre P04211 Ig lambda chain V region 4A 10,240 ESI Pre, Post PMF and P04259 Keratin, type II cytoskeletal 6B 59,936 Pre, Imm HPLC Frac ESI, PMF P04264 Keratin, type II cytoskeletal 1 65,908 and HPLC Pre, Imm Frac Ig kappa chain V-III region VG P04433 12,575 ESI Pre, Imm (Fragment) P04745 Alpha-amylase 1 55,909 all all MALDI, P04746 Pancreatic alpha-amylase 55,888 PMF and all HPLC Frac ESI and P05109 Protein S100-A8 10,835 all HPLC Frac P05164 Myeloperoxidase 66,107 ESI Pre, Imm Ig kappa chain V-IV region P06312 13,380 ESI Pre (Fragment) P06331 Ig heavy chain V-II region ARH-77 13,935 ESI Imm P06396 Gelsolin 82,959 ESI Pre P06702 Protein S100-A9 13,111 ESI Imm P06733 Alpha-enolase 47,038 ESI Pre, Imm P06870 Kallikrein-1 26,406 ESI all P07237 Protein disulfide-isomerase 55,294 ESI Pre

141

P07339 Cathepsin D 37,852 ESI Pre P07602 Prosaposin 56,493 ESI Pre, Imm P07737 Profilin-1 14,923 ESI Pre, Imm P08107 Heat shock 70 kDa protein 1A/1B 69,921 ESI Pre, Imm P08246 Neutrophil elastase 25,561 ESI Pre Monocyte differentiation antigen P08571 37,215 ESI Pre CD14 P09211 Glutathione S-transferase P 23,225 ESI Pre, Imm P09228 Cystatin-SA 14,350 all all P10599 Thioredoxin 11,606 ESI all P10909 Clusterin 50,063 ESI all P12273 Prolactin-inducible protein 13,523 all all ESI, PMF P13645 Keratin, type I cytoskeletal 10 58,827 and HPLC Imm Frac P13646 Keratin, type I cytoskeletal 13 49,588 PMF Imm, Post P13796 Plastin-2 70,157 ESI all P14780 Matrix metalloproteinase-9 66,609 ESI Pre, Imm P15515 Histatin-1 4,848 ESI Pre, Imm Immunoglobulin lambda-like P15814 19,135 ESI Imm polypeptide 1 P16870 Carboxypeptidase E 48,974 ESI Pre P18206 Vinculin 123,668 PMF all Interleukin-1 receptor antagonist P18510 17,126 ESI all protein P19012 Keratin, type I cytoskeletal 15 49,212 PMF Post P19013 Keratin, type II cytoskeletal 4 57,285 PMF Post Peptidyl-glycine alpha-amidating P19021 105,058 ESI Pre monooxygenase P19652 Alpha-1-acid glycoprotein 2 21,651 ESI Pre MALDI, P19961 Alpha-amylase 2B 55,891 PMF and all HPLC Frac P20061 Transcobalamin-1 45,597 ESI all P20160 Azurocidin 24,051 ESI Pre P20742 Pregnancy zone protein 161,057 ESI Pre ESI and P22079 Lactoperoxidase 71,568 all MALDI ESI, P23280 Carbonic anhydrase 6 33,570 MALDI all and PMF

142

ESI, P25311 Zinc-alpha-2-glycoprotein 32,145 MALDI all and PMF P25815 Protein S100-P 10,400 ESI Pre P27482 Calmodulin-like protein 3 16,891 ESI Pre P27797 Calreticulin 46,466 PMF Imm ESI and P28325 Cystatin-D 13,859 all MALDI P28799 Granulins 61,787 ESI Pre P29034 Protein S100-A2 11,117 ESI Pre Cellular retinoic acid-binding P29373 15,693 ESI Pre protein 2 Phosphatidylethanolamine-binding P30086 20,926 ESI Pre, Imm protein 1 P30740 Leukocyte elastase inhibitor 42,742 ESI Pre, Imm ESI, P31025 Lipocalin-1 17,446 MALDI all and PMF P31947 14-3-3 protein sigma 27,774 ESI Pre P31949 Protein S100-A11 11,740 ESI Pre, Post P32926 Desmoglein-3 101,843 ESI Pre P34931 Heat shock 70 kDa protein 1-like 70,375 ESI Post P35527 Keratin, type I cytoskeletal 9 62,064 PMF Imm Keratin, type II cytoskeletal 2 P35908 65,433 PMF Imm epidermal P36952 Serpin B5 42,100 ESI Pre P37837 Transaldolase 37,540 PMF Imm P42357 Histidine ammonia-lyase 72,698 ESI Pre P48668 Keratin, type II cytoskeletal 6C 59,893 PMF Imm Isoform 1 of Retinoic acid receptor P49788 33,285 ESI Post responder protein 1 Isoform 2 of 6-phosphogluconate P52209 53,009 ESI Pre dehydrogenase, decarboxylating P54108 Cysteine-rich secretory protein 3 25,502 ESI all P55058 Phospholipid transfer protein 53,070 ESI Pre ESI and P59665 Neutrophil 1 3,448 all MALDI P60174 Triosephosphate isomerase 30,791 ESI Pre ESI and P60709 Actin, cytoplasmic 1 41,737 all MALDI ESI, P61626 Lysozyme C 14,701 MALDI all and PMF 143

P61769 Beta-2-microglobulin 11,731 ESI Pre Peptidyl-prolyl cis-trans isomerase P62937 18,012 ESI Pre A Isoform 2 of 14-3-3 protein P63104 27,745 ESI Pre zeta/delta P63261 Actin, cytoplasmic 2 41,793 MALDI Pre, Imm P68032 Actin, alpha cardiac muscle 1 41,785 MALDI Pre, Imm P69905 Hemoglobin subunit alpha 15,126 ESI Pre Neutrophil gelatinase-associated P80188 20,548 ESI Pre, Imm lipocalin P80748 Ig lambda chain V-III region LOI 10,520 ESI all Fatty acid-binding protein, Q01469 15,033 ESI Pre, Post epidermal Adenylyl cyclase-associated protein Q01518 51,770 ESI Pre 1 Q02487 Desmocollin-2 84,728 ESI Pre Q02817 Mucin-2 538,420 ESI Pre Q02818 Nucleobindin-1 51,146 ESI Pre Q04118 Basic salivary proline-rich protein 3 29,370 HPLC Frac Pre, Imm Q07654 Trefoil factor 3 6,580 ESI all Q08380 Galectin-3-binding protein 63,277 ESI all Isoform 2 of Receptor-type Q13332 214,276 ESI Pre tyrosine-protein phosphatase S WAP four-disulfide core domain Q14508 10,036 ESI Imm, Post protein 2 Q14515 SPARC-like protein 1 73,569 ESI Pre, Imm Q5H9A7 Metalloproteinase inhibitor 1 9,343 ESI Pre, Imm POTE ankyrin domain family Q6S8J3 121,363 MALDI Pre, Imm member E Polypeptide N- Q7Z7M9 106,266 ESI Post acetylgalactosaminyltransferase 5 Bactericidal/permeability- ESI and Q8N4F0 47,126 all increasing protein-like 1 MALDI Q8NBJ4 Golgi membrane protein 1 45,333 ESI Pre Q8TAX7 Mucin-7 36,809 ESI all BPI fold-containing family B Q8TDL5 50,268 ESI Pre, Imm member 1 Q8WXG9 G-protein coupled receptor 98 690,080 ESI Pre Q8WZ42 Titin 3,816,030 ESI Pre, Post Q96BQ1 Protein FAM3D 22,087 ESI Pre Zymogen granule protein 16 Q96DA0 17,214 PMF Pre, Imm homolog B

144

Short palate, lung and nasal ESI and Q96DR5 epithelium carcinoma-associated 25,054 all MALDI protein 2 Q96QR1 Secretoglobin family 3A member 1 8,225 ESI Pre Q9H707 Zinc finger protein 552 46,198 ESI Post ESI and Q9HC84 Mucin-5B 593,844 all MALDI Q9NZW5 MAGUK p55 subfamily member 6 61,117 ESI Pre Q9P0T7 Transmembrane protein 9 18,568 PMF Imm Q9UBC9 Small proline-rich protein 3 18,023 ESI Pre Q9UBG3 Cornulin 53,533 ESI Pre Deleted in malignant brain tumors ESI and Q9UGM3 258,660 all 1 protein MALDI nanoHPLC Q9Y4L1 Hypoxia up-regulated protein 1 107,660 Pre, Imm -MALDI Q9Y6R7 IgGFc-binding protein 569,345 ESI Pre, Imm U3KPS2 Myeloblastin 22,717 ESI Pre, Imm V9GYC1 Apolipoprotein A-II (Fragment) 5,541 ESI Pre X6R3S7 Trefoil factor 3 4,923 ESI Pre X6RAH8 Histatin-3 4,802 ESI Pre X6RJP6 Transgelin-2 (Fragment) 12,819 ESI Pre a ESI: proteins identified by nanoHPLC-ESI-MS/MS of the trypsin-digest of whole saliva HPLC Frac: proteins identified by HPLC fractionation of whole saliva proteins MALDI: proteins identified by nanoHPLC-MALDI-MS/MS of the trypsin-digest of whole saliva PMF: proteins identified by peptide mass fingerprinting b Pre: saliva collected prior to the CPT Imm: saliva collected immediately after performing the CPT Post: saliva collected 20 minutes after performing the CPT

145

Appendix B

Tables of the raw and normalized band areas

146

Table B.1 Raw and normalized peak areas in arbitrary units of the alpha-amylase and ~26 kDa Coomassie-stained gel bands

Alpha-amylase ~26 kDa band

Resident Run Vol.* 50 kDa Wake Norma Preb Norma Postc Norma Wake Norma Preb Norma Postc Norma Wake Preb Postc Wake Preb Postc

1 1 25 27413 22225 0.032430 23363 0.034090 42998 0.062741 11408 0.016646 11277 0.016455 18693 0.027276

1 2 25 30211 32366 0.042853 34852 0.046144 76207 0.10089 5146.5 0.0068139 7720.6 0.010222 14694 0.019454

1 3 30 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a

1 4 30 4796.1 27376 0.19026 40573 0.28198 91705 0.63736 5531.5 0.038444 6775.6 0.047091 21432 0.14895

2 1 25 18125 33077 0.072998 22238 0.049078 40463 0.089297 25838 0.057022 12131 0.026771 21442 0.047319

2 2 25 40807 71853 0.070431 30493 0.029889 59663 0.058482 53416 0.052359 37973 0.037221 48399 0.047442

2 3 30 18130 81875 0.15052 51940 0.095492 68216 0.12541 103780 0.19079 65065 0.11962 81371 0.149600

2 4 30 12164 81174 0.22244 33905 0.092911 52778 0.14462 60702 0.16634 8261.9 0.022639 15134 0.041471

3 1 25 3327 30910 0.037161 51039 0.061362 86009 0.10340 15006 0.018041 19912 0.023939 28190 0.033891

3 2 25 24229 30800 0.050847 50609 0.083549 69740 0.11513 5514.2 0.0091032 6010.8 0.009923 10548 0.017414

3 3 30 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a

3 4 30 8081.0 33938 0.13999 35925 0.14818 57735 0.23815 12322 0.050827 14379 0.059315 12817 0.052871

4 1 25 40198 70669 0.070320 51398 0.051143 70287 0.069939 21366 0.021261 7963.9 0.0079245 15177 0.015102

4 2 25 24055 6667 0.011085 41814 0.069528 57440 0.095512 20720 0.034453 16123 0.026809 18832 0.031315

4 3 30 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a

4 4 30 8091.0 66966 0.27589 49687 0.20470 67502 0.27809 52667 0.21697 13196 0.054365 23186 0.095520

5 1 25 54457 41119 0.030203 40618 0.029835 87796 0.064488 32599 0.023945 19226 0.014122 29864 0.021936

5 2 25 21909 33515 0.061188 25694 0.046911 57023 0.10411 27232 0.0497169 14052 0.025655 21999 0.040164

147

5 3 30 4744.9 48455 0.34039 50378 0.35391 82606 0.58031 28272 0.19861 11125 0.078157 9343.2 0.065636

5 4 30 4650.4 31808 0.22799 32424 0.23241 57109 0.40934 22677 0.16254 11967 0.085779 17478 0.12528

6 1 25 30456 25497 0.033487 25773 0.033848 50109 0.065809 14539 0.019095 17617 0.023138 33799 0.044389

6 2 25 17108 17924 0.041907 13611 0.031822 31396 0.073404 5255.7 0.012288 6163.6 0.014411 9798.3 0.022908

6 3 30 8596.0 53859 0.20885 72482 0.28107 53802 0.20863 12220. 0.047386 15829 0.061382 9474.4 0.036739

6 4 30 8505.3 26667 0.10451 43029 0.16863 26269 0.10295 7309.5 0.028647 11030 0.043229 5252.9 0.02057

7 1 25 35900 29140 0.032468 39890. 0.044446 81971 0.091332 29492 0.032859 3331961 0.037124 54407 0.060621

7 2 25 22224 32851 0.059126 26951 0.048506 60118 0.10820 26975 0.048550 14284 0.025708 20919 0.037649

7 3 30 26613 109780 0.13750 117060 0.14662 153830 0.19267 80483 0.10080 98436 0.12329 113980 0.14276

7 4 30 8844.0 33624 0.12673 38910. 0.14665 44011 0.16588 21771 0.082054 16934 0.063826 8883.8 0.033483

8 1 25 37272 109820 0.11785 51727 0.055512 37392 0.040127 67284 0.072208 12217 0.013111 12077 0.012961

8 2 25 44277 96382 0.087071 73699 0.066579 75717 0.068403 89256 0.080633 47974 0.043334 32757 0.029592

8 3 30 27213 141340 0.17312 97842 0.119847 78919 0.096668 114160 0.139838 59099 0.072391 49718 0.060900

8 4 30 7911.7 62811 0.26463 29028 0.12230 25942 0.10929 46523 0.19601 15058 0.063442 10605 0.044680 * Total volume (µL) of sample loaded on gel a Norm refers to normalized area calculated by: ⁄ b Pre refers to saliva collected prior to the emergency medicine simulaton c Post refers to saliva collected after the emergency medicine simulation n/a: protein band either not sufficiently stained for analysis or there was spillover

148

Table B.2 Raw and normalized peak areas in arbitrary units of the cystatin and a low-molecular weight Coomassie stained gel bands

Cystatin Low-molecular weight band

a b a c a a b a c a Resident Run Vol.* 50 kDa Wake Norm Pre Norm Post Norm Wake Norm Pre Norm Post Norm b c b c Wake Pre Post Wake Pre Post

1 1 25 27413 13565 0.019793 11849 0.017289 14813 0.021614 7776.2 0.011346 7659.3 0.011176 12241 0.017861

1 2 25 30211 9054.9 0.011989 5118.5 0.0067769 19065 0.025242 2094.1 0.0027726 1312.2 0.0017374 4231.8 0.0056029

1 3 30 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a

1 4 30 4796.1 14322 0.099543 9698.7 0.067406 30396 0.21126 2507.4 0.017426 720.23 0.0050057 7831.1 0.054426

2 1 25 18125 15731 0.034717 12027.9 0.026544 19839 0.043783 4417.2 0.0097483 4457.4 0.0098369 11331 0.025006

2 2 25 40807 30016 0.029423 28479 0.027916 46787 0.045862 11229 0.011007 12879 0.012625 19600. 0.019213

2 3 30 18131 60582 0.11138 57036 0.10486 65380 0.12020 8414.4 0.015469 25223 0.046372 26022 0.047841

2 4 30 12164 36480 0.099966 30360 0.083196 40778 0.11174 4214.6 0.011549 5992.5 0.016421 16221 0.044449

3 1 25 33271 9743.0 0.011714 13934 0.016752 22427 0.026963 n/a n/a n/a n/a n/a n/a

3 2 25 24229 5825.6 0.0096173 8762.4 0.014466 14332 0.023661 n/a n/a n/a n/a n/a n/a

3 3 30 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a

3 4 30 8081.0 5858.3 0.024165 8939.2 0.036873 13959 0.057582 n/a n/a n/a n/a n/a n/a

4 1 25 40199 5597.12 0.0055694 9204.7 0.0091592 15423 0.015347 n/a n/a n/a n/a n/a n/a

4 2 25 24056 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a

4 3 30 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a

4 4 30 8091.0 19808 0.081606 17028 0.070152 22783 0.093860 2513.3 0.010354 8757.6 0.036079 9740.8 0.040130

5 1 25 54457 14037 0.010311 11437 0.0084006 20906 0.015356 n/a n/a n/a n/a n/a n/a

5 2 25 21909 17352 0.031679 15598 0.028477 27376 0.049979 n/a n/a n/a n/a n/a n/a

149

5 3 30 4744.9 2340.3 0.016441 2380.5 0.016723 2753.1 0.019341 n/a n/a n/a n/a n/a n/a

5 4 30 4650.5 9091.5 0.065165 8956.6 0.064198 16545 0.11859 n/a n/a n/a n/a n/a n/a

6 1 25 30456 8745.9 0.011486 14471 0.019005 35548 0.046686 n/a n/a n/a n/a n/a n/a

6 2 25 17109 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a

6 3 30 8596.0 2896.4 0.011232 8987.3 0.034851 2683.5 0.010406 n/a n/a n/a n/a n/a n/a

6 4 30 8505.3 3388.4 0.013279 14210. 0.055690 4474.2 0.017535 n/a n/a n/a n/a n/a n/a

7 1 25 35900. 13004 0.014489 14755 0.016441 29049 0.03237 n/a n/a n/a n/a n/a n/a

7 2 25 22224 18349 0.033024 16767 0.030178 25159 0.045283 n/a n/a n/a n/a n/a n/a

7 3 30 26614 37667 0.047177 50794 0.063619 59438 0.074445 7977.0 0.0099911 12266 0.015363 28712 0.035961

7 4 30 8844.0 10919 0.041154 14485 0.054594 16536 0.062326 1582.1 0.0059630 3256.9 0.012275 6005.9 0.022636

8 1 25 37273 20186 0.021663 13170 0.014134 17297 0.018562 n/a n/a n/a n/a n/a n/a

8 2 25 44277 66945 0.060478 34645 0.031298 71750. 0.064819 n/a n/a n/a n/a n/a n/a

8 3 30 27213 85406 0.10461 57946 0.070978 72752 0.089114 17814 0.021820 5278.2 0.0064653 12131 0.014859

8 4 30 7911.7 26613 0.11212 12071 0.050856 20491 0.086333 4127.5 0.017389 5287.1 0.022275 5018.6 0.021144 * Total volume (µL) of sample loaded on gel a Norm refers to normalized area calculated by: ⁄ b Pre refers to saliva collected prior to the emergency medicine simulaton c Post refers to saliva collected after the emergency medicine simulation n/a: protein band either not sufficiently stained for analysis or there was spillover

150

Appendix C

List of proteins identified by nanoHPLC-ESI-MS/MS from medical resident saliva

151

Accession Description Score Seq # Unique # MW Time point(s) Cov. Peptides Peptides [kDa] P04745 Alpha-amylase 1 1419.47 88.65% 12 77 57.7 Wake, pre- and post- simulation P19961 Alpha-amylase 2B 1199.41 77.50% 2 67 57.7 Post-simulation P02768 Serum albumin 332.95 91.95% 52 91 69.3 Wake, pre- and post- simulation P01876 Ig alpha-1 chain C region 221.3 67.14% 8 30 37.6 Wake, pre- and post- simulation P01877 Ig alpha-2 chain C region 164.64 57.65% 4 26 36.5 Wake, pre- and post- simulation P01037 Cystatin-SN 145.14 92.91% 9 17 16.4 Wake, pre- and post- simulation P01036 Cystatin-S 140.61 97.16% 10 19 16.2 Wake, pre- and post- simulation P01833 Polymeric immunoglobulin receptor 133.37 65.31% 29 50 83.2 Wake, pre- and post- simulation P01834 Ig kappa chain C region 97.67 100.00% 13 15 11.6 Wake, pre- and post- simulation P61626 Lysozyme C 81.56 90.54% 13 21 16.5 Wake, pre- and post- simulation P25311 Zinc-alpha-2-glycoprotein 57.12 68.79% 14 24 34.2 Wake, pre- and post- simulation P22079 Lactoperoxidase 50.05 60.11% 15 35 80.2 Wake, pre- and post- simulation P54108 Cysteine-rich secretory protein 3 47.2 78.37% 12 17 27.6 Wake, pre- and post- simulation P02814 Submaxillary gland androgen-regulated 45.56 65.82% 2 2 8.2 Post-simulation protein 3B P01591 Immunoglobulin J chain 39.34 42.14% 10 10 18.1 Wake, pre- and post- simulation Q8TAX7 Mucin-7 38.53 29.71% 9 14 39.1 Wake, pre- and post- simulation

152

Q9HC84 Mucin-5B 37.96 11.80% 9 32 596 Wake, pre- and post- simulation P12273 Prolactin-inducible protein 36.2 70.55% 6 8 16.6 Wake, pre- and post- simulation Q96DA0 Zymogen granule protein 16 homolog B 36.07 58.17% 8 13 22.7 Wake, pre- and post- simulation P09228 Cystatin-SA 31.72 69.50% 2 12 16.4 Wake, pre- and post- simulation P15516 Histatin-3 27.23 96.08% 3 9 6.1 Post-simulation P01034 Cystatin-C 23.67 60.27% 7 12 15.8 Wake, pre- and post- simulation Q8N4F0 BPI fold-containing family B member 2 21.19 33.41% 7 12 49.1 Wake, pre- and post- simulation P02787 Serotransferrin 19.6 34.53% 6 20 77 Wake, pre- and post- simulation P23280 Carbonic anhydrase 6 19.42 30.19% 6 8 35.3 Post-simulation P02812 Basic salivary proline-rich protein 2 17.94 49.76% 2 3 40.8 Post-simulation P15515 Histatin-1 13.95 85.96% 2 5 7 Wake, pre- and post- simulation P60709 Actin, cytoplasmic 1 12.74 38.13% 5 12 41.7 Wake, pre- and post- simulation P04080 Cystatin-B 12.15 60.20% 3 6 11.1 Wake, pre- and post- simulation P01857 Ig gamma-1 chain C region 12.04 46.97% 3 12 36.1 Wake, pre- and post- simulation P01617 Ig kappa chain V-II region TEW 12.02 100.00% 3 9 12.3 Pre- and post-simulation P01023 Alpha-2-macroglobulin 11.47 23.74% 7 26 163.2 Wake, pre- and post- simulation P28325 Cystatin-D 10.97 86.62% 3 7 16.1 Wake, pre- and post- simulation P04264 Keratin, type II cytoskeletal 1 9.83 31.99% 4 11 66 Wake, pre- and post- simulation P31025 Lipocalin-1 9.28 51.70% 6 7 19.2 Pre-simulation 153

P01623 Ig kappa chain V-III region WOL 8.62 79.82% 3 6 11.7 Wake, pre- and post- simulation P06702 Protein S100-A9 7.97 59.65% 4 6 13.2 Post-simulation P01859 Ig gamma-2 chain C region 7.57 21.78% 2 7 35.9 Pre- and post-simulation P03973 Antileukoproteinase 6.98 54.55% 4 6 14.3 Post-simulation P07339 Cathepsin D 6.55 38.59% 2 8 44.5 Pre- and post-simulation P14780 Matrix metalloproteinase-9 6.4 21.78% 2 12 78.4 Pre- and post-simulation P01781 Ig heavy chain V-III region GAL 6.36 61.21% 2 9 12.7 Wake, pre- and post- simulation P20061 Transcobalamin-1 5.69 21.25% 2 5 48.2 Wake, pre- and post- simulation P07737 Profilin-1 5.68 41.43% 2 5 15 Pre- and post-simulation P23280-2 Isoform 2 of Carbonic anhydrase 6 5.36 29.39% 4 8 35.3 Wake and pre-simulation P09211 Glutathione S-transferase P 5.12 25.71% 2 3 23.3 Pre- and post-simulation Q08380 Galectin-3-binding protein 4.98 21.54% 4 7 65.3 Wake, pre- and post- simulation P02647 Apolipoprotein A-I 4.84 25.47% 3 6 30.8 Pre-simulation P01620 Ig kappa chain V-III region SIE 4.68 24.77% 2 2 11.8 Post-simulation P01593 Ig kappa chain V-I region AG 4.4 31.48% 2 2 12 Wake P06870 Kallikrein-1 4.35 17.94% 2 3 28.9 Wake, pre- and post- simulation P37837 Transaldolase 3.87 32.05% 3 9 37.5 Post-simulation P01765 Ig heavy chain V-III region TIL 3.66 35.65% 2 3 12.3 Pre- and post-simulation Q02818 Nucleobindin-1 2.59 9.98% 2 4 53.8 Wake, pre- and post- simulation P60174-1 Isoform 2 of Triosephosphate isomerase 2.37 22.89% 2 3 26.7 Pre-simulation P01764 Ig heavy chain V-III region VH26 2.35 12.82% 2 2 12.6 Wake, pre- and post- simulation P10909-2 Isoform 2 of Clusterin 2.05 20.36% 2 6 57.8 Wake, pre- and post- simulation 154

Appendix D

Imaging Mass Spectrometry of Normal and Polycystic

Kidney Disease Mouse Kidney Tissues

D.1 Introduction to Polycystic Kidney Disease

As of 2003, polycystic kidney disease (PKD) affected approximately 12.5 million people globally, including 600,000 Americans. The National Kidney Foundation called it the most common life-threatening genetic disease, rendering it the most common life- threating genetic disease.1, 2 PKD results in the formation of cysts in the kidneys.

However, cysts can also develop in the liver pancreas, spleen and ovaries.1 These cysts increase in both size and number as the disease progresses. PKD is commonly diagnosed via ultrasound as it is reliable, non-invasive and inexpensive. CT scans can also be used.1

By age 60, approximately 50% of PKD patients develop kidney failure, and by age 70, approximately 60% of patients experience kidney failure.1 No cure exists for PKD, but some treatments slow down or prevent the loss of kidney function.1, 3

Polycystic kidney disease has two common forms, autosomal dominant polycystic kidney disease (ADPKD) and autosomal recessive polycystic kidney disease (ARPKD).

Its incidence spans all age groups.2 ADPKD is one of the most common genetic disorders

155

and is typically diagnosed in adults with an incidence of 1:400 to 1:1000.2, 4 It is characterized by the development and enlargement of renal cysts that result in kidney failure whose primary treatment is kidney replacement. However, cysts may also develop in other organs including the liver and pancreas.2, 4 Eventually it results in end-stage renal disease. In terms of end-stage renal disease, ADPKD is the third most common cause.5

Diagnosis is typically done by ultrasound imaging of the kidney, computed tomography

(CT) or magnetic resonance imaging (MRI) to visualize cysts and monitor their number and size.2, 4 CT and MRI are commonly used for the detection of cysts in the liver and pancreas. Additionally, MRI is commonly used to measure cyst volume for disease progression determination.5 The guidelines for diagnosis are age dependent. Specifically, the detection of at least three cysts in patients 15 to 39 years is necessary. However, only two cysts are necessary for diagnosis in 40 to 59 years.4 While genetic testing can be used for diagnosis, its use is limited by its cost and that it can detect mutations in up to 85% of individuals with PKD.4

The majority (85%) of ADPKD cases arise from mutations in the PKD1 gene with an incidence of 1:1,000.2, 4 The gene codes for the protein polycystin-1, a 4304 amino acid membrane protein, which is localized to the primary cilium, regulating the levels of intracellular cyclic adenosine monophosphate and mammalian target of rapamycin.2-4 The remaining 15% of ADPKD cases are caused by mutations in the PKD2 gene and occur with an incidence of 1:15,000.4 PKD2 codes for polycystin-2, a 968 amino acid membrane protein, which is also localized with polycystin-1 to the primary cilium. It is also localized in the endoplasmic reticulum and regulates intracellular levels of calcium.2-4 ADPKD as a result of mutations in PKD1 is more severe than the cases that 156

involve PKD2.2 It tends to involve the development of more cysts, not necessarily the growth of the cysts themselves. Additionally, end-stage renal disease typically occurs 20 years earlier in those with PKD1 mutations.2 All cases of ADPKD result in enlargement of the kidney, which can be used to differentiate it from other cystic diseases. A study performed by the Consortium for Radiologic Imaging in the Study of Polycystic Kidney

Disease found that in an eight year follow-up of those with ADPKD as a result of PKD1 mutations, the kidney volume had increased 55%, and the cysts accounted for at least

95% of the kidney volume.4 Those with ADPKD have a five to ten times higher frequency of kidney stones. Hypertension and lower urinary tract infections are also common in patients.4 While it is known that mutations in PKD1 are responsible for the onset of ADPKD, genetic testing is hampered by the large number of mutations associated with the disease. Specifically, 314 mutations in PKD1 and 91 mutations in

PKD2 have already been discovered. Additionally, at least 10% of cases can be traced to a new mutation instead of a hereditary component.2

In comparison, autosomal recessive PKD is typically diagnosed in utero or during the neonatal period. However, it can be diagnosed through adulthood. It has an incidence of 1:20,000.2, 4 Additionally, up to 30% of the cases of ARPKD die by the neonatal period, and renal replacement is necessary in up to one-third of child cases.2 ARPKD is caused in mutations in the gene PKHD1. It has 303 different mutations described, and one-third of the mutations are unique to a single family.2 The gene codes for the 4074 amino acid integral membrane protein, fibrocystin (also called polyductin), which is located in the primary cilium and is a receptor to regulate intracellular levels of cAMP.2, 4

Much like ADPKD, ARPKD is characterized by renal and liver cysts, but it has variable 157

severity depending on the mutation. Likewise, enlargement of the kidneys is a trademark symptom. Unlike ADPKD, progression to end-stage renal disease in the first decade occurs in 50% of ARPKD patients.4 Up to 80% of children with ARPKD also present hypertension.4

As with many other diseases, treatment regimens are most effective the earlier the diagnosis is made. Consequently, it would be optimal to diagnosis PKD in its early stages before symptoms become apparent. Given the current diagnostic techniques of kidney cyst imaging, this is challenging. Therefore, the discovery of biomarkers to aid in the diagnosis of PKD at an early, asymptomatic stage is essential. In order to discover putative protein biomarkers of early stage PKD, MALDI mass spectrometry imaging

(MALDI-MSI) has been utilized.

D.2 Principle of MALDI-MSI

Mass spectrometry imaging (MSI) is also commonly referred to as imaging mass spectrometry, but to avoid any abbreviation confusion with ion mobility spectrometry, the latter will not be used in this document. Mass spectrometry imaging is used to determine the composition, spatial distribution and relative abundance of biomolecules within thin tissue sections including peptides, proteins, lipids, xenobiotics and small molecules.6-10 While different ionization sources have been used, MALDI-MSI was first reported by Caprioli et al. in 1997.6 MALDI-MSI has several important clinical implications. Two of the most important ones include that it can be used in conjunction with histopathology for diagnosis, and it may aide in the early detection of disease. For example, it is useful in differentiating tumor boundaries where a normal histology may be

158

present.10, 11 Previous studies include the profiling of human gliomas, lung tumor biopsies, neuropeptides and mouse epididymis.12 In terms of MALDI-MSI of proteins, there are three main applications in regard to disease analysis. It can be used to distinguish healthy from non-healthy (e.g., tumor) tissues or to play a role in pathological diagnosis (i.e., short-term or long-term survival). Additionally, it may be used to predict a patient’s response to a specific drug treatment.13

The general MALDI-MSI procedure includes sectioning of either a resected organ or the entire body of a small organism (e.g., mouse or rat). Once sectioned, a washing step may be performed to minimize ion suppression by interfering compounds, such as salts. Afterwards, matrix is applied to the section and a set coordinate system is defined across the section to determine the laser raster positions for mass spectra acquisition.

Each x, y position in the coordinate system can be thought of as a pixel to generate images and density maps of selected ions across the tissues.9, 14 Additionally, the data can be processed to remove background noise and perform statistical analysis to differentiate tissue types. These steps will be described in further detail in the following sections for

MALDI-MSI of proteins and peptides.

There are several advantages of MALDI-MSI. For instance, compared to other proteomic techniques for biomarker discovery, MSI has the advantage that it preserves the spatial distribution of the analytes.15 Additionally, no prior knowledge of the sample is necessary.14 Consequently, in comparison to other discovery techniques, MALDI-MSI does not require target-specific markers.16 This also allows one to monitor many potential biomarkers simultaneously.17 It is capable of localizing hundreds of analytes in a single

159

experiment.18 Furthermore, local high concentrations of an analyte are preserved that would otherwise be lost if the tissue were subjected to homogenization.19

A limitation of MALDI-MSI of proteins is that it only provides the molecular weights of the proteins, not identification. Further analyses employing protein extraction and digestion for a ―bottom-up‖ approach must be performed to determine the identity of the proteins.9, 20 Another major limitation is that MALDI-MSI typically only detects proteins with an upper limit of 30-50 kDa.12, 17, 21 However, with special tuning of the instrument and specific sample preparation methods, this range can be extended.19

D.2.1 Sample Preparation

In order to generate high-quality images, one must pay careful attention to sample preparation and handling. Sample preparation encompasses all parts of the imaging workflow, beginning with euthanasia of the animal and resection of the organ(s) or tissue(s) of interest. Once removed, the tissues need to be properly preserved and sectioned. In order to minimize ion suppression and remove contaminants, the sections are typically washed for protein and peptide analyses. If a bottom-up approach is to be utilized, the enzyme also needs to be applied to the tissue section before matrix application.

After euthanasia, the tissue samples must immediately be preserved, typically by flash freezing, to prevent degradation of the analytes by endogenous enzymatic activity and delocalization of the analytes. At the same time, care must be taken to not deform the shape of the tissue during the freezing process. Specifically, small organs can be directly submersed in liquid nitrogen, whereas large organs can be frozen using dry-ice chilled 160

isopentane.7 During freezing, care should also be taken to avoid cracking or shattering of the tissue.17 In a complementary method, tissues can be formalin-fixed and paraffin embedded (FFPE). However, the main issue of MALDI-MSI of FFPE tissues is that extensive covalent crosslinking of the proteins occurs.7 This crosslinking is the result of methylene bridge formation, especially at primary amines.15, 20 The crosslinking allows

FFPE tissue may be stored at room temperature for several decades though.17, 20

For MALDI-MSI, the tissues must be thinly cut using a cryostat. Frozen tissues stored at -80 °C should be transported using dry-ice to minimize temperature differences that may induce freeze-thaw cycles.7 Multiple parameters regarding tissue sectioning should be considered. First, sectioning of the tissue allows the cells to be cut open to expose their components. The thickness of the sections is typically 3-20 μm for optimal results.17 It is easier to manipulate thicker sections. However, thick sections may not be electrically conductive, which is necessary for MALDI-MSI, and they may take longer to dry.22 However, too thin of tissues tear easily. Mammalian tissues are typically sectioned between 10 and 20 μm.23 Depending on the tissue being sectioned, the cryostat is generally set to between -12 and -30 °C. 17 In general, the higher the fat content, the lower the necessary temperature.22

Ideally, tissues should not be embedded for sectioning to avoid contamination. If the tissues need to be embedded, pure water gelatin or carboxymethylcellulose are recommended.7 Ice and gelatin are other mass spectrometry amenable media used for tissue embedding.22 However, optimal cutting temperature (OCT) media is traditionally used by histologists. It preserves both the quality and shape of the tissue, allowing it to be

161

cut to less than 5 μm thickness.17 OCT has several drawbacks for MSI as it easily contaminates the cutting blade and consequently the sample. Additionally, it is readily detected by mass spectrometry resulting in severe ion suppression and may cause mass interference.7, 17 If OCT is used, it should be removed by fixing the sectioned tissue using ethanol followed by washing with water.7 It can also be trimmed away.15 To avoid these drawbacks, a small drop of it can be used to attach frozen tissues to a cryostat anchor instead of fully embedding the tissue.17, 24 To correlate imaging data with histology data, adjacent tissue sections are typically used. This allows stains that are not mass spectrometry compatible, such as hematoxylin and eosin (H&E), to be used.

Alternatively, histological staining may be performed on the same section after MALDI-

MSI.11, 16

The tissue sections must be mounted on a conductive surface for MALDI-MSI.

Different types of surfaces can be used including glass slides, metal targets and metal- coated glass slides.7 The most common metal-coated glass slides are coated with indium- tin oxide (ITO), which retains the optical transparency of the glass slides. 7, 17 Of the various methods to attach the tissue sections, thaw mounting is one of the most common.

In this technique, the target is kept in the cryostat, and the section is transferred onto the slide using an artist’s brush where the slide surface is gently warmed. Specifically, once the tissue section is placed on the slide, a finger is applied to the underside of the slide which melts the tissue onto the slide.12, 23 Once sectioned and mounted, the tissues should be dried in a vacuum desiccator for 30 minutes to remove water, preserving the tissue and preventing analyte delocation.15, 20, 22

162

After sectioning, the tissues are typically processed to remove any analytes, such as salts and lipids, that cause ion-suppression of proteins and peptides. Different solvents can be used. Common washing protocols to remove salts and lipids used chilled 70-100% ethanol.7 One protocol that gives good results for protein and peptide analyses is washing with 70% isopropanol followed by washing with 90% isopropanol.17 An alternative protocol washes the tissue with 70% ethanol to remove salts and debris followed by washing in 90-100% ethanol to dehydrate and fix the tissues. For peptide analyses, a final

30 second wash of 90% ethanol, 9% glacial acetic acid, 1% deionized (DI) water is included.22 This step also serves to fix the proteins.10, 15 Washing with organic solvents removes lipids, increasing the sensitivity of protein analyses.19

Additionally, enzymes or derivatizing agents may also be used. For in situ enzymatic digestion, trypsin is either sprayed or microspotted on the tissue prior to matrix application. This methodology enables detection of large molecular weight or membrane bound proteins.10 However, the tissue should be at its biological hydration level for enzymatic digestion. This may result in diffusion of the proteins or peptides. It may also allow endogenous enzymes to resume their activity.7 At this point, matrix can be applied.

D.2.2 Matrix Application

Another critical step in the imaging workflow is matrix application. Specifically, the matrix should be deposited homogenously across the tissue surface. Matrix application also should not translocate the proteins in the thin tissue section while it wets the surface in order to co-crystallize with the proteins.6 Another important consideration

163

is that the image resolution for imaging should be greater than the size of the crystals.6

Using common techniques, the minimum crystal size is typically 10 μm.22

A number of different matrices for MALDI-MSI exist, but they all follow the same general guidelines. Ideally, a quickly evaporating solvent should be used for matrix application. Common solvents include acetonitrile, ethanol, methanol and isopropanol.7

For positive ion analyses, 0.1-1% TFA or a small amount of acetic acid is typically added.17, 22 For peptide analyses, 10-20 mg/mL of CHCA or 40 mg/mL of DHB are commonly used. DHB is also commonly used for lipid analyses.19, 22 For protein analyses, 10-30 mg/mL of SA is typically used.7, 22

Different methods of matrix application are utilized including manual and automated techniques. The primary matrix application techniques are spotting, spraying and sublimation. Spotting can be performed manually using a pipette or in an automated manner using a robot.15 Chemical printers can also be used for matrix deposition. Other manual deposition methods include using an artist paintbrush or a TLC sprayer.17

Commercial automated sprayers are used to spray a fine aerosol of matrix over the tissue section(s).17 Of these techniques, spraying typically results in the smallest matrix crystals, allowing greater imaging resolution. However, care needs to be taken as over wetting the tissue should be avoided as it can lead to analyte delocalization.15, 22 Another benefit of spraying is that automated sprayers are reproducible and result in uniform deposition of matrix across the tissue.22

The automated sprayer utilized in this study is Bruker’s ImagePrep device. This device vibrationally vaporizes the matrix within a closed chamber with a controlled

164

humidity level. The piezo-electric spray head moves a pinhole sheet next to a reservoir containing the matrix solution, ejecting small droplets.25 The average diameter of the matrix droplets is 20 μm.25 Therefore, using this methodology an imaging resolution of

20 μm can be obtained.17 Additionally, the thickness of the deposited matrix is observed by a light-scattering device within the chamber.17 This optical scattering-light sensor monitors the tissue wetness, matrix thickness and drying rate.25

D.2.3 Data Acquisition

For MALDI-MSI, triangulation marks need to be applied. Correction fluid is typically used to apply triangulation marks. Then, optical images must also be acquired.7

The purpose of the optical image is twofold. First, it allows comparisons between the image of the tissue and the mass spectrometry data. Second, it is necessary for the mass spectrometer to determine the location of the sample to be analyzed.7

For the MALDI-MSI experiment, a x,y coordinate system is generated. Each spot within this coordinate system is irradiated and ionized by the MALDI laser to generate a mass spectrum.15, 17 The surface of the tissue section is moved across the MALDI stage in fixed steps, and the laser beam desorbs and ionizes the sample.26 Each coordinate represents a pixel.17 The size of the pixel is determined by the spatial resolution, and the intensity of the pixel is a function of the ion intensity.19 Depending on the size of the matrix crystals and laser, the lateral resolution of MALDI-MSI is 10-100 μm.15 For large tissues, such as intact organs (10 mm x 12 mm), a spatial resolution of 75-200 μm is commonly used.19 In general, the mass range for imaging experiments ranges from m/z

400 to a maximum of m/z 30,000.11 The acquired data can be represented in 3D space.

165

D.2.4 Data Normalization

A x,y coordinate system is also used for data processing. In the x,y coordinate system is the m/z signal and the intensity of the ion which is proportional to its abundance. The intensity of each ion is visualized as an intensity plot using shades of grey or with different colors.10, 18 Bruker’s FlexImaging software uses color coding to visualize the ion distribution. It also is capable of overlaying the optical image with the

MS image.22 These imaging maps can be correlated to histology staining to determine anatomical structure. They can also be used in conjunction in which a pathologist who defines a region of interest.19

Once the data has been acquired it needs to be processed for reliable interpretation. Specifically, smoothing and background subtraction are used to minimize noise and variation across the section that is derived from sample preparation.19, 27

Normalization reduces the intensity axis range to a common intensity scale.28 The signal intensity depends on several factors including expression levels, ionization efficiency and desorption differences.9 Normalization removes systematic artifacts which can be introduced at multiple steps in the MALDI-MSI process. One of the primary sources of these artifacts is that the ion source becomes contaminated during the experiment resulting in a gradual decrease in ion transmission as the spectra acquisition progresses which causes image fading.28 Chemical inhomogeneities can also arise due to salt, pH gradients and other similar sources. Systematic artifacts may result from matrix crystal differences. However, normalization cannot correct for ion suppression.28

166

Several methods exist for normalization, the most common of which is total ion current (TIC).19 In this method, all the spectra are divided by their TIC.28 It also accounts for global effects, such as global intensity changes.27 However, this method assumes that there are a comparable number of signals present in each spectrum with a relative constant background.28 One of the benefits of normalization by TIC is that it is less prone to producing artifacts compared to other normalization methods.28

D.2.5 Statistical Analysis

In order to use imaging to differentiate between healthy and diseased tissue for biomarker discovery, statistical analyses must be employed. Overall, statistical analyses can be performed in either a supervised or an unsupervised manner. Supervised analysis is performed using prior knowledge regarding the samples to make a model. On the other hand, unsupervised analysis is used to find the unknown variables present in the data that may reveal unknown characteristics about the samples.27 Two of the most common statistics techniques used for imaging data is hierarchical clustering and principal component analysis, each of which will be discussed in further detail.

D.2.5.1 Hierarchical clustering

Hierarchical clustering is considered an unsupervised method of classification. It uses pair-wise clusters to create a dendogram that contains all the mass spectra.25 Each dendogram branch represents a class of spectra.25 As one progresses down the branches, mass spectra with the most similarity and more random differences are grouped together.29, 30 The top branches or levels will therefore have the largest differences and contain classes for different tissue types (e.g., tumor vs. non-tumor regions).29, 30

167

Therefore, by following branches in the dendogram, differences in spectra are explained.

This type of analysis can also be easily correlated to histological staining results.30

D.2.5.2 Principal Component Analysis

Another primary method used is principal component analysis (PCA). As with hierarchical clustering, it is considered an unsupervised technique.30 However, one of the drawbacks of PCA compared to hierarchical clustering is that it does not classify the spectra.25 PCA is a multivariate technique that is used to reduce the dimensionality of the data by transforming it from a relationship of peak intensities to the variance of the dataset (i.e., principal components).11, 25 This method removes redundancy from the data by grouping variables with high covariance together.27 The data obtained from MALDI-

MSI experiments are considered multidimensional where each m/z is considered one dimension.30 For PCA, a mass spectrum with n peaks is transformed into an n- dimensional coordinate system. In this system, each axis represents one peak with its corresponding intensity.30 Specifically, the x dimension is the peak intensity at measured mass 1. The y dimension is the intensity of the peak at measured mass 2, and the z dimension is the intensity of the peak at measured mass 3.27 This process is repeated for all the acquired mass spectra creating an n-dimensional space cloud in which similar spectra are spatially close together. This coordinate system is transformed by shifting and rotating the coordinates until one axis has the direction of the highest variance.30 This axis contains the most variance in the data, and is referred to as the first principal component (PC1).27, 30 It also typically correlates well with histology differences. This process is repeated such that the second axis has the second most variance. This second principal component is orthogonal to PC1 which maximizes the remaining variance.27, 30 168

Loading plots can also be constructed to show the contribution of the original coordinate system (i.e., peak masses) to each of the principal components. If these plots are differentiated into two separate regions, they show which peaks in the original spectra contributed to the difference.27

D.3 Experimental

D.3.1 Chemicals and Materials

HPLC grade acetonitrile, methanol and water were purchased from Fisher

Scientific (Pittsburgh, PA). Glacial acetic acid, hydrochloric acid, hematoxylin gill 3x and eosin Y were also from Fisher. Formic acid (FA), trifluoroacetic acid (TFA), ammonium bicarbonate, ammonium phosphate monobasic (NH4H2PO4) and proteomics grade, dimethylated trypsin were purchased from Sigma (St. Louis, MO). The matrices

DHB, CHCA and SA along with ITO glass slides were from Bruker Daltonics (Bremen,

Germany). A standard bovine cytochrome c tryptic digest was acquired from Dionex

(Sunnyvale, CA). Histoclear was purchased from National Diagnostics (Atlanta,

Georgia), and mounting medium was from Vector Laboratories (Burlingame, CA). Both

C18 ZipTips (EMD Millipore, Billerica, MA) and StageTips (Thermo, West Palm Beach,

FL) were used.

Additionally, a water filtration system in the pharmacology department was used to obtain MilliQ water. A Leica CM1950 cryostat (Wetzlar, Germany) was used for sectioning. Bruker’s ImagePrep was used for application of matrix and trypsin. Bruker’s

MALDI UltrafleXtreme mass spectrometer was used. It is equipped with a pulsed

Smartbeam II Nd:YAG 355 nm laser. For peptide analyses, Bruker’s peptide calibration

169

standard was used. It contains angiotensin II (m/z 1046.5418), angiotensin I (m/z

1296.6848), substance P (m/z 1347.7354), bombesin (m/z 1619.8223), ACTH clip 1-17

(m/z 2093.0862), ACTH clip 18-39 (m/z 2465.1983) and somatostatin 28 (m/z

3147.4710). For protein analyses, Bruker’s protein standard I was used. It contains insulin (m/z 5734.51), ubiquitin (m/z 8565.76), cytochrome c (m/z 6180.99 and 12360.97) and myoglobin (m/z 8476.65 and 16952.30).

D.3.2 Animal Samples

Normal and PKD kidney tissues from six month old mice were provided by Dr.

Surya Nauli’s laboratory (Department of Pharmacology, University of Toledo). The mice were sacrificed, and the kidneys were excised following a previously approved IRB protocol in their lab. The excised kidneys were flash frozen and stored at -80 °C until sectioning.

D.3.3 Tissue Preparation

The mouse kidney tissues were prepared for MALDI-MSI in accordance with

Andersson et al.31 Details of the procedure are described by in Yang Xu’s thesis.32 An overview of the MALDI-MSI workflow is provided in Figure D-1. In brief, First, the frozen kidney was mounted on a cryostat block holder using a minimal amount of Cryo-

Gel embedding medium. Both the normal and early-stage PKD kidneys were sliced to 10

μm thickness. However, the late-stage PKD kidneys were sliced to 15 μm thickness to avoid tearing of the sections as the large cysts made the tissue more fragile. The sections were transported on dry ice before storage at -80 °C until further analyzed.

170

Figure D-1 Overview of the MALDI mass spectrometry imaging workflow for the analysis of proteins and peptides

Before use, the stored samples should be gradually warmed to room temperature prior to washing as described by Yang Xu.32 Afterwards, a fixation and washing step to remove salts, lipids and other small molecules was performed using a series of alcohol washings. First, the kidney sections were washed with 70% ice-cold isopropanol for 30 seconds, followed by drying in a vacuum desiccator. Next, the sections were washed in

95% ice-cold isopropanol for 30 seconds, followed by drying. This step was repeated. If a trypsin digest for peptide analysis is to be performed, these washes were decreased to 10

171

seconds. A final wash using ice-cold 90% isopropanol, 9% glacial acetic acid and 1%

HPLC grade water with subsequent drying was also used for peptide analysis.

In-situ trypsin digestion was performed on several of the kidney sections for peptide analysis. Before matrix application, teaching marks for the FlexImaging software were placed at the corners of each kidney tissue section on the slide using a white-out marker pen. Additionally, the appropriate calibration mixture (either peptide or protein) was spotted on the slide. The final step before MALDI-MSI was matrix application.

These steps are described in greater detail by Yang Xu.32

D.3.4 MALDI-MS Imaging

For MALDI-MSI, the scanned images of the kidney tissue slides were imported into the FlexImaging software (version 3.0, Bruker) in order to appropriately locate the sections for imaging. The edge of the kidney tissue was defined as the area to be imaged.

For protein imaging, mass spectra were acquired from m/z 3000 to m/z 30,200. For peptide imaging, mass spectra were acquired from m/z 800 to m/z 4,000. Prior to data acquisition, the instrument was calibrated using the spotted peptide or protein calibration mixture. A raster width of 100 µm was used, and 500 laser shots were accumulated per spot with a 500 hertz repetition frequency (FlexControl version 3.3, Bruker). The spectra were analyzed using FlexImaging and ClinProTools (version 2.2, Bruker) to perform peak smoothing and baseline subtraction, to create distribution and density maps, and to perform PCA.

172

D.3.5 nanoHPLC-MALDI-MS/MS

After imaging, the tryptic peptides from the kidney sections were extracted from the tissue on the ITO slides and purified prior to nanoHPLC-MALDI-MS/MS for protein identification. For extraction, 20 μL of 0.1% TFA was applied to each kidney section on a slide. The pipette tip was used to help break up the kidney tissue. The liquid from all six sections on a single ITO slide were combined together. The solution was centrifuged for

20-30 minutes at 8000 rpm to remove any cell debris. Then, the samples were concentrated using a vacuum evaporator to remove excess solvent. After drying, the peptides were reconstituted in 10 μL of 0.1% TFA. Afterwards, the samples were cleaned up to remove matrix, salts and other small particulates by using ZipTips for solid-phase extraction. Briefly, the ZipTip was wetted using 10 μL of a wetting solution containing

100% acetonitrile, aspirating the solution into the tip five times. Then, the tip was equilibrated by passing equilibration solution of 0.1% TFA over the tip three times. At this point, the sample was aspirated and dispensed for three to ten cycles to maximize binding of the peptides to the media. The sample was subsequently washed using five cycles of the wash solution of 0.1% TFA. Finally, the peptides were eluted using an acetonitrile gradient. First, they were eluted using 10 μL of 30:70 (v:v) acetonitrile:H2O with 0.1% TFA, followed by 10 μL of 70:30 (v:v) acetonitrile:H2O with 0.1% TFA and

10 μL of acetonitrile. The eluted samples were combined together for nanoHPLC-

MALDI-MS/MS analyses.

StageTips were also used and gave similar results to purification using the

ZipTips. The StageTip was wetted using 20 µL of 80% acetonitrile, 5% TFA. Re- equilibration of the tip was performed using 5% TFA. The peptides in 5% TFA were 173

loaded twice to maximize binding. Then the sample was washed using 5% TFA. Desalted peptides were eluted using 20 µL of 80% acetonitrile, 5% TFA. For nano-HPLC analyses, 10 µL of the desalted sample was diluted to 50 µL with 0.1% TFA.

The extracted tryptic mouse kidney peptides were separated using an Ultimate

3000 nanoHPLC (Dionex) system equipped with an Acclaim PepMap 100 C18 column

(75 µm x 15 cm, 2 µm, 100 Å). Mobile phase A was H2O + 0.05% TFA, and mobile phase B was 0.05% TFA in 80:20 (v:v) acetonitrile: H2O. The column oven was maintained at 25 °C. The flow rate was 200 nL/min, and eluting peptides were monitored using a UV detector at 214 nm. Gradients were optimized for each sample. One common separation method consisted of a linear gradient from 2-60% B in 60 min, followed by

60-100% B in 1 minute.

The nanoHPLC system was connected to a Proteineer fc (Bruker) spotting device for MALDI-MS/MS analyses. This spotting device utilizes a micro tee chamber to mix the eluting peptides with CHCA matrix supplied from an external syringe pump at a rate of 100 µL/hour. The CHCA matrix consisted of 712 µL of 95:5 (v:v) acetonitrile: H2O with 0.1% TFA, 72 µL of saturated CHCA in 90:10 (v:v) acetonitrile: H2O with 0.1%

TFA, 8 µL of 10% TFA and 8 µL of 100 mM NH4H2PO4. Every 15 seconds, the

Proteineer fc spotted the eluting peptide and matrix mixture onto a 384-spot Anchorchip target plate (Bruker). Before analyses, 0.1 µL of a peptide calibration mixture was applied to the calibration spots on the plate. The calibration mixture consisted of 0.6 µL of the peptide standard and 5.4 µL of the aforementioned CHCA matrix.

174

The spotted, separated peptides were analyzed by MALDI-MS/MS using an automated sequence setup in the WARP software (version 1.2, Bruker). Calibration was performed prior to the automated analysis. For MALDI-MS, reflection and positive ion modes were used with a detection range of m/z 400 to m/z 3,800. Using the WARP software, selection of ions for MS/MS was determined using preset parameters in which any m/z ion with a signal-noise ratio greater than 20 was selected. MS/MS spectra were analyzed using BioTools (Bruker) for a Mascot (Matrix Sciences) database search of Mus musculus proteins. Methionine oxidation was set as a variable modification, and two missed cleavages were allowed.

D.4 Results and Discussion

D.4.1 Normal and PKD Kidney Protein Distribution Imaging

For each kidney tissue section imaged, TIC normalization was utilized for background smoothing and subtraction. An average spectrum for each kidney section was generated using FlexImaging. This average spectrum can be used to construct protein mass images to depict the localization of various proteins across the kidney. Overall, no significant localization of the proteins was observed in either normal or early-stage PKD mouse kidney tissues. Most of the proteins were present fairly uniformly across the entire tissue section. Representative protein distribution images for these tissues have been shown by Yang Xu.32

A basic normalization of the average spectra was performed to compare disease states. This normalization was achieved by dividing the intensity of each m/z by the intensity of the base peak, such that intensity of the base peak was normalized to one.

175

This normalized average mass spectrum was compared to other kidney tissue sections mounted on the same ITO slide that were from the same region of the kidney to verify the reproducibility of the imaging procedure. For example, six normal kidney tissue sections mounted on the same slide were compared. As shown in Figure D-2, these six average spectra were nearly identical. Specifically, the majority of observed peak masses were the same, and the peak masses had a similar intensity across the sections, which is in good agreement with previous results.32 Average mass spectra from different disease states were also compared, and several differences in the peak masses and ion intensities between the normal and early-stage PKD tissues were observed.

176

Figure D-2 Normalized average spectra (A-F) of six normal mouse kidney tissues from the same region of the kidney mounted on the same ITO slide

D.4.2 Normal and PKD Kidney Peptide Distribution Imaging

Proteins with a mass greater than 16,000 Da were not readily detected.

Furthermore, direct protein identification using only MS spectra is challenging.

177

Consequently, an on-tissue enzymatic digest using trypsin was performed to facilitate protein identification. Tissue sections were sprayed with a solution containing trypsin prior to matrix application.

Peptide distribution images were also constructed as density maps to determine the localization of various peptides. Peptide density maps for a female normal kidney tissue digested with trypsin and sprayed with DHB matrix were generated. Several peptide ions were determined to be distinctly localized to particular regions of the kidney tissue (e.g., pelvis, medulla, cortex and medullary-cortex junction) in accordance with earlier results.32

D.4.3 PCA

In order to discover putative biomarkers of PKD, a large scale analysis of all the collected spectra from tissue sections, not just the average spectra, should be evaluated.

However, this process is impractical as it becomes very time-consuming owing to the thousands of spectra that are typically generated for a single tissue section. Another major constraint is that the person performing this analysis should be proficient in both mass spectrometry and pathology. However, these constraints can be overcome using PCA to perform statistical analysis in order to determine meaningful differences between normal and diseased tissues.

PCA was performed using ClinProTools (Bruker), whose main limitation is the size of the dataset that can be analyzed was limited by the available computing power.

Using this software only two classes, or two MS imaging data sets, were able to be compared at a time by the software. This gave three possibilities of tissue type 178

combinations to compare: normal vs. early-stage PKD, normal vs. late-stage PKD, and early-stage vs. late-stage PKD.

Initially, normal and late-stage PKD mouse kidney tissues were compared using

PCA. All the acquired spectra from a single normal kidney section and from a single late- stage PKD kidney section from the same region were subjected to PCA. The principal components and loading plots are depicted in Figure D-3. These two datasets were well separated by the first two principal components (PC1 and PC2) indicating distinguishable features existed in their respective mass spectra (Figure D-3A and B). In order to determine which peak masses contribute the most to the separation, loading plots were also created (Figure D-3C and D). In the loading plot, each spot represents a single m/z value. The distance of the spot to the origin denotes the contribution of the ion to the separation of the datasets, where an increased distance indicates an increased contribution. For example, m/z 3698.95 does not greatly contribute to the separation between normal and late-stage PKD tissues as it is relatively close to the origin.

Conversely, m/z 3360.47 and 14981.49 are clearly separated as belonging to normal kidney tissue, and m/z 954.29 clearly belongs to late-stage PKD mouse kidney tissue.

These contributions were confirmed by comparing the respective peaks in the average mass spectra.

179

Figure D-3 PCA of MALDI-MSI data. All spectra from a single healthy kidney section and from a late-stage PKD kidney section were loaded for PCA. (A) PCs in 3D plot, (B) PCs in 2D plot of the first two principal components, (C) 3D loading plot, (D) 2D loading plot

D.4.4 nanoHPLC-MALDI-MS/MS of Extracted Peptides

In order to identify the proteins present in the mouse kidney tissues and to determine putative biomarkers of PKD, a bottom-up proteomics method was used. In this method the tissue sections were digested with trypsin prior to MALDI-MSI. Once imaging was complete, the tryptic peptides were extracted from the tissue sections and purified using solid-phase extraction to remove any remaining salts and the matrix. These

180

extracted peptides were subjected to nanoHPLC for separation, and the eluting peptides were co-spotted with CHCA matrix on a MALDI plate. MALDI-MS/MS was performed on the collected fractions, and the spectra were subjected to database searching using

MASCOT for protein identification.

From most sections, very few peaks were present in the chromatogram of the nanoLC separation. The UV chromatogram at 214 nm from peptides extracted from six imaged sections of a normal, male mouse kidney is shown in Figure D-4. A linear gradient of 2-60% B in 60 minutes was used for peptide separation, followed by a 10 minute wash at 100% B to elute hydrophobic peptides. Overall, a relatively low number of peptides from the kidney tissue were efficiently extracted. However, several peptide peaks can be observed on the chromatogram after 35 minutes. Consequently, fractions were collected from 30 to 75 minutes with spotting on the MALDI plate at 15 seconds per spot. These spots were analyzed by MALDI-MS/MS for peptide identification.

181

Figure D-4 nanoHPLC of peptides extracted from six imaged sections of normal mouse kidney tissue Thus far, only two abundant proteins have been identified from the nanoHPLC-

MALDI-MS/MS analyses of the mouse kidney extracts. The MS and MS/MS spectra for one identified protein, hemoglobin subunit alpha, are shown in Figures D-5 and D-6, respectively. In the MALDI-MS, only two major ions are present, m/z 568.2 and m/z

1529.8. The ion at m/z 568.2 is likely due to the matrix. The ion at m/z 1529.8 was selected for MS/MS. The MS/MS spectrum (Figure D-6) is annotated with the b and y fragment ions. A MASCOT search of the Mus musculus proteins identified the peptide ion with m/z 1529.8 as hemoglobin subunit alpha with a score of 137 where ion scores greater than 29 indicate identity or extensive homology. The complete y series of ions is observed in the MS/MS spectrum as well as the majority of the b series of ions. A peptide ion with m/z 1168.7 was also selected for MS/MS (Figure D-7). It was identified as a peptide from cytochrome c, somatic with a score of 33 where ion scores greater than 28 indicate identity or extensive homology.

182

Figure D-5 nanoHPLC-MALDI-MS of the fraction containing the peptide ion with m/z 1529.8 which was selected for MS/MS analysis for protein identification

Figure D-6 nanoHPLC-MALDI-MS/MS of the peptide ion with m/z 1529.8 identified as hemoglobin subunit alpha by Mascot

183

Figure D-7 nanoHPLC-MALDI-MS/MS of peptide ion with m/z 1168 identified as cytochrome c by Mascot

D.5 Conclusions and Future Directions

Proteins and tryptic peptides were imaged from normal and PKD mouse kidney tissues. Overall, proteins with MWs <20 kDa were efficiently ionized by MALDI-MSI.

From these imaging experiments, distribution and density maps were generated. Many of the proteins were detected across the entire kidney tissue section, and very little localization to specific regions of the kidney was observed. In contrast, several of the imaged tryptic peptides from normal mouse kidneys could be localized to the pelvis, medulla, cortex or medullary-cortex junction regions of the kidney.

Additionally, average mass spectra of proteins originating from normal and PKD mouse kidney tissues were compared. Spectra from the multiple sections mounted on a 184

single ITO slide were fairly reproducible exhibiting the same protein ions in the same relative intensities. However, differences between peak masses and ion intensities were observed between normal and PKD mouse kidney tissues. These differences were also seen using PCA. PCA clearly grouped protein spectra corresponding to normal and late- stage PKD mouse kidney tissues. The loading plots showed which ions were responsible for the differentiation of spectra as either normal or PKD.

As protein identification by MALDI-MSI is challenging, nanoLC-MALDI-

MS/MS was performed on peptides extracted from the imaged sections. A relatively low number of peptides were extracted from kidney tissue sections. Despite, this low extraction efficiency from the ITO slides, two abundant proteins (i.e., hemoglobin subunit alpha and cytochrome c) were identified using this bottom-up proteomics workflow. The complementary technique of coupling nanoHPLC to ESI-MS/MS for online detection may produce additional identifications.

Overall, these results indicate that MALDI-MSI and PCA can be used to further understand differences in protein expression leading to PKD biomarker discovery. The current analyses should also be expanded to compare normal and early-stage PKD mouse kidney tissues to facilitate detection before cysts are detectable in the kidneys. To aide in the identification of proteins observed by MALDI-MSI, more efficient peptide extraction techniques are needed. In addition, differences seen by PCA may be confirmed using label-free quantification of the tryptic peptides using nanoHPLC-ESI-MS/MS.

185

Appendix References

1. Polycystic Kidney Disease; 2003.

2. Harris, P. C.; Torres, V. E. Polycystic Kidney Disease. In Annual Review of Medicine, Annual Reviews: Palo Alto, 2009; Vol. 60, pp 321-337.

3. Torres, V. E.; Harris, P. C. Polycystic kidney disease: , proteins, animal models, disease mechanisms and therapeutic opportunities. Journal of Internal Medicine 2007, 261 (1), 17-31.

4. Rizk, D. V.; Chapman, A. B. 41 - Polycystic and Other Cystic Kidney Diseases A2 - Gilbert, Scott J. In National Kidney Foundation Primer on Kidney Diseases (Sixth Edition), Weiner, D. E., Ed. W.B. Saunders: Philadelphia, 2014; pp 362- 370.

5. Arif-Tiwari, H.; Kalb, B.; Semelka, R. C.; Martin, D. R. 6 - Kidney Imaging A2 - Gilbert, Scott J. In National Kidney Foundation Primer on Kidney Diseases (Sixth Edition), Weiner, D. E., Ed. W.B. Saunders: Philadelphia, 2014; pp 51-59.

6. Chaurand, P.; Schwartz, S. A.; Caprioli, R. M. Imaging mass spectrometry: a new tool to investigate the spatial organization of peptides and proteins in mammalian tissue sections. Current Opinion in Chemical Biology 2002, 6 (5), 676-81.

7. Goodwin, R. J. Sample preparation for mass spectrometry imaging: small mistakes can lead to big consequences. Journal of Proteomics 2012, 75 (16), 4893-911.

8. Caprioli, R. M.; Farmer, T. B.; Gile, J. Molecular imaging of biological samples: localization of peptides and proteins using MALDI-TOF MS. Analytical Chemistry 1997, 69 (23), 4751-60.

9. Groseclose, M. R.; Andersson, M.; Hardesty, W. M.; Caprioli, R. M. Identification of proteins directly from tissue: in situ tryptic digestions coupled 186

with imaging mass spectrometry. Journal of Mass Spectrometry 2007, 42 (2), 254-262.

10. Chaurand, P. Imaging mass spectrometry of thin tissue sections: a decade of collective efforts. Journal of Proteomics 2012, 75 (16), 4883-92.

11. Franck, J.; Arafah, K.; Elayed, M.; Bonnel, D.; Vergara, D.; Jacquet, A.; Vinatier, D.; Wisztorski, M.; Day, R.; Fournier, I.; Salzet, M. MALDI imaging mass spectrometry: state of the art technology in clinical proteomics. Molecular & Cellular Proteomics 2009, 8 (9), 2023-33.

12. Chaurand, P.; Schwartz, S. A.; Billheimer, D.; Xu, B. J.; Crecelius, A.; Caprioli, R. M. Integrating Histology and Imaging Mass Spectrometry. Analytical Chemistry 2004, 76 (4), 1145-1155.

13. Schwamborn, K.; Caprioli, R. M. Molecular imaging by mass spectrometry — looking beyond classical histology. Nature Reviews Cancer 2010, 10 (9), 639- 646.

14. Cornett, D. S.; Reyzer, M. L.; Chaurand, P.; Caprioli, R. M. MALDI imaging mass spectrometry: molecular snapshots of biochemical systems. Nature Methods 2007, 4 (10), 828-833.

15. Seeley, E. H.; Caprioli, R. M. Molecular imaging of proteins in tissues by mass spectrometry. Proceedings of the National Academy of Sciences 2008.

16. Deutskens, F.; Yang, J.; Caprioli, R. M. High Spatial Resolution Imaging Mass Spectrometry and Classical Histology on a Single Tissue Section. Journal of Mass Spectrometry 2011, 46 (6), 568-571.

17. Thomas, A.; Chaurand, P. Advances in tissue section preparation for MALDI imaging MS. Bioanalysis 2014, 6 (7), 967-82.

18. Rubakhin, S.; Sweedler, J. A Mass Spectrometry Primer for Mass Spectrometry Imaging. In Mass Spectrometry Imaging, Rubakhin, S. S.; Sweedler, J. V., Eds. Humana Press: 2010; Vol. 656, pp 21-49.

187

19. Angel, P. M.; Caprioli, R. M. Matrix-Assisted Laser Desorption Ionization Imaging Mass Spectrometry: In Situ Molecular Mapping. Biochemistry 2013, 52 (22), 3818-3828.

20. Wisztorski, M.; Franck, J.; Salzet, M.; Fournier, I. MALDI Direct Analysis and Imaging of Frozen Versus FFPE Tissues: What Strategy for Which Sample? In Mass Spectrometry Imaging, Rubakhin, S. S.; Sweedler, J. V., Eds. Humana Press: 2010; Vol. 656, pp 303-322.

21. Balluff, B.; Schöne, C.; Höfler, H.; Walch, A. MALDI imaging mass spectrometry for direct tissue analysis: technological advancements and recent applications. Histochemistry and Cell Biology 2011, 136 (3), 227-244.

22. Chughtai, K.; Heeren, R. M. A. Mass Spectrometric Imaging for Biomedical Tissue Analysis. Chemical Reviews 2010, 110 (5), 3237-3277.

23. Schwartz, S. A.; Reyzer, M. L.; Caprioli, R. M. Direct tissue analysis using matrix-assisted laser desorption/ionization mass spectrometry: practical aspects of sample preparation. Journal of Mass Spectrometry 2003, 38 (7), 699-708.

24. Frese, C. K.; Zhou, H.; Taus, T.; Altelaar, A. F. M.; Mechtler, K.; Heck, A. J. R.; Mohammed, S. Unambiguous Phosphosite Localization using Electron- Transfer/Higher-Energy Collision Dissociation (EThcD). Journal of Proteome Research 2013, 12 (3), 1520-1525.

25. Walch, A.; Rauser, S.; Deininger, S.-O.; Höfler, H. MALDI imaging mass spectrometry for direct tissue analysis: a new frontier for molecular histology. Histochemistry and Cell Biology 2008, 130 (3), 421-434.

26. Spengler, B. MALDI-Mass Spectrometry Imaging. In MALDI MS, Wiley-VCH Verlag GmbH & Co. KGaA: 2013; pp 133-167.

27. Jones, E. A.; Deininger, S. O.; Hogendoorn, P. C.; Deelder, A. M.; McDonnell, L. A. Imaging mass spectrometry statistical analysis. Journal of Proteomics 2012, 75 (16), 4962-89.

28. Deininger, S.-O.; Cornett, D.; Paape, R.; Becker, M.; Pineau, C.; Rauser, S.; Walch, A.; Wolski, E. Normalization in MALDI-TOF imaging datasets of

188

proteins: practical considerations. Analytical and Bioanalytical Chemistry 2011, 401 (1), 167-181.

29. Deininger, S. r.-O.; Ebert, M. P.; Fütterer, A.; Gerhard, M.; Röcken, C. MALDI Imaging Combined with Hierarchical Clustering as a New Tool for the Interpretation of Complex Human Cancers. Journal of Proteome Research 2008, 7 (12), 5230-5236.

30. Deininger, S.-O.; Becker, M.; Suckau, D. Tutorial: Multivariate Statistical Treatment of Imaging Data for Clinical Biomarker Discovery. Mass Spectrometry Imaging. Rubakhin, S. S.; Sweedler, J. V., Eds. Humana Press: 2010; Vol. 656, pp 385-403. 31. Andersson, M.; Groseclose, M. R.; Deutch, A. Y.; Caprioli, R. M. Imaging mass spectrometry of proteins and peptides: 3D volume reconstruction. Nature Methods 2008, 5 (1), 101-108.

32. Xu, Y. Multimodal Spectral Microscopy and Imaging Mass Spectrometry of Biomolecules in Cells and Tissues. University of Toledo, 2012.

189