We are IntechOpen, the world's leading publisher of Open Access books Built by scientists, for scientists

4,000+

Open access books available

116,000+

International authors and editors

120M+

Downloads

Our authors are among the

Top 1%

most cited scientists

12.2%

Contributors from top 500 universities

Selection of our books indexed in the Book Citation Index in Web of Science™ Core Collection (BKCI)

### Interested in publishing with us? Contact book.department@intechopen.com

Numbers displayed above are based on latest data collected. For more information visit www.intechopen.com

## **Meet the editors**

Dr. Tsz-Kwong Man received his PhD degree from the University of Texas Health Science Center at Houston, Texas, and subsequently completed his postdoctoral training at the Washington University at St. Louis, Missouri. He has more than 10 years of research experience in genomic, proteomic and bioinformatic characterizations of pediatric cancers and has published more than

30 original research articles and two book chapters. Dr Man is currently an associate professor of pediatrics at the Baylor College of Medicine in the fields of cancer proteomics and biomarker discovery. He is a member of several national organizations, including American Association for Cancer Research, and a steering committee member of the TARGET consortium for pediatric cancers in the National Cancer Institute.

Dr. Ricardo J. Flores is currently a Pediatrics Hematology/Oncology instructor at the Baylor College of Medicine. He received his Medical Doctorate degree from the School of Medicine, University of Puerto Rico. He subsequently completed his Pediatrics residency at the University Pediatric Hospital of Puerto Rico Medical Center, followed by a clinical postdoctoral fellowship in

Pediatric Hematology/Oncology at the Texas Children's Cancer and Hematology Centers and Baylor College of Medicine. His research interest is applying genomic, proteomic, and bioinformatic approaches to characterize pediatric sarcomas. Dr Flores has coauthored two original research articles and three book chapters. He is a member of several national research and clinical organizations, including American Association for Cancer Research and American Society of Clinical Oncology.

Contents

**Preface IX** 

**Part 1 Proteomic Discovery of Disease Biomarkers 1** 

**Discovery of Vascular Biomarkers of Atherosclerosis 3** 

Chapter 1 **Overview of Current Proteomic Approaches for** 

Zinellu Elisabetta and Formato Marilena

**Diagnosis of Lung Surgery-Induced Injury 33** 

Chapter 3 **Urinary Exosomes for Protein Biomarker Research 49**  Delfin Albert Amal Raj, Immacolata Fiume, Giovambattista Capasso and Gabriella Pocsfalvi

**for Discovery of Biomarkers of Heart Disease 65** 

Chapter 4 **Circadian Proteomics and Its Unique Advantage** 

Gordon M. Kirby and Tami A. Martino

**for the Diagnosis and Management of Traumatic Brain Injury Patients 89** 

**Part 2 Proteomic Analysis of Protein Functions 107** 

**Function of a Novel Gene Called BRE 109** 

John Yeuk-Hon Chan, Yiu Loon Chui, Elve Chen, Yao Yao, Olivia Miu Yung Ngan and Henry Siu Sum Lee

**An Approach to Elucidating the** 

Kenneth Ka Ho Lee, Mei KuenTang,

Lepedda Antonio Junior,

**Clinical Evaluation for Early** 

Mei-Ling Tsai, Shu-Hui Chen, Chih-Ching Chang and Ming-Ho Wu

Chapter 2 **From Biomarker Discovery to** 

Peter S. Podobed,

Linda Papa

Chapter 6 **Comparative Proteomics:** 

Chapter 5 **Exploring the Role of Biomarkers** 

### Contents

#### **Preface** XIII

	- **Part 2 Proteomic Analysis of Protein Functions 107**

X Contents


Contents VII

**Part 4 Organelles and Secretome Proteomics 347** 

**Quantitative Mass Spectrometry Based Proteomics 349**  Florian Fröhlich, Tobias C. Walther and Romain Christiano

**Proteins in an** *In Vitro* **Blood-Brain Barrier Model 391**  Sophie Duban-Deweer, Johan Hachani, Barbara Deracinois, Roméo Cecchelli, Christophe Flahaut and Yannis Karamanos

**Secreted Factors: Focus on Muscle Secretome 417**  Jeanette Henningsen, Blagoy Blagoev and Irina Kratchmarova

Chapter 17 **Mitochondrial Proteomics: From Structure to Function 369**  Bernardo A. Petriz, Jeeser A. Almeida, Mirna S. Freire, Luiz A. O. Rocha, Taia M. B. Rezende and Octavio L. Franco

Chapter 16 **Analysis of Organelle Dynamics by** 

Chapter 18 **Proteomic Analysis of Plasma Membrane**

Chapter 19 **Quantitative Proteomics for Investigation of** 

#### **Part 4 Organelles and Secretome Proteomics 347**

VI Contents

Chapter 7 **Proteomic Approaches to** 

Federica Dabbeni-Sala,

Noriko Yokoyama

Chapter 11 **Posttranslational Modifications of** 

Yi-Jun Qi and Jen-Fu Chiu

Chapter 13 **Multidimensional Proteomics for the** 

Chapter 14 **The Microtubule-Dissociating** 

Chapter 15 **Identification of Factors Involved in** 

Chapter 10 **Identification of the Novel** 

Chapter 12 **Proteomic Study of** 

Chapter 8 **F0F1 ATP Synthase:** 

**Unraveling the RB/E2F Regulatory Pathway 135** Jone Mitxelena, Nerea Osinalde, Jesus M. Arizmendi,

**A Fascinating Challenge for Proteomics 161** 

**Dishevelled-Based Supermolecular Complexes 189**

Lindsey A. Miles, Nicholas M. Andronicos, Emily I. Chen, Nagyung Baik, Hongdong Bai, Caitlin M. Parmer, Shahrzad Lighvani, Samir Nangia, William B. Kiosses, Mark P. Kamps, John R. Yates III and Robert J. Parmer

**Myosin Light Chains Determine the Protein Fate 239**

**Part 3 Proteomic Approaches to Dissecting Disease Processes 255**

**Signals of Importance in Vascular Remodeling 275** Isabelle Sirois, Alexey V. Pshezhetsky and Marie-Josée Hébert

**Neurogenesis Recovery After Irradiation of the** 

François Chevalier, Alexandra Chicheportiche, Mathieu Daynac, Jordane Depagne, Pascale Bertrand, François D. Boussin and Marc-André Mouthon

**Adult Mouse Subventricular Zone: A Preliminary Study 327**

**Esophageal Squamous Cell Carcinoma 257** 

**Identification of Endothelial Post Mortem** 

**Tau in Neurological Disorders 291**  Francisco José Fernández-Gómez, Susanna Schraen-Maschke and Luc Buée

Asier Fullaondo and Ana M. Zubiaga

Amit Kumar Rai and Giovanna Lippe

**Plasminogen Receptor, Plg-RKT 219** 

Virgilio J. J. Cadete and Grzegorz Sawicki

Chapter 9 **Proteomic Analysis of Wnt-Dependent** 


Preface

cancer patients.

Advances in the biological and computational fields during the past two decades have unsealed new realms of possibilities embodied in the Omics fields. Genomics, the study of the genome, has made great strides toward unraveling and characterizing important gene functions and regulations of numerous organisms. However, because of the lack of correlation between mRNA and proteins, and the importance of posttranscriptional regulations and protein modifications in protein functions and human diseases, proteomics has become increasingly important in the research field. Proteomics is the study of the proteome in the cell, which represents the complete set of proteins encoded by the genome. Since the introduction of gel electrophoresis for protein separation in the 1960's, the methods for protein collection, identification, and quantification have continued to rapidly evolve and be refined. Protein research has expanded from the biochemical characterization of individual proteins to the highthroughput proteomics analysis of a cell, complex cell populations, and even an entire organism. This remarkable development highlights the potential of using proteomics methods to study protein functions and human diseases. New generations of mass spectrometry with higher resolutions and better quantification capabilities have also fueled the use of proteomics in biomarker and functional research. Proteomics approaches have been commonly used in the recent literature to identify biomarkers for disease screening, diagnosis, classification and monitoring. A potential application of proteomics in the field of oncology is in the discovery and validation of prognostic and predictive biomarkers, which play a fundamental role in personalized therapy for

The goal of this book is to provide a succinct overview of proteomics advances, including descriptions of the challenges that have been conquered and those yet to be resolved. The intended readers of this book include scientists and students involved in protein research from a basic, translational, or clinical perspective. The book consists of 19 chapters written by leaders in their fields and is organized into four major sections. The first section is comprised of five chapters on proteomics research for disease biomarkers, which include discovery of atherosclerosis biomarkers, proteomics analysis of bronchial fluids after lung injury, proteomics approaches for urinary exosome characterization, traumatic brain injury diagnosis and management, and circadian proteomics. The second section focuses on the use of proteomics to unravel and characterize important human protein functions. The section consists of

### Preface

Advances in the biological and computational fields during the past two decades have unsealed new realms of possibilities embodied in the Omics fields. Genomics, the study of the genome, has made great strides toward unraveling and characterizing important gene functions and regulations of numerous organisms. However, because of the lack of correlation between mRNA and proteins, and the importance of posttranscriptional regulations and protein modifications in protein functions and human diseases, proteomics has become increasingly important in the research field. Proteomics is the study of the proteome in the cell, which represents the complete set of proteins encoded by the genome. Since the introduction of gel electrophoresis for protein separation in the 1960's, the methods for protein collection, identification, and quantification have continued to rapidly evolve and be refined. Protein research has expanded from the biochemical characterization of individual proteins to the highthroughput proteomics analysis of a cell, complex cell populations, and even an entire organism. This remarkable development highlights the potential of using proteomics methods to study protein functions and human diseases. New generations of mass spectrometry with higher resolutions and better quantification capabilities have also fueled the use of proteomics in biomarker and functional research. Proteomics approaches have been commonly used in the recent literature to identify biomarkers for disease screening, diagnosis, classification and monitoring. A potential application of proteomics in the field of oncology is in the discovery and validation of prognostic and predictive biomarkers, which play a fundamental role in personalized therapy for cancer patients.

The goal of this book is to provide a succinct overview of proteomics advances, including descriptions of the challenges that have been conquered and those yet to be resolved. The intended readers of this book include scientists and students involved in protein research from a basic, translational, or clinical perspective. The book consists of 19 chapters written by leaders in their fields and is organized into four major sections. The first section is comprised of five chapters on proteomics research for disease biomarkers, which include discovery of atherosclerosis biomarkers, proteomics analysis of bronchial fluids after lung injury, proteomics approaches for urinary exosome characterization, traumatic brain injury diagnosis and management, and circadian proteomics. The second section focuses on the use of proteomics to unravel and characterize important human protein functions. The section consists of six chapters, which include elucidation of the function of a novel gene (BRE), characterization of the RB/E2F transcriptional regulatory pathway, analyses of essential proteins such as ATP synthetase and Wnt-dependent Disheveled-based supermolecular complexes, identification of a novel plasminogen receptor (Plg-RKT), and the study of posttranslational modifications of myosin light chains. The third section contains four chapters that recapitulate proteomics efforts in dissecting human disease processes. The first two chapters describe proteomics studies on esophageal squamous cell carcinoma and identification of endothelial signals for vascular remodeling. The last two chapters in this section focus on the role of microtubuledissociating Tau in neurological disorders, and identification of factors involved in neurogenesis recovery following adult mouse brain irradiation. The final section reports recent proteomics research on specific subproteomes and cellular organelles. The section consists of four chapters, which detail the analyses on organelle dynamics, mitochondrial proteome, plasma membrane of the blood brain barrier, and muscle secretome.

Finally, we would like to thank everyone who has made this book a reality and helped us to serve as subject editors of the InTech Proteomics Book series, and as editors of this volume. The whole experience has been very rewarding and exciting for us. This book would not have been created without the constant assistance and support of InTech staff members, especially our Publishing Process Manager, Ms. Martina Durovic. We truly appreciate the kind assistance from all of you!

> **Tsz-Kwong Man, PhD** and **Ricardo J. Flores, MD**  Baylor College of Medicine, Texas Children's Hospital, USA

## **Part 1**

### **Proteomic Discovery of Disease Biomarkers**

**1** 

*Italy* 

**Overview of Current Proteomic** 

**Biomarkers of Atherosclerosis** 

**Approaches for Discovery of Vascular** 

Lepedda Antonio Junior, Zinellu Elisabetta and Formato Marilena *University of Sassari/Dipartimento di Scienze Fisiologiche, Biochimiche e Cellulari* 

Cardiovascular diseases are the leading cause of mortality and morbidity in developed countries being atherosclerosis the major contributor. Atherosclerosis is a form of chronic inflammation characterized by the accumulation of lipids and fibrous elements in medium and large arteries (Libby, 2002). The retention of apoB-100 containing lipoproteins (mainly LDL and Lp(a)) in the subendothelial space and their subsequent oxidation is thought to be the leading event in the development of atherosclerotic lesions (Williams & Tabas, 1995). The degree of inflammation, proteolysis, calcification and neovascularization affects the stability of advanced lesions. Plaque rupture and thrombosis are the most important clinical complications in the pathogenesis of stroke, coronary arteries and peripheral vascular diseases (Lutgens et al., 2003). So, the identification of early biomarkers of plaque presence and susceptibility to ulceration could be of primary importance in preventing such a lifethreatening event. Disease aetiology is very complex and includes several important environmental and genetic risk factors such as hyperlipidemia, diabetes, and hypertension. In this regard elevated plasma levels of LDL cholesterol and low levels of HDL cholesterol have been long associated with the onset and development of atherosclerotic lesions. Although enormous efforts have been done to elucidate the molecular mechanisms underlying plaque formation and progression, they are not yet completely understood. In the last years, proteomic studies have been undertaken to both elucidate pathways of atherosclerotic degeneration and individuate new circulating markers to be utilized either as

This chapter will provide an overview of latest advances in proteomic studies on atherosclerosis and some related diseases, with particular emphasis on vascular tissue

Atherosclerosis is a very complex pathology in terms of cell types involved, inflammatory mechanisms and multifactorial aetiology. Many efforts have been done to shed light on the mechanisms underlying atherogenesis and to identify new circulating biomarkers which, along with traditional risk factors, will help in early diagnosis and prevention as well as in

**2. Application of proteomic technologies to the study of atherosclerosis** 

early diagnostic traits or as targets for new drug therapies.

proteomics and lipoproteomics.

**1. Introduction** 

### **Overview of Current Proteomic Approaches for Discovery of Vascular Biomarkers of Atherosclerosis**

Lepedda Antonio Junior, Zinellu Elisabetta and Formato Marilena *University of Sassari/Dipartimento di Scienze Fisiologiche, Biochimiche e Cellulari Italy* 

#### **1. Introduction**

Cardiovascular diseases are the leading cause of mortality and morbidity in developed countries being atherosclerosis the major contributor. Atherosclerosis is a form of chronic inflammation characterized by the accumulation of lipids and fibrous elements in medium and large arteries (Libby, 2002). The retention of apoB-100 containing lipoproteins (mainly LDL and Lp(a)) in the subendothelial space and their subsequent oxidation is thought to be the leading event in the development of atherosclerotic lesions (Williams & Tabas, 1995). The degree of inflammation, proteolysis, calcification and neovascularization affects the stability of advanced lesions. Plaque rupture and thrombosis are the most important clinical complications in the pathogenesis of stroke, coronary arteries and peripheral vascular diseases (Lutgens et al., 2003). So, the identification of early biomarkers of plaque presence and susceptibility to ulceration could be of primary importance in preventing such a lifethreatening event. Disease aetiology is very complex and includes several important environmental and genetic risk factors such as hyperlipidemia, diabetes, and hypertension. In this regard elevated plasma levels of LDL cholesterol and low levels of HDL cholesterol have been long associated with the onset and development of atherosclerotic lesions. Although enormous efforts have been done to elucidate the molecular mechanisms underlying plaque formation and progression, they are not yet completely understood. In the last years, proteomic studies have been undertaken to both elucidate pathways of atherosclerotic degeneration and individuate new circulating markers to be utilized either as early diagnostic traits or as targets for new drug therapies.

This chapter will provide an overview of latest advances in proteomic studies on atherosclerosis and some related diseases, with particular emphasis on vascular tissue proteomics and lipoproteomics.

#### **2. Application of proteomic technologies to the study of atherosclerosis**

Atherosclerosis is a very complex pathology in terms of cell types involved, inflammatory mechanisms and multifactorial aetiology. Many efforts have been done to shed light on the mechanisms underlying atherogenesis and to identify new circulating biomarkers which, along with traditional risk factors, will help in early diagnosis and prevention as well as in

Overview of Current Proteomic Approaches

stages of lesion formation.

**2.1.1 Studies on animal models** 

et al. analysed aortic lesions from apolipoprotein E-/-

for Discovery of Vascular Biomarkers of Atherosclerosis 5

diagnosis and patients treatment. Compared to tissue specimens of human origin, animal models, mainly rodents, have been utilized to study the mechanisms underlying the early

Apolipoprotein E-deficient mouse is the most popular murine model in cardiovascular research and has revealed important insights into mechanisms affecting atherogenesis. Mayr

medium, and severe according to lesion-covered areas on the aortic surface (Mayr et al., 2005). As expected, authors found an increase of inflammatory cells, a decrease of VSMCs, and an accumulation of serum proteins associated to an impaired endothelial barrier function with lesion progression. Interestingly, immunoglobulins, that were barely detectable in apolipoprotein E+/+ mice, accumulated even in aortas of young apolipoprotein E-/- mice. The authors identified 79 differentially expressed spots. Moreover, they suggested an increase in oxidative stress with lesion progression evaluating the ratio between the oxidized and the reduced forms of peroxiredoxin, the former resulting in a charge shift toward a more acidic isoelectric point. Overall, they found a linear relationship between the degree of peroxiredoxin-Cys oxidation and the extent of lesion formation in aortas of apolipoprotein E-deficient mice. Almofti et al. applied 2DE coupled to matrix-assisted laser desorption/ionization time of flight (MALDI TOF) MS analysis to a rat model of atherosclerosis. They induced atherosclerosis by a single dose of vitamin D3 associated with a high fat diet and identified 46 proteins differently expressed in diseased tissues. Among them, 18 proteins, including a group of oxidization-related enzymes, were found to be upregulated, while 28 proteins were found down-regulated (Almofti et al., 2006). Vascular endothelium plays important physiological roles in vascular homeostasis, coagulation, inflammation, as well as tissue growth and repair. Impairment of the endothelial function is an early event in atherosclerotic lesion formation leading to overexpression of adhesion molecules as well as secretion of pro-inflammatory and chemotactic cytokines. An affinitybased proteomic approach was used by Wu et al. (Wu et al., 2007) to identify vascular endothelial surface proteins differentially expressed in aortic tissues of apolipoprotein E deficient mice. After *in situ* perfusion of vascular bed with a solution containing a biotinderivative, biotinylated endothelial proteins were extracted, purified by affinity enrichment with streptavidin-agarose beads, and resolved by SDS-PAGE. The whole gel lanes were cut into slices that were subjected to tryptic digestion for nano liquid chromatography (LC) MS/MS analysis. In this way, 454 proteins, mainly extracellular or associated to cell membrane, were identified. Among them, there were cell adhesion molecules, accounting for the largest category, followed by proteins involved in signal transduction and transport. Interestingly, proteins associated with immune and inflammatory responses were more than doubled in atherosclerotic aorta (13%) in comparison to normal aorta (6%). On the other hand, proteins involved in lipid metabolism were decreased by 34% in atherosclerotic aorta. A rat model has been recently used for a proteomic study on the effects of blood shear stress on atherogenesis (Qi et al., 2008). It is well known that blood shear stress affects endothelial cell shape and orientation, as well as vascular wall permeability. Indeed, regions of arterial branching or curvature, where blood flow is not uniform, are preferential sites for lesion formation. By comparing homogenates of aortas kept under two levels of shear stress in a perfusion culture system for 24 hours, Qi et al. detected a reduced expression of protein Rho-GDP dissociation inhibitor alpha (Rho-GDIα) in low shear stress conditions and

and wild type mice classified as light,

monitoring the effects of pharmacological agents. To address these issues, proteomic studies have been focused on different matrices such as vascular cell/tissues, looking at both proteomes and secretomes, plasma/serum, urine, and purified plasma lipoprotein fractions (fig. 1).

Fig. 1. Overview of the main targets of proteomic studies searching for both mechanisms of atherogenesis and biomarkers of atherosclerotic lesion presence and progression. Dotted lines represent almost unexplored paths. LCMs, laser-captured microdissections

To date, several proteomic approaches, such as 1D-2D electrophoresis (1DE-2DE) followed by mass spectrometry (MS) analyses, western arrays, protein arrays, and gel-free MS based proteomics, have been applied in the search of vascular biomarkers of atherosclerosis. Often, classical biochemical methods, mainly western blotting (WB), ELISA, and immunohistochemistry (IH) have been used to validate the proteomic results.

#### **2.1 Vascular tissue proteomics**

Even though tissue analyses frequently provide useful data, there are major drawbacks in analysing human atherosclerotic specimens. Atherosclerotic plaques are quite complex in terms of vascular cells and extracellular components. In this respect, besides vascular smooth muscle cells (VSMCs) and endothelial cells (ECs), they are composed of inflammatory cells, filtered plasma proteins, new-formed extracellular matrix, cellular debris and end-products of lipid and protein oxidation. Another critical point in the *in situ* analysis of protein expression within atherosclerotic plaques is the choice of the appropriate control. It would be desirable to utilize control specimens from the same vascular district of the same patient, in order to minimize intrinsic tissue differences, and from surgical endarterectomy rather than from post-mortem material, to avoid the occurrence of proteolytic modifications prior to analysis. Also the availability of a significant number of human specimens could be limiting. Because of the complexity of advanced lesions (Stary, 2000; Virmani et al., 2000) in terms of necrotic core dimension, fibrous cap thickness, inflammatory and proteolytic components, careful histochemical classification is needed. Moreover, results from different advanced lesion typologies are difficult to interpret because they could be either associated to the lesion development or merely a consequence of the advanced condition. In the latest years proteomic technologies have been applied to human diseased tissues to both characterize mechanisms of advanced atherosclerotic plaque development, mainly those responsible for its instability, and to identify markers useful in diagnosis and patients treatment. Compared to tissue specimens of human origin, animal models, mainly rodents, have been utilized to study the mechanisms underlying the early stages of lesion formation.

#### **2.1.1 Studies on animal models**

4 Proteomics – Human Diseases and Protein Functions

monitoring the effects of pharmacological agents. To address these issues, proteomic studies have been focused on different matrices such as vascular cell/tissues, looking at both proteomes and secretomes, plasma/serum, urine, and purified plasma lipoprotein fractions

Fig. 1. Overview of the main targets of proteomic studies searching for both mechanisms of atherogenesis and biomarkers of atherosclerotic lesion presence and progression. Dotted

To date, several proteomic approaches, such as 1D-2D electrophoresis (1DE-2DE) followed by mass spectrometry (MS) analyses, western arrays, protein arrays, and gel-free MS based proteomics, have been applied in the search of vascular biomarkers of atherosclerosis. Often, classical biochemical methods, mainly western blotting (WB), ELISA, and

Even though tissue analyses frequently provide useful data, there are major drawbacks in analysing human atherosclerotic specimens. Atherosclerotic plaques are quite complex in terms of vascular cells and extracellular components. In this respect, besides vascular smooth muscle cells (VSMCs) and endothelial cells (ECs), they are composed of inflammatory cells, filtered plasma proteins, new-formed extracellular matrix, cellular debris and end-products of lipid and protein oxidation. Another critical point in the *in situ* analysis of protein expression within atherosclerotic plaques is the choice of the appropriate control. It would be desirable to utilize control specimens from the same vascular district of the same patient, in order to minimize intrinsic tissue differences, and from surgical endarterectomy rather than from post-mortem material, to avoid the occurrence of proteolytic modifications prior to analysis. Also the availability of a significant number of human specimens could be limiting. Because of the complexity of advanced lesions (Stary, 2000; Virmani et al., 2000) in terms of necrotic core dimension, fibrous cap thickness, inflammatory and proteolytic components, careful histochemical classification is needed. Moreover, results from different advanced lesion typologies are difficult to interpret because they could be either associated to the lesion development or merely a consequence of the advanced condition. In the latest years proteomic technologies have been applied to human diseased tissues to both characterize mechanisms of advanced atherosclerotic plaque development, mainly those responsible for its instability, and to identify markers useful in

lines represent almost unexplored paths. LCMs, laser-captured microdissections

immunohistochemistry (IH) have been used to validate the proteomic results.

**2.1 Vascular tissue proteomics** 

(fig. 1).

Apolipoprotein E-deficient mouse is the most popular murine model in cardiovascular research and has revealed important insights into mechanisms affecting atherogenesis. Mayr et al. analysed aortic lesions from apolipoprotein E-/- and wild type mice classified as light, medium, and severe according to lesion-covered areas on the aortic surface (Mayr et al., 2005). As expected, authors found an increase of inflammatory cells, a decrease of VSMCs, and an accumulation of serum proteins associated to an impaired endothelial barrier function with lesion progression. Interestingly, immunoglobulins, that were barely detectable in apolipoprotein E+/+ mice, accumulated even in aortas of young apolipoprotein E-/- mice. The authors identified 79 differentially expressed spots. Moreover, they suggested an increase in oxidative stress with lesion progression evaluating the ratio between the oxidized and the reduced forms of peroxiredoxin, the former resulting in a charge shift toward a more acidic isoelectric point. Overall, they found a linear relationship between the degree of peroxiredoxin-Cys oxidation and the extent of lesion formation in aortas of apolipoprotein E-deficient mice. Almofti et al. applied 2DE coupled to matrix-assisted laser desorption/ionization time of flight (MALDI TOF) MS analysis to a rat model of atherosclerosis. They induced atherosclerosis by a single dose of vitamin D3 associated with a high fat diet and identified 46 proteins differently expressed in diseased tissues. Among them, 18 proteins, including a group of oxidization-related enzymes, were found to be upregulated, while 28 proteins were found down-regulated (Almofti et al., 2006). Vascular endothelium plays important physiological roles in vascular homeostasis, coagulation, inflammation, as well as tissue growth and repair. Impairment of the endothelial function is an early event in atherosclerotic lesion formation leading to overexpression of adhesion molecules as well as secretion of pro-inflammatory and chemotactic cytokines. An affinitybased proteomic approach was used by Wu et al. (Wu et al., 2007) to identify vascular endothelial surface proteins differentially expressed in aortic tissues of apolipoprotein E deficient mice. After *in situ* perfusion of vascular bed with a solution containing a biotinderivative, biotinylated endothelial proteins were extracted, purified by affinity enrichment with streptavidin-agarose beads, and resolved by SDS-PAGE. The whole gel lanes were cut into slices that were subjected to tryptic digestion for nano liquid chromatography (LC) MS/MS analysis. In this way, 454 proteins, mainly extracellular or associated to cell membrane, were identified. Among them, there were cell adhesion molecules, accounting for the largest category, followed by proteins involved in signal transduction and transport. Interestingly, proteins associated with immune and inflammatory responses were more than doubled in atherosclerotic aorta (13%) in comparison to normal aorta (6%). On the other hand, proteins involved in lipid metabolism were decreased by 34% in atherosclerotic aorta. A rat model has been recently used for a proteomic study on the effects of blood shear stress on atherogenesis (Qi et al., 2008). It is well known that blood shear stress affects endothelial cell shape and orientation, as well as vascular wall permeability. Indeed, regions of arterial branching or curvature, where blood flow is not uniform, are preferential sites for lesion formation. By comparing homogenates of aortas kept under two levels of shear stress in a perfusion culture system for 24 hours, Qi et al. detected a reduced expression of protein Rho-GDP dissociation inhibitor alpha (Rho-GDIα) in low shear stress conditions and

Overview of Current Proteomic Approaches

10 coronary arteries from CAD patients vs 7 normal autoptic coronary

(2DE of homogenates, LC-MS/MS,

6 carotid plaques containing a thrombus vs 5 advanced stable lesions

(2DE of homogenates, MALDI-TOF/TOF MS, LC-MS/MS, WB, IH)

7 atherosclerotic aortic specimens vs biopsies of the normal tissue from the

(2DE of homogenates, MALDI-TOF

35 atherosclerotic endarterectomies (10 femoral, 25 carotids) vs 36 control endarteries (24 mammary, 12 radial)

(2DE of secretomes, MALDI-TOF MS, LC-MS/MS and IMAC combined with MALDI Q-TOF MS/MS, WB, ELISA,

(2DE of secretomes, MALDI-TOF MS,

21 stenosing complicated carotid regions with/without atorvastatin treatement vs fibrous regions (ex vivo)

29 unstable carotid plaques vs 19

(2DE of extracts, MALDI-TOF MS,

10 complicated segments in the internal carotid artery (ICA) vs 10 stable segments in the common carotid

(2-D DIGE of homogenates, LC-

artery (CCA)

MS/MS, IH)

arteries

WB, rt-PCR)

same patients

MS, WB)

IH)

WB)

LC MS/MS, WB)

stable carotid plaques

for Discovery of Vascular Biomarkers of Atherosclerosis 7

**Human tissues (Methods) Results Known functions Ref.** 

↑ ferritin light chain ↓ ferritin light chain mRNA

> ↑ 39 proteins (27 identified)

↓↓ HSP27 secretion ↓↓ HSP27 plasma levels

↑ 24 proteins ↓ 20 proteins Treatment reverts the differential protein secretion

↑ ferritin light subunit ↑ superoxide dismutase 2 ↑ fibrinogen fragment D ↓ superoxide dismutase 3 ↓ glutathione Stransferase ↓ Rho GDP-dissociation inhibitor 1 ↓ annexin A10 ↓ HSP 20 ↓ HSP 27

↑ 6 proteins ↓ 11 proteins 2 proteins with isoform dependent distributions

modulation of oxidation

↑ α1-antitrypsin acute-phase protein Donners et al.,

signal transduction angiogenesis MMP activation regulation of proinflammatory cytokines

anti-inflammatory down-regulation of the apoptotic signaling pathway

modulation of oxidation and inflammation structural signaling pathway cholesterol metabolism

> modulation of inflammation and oxidative stress

signal transduction transport cell growth metabolism

You et al., 2003

2005

Sung et al., 2006

Duran et al., 2003; Martin-Ventura et al., 2004

Durán et al., 2007

Lepedda et al., 2009

Olson et al., 2010

demonstrated, by siRNA technology, that this reduction enhances VSMC migration and apoptosis.

#### **2.1.2 Studies on human tissues**

As from 2003, 14 researches on human atherosclerotic plaque proteomics have been published; the diseased tissues used as matrices were coronary arteries (2/14), carotid arteries (11/14), and aortas (1/14).

Most of them were conducted by using two-dimensional electrophoresis coupled to mass spectrometry as analytical method (10/14). The sample source, the methodology applied, and the most relevant findings of these studies are summarized in table 1.

In 2003, You et al., by analysing 10 diseased (coronary artery disease, CAD) and 7 normal autoptic coronary arteries, reported about 2 fold increase of the ferritin light chain in the pathological specimens (You et al., 2003). Quantitative analysis by real-time PCR showed a decrease in ferritin light chain mRNA expression in diseased tissues suggesting that the increased expression of ferritin light chain in CAD coronary arteries may be related to increased protein stability. This result highlights the importance of protein expression analysis in studying disease-associated gene expression. Donners et al. analysed 5 stable plaques and 6 lesions with a thrombus from patients undergoing carotid endarterectomy, classified according to Virmani et al. (Donners et al., 2005). By 2DE analysis, they identified vinexin-β and α1-antitrypsin as differentially expressed. However, neither immunohistochemistry nor western blotting confirmed vinexin-β differential expression underlining the importance of validating proteomic results by other biochemical methods. Conversely, western blotting of 2D gels revealed, in lesions with a thrombus, the expression of six isoforms of the acute phase protein α1-antitrypsin, one of which was uniquely expressed in thrombus-containing plaques. Sung et al. analysed non-diseased and atherosclerotic specimens from 7 patients undergoing aorta bypass surgery. They identified a panel of 27 proteins differentially expressed in the atherosclerotic aorta involved in a number of biological processes, including calcium-mediated processes, migration of VSMCs, matrix metalloproteinase activation and regulation of pro-inflammatory cytokines (Sung et al., 2006). A different approach was adopted by Martin-Ventura et al. who analysed the protein secretion profiles obtained from 35 cultured atherosclerotic plaques (10 femoral, 25 carotids) and 36 control arteries (24 mammary, 12 radial) in the search of new biological markers potentially released by the arterial wall into the plasma (Duran et al., 2003; Martin-Ventura et al., 2004). In particular, they isolated and analysed the secretomes from noncomplicated and ruptured/thrombosed areas of the same cultured carotid plaque so avoiding the variability of the control specimens. They showed that, compared to control arteries, heat shock protein 27 (HSP27) secretion into the culture medium was significantly lower in atherosclerotic plaques and barely detectable in complicated plaque supernatants, as confirmed by WB analysis. They also evidenced a 20-fold reduction in HSP27 levels in the plasma of patients with carotid stenosis respect to healthy controls so identifying HSP27 as a possible marker of atherosclerosis. The same research group evaluated the effects of incubation with atorvastatin, a 3-hydroxy-3-methylglutaryl CoenzymeA reductase inhibitor, on the secretomes of cultured atherosclerotic plaques (Durán et al., 2007). They identified 24 proteins that were increased and 20 proteins that were decreased in atherosclerotic plaque supernatants compared to controls. Interestingly, the presence of atorvastatin in culture medium reverted secretion of 66% proteins to control values. In this report, authors identified cathepsin D as a potential target for therapeutical treatment of atherosclerosis.

demonstrated, by siRNA technology, that this reduction enhances VSMC migration and

As from 2003, 14 researches on human atherosclerotic plaque proteomics have been published; the diseased tissues used as matrices were coronary arteries (2/14), carotid

Most of them were conducted by using two-dimensional electrophoresis coupled to mass spectrometry as analytical method (10/14). The sample source, the methodology applied,

In 2003, You et al., by analysing 10 diseased (coronary artery disease, CAD) and 7 normal autoptic coronary arteries, reported about 2 fold increase of the ferritin light chain in the pathological specimens (You et al., 2003). Quantitative analysis by real-time PCR showed a decrease in ferritin light chain mRNA expression in diseased tissues suggesting that the increased expression of ferritin light chain in CAD coronary arteries may be related to increased protein stability. This result highlights the importance of protein expression analysis in studying disease-associated gene expression. Donners et al. analysed 5 stable plaques and 6 lesions with a thrombus from patients undergoing carotid endarterectomy, classified according to Virmani et al. (Donners et al., 2005). By 2DE analysis, they identified vinexin-β and α1-antitrypsin as differentially expressed. However, neither immunohistochemistry nor western blotting confirmed vinexin-β differential expression underlining the importance of validating proteomic results by other biochemical methods. Conversely, western blotting of 2D gels revealed, in lesions with a thrombus, the expression of six isoforms of the acute phase protein α1-antitrypsin, one of which was uniquely expressed in thrombus-containing plaques. Sung et al. analysed non-diseased and atherosclerotic specimens from 7 patients undergoing aorta bypass surgery. They identified a panel of 27 proteins differentially expressed in the atherosclerotic aorta involved in a number of biological processes, including calcium-mediated processes, migration of VSMCs, matrix metalloproteinase activation and regulation of pro-inflammatory cytokines (Sung et al., 2006). A different approach was adopted by Martin-Ventura et al. who analysed the protein secretion profiles obtained from 35 cultured atherosclerotic plaques (10 femoral, 25 carotids) and 36 control arteries (24 mammary, 12 radial) in the search of new biological markers potentially released by the arterial wall into the plasma (Duran et al., 2003; Martin-Ventura et al., 2004). In particular, they isolated and analysed the secretomes from noncomplicated and ruptured/thrombosed areas of the same cultured carotid plaque so avoiding the variability of the control specimens. They showed that, compared to control arteries, heat shock protein 27 (HSP27) secretion into the culture medium was significantly lower in atherosclerotic plaques and barely detectable in complicated plaque supernatants, as confirmed by WB analysis. They also evidenced a 20-fold reduction in HSP27 levels in the plasma of patients with carotid stenosis respect to healthy controls so identifying HSP27 as a possible marker of atherosclerosis. The same research group evaluated the effects of incubation with atorvastatin, a 3-hydroxy-3-methylglutaryl CoenzymeA reductase inhibitor, on the secretomes of cultured atherosclerotic plaques (Durán et al., 2007). They identified 24 proteins that were increased and 20 proteins that were decreased in atherosclerotic plaque supernatants compared to controls. Interestingly, the presence of atorvastatin in culture medium reverted secretion of 66% proteins to control values. In this report, authors identified cathepsin D as a potential target for therapeutical treatment of atherosclerosis.

and the most relevant findings of these studies are summarized in table 1.

apoptosis.

**2.1.2 Studies on human tissues** 

arteries (11/14), and aortas (1/14).


Overview of Current Proteomic Approaches

**Human tissues (Methods)** 

atherosclerotic mammary arteries

(Western array (823 Abs), WB, rt-

4 pooled unstable carotid plaques vs 4 pooled stable carotid plaques

(protein microarray analysis of the expression of 512 proteins)

Histological sections from 35 coronary vessels in paraffin or

(direct tissue proteomics AQUA methodology)

that had a secondary

Carotid plaques from 80 patients

human advanced atherosclerotic plaques.

cardiovascular event vs 80 sex and age matched event-free patients (during a 3-year follow-

frozen blocks

up)

(LC MS/MS)

↑, increase. ↓, decrease.

12 carotid endarterectomy specimens vs 7 non-

PCR, IH)

for Discovery of Vascular Biomarkers of Atherosclerosis 9

↓↓ apoptosis-linked gene 2

↑↑ Thrombospondin-2, Mn superoxide dismutase, apolipoprotein B-100, protein-tyrosine phosphatase 1C, apolipoprotein E

↓↓ glycogen synthase kinase-3β

> ↑ 21 proteins ↓ 3 proteins

806 unique proteins identified with high confidence

Strong positive association between osteopontin and the occurrence of new vascular complications

High-throughput western blot analysis, also called western array, was used to screen cell lysates from 12 carotid endarterectomy specimens and 7 non-atherosclerotic mammary arteries, obtained during bypass surgery, with 823 monoclonal antibodies mainly directed against signal-transducing proteins (Martinet et al., 2003). Western arrays showed a highly reproducible pattern of protein expression but also a high rate of false-positive signals (differential protein expression of only 7 of the 15 proteins detected by using this method was confirmed by standard immunoblot assay). A strong down regulation of apoptosislinked gene 2 (ALG-2) was found, suggesting a novel mechanism inhibiting cell death in

Table 2. Alternative approaches in proteomics of the atherosclerotic plaque.

**Results Known** 

**functions** 

mediator of apoptosis

modulation of inflammatory, angiogenic, proliferative, and apoptotic pathways

**Ref.** 

Martinet et al., 2003

Slevin et al., 2006

Bagnato et al., 2007

de Kleijn et al., 2010


IMAC, immobilized metal affinity chromatography. ↑, increase. ↓, decrease.

Table 1. 2DE coupled to MS studies on the human atherosclerotic plaque.

Since carotid plaque rupture and thrombosis are the most important clinical complications in the pathogenesis of cerebro-vascular diseases, many efforts have been done to elucidate mechanisms underlying plaque vulnerability and to identify reliable specific markers of plaques prone to rupture. In a previous study we provided evidence for a wide fragmentation of some apolipoproteins and arterial proteoglycans and for a proinflammatory microenvironment in unstable and much less in stable endarterectomy carotid plaques (Formato et al., 2004). Recently, we evaluated differential protein expression in a considerable number (n=48) of plaques obtained from carotid endarterectomy classified by immunohistochemistry in stable and unstable (Lepedda et al., 2009). Our study was carried out on the premise that plaque stability/instability is associated with distinct patterns of protein expression. We analysed extracts from finely minced tissues in order to allow enrichment in both topically expressed and filtered/retained proteins. A total of 57 distinct spots corresponding to 33 different proteins were identified in both stable and unstable plaques by peptide mass fingerprinting (PMF) analysis, most of which were of plasma origin (about 70%). This suggested the existence of an impaired endothelial barrier function independent from plaque typology. Compared to stable plaques, unstable ones showed reduced abundance of protective enzymes superoxide dismutase 3 and glutathione Stransferase, small HSP 27 and 20, annexin A10, and Rho GDP-dissociation inhibitor and a higher abundance of ferritin light subunit, superoxide dismutase 2 and fibrinogen fragment D. These proteins are described to play a role in either oxidative or inflammatory processes and in the formation and progression of the atherosclerotic plaque. Our proteomic approach, trying to differentiate unstable from stable human carotid plaques, identified, in the former, a panel of proteins with pro-oxidant and pro-inflammatory potentials according to our current understanding of the molecular basis of the atherosclerotic process.

To overcome inter-individual variations in protein expression, Olson et al. applied 2-D differential in gel electrophoresis (2D DIGE) in combination with MS/MS to compare protein distribution in 10 complicated segments located in the internal carotid artery (ICA) with that in 10 more stable segments in the common carotid artery (CCA) from the same patient (Olson et al., 2010). In this way, they identified 19 proteins with differential distribution between ICA and CCA segments. To overcome the problem of plaque heterogeneity, Terzuoli et al. proposed a method for selecting proteins exclusive to plaque by constructing a reference synthetic gel (Terzuoli et al., 2007; Porcelli et al., 2010). This gel, obtained by averaging the positions, shapes and optical densities of spots in 2DE maps from 10 carotid plaque samples was compared with an equivalent synthetic gel constructed using 10 plasma samples from the same carotid surgery patients. The comparison allowed discriminating between plasma and plaque proteins, the latter being potential markers of plaque vulnerability.

Some alternative proteomic approaches have been applied to date in the search of new biomarkers of the atherosclerotic process (table 2).

**Human tissues (Methods) Results Known functions Ref.** 

Identification of proteins exclusive to plaque

Since carotid plaque rupture and thrombosis are the most important clinical complications in the pathogenesis of cerebro-vascular diseases, many efforts have been done to elucidate mechanisms underlying plaque vulnerability and to identify reliable specific markers of plaques prone to rupture. In a previous study we provided evidence for a wide fragmentation of some apolipoproteins and arterial proteoglycans and for a proinflammatory microenvironment in unstable and much less in stable endarterectomy carotid plaques (Formato et al., 2004). Recently, we evaluated differential protein expression in a considerable number (n=48) of plaques obtained from carotid endarterectomy classified by immunohistochemistry in stable and unstable (Lepedda et al., 2009). Our study was carried out on the premise that plaque stability/instability is associated with distinct patterns of protein expression. We analysed extracts from finely minced tissues in order to allow enrichment in both topically expressed and filtered/retained proteins. A total of 57 distinct spots corresponding to 33 different proteins were identified in both stable and unstable plaques by peptide mass fingerprinting (PMF) analysis, most of which were of plasma origin (about 70%). This suggested the existence of an impaired endothelial barrier function independent from plaque typology. Compared to stable plaques, unstable ones showed reduced abundance of protective enzymes superoxide dismutase 3 and glutathione Stransferase, small HSP 27 and 20, annexin A10, and Rho GDP-dissociation inhibitor and a higher abundance of ferritin light subunit, superoxide dismutase 2 and fibrinogen fragment D. These proteins are described to play a role in either oxidative or inflammatory processes and in the formation and progression of the atherosclerotic plaque. Our proteomic approach, trying to differentiate unstable from stable human carotid plaques, identified, in the former, a panel of proteins with pro-oxidant and pro-inflammatory potentials according

IMAC, immobilized metal affinity chromatography. ↑, increase. ↓, decrease.

Table 1. 2DE coupled to MS studies on the human atherosclerotic plaque.

to our current understanding of the molecular basis of the atherosclerotic process.

To overcome inter-individual variations in protein expression, Olson et al. applied 2-D differential in gel electrophoresis (2D DIGE) in combination with MS/MS to compare protein distribution in 10 complicated segments located in the internal carotid artery (ICA) with that in 10 more stable segments in the common carotid artery (CCA) from the same patient (Olson et al., 2010). In this way, they identified 19 proteins with differential distribution between ICA and CCA segments. To overcome the problem of plaque heterogeneity, Terzuoli et al. proposed a method for selecting proteins exclusive to plaque by constructing a reference synthetic gel (Terzuoli et al., 2007; Porcelli et al., 2010). This gel, obtained by averaging the positions, shapes and optical densities of spots in 2DE maps from 10 carotid plaque samples was compared with an equivalent synthetic gel constructed using 10 plasma samples from the same carotid surgery patients. The comparison allowed discriminating between plasma and plaque proteins, the latter being potential markers of

Some alternative proteomic approaches have been applied to date in the search of new

Terzuoli et al., 2007; Porcelli et al., 2010)

10 carotid plaques vs reference

plaque vulnerability.

biomarkers of the atherosclerotic process (table 2).

(2DE of homogenates, LC-MS/MS)

synthetic gel


↑, increase. ↓, decrease.

Table 2. Alternative approaches in proteomics of the atherosclerotic plaque.

High-throughput western blot analysis, also called western array, was used to screen cell lysates from 12 carotid endarterectomy specimens and 7 non-atherosclerotic mammary arteries, obtained during bypass surgery, with 823 monoclonal antibodies mainly directed against signal-transducing proteins (Martinet et al., 2003). Western arrays showed a highly reproducible pattern of protein expression but also a high rate of false-positive signals (differential protein expression of only 7 of the 15 proteins detected by using this method was confirmed by standard immunoblot assay). A strong down regulation of apoptosislinked gene 2 (ALG-2) was found, suggesting a novel mechanism inhibiting cell death in human advanced atherosclerotic plaques.

Overview of Current Proteomic Approaches

**2.3 Plasma/serum proteomics** 

for Discovery of Vascular Biomarkers of Atherosclerosis 11

of tissue analyses, allowing researchers to study single aspects of the atherosclerotic process in very controlled conditions. In the last years, studies on proteome (the intracellular proteins) and secretome (the proteins released into the cell culture medium) by 2DE and MS of ECs (Bruneel et al., 2003; Chen et al., 2007; Tunica et al., 2009), VSMCs (McGregor et al., 2001; Dupont et al., 2005; Lee et al., 2006), and monocytes/macrophages (Dupont et al., 2004; Fach et al., 2004; Slomianny et al., 2006; Zhang et al., 2007; Zhao et al., 2009) have been performed. Moreover, in the attempt to help in elucidating the mechanisms of atherogenesis, several proteomic studies have been carried out on vascular cells cultured in different experimental conditions (table 3). The most applied proteomic methodologies were 1DE-2DE coupled to MS analyses. Very few studies applied gel free proteomic approaches such as LC MS/MS (Fach et al., 2004; X.L. Wang et al., 2007; Zhao et al., 2009; Tunica et al.,

Plasma is one of the most useful matrices to investigate for identifying new biomarkers. It carries resident proteins that represent the majority of plasma proteins, together with proteins released from vascular cells and other tissues. In this respect, variations in plasma proteome could reflect directly or indirectly a cardiovascular disease or other pathological conditions. Moreover, monitoring plasma proteome could be successful in patients follow up and in relation to drug therapies. Besides plasma, also serum is widely investigated, although its proteome is known to be drastically affected by blood coagulation. Plasma proteomic studies are challenging due to both the high dynamic range of protein concentrations and the number of low expressed proteins. Plasma is composed for up to 99% by 21 most abundant proteins and for the remaining 1% by about 50,000 low expressed protein variants, representing the so called "deep proteome" (Righetti et al., 2005). Plasma protein levels range from 40-50 mg/ml for albumin to less than 10 ng/ml for interleukins, chatepsins and peptide hormones. Therefore, differential proteomics of unfractionated plasma provides only limited information. In this regard, several pre-analytical depletion systems that imply affinity/immunoaffinity steps for simultaneous removal of the most represented plasma proteins have been set up. However, due to non-specific binding, many other proteins could be depleted. Besides depletion systems, also protein enrichment technology has been developed. This relies on solid-phase combinatorial ligand libraries, made of hexapeptides, to reduce the high dynamic range of protein concentration preventing the co-depletion of low-abundance proteins (Boschetti et al., 2007). Since variations in plasma protein expression can provide useful information on both physiological and pathological conditions of the different tissues in the body, many efforts have been made in characterizing the entire human plasma proteome. In this regard, during 2003-2005, 55 laboratories worldwide participating to the Human Plasma Proteome Project (HPPP) analysed reference specimens by using emerging technologies in the field of proteomics, and generated integrated databases for proteins detectable and identifiable in human serum and plasma (http://www.hupo.org/research/hppp/). They confidently identified 3020 proteins with a minimum of two high-scoring MS/MS spectra that have been searched for relevance to cardiovascular function and disease using PubMed search engine and specific keywords. On the basis of the current knowledge, the study individuated a subset of 345 proteins showing cardiovascular-related functions (markers of inflammation and/or cardiovascular disease, proteins implicated in coagulation, signalling,

2009; Zimman et al., 2010) and microarrays (Sukhanov et al., 2005).

growth, differentiation, and vascular remodelling) (Berhane et al., 2005).

By using microarray technology, Slevin et al. compared the expression of 512 proteins associated with inflammatory, pro/anti-apoptotic, and angiogenesis pathways between 4 pooled fibrous stable carotid plaques and 4 pooled ulcerated, hemorrhagic unstable plaques (Slevin et al., 2006). In spite of the high sensitivity of protein microarrays, allowing detecting nanogram quantities, errors can occur because of weakly reacting or nonspecific antibodies, degraded proteins, and/or the efficiency of sample fluorescent labelling. However, western blotting analyses confirmed differential expression between stable and unstable plaque pools for all 11 proteins selected suggesting a high level of specificity of the array antibodies and the usefulness of this proteomic approach in the study of plaque pathogenesis.

Atherosclerotic tissue could also be laser-microdissected, which would allow one to compare different areas of the plaque such as necrotic core and shoulders/fibrous cap, providing valuable spatial information. Bagnato et al. applied the direct tissue proteomic (DTP) approach to paraffin or frozen blocks from 35 coronary atherosclerotic lesions classified by histopathological examination in early, intermediate, and advanced (Bagnato et al., 2007). In particular, different plaque regions were laser-microdissected (LCMs) from both paraffin and frozen sections and subjected to tryptic digestion followed by LC MS/MS to obtain area-specific proteomic information. Frozen sections were also homogenized and proteins resolved by SDS PAGE and analysed by LC MS/MS. Moreover, they used AQUA (absolute quantitation) methodology to quantify Stromal Cell-derived Factor 1 α (SDF1-α) and growth factors not detected by the above mentioned methods. These multiple approaches allowed them to identify 806 unique proteins with high confidence so obtaining a large scale protein profile of human atherosclerotic coronary arteries.

Recently, de Kleijn et al. analysed carotid endarterectomy specimens from 80 patients that had a secondary cardiovascular event in the 3-year follow-up and 80 sex and age matched eventfree patients, by two HPLC fractionations coupled to MS (de Kleijn et al., 2010). They identified osteopontin as potential biomarker and validated data by assaying its level in a group of 574 patients that underwent carotid endarterectomy and a group of 151 patients that underwent femoral endarterectomy included in the follow-up. Osteopontin resulted highly predictive for secondary manifestations of cardiovascular events in other vascular territories.

Atherosclerosis is characterized by high oxidative and proteolytic activities. This process may lead to a pathological remodelling of aorta characterized by dilatation that can evolve toward vessel wall rupture (aortic aneurysm). Recently, Dejouvencel et al. focused their attention on intraluminal mural thrombus that develops in human abdominal aortic aneurysm (Dejouvencel et al., 2010). In particular, they analysed the protein released from three different aortic layers of the intramural thrombus (luminal, intermediate and abluminal), after 24 hours incubation on RPMI medium, by surface-enhanced laser desorption/ionization (SELDI) TOF MS profiling. They identified a peptide that was largely abundant in newly formed luminal layer respect to the other areas as hemorphin 7, a proteolytic fragment of the hemoglobin. The levels of this peptide were confirmed (by ELISA) to be higher in sera of abdominal aortic aneurysm patients respect to controls, and positively correlated with the volume of the thrombus. This peptide has been suggested as a potential marker of pathological vascular remodelling.

#### **2.2 Vascular cell proteomics**

As mentioned above, the study of atherosclerotic specimens are complicated by both the heterogeneous cellular composition and the inflammatory/proteolytic environment. In this respect, cell culture systems could represent a useful tool to partially overcome drawbacks

of tissue analyses, allowing researchers to study single aspects of the atherosclerotic process in very controlled conditions. In the last years, studies on proteome (the intracellular proteins) and secretome (the proteins released into the cell culture medium) by 2DE and MS of ECs (Bruneel et al., 2003; Chen et al., 2007; Tunica et al., 2009), VSMCs (McGregor et al., 2001; Dupont et al., 2005; Lee et al., 2006), and monocytes/macrophages (Dupont et al., 2004; Fach et al., 2004; Slomianny et al., 2006; Zhang et al., 2007; Zhao et al., 2009) have been performed. Moreover, in the attempt to help in elucidating the mechanisms of atherogenesis, several proteomic studies have been carried out on vascular cells cultured in different experimental conditions (table 3). The most applied proteomic methodologies were 1DE-2DE coupled to MS analyses. Very few studies applied gel free proteomic approaches such as LC MS/MS (Fach et al., 2004; X.L. Wang et al., 2007; Zhao et al., 2009; Tunica et al., 2009; Zimman et al., 2010) and microarrays (Sukhanov et al., 2005).

#### **2.3 Plasma/serum proteomics**

10 Proteomics – Human Diseases and Protein Functions

By using microarray technology, Slevin et al. compared the expression of 512 proteins associated with inflammatory, pro/anti-apoptotic, and angiogenesis pathways between 4 pooled fibrous stable carotid plaques and 4 pooled ulcerated, hemorrhagic unstable plaques (Slevin et al., 2006). In spite of the high sensitivity of protein microarrays, allowing detecting nanogram quantities, errors can occur because of weakly reacting or nonspecific antibodies, degraded proteins, and/or the efficiency of sample fluorescent labelling. However, western blotting analyses confirmed differential expression between stable and unstable plaque pools for all 11 proteins selected suggesting a high level of specificity of the array antibodies

Atherosclerotic tissue could also be laser-microdissected, which would allow one to compare different areas of the plaque such as necrotic core and shoulders/fibrous cap, providing valuable spatial information. Bagnato et al. applied the direct tissue proteomic (DTP) approach to paraffin or frozen blocks from 35 coronary atherosclerotic lesions classified by histopathological examination in early, intermediate, and advanced (Bagnato et al., 2007). In particular, different plaque regions were laser-microdissected (LCMs) from both paraffin and frozen sections and subjected to tryptic digestion followed by LC MS/MS to obtain area-specific proteomic information. Frozen sections were also homogenized and proteins resolved by SDS PAGE and analysed by LC MS/MS. Moreover, they used AQUA (absolute quantitation) methodology to quantify Stromal Cell-derived Factor 1 α (SDF1-α) and growth factors not detected by the above mentioned methods. These multiple approaches allowed them to identify 806 unique proteins with high confidence so obtaining

Recently, de Kleijn et al. analysed carotid endarterectomy specimens from 80 patients that had a secondary cardiovascular event in the 3-year follow-up and 80 sex and age matched eventfree patients, by two HPLC fractionations coupled to MS (de Kleijn et al., 2010). They identified osteopontin as potential biomarker and validated data by assaying its level in a group of 574 patients that underwent carotid endarterectomy and a group of 151 patients that underwent femoral endarterectomy included in the follow-up. Osteopontin resulted highly predictive for

Atherosclerosis is characterized by high oxidative and proteolytic activities. This process may lead to a pathological remodelling of aorta characterized by dilatation that can evolve toward vessel wall rupture (aortic aneurysm). Recently, Dejouvencel et al. focused their attention on intraluminal mural thrombus that develops in human abdominal aortic aneurysm (Dejouvencel et al., 2010). In particular, they analysed the protein released from three different aortic layers of the intramural thrombus (luminal, intermediate and abluminal), after 24 hours incubation on RPMI medium, by surface-enhanced laser desorption/ionization (SELDI) TOF MS profiling. They identified a peptide that was largely abundant in newly formed luminal layer respect to the other areas as hemorphin 7, a proteolytic fragment of the hemoglobin. The levels of this peptide were confirmed (by ELISA) to be higher in sera of abdominal aortic aneurysm patients respect to controls, and positively correlated with the volume of the thrombus. This peptide has been suggested as a

As mentioned above, the study of atherosclerotic specimens are complicated by both the heterogeneous cellular composition and the inflammatory/proteolytic environment. In this respect, cell culture systems could represent a useful tool to partially overcome drawbacks

and the usefulness of this proteomic approach in the study of plaque pathogenesis.

a large scale protein profile of human atherosclerotic coronary arteries.

secondary manifestations of cardiovascular events in other vascular territories.

potential marker of pathological vascular remodelling.

**2.2 Vascular cell proteomics** 

Plasma is one of the most useful matrices to investigate for identifying new biomarkers. It carries resident proteins that represent the majority of plasma proteins, together with proteins released from vascular cells and other tissues. In this respect, variations in plasma proteome could reflect directly or indirectly a cardiovascular disease or other pathological conditions. Moreover, monitoring plasma proteome could be successful in patients follow up and in relation to drug therapies. Besides plasma, also serum is widely investigated, although its proteome is known to be drastically affected by blood coagulation. Plasma proteomic studies are challenging due to both the high dynamic range of protein concentrations and the number of low expressed proteins. Plasma is composed for up to 99% by 21 most abundant proteins and for the remaining 1% by about 50,000 low expressed protein variants, representing the so called "deep proteome" (Righetti et al., 2005). Plasma protein levels range from 40-50 mg/ml for albumin to less than 10 ng/ml for interleukins, chatepsins and peptide hormones. Therefore, differential proteomics of unfractionated plasma provides only limited information. In this regard, several pre-analytical depletion systems that imply affinity/immunoaffinity steps for simultaneous removal of the most represented plasma proteins have been set up. However, due to non-specific binding, many other proteins could be depleted. Besides depletion systems, also protein enrichment technology has been developed. This relies on solid-phase combinatorial ligand libraries, made of hexapeptides, to reduce the high dynamic range of protein concentration preventing the co-depletion of low-abundance proteins (Boschetti et al., 2007). Since variations in plasma protein expression can provide useful information on both physiological and pathological conditions of the different tissues in the body, many efforts have been made in characterizing the entire human plasma proteome. In this regard, during 2003-2005, 55 laboratories worldwide participating to the Human Plasma Proteome Project (HPPP) analysed reference specimens by using emerging technologies in the field of proteomics, and generated integrated databases for proteins detectable and identifiable in human serum and plasma (http://www.hupo.org/research/hppp/). They confidently identified 3020 proteins with a minimum of two high-scoring MS/MS spectra that have been searched for relevance to cardiovascular function and disease using PubMed search engine and specific keywords. On the basis of the current knowledge, the study individuated a subset of 345 proteins showing cardiovascular-related functions (markers of inflammation and/or cardiovascular disease, proteins implicated in coagulation, signalling, growth, differentiation, and vascular remodelling) (Berhane et al., 2005).

Overview of Current Proteomic Approaches

dyslipidemia, hypertension and diabetes.

**2.4 Urine proteomics** 

**2.5 Lipoproteomics** 

for Discovery of Vascular Biomarkers of Atherosclerosis 13

elucidate effects on plasma proteome of different pharmacological treatments (Alonso-Orgaz et al., 2006; López-Farré et al., 2007). In this regard, the most applied proteomic methodologies were gel electrophoresis coupled to mass spectrometry analyses followed by gel free proteomic approaches such as LC MS/MS (Donahue et al., 2006; Wilson et al., 2007; Prentice et al., 2010) and microarrays (Tabibiazar et al., 2006). Almost all of these analytical approaches were preceded by one or more fractionation steps, mainly immunoaffinity depletion of the most abundant plasma protein species and ion exchange chromatography, to reduce the high complexity of plasma samples in terms of both number and dynamic range of protein species. Since many genetic and environmental factors affect atherosclerosis aetiology, one of the main drawbacks in differential analysis is the choice of a proper control group. In particular, results could be affected by coexisting pathological conditions such as

Urine is an easily accessible body fluid, stable against proteolytic degradation even after long storage times, and it represents a rich source of information. Urine protein and peptide composition results from glomerular filtration and proximal tubular absorption of circulating proteins (30%) and from the kidney and the urinary tract (70%) (Decramer et al., 2008). Urine proteomics is emerging as a powerful tool for identifying new biomarkers useful in diagnosis and monitoring of several human diseases. However, the urinary proteome analysis is not a simple task because the urine shows low protein concentration and high levels of salts or other interfering compounds. Moreover, urinary proteome is highly influenced by both inter-individual and intra-individual variability, the latter due to physical training, diet, drugs, caffeine consumption, etc. One of the priorities in this field during the coming years is to optimize sample preparation methods for urine proteomics (Thongboonkerd V., 2007). 2DE coupled with MS has represented for years the technique of choice for the analysis of urine proteins. Recently, Candiano et al. resolved 1118 spots in normal urine samples, 275 of which were characterized as isoforms of 82 proteins, 30 (108 spots) corresponding to typical plasma components (Candiano et al., 2010). However, the identity of most of the proteins found in normal urine by 2DE remains to be determined, the majority being low-molecular weight proteins (<30 kDa). By means of 1DE and HPLC as fractionation methods, and nano LC MS/MS and MS3 as analytical methods, Adachi et al. identified 8041 peptides corresponding to 1543 proteins, probably representing the most

advanced proteomic approach to urine characterization (Adachi et al., 2006).

identifies patients with high sensitivity and specificity has been defined.

The study of urinary proteome in relation to atherosclerosis is in its infancy but, in the last years, many efforts have been done. In particular, multiple urinary biomarkers of CAD have been described (Zimmerli et al., 2008; von Zur Muhlen et al., 2009; Delles et al., 2010) by means of capillary electrophoresis coupled online to micro time-of-flight mass spectrometry. A pattern of 238 CAD-specific polypeptides (49 of which have been sequenced) that

In 1995 Williams and Tabas published the "response-to-retention" hypothesis. According to this theory, early events in atherogenesis are related to a selective retention of LDL in the sub-endothelial space by means of specific interactions with some extracellular matrix components (Williams & Tabas, 1995). The molecular mechanisms underlying these processes are not completely understood, but knowledge of lipoprotein structure,


ACS, acute coronary syndrome; CAD, coronary artery disease

Table 3. Literature overview of experimental conditions adopted in proteomic studies on ECs, VSMCs and monocytes/macrophages.

In the latest years, proteomics has been applied to plasma/serum to identify early diagnostic markers in relation to several cerebro-cardiovascular pathologies such as stroke (Allard et al., 2004; Kiga et al., 2008; Brea et al., 2009; Prentice et al., 2010), acute coronary syndrome (Mateos-Cáceres et al., 2004; Dardé et al., 2010), angiographic coronary disease (Donahue et al., 2006), acute myocardial infarction (AMI) (Distelmaier et al., 2009), coronary heart disease (CHD) (Prentice et al., 2010), coronary artery bypass grafting (CABG) (Banfi et al., 2010), aortic atherosclerotic plaque (Tabibiazar et al., 2006), and peripheral arterial disease (Wilson et al., 2007). Moreover, differential proteomic analysis has been addressed to elucidate effects on plasma proteome of different pharmacological treatments (Alonso-Orgaz et al., 2006; López-Farré et al., 2007). In this regard, the most applied proteomic methodologies were gel electrophoresis coupled to mass spectrometry analyses followed by gel free proteomic approaches such as LC MS/MS (Donahue et al., 2006; Wilson et al., 2007; Prentice et al., 2010) and microarrays (Tabibiazar et al., 2006). Almost all of these analytical approaches were preceded by one or more fractionation steps, mainly immunoaffinity depletion of the most abundant plasma protein species and ion exchange chromatography, to reduce the high complexity of plasma samples in terms of both number and dynamic range of protein species. Since many genetic and environmental factors affect atherosclerosis aetiology, one of the main drawbacks in differential analysis is the choice of a proper control group. In particular, results could be affected by coexisting pathological conditions such as dyslipidemia, hypertension and diabetes.

#### **2.4 Urine proteomics**

12 Proteomics – Human Diseases and Protein Functions

**cells** 

McGregor et al.,

Boccardi et al.,

Delafontaine, 2005; Padró et al.,

Jang et al., 2004; Lee et al., 2006

**Monocytes/macrophages** 

Fach et al., 2004; Conway & Kinter, 2005; Dupont et al., 2004; Kang et al., 2009; Burillo et al., 2009; Y.L. Yu et al., 2003a; Y.L.Yu et al.,

2003b

Won et al., 2011 Barderas et al., 2009

Gadgil et al., 2003; Sintiprungrat et al., 2010

**Endothelial cells Smooth muscle** 

2004

2007

2008

2010

**ACS vs CAD** Barderas et al., 2007

Table 3. Literature overview of experimental conditions adopted in proteomic studies on

In the latest years, proteomics has been applied to plasma/serum to identify early diagnostic markers in relation to several cerebro-cardiovascular pathologies such as stroke (Allard et al., 2004; Kiga et al., 2008; Brea et al., 2009; Prentice et al., 2010), acute coronary syndrome (Mateos-Cáceres et al., 2004; Dardé et al., 2010), angiographic coronary disease (Donahue et al., 2006), acute myocardial infarction (AMI) (Distelmaier et al., 2009), coronary heart disease (CHD) (Prentice et al., 2010), coronary artery bypass grafting (CABG) (Banfi et al., 2010), aortic atherosclerotic plaque (Tabibiazar et al., 2006), and peripheral arterial disease (Wilson et al., 2007). Moreover, differential proteomic analysis has been addressed to

Chen et al., 2007 Sukhanov &

**Experimental conditions** 

**Pro-inflammatory conditions** 

**Oxidized/aggregated** 

**Lipopolysaccharide or phorbol myristate** 

**LDL** 

**Senescence** Kamino et al.,

**Shear stress** X.L. Wang et al.,

**Antioxidant/oxidant** Ha et al., 2005;

**Cholesterol loading** T. Wang et al.,

**HSP27 over-expression** Trott et al., 2009 **Drug treatment** M. Yu et al., 2004;

2003

2009

2007; Huang et al.,

Lomnytska et al., 2004; Pawlowska et al., 2005; González-Cabrero

et al., 2007

Zimman et al.,

Bieler et al., 2009; Millioni et al.,

2010

2006

2010

ACS, acute coronary syndrome; CAD, coronary artery disease

ECs, VSMCs and monocytes/macrophages.

**PKCδ-/-** Mayr et al., 2004 **Hyperinsulinemia** Y. Wang et al.,

Urine is an easily accessible body fluid, stable against proteolytic degradation even after long storage times, and it represents a rich source of information. Urine protein and peptide composition results from glomerular filtration and proximal tubular absorption of circulating proteins (30%) and from the kidney and the urinary tract (70%) (Decramer et al., 2008). Urine proteomics is emerging as a powerful tool for identifying new biomarkers useful in diagnosis and monitoring of several human diseases. However, the urinary proteome analysis is not a simple task because the urine shows low protein concentration and high levels of salts or other interfering compounds. Moreover, urinary proteome is highly influenced by both inter-individual and intra-individual variability, the latter due to physical training, diet, drugs, caffeine consumption, etc. One of the priorities in this field during the coming years is to optimize sample preparation methods for urine proteomics (Thongboonkerd V., 2007). 2DE coupled with MS has represented for years the technique of choice for the analysis of urine proteins. Recently, Candiano et al. resolved 1118 spots in normal urine samples, 275 of which were characterized as isoforms of 82 proteins, 30 (108 spots) corresponding to typical plasma components (Candiano et al., 2010). However, the identity of most of the proteins found in normal urine by 2DE remains to be determined, the majority being low-molecular weight proteins (<30 kDa). By means of 1DE and HPLC as fractionation methods, and nano LC MS/MS and MS3 as analytical methods, Adachi et al. identified 8041 peptides corresponding to 1543 proteins, probably representing the most advanced proteomic approach to urine characterization (Adachi et al., 2006).

The study of urinary proteome in relation to atherosclerosis is in its infancy but, in the last years, many efforts have been done. In particular, multiple urinary biomarkers of CAD have been described (Zimmerli et al., 2008; von Zur Muhlen et al., 2009; Delles et al., 2010) by means of capillary electrophoresis coupled online to micro time-of-flight mass spectrometry. A pattern of 238 CAD-specific polypeptides (49 of which have been sequenced) that identifies patients with high sensitivity and specificity has been defined.

#### **2.5 Lipoproteomics**

In 1995 Williams and Tabas published the "response-to-retention" hypothesis. According to this theory, early events in atherogenesis are related to a selective retention of LDL in the sub-endothelial space by means of specific interactions with some extracellular matrix components (Williams & Tabas, 1995). The molecular mechanisms underlying these processes are not completely understood, but knowledge of lipoprotein structure,

Overview of Current Proteomic Approaches

**VLDL**  1 hyperlipidemic subject vs 3 healthy subjects

**LDL** 

metabolic syndrome and subclinical carotid atherosclerosis vs 10 healthy controls

21 patients with type 2

**HDL**  20 control subjects for total HDL analysis

7 CAD subjects vs 6 control subjects for HDL3

6 CAD subjects treated with niacin and atorvastatin for 12 months vs 6 non treated

18 men with established CAD vs 20 apparently

7 hypercholesterolemic

normolipidemic subjects

3 subjects having low HDL-cholesterol vs 3 subjects having high HDL-cholesterol

CAD subjects

healthy men

subjects vs 9

analysis

10 subjects with

diabetes and atherosclerosis vs 23 healthy controls.

**Subjects analysed Purification methods** 

for Discovery of Vascular Biomarkers of Atherosclerosis 15

**Results Ref.** 

Proteins differentially expressed in small dense LDL: ↑ apo C-III (3 isoforms), ↓ apo C-I (2isoforms), ↓ apo A-I, ↓ apo E

48 proteins identified in total HDL HDL3 analysis: ↑apo C-IV, ↑PON1, ↑complement C3, ↑apo A-IV, ↑apo E

> ↑PLTP, ↑apo F, ↑apo J, ↓apo E

↑apo C-III, ↓apo C-I ↑apo A-I peptides containing oxidized methionine

> ↓apo A-I, ↑apo C-I, ↑apo C-III, ↑apo E

380 peaks ↑ two forms of apo C-III

↑ apo C-III Bondarenko et al.,

1999

Davidsson et al., 2005

Vaisar et al., 2007

Green et al., 2008

Vaisar et al., 2010

Heller et al., 2007

Mazur et al., 2010

**Proteomic methods** 

ultracentrifugation in sucrose or in NaBr density gradient

MALDI-TOF and ESI-TOF MS

ultracentrifugation in D2O density gradient (small dense LDL)

SELDI-TOF MS 1DE MALDI-TOF/TOF WB

ultracentrifugation or affinity chromatography

LC-ESI MS/MS

sequential salt ultracentrifugation (HDL3)

LC–Fourier transform Ion Cyclotron Resonance –MS

sequential salt ultracentrifugation (HDL2)

MALDI-TOF MS and pattern recognition analysis LC-MALDI-TOF/TOF

ultracentrifugation in salt density gradient

Shotgun LC-ESI MS/MS

ultracentrifugation in salt density gradient (HDL3)

Top-down Differential Mass Spectrometry

apolipoprotein composition and their post-translational modifications could help in this respect. There are several types of lipoproteins differing for chemical compositions, physical properties and metabolic functions. They may be classified according to their densities in chylomicrons (d< 0.95 g/ml), very low density lipoproteins (VLDL, d< 1.006 g/ml), low density lipoproteins (LDL, 1.019<d<1.063 g/ml), high density lipoproteins (HDL, 1.063<d<1.21 g/ml) (Vance & Vance, 2008). A further class of lipoprotein particles is known as lipoprotein(a) (Lp(a)). Lp(a) is a LDL-like particle that carries, linked to apoB100 by a single disulfide bond, an heavily glycosylated multi-kringle protein named apolipoprotein(a). The physiological role of Lp(a) is unknown, although it is considered a risk factor for cardiovascular disease. Lipoproteins have attracted a great deal of interest because of their implication in the development of cardiovascular diseases, such as atherosclerosis. Although it is well known that high LDL-cholesterol and low HDLcholesterol are positively correlated with the risk for the development of cardiovascular disease, clinical studies suggest that levels of apo B-100 and apo A-I may be better predictors (Walldius et al., 2001). Since the protein component of these particles is largely responsible for carrying out their various functions, detailed information about the apolipoprotein composition and structure may contribute to reveal their role in atherogenesis and to develop new therapeutic strategies for the treatment of lipoprotein-associated disorders. Applying proteomics to the study of lipoproteins seems to contribute significantly to the achievement of this goal. Indeed, recent proteomic studies have revealed that lipoproteins carry an array of proteins previously unsuspected. Among proteomic approaches, 2DE was applied to the study of lipoprotein particles for the first time in the 1970s (Emes et al., 1976) and allowed to reveal several protein isoforms (Zannis, 1986). With the improvement of 2DE technologies, due to the advent of immobilized pH gradient strips, in the last ten years several studies have been done in the attempt to elucidate the apolipoprotein cargo of the different lipoprotein species. In this respect, besides 2DE, several gel-free mass spectrometry based proteomics have been applied. So far, 31 proteomic studies on VLDL, LDL and HDL have been published, while no proteomic studies on chylomicrons or Lp(a) are present in literature. Among them, only 9 focused on lipoproteomics in relation to atherosclerosis in humans (table 4).

Before overviewing lipoproteomic studies, it is worth mentioning that the method used to isolate lipoproteins significantly affects the protein content of the resulting particles. Traditional methods, established in the 1950s (Havel et al., 1955), imply ultracentrifugation in high-salt media containing KBr or NaBr. Several lipoproteomic studies have been published using these procedures of lipoprotein isolation (Banfi et al., 2009; Davidson et al., 2009; Green et al., 2008; Heller et al., 2005, 2007; Hortin et al., 2006; Karlsson et al., 2005a, 2005b; Khovidhunkit et al., 2004; Mancone et al., 2007; Mazur et al., 2010; Rezaee et al., 2006; Vaisar et al., 2007, 2010; Alwaili et al, 2011). However, the high ionic strength and the high centrifugal field forces might cause either the dissociation of proteins or their exchange between different lipoprotein classes, altering the pattern of associated exchangeable apolipoproteins. Indeed, some of these studies reported a loss of proteins after a second step of ultracentrifugation (Banfi et al., 2009; Davidson et al., 2009; Mancone et al., 2007). Some others employed two ultracentrifuge procedures, using both salts and other compounds, such as sucrose and iodixanol (Bondarenko et al., 1999; Sun et al., 2010), reporting comparable results. By the way, Stahlman et al. reported that deuterium oxide (D2O) is to be preferred over salts at least for LDL and HDL, since for VLDL isolation, the ionic strength of the solution is not so relevant (Ståhlman et al., 2008). Alternatively, lipoprotein can be isolated by means of

apolipoprotein composition and their post-translational modifications could help in this respect. There are several types of lipoproteins differing for chemical compositions, physical properties and metabolic functions. They may be classified according to their densities in chylomicrons (d< 0.95 g/ml), very low density lipoproteins (VLDL, d< 1.006 g/ml), low density lipoproteins (LDL, 1.019<d<1.063 g/ml), high density lipoproteins (HDL, 1.063<d<1.21 g/ml) (Vance & Vance, 2008). A further class of lipoprotein particles is known as lipoprotein(a) (Lp(a)). Lp(a) is a LDL-like particle that carries, linked to apoB100 by a single disulfide bond, an heavily glycosylated multi-kringle protein named apolipoprotein(a). The physiological role of Lp(a) is unknown, although it is considered a risk factor for cardiovascular disease. Lipoproteins have attracted a great deal of interest because of their implication in the development of cardiovascular diseases, such as atherosclerosis. Although it is well known that high LDL-cholesterol and low HDLcholesterol are positively correlated with the risk for the development of cardiovascular disease, clinical studies suggest that levels of apo B-100 and apo A-I may be better predictors (Walldius et al., 2001). Since the protein component of these particles is largely responsible for carrying out their various functions, detailed information about the apolipoprotein composition and structure may contribute to reveal their role in atherogenesis and to develop new therapeutic strategies for the treatment of lipoprotein-associated disorders. Applying proteomics to the study of lipoproteins seems to contribute significantly to the achievement of this goal. Indeed, recent proteomic studies have revealed that lipoproteins carry an array of proteins previously unsuspected. Among proteomic approaches, 2DE was applied to the study of lipoprotein particles for the first time in the 1970s (Emes et al., 1976) and allowed to reveal several protein isoforms (Zannis, 1986). With the improvement of 2DE technologies, due to the advent of immobilized pH gradient strips, in the last ten years several studies have been done in the attempt to elucidate the apolipoprotein cargo of the different lipoprotein species. In this respect, besides 2DE, several gel-free mass spectrometry based proteomics have been applied. So far, 31 proteomic studies on VLDL, LDL and HDL have been published, while no proteomic studies on chylomicrons or Lp(a) are present in literature. Among them, only 9 focused on lipoproteomics in relation to atherosclerosis in

Before overviewing lipoproteomic studies, it is worth mentioning that the method used to isolate lipoproteins significantly affects the protein content of the resulting particles. Traditional methods, established in the 1950s (Havel et al., 1955), imply ultracentrifugation in high-salt media containing KBr or NaBr. Several lipoproteomic studies have been published using these procedures of lipoprotein isolation (Banfi et al., 2009; Davidson et al., 2009; Green et al., 2008; Heller et al., 2005, 2007; Hortin et al., 2006; Karlsson et al., 2005a, 2005b; Khovidhunkit et al., 2004; Mancone et al., 2007; Mazur et al., 2010; Rezaee et al., 2006; Vaisar et al., 2007, 2010; Alwaili et al, 2011). However, the high ionic strength and the high centrifugal field forces might cause either the dissociation of proteins or their exchange between different lipoprotein classes, altering the pattern of associated exchangeable apolipoproteins. Indeed, some of these studies reported a loss of proteins after a second step of ultracentrifugation (Banfi et al., 2009; Davidson et al., 2009; Mancone et al., 2007). Some others employed two ultracentrifuge procedures, using both salts and other compounds, such as sucrose and iodixanol (Bondarenko et al., 1999; Sun et al., 2010), reporting comparable results. By the way, Stahlman et al. reported that deuterium oxide (D2O) is to be preferred over salts at least for LDL and HDL, since for VLDL isolation, the ionic strength of the solution is not so relevant (Ståhlman et al., 2008). Alternatively, lipoprotein can be isolated by means of

humans (table 4).


Overview of Current Proteomic Approaches

subject (Bondarenko et al., 1999).

**2.5.2 LDL** 

many of which were ribosomal proteins (Rashid et al., 2002).

for Discovery of Vascular Biomarkers of Atherosclerosis 17

metabolic changes, in terms of protein composition, during physiological VLDL to LDL transition (Sun et al., 2010). While the studies described above focused on the human lipoproteome of mature VLDL particles, other researches tried to shed light on VLDL assembly and maturation in animal models. For example, Rashid et al. immunopurified apo B from rat liver microsomes treated with chemical crosslinkers. Then, using LC MS/MS technology, they identified 99 unique proteins that co-immunoprecipitated with apo B,

Bondarenko et al. applied MALDI TOF and ESI TOF MS techniques to identify low molecular weight proteins constituting VLDL from 3 healthy subjects and 1 hyperlipidemic subject without previous tryptic digestion. By this approach they identified 15 apo C isoforms and 1 apo E isoform and observed higher level of apo C-III in the hyperlipidemic

LDL particles have been studied using different proteomic approaches. Karlsson et al. provided a 2DE map of LDL-associated proteins from a pooled plasma of 4 healthy subjects purified by KBr density gradient ultracentrifugation. Their results confirmed the presence of proteins known to be associated with LDL particles, showing that many of these were present in different isoforms. In particular, they detected three proteins not previously identified in LDL: serum amyloid A-IV, calgranulin A, and lysozyme C. To confirm that the proteins identified were truly associated with LDL rather than adsorbed during the isolation procedure, LDL was also purified by size-exclusion chromatography (Karlsson et al., 2005a). Moreover, they described three isoforms of apo M that were characterized for PTMs in a following work (Karlsson et al., 2006). Stahlman et al. applied 2DE coupled to MALDI TOF/TOF and SELDI TOF MS as well, to compare VLDL, LDL and HDL isolated from pooled plasma of 5 healthy donors by ultracentrifugation using either KBr or D2O/sucrose to generate the gradient. VLDL profiles obtained with the two procedure of isolation were almost identical. Conversely, 2DE maps and SELDI TOF profiles of LDL and HDL were qualitatively similar, but differed in relative abundance of some protein species. Moreover, a reduced protein-lipids ratio was detected in LDL and HDL fractions purified by using KBr indicating that in the D2O buffer the lipoproteins retained a higher content of exchangeable apoproteins (Ståhlman et al., 2008). LDL-associated proteins have also been studied using other proteomic approaches. Banfi et al. applied liquid-phase IEF and 1DE coupled with LC MS/MS to characterize the proteome of LDL isolated by density gradient ultracentrifugation from healthy subjects. They identified LDL-associated proteins not previously described, including prenylcysteine lyase (PCL1), orosomucoid, retinol-binding protein, and paraoxonase-1. The authors analysed PCL1 distribution in all the lipoprotein classes isolated by ultracentrifugation from 6 healthy subjects showing a decline from VLDL to LDL to HDL and its absence in lipoprotein-depleted plasma. Due to the oxidizing role of PCL1, they hypothesized that lipoproteins can themselves generate pro-oxidant species, thus suggesting a new role for lipoprotein in the development of atherosclerosis (Banfi et al., 2009). Bancells et al. analysed the proteome of LDL subfractions isolated by anion exchange chromatography after sequential ultracentrifugation of pooled healthy subjects plasma. Proteomic analysis, performed by LC MS/MS method, revealed the presence of 28 proteins most of which were involved in inflammation, coagulation and innate immunity, besides apolipoproteins involved in lipid metabolism. They observed that electronegative LDL, a minor subfraction of LDL fraction, has a higher content of minor proteins, especially apo F


↑, increase. ↓, decrease.

Table 4. Overview of the lipoproteomic studies related to atherosclerosis in humans reviewed in the chapter.

immunopurification methods that rely on antibodies specific for the dominant protein of each class (Levels et al., 2007, 2011; Ogorzalek Loo et al., 2004; Rashid et al., 2002; Rezaee et al., 2006). Although this procedure does not lead to loss of weakly associated protein, it tends to nonspecifically co-purify associated proteins as serum contaminants and other lipoprotein fractions having the same antibody target (e.g. apolipoprotein A-I is the main HDL apolipoprotein but it is also present in both VLDL and LDL fractions). Other lipoprotein isolation methods that have been applied in lipoproteomic studies, involve electrophoretic techniques, specifically free solution isotachophoresis (Böttcher et al., 2000), and chromatographic techniques, such as fast protein liquid chromatography (Collins & Olivier, 2010; Richardson et al., 2009) and size exclusion/affinity chromatographies (Gordon et al., 2010).

#### **2.5.1 VLDL**

Both 2DE coupled to MS and gel-free MS approaches have been applied to the study of VLDL protein composition. Mancone et al., by using 2DE coupled to MALDI TOF/TOF MS analysis, provided a detailed map of VLDL, isolated by the classical ultracentrifugation method, from a plasma pool of 3 healthy volunteers. They identified two newly VLDLassociated proteins, namely apo L-I and prenylcysteine lyase that were known to be associated with HDL, and some post-translational modifications of Apo E (Thr212glycosylations) and apo L-I (Ser296phosphorylation) (Mancone et al., 2007). Sun et al. used two different analytical approaches to compare the protein content of VLDL and LDL isolated from pooled samples of healthy subjects by either NaBr or iodixanol gradient ultracentrifugation. By using a gel-free approach based on LC coupled to MS/MS analysis of tryptic digests labeled with iTRAQ (isobaric tag for relative and absolute quantitation) tags, they revealed 15 proteins differentially expressed in the two classes of lipoproteins. By using 2DE coupled with LC MS/MS, they further revealed 6 proteins differentially expressed as well. Moreover, the 5 apo A-I isoforms were found to be phosphorylated. This study, besides describing the VLDL and LDL lipoproteomes, provided insights into the metabolic changes, in terms of protein composition, during physiological VLDL to LDL transition (Sun et al., 2010). While the studies described above focused on the human lipoproteome of mature VLDL particles, other researches tried to shed light on VLDL assembly and maturation in animal models. For example, Rashid et al. immunopurified apo B from rat liver microsomes treated with chemical crosslinkers. Then, using LC MS/MS technology, they identified 99 unique proteins that co-immunoprecipitated with apo B, many of which were ribosomal proteins (Rashid et al., 2002).

Bondarenko et al. applied MALDI TOF and ESI TOF MS techniques to identify low molecular weight proteins constituting VLDL from 3 healthy subjects and 1 hyperlipidemic subject without previous tryptic digestion. By this approach they identified 15 apo C isoforms and 1 apo E isoform and observed higher level of apo C-III in the hyperlipidemic subject (Bondarenko et al., 1999).

#### **2.5.2 LDL**

16 Proteomics – Human Diseases and Protein Functions

**Results Ref.** 

Levels et al., 2011

Alwaili et al., 2011

profound changes in 21 markers in both groups

> 67 proteins identified

↓ apo A-IV ↑SAA ↑complement C3

**Proteomic methods** 

apo A-I immunocapturing

SELDI-TOF MS

KBr sequential ultracentrifugation

1D LC-MS/MS WB ELISA

Table 4. Overview of the lipoproteomic studies related to atherosclerosis in humans

immunopurification methods that rely on antibodies specific for the dominant protein of each class (Levels et al., 2007, 2011; Ogorzalek Loo et al., 2004; Rashid et al., 2002; Rezaee et al., 2006). Although this procedure does not lead to loss of weakly associated protein, it tends to nonspecifically co-purify associated proteins as serum contaminants and other lipoprotein fractions having the same antibody target (e.g. apolipoprotein A-I is the main HDL apolipoprotein but it is also present in both VLDL and LDL fractions). Other lipoprotein isolation methods that have been applied in lipoproteomic studies, involve electrophoretic techniques, specifically free solution isotachophoresis (Böttcher et al., 2000), and chromatographic techniques, such as fast protein liquid chromatography (Collins & Olivier, 2010; Richardson et al., 2009) and size exclusion/affinity chromatographies (Gordon

Both 2DE coupled to MS and gel-free MS approaches have been applied to the study of VLDL protein composition. Mancone et al., by using 2DE coupled to MALDI TOF/TOF MS analysis, provided a detailed map of VLDL, isolated by the classical ultracentrifugation method, from a plasma pool of 3 healthy volunteers. They identified two newly VLDLassociated proteins, namely apo L-I and prenylcysteine lyase that were known to be associated with HDL, and some post-translational modifications of Apo E (Thr212glycosylations) and apo L-I (Ser296phosphorylation) (Mancone et al., 2007). Sun et al. used two different analytical approaches to compare the protein content of VLDL and LDL isolated from pooled samples of healthy subjects by either NaBr or iodixanol gradient ultracentrifugation. By using a gel-free approach based on LC coupled to MS/MS analysis of tryptic digests labeled with iTRAQ (isobaric tag for relative and absolute quantitation) tags, they revealed 15 proteins differentially expressed in the two classes of lipoproteins. By using 2DE coupled with LC MS/MS, they further revealed 6 proteins differentially expressed as well. Moreover, the 5 apo A-I isoforms were found to be phosphorylated. This study, besides describing the VLDL and LDL lipoproteomes, provided insights into the

**Subjects analysed Purification methods** 

10 subjects having low HDL-cholesterol vs 10 subjects having high HDL cholesterol challenged with lipopolysaccharide (24 hours follow up)

10 ACS subjects vs 10 stable CAD vs 10 healthy control

↑, increase. ↓, decrease.

reviewed in the chapter.

subjects

et al., 2010).

**2.5.1 VLDL** 

LDL particles have been studied using different proteomic approaches. Karlsson et al. provided a 2DE map of LDL-associated proteins from a pooled plasma of 4 healthy subjects purified by KBr density gradient ultracentrifugation. Their results confirmed the presence of proteins known to be associated with LDL particles, showing that many of these were present in different isoforms. In particular, they detected three proteins not previously identified in LDL: serum amyloid A-IV, calgranulin A, and lysozyme C. To confirm that the proteins identified were truly associated with LDL rather than adsorbed during the isolation procedure, LDL was also purified by size-exclusion chromatography (Karlsson et al., 2005a). Moreover, they described three isoforms of apo M that were characterized for PTMs in a following work (Karlsson et al., 2006). Stahlman et al. applied 2DE coupled to MALDI TOF/TOF and SELDI TOF MS as well, to compare VLDL, LDL and HDL isolated from pooled plasma of 5 healthy donors by ultracentrifugation using either KBr or D2O/sucrose to generate the gradient. VLDL profiles obtained with the two procedure of isolation were almost identical. Conversely, 2DE maps and SELDI TOF profiles of LDL and HDL were qualitatively similar, but differed in relative abundance of some protein species. Moreover, a reduced protein-lipids ratio was detected in LDL and HDL fractions purified by using KBr indicating that in the D2O buffer the lipoproteins retained a higher content of exchangeable apoproteins (Ståhlman et al., 2008). LDL-associated proteins have also been studied using other proteomic approaches. Banfi et al. applied liquid-phase IEF and 1DE coupled with LC MS/MS to characterize the proteome of LDL isolated by density gradient ultracentrifugation from healthy subjects. They identified LDL-associated proteins not previously described, including prenylcysteine lyase (PCL1), orosomucoid, retinol-binding protein, and paraoxonase-1. The authors analysed PCL1 distribution in all the lipoprotein classes isolated by ultracentrifugation from 6 healthy subjects showing a decline from VLDL to LDL to HDL and its absence in lipoprotein-depleted plasma. Due to the oxidizing role of PCL1, they hypothesized that lipoproteins can themselves generate pro-oxidant species, thus suggesting a new role for lipoprotein in the development of atherosclerosis (Banfi et al., 2009). Bancells et al. analysed the proteome of LDL subfractions isolated by anion exchange chromatography after sequential ultracentrifugation of pooled healthy subjects plasma. Proteomic analysis, performed by LC MS/MS method, revealed the presence of 28 proteins most of which were involved in inflammation, coagulation and innate immunity, besides apolipoproteins involved in lipid metabolism. They observed that electronegative LDL, a minor subfraction of LDL fraction, has a higher content of minor proteins, especially apo F

Overview of Current Proteomic Approaches

al., 2004).

for Discovery of Vascular Biomarkers of Atherosclerosis 19

mainly in iHDL and sHDL with only little apo C-III in fHDL (Böttcher et al., 2000). A study performed by Ogorzalek Loo et al. suggests that a synergy between classical 2D gels and virtual 2D gels can be useful for studying HDL protein composition. Virtual 2DE is based on combining a first-dimensional isoelectric focusing (IEF) separation on polyacrylamide gels with MALDI MS surface scanning of the dried gel. In such a way, a virtual 2D gel can be created, generating an image in which mass spectrometry substitutes the second-dimension SDS-PAGE separation. By this approach the authors examined HDL isolated from human sera by selected-affinity immunosorption of apo A-I and revealed 42 unique masses for protein species with isoelectric points between pH 5.47–5.04 (Ogorzalek Loo et al., 2004). Heller et al. by using multiple proteomic approaches such as native or denaturing PAGE coupled with LC MS/MS, shotgun LC MS/MS and MALDI TOF MS profiling, analysed the protein complement of HDL3, HDL2, HDL2/LDL and LDL/VLDL enriched fractions, isolated from a plasma pool of 10,000 healthy donors by density gradient ultracentrifugation. Therefore, they were able to characterize comprehensively the protein composition of the purified lipoprotein fractions (Heller et al., 2005). Karlsson et al. provided a detailed 2DE map of HDL2 and HDL3 isolated by salt gradient ultracentrifugation from pooled plasma of 4 healthy volunteers. Besides several isoforms of apolipoproteins already described to associate with HDL, they identified new proteins such as α-1-antitrypsin, two isoforms of salivary-α-amylase in HDL2 and a glycosylated apoAII in HDL3 (Karlsson et al., 2005b). By using 1DE and 2DE MALDI TOF MS and isotope-coded affinity tag (ICAT), Rezaee et al. detected many more proteins than Karlsson et al. in ultracentrifugally isolated HDL. This was the first study employing an ICAT method to identify lower abundance proteins. The overall identified proteins are known to be involved in different functions, such as lipid transport and metabolism, inflammation, immune system, hemostasis and thrombosis (Rezaee et al., 2006). The higher number of identified proteins could be ascribed to the use of ICAT method, that improve the sensitivity of the detection, as well as to the use of a single step of ultracentrifugation to isolate HDL. Khovidhunkit et al. investigated changes in proteins associated to HDL during inflammation by means of 2DE and LC MS/MS in an animal model. For this purpose, they analysed HDL isolated by salt gradient ultracentrifugation from sera of mice injected with normal saline or with endotoxin so detecting increased levels of SAA, apo E, apo A-IV and apo A-V and decreased levels of apo A-I and apo A-II in acute-phase HDL (Khovidhunkit et

Besides 2DE analyses, several groups have applied different gel-free proteomic approaches to characterize HDL proteome in healthy subjects. One of the first attempts was performed by Bondarenko et al. that used MALDI TOF MS and ESI TOF MS techniques to the analysis of intact protein of HDL isolated by density gradient ultracentrifugation in sucrose solution. They observed forty-nine peaks in the MALDI spectrum and 11 species in the ESI MS spectrum corresponding to the most abundant apolipoproteins, such as apo A-I, apo A-II, apo C-I, apo C-II, and apo C-III showing different isoforms due to post-translational modifications (Bondarenko et al., 2002). Applying immobilized pH gradient isoelectric focusing coupled with MALDI TOF MS, Farwig et al. were able to detect also SAA-IV in HDL isolated by ultracentrifugation in CsBiEDTA. They developed a successful method for recovering the apolipoproteins from immobilized pH gradient gels prior to MALDI analysis, demonstrating the analytical power of linking the IPG pI profile with MALDI TOF MS analysis (Farwig et al., 2003). Hortin et al. focused on HDL-associated low molecular weight peptides. By using HPLC and MALDI TOF MS or HPLC-ion trap mass spectrometry,

and apo J, compared to electropositive LDL (Bancells et al, 2010). Collins et al. performed a proteomic analysis, applying LC MS/MS, to compare the lipoprotein-associated proteins derived from plasma and serum samples. They isolated both HDL and LDL from healthy subjects by means of fast protein liquid chromatography-size exclusion chromatography (FPLC-SEC). 16 proteins, several of which were complement subcomponents, were found only in the LDL fraction. 65 proteins were identified to be unique to HDL, while another list of proteins was found to overlap between the two lipoprotein fractions. Regarding the differences between plasma- and serum-derived LDL and HDL particles, the authors reported that the most relevant differences regarded fibrinogen proteins which were depleted in serum. Therefore, they stated that, apart from significantly higher levels of apo B-100 in LDL purified from serum samples, comparative proteomic analysis of plasma and serum gives similar results (Collins & Olivier, 2010).

Up to date, only few studies on LDL proteomics and atherosclerosis have been reported. 2DE coupled with LC MS/MS and label-free quantitative MS (LFQMS) was applied by Richardson et al. to the analysis of LDL in the early stages of atherosclerosis in an animal model. LDL was isolated by fast protein LC (FPLC) from non-diabetic hyperlipidemic, diabetic dyslipidemic, diabetic dyslipidemic under exercise training, and healthy Yucatan pigs (Richardson et al., 2009). They identified 28 unique proteins and detected several differential expression patterns for apo E, A-I, C-III, fibrinogen, apo B, adiponectin, alpha-2 macroblobulin, complement C1q, ficolin, and apo J. Since LDL was isolated from pigs in the early stages of atherosclerosis, the alterations observed might be involved in the initiating stages of the disease. LDL-associated proteins have also been studied using other proteomic approaches. For example Davidsson et al. applied SELDI TOF technologies to compare LDL associated proteins from atherosclerotic patients (having either metabolic syndrome or diabetes) to that from healthy subjects. They focused on small dense LDL isolated by means of gradient ultracentrifugation using D2O. The results showed that LDL from patients had lower content of apo A-I, apo C-I and apo E and higher content of apo C-III, the latter responsible for higher affinity for arterial proteoglycans that could facilitate LDL in situ oxidative modifications (Davidsson et al., 2005).

#### **2.5.3 HDL**

HDL is the most studied among lipoprotein particles, probably because of its antiatherogenic functions. Proteomic studies in humans succeeded in identifying, besides the known apolipoproteins involved in the lipoprotein metabolism, other associated proteins such as acute-phase response proteins, proteinase inhibitors, and members of the complement activation. Therefore, characterizing the HDL proteome should help in the identification of novel anti-inflammatory and cardioprotective actions of HDL and could provide insights into lipid therapy. The most used among the several proteomic approaches that have been applied to characterize the HDL-associated proteins is 2DE coupled with MS. Böttcher et al. applied two-dimensional non-denaturing gradient gel electrophoresis (2D-GGE) and immunoblotting to analyse HDL subfractions isolated from healthy subjects. By means of free solution isotachophoresis (FS-ITP), they separated 3 HDL subfractions, namely fast (fHDL), intermediate (iHDL) and slow-migrating (sHDL). Proteomic analysis showed compositional differences in HDL subfractions. In particular, they observed that fHDL and iHDL contained the bulk of HDL and of apo A-I. Apolipoproteins other than apo A-I and apo A-II were not detectable in fHDL, while sHDL contained several minor apolipoproteins such as apo A-IV, apo D, apo E, apo J, and factor H. Apo C-III was found

and apo J, compared to electropositive LDL (Bancells et al, 2010). Collins et al. performed a proteomic analysis, applying LC MS/MS, to compare the lipoprotein-associated proteins derived from plasma and serum samples. They isolated both HDL and LDL from healthy subjects by means of fast protein liquid chromatography-size exclusion chromatography (FPLC-SEC). 16 proteins, several of which were complement subcomponents, were found only in the LDL fraction. 65 proteins were identified to be unique to HDL, while another list of proteins was found to overlap between the two lipoprotein fractions. Regarding the differences between plasma- and serum-derived LDL and HDL particles, the authors reported that the most relevant differences regarded fibrinogen proteins which were depleted in serum. Therefore, they stated that, apart from significantly higher levels of apo B-100 in LDL purified from serum samples, comparative proteomic analysis of plasma and

Up to date, only few studies on LDL proteomics and atherosclerosis have been reported. 2DE coupled with LC MS/MS and label-free quantitative MS (LFQMS) was applied by Richardson et al. to the analysis of LDL in the early stages of atherosclerosis in an animal model. LDL was isolated by fast protein LC (FPLC) from non-diabetic hyperlipidemic, diabetic dyslipidemic, diabetic dyslipidemic under exercise training, and healthy Yucatan pigs (Richardson et al., 2009). They identified 28 unique proteins and detected several differential expression patterns for apo E, A-I, C-III, fibrinogen, apo B, adiponectin, alpha-2 macroblobulin, complement C1q, ficolin, and apo J. Since LDL was isolated from pigs in the early stages of atherosclerosis, the alterations observed might be involved in the initiating stages of the disease. LDL-associated proteins have also been studied using other proteomic approaches. For example Davidsson et al. applied SELDI TOF technologies to compare LDL associated proteins from atherosclerotic patients (having either metabolic syndrome or diabetes) to that from healthy subjects. They focused on small dense LDL isolated by means of gradient ultracentrifugation using D2O. The results showed that LDL from patients had lower content of apo A-I, apo C-I and apo E and higher content of apo C-III, the latter responsible for higher affinity for arterial proteoglycans that could facilitate LDL in situ

HDL is the most studied among lipoprotein particles, probably because of its antiatherogenic functions. Proteomic studies in humans succeeded in identifying, besides the known apolipoproteins involved in the lipoprotein metabolism, other associated proteins such as acute-phase response proteins, proteinase inhibitors, and members of the complement activation. Therefore, characterizing the HDL proteome should help in the identification of novel anti-inflammatory and cardioprotective actions of HDL and could provide insights into lipid therapy. The most used among the several proteomic approaches that have been applied to characterize the HDL-associated proteins is 2DE coupled with MS. Böttcher et al. applied two-dimensional non-denaturing gradient gel electrophoresis (2D-GGE) and immunoblotting to analyse HDL subfractions isolated from healthy subjects. By means of free solution isotachophoresis (FS-ITP), they separated 3 HDL subfractions, namely fast (fHDL), intermediate (iHDL) and slow-migrating (sHDL). Proteomic analysis showed compositional differences in HDL subfractions. In particular, they observed that fHDL and iHDL contained the bulk of HDL and of apo A-I. Apolipoproteins other than apo A-I and apo A-II were not detectable in fHDL, while sHDL contained several minor apolipoproteins such as apo A-IV, apo D, apo E, apo J, and factor H. Apo C-III was found

serum gives similar results (Collins & Olivier, 2010).

oxidative modifications (Davidsson et al., 2005).

**2.5.3 HDL** 

mainly in iHDL and sHDL with only little apo C-III in fHDL (Böttcher et al., 2000). A study performed by Ogorzalek Loo et al. suggests that a synergy between classical 2D gels and virtual 2D gels can be useful for studying HDL protein composition. Virtual 2DE is based on combining a first-dimensional isoelectric focusing (IEF) separation on polyacrylamide gels with MALDI MS surface scanning of the dried gel. In such a way, a virtual 2D gel can be created, generating an image in which mass spectrometry substitutes the second-dimension SDS-PAGE separation. By this approach the authors examined HDL isolated from human sera by selected-affinity immunosorption of apo A-I and revealed 42 unique masses for protein species with isoelectric points between pH 5.47–5.04 (Ogorzalek Loo et al., 2004). Heller et al. by using multiple proteomic approaches such as native or denaturing PAGE coupled with LC MS/MS, shotgun LC MS/MS and MALDI TOF MS profiling, analysed the protein complement of HDL3, HDL2, HDL2/LDL and LDL/VLDL enriched fractions, isolated from a plasma pool of 10,000 healthy donors by density gradient ultracentrifugation. Therefore, they were able to characterize comprehensively the protein composition of the purified lipoprotein fractions (Heller et al., 2005). Karlsson et al. provided a detailed 2DE map of HDL2 and HDL3 isolated by salt gradient ultracentrifugation from pooled plasma of 4 healthy volunteers. Besides several isoforms of apolipoproteins already described to associate with HDL, they identified new proteins such as α-1-antitrypsin, two isoforms of salivary-α-amylase in HDL2 and a glycosylated apoAII in HDL3 (Karlsson et al., 2005b). By using 1DE and 2DE MALDI TOF MS and isotope-coded affinity tag (ICAT), Rezaee et al. detected many more proteins than Karlsson et al. in ultracentrifugally isolated HDL. This was the first study employing an ICAT method to identify lower abundance proteins. The overall identified proteins are known to be involved in different functions, such as lipid transport and metabolism, inflammation, immune system, hemostasis and thrombosis (Rezaee et al., 2006). The higher number of identified proteins could be ascribed to the use of ICAT method, that improve the sensitivity of the detection, as well as to the use of a single step of ultracentrifugation to isolate HDL. Khovidhunkit et al. investigated changes in proteins associated to HDL during inflammation by means of 2DE and LC MS/MS in an animal model. For this purpose, they analysed HDL isolated by salt gradient ultracentrifugation from sera of mice injected with normal saline or with endotoxin so detecting increased levels of SAA, apo E, apo A-IV and apo A-V and decreased levels of apo A-I and apo A-II in acute-phase HDL (Khovidhunkit et al., 2004).

Besides 2DE analyses, several groups have applied different gel-free proteomic approaches to characterize HDL proteome in healthy subjects. One of the first attempts was performed by Bondarenko et al. that used MALDI TOF MS and ESI TOF MS techniques to the analysis of intact protein of HDL isolated by density gradient ultracentrifugation in sucrose solution. They observed forty-nine peaks in the MALDI spectrum and 11 species in the ESI MS spectrum corresponding to the most abundant apolipoproteins, such as apo A-I, apo A-II, apo C-I, apo C-II, and apo C-III showing different isoforms due to post-translational modifications (Bondarenko et al., 2002). Applying immobilized pH gradient isoelectric focusing coupled with MALDI TOF MS, Farwig et al. were able to detect also SAA-IV in HDL isolated by ultracentrifugation in CsBiEDTA. They developed a successful method for recovering the apolipoproteins from immobilized pH gradient gels prior to MALDI analysis, demonstrating the analytical power of linking the IPG pI profile with MALDI TOF MS analysis (Farwig et al., 2003). Hortin et al. focused on HDL-associated low molecular weight peptides. By using HPLC and MALDI TOF MS or HPLC-ion trap mass spectrometry,

Overview of Current Proteomic Approaches

samples (Mazur et al., 2010).

subjects (paper in preparation).

western blotting and ELISA (Alwaili et al, 2011).

for Discovery of Vascular Biomarkers of Atherosclerosis 21

and phospholipid transfer protein levels (Green et al., 2008). In a successive study, they investigated if protein composition was altered in HDL2 isolated from CAD patients. Ultracentrifugally isolated HDL2 was digested with trypsin and analysed by MALDI TOF MS and pattern recognition analysis. The most significant informative features were then subjected to LC MALDI MS/MS for identification. This analysis revealed that HDL2 of CAD subjects carried a distinct protein cargo with increased levels of apo C-III and decreased levels of apo C-I, two apolipoproteins involved in the metabolism of HDL particles. Moreover, they found increased levels of apo A-I peptides containing oxidized methionine indicating the occurence of oxidative processes in CAD patients (Vaisar et al., 2010). Heller et al. used a shotgun LC MS/MS approach to characterize HDL protein composition of 7 hypercholesterolemic subjects and 9 normolipidemic ones. They used the peptide match score summation index, based on probabilistic peptide scores for absolute protein quantitation. By this approach, they found that in hypercholesterolemic subjects apo A-I levels were reduced while apo C-I, apo C-III, and apo E levels were increased, suggesting that HDL protein composition could be altered in lipemic disease (Heller et al., 2007). Mazur et al. applied differential top-down mass spectrometry to compare HDL3 protein profiles between 3 subjects having low HDL cholesterol and 3 subjects having high HDL cholesterol. Differently from the so called "bottom up" proteomic methods that are based on the digestion of proteins into short peptides, "top-down" proteomic techniques characterize intact proteins. In this study HDL3 samples were analysed by a reverse-phase nano-HPLC coupled to a linear trap quadrupole Fourier transform (LTQ-FT) hybrid mass spectrometer. The authors found 380 peaks that changed significantly in protein abundance between high HDL-c and low HDL-c subject groups demonstrating that this approach is suitable for the detection of quantitative differences in proteins and protein isoforms in human HDL

Very recently, Levels et al., applied SELDI TOF MS to HDL isolated, by apo A-I immunocapturing, from healthy subjects having low HDL-c and high HDL-c challenged with an endotoxin for 24 hours. Overall they observed profound changes in 21 markers in both study groups proteome irrespective of HDL cholesterol levels (Levels et al., 2011). Alwaili et al. applied 1D followed by LC-MS/MS to HDL isolated by sequential ultracentrifugation from male control, stable CAD, and ACS subjects (n=10/group). They identified 67 HDL-associated proteins involved in lipid binding, acute-phase response, immune response, and endopeptidase/protease inhibition. By means of spectral counting they found that nine proteins were differently abundant. Among them, apo A-IV was significantly reduced, whereas serum amyloid A and complement C3 were significantly increased in ACS patients compared to either controls or CAD subjects, as confirmed by

Recently, our research group started to study lipoproteomic profiles in relation to atherosclerosis (Formato et al., 2011). For this purpose we purified plasma VLDL, LDL, and HDL from patients undergoing carotid endarterectomy and from healthy normolipidemic donors by single isopycnic salt density gradient ultracentrifugation, followed by a second step of ultracentrifugation. Samples were subjected to 2DE followed by PMF analysis as reported in figure 2. In this way, we identified 21 spots corresponding to about 96% of 52 protein spots detected in VLDL, 22 spots corresponding to about 92% of 43 spots in LDL, and 20 spots corresponding to about 96% of 60 spots in HDL. The relative abundance of several identified lipoprotein-associated proteins differed between patients and healthy

68 peptides in the 1-5 kDa size range were identified in ultracentrifugally isolated HDL. Among these, 19 were fragments derived from well-known HDL-associated protein while others were derived from non-lipoprotein plasma proteins as fibrinogen, α1-proteinase inhibitor, and transthyretin, suggesting that HDL particles may represent significant reservoirs of small peptides in the circulation (Hortin et al., 2006). Levels et al. applied SELDI TOF MS technologies to HDL isolated from normolipidemic individuals by means of immunocapturing directly on a SELDI protein chip covalently bound with anti-apo AI or anti apo-AII antibodies. In this way, 95 peaks in the 3–50 kDa molecular mass range and 27 more peaks between 50 and 160 kDa were detected (Levels et al., 2007). Gordon et al. applied MS-based proteomic approaches to the analysis of HDL purified from healthy subjects by means of gel filtration chromatography. To overcome problems related to nonspecific co-purification, they isolated only phospholipid-containing particles using calcium silicate hydrate (CSH), that were subjected to trypsin digestion while still bound to the CSH for identification by means of LC MS/MS. By this approach 47 proteins were identified. Among these, 14 were described as newly discovered HDL-associated proteins that support roles for HDL in complement regulation and protease inhibition (Gordon et al., 2010). To investigate the role of specific subspecies in the anti-atherogenic effects of HDL, Davidson et al. applied LC MS/MS to investigate the distribution of associated proteins across 5 subpopulations of HDL from healthy human volunteers. Subjecting one set of samples to sequential ultracentrifugation followed by salt gradient ultracentrifugation, and the other one to a single step of salt gradient ultracentrifugation they identified 22 and 28 proteins, respectively. Among them, the majority were apolipoproteins already known to be associated with ultracentrifugally-isolated HDL, while several complement factors and protease inhibitors already documented in other proteomic studies were not detected. By using peptide counts determined by MS, they monitored the relative abundance of a given protein across the HDL subfractions. Some proteins were found to associate preferentially to a specific subclass, while others were uniformly distributed across the subpopulations. This finding supports the proposal that HDL is composed of distinct subpopulations of particles that have discreet biological properties (Davidson et al., 2009).

A limited number of studies have focused on HDL proteomes in relation to atherosclerosis. Vaisar et al. used a shotgun LC MS/MS approach to identify proteins associated to total plasma HDL isolated from 20 healthy individuals. In this way, they described 48 proteins, 13 of which not yet known to associate to HDL. Moreover, they compared plasma HDL3 fraction isolated from 6 healthy donors and 7 CAD patients. By means of Gene Ontology (GO) Consortium analysis, they were able to associate the array of HDL proteins to biological processes. Members of the complement pathway and endopeptidase inhibitors were found, suggesting that HDL plays also roles in regulating the complement system and protecting tissue from proteolysis. Thereafter, they found that some proteins associated to HDL3 were upregulated in CAD patients, in particular apo C-IV, PON1, complement C3, apo A-IV, and apo E. Interestingly, they found three of these proteins also in HDL isolated from human carotid atherosclerotic tissues, being apo E the most abundant (Vaisar et al., 2007). In another study, they investigated whether combined statin and niacin therapy, which increase HDL cholesterol levels and reduce CAD risk, could reverse the changes in the protein composition observed in HDL3. For this purpose HDL3, isolated from 6 CAD patients before and 1 year after combined therapy, were subjected to LC–Fourier Transform Ion Cyclotron Resonance MS. By means of spectral counting and extracted ion chromatograms they found that treatment decreased apo E levels and increased apo J, apo F,

68 peptides in the 1-5 kDa size range were identified in ultracentrifugally isolated HDL. Among these, 19 were fragments derived from well-known HDL-associated protein while others were derived from non-lipoprotein plasma proteins as fibrinogen, α1-proteinase inhibitor, and transthyretin, suggesting that HDL particles may represent significant reservoirs of small peptides in the circulation (Hortin et al., 2006). Levels et al. applied SELDI TOF MS technologies to HDL isolated from normolipidemic individuals by means of immunocapturing directly on a SELDI protein chip covalently bound with anti-apo AI or anti apo-AII antibodies. In this way, 95 peaks in the 3–50 kDa molecular mass range and 27 more peaks between 50 and 160 kDa were detected (Levels et al., 2007). Gordon et al. applied MS-based proteomic approaches to the analysis of HDL purified from healthy subjects by means of gel filtration chromatography. To overcome problems related to nonspecific co-purification, they isolated only phospholipid-containing particles using calcium silicate hydrate (CSH), that were subjected to trypsin digestion while still bound to the CSH for identification by means of LC MS/MS. By this approach 47 proteins were identified. Among these, 14 were described as newly discovered HDL-associated proteins that support roles for HDL in complement regulation and protease inhibition (Gordon et al., 2010). To investigate the role of specific subspecies in the anti-atherogenic effects of HDL, Davidson et al. applied LC MS/MS to investigate the distribution of associated proteins across 5 subpopulations of HDL from healthy human volunteers. Subjecting one set of samples to sequential ultracentrifugation followed by salt gradient ultracentrifugation, and the other one to a single step of salt gradient ultracentrifugation they identified 22 and 28 proteins, respectively. Among them, the majority were apolipoproteins already known to be associated with ultracentrifugally-isolated HDL, while several complement factors and protease inhibitors already documented in other proteomic studies were not detected. By using peptide counts determined by MS, they monitored the relative abundance of a given protein across the HDL subfractions. Some proteins were found to associate preferentially to a specific subclass, while others were uniformly distributed across the subpopulations. This finding supports the proposal that HDL is composed of distinct subpopulations of particles

that have discreet biological properties (Davidson et al., 2009).

A limited number of studies have focused on HDL proteomes in relation to atherosclerosis. Vaisar et al. used a shotgun LC MS/MS approach to identify proteins associated to total plasma HDL isolated from 20 healthy individuals. In this way, they described 48 proteins, 13 of which not yet known to associate to HDL. Moreover, they compared plasma HDL3 fraction isolated from 6 healthy donors and 7 CAD patients. By means of Gene Ontology (GO) Consortium analysis, they were able to associate the array of HDL proteins to biological processes. Members of the complement pathway and endopeptidase inhibitors were found, suggesting that HDL plays also roles in regulating the complement system and protecting tissue from proteolysis. Thereafter, they found that some proteins associated to HDL3 were upregulated in CAD patients, in particular apo C-IV, PON1, complement C3, apo A-IV, and apo E. Interestingly, they found three of these proteins also in HDL isolated from human carotid atherosclerotic tissues, being apo E the most abundant (Vaisar et al., 2007). In another study, they investigated whether combined statin and niacin therapy, which increase HDL cholesterol levels and reduce CAD risk, could reverse the changes in the protein composition observed in HDL3. For this purpose HDL3, isolated from 6 CAD patients before and 1 year after combined therapy, were subjected to LC–Fourier Transform Ion Cyclotron Resonance MS. By means of spectral counting and extracted ion chromatograms they found that treatment decreased apo E levels and increased apo J, apo F, and phospholipid transfer protein levels (Green et al., 2008). In a successive study, they investigated if protein composition was altered in HDL2 isolated from CAD patients. Ultracentrifugally isolated HDL2 was digested with trypsin and analysed by MALDI TOF MS and pattern recognition analysis. The most significant informative features were then subjected to LC MALDI MS/MS for identification. This analysis revealed that HDL2 of CAD subjects carried a distinct protein cargo with increased levels of apo C-III and decreased levels of apo C-I, two apolipoproteins involved in the metabolism of HDL particles. Moreover, they found increased levels of apo A-I peptides containing oxidized methionine indicating the occurence of oxidative processes in CAD patients (Vaisar et al., 2010). Heller et al. used a shotgun LC MS/MS approach to characterize HDL protein composition of 7 hypercholesterolemic subjects and 9 normolipidemic ones. They used the peptide match score summation index, based on probabilistic peptide scores for absolute protein quantitation. By this approach, they found that in hypercholesterolemic subjects apo A-I levels were reduced while apo C-I, apo C-III, and apo E levels were increased, suggesting that HDL protein composition could be altered in lipemic disease (Heller et al., 2007). Mazur et al. applied differential top-down mass spectrometry to compare HDL3 protein profiles between 3 subjects having low HDL cholesterol and 3 subjects having high HDL cholesterol. Differently from the so called "bottom up" proteomic methods that are based on the digestion of proteins into short peptides, "top-down" proteomic techniques characterize intact proteins. In this study HDL3 samples were analysed by a reverse-phase nano-HPLC coupled to a linear trap quadrupole Fourier transform (LTQ-FT) hybrid mass spectrometer. The authors found 380 peaks that changed significantly in protein abundance between high HDL-c and low HDL-c subject groups demonstrating that this approach is suitable for the detection of quantitative differences in proteins and protein isoforms in human HDL samples (Mazur et al., 2010).

Very recently, Levels et al., applied SELDI TOF MS to HDL isolated, by apo A-I immunocapturing, from healthy subjects having low HDL-c and high HDL-c challenged with an endotoxin for 24 hours. Overall they observed profound changes in 21 markers in both study groups proteome irrespective of HDL cholesterol levels (Levels et al., 2011). Alwaili et al. applied 1D followed by LC-MS/MS to HDL isolated by sequential ultracentrifugation from male control, stable CAD, and ACS subjects (n=10/group). They identified 67 HDL-associated proteins involved in lipid binding, acute-phase response, immune response, and endopeptidase/protease inhibition. By means of spectral counting they found that nine proteins were differently abundant. Among them, apo A-IV was significantly reduced, whereas serum amyloid A and complement C3 were significantly increased in ACS patients compared to either controls or CAD subjects, as confirmed by western blotting and ELISA (Alwaili et al, 2011).

Recently, our research group started to study lipoproteomic profiles in relation to atherosclerosis (Formato et al., 2011). For this purpose we purified plasma VLDL, LDL, and HDL from patients undergoing carotid endarterectomy and from healthy normolipidemic donors by single isopycnic salt density gradient ultracentrifugation, followed by a second step of ultracentrifugation. Samples were subjected to 2DE followed by PMF analysis as reported in figure 2. In this way, we identified 21 spots corresponding to about 96% of 52 protein spots detected in VLDL, 22 spots corresponding to about 92% of 43 spots in LDL, and 20 spots corresponding to about 96% of 60 spots in HDL. The relative abundance of several identified lipoprotein-associated proteins differed between patients and healthy subjects (paper in preparation).

Overview of Current Proteomic Approaches

by the several proteomic studies.

**4. Acknowledgments** 

(Grant N° 536/2011.732).

**5. References** 

**3. Conclusions** 

for Discovery of Vascular Biomarkers of Atherosclerosis 23

Plaque rupture and thrombosis are the most important clinical complications in the pathogenesis of vascular diseases. To date, many efforts have been done to elucidate mechanisms underlying plaque vulnerability and to identify reliable specific markers of plaques prone to rupture. In the last years, with the improvement of proteomic tools, largescale technologies have been proved valuable in attempting to unravel pathways of complex diseases and biomarkers for early diagnosis and patients follow up. Collecting multiple biomarkers would be preferable over single markers in terms of higher sensitivity and specificity for the diagnosis of cardiovascular diseases. In this chapter, we have reviewed a great deal of information obtained by applying proteomics to the study of proteome/secretome from atherosclerotic tissues and plasma lipoproteins. In tissue proteomics, major drawbacks such as the plaque complexity, tissue sampling and availability, and the choice of the proper controls could affect the analysis. Even though results reported above seem to be quite promising, large-scale clinical studies are required to validate the usefulness of newly identified biomarkers. Moreover, there are several aspects of the atherosclerotic process that deserve further investigation. The analysis of laser captured microdissections by proteomics is still in its infancy but it could reveal valuable topological differences between specific areas of such a heterogeneous environment. Atherosclerotic plaques are characterized by the presence of an imbalance between oxidant and antioxidant species toward the former, leading to deep protein modifications. In this respect, recent advances in protein post-translational modifications analysis by mass spectrometry could be helpful. To date, many studies have been performed on proteome of purified plasma lipoproteins focusing mainly on HDL and LDL due to their association with atherosclerosis. As far as we know, no proteomic analyses have been performed on Lp(a). Since it is well known that elevated Lp(a) plasma levels are an important risk factor in atherogenesis, it would be of great interest to elucidate its apolipoprotein composition in relation to cardiovascular diseases. Another promising topic for future investigations is the characterization of proteomes of lipoproteins retained in atherosclerotic plaque. Finally, besides the great deal of work to be done in the future in both improving proteomic technologies and providing clues for the many aspects not yet investigated, it will also be necessary to put efforts on a comprehensive analysis of the huge quantity of data provided

This work has been supported by Fondazione Banco di Sardegna (Sassari, Italy) grant

Adachi, J.; Kumar, C.; Zhang, Y.; Olsen, J.V. & Mann, M. (2006). The human urinary

Allard, L.; Lescuyer, P.; Burgess, J.; Leung, K.Y.; Ward, M.; Walter, N.; Burkhard, P.R.;

membrane proteins, *Genome Biol*, Vol.7, No. 9, pp. R80

proteome contains more than 1500 proteins, including a large proportion of

Corthals, G.; Hochstrasser, D.F. & Sanchez, J.C. (2004). ApoC-I and ApoC-III as

Fig. 2. Schematic workflow of the proteomic analysis of plasma lipoproteins adopted in our laboratory. Representative 2D maps of isolated VLDL and whole plasma are reported.

#### **3. Conclusions**

22 Proteomics – Human Diseases and Protein Functions

Fig. 2. Schematic workflow of the proteomic analysis of plasma lipoproteins adopted in our laboratory. Representative 2D maps of isolated VLDL and whole plasma are reported.

Plaque rupture and thrombosis are the most important clinical complications in the pathogenesis of vascular diseases. To date, many efforts have been done to elucidate mechanisms underlying plaque vulnerability and to identify reliable specific markers of plaques prone to rupture. In the last years, with the improvement of proteomic tools, largescale technologies have been proved valuable in attempting to unravel pathways of complex diseases and biomarkers for early diagnosis and patients follow up. Collecting multiple biomarkers would be preferable over single markers in terms of higher sensitivity and specificity for the diagnosis of cardiovascular diseases. In this chapter, we have reviewed a great deal of information obtained by applying proteomics to the study of proteome/secretome from atherosclerotic tissues and plasma lipoproteins. In tissue proteomics, major drawbacks such as the plaque complexity, tissue sampling and availability, and the choice of the proper controls could affect the analysis. Even though results reported above seem to be quite promising, large-scale clinical studies are required to validate the usefulness of newly identified biomarkers. Moreover, there are several aspects of the atherosclerotic process that deserve further investigation. The analysis of laser captured microdissections by proteomics is still in its infancy but it could reveal valuable topological differences between specific areas of such a heterogeneous environment. Atherosclerotic plaques are characterized by the presence of an imbalance between oxidant and antioxidant species toward the former, leading to deep protein modifications. In this respect, recent advances in protein post-translational modifications analysis by mass spectrometry could be helpful. To date, many studies have been performed on proteome of purified plasma lipoproteins focusing mainly on HDL and LDL due to their association with atherosclerosis. As far as we know, no proteomic analyses have been performed on Lp(a). Since it is well known that elevated Lp(a) plasma levels are an important risk factor in atherogenesis, it would be of great interest to elucidate its apolipoprotein composition in relation to cardiovascular diseases. Another promising topic for future investigations is the characterization of proteomes of lipoproteins retained in atherosclerotic plaque. Finally, besides the great deal of work to be done in the future in both improving proteomic technologies and providing clues for the many aspects not yet investigated, it will also be necessary to put efforts on a comprehensive analysis of the huge quantity of data provided by the several proteomic studies.

#### **4. Acknowledgments**

This work has been supported by Fondazione Banco di Sardegna (Sassari, Italy) grant (Grant N° 536/2011.732).

#### **5. References**


Overview of Current Proteomic Approaches

pp. 277-90

No. 2, pp. 561-7

No. 10, pp. 1522-40

No.6, pp.870-6

*Res*, Vol.9, No. 9, pp. 4420-32

*Proteomics*, Vol.3, No.5, pp.714-23

*Atherosclerosis*, Vol.207, No. 1, pp. 32-7

we stand?, *J Proteomics*, Vol.73, No. 5, pp. 829-44

endothelial cell, *Life Sci*, Vol.80, No.26, pp.2469-80

spectrometry, *Proteome Sci*, Vol.8, No. 42, pp.1-9

for Discovery of Vascular Biomarkers of Atherosclerosis 25

Bondarenko, P.V.; Cockrill, S.L.; Watkins, L.K.; Cruzado, I.D. & Macfarlane, R.D. (1999).

Bondarenko, P.V.; Farwig, Z.N.; McNeal, C.J. & Macfarlane, R.D. (2002). MALDI- and ESI-

Boschetti, E.; Lomas, L.; Citterio, A. & Righetti, P.G. (2007). Romancing the "hidden

Böttcher, A.; Schlosser, J.; Kronenberg, F.; Dieplinger, H.; Knipping, G.; Lackner, K.J. &

Brea, D.; Sobrino, T.; Blanco, M.; Fraga, M.; Agulla, J.; Rodríguez-Yáñez, M.; Rodríguez-

Bruneel, A.; Labas, V.; Mailloux, A.; Sharma, S.; Vinh, J.; Vaubourdolle, M. & Baudin, B.

Burillo, E.; Recalde, D.; Jarauta, E.; Fiddyment, S.; Garcia-Otin, A.L.; Mateo-Gallego, R.;

Candiano, G.; Santucci, L.; Petretto, A.; Bruschi, M.; Dimuccio, V.; Urbani, A.; Bagnasco, S. &

Chen, C.Y.; Lee, C.M.; Hsu, H.C.; Yang, C.Y.; Chow, L.P. & Lee, Y.T. (2007). Proteomic

Collins, L.A. & Olivier M. (2010). Quantitative comparison of lipoprotein fractions derived

Conway, J.P. & Kinter, M. (2005). Proteomic and transcriptomic analyses of macrophages

Dardé, V.M.; de la Cuesta, F.; Dones, F.G.; Alvarez-Llamas, G.; Barderas, M.G. & Vivanco, F.

Davidson, W.S.; Silva, R.A.; Chantepie, S.; Lagor, W.R.; Chapman, M.J. & Kontush, A. (2009).

lipoprotein, *J Lipid Res*, Vol.40, No.3, pp.543-55

*Mass Spectrometry,* Vol.219, No.3, pp.671–80

subfractions, *J Lipid Res*. Vol.41, No.6, pp.905-15

Mass spectral study of polymorphism of the apolipoproteins of very low density

MS of the HDL apolipoproteins; new isoforms of apoA-I, II, *International Journal of* 

proteome", Anno Domini two zero zero seven, *J Chromatogr A*, Vol.1153, No. 1-2,

Schmitz, G. (2000). Preparative free-solution isotachophoresis for separation of human plasma lipoproteins: apolipoprotein and lipid composition of HDL

González, R.; Pérez de la Ossa, N.; Leira, R.; Forteza, J.; Dávalos, A. & Castillo, J. (2009). Usefulness of haptoglobin and serum amyloid A proteins as biomarkers for atherothrombotic ischemic stroke diagnosis confirmation, *Atherosclerosis*, Vol.205,

(2003). Proteomic study of human umbilical vein endothelial cells in culture,

Cenarro, A. & Civeira, F. (2009). Proteomic study of macrophages exposed to oxLDL identifies a CAPG polymorphism associated with carotid atherosclerosis,

Ghiggeri, G.M. (2010). 2D-electrophoresis and the urine proteome map: where do

approach to study the effects of various oxidatively modified low-density lipoprotein on regulation of protein expression in human umbilical vein

from human plasma and serum by liquid chromatography-tandem mass

with an increased resistance to oxidized low density lipoprotein (oxLDL)-induced cytotoxicity generated by chronic exposure to oxLDL, *Mol Cell Proteomics*, Vol.4,

(2010). Analysis of the plasma proteome associated with acute coronary syndrome: does a permanent protein signature exist in the plasma of ACS patients?, *J Proteome* 

Proteomic analysis of defined HDL subpopulations reveals particle-specific protein clusters: relevance to antioxidative function, *Arterioscler Thromb Vasc Biol*, Vol.29,

potential plasmatic markers to distinguish between ischemic and hemorrhagic stroke, *Proteomics*, Vol.4, No. 8, pp. 2242-51


Almofti, M.R.; Huang, Z.; Yang, P.; Rui, Y. & Yang, P. (2006). Proteomic analysis of rat aorta

Alonso-Orgaz, S.; Moreno, L.; Macaya, C.; Rico, L.; Mateos-Cáceres, P.J.; Sacristán, D.; Pérez-

Alwaili, K.; Bailey, D.; Awan, Z.; Bailey, S.D.; Ruel, I.; Hafiane, A.; Krimbou, L.; Laboissiere,

Bagnato, C.; Thumar, J.; Mayya, V.; Hwang, S.I.; Zebroski, H.; Claffey, K.P.; Haudenschild,

Bancells, C.; Canals, F.; Benítez, S.; Colomé, N.; Julve, J.; Ordóñez-Llanos, J. & Sánchez-

Banfi, C.; Brioschi, M.; Barcella, S.; Wait, R.; Begum, S.; Galli, S.; Rizzi, A. & Tremoli, E.

Banfi, C.; Parolari, A.; Brioschi, M.; Barcella, S.; Loardi, C.; Centenaro, C.; Alamanni, F.;

Barderas, M.G.; Tuñón, J.; Dardé, V.M.; De la Cuesta, F.; Durán, M.C.; Jiménez-Nácher, J.J.;

Barderas, M.G.; Tuñón, J.; Dardé, V.M.; De la Cuesta, F.; Jiménez-Nácher, J.J.; Tarín, N.;

Berhane, B.T.; Zong, C.; Liem, D.A.; Huang, A.; Le, S.; Edmondson, R.D.; Jones, R.C.; Qiao,

Bieler, S.; Meiners, S.; Stangl, V.; Pohl, T. & Stangl, K. (2009). Comprehensive proteomic and

stroke, *Proteomics*, Vol.4, No. 8, pp. 2242-51

pp. 2301-8

No. 6, pp. 1088-10

No.5, pp.1344-52

5, pp. 2347-57

No. 1, pp. 185-95

*J Lipid Res*, Vol. 51, No.12, pp.3508-15

profile, *J Proteome Res*, Vol.6, No. 2, pp. 876-86

phase, *Proteomics*, Vol.5, No. 13, pp. 3520-30

*Proteomics*, Vol.9, No. 7, pp. 1982-93

*Clin Exp Pharmacol Physiol*, Vol.33, No. 4, pp. 305-9

inflammatory profile, *Biochim Biophys Acta*, in press

potential plasmatic markers to distinguish between ischemic and hemorrhagic

during atherosclerosis induced by high cholesterol diet and injection of vitamin D3,

Vizcaíno, F.; Segura, A.; Tamargo, J. & López-Farré, A. (2006). Proteomic study of plasma from moderate hypercholesterolemic patients, *J Proteome Res,* Vol.5, No. 9,

S. & Genest, J. (2011). The HDL proteome in acute coronary syndromes shifts to an

C.; Eng, J.K.; Lundgren, D.H. & Han, D.K. (2007). Proteomics analysis of human coronary atherosclerotic plaque: a feasibility study of direct tissue proteomics by liquid chromatography and tandem mass spectrometry, *Mol Cell Proteomics*, Vol.6,

Quesada, J.L. (2010). Proteomic analysis of electronegative low-density lipoprotein,

(2009). Proteomic analysis of human low-density lipoprotein reveals the presence of prenylcysteine lyase, a hydrogen peroxide-generating enzyme, *Proteomics*. Vol.9,

Mussoni, L. & Tremoli, E. (2010). Proteomic analysis of plasma from patients undergoing coronary artery bypass grafting reveals a protease/antiprotease imbalance in favor of the serpin alpha1-antichymotrypsin, *J Proteome Res*, Vol.9, No.

Tarín, N.; López-Bescós, L.; Egido, J. & Vivanco, F. (2007). Circulating human monocytes in the acute coronary syndrome express a characteristic proteomic

López-Bescós, L.; Egido, J. & Vivanco, F. (2009). Atorvastatin modifies the protein profile of circulating human monocytes after an acute coronary syndrome,

X.; Whitelegge, J.P.; Ping, P. & Vondriska, T.M. (2005). Cardiovascular-related proteins identified in human plasma by the HUPO Plasma Proteome Project pilot

transcriptomic analysis reveals early induction of a protective anti-oxidative stress response by low-dose proteasome inhibition. *Proteomics*, Vol.9, No.12, pp.3257-67 Boccardi, C.; Cecchettini, A.; Caselli, A.; Camici, G.; Evangelista, M.; Mercatanti, A.;

Rainaldi, G. & Citti, L. (2007). A proteomic approach to the investigation of early events involved in vascular smooth muscle cell activation, *Cell Tissue Res*, Vol.328,


Overview of Current Proteomic Approaches

pp. 1767-80

No12, pp.1259-67

No.9, pp. 1345-53.

Vol. 340, No.3, pp.909-15

for Discovery of Vascular Biomarkers of Atherosclerosis 27

Emes, A.V.; Latner, A.L.; Rahbani-Nobar, M. & Tan, B.H. (1976). The separation of plasma

Fach, E.M.; Garulacan, L.A.; Gao, J.; Xiao, Q.; Storm, S.M.; Dubaquie, Y.P.; Hefta, S.A. &

Farwig, Z.N.; Campbell, A.V. & Macfarlane, R.D. (2003). Analysis of high-density

Formato, M.; Farina, M.; Spirito, R.; Maggioni, M.; Guarino, A.; Cherchi, G.M.; Biglioli, P.;

Formato, M.; Lepedda, A.J.; Zinellu, E.; Cigliano, A.; Piredda, F.; Bacciu, P.P.; Guarino, A. &

González-Cabrero, J.; Pozo, M.; Durán, M.C.; de Nicolás, R.; Egido, J. & Vivanco, F. (2007). The proteome of endothelial cells, *Methods Mol Biol*, Vol.357, pp.181-98 Gordon, S.M.; Deng, J.; Lu, L.J. & Davidson, W.S. (2010). Proteomic characterization of

Green, P.S.; Vaisar, T.; Pennathur, S.; Kulstad, J.J.; Moore, A.B.; Marcovina, S.; Brunzell, J.;

Ha, M.K.; Chung, K.Y.; Bang, D.; Park, Y.K. & Lee, K.H. (2005). Proteomic analysis of the

Havel, R. J.; Eder, H.A. & Bragdon, J.H. (1955). The distribution and chemical composition of

Heller, M.; Stalder, D.; Schlappritzi, E.; Hayn, G; Matter, U. & Haeberli, A. (2005). Mass

Heller, M.; Schlappritzi, E.; Stalder, D.; Nuoffer, J.M. & Haeberli, A. (2007) Compositional

Hortin, G.L.; Shen, R.F.; Martin, B.M. & Remaley, A.T. (2006). Diverse range of small

Huang, B.; Chen, S.C. & Wang, D.L. Shear flow increases S-nitrosylation of proteins in

microvascular endothelial cells, *Proteomics,* Vol.5, No.6, pp.1507-19

human plasma lipoproteins, *Proteomics*, Vol.5, No.10, pp. 2619-30

endothelial cells, (2009) *Cardiovasc Res*, Vol.83, No.3, pp.536-46

79th European Atherosclerosis Society Congress, Goteborg, June, 2011 Gadgil, H.S.; Pabst, K.M.; Giorgianni, F.; Umstot, E.S.; Desiderio, D.M.; Beranova-

electrophoresis, *Clin Chim Acta*, Vol.71, No.2, pp.293-301

*Mol Cell Proteomics*, Vol.3, No. 12, pp. 1200-10

spectrometry, *Anal. Chem,* Vol.75, pp.3823-30

*Arterioscler Thromb Vasc Biol*, Vol.24, No. 1, pp. 129-35

chromatography, *J Proteome Res*, Vol.9, No.10, pp.5239-49

lipoproteins using gel electrofocusing and polyacrylamide gradient gel

Opiteck, G.J. (2004). In vitro biomarker discovery for atherosclerosis by proteomics,

lipoprotein apolipoproteins recovered from specific immobilized pH gradient gel pI domains by matrix-assisted laser desorption/ionization time-of-flight mass

Edelstein, C. & Scanu, A.M. (2004). Evidence for a proinflammatory and proteolytic environment in plaques from endarterectomy segments of human carotid arteries,

Spirito R. (2011). Apolipoprotein profiles in atherosclerotic disease, Proceedings of

Giorgianni, S.; Gerling, I.C. & Pabst, M.J. (2003). Proteome of monocytes primed with lipopolysaccharide: analysis of the abundant proteins, *Proteomics*, Vol.3, No. 9,

human plasma high density lipoprotein fractionated by gel filtration

Knopp, R.H.; Zhao, X.Q. & Heinecke, J.W. (2008) Combined statin and niacin therapy remodels the high-density lipoprotein proteome, *Circulation*, Vol.118,

proteins expressed by hydrogen peroxide treated cultured human dermal

ultracentrifugally separated lipoproteins in human serum, *J. Clin. Invest*, Vol.34,

spectrometry-based analytical tools for the molecular protein characterization of

protein analysis of high density lipoproteins in hypercholesterolemia by shotgun LC-MS/MS and probabilistic peptide scoring, *Mol Cell Proteomics*, Vol.6, No.6, pp.1059-72

peptides associated with high-density lipoprotein, *Biochem Biophys Res Commun*,


Davidsson, P.; Hulthe, J.; Fagerberg, B.; Olsson, B.M.; Hallberg, C.; Dahllöf, B. & Camejo, G.

de Kleijn, D.P.V.; Moll, F.L.; Hellings, W.E.; Ozsarlak-Sozer, G.; de Bruin, P.; Doevendans,

Decramer, S.; Gonzalez de Peredo, A.; Breuil, B.; Mischak, H.; Monsarrat, B.; Bascands, J.L. &

Dejouvencel, T.; Féron, D.; Rossignol, P.; Sapoval, M.; Kauffmann, C.; Piot, J.M.; Michel, J.B.;

Delles, C.; Schiffer, E.; von Zur Muhlen, C.; Peter, K.; Rossing, P.; Parving, H.H.; Dymott,

Distelmaier, K.; Adlbrecht, C.; Jakowitsch, J.; Winkler, S.; Dunkler, D.; Gerner, C.; Wagner,

Donahue, M.P.; Rose, K.; Hochstrasser, D.; Vonderscher, J.; Grass, P.; Chibout, S.D.; Nelson,

Donners, M.M.; Verluyten, M.J.; Bouwman, F.G.; Mariman, E.C.; Devreese, B.; Vanrobaeys,

Dupont, A.; Tokarski, C.; Dekeyzer, O.; Guihot, A.L.; Amouyel, P.; Rolando, C. & Pinet, F.

Dupont, A.; Corseaux, D.; Dekeyzer, O.; Drobecq, H.; Guihot, A.L.; Susen, S.; Vincentelli, A.;

Duran, M.C.; Mas, S.; Martin-Ventura, J.L.; Meilhac, O.; Michel, J.B.; Gallego-Delgado, J.;

analysis of pooled plasma, *Am Heart J*, Vol.152, No. 3, pp. 478-85

arterial smooth muscle cells, *Proteomics*, Vol.5, No. 2, pp. 585-96

infarction, *Thromb Haemost*, Vol.102, No. 3, pp. 564-72

plaque progression, *J Pathol*, Vol.206, No.1, pp. 39-45

and secretome, *Proteomics*, Vol.4, No. 6, pp. 1761-78

plaques, *Eur J Pharmacol*, Vol.562, No. 1-2, pp. 119-29

*Thromb Vasc Biol*, Vol.30, pp. 612-19

10, pp. 1850-62

2, pp. 269-75

Vol.28, No. 11, pp. 2316-22

(2005) A proteomic study of the apolipoproteins in LDL subclasses in patients with the metabolic syndrome and type 2 diabetes, *J Lipid Res*, Vol.46, No.9, pp.1999-2006

P.A.; Vink, A.; Catanzariti, L.M.; Schoneveld, A.H.; Algra, A.; Daemen, M.J.; Biessen, E.A.; de Jager, W.; Zhang, H.; de Vries, J.; Falk, E.; Lim, S.K.; van der Spek, P.J.; Kwan Sze, S. & Pasterkamp, G. (2010). Local atherosclerotic plaques are a source of prognostic biomarkers for adverse cardiovascular events, *Arterioscler* 

Schanstra, J.P. (2008). Urine in clinical proteomics, *Mol Cell Proteomics*, Vol.7, No.

Fruitier-Arnaudin, I. & Meilhac, O. (2010). Hemorphin 7 reflects hemoglobin proteolysis in abdominal aortic aneurysm, *Arterioscler Thromb Vasc Biol*, Vol.30, No.

J.A.; Neisius, U.; Zimmerli, L.U.; Snell-Bergeon, J.K.; Maahs, D.M.; Schmieder, R.E.; Mischak, H. & Dominiczak, A.F. (2010). Urinary proteomic diagnosis of coronary artery disease: identification and clinical validation in 623 individuals, *J Hypertens*,

O.; Lang, I.M. & Kubicek, M. (2009). Local complement activation triggers neutrophil recruitment to the site of thrombus formation in acute myocardial

C.L.; Sinnaeve, P.; Goldschmidt-Clermont, P.J. & Granger, C.B. (2006). Discovery of proteins related to coronary artery disease using industrial-scale proteomics

F.; van Beeumen, J.; van den Akker, L.H.; Daemen, M.J. & Heeneman, S. (2005). Proteomic analysis of differential protein expression in human atherosclerotic

(2004). Two-dimensional maps and databases of the human macrophage proteome

Amouyel, P.; Jude, B. & Pinet, F. (2005). The proteome and secretome of human

Lázaro, A.; Tuñon, J.; Egido, J. & Vivanco, F. (2003). Proteomic analysis of human vessels: application to atherosclerotic plaques, *Proteomics*, Vol.3, No.6, pp. 973-8 Durán,M.C.; Martín-Ventura, J.L.; Mohammed, S.; Barderas, M.G.; Blanco-Colio, L.M.; Mas,

S.; Moral, V.; Ortega, L.; Tuñón, J.; Jensen, O.N.; Vivanco, F. & Egido, J. (2007). Atorvastatin modulates the profile of proteins released by human atherosclerotic


Overview of Current Proteomic Approaches

pp.143-54

Vol.94, No. 10, pp. e87-96

*Proteomics*, Vol.1, No. 11, pp. 1405-14

No. 2, pp. 115-24

No.1, pp.578-84

No. 4, pp. 574-80

for Discovery of Vascular Biomarkers of Atherosclerosis 29

ischemic patients: a proteomic study, *J Proteome Res*, Vol.6, No. 7, pp. 2481-7 Lutgens, E.; van Suylen, R.J.; Faber, B.C.; Gijbels, M.J.; Eurlings, P.M.; Bijnens, A.P.; Cleutjens,

systemic process?, *Arterioscler Thromb Vasc Biol*, Vol.23, No.12, pp. 2123-30 Mancone, C.; Amicone, L.; Fimia, G.M.; Bravo, E.; Piacentini, M.; Tripodi, M. & Alonzi, T.

Martinet, W.; Schrijvers, D.M.; De Meyer, G.R.; Herman, A.G. & Kockx, M.M. (2003).

Martin-Ventura, J.L.; Duran, M.C.; Blanco-Colio, L.M.; Meilhac, O.; Leclercq, A.; Michel, J.B.;

potential marker of atherosclerosis, *Circulation*, Vol.110, No.15, pp. 2216-9 Mateos-Cáceres, P.J.; García-Méndez, A.; López Farré, A.; Macaya, C.; Núñez, A.; Gómez, J.;

Mayr, M.; Siow, R.; Chung, Y.L.; Mayr, U.; Griffiths, J.R. & Xu, Q. (2004). Proteomic and

Mayr, M.; Chung, Y.L.; Mayr, U.; Yin, X.; Ly, L.; Troy, H.; Fredericks, S.; Hu, Y.; Griffiths, J.R. &

McGregor, E.; Kempster, L.; Wait, R.; Welson, S.Y.; Gosling, M.; Dunn, M.J. & Powel, J.T.

McGregor, E.; Kempster, L.; Wait, R.; Gosling, M.; Dunn, M.J. & Powell, J.T. (2004). F-actin

Millioni, R.; Puricelli, L.; Iori, E.; Arrigoni, G. & Tessari, P. (2010). The effects of rosiglitazone

Ogorzalek Loo, R.R.; Yam, L.; Loo, J.A. & Schumaker, V.N. (2004). Virtual two-dimensional gel

Olson, F.J.; Sihlbom, C.; Davidsson, P.; Hulthe, J.; Fagerberg, B. & Bergström, G. (2010).

apoptosis-linked gene 2, *Cardiovasc Res*, Vol.60, No. 2, pp. 259-67

coronary syndrome, *J Am Coll Cardiol*, Vol.44, No. 8, pp. 1578-83

Relationship between vitamin D binding protein and aspirin resistance in coronary

K.B.; Heeneman, S. & Daemen, M.J. (2003). Atherosclerotic plaque rupture: local or

(2007). Proteomic analysis of human very low-density lipoprotein by twodimensional gel electrophoresis and MALDI-TOF/TOF, *Proteomics*, Vol.7, No.1,

Western array analysis of human atherosclerotic plaques: downregulation of

Jensen, O.N.; Hernandez-Merida, S.; Tuñón, J.; Vivanco, F. & Egido, J. (2004). Identification by a differential proteomic approach of heat shock protein 27 as a

Alonso-Orgaz, S.; Carrasco, C.; Burgos, M.E.; de Andrés, R.; Granizo, J.J.; Farré, J. & Rico, L.A. (2004). Proteomic analysis of plasma from patients during an acute

metabolomic analysis of vascular smooth muscle cells: role of PKCdelta, *Circ Res*,

Xu, Q. (2005). Proteomic and metabolomic analyses of atherosclerotic vessels from apolipoprotein E-deficient mice reveal alterations in inflammation, oxidative stress, and energy metabolism, *Arterioscler Thromb Vasc Biol*, Vol.25, No. 10, pp. 2135-42 Mazur, M.T.; Cardasis, H.L.; Spellman, D.S.; Liaw, A.; Yates, N.A. & Hendrickson, R.C.

(2010). Quantitative analysis of intact apolipoproteins in human HDL by top-down differential mass spectrometry. *Proc Natl Acad Sci U S A.* Vol.107, No.17, pp.7728-33

(2001). Identification and mapping of human saphenous vein medial smooth muscle proteins by two-dimensional polyacrylamide gel electrophoresis,

capping (CapZ) and other contractile saphenous vein smooth muscle proteins are altered by hemodynamic stress: a proteonomic approach, *Mol Cell Proteomics*, Vol.3,

and high glucose on protein expression in endothelial cells, *J Proteome Res*, Vol.9,

electrophoresis of high-density lipoproteins, *Electrophoresis.* Vol. 25, No.14, pp.2384-91

Consistent differences in protein distribution along the longitudinal axis in symptomatic carotid atherosclerotic plaques, *Biochem Biophys Res Commun*, Vol.401,


Jang, W.G.; Kim, H.S.; Park, K.G.; Park, Y.B.; Yoon, K.H.; Han, S.W.; Hur, S.H.; Park, K.S. &

Kamino, H.; Hiratsuka, M.; Toda, T.; Nishigaki, R.; Osaki, M.; Ito, H.; Inoue, T. & Oshimura,

Kang, J.H.; Ryu, H.S.; Kim, H.T.; Lee, S.J.; Choi, U.K.; Park, Y.B.; Huh, T.L.; Choi, M.S.; Kang,

Karlsson, H.; Leanderson, P.; Tagesson, C. & Lindahl, M. (2005a). Lipoproteomics I:

electrophoresis and mass spectrometry, *Proteomics*, Vol.5, No.2, pp. 551-65 Karlsson, H.; Leanderson, P.; Tagesson, C. & Lindahl, M. (2005b). Lipoproteomics II:

electrophoresis and mass spectrometry, *Proteomics*. Vol.5, No.5, pp.1431-45 Karlsson, H.; Lindqvist, H.; Tagesson, C. & Lindahl, M. (2006). Characterization of

Khovidhunkit, W.; Duchateau, P.N.; Medzihradszky, K.F.; Moser, A.H.; Naya-Vigne, J.;

Kiga, C.; Sakurai, H.; Goto, H.; Hayashi, K.; Shimada, Y. & Saiki, I. (2008). Proteomic

Lepedda, A.J.; Cigliano, A.; Cherchi, G.M.; Spirito, R.; Maggioni, M.; Carta, F.; Turrini, F.;

Levels, J.H.; Bleijlevens, B.; Rezaee, F.; Aerts, J.M. & Meijers, J.C. (2007). SELDI-TOF mass spectrometry of High-Density Lipoprotein, *Proteome Sci*, Vol. 6, pp.5-15 Levels, J.H.; Geurts, P.; Karlsson, H; Marée, R.; Ljunggren, S.; Fornander, L.; Wehenkel, L.;

Lomnytska, M.; Lukiyanchuk, V.; Hellman, U. & Souchelnytskyi, S. (2004). Transforming

López-Farré, A.J.; Mateos-Cáceres, P.J.; Sacristán, D.; Azcona, L.; Bernardo, E.; de Prada, T.P.;

Libby P. (2002). Inflammation in atherosclerosis, *Nature*, Vol.420, No.6917, pp. 868-74

hypertensive stroke-prone rats, *Life Sci*, Vol.83, No. 17-18, pp. 625-31 Lee, C.K.; Park H.J.; So, H.H.; Kim, H.J.; Lee, K.S.; Choi, W.S.; Lee, H.M.; Won, K.J.; Yoon,

carotid arteries, *Atherosclerosis*, Vol.203, No. 1, pp. 112-8

*Proteomics*, Vol.4, No. 11, pp. 3383-93

Vol.1794, No. 3, pp. 446-58

pp.2685-90

No.1, pp. 37-44

pp. 6455-75

No.4, pp.995-1006

*Cell Struct Funct,* Vol.28, No.6, pp.495-503

Lee, I.K. (2004). Analysis of proteome and transcriptome of tumor necrosis factor alpha stimulated vascular smooth muscle cells with or without alpha lipoic acid,

M. (2003). Searching for genes involved in arteriosclerosis: proteomic analysis of cultured human umbilical vein endothelial cells undergoing replicative senescence,

T.C.; Choi, S.Y. & Kwon, O.S. (2009). Proteomic analysis of human macrophages exposed to hypochlorite-oxidized low-density lipoprotein, *Biochim Biophys Acta*,

mapping of proteins in low-density lipoprotein using two-dimensional gel

mapping of proteins in high-density lipoprotein using two-dimensional gel

apolipoprotein M isoforms in low-density lipoprotein, *J Proteome Res*, Vol.5, No.10,

Shigenaga, J.K.; Kane, J.P.; Grunfeld, C. & Feingold, K.R. (2004). Apolipoproteins A-IV and A-V are acute-phase proteins in mouse HDL, *Atherosclerosis*, Vol. 176,

identification of haptoglobin as a stroke plasma biomarker in spontaneously

T.J.; Park, T.K. & Kim, B. (2006). Proteomic profiling and identification of cofilin responding to oxidative stress in vascular smooth muscle, *Proteomics*, Vol.6, No. 24,

Edelstein, C.; Scanu, A.M. & Formato, M. (2009). A proteomic approach to differentiate histologically classified stable and unstable plaques from human

Lindahl, M.; Stroes, E.S.; Kuivenhoven, J.A. & Meijers, J.C. (2011). High-density lipoprotein proteome dynamics in human endotoxemia, *Proteome Sci,* Vol.9, No.34

growth factor-beta1-regulated proteins in human endothelial cells identified by two-dimensional gel electrophoresis and mass spectrometry, *Proteomics*, Vol.4,

Alonso-Orgaz, S.; Fernández-Arquero, M.; Fernández-Ortiz, A. & Macaya, C. (2007).

Relationship between vitamin D binding protein and aspirin resistance in coronary ischemic patients: a proteomic study, *J Proteome Res*, Vol.6, No. 7, pp. 2481-7


Overview of Current Proteomic Approaches

pp. 637-42

10, pp. 3881-90

for Discovery of Vascular Biomarkers of Atherosclerosis 31

Sukhanov, S. & Delafontaine, P. (2005). Protein chip-based microarray profiling of oxidized low density lipoprotein-treated cells, *Proteomics*, Vol.5, No. 5, pp. 1274-80 Sun, H.Y.; Chen, S.F.; Lai, M.D.; Chang, T.T.; Chen, T.L.; Li, P.Y.; Shieh, D.B. & Young, K.C.

Sung, H.J.; Ryang, Y.S.; Jang, S.W.; Lee, C.W.; Han, K.H. & Ko, J. (2006). Proteomic analysis of differential protein expression in atherosclerosis, *Biomarkers*, Vol.11, No.3, pp. 279-90 Tabibiazar, R.; Wagner, R.A.; Deng, A.; Tsao, P.S. & Quertermous, T. (2006). Proteomic

Terzuoli, L.; Felici, C.; Ciari, I.; Guerranti, R.; Pagani, R.; Marinello, E. & Porcelli, B. (2007).

Thongboonkerd V. (2007). Practical points in urinary proteomics, *J Proteome Res*, Vol.6, No.

Trott, D.; McManus, C.A.; Martin, J.L.; Brennan, B.; Dunn, M.J. & Rose, M.L. (2009). Effect of

Tunica, D.G.; Yin, X.; Sidibe, A.; Stegemann, C.; Nissum, M.; Zeng, L.; Brunet, M. & Mayr,

Vaisar, T.; Pennathur, S.; Green, P.S.; Gharib, S.A.; Hoofnagle, A.N.; Cheung, M.C.; Byun, J.;

Vance, D.E. & Vance J.E. (2008). Biochemistry of lipids, lipoprotein and membranes (Fifth

Virmani, R.; Kolodgie, F.D.; Burke, A.P.; Farb, A. & Schwartz, S.M. (2000). Lessons from

Walldius, G.; Jungner, I.; Holme, I.; Aastveit, A.H.; Kolar, W. & Steiner, E. (2001). High

Wang, T.; Chen, Z.; Wang, X.; Shyy, J.Y. & Zhu Y. (2006). Cholesterol loading increases the

Wang, X.L.; Fu, A.; Raghavakaimal, S. & Lee, H.C. (2007). Proteomic analysis of vascular

endothelial cells, *Biochim Biophys Acta*, Vol.1761, No.10, pp.1182-90

density lipoproteins, *Clin Chim Acta,* Vol*.*411, No.5-6, pp.336-44

*Physiol Genomics*, Vol.25, No. 2, pp. 194-202

cell, *Proteomics*, Vol.9, No.12, pp.3383-94

*Proteomics*, Vol.9, No.21, pp.4991-6

Vol.411, No.13-14, pp.972-9

No. 9298, pp. 2026-33

Edition), Elsevier B.V. ISBN: 978-0-444-53219-0

(2010). Comparative proteomic profiling of plasma very-low-density and low-

profiles of serum inflammatory markers accurately predict atherosclerosis in mice,

Synthetic gel of carotid artery plaque, *Int J Immunopathol Pharmacol*, Vol.20, No. 3,

phosphorylated hsp27 on proliferation of human endothelial and smooth muscle

M. (2009) Proteomic analysis of the secretome of human umbilical vein endothelial cells using a combination of free-flow electrophoresis and nanoflow LC-MS/MS,

Vuletic, S.; Kassim, S.; Singh, P.; Chea, H.; Knopp, R.H.; Brunzell, J.; Geary, R.; Chait, A.; Zhao, X.Q.; Elkon, K.; Marcovina, S.; Ridker, P.; Oram, J.F. & Heinecke, J.W. (2007). Shotgun proteomics implicates protease inhibition and complement activation in the antiinflammatory properties of HDL, *J Clin Invest*, Vol.117, No.3, pp.746-56 Vaisar, T.; Mayer, P.; Nilsson, E.; Zhao, X.Q.; Knopp, R. & Prazen, B.J. (2010) HDL in

humans with cardiovascular disease exhibits a proteomic signature, *Clin Chim Acta.*

sudden coronary death: a comprehensive morphological classification scheme for atherosclerotic lesions, *Arterioscler Thromb Vasc Biol*, Vol.20, No.5, pp. 1262-75 von Zur Muhlen, C.; Schiffer, E.; Zuerbig, P.; Kellmann, M.; Brasse, M.; Meert, N.; Vanholder,

R.C.; Dominiczak, A.F.; Chen, Y.C.; Mischak, H.; Bode, C. & Peter, K. (2009). Evaluation of urine proteome pattern analysis for its potential to reflect coronary artery atherosclerosis in symptomatic patients, *J Proteome Res*, Vol.8, No. 1, pp. 335-45

apolipoprotein B, low apolipoprotein A-I, and improvement in the prediction of fatal myocardial infarction (AMORIS study): a prospective study, *Lancet*, Vol.358,

translocation of ATP synthase beta chain into membrane caveolae in vascular

endothelial cells in response to laminar shear stress, *Proteomics*, Vol.7, No.4, pp.588-96


Padró, T.; Peña, E.; García-Arguinzonis, M.; Llorente-Cortes, V. & Badimon, L. (2008). Low-

Pawlowska, Z.; Baranska, P.; Jerczynska, H.; Koziolkiewicz, W. & Cierniewski, C.S. (2005).

Porcelli, B.; Ciari, I.; Felici, C.; Pagani, R.; Banfi, C.; Brioschi, M.; Giubbolini, M.; de Donato,

Prentice, R.L.; Paczesny, S.; Aragaki, A.; Amon, L.M.; Chen, L.; Pitteri, S.J.; McIntosh, M.;

Rashid, K.A.; Hevi, S.; Chen, Y.; Le Cahérec, F. & Chuck, S.L. (2002). A proteomic approach

Rezaee, F.; Casetta, B.; Levels, J.H.; Speijer, D. & Meijers, J.C. (2006). Proteomic analysis of

Richardson, M.R.; Lai, X.; Dixon, J.L.; Sturek, M. & Witzmann, F.A. (2009). Diabetic

Righetti, P.G.; Castagna, A.; Antonucci, F.; Piubelli, C.; Cecconi, D.; Campostrini, N.;

Sintiprungrat, K.; Singhto, N.; Sinchaikul, S.; Chen, S.T. & Thongboonkerd, V. (2010).

Slevin, M.; Elasbali, A.B.; Miguel Turu, M.; Krupinski, J.; Badimon, L. & Gaffney, J. (2006).

Stary, HC. (2000). Natural history and histological classification of atherosclerotic lesions: an

unstable human carotid plaques, *Am J Pathol*, Vol.168, No. 3, pp. 1004-21 Slomianny M.C.; Dupont A.; Bouanou F.; Beseme O.; Guihot A.L.; Amouyel P.; Michalski

comparison of two approaches, *Proteomics*, Vol.6, No. 8, pp. 2365-75 Ståhlman, M.; Davidsson, P.; Kanmert, I.; Rosengren, B.; Borén, J.; Fagerberg, B. & Camejo,

in D2O/sucrose or in KBr, *J Lipid Res*, Vol. 49, No.2, pp.481-90

update, *Arterioscler Thromb Vasc Biol*, Vol.20, No.5, pp. 1177-8

high-density lipoprotein, *Proteomics*. Vol.6, No.2, pp.721-30

in-depth plasma proteome profiling, *Genome Med*, Vol.2, No. 7, pp. 48 Qi, Y.X.; Qu, M.J.; Long, D.K.; Liu, B.; Yao, Q.P.; Chien, S. & Jiang, Z.L. (2008). Rho-GDP

human endothelial cells, *Proteomics*, Vol.5, No.5, pp.1217-27

*Biomed Pharmacother*, Vol.64, No. 5, pp. 369-72

*Res*, Vol.77, No. 1, pp. 211-20

Vol.80, No. 1, pp. 114-22

Vol. 277, No.24, pp.22010-7

pigs, *Proteomics*, Vol.9, No.9, pp.2468-83

*Chim Acta*, Vol.357, No. 2, pp. 123-39

processes, *J Proteomics*, Vol.73, No. 3, pp. 602-18

density lipoproteins impair migration of human coronary vascular smooth muscle cells and induce changes in the proteomic profile of myosin light chain, *Cardiovasc* 

Heat shock proteins and other components of cellular machinery for protein synthesis are up-regulated in vascular endothelial cell growth factor-activated

G.; Setacci, C. & Terzuoli, L. (2010). Proteomic analysis of atherosclerotic plaque,

Wang, P.; Buson Busald, T.; Hsia, J.; Jackson, R.D.; Rossouw, J.E.; Manson, J.E.; Johnson, K.; Eaton, C. & Hanash, S.M. (2010). Novel proteins associated with risk for coronary heart disease or stroke among postmenopausal women identified by

dissociation inhibitor alpha downregulated by low shear stress promotes vascular smooth muscle cell migration and apoptosis: a proteomic analysis, *Cardiovasc Res*,

identifies proteins in hepatocytes that bind nascent apolipoprotein B, *J Biol Chem*,

dyslipidemia and exercise alter the plasma low-density lipoproteome in Yucatan

Rustichelli, C.; Antonioli, P.; Zanusso, G.; Monaco, S.; Lomas, L. & Boschetti, E. (2005). Proteome analysis in the clinical chemistry laboratory: myth or reality?, *Clin* 

Alterations in cellular proteome and secretome upon differentiation from monocyte to macrophage by treatment with phorbol myristate acetate: insights into biological

Identification of differential protein expression associated with development of

J.C. & Pinet F. (2006). Profiling of membrane proteins from human macrophages:

G. (2008). Proteomics and lipids of lipoproteins isolated at low salt concentrations


**2** 

**From Biomarker Discovery to Clinical** 

Lung cancer is one of the most common cancers in the world (Chiang et al., 2010; Landis et al., 1998). Surgical removal of the tumor mass offers the best chance for a cure in patients with non-small-cell lung cancer. A tumor in stages I (confined to the lung without nodal or distant metastasis), II (involvement of only lymph nodes within the lung), and IIIA (involvement of nodes on the same side as the tumor) is considered potentially resectable

Based on tumor size and location, lung surgery is mainly divided into three types: wedge resection (removal of a small area in one lobe of either right or left lung), lobectomy (removal of one lobe from a right or left lung), and pneumonectomy (removal of an entire right or left lung). The mortality rate is much higher after pneumonectomy (61%) than lobectomy (35%) (Gunluoglu et al., 2011). Among the post-surgical factors, aberrant local inflammation and abnormal fluid drainage are the most common for inducing pulmonary edema. Excessive accumulation of fluid in the alveoli causes lung injury and hinders functional recovery. A severe form of acute lung injury results in acute respiratory distress syndrome. Sudden and life-threatening lung failure is the most detrimental factor in post-

With the progression from lung injury to acute respiratory distress syndrome, proinflammatory cytokines are increased, such as interleukin-1β (Donnelly et al., 1996; Geiser et al., 2001) and tumor necrosis factor-α (Tremblay et al., 2002). However, the production of vascular endothelial growth factor is reduced in the early stage but not altered in the late stage (Medfor & Millar, 2006). Interleukin-1β and tumor necrosis factorα are further elevated in the sustained phase (Bhatia & Moochhala, 2004). The increases in proinflammatory cytokines are not correlated with injury-induced mortality (Donnelly et al., 1996). Corticosteroid which alters host inflammatory responses do not show beneficial effects in the early stage of acute respiratory distress syndrome (Kollef et al., 1995). Current reviews suggest that activation of inflammation-independent pathways in the early stage and inflammation-dependent pathways in the late stage contribute to the development of acute respiratory distress syndrome (Spragg et al., 2010; Bhatia &

**1. Introduction** 

for cure (Martini et al., 1995).

surgical mortality (Jordan et al., 2000).

Moochhala, 2004).

**Evaluation for Early Diagnosis of** 

**Lung Surgery-Induced Injury** 

Mei-Ling Tsai, Shu-Hui Chen,

*National Cheng Kung University,* 

*Taiwan, Republic of China* 

Chih-Ching Chang and Ming-Ho Wu


### **From Biomarker Discovery to Clinical Evaluation for Early Diagnosis of Lung Surgery-Induced Injury**

Mei-Ling Tsai, Shu-Hui Chen, Chih-Ching Chang and Ming-Ho Wu *National Cheng Kung University, Taiwan, Republic of China* 

#### **1. Introduction**

32 Proteomics – Human Diseases and Protein Functions

Wang, Y.; Zhang, B.; Bai, Y.; Zeng, C. & Wang, X. (2010). Changes in proteomic features

Won, K.J.; Lee, P.; Jung, S.H.; Jiang, X.; Lee, C.K.; Lin, H.Y.; Kang, H.; Lee, H.M.; Kim, J.;

You, S.A.; Archacki, S.R.; Angheloiu, G.; Moravec, C.S.; Rao, S.; Kinter, M.; Topol, E.J. &

Yu, M.; Chen, D.M.; Hu, G. & Wang, H. (2004). Proteomic response analysis of endothelial

Yu, Y.L.; Huang, Z.Y.; Yang, P.Y.; Rui, Y.C. & Yang, P.Y. (2003a). Proteomic studies of

Yu, Y.L.; Yang, P.Y.; Fan, H.Z.; Huang, Z.Y.; Rui, Y.C. & Yang, P.Y. (2003b). Protein

dimensional gel electrophoresis, *Acta Pharmacol Sin*, Vol.24, No. 9, pp. 873-7 Zannis, V.I. (1986). Genetic polymorphism in human apolipoprotein E, *Methods Enzymol*,

Zhang, L.; Lun, Y.; Yan, D.; Yu, L.; Ma, W.; Du, B. & Zhu, X. (2007). Proteomic analysis of

Zhao, C.; Zhang, H.; Wong, W.C.; Sem, X.; Han, H.; Ong, S.M.; Tan, Y.C.; Yeap, W.H.; Gan,

and transcriptomic methods, *J Proteome Res*, Vol.8, No. 8, pp. 4028-38 Zimman, A.; Chen, S.S.; Komisopoulou, E.; Titz, B.; Martínez-Pinna, R.; Kafi, A.; Berliner, J.A.

smooth muscle cell migration, *Proteomics*, Vol.11, No. 2, pp. 193-201 Wu, J.; Liu, W.; Sousa, E.; Qiu, Y.; Pittman, D.D.; Maganti, V.; Feldman, J.; Gill, D.; Lu, Z.;

atherosclerosis, *Physiol Genomics*, Vol.13, No.1, pp. 25-30

hypertensive rats in vitro, *Cell Biochem Biophys*, Vol.58, No. 2, pp. 97-106 Williams, K.J. & Tabas, I. (1995). The response-to-retention hypothesis of early atherogenesis, *Arterioscler Thromb Vasc Biol*, Vol.15, No.5, pp. 551-61 Wilson, A.M.; Kimura,E.; Harada, R.K.; Nair, N.; Narasimhan, B.; Meng, X.Y.; Zhang, F.;

*Circulation*, Vol.116, No. 12, pp. 1396-403

*Proteome Res*, Vol.6, No. 12, pp. 4728-36

Vol.25, No.9, pp.1124-30

No. 6, pp. 782-9

Vol.128, pp.823-51

Vol.321, No. 1-2, pp. 80-5

*Mol Cell Proteomics*, Vol.7, No. 2, pp. 290-8

induced by insulin on vascular smooth muscle cells from spontaneous

Beck, K.R.; Olin, J.W.; Fung, E.T. & Cooke, J.P. (2007). Beta2-microglobulin as a biomarker in peripheral arterial disease: proteomic profiling and clinical studies,

Toyokuni, S. & Kim, B. (2011). 3-morpholinosydnonimine participates in the attenuation of neointima formation via inhibition of annexin A2-mediated vascular

Dorner, A.J.; Schaub, R. & Tan, X.Y. (2007). Proteomic identification of endothelial proteins isolated in situ from atherosclerotic aorta via systemic perfusion, *J* 

Wang, Q. (2003). Proteomic approach to coronary atherosclerosis shows ferritin light chain as a significant marker: evidence consistent with iron hypothesis in

cells of human coronary artery to stimulation with carbachol, *Acta Pharmacol Sin*,

macrophage-derived foam cell from human U937 cell line using two-dimensional gel electrophoresis and tandem mass spectrometry, *J Cardiovasc Pharmacol*, Vol.42,

expressions in macrophage-derived foam cells: comparative analysis by two-

macrophages: a new way to identify novel cell-surface antigens, *J Immunol Methods*,

C.S.; Ng, K.Q.; Koh, M.B.; Kourilsky, P.; Sze, S.K. & Wong, S.C. (2009). Identification of novel functional differences in monocyte subsets using proteomic

& Graeber, T.G. (2010). Activation of aortic endothelial cells by oxidized phospholipids: a phosphoproteomic analysis, *J Proteome Res*, Vol.9, No.6, pp.2812-24 Zimmerli, L.U.; Schiffer, E.; Zürbig, P.; Good, D.M.; Kellmann, M.; Mouls, L.; Pitt, A.R.;

Coon, J.J.; Schmieder, R.E.; Peter, K.H.; Mischak, H.; Kolch, W.; Delles, C. & Dominiczak, A.F. (2008). Urinary proteomic biomarkers in coronary artery disease, Lung cancer is one of the most common cancers in the world (Chiang et al., 2010; Landis et al., 1998). Surgical removal of the tumor mass offers the best chance for a cure in patients with non-small-cell lung cancer. A tumor in stages I (confined to the lung without nodal or distant metastasis), II (involvement of only lymph nodes within the lung), and IIIA (involvement of nodes on the same side as the tumor) is considered potentially resectable for cure (Martini et al., 1995).

Based on tumor size and location, lung surgery is mainly divided into three types: wedge resection (removal of a small area in one lobe of either right or left lung), lobectomy (removal of one lobe from a right or left lung), and pneumonectomy (removal of an entire right or left lung). The mortality rate is much higher after pneumonectomy (61%) than lobectomy (35%) (Gunluoglu et al., 2011). Among the post-surgical factors, aberrant local inflammation and abnormal fluid drainage are the most common for inducing pulmonary edema. Excessive accumulation of fluid in the alveoli causes lung injury and hinders functional recovery. A severe form of acute lung injury results in acute respiratory distress syndrome. Sudden and life-threatening lung failure is the most detrimental factor in postsurgical mortality (Jordan et al., 2000).

With the progression from lung injury to acute respiratory distress syndrome, proinflammatory cytokines are increased, such as interleukin-1β (Donnelly et al., 1996; Geiser et al., 2001) and tumor necrosis factor-α (Tremblay et al., 2002). However, the production of vascular endothelial growth factor is reduced in the early stage but not altered in the late stage (Medfor & Millar, 2006). Interleukin-1β and tumor necrosis factorα are further elevated in the sustained phase (Bhatia & Moochhala, 2004). The increases in proinflammatory cytokines are not correlated with injury-induced mortality (Donnelly et al., 1996). Corticosteroid which alters host inflammatory responses do not show beneficial effects in the early stage of acute respiratory distress syndrome (Kollef et al., 1995). Current reviews suggest that activation of inflammation-independent pathways in the early stage and inflammation-dependent pathways in the late stage contribute to the development of acute respiratory distress syndrome (Spragg et al., 2010; Bhatia & Moochhala, 2004).

From Biomarker Discovery to

market.

legal guardians .

found in all samples studied.

Clinical Evaluation for Early Diagnosis of Lung Surgery-Induced Injury 35

with molecular weight greater than 100 kDa (such as α2-macroglobulin). The discovery of hundreds of proteins in bronchoalveolar lavage fluid demonstrates its feasibility in

To accelerate the translation of biomarker discovery from bench to bedside, the development of techniques has been divided into 5 stages (Pepe et al., 2001). In Phase 1, potential biomarkers are discovered by various approaches, such as proteomic analysis. After the leads are identified by biochemical studies, measurable classifiers or outcomes are developed in Phase 2. Based on the analysis of their specificity and sensitivity, the cutoff point of the measurable outcome is determined and used in Phase 3. Based on patient history and clinical data, the number and nature of clinical cases is well defined in Phase 3. Suitable criteria for a clinical trial are determined in Phase 4. Phase 5 is a randomized trial to compare the specificity and sensitivity of the leads with those of current biomarkers in the

Today, the importance of sensitivity and specificity in biomarker selection has shifted proteomic studies from large-scale analysis to clinically-relevant validation. In addition to large-scale analysis in protein or metabolite identification (Mou et al., 2011; Huang et al.,

The purpose of this study was to discover potential biomarkers for the early detection of acute respiratory distress syndrome. To avoid sampling-induced complications, bronchial washings from lung cancer patients before and after surgical therapy (lobectomy) were collected. To reduce population heterogeneity, cancer stage, hormonal variation, and tumor location were well-defined. Only patients older than 60 years, had right lung cancer at

Those patients who met the inclusion and exclusion criteria were selected as controls. The inclusion criteria were: defined cancer in any lobe of the right lung, non-smoker, age ≥60 years, elective operation, operation period <210 min, forced expiratory volume in 1 s (FEV1) >80%, and no prior major lung resection or thoracic irradiation. Exclusion criteria were: age <60 years, operation period >210 min, FEV1 <60%, emergency or urgent operation, and prior major lung resection or distant thoracic irradiation. Based on our criteria, 7 patients (5 females and 2 males with ages ranging between 61 and 77 years) were included as controls. A review was conducted from the medical records and prospective database. The study protocol was approved by the Human Medical Studies Committee at National Cheng Kung University Medical College Hospital. Informed consent was given by all participants or their

Prior to a large-scale analysis of proteins in bronchial washings, the protein profiles of washings from different compartments of the lungs before and after lobectomy were compared. In the right lung where tumor tissues were identified, no clear bands at molecular weights >75 kDa were found in washings collected before or after lobectomy. In the left lung, no clear bands at molecular weights >75 kDa were found in the washings collected before lobectomy. After lobectomy, more bands at molecular weights >75 kDa were found. The intensity of each band was much greater (Fig. 1). Similar patterns were

**3. Translational study from protein identification to clinical application** 

stages IA and IB, and agreed to receive right lung lobectomy were recruited.

**2.5 Sensitivity and specificity of the lead proteins after proteomic analysis** 

2011), the leads are selected based on their sensitivity and specificity.

biomarker identification (Wu et al., 2005; Chang et al., 2007).

To effectively reduce post-surgical mortality, early detection of acute respiratory distress syndrome may provide in-depth information for the design of management plans, including non-pharmacological therapies (Villar et al., 2011).

### **2. Proteomic analysis of bronchoalveolar lavage fluid in biomarker studies**

To effectively identify the biomarkers of various lung diseases, bronchoalveolar lavage fluid from the lower airways and alveoli is collected for genomic or cytological analysis of cellular components (Meyer, 2007). This lung-specific fluid can be used for protein analysis. Identification of the non-cellular components in bronchoalveolar lavage fluid may provide valuable data for the early detection of acute lung injury.

#### **2.1 Current advances in proteomic analysis of bronchoalveolar lavage fluid**

In the past decades, over 100 human proteins or protein isoforms have been identified in bronchoalveolar lavage fluid from patients with various lung diseases (Wattiez et al., 1999; Lenz et al., 1993; Sadaghdar et al., 1992; Sabounchi-Schütt et al., 2001; Vesterberg et al., 2001). The major challenge today is to identify the lead proteins and validate the potential biomarkers (Turtoi et al., 2011a). To overcome this difficulty, integration of clinical studies with proteomic analysis of bronchoalveolar fluid is a potential solution (Turtoi et al., 2011b).

#### **2.2 Sampling concern in proteomic analysis of bronchoalveolar lavage fluid**

In clinical proteomics, the most difficult challenge before sample analysis is patient selection and sample collection (Apweiler et al., 2009). In the case of lung cancer patients, the major concern is to collect bronchoalveolar lavage fluid from those who may develop post-surgical lung edema. Although both bronchoalveolar lavage fluid and bronchial washings are collected using similar procedures, the former is collected from terminal alveoli after instilling more than 140 ml of sterile saline and the latter is collected from major airways after instilling less than 140 ml of saline. Because of the concern that excessive fluid accumulation may cause the complication of lung edema, bronchial washing is a better choice for conducting clinical proteomics.

#### **2.3 Technical limitations in proteomic analysis of bronchoalveolar lavage fluid**

In conventional proteomic analysis, two-dimensional gel electrophoresis provides good protein separation. However, it restricts the discovery of proteins with extreme biochemical properties such as size, isoelectric point, and solubility (Rabilloud, 2002). In comparison, one-dimensional gel electrophoresis provides easy comparison of banding patterns in protein profiling but is less efficient in protein separation. Moreover, the high salt concentration in the bronchoalveolar lavage fluid interferes with the resolution of protein separation to a lesser extent in one-dimensional gel electrophoresis (Plymoth, 2003).

#### **2.4 Application of 1D gel with liquid chromatography and MS/MS in biomarker discovery**

The rapid development of LC/MS/MS offers a better solution to one-dimensional gel electrophoresis (Schirle et al., 2003). A similar approach has been used to discover proteins

To effectively reduce post-surgical mortality, early detection of acute respiratory distress syndrome may provide in-depth information for the design of management plans, including

**2. Proteomic analysis of bronchoalveolar lavage fluid in biomarker studies**  To effectively identify the biomarkers of various lung diseases, bronchoalveolar lavage fluid from the lower airways and alveoli is collected for genomic or cytological analysis of cellular components (Meyer, 2007). This lung-specific fluid can be used for protein analysis. Identification of the non-cellular components in bronchoalveolar lavage fluid may provide

**2.1 Current advances in proteomic analysis of bronchoalveolar lavage fluid** 

**2.2 Sampling concern in proteomic analysis of bronchoalveolar lavage fluid** 

**2.3 Technical limitations in proteomic analysis of bronchoalveolar lavage fluid** 

**2.4 Application of 1D gel with liquid chromatography and MS/MS in biomarker** 

The rapid development of LC/MS/MS offers a better solution to one-dimensional gel electrophoresis (Schirle et al., 2003). A similar approach has been used to discover proteins

In conventional proteomic analysis, two-dimensional gel electrophoresis provides good protein separation. However, it restricts the discovery of proteins with extreme biochemical properties such as size, isoelectric point, and solubility (Rabilloud, 2002). In comparison, one-dimensional gel electrophoresis provides easy comparison of banding patterns in protein profiling but is less efficient in protein separation. Moreover, the high salt concentration in the bronchoalveolar lavage fluid interferes with the resolution of protein separation to a lesser extent in one-dimensional gel electrophoresis (Plymoth,

In the past decades, over 100 human proteins or protein isoforms have been identified in bronchoalveolar lavage fluid from patients with various lung diseases (Wattiez et al., 1999; Lenz et al., 1993; Sadaghdar et al., 1992; Sabounchi-Schütt et al., 2001; Vesterberg et al., 2001). The major challenge today is to identify the lead proteins and validate the potential biomarkers (Turtoi et al., 2011a). To overcome this difficulty, integration of clinical studies with proteomic analysis of bronchoalveolar fluid is a potential solution

In clinical proteomics, the most difficult challenge before sample analysis is patient selection and sample collection (Apweiler et al., 2009). In the case of lung cancer patients, the major concern is to collect bronchoalveolar lavage fluid from those who may develop post-surgical lung edema. Although both bronchoalveolar lavage fluid and bronchial washings are collected using similar procedures, the former is collected from terminal alveoli after instilling more than 140 ml of sterile saline and the latter is collected from major airways after instilling less than 140 ml of saline. Because of the concern that excessive fluid accumulation may cause the complication of lung edema, bronchial washing is a better

non-pharmacological therapies (Villar et al., 2011).

valuable data for the early detection of acute lung injury.

(Turtoi et al., 2011b).

2003).

**discovery** 

choice for conducting clinical proteomics.

with molecular weight greater than 100 kDa (such as α2-macroglobulin). The discovery of hundreds of proteins in bronchoalveolar lavage fluid demonstrates its feasibility in biomarker identification (Wu et al., 2005; Chang et al., 2007).

#### **2.5 Sensitivity and specificity of the lead proteins after proteomic analysis**

To accelerate the translation of biomarker discovery from bench to bedside, the development of techniques has been divided into 5 stages (Pepe et al., 2001). In Phase 1, potential biomarkers are discovered by various approaches, such as proteomic analysis. After the leads are identified by biochemical studies, measurable classifiers or outcomes are developed in Phase 2. Based on the analysis of their specificity and sensitivity, the cutoff point of the measurable outcome is determined and used in Phase 3. Based on patient history and clinical data, the number and nature of clinical cases is well defined in Phase 3. Suitable criteria for a clinical trial are determined in Phase 4. Phase 5 is a randomized trial to compare the specificity and sensitivity of the leads with those of current biomarkers in the market.

Today, the importance of sensitivity and specificity in biomarker selection has shifted proteomic studies from large-scale analysis to clinically-relevant validation. In addition to large-scale analysis in protein or metabolite identification (Mou et al., 2011; Huang et al., 2011), the leads are selected based on their sensitivity and specificity.

#### **3. Translational study from protein identification to clinical application**

The purpose of this study was to discover potential biomarkers for the early detection of acute respiratory distress syndrome. To avoid sampling-induced complications, bronchial washings from lung cancer patients before and after surgical therapy (lobectomy) were collected. To reduce population heterogeneity, cancer stage, hormonal variation, and tumor location were well-defined. Only patients older than 60 years, had right lung cancer at stages IA and IB, and agreed to receive right lung lobectomy were recruited.

Those patients who met the inclusion and exclusion criteria were selected as controls. The inclusion criteria were: defined cancer in any lobe of the right lung, non-smoker, age ≥60 years, elective operation, operation period <210 min, forced expiratory volume in 1 s (FEV1) >80%, and no prior major lung resection or thoracic irradiation. Exclusion criteria were: age <60 years, operation period >210 min, FEV1 <60%, emergency or urgent operation, and prior major lung resection or distant thoracic irradiation. Based on our criteria, 7 patients (5 females and 2 males with ages ranging between 61 and 77 years) were included as controls. A review was conducted from the medical records and prospective database. The study protocol was approved by the Human Medical Studies Committee at National Cheng Kung University Medical College Hospital. Informed consent was given by all participants or their legal guardians .

Prior to a large-scale analysis of proteins in bronchial washings, the protein profiles of washings from different compartments of the lungs before and after lobectomy were compared. In the right lung where tumor tissues were identified, no clear bands at molecular weights >75 kDa were found in washings collected before or after lobectomy. In the left lung, no clear bands at molecular weights >75 kDa were found in the washings collected before lobectomy. After lobectomy, more bands at molecular weights >75 kDa were found. The intensity of each band was much greater (Fig. 1). Similar patterns were found in all samples studied.

From Biomarker Discovery to

8 1483187 inter-α-trypsin inhibitor family

9 4507021 solute carrier family 4, anion

13 4504489 histidine-rich glycoprotein precursor

21 123510 haptoglobin-related protein precursor

right lung lobectomy.

exudation and leukocyte infiltration.

exchanger, member 1

heavy chain-related protein (IHRP)

**No GI** 

**number** 

Clinical Evaluation for Early Diagnosis of Lung Surgery-Induced Injury 37

**(kDa)** 

**No. of matched peptides**

103.553 2 2% 82

102.017 3 4% 213

60.527 2 4% 69

39.505 2 6% 114

**Sequence coverage**  **Score** 

**Protein name MW** 

1 28780 apo-B100 precursor 516.384 6 2% 285 2 1174412 spectrin α chain, erythrocyte 280.904 5 3% 257 3 134798 spectrin β chain, erythrocyte 247.040 7 4% 291 4 179674 complement component C4A 194.365 5 4% 216 5 4557385 complement component 3 precursor 188.612 8 7% 451 6 224053 α2-macroglobulin 162.096 2 1% 86 7 4557485 ceruloplasmin (ferroxidase) 122.998 2 2% 91

10 6005942 valosin-containing protein 89.962 4 9% 130 11 28592 serum albumin 71.351 12 19% 615 12 3287489 hsp89-α-δ-N 63.850 2 5% 85

14 553788 transferrin 55.233 4 10% 177 15 69990 α1-glycoprotein 52.488 2 6% 84 16 386789 hemopexin precursor 52.266 2 2% 81 17 38408 immunoglobulin M heavy chain 50.135 3 9% 112 18 229601 Ig G1 H Nie 49.812 3 12% 127 19 177827 α1-antitrypsin 46.790 2 7% 67 20 10334547 immunoglobulin heavy chain 42.319 2 8% 57

22 121039 Ig gamma-1 chain C region 36.605 3 14% 118 23 121043 Ig gamma-2 chain C region 36.500 2 11% 55 24 183817 β-globin 19.209 4 33% 221 25 442753 Chain D, Hemoglobin Ypsilanti 16.021 5 50% 206 Table 1. Proteins identified in bronchial washings from the left lung of a patient receiving

**3.1 Vascular endothelial growth factor and lobectomy-induced inflammation** 

Since vascular endothelial growth factor is a potent inducer of vascular permeability (Lee, 2005) and its expression is positively correlated with inflammation-induced protein exudation and leukocyte infiltration (Chang et al., 2005), it is plausible to suggest that an increase in vascular endothelial growth factor is associated with surgery-induced protein

Fig. 1. Protein profiling of bronchial washes from right (RL) and left lungs (LL) from patients before (Pre-Op) and after (Post-Op) right lung lobectomy. Bovine serum albumin (BSA) was used a positive control because albumin was identified in various bands.

The banding pattern allowed us to hypothesize that the proteins at molecular weights >75 kDa are exuded into alveoli after surgery. One-dimensional gel electrophoresis coupled with LC/MS/MS allowed us to identify the proteins in 13 major bands. As listed in Table 1, 8 proteins had molecular weights >100 kDa, including α2-macroglobulin.

To test our hypothesis that protein exudation is surgery-dependent, the relative abundance of α1-antitrypsin (47 kDa) and α2-macroglobulin (162 kDa) in bronchial washings was measured by Western blot analysis. α1-antitrypsin was found in washings collected before and after lobectomy (data not shown) but α2-macroglobulin was only found after lobectomy (Fig. 2).

Fig. 2. Relative abundance of α2-macroglobulin in bronchial washings before (Pre) and after (Post) lobectomy.

Fig. 1. Protein profiling of bronchial washes from right (RL) and left lungs (LL) from patients before (Pre-Op) and after (Post-Op) right lung lobectomy. Bovine serum albumin (BSA) was used a positive control because albumin was identified in various bands.

proteins had molecular weights >100 kDa, including α2-macroglobulin.

α2-macroglobulin→ α2-macroglobulin→ α2-macroglobulin→

(Fig. 2).

(Post) lobectomy.

The banding pattern allowed us to hypothesize that the proteins at molecular weights >75 kDa are exuded into alveoli after surgery. One-dimensional gel electrophoresis coupled with LC/MS/MS allowed us to identify the proteins in 13 major bands. As listed in Table 1, 8

To test our hypothesis that protein exudation is surgery-dependent, the relative abundance of α1-antitrypsin (47 kDa) and α2-macroglobulin (162 kDa) in bronchial washings was measured by Western blot analysis. α1-antitrypsin was found in washings collected before and after lobectomy (data not shown) but α2-macroglobulin was only found after lobectomy

Fig. 2. Relative abundance of α2-macroglobulin in bronchial washings before (Pre) and after

Patient 3 Patient 4 Patient 5 Pre Post Pre Post Pre Post


Table 1. Proteins identified in bronchial washings from the left lung of a patient receiving right lung lobectomy.

#### **3.1 Vascular endothelial growth factor and lobectomy-induced inflammation**

Since vascular endothelial growth factor is a potent inducer of vascular permeability (Lee, 2005) and its expression is positively correlated with inflammation-induced protein exudation and leukocyte infiltration (Chang et al., 2005), it is plausible to suggest that an increase in vascular endothelial growth factor is associated with surgery-induced protein exudation and leukocyte infiltration.

From Biomarker Discovery to

α2-macroglobulin) into alveoli.

and met these criteria were studied.

syndrome (ARDS).

\*Significant difference from pre-op.

Total cell number (x10,000)

**with acute respiratory distress syndrome** 

contribute to the development of this syndrome.

pre-op post-op ARDS

\*

Clinical Evaluation for Early Diagnosis of Lung Surgery-Induced Injury 39

Likewise, the relative expression of α1-antitrypsin at bands 5, 7, and 8 from bronchial washing was positively correlated with protein concentration, leukocyte number, and the level of vascular endothelial growth factor (data not shown). These data supported our hypothesis that the increase of vascular endothelial growth factor after surgery facilitates leukocyte infiltration and the exudation of acute-phase proteins (such as α1-antitrypsin and

**3.3 Characterization of α2-macroglobulin and α1-antitrypsin in lobectomized patients** 

Based on the report of the joint American–European Consensus Conference, the acute respiratory distress syndrome is well defined as follows: bilateral infiltrates on frontal chest radiography, the absence of left atrial hypertension (pulmonary capillary wedge pressure <18 mmHg or no clinical signs of left ventricular failure), and severe hypoxemia with a PaO2/FiO2 ratio <200 mmHg (Bernard et al., 1994). Five patients who received lung surgery

The group with lobectomy free of complications had levels of total protein and total leukocyte numbers in their bronchial washings similar to those who developed acute respiratory distress syndrome (P >0.05, Fig. 4). These data indicate that lung surgery induces inflammation (leukocyte infiltration and protein exudation) in the groups with and without the complication of acute respiratory distress syndrome. So, factors other than inflammation

0

pre-op post-op ARDS

\*

\*

4

Total proteins (g/L)

Total protein (g/L)

Fig. 4. Total leukocyte number and protein concentration in patients before (pre-op) and after lobectomy (post-op) with no complication and those with acute respiratory distress

surgery does not contribute to surgery-induced acute respiratory distress syndrome.

In lung cancer patients, an increase of vascular endothelial growth factor is positively associated with poor prognosis (*P* = 0.018; Han et al., 2001) but not with a worse postoperative year-survival rate (*P* = 0.0643; Liao et al., 2001). These reports are also consistent with our finding that the increase of vascular endothelial growth factor after lung

8

**3.3.1 Characterization of patients with acute respiratory distress syndrome** 

\*

As shown in Table 2, the vascular endothelial growth factor level was positively correlated with total protein concentration (y = 0.0025x + 1.0755, R2 = 0.7359, *P* <0.05) and cell count (y = 0.0696x - 0.6441, R2 = 0.8463, *P* <0.05) but not with operation duration or PaO2/FiO2. The correlation analysis supported the hypothesis that the induction of vascular endothelial growth factor after surgery contributes to leukocyte infiltration and protein exudation.


PaO2: arterial partial pressure of oxygen; FiO2: inspired oxygen fraction; PaO2/FiO2: oxygenation index Table 2. Clinical data of 7 patients who received right lung lobectomy without complications

#### **3.2 α2-macroglobulin and α1- antitrypsin in lobectomy-induced inflammation**

The relative expression of α2-macroglobulin at bands 2, 4, and 5 from bronchial washings was correlated with protein concentration, leukocyte number, and the level of vascular endothelial growth factor (Fig. 3).

Fig. 3. Correlation analysis of α2-macroglobulin and VEGF/total cells in bronchial washings.

As shown in Table 2, the vascular endothelial growth factor level was positively correlated with total protein concentration (y = 0.0025x + 1.0755, R2 = 0.7359, *P* <0.05) and cell count (y = 0.0696x - 0.6441, R2 = 0.8463, *P* <0.05) but not with operation duration or PaO2/FiO2. The correlation analysis supported the hypothesis that the induction of vascular endothelial growth factor after surgery contributes to leukocyte infiltration and protein exudation.

PaO2/FiO2 Cell count

**3.2 α2-macroglobulin and α1- antitrypsin in lobectomy-induced inflammation** 

The relative expression of α2-macroglobulin at bands 2, 4, and 5 from bronchial washings was correlated with protein concentration, leukocyte number, and the level of vascular

Fig. 3. Correlation analysis of α2-macroglobulin and VEGF/total cells in bronchial washings.

1 152 270.167 2.5 0.641 162.80 2 191 202.500 67.0 3.412 613.68 3 234 435.000 100.0 4.826 1517.40 4 160 220.800 22.5 1.737 89.95 5 196 336.000 3.5 2.462 99.79 6 140 338.333 0.5 0.660 109.64 7 221 505.600 5.0 1.302 359.69 PaO2: arterial partial pressure of oxygen; FiO2: inspired oxygen fraction; PaO2/FiO2: oxygenation index Table 2. Clinical data of 7 patients who received right lung lobectomy without complications

(104 cells/ml)

Protein conc. (μg/ml)

VEGF (pg/ml)

Patient No

Operation duration (min)

endothelial growth factor (Fig. 3).

Likewise, the relative expression of α1-antitrypsin at bands 5, 7, and 8 from bronchial washing was positively correlated with protein concentration, leukocyte number, and the level of vascular endothelial growth factor (data not shown). These data supported our hypothesis that the increase of vascular endothelial growth factor after surgery facilitates leukocyte infiltration and the exudation of acute-phase proteins (such as α1-antitrypsin and α2-macroglobulin) into alveoli.

#### **3.3 Characterization of α2-macroglobulin and α1-antitrypsin in lobectomized patients with acute respiratory distress syndrome**

Based on the report of the joint American–European Consensus Conference, the acute respiratory distress syndrome is well defined as follows: bilateral infiltrates on frontal chest radiography, the absence of left atrial hypertension (pulmonary capillary wedge pressure <18 mmHg or no clinical signs of left ventricular failure), and severe hypoxemia with a PaO2/FiO2 ratio <200 mmHg (Bernard et al., 1994). Five patients who received lung surgery and met these criteria were studied.

#### **3.3.1 Characterization of patients with acute respiratory distress syndrome**

The group with lobectomy free of complications had levels of total protein and total leukocyte numbers in their bronchial washings similar to those who developed acute respiratory distress syndrome (P >0.05, Fig. 4). These data indicate that lung surgery induces inflammation (leukocyte infiltration and protein exudation) in the groups with and without the complication of acute respiratory distress syndrome. So, factors other than inflammation contribute to the development of this syndrome.

\*Significant difference from pre-op.

Fig. 4. Total leukocyte number and protein concentration in patients before (pre-op) and after lobectomy (post-op) with no complication and those with acute respiratory distress syndrome (ARDS).

In lung cancer patients, an increase of vascular endothelial growth factor is positively associated with poor prognosis (*P* = 0.018; Han et al., 2001) but not with a worse postoperative year-survival rate (*P* = 0.0643; Liao et al., 2001). These reports are also consistent with our finding that the increase of vascular endothelial growth factor after lung surgery does not contribute to surgery-induced acute respiratory distress syndrome.

From Biomarker Discovery to

respiratory distress syndrome (ARDS).

**acute respiratory distress syndrome** 

(3/7).

could be due to functional changes in lung epithelial cells.

negative among those without the condition) can be calculated.

was considered an indication of acute respiratory distress syndrome.

Clinical Evaluation for Early Diagnosis of Lung Surgery-Induced Injury 41

Fig. 6. Relative expression of α1-antitrypsin and α2-macroglobulin (macroglobulin) in the lobectomized group without complications (lobectomy) and in the group with acute

α1-antitrypsin in A549, a lung epithelial cell line. The changes in α1-antitrypsin variants

It is of importance to turn the relative expression of α1-antitrypsin in bronchial washings into a measurable outcome because only the measurable outcome is used to determine the cutoff value. Based on the cutoff value, sensitivity (the proportion of subjects who test positive among those with the condition) and specificity (the proportion of subjects who test

As shown in Fig. 6, α1-antitrypsin variants at bands 7 (47 kDa) and 8 (40 kDa) had a lower abundance in the group without complications than the group with acute respiratory syndrome. To avoid variations in sample loading and the intensity in each calculation, the ratio of the expression of α1-antitrypsin at band 5 (70 kDa) to that at bands 7 and 8 was used as the measurable outcome. Based on this calculation, the cutoff value was 0.5. A ratio <0.5

Table 3 shows the ratio for each patient from the complication-free group. Four out of 7 patients had a ratio <0.5. The specificity of α1-antitrypsin for true negative patients was 0.43

Table 4 shows the ratio for each patient from the complication group. Three out of 5 patients had a ratio <0.5. The sensitivity of α1-antitrypsin for true positive patients was 0.6 (3/5).

**3.4 Specificity and sensitivity of α1-antitrypsin variants as potential biomarkers for** 

#### **3.3.2 Protein profiling of bronchial washings from lobectomized patients with acute respiratory distress syndrome**

Unlike patients with no complications, those with acute respiratory distress syndrome showed white or gray patches on the chest X-ray. In one-dimensional gel electrophoresis, the protein profiling of bronchial washings from patients without complications showed a much clearer banding pattern than those from patients with acute respiratory distress syndrome (Fig. 5). Eight bands from each gel were cut and subjected to LC/MS/MS for protein identification. No protein was identified in Lane 1. The most significant difference was that albumin appeared in almost every band of the samples from patients without complications but not in those with acute respiratory distress syndrome. In contrast, α1 antitrypsin was identified only in bands 6 and 7 from the group without complications but was found in bands 2, 3, 4, 5, 6, and 7 in the group with the complication (Fig. 5).

Fig. 5. Comparison of chest X-rays and protein profiling of bronchial washings in lobectomized patients with no complications (lobectomy, Lob) and those with acute respiratory distress syndrome (ARDS).

#### **3.3.3 α2-macroglobulin and α1-antitrypsin in bronchial washings from lobectomized patients with acute respiratory distress syndrome**

As shown in Fig. 6, both α2-macroglobulin and α1-antitrypsin were detected in bronchial washings after surgery.

After quantification, the total amounts of α2-macroglobulin at bands 2, 4, and 5 and α1 antitrypsin at bands 5, 7, and 8 did not show any statistical difference between the groups with and without complications. The most important finding was lower levels of α1 antytrypsin at bands 7 and 8 in the group without complications than the acute respiratory distress syndrome group (Fig. 6). It is likely that α1-antitrypsin variants at bands 5, 7, and 8 can be used as biomarkers for the early detection of acute respiratory distress syndrome.

In bronchial washings collected from the patients with acute respiratory distress syndrome, leukocyte number was not correlated with the total amounts of α2-macroglobulin or α1 antitrypsin. Our analyses again supported the notion that surgery-induced inflammation is not an important indicator in the early phase of acute respiratory distress syndrome.

It has been reported that α1-antitrypsin can be produced by lung epithelial cells (Venember et al., 1994) but α2-macroglobulin cannot. Our preliminary data confirmed the expression of

Unlike patients with no complications, those with acute respiratory distress syndrome showed white or gray patches on the chest X-ray. In one-dimensional gel electrophoresis, the protein profiling of bronchial washings from patients without complications showed a much clearer banding pattern than those from patients with acute respiratory distress syndrome (Fig. 5). Eight bands from each gel were cut and subjected to LC/MS/MS for protein identification. No protein was identified in Lane 1. The most significant difference was that albumin appeared in almost every band of the samples from patients without complications but not in those with acute respiratory distress syndrome. In contrast, α1 antitrypsin was identified only in bands 6 and 7 from the group without complications but

Marker Lob ARDS

1

2 α1-antitrypsin 3 α1-antitrypsin 4 α1-antitrypsin

α1-antitrypsin α1-antitrypsin α1-antitrypsin 8 β-actin

**3.3.2 Protein profiling of bronchial washings from lobectomized patients with acute** 

was found in bands 2, 3, 4, 5, 6, and 7 in the group with the complication (Fig. 5).

50

35

Fig. 5. Comparison of chest X-rays and protein profiling of bronchial washings in lobectomized patients with no complications (lobectomy, Lob) and those with acute

**3.3.3 α2-macroglobulin and α1-antitrypsin in bronchial washings from lobectomized** 

As shown in Fig. 6, both α2-macroglobulin and α1-antitrypsin were detected in bronchial

After quantification, the total amounts of α2-macroglobulin at bands 2, 4, and 5 and α1 antitrypsin at bands 5, 7, and 8 did not show any statistical difference between the groups with and without complications. The most important finding was lower levels of α1 antytrypsin at bands 7 and 8 in the group without complications than the acute respiratory distress syndrome group (Fig. 6). It is likely that α1-antitrypsin variants at bands 5, 7, and 8 can be used as biomarkers for the early detection of acute respiratory distress syndrome. In bronchial washings collected from the patients with acute respiratory distress syndrome, leukocyte number was not correlated with the total amounts of α2-macroglobulin or α1 antitrypsin. Our analyses again supported the notion that surgery-induced inflammation is

not an important indicator in the early phase of acute respiratory distress syndrome.

It has been reported that α1-antitrypsin can be produced by lung epithelial cells (Venember et al., 1994) but α2-macroglobulin cannot. Our preliminary data confirmed the expression of

**respiratory distress syndrome** 

Lobectomy

Acute respiratory distress syndrome

respiratory distress syndrome (ARDS).

washings after surgery.

**patients with acute respiratory distress syndrome** 

Fig. 6. Relative expression of α1-antitrypsin and α2-macroglobulin (macroglobulin) in the lobectomized group without complications (lobectomy) and in the group with acute respiratory distress syndrome (ARDS).

α1-antitrypsin in A549, a lung epithelial cell line. The changes in α1-antitrypsin variants could be due to functional changes in lung epithelial cells.

#### **3.4 Specificity and sensitivity of α1-antitrypsin variants as potential biomarkers for acute respiratory distress syndrome**

It is of importance to turn the relative expression of α1-antitrypsin in bronchial washings into a measurable outcome because only the measurable outcome is used to determine the cutoff value. Based on the cutoff value, sensitivity (the proportion of subjects who test positive among those with the condition) and specificity (the proportion of subjects who test negative among those without the condition) can be calculated.

As shown in Fig. 6, α1-antitrypsin variants at bands 7 (47 kDa) and 8 (40 kDa) had a lower abundance in the group without complications than the group with acute respiratory syndrome. To avoid variations in sample loading and the intensity in each calculation, the ratio of the expression of α1-antitrypsin at band 5 (70 kDa) to that at bands 7 and 8 was used as the measurable outcome. Based on this calculation, the cutoff value was 0.5. A ratio <0.5 was considered an indication of acute respiratory distress syndrome.

Table 3 shows the ratio for each patient from the complication-free group. Four out of 7 patients had a ratio <0.5. The specificity of α1-antitrypsin for true negative patients was 0.43 (3/7).

Table 4 shows the ratio for each patient from the complication group. Three out of 5 patients had a ratio <0.5. The sensitivity of α1-antitrypsin for true positive patients was 0.6 (3/5).

From Biomarker Discovery to

optimize the cutoff values.

**respiratory distress syndrome** 

factor-mediated permeability.

**4.2 Limitations of this study** 

acute respiratory distress syndrome.

diagnosis of acute respiratory distress syndrome.

distress syndrome.

further explored.

**and α1-antitrypsin variants as an example** 

was improved.

Clinical Evaluation for Early Diagnosis of Lung Surgery-Induced Injury 43

the sample expressed detectable α2-macroglobulin in bronchial washings. Accordingly, the specificity for true negative patients changed to 0.71 (5/7). The prediction for true negatives

**4. From identification of leads to further validation using α2-macroglobulin** 

**4.1 Contribution of this study to the discovery of biomarkers for detecting acute** 

After the discovery of potential biomarkers by proteomic analysis in this study, the first challenge was to identify the leads from the proteins discovered after developing a quick screening test. After Phase 1, the second challenge was to provide clear justification to

Ideally, quantitative proteomic analysis should be used to reveal lobectomy-induced changes of all proteins in bronchial washings. However, the unique compartment of the lung allowed us to analyze exudate components which may not exist before surgery, such as α2-macroglobulin. Based on the important mechanism of surgery-induced inflammation in the early phase of lung injury, one-dimensional gel electrophoresis in this study was an easy and suitable tool to identify α2-macroglobulin as an indicator of vascular endothelial growth

The second contribution of this study was to take advantage of one-dimensional gel electrophoresis with pattern analysis to reveal the pattern changes of α1-antitrypsin between the groups with and without post-surgical complications. The difference found allowed us to identify α1-antytripsin variants as biomarkers for the early detection of acute respiratory

In this study, α1-antitrypsin variants were considered as biomarkers for acute respiratory distress. No mechanistic data are provided to explain why and how the formation of α1 antitrypsin variants are related to the progression from surgery-induced inflammation to

The association between α1-antitrypsin variants and infection was first reported in 2010 (Zhang et al., 2010). The decrease of the α1-antitrypsin variant at 130 kDa and the increase of the variant at 40 kDa is associated with human immunodeficiency virus-induced infection. Glycoproteomic analysis shows that changes in α1-antitrypsin variants may be due to a shift of glycosylation. In future, glycoproteomic analysis of α1-antitrypsin variants should be

Although the analysis of their specificity and sensitivity, the cutoff point of the measurable outcome, and criteria for patient selection are clearly and easily determined, the small number of clinical cases in this study limits the generalization of α2-macroglobulin and α1 antitrypsin as markers for acute respiratory distress syndrome. To use them as measurable biomarkers in Phase 3, it is necessary to increase the number and the complexity of clinical cases for further validation on whether the cutoff points determined are suitable for early

One-dimensional gel electrophoresis does not offer a good way for protein separation. Comparative proteomic analysis only compares the intensity of each spot. These two


Table 3. Ratio of the expression of α1-antitrypsin at band 5 to that at bands 7 and 8 in the lobectomized patients without acute respiratory distress syndrome.


Table 4. Ratio of the expression of α1-antitrypsin at band 5 to that at bands 7 and 8 in lobectomized patients with acute respiratory distress syndrome.

#### **3.5 Further improvement of specificity and sensitivity for detecting acute respiratory distress syndrome using dual biomarkers**

As shown in Tables 3 and 4, the sensitivity of α1-antitrypsin variants for detecting acute respiratory distress syndrome (0.6) was better than the specificity (0.43). The major concern is how to optimize the cutoff value and improve the specificity. In table 3, patients 1 and 6 with ratios <0.5 showed the lowest values in cell counts and protein concentration. Meanwhile, the expression of α2-macroglobulin was almost undetectable, which indicates minor inflammation in the patients. The lower ratio of relative expression of α1-antitrypsin at band 5 to that at bands 7 and 8 was false-positive.

α1-antitrypsin was found in the lungs before and after surgery; α2-macroglobulin only occurred in the lungs after surgery. To avoid the lower levels of α1-antitrypsin variants which may create a false-positive result, α2-macroglobulin can be recruited as a second biomarker. The ratio of α1-antitrypsin variants was considered as a true result only when

Cutoff value = 0.5

Cutoff value = 0.5

Ratio of expression of α1-antitrypsin at band 5 to that at bands 7 and 8

1 0.000: 0.027 <0.5 2 0.043: 0.099 <0.5 3 0.019: 0.024 >0.5 4 0.017: 0.023 >0.5 5 0.018: 0.087 <0.5 6 0.000: 0.006 <0.5 7 0.042: 0.053 >0.5

Table 3. Ratio of the expression of α1-antitrypsin at band 5 to that at bands 7 and 8 in the

A 0.081: 0.177 <0.5 B 0.043: 0.199 <0.5 C 0.081: 0.086 >0.5 D 0.015: 0.040 <0.5 E 0.025: 0.048 >0.5

Table 4. Ratio of the expression of α1-antitrypsin at band 5 to that at bands 7 and 8 in

**3.5 Further improvement of specificity and sensitivity for detecting acute respiratory** 

As shown in Tables 3 and 4, the sensitivity of α1-antitrypsin variants for detecting acute respiratory distress syndrome (0.6) was better than the specificity (0.43). The major concern is how to optimize the cutoff value and improve the specificity. In table 3, patients 1 and 6 with ratios <0.5 showed the lowest values in cell counts and protein concentration. Meanwhile, the expression of α2-macroglobulin was almost undetectable, which indicates minor inflammation in the patients. The lower ratio of relative expression of α1-antitrypsin

α1-antitrypsin was found in the lungs before and after surgery; α2-macroglobulin only occurred in the lungs after surgery. To avoid the lower levels of α1-antitrypsin variants which may create a false-positive result, α2-macroglobulin can be recruited as a second biomarker. The ratio of α1-antitrypsin variants was considered as a true result only when

lobectomized patients without acute respiratory distress syndrome.

5 to that at bands 7 and 8

Patient No Ratio of expression of α1-antitrypsin at band

lobectomized patients with acute respiratory distress syndrome.

**distress syndrome using dual biomarkers** 

at band 5 to that at bands 7 and 8 was false-positive.

Patient No

the sample expressed detectable α2-macroglobulin in bronchial washings. Accordingly, the specificity for true negative patients changed to 0.71 (5/7). The prediction for true negatives was improved.

#### **4. From identification of leads to further validation using α2-macroglobulin and α1-antitrypsin variants as an example**

After the discovery of potential biomarkers by proteomic analysis in this study, the first challenge was to identify the leads from the proteins discovered after developing a quick screening test. After Phase 1, the second challenge was to provide clear justification to optimize the cutoff values.

#### **4.1 Contribution of this study to the discovery of biomarkers for detecting acute respiratory distress syndrome**

Ideally, quantitative proteomic analysis should be used to reveal lobectomy-induced changes of all proteins in bronchial washings. However, the unique compartment of the lung allowed us to analyze exudate components which may not exist before surgery, such as α2-macroglobulin. Based on the important mechanism of surgery-induced inflammation in the early phase of lung injury, one-dimensional gel electrophoresis in this study was an easy and suitable tool to identify α2-macroglobulin as an indicator of vascular endothelial growth factor-mediated permeability.

The second contribution of this study was to take advantage of one-dimensional gel electrophoresis with pattern analysis to reveal the pattern changes of α1-antitrypsin between the groups with and without post-surgical complications. The difference found allowed us to identify α1-antytripsin variants as biomarkers for the early detection of acute respiratory distress syndrome.

#### **4.2 Limitations of this study**

In this study, α1-antitrypsin variants were considered as biomarkers for acute respiratory distress. No mechanistic data are provided to explain why and how the formation of α1 antitrypsin variants are related to the progression from surgery-induced inflammation to acute respiratory distress syndrome.

The association between α1-antitrypsin variants and infection was first reported in 2010 (Zhang et al., 2010). The decrease of the α1-antitrypsin variant at 130 kDa and the increase of the variant at 40 kDa is associated with human immunodeficiency virus-induced infection. Glycoproteomic analysis shows that changes in α1-antitrypsin variants may be due to a shift of glycosylation. In future, glycoproteomic analysis of α1-antitrypsin variants should be further explored.

Although the analysis of their specificity and sensitivity, the cutoff point of the measurable outcome, and criteria for patient selection are clearly and easily determined, the small number of clinical cases in this study limits the generalization of α2-macroglobulin and α1 antitrypsin as markers for acute respiratory distress syndrome. To use them as measurable biomarkers in Phase 3, it is necessary to increase the number and the complexity of clinical cases for further validation on whether the cutoff points determined are suitable for early diagnosis of acute respiratory distress syndrome.

One-dimensional gel electrophoresis does not offer a good way for protein separation. Comparative proteomic analysis only compares the intensity of each spot. These two

From Biomarker Discovery to

1073-449X

1068-9265

1615-9861

pp.229-35, ISSN 1341-1098

Clinical Evaluation for Early Diagnosis of Lung Surgery-Induced Injury 45

Bhatia, M., & Moochhala, S. (2004). Role of inflammatory mediators in the pathophysiology

Chang, CC., Chiu, HF., Wu, YS., Li, YC., Tsai, ML., Shen, CK., & Yang, CY. (2005). The

Chang, CC., Chen, SH., Ho, SH., Yang, CY., Wang, HD., & Tsai, ML. (2007). Proteomic

Chiang, CJ., Chen, YC., Chen, CJ., You, SL., & Lai, MS. (2010). Cancer trends in Taiwan.

Donnelly, SC., Strieter, RM., Reid, PT., Kunkel, SL., Burdick, MD., Armstrong, I., Mackenzie,

Geiser, T., Atabai, K., Jarreau, PH., Ware, LB., Pugin, J., & Matthay, MA. (2001). Pulmonary

Gunluoglu, MZ., Demir, A., Turna, A., Sansar, D., Melek, H., Dincer, SI., & Gurses, A. (2011).

Han, H., Silverman, JF., Santucci, TS., Macherey, RS., d'Amato, TA., Tung, MY., Weyant, RJ.,

Huang, SY., Tsai, ML., Tsai, CJ., Wu, JL., Hsu, JL., Ho, SH., & Chen SH. (2006). Quantitation

Huang, Z., Lin, L., Gao, Y., Chen, Y., Yan, X., Xing, J., & Hang, W. (2011). Bladder cancer

press). *Molecular & Cellular Proteomics,* ISSN 1535-9484

*Medicine,* Vol.125, No.3, (August 1996), pp.191–6, ISSN 1539-3704

*Perspectives*, Vol.113, No.4, (April 2005), pp.454-60, ISSN 0091-6765

3 Pt 1, (March 1994), pp. 818-24, ISSN 1073-449X

No.23, (December 2007), pp.4388-97, ISSN 1615-9861

No.10, (October 2010), pp.897-904, ISSN 0368-2811

2004), pp. 145-56, ISSN 1096-9896

coordination. *American Journal of Respiratory and Critical Care Medicine*, Vol.149, No.

of acute respiratory distress syndrome. *Journal of Patholog,* Vol.202, No.2, (February

induction of vascular endothelial growth factor by ultrafine carbon black contributes to the increase of alveolar-capillary permeability. *Environmental Health* 

analysis of proteins from bronchoalveolar lavage fluid reveals the action mechanism of ultrafine carbon black-induced lung injury in mice. *Proteomics,* Vol.7,

Taiwan Cancer Registry Task Force. *Japanese Journal of Clinical Oncology,* Vol.40,

A., & Haslett, C. (1996). The association between mortality rates and decreased concentrations of interleukin-10 and interleukin-1 receptor antagonist in the lung fluids of patients with the adult respiratory distress syndrome. *Annals of Internal* 

edema fluid from patients with acute lung injury augments in vitro alveolar epithelial repair by an IL-1beta-dependent mechanism. *American Journal of Respiratory and Critical Care Medicine,* Vol.163, No.6, (May 2001), pp.1384–8, ISSN

Extent of lung resection in non-small lung cancer with interlobar lymph node involvement. *Annals of Thoracic and Cardiovascular Surgery*, Vol.17, No.3, (June 2011),

& Landreneau, RJ. (2001). Vascular endothelial growth factor expression in stage I non-small cell lung cancer correlates with neoangiogenesis and a poor prognosis. *Annals of Surgical Oncology*, Vol.8, No.1, (January-February 2001), pp.72-9, ISSN

of protein phosphorylation in pregnant rat uteri using stable isotope dimethyl labeling coupled with IMAC. *Proteomics*, Vol.6, No.6, (March 2006), pp.1-12, ISSN

determination via two urinary metabolites: a biomarker pattern approach. (in

approaches may our discovery of new proteins. The technology of stable isotope dimethyl labeling coupled with LC/MS/MS permits further quantification of specific peptides of each protein and provides a better quantification tool after one-dimensional electrophoresis (Huang et al., 2006). This approach then compensates for the limitation of one-dimensional gel electrophoresis.

#### **5. Conclusion**

Both inflammation -dependent and -independent mechanisms contribute to the progression from lung injury to acute respiratory distress syndrome. Stage-dependent changes in biomarkers allow us to monitor the progression of the diseases and develop new treatments in a stage-dependent manner.

In this study, α2-macroglobulin and α1-antitrypsin were positively correlated with vascular endothelial growth factor, clearly showing lobectomy-induced inflammation. The total amount of α1-macroglobulin can be used as a biomarker of increased vascular permeability in the lung. The severity of lobectomy-induced inflammation is similar to that of inflammation in acute respiratory distress syndrome but respiratory function becomes much worse in patients with the syndrome. Concomitantly, the patients with acute respiratory distress syndrome had lower levels of α1-antitrypsin at higher molecular weights and higher levels of α1-antitrypsin at lower molecular weights. Similarly, human immunodeficiency virus-induced infection is associated with the decreased abundance of α1-antitrypsin at higher molecular weights and the increased abundance of α1-antitrypsin at lower molecular weights (Zhang et al., 2010). Because α1-antitrypsin exists in lung epithelial cells (Venember et al., 1994), the changes of α1-antitrypsin variants in the patients with acute respiratory distress may reflect lung epithelial damage.

#### **6. Acknowledgment**

The authors appreciate the technical support of Shih-Hsin Ho, Hong-Da Wang, and Yan-Jie Chen, clinical sample collections by Drs. Jia-Ming Chang and Chang-Wen Chen, and grant support from the National Science Council, Taiwan (NSC-95-2314-B-006-125-MY2 and NSC-95-2323-B-006-004).

#### **7. References**


approaches may our discovery of new proteins. The technology of stable isotope dimethyl labeling coupled with LC/MS/MS permits further quantification of specific peptides of each protein and provides a better quantification tool after one-dimensional electrophoresis (Huang et al., 2006). This approach then compensates for the limitation of one-dimensional gel

Both inflammation -dependent and -independent mechanisms contribute to the progression from lung injury to acute respiratory distress syndrome. Stage-dependent changes in biomarkers allow us to monitor the progression of the diseases and develop new treatments

In this study, α2-macroglobulin and α1-antitrypsin were positively correlated with vascular endothelial growth factor, clearly showing lobectomy-induced inflammation. The total amount of α1-macroglobulin can be used as a biomarker of increased vascular permeability in the lung. The severity of lobectomy-induced inflammation is similar to that of inflammation in acute respiratory distress syndrome but respiratory function becomes much worse in patients with the syndrome. Concomitantly, the patients with acute respiratory distress syndrome had lower levels of α1-antitrypsin at higher molecular weights and higher levels of α1-antitrypsin at lower molecular weights. Similarly, human immunodeficiency virus-induced infection is associated with the decreased abundance of α1-antitrypsin at higher molecular weights and the increased abundance of α1-antitrypsin at lower molecular weights (Zhang et al., 2010). Because α1-antitrypsin exists in lung epithelial cells (Venember et al., 1994), the changes of α1-antitrypsin variants in the patients with acute

The authors appreciate the technical support of Shih-Hsin Ho, Hong-Da Wang, and Yan-Jie Chen, clinical sample collections by Drs. Jia-Ming Chang and Chang-Wen Chen, and grant support from the National Science Council, Taiwan (NSC-95-2314-B-006-125-MY2 and NSC-

Apweilerm, R., Aslanidis, C., Deufel, T., Gerstner, A., Hansen, J., Hochstrasser, D., Kellner,

Bernard, GR., Artigas, A., Brigham, KL., Carlet, J., Falke, K., Hudson, L., Lamy, M., Legall,

R., Kubicek, M., Lottspeich, F., Maser, E., Mewes, HW., Meyer, HE., Müllner, S., Mutter, W., Neumaier, M., Nollau, P., Nothwang, HG., Ponten, F., Radbruch, A., Reinert, K., Rothe, G., Stockinger, H., Tárnok, A., Taussig, MJ., Thiel, A., Thiery, J., Ueffing, M., Valet, G., Vandekerckhove, J., Wagener, C., Wagner, O., & Schmitz, G. (2009). Approaching clinical proteomics: current state and future fields of application in cellular proteomics. *Cytometry A*, Vol.75, No.10, (October 2009), pp.

JR., Morris, A., & Spragg, R. (1994). The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial

electrophoresis.

**5. Conclusion** 

in a stage-dependent manner.

**6. Acknowledgment** 

95-2323-B-006-004).

816-32, ISSN 1552-4930

**7. References** 

respiratory distress may reflect lung epithelial damage.

coordination. *American Journal of Respiratory and Critical Care Medicine*, Vol.149, No. 3 Pt 1, (March 1994), pp. 818-24, ISSN 1073-449X


From Biomarker Discovery to

1522-2683

1535-9484

0002-9440

1992), pp.63-9. ISSN 1931-3543

(August 2002), pp.1693–1700, ISSN 1530-0293

No.2-3, (June 1994), pp.171-4. ISSN 0014-5793

Vol.*74*, No.4, (May 2001), pp.249-54, ISSN 1432-1246

2011), pp.3160-82, ISSN 1535-3907

2011), pp.647-53, ISSN 1827-1596

(June 1999), pp.1634-45, ISSN 1522-2683

Clinical Evaluation for Early Diagnosis of Lung Surgery-Induced Injury 47

Sadaghdar, H., Huang, ZB., & Eden, E. (1992). Correlation of bronchoalveolar lavage

Schirle, M., Heurtier, MA., & Kuster, B. (2003). Profiling core proteomes of human cell lines

Spragg, RG., Bernard, GR., Checkley, W., Curtis, JR., Gajic, O., Guyatt, G., Hall, J., Israel, E.,

Turtoi, A., De Pauw, E., & Castronovo, V. (2011a). Innovative proteomics for the discovery

Turtoi, A., Dumont, B., Greffe, Y., Blomme, A., Mazzucchelli, G., Delvenne, P., Mutijima,

Venembre, P., Boutten, A., Seta, N., Dehoux, MS., Crestani, B., Aubier, M., & Durand, G.

Vesterberg, O., Palmberg, L., & Larsson, K. (2001). Albumin, transferrin and alpha2-

Villar, J., Blanco, J., Zhang, H., & Slutsky, AS. (2011). Ventilator-induced lung injury and

Wattiez, R., Hermans, C., Bernard, A., Lesur, O., & Falmagne, P. (1999). Human

Wu, J., Kobayashi, M., Sousa, EA., Liu, W., Cai, J., Goldman, SJ., Dorner, AJ., Projan, SJ.,

application method. *Electrophoresis*, Vol.22, No.9, (May 2001), pp.1851–60, ISSN

findings to severity of Pneumocystis carinii pneumonia in AIDS. Evidence for the development of high-permeability pulmonary edema. *Chest* Vol.*102*, No.1, (July

by one-dimensional PAGE and liquid chromatography-tandem mass spectrometry. *Molecular & Cellular Proteomics,* Vol.2, No.12, (December 2003), pp.1297-305, ISSN

Jain, M., Needham, DM., Randolph, AG., Rubenfeld, GD., Schoenfeld, D., Thompson, BT., Ware, LB., Young, D., & Harabin, AL. (2010). Beyond mortality: future clinical research in acute lung injury. *American Journal of Respiratory and Critical Care Medicine*, Vol.181, No.10, (May 2010), pp.1121-7, ISSN 1073-449X Tremblay, LN., Miatto, D., Hamid, Q., Govindarajan, A., & Slutsky, AS. (2002). Injurious

ventilation induces widespread pulmonary epithelial expression of tumor necrosis factor-alpha and interleukin-6 messenger RNA. *Critical Care Medicine,* Vol.30, No.8,

of systemically accessible cancer biomarkers suitable for imaging and targeted therapies. *American Journal of Pathology*, Vol.178, No.1, (January 2011), pp.12-8 ISSN

EN., Lifrange, E., De Pauw, E., & Castronovo, V. (2011b). Novel Comprehensive Approach for Accessible Biomarker Identification and Absolute Quantification from Precious Human Tissues. *Journal of Proteome Research,* Vol.10, No.7, (July

Secretion of alpha 1-antitrypsin by alveolar epithelial cells. FEBS Letters, Vol.346,

macroglobulin in bronchoalveolar lavage fluid following exposure to organic dust in healthy subjects. *International Archives of Occupational and Environmental Health*,

sepsis: two sides of the same coin? *Minerva Anestesiologica,* Vol.77, No.6, (June

bronchoalveolar lavage fluid: two-dimensional gel electrophoresis, amino acid microsequencing and identification of major proteins. *Electrophoresis,* Vol.20, No.7,

Kavuru, MS., Qiu, Y., & Thomassen, MJ. (2005). Differential proteomic analysis of


Jordan, S., Mitchell, JA., Quinlan, GJ., Goldstraw, P., & Evans, TW. (2000). The pathogenesis

Kollef, MH., & Schuster, DP. (1998). The acute respiratory distress syndrome. *The New* 

Landis, SH., Murray, T., Bolden, S., & Wingo, PA. (1998). Cancer statistics, *CA-a Cancer* 

Lee, YC. (2005). The involvement of VEGF in endothelial permeability: a target for anti-

Lenz, AG, Meyer, B., Costabel, U., & Maier, K. (1993). Bronchoalveolar lavage fluid proteins

Martini, N., Bains, MS., Burt, ME., Zakowski, MF., McCormack, P., Rusch, VW., & Ginsberg,

Medford, AR., & Millar, AB. (2006). Vascular endothelial growth factor (VEGF) in acute lung

Pepe, MS., Etzioni, R., Feng, Z., Potter, JD., Thompson, M., Thornquist, M., Winget, M., &

Plymoth, A., Löfdahl, CG., Ekberg-Jansson, A., Dahlbäck, M., Lindberg, H., Fehniger, TE., &

Rabilloud, T. (2002). Two-dimensional gel electrophoresis in proteomics: old, old fashioned,

Sabounchi-Schütt, F., Aström, J., Eklund, A., Grunewald, J., & Bjellqvist, B. (2001).

paradigm? *Thorax*, Vol.61, No.7, (July 2006), pp.621-6, ISSN 1468-3296 Meyer, KC. (2007). Bronchoalveolar lavage as a diagnostic tool. *Seminars in Respiratory and Critical Care Medicine*, Vol.28, No.5, (October 2007), pp.546-60, ISSN 1069-3424 Mou, Y., Xing, R., & Liu, C. (2011). Diagnosis of Gallbladder Cancer Using Matrix-Assisted

*Electrophoresis,* Vol.14, No. 3, (March 1993), pp. 242-4, ISSN 1522-2683 Liao, M., Wang, H., Lin, Z., Feng, J., & Zhu, D. (2001). Vascular endothelial growth factor

No.4, (April 2000), pp.790-9, ISSN 1399-3003

(November 2005), pp.1124-30, ISSN 1472-4472

(January 1995), pp.120-9, ISSN 0022-5223

*Journal of the Medical Sciences,* ISSN 1538-2990

pp.125-32, ISSN 0169-5002

1460-2105

pp.962-72 ISSN 1615-9861

ISSN 1615-9861

4863

of lung injury following pulmonary resection. *European Respiratory Journal,* Vol.15,

*England Journal of Medicine*, Vol.332, No.1, (January 1995), pp.27-37, ISSN 1533-4406

*Journal for Clinicians.* Vol.48, No.1, (January-February 1998), pp.6-29, ISSN 1542-

inflammatory therapy. *Current Opinion in Investigational Drugs*, Vol.6, No.11,

in human lung disease: analysis by two-dimensional electrophoresis.

and other biological predictors related to the postoperative survival rate on nonsmall cell lung cancer. *Lung Cancer,* Vol.33, No.2-3, (August-September 2001),

RJ. (1995). Incidence of local recurrence and second primary tumors in resected stage I lung cancer. *The Journal of Thoracic and Cardiovascular Surgery,* Vol.109, No.1,

injury (ALI) and acute respiratory distress syndrome (ARDS): paradox or

Laser Desorption/Ionization Time-of-Flight Profiling. (in press). *The American* 

Yasui, Y. (2001). Phases of biomarker development for early detection of cancer. *Journal of the National Cancer Institute.* Vol.93, No.14, (July 2001), pp.1054–61. ISSN

Marko-Varga, G. (2003). Human bronchoalveolar lavage: biofluid analysis with special emphasis on sample preparation. *Proteomics,* Vol.3, No.6, (June 2003),

but it still climbs up the mountains. *Proteomics,* Vol.*2*, No.1, (January 2002), pp.3–10,

Detection and identification of human bronchoalveolar lavage proteins using narrow-range immobilized pH gradient DryStrip and the paper bridge sample application method. *Electrophoresis*, Vol.22, No.9, (May 2001), pp.1851–60, ISSN 1522-2683


**3** 

*Italy* 

**Urinary Exosomes for** 

*1Mass Spectrometry and Proteomics,* 

**Protein Biomarker Research** 

Delfin Albert Amal Raj1,2, Immacolata Fiume1, Giovambattista Capasso2 and Gabriella Pocsfalvi1

*2Department of Internal Medicine, Chair of Nephrology, Faculty of Medicine, Second University of Naples, Naples* 

Exosomes represent a distinct class of membrane nanovesicles of endocytic origin that are released to the extracellular microenvironment from diverse cell types under both physiological and pathological conditions. Remarkable roles of exosomes have been revealed in intercellular communication, immune regulation, infection, aging and cancer. Exosomes carry and transfer proteins, nucleic acids and lipids, and are ubiquitous in most biofluids, such as urine, plasma, cerebrospinal fluid, etc. Membrane vesicles secreted by the epithelial cells of the urinary tract hold the promise to be an excellent source of disease relevant cargo proteins. In clinical proteomics urine is one of the most attractive biofluids as it can be obtained non-invasively, in large quantities and is relatively stable. Current isolation methods however are not sufficiently proficient to produce urinary exosomes (UEs) at a purity grade and with reproducibility suitable for downstream LC-MS based quantitative proteomics applications. Consequently urinary exosome based protein

biomarker research today exclusively relies on targeted protein studies (Table 1).

**2. Cell-derived exosomes: Biogenesis, composition and biological role** 

Cells rely on two basic mechanisms for active, vesicle-mediated macromolecular transport through the cellular plasma membrane: exocytosis and endocytosis (Figure 1). Both make use of membrane vesicles for the packaging and trafficking of molecules. While endocytosis is the process in which the extracellular substances enter into a cell without directly passing

LC-MS-based quantitative proteomics workflow.

This chapter describes the current state-of-the-art in exosome research in general and urinary exosomes in particular with a special focus on the potential of UEs in protein biomarker discovery. Recently we have developed an improved isolation/purification method based on double-cushion sucrose/D2O ultracentrifugation (Raj et al., 2011b). The method relies on the solubilization of the major impurities associated with UEs in a carefully selected buffer solution. The new method separates exosomes from the heavier membrane fragments and/or vesicles more efficiently than current protocols and is compatible with

**1. Introduction** 

*Institute of Protein Biochemistry – CNR, Naples* 

bronchoalveolar lavage fluid in asthmatics following segmental antigen challenge. *Molecular & Cellular Proteomics,* Vol.4, No.9, (September 2005), pp.1251-64, ISSN 1535-9484

Zhang, L., Jia, X., Zhang, X., Cao, J., Yang, P., Qiu, C., Shen, Y., Ma, F., Liu, L., Sun, J., Shen, F., Yin, L., Liu, L., Yao, Y., & Lu, H. (2010). Alpha-1 antitrypsin variants in plasma from HIV-infected patients revealed by proteomic and glycoproteomic analysis. *Electrophoresis*, Vol.31, No.20, (October 2010), pp.3437-45, ISSN 1522-2683

### **Urinary Exosomes for Protein Biomarker Research**

Delfin Albert Amal Raj1,2, Immacolata Fiume1, Giovambattista Capasso2 and Gabriella Pocsfalvi1 *1Mass Spectrometry and Proteomics, Institute of Protein Biochemistry – CNR, Naples 2Department of Internal Medicine, Chair of Nephrology, Faculty of Medicine, Second University of Naples, Naples Italy* 

#### **1. Introduction**

48 Proteomics – Human Diseases and Protein Functions

Zhang, L., Jia, X., Zhang, X., Cao, J., Yang, P., Qiu, C., Shen, Y., Ma, F., Liu, L., Sun, J., Shen,

*Electrophoresis*, Vol.31, No.20, (October 2010), pp.3437-45, ISSN 1522-2683

1535-9484

bronchoalveolar lavage fluid in asthmatics following segmental antigen challenge. *Molecular & Cellular Proteomics,* Vol.4, No.9, (September 2005), pp.1251-64, ISSN

F., Yin, L., Liu, L., Yao, Y., & Lu, H. (2010). Alpha-1 antitrypsin variants in plasma from HIV-infected patients revealed by proteomic and glycoproteomic analysis.

> Exosomes represent a distinct class of membrane nanovesicles of endocytic origin that are released to the extracellular microenvironment from diverse cell types under both physiological and pathological conditions. Remarkable roles of exosomes have been revealed in intercellular communication, immune regulation, infection, aging and cancer. Exosomes carry and transfer proteins, nucleic acids and lipids, and are ubiquitous in most biofluids, such as urine, plasma, cerebrospinal fluid, etc. Membrane vesicles secreted by the epithelial cells of the urinary tract hold the promise to be an excellent source of disease relevant cargo proteins. In clinical proteomics urine is one of the most attractive biofluids as it can be obtained non-invasively, in large quantities and is relatively stable. Current isolation methods however are not sufficiently proficient to produce urinary exosomes (UEs) at a purity grade and with reproducibility suitable for downstream LC-MS based quantitative proteomics applications. Consequently urinary exosome based protein biomarker research today exclusively relies on targeted protein studies (Table 1).

> This chapter describes the current state-of-the-art in exosome research in general and urinary exosomes in particular with a special focus on the potential of UEs in protein biomarker discovery. Recently we have developed an improved isolation/purification method based on double-cushion sucrose/D2O ultracentrifugation (Raj et al., 2011b). The method relies on the solubilization of the major impurities associated with UEs in a carefully selected buffer solution. The new method separates exosomes from the heavier membrane fragments and/or vesicles more efficiently than current protocols and is compatible with LC-MS-based quantitative proteomics workflow.

#### **2. Cell-derived exosomes: Biogenesis, composition and biological role**

Cells rely on two basic mechanisms for active, vesicle-mediated macromolecular transport through the cellular plasma membrane: exocytosis and endocytosis (Figure 1). Both make use of membrane vesicles for the packaging and trafficking of molecules. While endocytosis is the process in which the extracellular substances enter into a cell without directly passing

Urinary Exosomes for Protein Biomarker Research 51

Fig. 1. Schematic representation of extracellular vesicles biogenesis. The formation, release and cargo sorting into vesicles destined to be exosomes may involve: i) ESCRT dependent pathway – involving the ubiquitination and ESCRT protein complexes and ii) ESCRT – independent pathway – like ceramide mediated budding. Microvesicles, membrane

particles and exosome like vesicles are secreted by outward budding or fission from the cell

the other hand could also be identified in exosome preparations (van Niel et al., 2006). Protein contents of exosomes from different cells have been mapped by proteomics and the most of the data obtained has been catalogued in Exocarta database (Mathivanan et al.,

Despite their role in immune system modulation (Li et al., 2006), the biological role of exosome secretion remained largely elusive until recent years when Lötvall's group demonstrated that exosomes can transfer genetic information from one cell to another (Valadi et al., 2007, Taylor, 2010). Since then several mechanisms have been proposed to describe exosome-cell interactions: (i) cellular binding via conventional receptor–ligand interactions, similar to cell–cell communication. (ii) attaching/fusing with target cell membrane and (iii) internalization by recipient cells by endocytosis in a transcytotic manner. Besides the physiological roles of exosomes to remove the unwanted cellular debris, recent findings uncover an entirely new and exciting modes of cell–cell communication and paracrine signalling mediated by exosomes (Thery et al., 2002, Camussi et al., 2011). Emerging data shows their involvement in different diseases including inflammation, renal diseases, Alzheimer diseases, aging, bacterial and viral infections, allergies and cancer. Using different sources of tumor-derived exosomes, several groups claim that exosomes can prevent tumor development, induce tumor specific immunity, and provide a possible strategy for therapeutic tumor vaccination reviewed by van Niel et al. (van Niel et al., 2006).

surface.

2009).

through the cell membrane, exocytosis is the primary means of cellular secretion. During both constitutive and regulated exocytosis the secretory-vesicles dock and/or fuse with the plasma membrane. Endocytic pathway (EP), which is primarily responsible for the uptake, trafficking and sorting of internalized proteins has a role in vesicle secretion too (Thery et al., 2002). In the EP, transmembrane proteins are sorted into lumenal vesicles of multivesicular bodies (MVBs). MVBs can have different destinies: they can fuse or mature with lysosomes where the degradation of their protein cargo takes place, or can fuse with the cell membrane to secrete the intraluminal vesicles (ILVs) into the extracellular space. These extracellularly released ILVs are called exosomes (Gruenberg et al., 2004, Keller et al., 2006). During this process, the second inward budding of the endosome membrane results in a positive orientation of the ILVs lipid membrane. Thus when the ILVs are released to the extracellular environment, they have the same orientation as the cell membrane and have been shown to display many of the surface markers from their cell of origin (Thery et al., 2002). The sorting process of membrane proteins during ILV formation is considered to be an active process and thus, exosomal surface proteins seem not to be a plain one-to-one representation of the surface markers for the cell of origin.

While the regulation of endocytic cargo sorting and its delivery to lysosomes have been extensively studied (Williams et al., 2007) relatively less is known about the factors which regulate the formation, the release and the cargo sorting into vesicles destined to be exosomes. The involvement of ubiquitinization and ESCRT (endosomal sorting complex required for transport) protein complexes have been shown by different groups (Gan et al., 2011, Shen et al., 2011). Though, ESCRT-independent mechanisms by means of ceramidemediated budding of exosomes into ILVs within the MVBs have also been identified (Marsh et al., 2008, Trajkovic et al., 2008). Further evidence of ESCRT-independent pathway of ILV formation has come from studying the protein Pmel17, a main component of the c fibrils of pre-melanosomes, which is targeted to intraluminal vesicles of MVBs independently of ubiquitination, ESCRT0 and ESCRTI (Raposo et al., 2001). The most recent model on the formation of ILVs combines the lipid-driven membrane deformation theory with the ESCRT-regulated sorting mechanism (Babst, 2011).

Microvesicles (MVs) are generated by the outward budding and fission of membrane vesicles from the cell surface (Fig. 1) (Lee et al., 2011). MVs (100–1000 nm) are generally bigger in size than exosomes (30-100 nm). Yet due to the analytical difficulties in distinguishing between exosomes and MVs, which are also shed by normal and diseased cells, they are often grouped together.

Many mammalian cells like dendritic, mast, epithelial, neural, stem and hematopoietic cells, reticulocytes, astrocytes, adipocytes, and tumor cells have been reported to release exosomes (Denzer et al., 2000, van Niel et al., 2006). Exosomes purified from the cell culture supernatants are usually heterogeneous in size and contain functional mRNA translatable to proteins, mature microRNAs, lipids and proteins. Proteins of exosomes have been analyzed both by proteomics and targeted immunochemical methods, like Western-blot, FACS with immunolabeling, and immunoelectron microscopy. Protein composition analysis of exosomes shows a rather limited sub-cellular localization for the exosomal proteins. In fact, usually the preparations of exosomes are mostly enriched in cytosolic and membrane proteins and contain less proteins of nuclear, mitochondrial, endoplasmic-reticulum or Golgi-apparatus origin. Secondly, exosomes express a common set of proteins. These are structural components and proteins with a role in exosome biogenesis and trafficking. Cell type specific components which presumably reflect the biological function of the parent cell on

through the cell membrane, exocytosis is the primary means of cellular secretion. During both constitutive and regulated exocytosis the secretory-vesicles dock and/or fuse with the plasma membrane. Endocytic pathway (EP), which is primarily responsible for the uptake, trafficking and sorting of internalized proteins has a role in vesicle secretion too (Thery et al., 2002). In the EP, transmembrane proteins are sorted into lumenal vesicles of multivesicular bodies (MVBs). MVBs can have different destinies: they can fuse or mature with lysosomes where the degradation of their protein cargo takes place, or can fuse with the cell membrane to secrete the intraluminal vesicles (ILVs) into the extracellular space. These extracellularly released ILVs are called exosomes (Gruenberg et al., 2004, Keller et al., 2006). During this process, the second inward budding of the endosome membrane results in a positive orientation of the ILVs lipid membrane. Thus when the ILVs are released to the extracellular environment, they have the same orientation as the cell membrane and have been shown to display many of the surface markers from their cell of origin (Thery et al., 2002). The sorting process of membrane proteins during ILV formation is considered to be an active process and thus, exosomal surface proteins seem not to be a plain one-to-one

While the regulation of endocytic cargo sorting and its delivery to lysosomes have been extensively studied (Williams et al., 2007) relatively less is known about the factors which regulate the formation, the release and the cargo sorting into vesicles destined to be exosomes. The involvement of ubiquitinization and ESCRT (endosomal sorting complex required for transport) protein complexes have been shown by different groups (Gan et al., 2011, Shen et al., 2011). Though, ESCRT-independent mechanisms by means of ceramidemediated budding of exosomes into ILVs within the MVBs have also been identified (Marsh et al., 2008, Trajkovic et al., 2008). Further evidence of ESCRT-independent pathway of ILV formation has come from studying the protein Pmel17, a main component of the c fibrils of pre-melanosomes, which is targeted to intraluminal vesicles of MVBs independently of ubiquitination, ESCRT0 and ESCRTI (Raposo et al., 2001). The most recent model on the formation of ILVs combines the lipid-driven membrane deformation theory with the

Microvesicles (MVs) are generated by the outward budding and fission of membrane vesicles from the cell surface (Fig. 1) (Lee et al., 2011). MVs (100–1000 nm) are generally bigger in size than exosomes (30-100 nm). Yet due to the analytical difficulties in distinguishing between exosomes and MVs, which are also shed by normal and diseased

Many mammalian cells like dendritic, mast, epithelial, neural, stem and hematopoietic cells, reticulocytes, astrocytes, adipocytes, and tumor cells have been reported to release exosomes (Denzer et al., 2000, van Niel et al., 2006). Exosomes purified from the cell culture supernatants are usually heterogeneous in size and contain functional mRNA translatable to proteins, mature microRNAs, lipids and proteins. Proteins of exosomes have been analyzed both by proteomics and targeted immunochemical methods, like Western-blot, FACS with immunolabeling, and immunoelectron microscopy. Protein composition analysis of exosomes shows a rather limited sub-cellular localization for the exosomal proteins. In fact, usually the preparations of exosomes are mostly enriched in cytosolic and membrane proteins and contain less proteins of nuclear, mitochondrial, endoplasmic-reticulum or Golgi-apparatus origin. Secondly, exosomes express a common set of proteins. These are structural components and proteins with a role in exosome biogenesis and trafficking. Cell type specific components which presumably reflect the biological function of the parent cell on

representation of the surface markers for the cell of origin.

ESCRT-regulated sorting mechanism (Babst, 2011).

cells, they are often grouped together.

Fig. 1. Schematic representation of extracellular vesicles biogenesis. The formation, release and cargo sorting into vesicles destined to be exosomes may involve: i) ESCRT dependent pathway – involving the ubiquitination and ESCRT protein complexes and ii) ESCRT – independent pathway – like ceramide mediated budding. Microvesicles, membrane particles and exosome like vesicles are secreted by outward budding or fission from the cell surface.

the other hand could also be identified in exosome preparations (van Niel et al., 2006). Protein contents of exosomes from different cells have been mapped by proteomics and the most of the data obtained has been catalogued in Exocarta database (Mathivanan et al., 2009).

Despite their role in immune system modulation (Li et al., 2006), the biological role of exosome secretion remained largely elusive until recent years when Lötvall's group demonstrated that exosomes can transfer genetic information from one cell to another (Valadi et al., 2007, Taylor, 2010). Since then several mechanisms have been proposed to describe exosome-cell interactions: (i) cellular binding via conventional receptor–ligand interactions, similar to cell–cell communication. (ii) attaching/fusing with target cell membrane and (iii) internalization by recipient cells by endocytosis in a transcytotic manner. Besides the physiological roles of exosomes to remove the unwanted cellular debris, recent findings uncover an entirely new and exciting modes of cell–cell communication and paracrine signalling mediated by exosomes (Thery et al., 2002, Camussi et al., 2011). Emerging data shows their involvement in different diseases including inflammation, renal diseases, Alzheimer diseases, aging, bacterial and viral infections, allergies and cancer. Using different sources of tumor-derived exosomes, several groups claim that exosomes can prevent tumor development, induce tumor specific immunity, and provide a possible strategy for therapeutic tumor vaccination reviewed by van Niel et al. (van Niel et al., 2006).

Urinary Exosomes for Protein Biomarker Research 53

mRNA transcripts encoding specific genes from various regions of the nephron, the collecting duct, the prostate and the bladder have been isolated from urinary exosome preparations (Miranda et al., 2010, Keller et al., 2011). Interestingly, RNA of UEs was found to be protected from RNase degradation which may suggest a functional role for the nucleic acids present in exosome (Keller et al., 2011). In the mRNA sample isolated from the urinary exosomes of prostate cancer patients PCA-3 and TMPRSS2:ERG, two known prostate cancer related biomarkers were detected (Nilsson et al., 2009). Urinary exosomes seem to be particularly rich in miRNAs too. The use of miRNA as diagnostic biomarkers in exosome research is an emerging field due to important potential advantages over standard mRNA

There are over a thousand proteins identified from UE preparations published in the Exocarta (Mathivanan et al., 2009) and the Urinary Exosome Protein Database (Pisitkun et al., 2004) including the six exosome markers commonly used in exosome research (Alix, Tsg101, CD63, CD9, CD81, HSP70). Proteins of UEs show a different profile from that of total urinary proteins but with a high degree of overlap. UEs are enriched in membrane and cytosolic cargo proteins from the different epithelial cells lining the urinary tract (Pisitkun et al., 2004, Gonzales et al., 2009). For clinical biomarker discovery, LC-MS based large-scale quantitative proteomic analysis would be the method of choice. However, at the urinary exosome level it is still a daunting task (Gonzales et al., 2008, Mitchell et al., 2009, Keller et al., 2011). Therefore, protein quantitation and expression analysis has mainly been performed by targeted studies like antibody-based Western blot analysis (Table 1). For this

reason only a few protein biomarker candidates have so far been identified in UEs.

Protocols for collection, storage and processing of human urine for exosome isolation and protein characterization have recently been published (Zhou et al., 2006b). Concerning the isolation of UEs, current methods rely on ultracentrifugation or filtration, or the combination of these two. The majority of the studies use a two-step differential centrifugation protocol developed by Pisitkun et al (Pisitkun et al., 2004). The initial step is a low velocity sequential centrifugation which serves to remove cells and cellular debris (urinary sediment) from urine, leaving the exosomes in the supernatant. The second step is the ultracentrifugation for 1h to overnight of the supernatant at 100,000-200,000g velocity to sediment exosomes. The major short comings of this process are the high level of contamination from uromodulin (see later) and the lack of separation of exosomes from the

To obtain higher purity grade UEs, the crude preparation obtained by the two-step differential centrifugation method can be further processed using the sucrose gradient or the sucrose cushion centrifugation. Sucrose gradient centrifugation can be performed on linear or step gradients typically using sucrose concentrations between 2.0 M – 0.25 M (Keller et al., 2007, Hogan et al., 2009, Simpson et al., 2009, Mathivanan et al., 2010). Instead of gradient, a small density cushion typically composed of 30% sucrose in deuterium oxide (D2O), can also be employed for the purification of UEs (Mitchell et al., 2009, Simpson et al., 2009, Welton et al., 2010). In the sucrose cushion, formation of a mini density gradient takes place in the range of 1.10-1.18 g/cm3. This range was shown to be suitable to enrich and purify exosomes preventing vesicle aggregation that pelleting could cause. Sucrose gradient and cushion centrifugations thus allow a better separation of exosomes from the vesicles of different densities respect to the differential centrifugation

(Li et al., 2010).

**3.2 Isolation and purification** 

other MVs and membrane particles.

#### **3. Urinary exosomes**

#### **3.1 mRNA, miRNA and protein biomarkers in urinary exosomes**

Urinary exosomes originate from those ILVs that are shed into the urinary space by the fusion of the outer membrane of MVBs with the apical plasma membrane of cells lining the urinary tract, including glomerular podocytes, renal tubule cells, and bladder. The number, and the physical, chemical and biological properties of UEs may change over time in association with disorders that affect the urinary system. Respect to the total urine sample, UEs result in a remarkable enrichment of low-abundance biomolecules with potentially high diagnostic value regarding the physiological and pathological state of the renal system. Therefore, it is not surprising that there is a great interest in the use of UEs as a novel biomarker source for early disease detection, classification, prediction severity, outcome and response to treatment. Since the first publication on proteomic profiling of UEs by the group of Knepper (Pisitkun et al., 2004), an increasing number of articles with keywords "exosome and urine" are to be found in the PubMed database. The principal aim of urinary exosome research today is to discover mRNA, microRNA and protein biomarkers.


AKI - acute kidney injury

FSGS - focal segmental glomerulosclerosis

BC - bladder cancer

PC - prostate cancer

I/R - renal ischemia/reperfusion

GKD - glomural kidney disease

NSCL – non-small cell lung cancer

Table 1. Different isolation/purification, protein separation, identification and quantitation methods used in urinary exosome related targeted protein biomarker studies.

Urinary exosomes originate from those ILVs that are shed into the urinary space by the fusion of the outer membrane of MVBs with the apical plasma membrane of cells lining the urinary tract, including glomerular podocytes, renal tubule cells, and bladder. The number, and the physical, chemical and biological properties of UEs may change over time in association with disorders that affect the urinary system. Respect to the total urine sample, UEs result in a remarkable enrichment of low-abundance biomolecules with potentially high diagnostic value regarding the physiological and pathological state of the renal system. Therefore, it is not surprising that there is a great interest in the use of UEs as a novel biomarker source for early disease detection, classification, prediction severity, outcome and response to treatment. Since the first publication on proteomic profiling of UEs by the group of Knepper (Pisitkun et al., 2004), an increasing number of articles with keywords "exosome and urine" are to be found in the PubMed database. The principal aim of urinary exosome research today is to discover mRNA, microRNA and protein

Table 1. Different isolation/purification, protein separation, identification and quantitation

methods used in urinary exosome related targeted protein biomarker studies.

**3.1 mRNA, miRNA and protein biomarkers in urinary exosomes** 

**3. Urinary exosomes** 

biomarkers.

AKI - acute kidney injury

I/R - renal ischemia/reperfusion GKD - glomural kidney disease NSCL – non-small cell lung cancer

BC - bladder cancer PC - prostate cancer

FSGS - focal segmental glomerulosclerosis

mRNA transcripts encoding specific genes from various regions of the nephron, the collecting duct, the prostate and the bladder have been isolated from urinary exosome preparations (Miranda et al., 2010, Keller et al., 2011). Interestingly, RNA of UEs was found to be protected from RNase degradation which may suggest a functional role for the nucleic acids present in exosome (Keller et al., 2011). In the mRNA sample isolated from the urinary exosomes of prostate cancer patients PCA-3 and TMPRSS2:ERG, two known prostate cancer related biomarkers were detected (Nilsson et al., 2009). Urinary exosomes seem to be particularly rich in miRNAs too. The use of miRNA as diagnostic biomarkers in exosome research is an emerging field due to important potential advantages over standard mRNA (Li et al., 2010).

There are over a thousand proteins identified from UE preparations published in the Exocarta (Mathivanan et al., 2009) and the Urinary Exosome Protein Database (Pisitkun et al., 2004) including the six exosome markers commonly used in exosome research (Alix, Tsg101, CD63, CD9, CD81, HSP70). Proteins of UEs show a different profile from that of total urinary proteins but with a high degree of overlap. UEs are enriched in membrane and cytosolic cargo proteins from the different epithelial cells lining the urinary tract (Pisitkun et al., 2004, Gonzales et al., 2009). For clinical biomarker discovery, LC-MS based large-scale quantitative proteomic analysis would be the method of choice. However, at the urinary exosome level it is still a daunting task (Gonzales et al., 2008, Mitchell et al., 2009, Keller et al., 2011). Therefore, protein quantitation and expression analysis has mainly been performed by targeted studies like antibody-based Western blot analysis (Table 1). For this reason only a few protein biomarker candidates have so far been identified in UEs.

#### **3.2 Isolation and purification**

Protocols for collection, storage and processing of human urine for exosome isolation and protein characterization have recently been published (Zhou et al., 2006b). Concerning the isolation of UEs, current methods rely on ultracentrifugation or filtration, or the combination of these two. The majority of the studies use a two-step differential centrifugation protocol developed by Pisitkun et al (Pisitkun et al., 2004). The initial step is a low velocity sequential centrifugation which serves to remove cells and cellular debris (urinary sediment) from urine, leaving the exosomes in the supernatant. The second step is the ultracentrifugation for 1h to overnight of the supernatant at 100,000-200,000g velocity to sediment exosomes. The major short comings of this process are the high level of contamination from uromodulin (see later) and the lack of separation of exosomes from the other MVs and membrane particles.

To obtain higher purity grade UEs, the crude preparation obtained by the two-step differential centrifugation method can be further processed using the sucrose gradient or the sucrose cushion centrifugation. Sucrose gradient centrifugation can be performed on linear or step gradients typically using sucrose concentrations between 2.0 M – 0.25 M (Keller et al., 2007, Hogan et al., 2009, Simpson et al., 2009, Mathivanan et al., 2010). Instead of gradient, a small density cushion typically composed of 30% sucrose in deuterium oxide (D2O), can also be employed for the purification of UEs (Mitchell et al., 2009, Simpson et al., 2009, Welton et al., 2010). In the sucrose cushion, formation of a mini density gradient takes place in the range of 1.10-1.18 g/cm3. This range was shown to be suitable to enrich and purify exosomes preventing vesicle aggregation that pelleting could cause. Sucrose gradient and cushion centrifugations thus allow a better separation of exosomes from the vesicles of different densities respect to the differential centrifugation

Urinary Exosomes for Protein Biomarker Research 55

**Tryptic digestion**

Fig. 2. Scheme of the MudPIT based 4-plex iTRAQ quantitative analysis comparing the

**4.1 A novel isolation/purification method based on uromodulin solubilization and** 

The urinary exosome isolation/purification method which we have recently developed (Raj et al., 2011b) employs a double-cushion ultracentrifugation step performed in a carefully chosen buffer solution. Respect to other ultracentrifugation based methods which generally use a PBS buffer (150 mM NaCl at pH 7.2) the novel method employs a solubilising buffer composed of 20 mM Tris at pH 8.6. We have found that Tris buffer efficiently solubilizes uromodulin aggregates, keeps uromodulin in solution and does not lyses exosomes. This is in accordance with a previous in vitro study on uromodulin solubility which underlines the importance of alkaline pH, low sodium and calcium concentrations and sample dilution to prevent the formation of uromodulin aggregates (Kobayashi et al., 2001). After solubilizing the pellet obtained in the differential ultracentrifugation step, double-cushion ultracentrifugation is performed. The double-cushion is made of sucrose 1 M and sucrose 2 M prepared in 20 mM Tris pH 8.6 in D2O and subsequently under layered below the sample in the centrifuge tube. This step was found to considerably improve the separation of

cushion ultracentrifugation methods. Simultaneously, we compared samples obtained from a single person with a pool of healthy volunteers divided into two age groups (25-50 years and 50-70 years) in order to study feasibility of analysis of single patient versus pooled

double-cushion ultracentrifugation method with that of single-cushion.

samples in the discovery phase of protein biomarker research.

exosomes from the heavier vesicles and/or membrane fragments.

**4.2 Analysis of urinary vesicles at the various steps of isolation/purification** 

Exosomes were purified from pooled urine samples of ten healthy donors and separated on 4-12% gradient polyacrylamide gel then stained with colloidal Coomassie blue. SDS-PAGE analysis at the various phases of the isolation/purification process is shown in Figure 3. Total urinary protein profiles before (Figure 3.A, Lane 1) and after exosome depletion (Figure 3.A, Lane 2) do not markedly differ from each other and show the typical pattern of

**iTRAQ 115**

**iTRAQ 114**

**iTRAQ 116**

**Combine labeled**

 **peptides** **MudPIT**

**iTRAQ labeling**

**iTRAQ 117**

Urine samples Crude exosomes Proteins Tryptic peptides Labeled peptides

**Lysis, reduction and alkylation**

**Differential centrifugation**

**Double-cushion** 

**Single-cushion** 

**centrifugation**

**double-cushion ultracentrifugation** 

**centrifugation**

Vesicles

LC-MS/MS

Reversed phase chromatography

**SCX** . . . . . . . . . . . .

> **Quantify** iTRAQ reporter ions

**Identify** MS/MS fragmentation

Data normalisation and statistical analysis Protein ID and quantitation by Mascot

method, however it does not seem to eliminate the problem of the co-purifying uromodulin (Hogan et al., 2009).

Filtration-based protocols generally use polyether sulfone nano-membranes in a spin concentrator to isolate urinary exosomes (Cheruvanky et al., 2007). The method is simple, fast and is capable to isolate UEs from small volumes of urine (0.5–10 mL). Therefore it is very promising, especially for mRNA and miRNA based exosome biomarker research. Drawbacks of this method for protein biomarker research are the low yield and the high level of contamination caused by urinary proteins binding to the filter. To overcome this, recently a low protein binding membrane (hydrophilized polyvinylidene difluoride) has been used to isolate urinary exosomes (Merchant et al., 2010).

#### **3.3 The uromodulin problem**

Current methods are characterized by a high and variable level of uromodulin contamination (Hogan et al., 2009, Fernandez-Llama et al., 2010, Rood et al., 2010). Uromodulin, also referred to as Tamm–Horsfall glycoprotein, is a major glycoprotein produced by kidney cells. Uromodulin assembles into intracellular filaments in urine (Porter et al., 1955, Schaeffer et al., 2009). The filaments have an average width and length of 100 Ǻ and 2.5 µm, respectively and tend to form a three-dimensional matrix with pores as shown by electron microscopy (Porter et al., 1955). This filament network traps exosomes and prevents their efficient isolation and purification by traditional methods. The uromodulin problem is one of the bottle neck of UE protein research because it considerably reduces sample yield and reproducibility (Fernandez-Llama et al., 2010). In order to facilitate the removal of high molecular weight aggregates recently, dithiothreitol (DTT) was applied to reduce the intermolecular disulfide bonds of uromodulin (Pisitkun et al., 2004, Fernandez-Llama et al., 2010). Treatment with DTT result in a higher yield of urinary exosomes. Notwithstanding it does not solve the problem efficiently. For this reason, urinary exosome samples prepared by the current methods are far from being ideal for quantitative proteomic analysis.

#### **4. Interfacing urinary exosome isolation/purification and lysis with quantitative proteomics for protein biomarker research**

Biomarkers support the diagnosis and medical management of various disorders. The remarkable progress made in proteomic technologies in the past decade have enabled researchers to consider designing studies to identify diagnostic and therapeutic biomarkers by analyzing complex proteome samples using unbiased mass spectrometry based methods. In urinary exosome research this has been hampered by the high and variable concentration of uromodulin causing low sample quantity, quality and low reproducibility. To meet the need of a global protein biomarker discovery platform we have set-up new protocols for the isolation/purification and also for the lysis and subsequent solubilization of membrane proteins. Paragraph 4.1 describes a novel urinary exosome preparation called doublecushion ultracentrifugation method and paragraph 4.2 shows its compatibility with downstream analysis.

We have employed a multiplex quantitative proteomics method, iTRAQ (isobaric Tagging for Relative and Absolute protein Quantification), in conjunction with multidimensional chromatography, followed by tandem mass spectrometry (MS/MS), to measure relative differences in the protein composition of urinary exosome samples (Figure 2). The aim of this work was to compare the protein content of UEs obtained by single- and double-

method, however it does not seem to eliminate the problem of the co-purifying

Filtration-based protocols generally use polyether sulfone nano-membranes in a spin concentrator to isolate urinary exosomes (Cheruvanky et al., 2007). The method is simple, fast and is capable to isolate UEs from small volumes of urine (0.5–10 mL). Therefore it is very promising, especially for mRNA and miRNA based exosome biomarker research. Drawbacks of this method for protein biomarker research are the low yield and the high level of contamination caused by urinary proteins binding to the filter. To overcome this, recently a low protein binding membrane (hydrophilized polyvinylidene difluoride) has

Current methods are characterized by a high and variable level of uromodulin contamination (Hogan et al., 2009, Fernandez-Llama et al., 2010, Rood et al., 2010). Uromodulin, also referred to as Tamm–Horsfall glycoprotein, is a major glycoprotein produced by kidney cells. Uromodulin assembles into intracellular filaments in urine (Porter et al., 1955, Schaeffer et al., 2009). The filaments have an average width and length of 100 Ǻ and 2.5 µm, respectively and tend to form a three-dimensional matrix with pores as shown by electron microscopy (Porter et al., 1955). This filament network traps exosomes and prevents their efficient isolation and purification by traditional methods. The uromodulin problem is one of the bottle neck of UE protein research because it considerably reduces sample yield and reproducibility (Fernandez-Llama et al., 2010). In order to facilitate the removal of high molecular weight aggregates recently, dithiothreitol (DTT) was applied to reduce the intermolecular disulfide bonds of uromodulin (Pisitkun et al., 2004, Fernandez-Llama et al., 2010). Treatment with DTT result in a higher yield of urinary exosomes. Notwithstanding it does not solve the problem efficiently. For this reason, urinary exosome samples prepared by the current methods are far from being

**4. Interfacing urinary exosome isolation/purification and lysis with** 

Biomarkers support the diagnosis and medical management of various disorders. The remarkable progress made in proteomic technologies in the past decade have enabled researchers to consider designing studies to identify diagnostic and therapeutic biomarkers by analyzing complex proteome samples using unbiased mass spectrometry based methods. In urinary exosome research this has been hampered by the high and variable concentration of uromodulin causing low sample quantity, quality and low reproducibility. To meet the need of a global protein biomarker discovery platform we have set-up new protocols for the isolation/purification and also for the lysis and subsequent solubilization of membrane proteins. Paragraph 4.1 describes a novel urinary exosome preparation called doublecushion ultracentrifugation method and paragraph 4.2 shows its compatibility with

We have employed a multiplex quantitative proteomics method, iTRAQ (isobaric Tagging for Relative and Absolute protein Quantification), in conjunction with multidimensional chromatography, followed by tandem mass spectrometry (MS/MS), to measure relative differences in the protein composition of urinary exosome samples (Figure 2). The aim of this work was to compare the protein content of UEs obtained by single- and double-

**quantitative proteomics for protein biomarker research** 

uromodulin (Hogan et al., 2009).

**3.3 The uromodulin problem** 

ideal for quantitative proteomic analysis.

downstream analysis.

been used to isolate urinary exosomes (Merchant et al., 2010).

Fig. 2. Scheme of the MudPIT based 4-plex iTRAQ quantitative analysis comparing the double-cushion ultracentrifugation method with that of single-cushion.

cushion ultracentrifugation methods. Simultaneously, we compared samples obtained from a single person with a pool of healthy volunteers divided into two age groups (25-50 years and 50-70 years) in order to study feasibility of analysis of single patient versus pooled samples in the discovery phase of protein biomarker research.

#### **4.1 A novel isolation/purification method based on uromodulin solubilization and double-cushion ultracentrifugation**

The urinary exosome isolation/purification method which we have recently developed (Raj et al., 2011b) employs a double-cushion ultracentrifugation step performed in a carefully chosen buffer solution. Respect to other ultracentrifugation based methods which generally use a PBS buffer (150 mM NaCl at pH 7.2) the novel method employs a solubilising buffer composed of 20 mM Tris at pH 8.6. We have found that Tris buffer efficiently solubilizes uromodulin aggregates, keeps uromodulin in solution and does not lyses exosomes. This is in accordance with a previous in vitro study on uromodulin solubility which underlines the importance of alkaline pH, low sodium and calcium concentrations and sample dilution to prevent the formation of uromodulin aggregates (Kobayashi et al., 2001). After solubilizing the pellet obtained in the differential ultracentrifugation step, double-cushion ultracentrifugation is performed. The double-cushion is made of sucrose 1 M and sucrose 2 M prepared in 20 mM Tris pH 8.6 in D2O and subsequently under layered below the sample in the centrifuge tube. This step was found to considerably improve the separation of exosomes from the heavier vesicles and/or membrane fragments.

#### **4.2 Analysis of urinary vesicles at the various steps of isolation/purification**

Exosomes were purified from pooled urine samples of ten healthy donors and separated on 4-12% gradient polyacrylamide gel then stained with colloidal Coomassie blue. SDS-PAGE analysis at the various phases of the isolation/purification process is shown in Figure 3. Total urinary protein profiles before (Figure 3.A, Lane 1) and after exosome depletion (Figure 3.A, Lane 2) do not markedly differ from each other and show the typical pattern of

Urinary Exosomes for Protein Biomarker Research 57

Western blot analysis was performed to monitor the enrichment in exosomes and the reproducibility of sample preparation by the double-cushion ultracentrifugation. Exosomal proteins were separated on 4-12% gradient SDS-PAGE and electro blotted to PVDF membrane. Blots were probed with antibodies against two known exosome markers Alix and TSG101, together with NKCC2 a renal sodium transporter known to be present in urinary exosomes (Figure 4.). The enrichment of exosomes is excellent in the samples prepared by the doublecushion (Figure 4., lane 4-6) respect to the starting and exosome depleted urine samples (Figure 4., lane 2-3) and also to the sample prepared by the differential centrifugation method (Figure 4., lane 1). Importantly a very high degree of reproducibility was achieved in three

Fig. 4. Western blot analysis of urinary exosomes prepared by two different methods. Lane 1- Exosome purified by differential centrifugation; Lane 2- Total urine; Lane 3- Exosome depleted urine; Lane 4-6 – Exosomes purified in three independent experiments from pooled urine samples of ten healthy volunteers by the double-cushion method.

Exosome-like vesicles isolated from culture supernatant are limited by a lipid bilayer and in literature often described as saucer- or cup-shaped particles. Urinary exosomes isolated by the double-cushion ultracentrifugation method have a similar morphology as single cell line derived exosomes. The transmission electron microscopy (TEM) image shows (Figure 5) that diameters of the vesicles purified in the 1 M fraction are between 30 and 80 nm. Interestingly, the shape of the exosomes appeared to be nearly spherical with only a few elongated or cup-shaped specimens. After the double-cushion ultracentrifugation the sample is basically free from the long uromodulin filaments known to contaminate UEs

Fig. 5. Transmission electron microscopy image of urinary exosomes isolated and purified by the double-cushion ultracentrifugation method (1 M fraction). The image shows the typical morphology and size distribution of the vesicles. Frame shows the enlarged image

(central) and the arrow shows a single vesicle enlarged on the right image.

independent urinary exosome preparations (Figure 4., lanes 4-6).

prepared by traditional methods.

Fig. 3. SDS PAGE analyses A) at the different stages of urinary exosome isolation/purification through the double-cushion (lanes 1-7) and the single-cushion (lane 9) methods and, B) of the 1 M and 2 M sucrose fractions obtained after the double-cushion ultracentrifugation method (major proteins identified by *in-gel* digestion proteomics are indicated next to the band). Lanes in Figure A as follow: 1- Total urine; 2- Exosome depleted urine; 3- Crude exosome fraction after differential centrifugation; 4- 15,000g pellet; 5- 15,000g supernatant; 6- Purified exosomes (1 M sucrose fraction); 7- 2 M sucrose fraction; M- Protein molecular weight markers (kDa); 9- Urinary exosomes prepared by the single sucrose/D2O cushion method. Lanes in Figure B are as follow: 1- 1 M sucrose fraction and 2- 2 M sucrose fraction and M- Protein molecular weight markers (kDa).

the major urinary proteins, like albumin, various IgG chains, uromodulin etc. After the twostep differential centrifugation the crude exosome pellet (Figure 3.A, lane 3) still contains a considerable amount of contaminating urinary proteins and in particular uromodulin at 85 kDa. These are in part removed after the solubilization step by low-speed centrifugation (Figure 3.A, lane 4-5) and, in part by the double-cushion ultracentrifugation. The later yields two fractions: the 1 M sucrose fraction which contains the exosome vesicles (Figure 3.A, lane 6) and the 2 M fraction which contains vesicles heavier than exosomes (Figure 3.A, lane 7). The efficiency of the uromodulin removal by the double-cushion sucrose ultracentrifugation methods can be appreciated by comparing the 1 M fraction (Figure 3.A, lane 6) with the crude exosome fraction (Figure 3.A, lane 3) and with the exosomes purified by the singlecushion method (Figure 3.A, lane 9). In Figure 3.B SDS-PAGE image of the two vesicle containing fractions, 1 M (lane 1) and 2 M (lane 2) are shown together with the major proteins identified in the gel bands. It is of note that not only the protein pattern but also the proteins identified in the major SDS-PAGE bands were found to be different, indicating the presence of two different types of vesicles in the two fractions. Semenogelin 1 and semenogelin 2 and olfactomedin for example have previously been identified in prostasomes, i.e. the secretory particles in human seminal fluid (Utleg et al., 2003). Therefore it is plausible to presume that the 2 M sucrose fraction contains heavier vesicles, like urinary secreted prostasomes.

Fig. 3. SDS PAGE analyses A) at the different stages of urinary exosome

2- 2 M sucrose fraction and M- Protein molecular weight markers (kDa).

secreted prostasomes.

isolation/purification through the double-cushion (lanes 1-7) and the single-cushion (lane 9) methods and, B) of the 1 M and 2 M sucrose fractions obtained after the double-cushion ultracentrifugation method (major proteins identified by *in-gel* digestion proteomics are indicated next to the band). Lanes in Figure A as follow: 1- Total urine; 2- Exosome depleted

5- 15,000g supernatant; 6- Purified exosomes (1 M sucrose fraction); 7- 2 M sucrose fraction; M- Protein molecular weight markers (kDa); 9- Urinary exosomes prepared by the single sucrose/D2O cushion method. Lanes in Figure B are as follow: 1- 1 M sucrose fraction and

the major urinary proteins, like albumin, various IgG chains, uromodulin etc. After the twostep differential centrifugation the crude exosome pellet (Figure 3.A, lane 3) still contains a considerable amount of contaminating urinary proteins and in particular uromodulin at 85 kDa. These are in part removed after the solubilization step by low-speed centrifugation (Figure 3.A, lane 4-5) and, in part by the double-cushion ultracentrifugation. The later yields two fractions: the 1 M sucrose fraction which contains the exosome vesicles (Figure 3.A, lane 6) and the 2 M fraction which contains vesicles heavier than exosomes (Figure 3.A, lane 7). The efficiency of the uromodulin removal by the double-cushion sucrose ultracentrifugation methods can be appreciated by comparing the 1 M fraction (Figure 3.A, lane 6) with the crude exosome fraction (Figure 3.A, lane 3) and with the exosomes purified by the singlecushion method (Figure 3.A, lane 9). In Figure 3.B SDS-PAGE image of the two vesicle containing fractions, 1 M (lane 1) and 2 M (lane 2) are shown together with the major proteins identified in the gel bands. It is of note that not only the protein pattern but also the proteins identified in the major SDS-PAGE bands were found to be different, indicating the presence of two different types of vesicles in the two fractions. Semenogelin 1 and semenogelin 2 and olfactomedin for example have previously been identified in prostasomes, i.e. the secretory particles in human seminal fluid (Utleg et al., 2003). Therefore it is plausible to presume that the 2 M sucrose fraction contains heavier vesicles, like urinary

urine; 3- Crude exosome fraction after differential centrifugation; 4- 15,000g pellet;

Western blot analysis was performed to monitor the enrichment in exosomes and the reproducibility of sample preparation by the double-cushion ultracentrifugation. Exosomal proteins were separated on 4-12% gradient SDS-PAGE and electro blotted to PVDF membrane. Blots were probed with antibodies against two known exosome markers Alix and TSG101, together with NKCC2 a renal sodium transporter known to be present in urinary exosomes (Figure 4.). The enrichment of exosomes is excellent in the samples prepared by the doublecushion (Figure 4., lane 4-6) respect to the starting and exosome depleted urine samples (Figure 4., lane 2-3) and also to the sample prepared by the differential centrifugation method (Figure 4., lane 1). Importantly a very high degree of reproducibility was achieved in three independent urinary exosome preparations (Figure 4., lanes 4-6).

Fig. 4. Western blot analysis of urinary exosomes prepared by two different methods. Lane 1- Exosome purified by differential centrifugation; Lane 2- Total urine; Lane 3- Exosome depleted urine; Lane 4-6 – Exosomes purified in three independent experiments from pooled urine samples of ten healthy volunteers by the double-cushion method.

Exosome-like vesicles isolated from culture supernatant are limited by a lipid bilayer and in literature often described as saucer- or cup-shaped particles. Urinary exosomes isolated by the double-cushion ultracentrifugation method have a similar morphology as single cell line derived exosomes. The transmission electron microscopy (TEM) image shows (Figure 5) that diameters of the vesicles purified in the 1 M fraction are between 30 and 80 nm. Interestingly, the shape of the exosomes appeared to be nearly spherical with only a few elongated or cup-shaped specimens. After the double-cushion ultracentrifugation the sample is basically free from the long uromodulin filaments known to contaminate UEs prepared by traditional methods.

Fig. 5. Transmission electron microscopy image of urinary exosomes isolated and purified by the double-cushion ultracentrifugation method (1 M fraction). The image shows the typical morphology and size distribution of the vesicles. Frame shows the enlarged image (central) and the arrow shows a single vesicle enlarged on the right image.

Urinary Exosomes for Protein Biomarker Research 59

peptides and the proteins from which they originate by Mascot iTRAQ 4-plex quantification method. Proteins which were quantified with a minimum of two unique peptides and

More than hundred proteins were quantified in the iTRAQ analysis. Table 3. shows the weighted median ratios of the first 25 proteins ranked by Mascot protein score. Expression level of the major proteins isolated and purified by the double-cushion method are different from those purified with the single-cushion protocol (Table 3., ratio 117/114). In particular, cytoskeletal proteins (cubulin, megalin, actin, cofillin, moesin, tubulin, etc.) seem to be less abundant in the sample. They may be due to heterogeneous constituents of the cytoskeleton filaments present in urine which co-purify with the UEs in traditional methods. On the other hand a marked enrichment was observed in proteins which are related to the VPS4 complex of ESCRT machinery (IST1, VPS4A, and VPS4B), its associated proteins (CHM2A, CHMP5,

AMPN\_HUMAN Aminopeptidase 1315 0.845 1.306 0.237 IST1\_HUMAN IST1 homolog 656 1.037 0.706 0.406 ACTB\_HUMAN Actin, cytoplasmic 430 1.408 1.631 0.733 DPEP1\_HUMAN Dipeptidase 1 370 0.857 0.700 0.430

CHM2A\_HUMAN Charged multivesicular body protein 2a 341 1.242 1.493 0.595 UROM\_HUMAN Uromodulin 282 0.958 0.578 22.450 CHMP5\_HUMAN Charged multivesicular body protein 5 279 1.082 0.569 0.161 RS27A\_HUMAN Ubiquitin-40S ribosomal protein S27a 275 1.051 0.764 0.445 GGT1\_HUMAN Gamma-glutamyltranspeptidase 267 0.840 1.045 0.342 NEP\_HUMAN Neprilysin 258 0.897 0.995 0.381 EZRI\_HUMAN Ezrin 254 1.245 1.093 0.776 ANX11\_HUMAN Annexin A11 252 1.393 0.435 0.601 PSCA\_HUMAN Prostate stem cell antigen 231 1.415 5.214 0.345 HSP7C\_HUMAN Heat shock cognate 71 kDa protein 208 1.186 0.990 0.428

CDC42\_HUMAN Cell division control protein 42 homolog 208 1.044 1.773 0.190

CHM4B\_HUMAN Charged multivesicular body protein 4b 181 1.308 0.923 0.559 POTEF\_HUMAN POTE ankyrin domain family member F 176 1.394 1.404 0.510 DPP4\_HUMAN Dipeptidyl peptidase 4 175 0.931 0.837 0.320 AQP1\_HUMAN Aquaporin-1 167 0.898 0.679 0.573 THY1\_HUMAN Thy-1 membrane glycoprotein 151 1.316 2.332 0.354 MUC1\_HUMAN Mucin-1 145 0.945 0.523 0.743 PROM1\_HUMAN Prominin-1 140 1.014 0.655 0.500 Table 3. The weighted median ratios of the 25 top-ranking proteins in the MudPIT based 4 plex iTRAQ experiment. 114, 115, 116 and 117 indicate sample-labeling by iTRAQ according

**score** 

**115/114 116/114 117/114** 

350 1.150 0.742 0.473

231 1.093 0.684 0.500

195 1.063 0.601 0.500

p<0.05 significance threshold using MudPIT scoring have been considered.

**UniProt ID Protein Name Mascot** 

VPS4A\_HUMAN Vacuolar protein sorting-associated protein

PDC6I\_HUMAN Programmed cell death 6-interacting

VPS4B\_HUMAN Vacuolar protein sorting-associated protein

protein

4B

to Table 2.

4A

#### **5. Quantitative proteomics of urinary exosomes for protein biomarker discovery**

Recently, we have developed protocols for lysis, protein extraction and *in-solution* digestion of UEs for MudPIT application to quantitative proteomics (Raj et al., 2011a). For the solubilization of exosomal membrane proteins the use of an acid cleavable detergent was found to be particularly useful. In a preliminary study four exosomal protein samples were prepared in parallel (Table 2) according to single- (sample 4) and double-cushion protocols (sample 1) from a pooled urine sample of 20 healthy donors (male, age group 25-45 years). Effects of age (sample 2) and sample pooling (sample 3) on the protein expression were also monitored in the same experiment.


Table 2. Samples analysed by *in-solution* digestion based MudPIT proteomics and iTRAQ labeling.

The 4-plex iTRAQ method (Ross et al., 2004) based on covalent labeling of the N-terminus and side-chain amines of peptides with four tags of varying mass was used for the protein quantitation (Figure 2.).

Protein samples were denatured, reduced, alkylated, enzymatically digested by trypsin and then labeled according to the manufacturer's protocol (iTRAQ reagent kit, Applied Biosystems). After iTRAQ labeling equal amounts of each sample (100 µg) were mixed, vacuum dried, detergent was acid cleaved and the resulting sample was desalted. The purified sample was then separated by two-dimensional HPLC. For strong cation-exchange (SCX) chromatography, in the first dimension, the following conditions were used: 95% solvent A (20% acetonitrile, 0.05% formic acid) and 5% solvent B (20% acetonitrile, 0.05% formic acid, 500mM KCl) for 3 min, solvent B ramped up to 90% in 40 min and maintained at 100% for 7 min. 47 fractions were collected between 0-55 min. Fractions were further separated in the second dimension on a reversed phase monolithic nano column using the following conditions: 95% solvent C (2% acetonitrile, 0.1% formic acid) and 5% solvent D (98% acetonitrile, 0.1% formic acid) for 5 min, ramp to 50% solvent D in 90 min and in 6 sec to 98% solvent D for 10 min. Eluting peptides were analyzed online by a QTOF type of tandem mass spectrometer (Qstar Elite) in an information dependent acquisition mode which facilitates both the protein identification and the multiplex quantitative analysis of the four samples. Tandem mass spectra were extracted and peak lists were generated by Analyst QS 2.0 software using the default parameters. Peak lists containing all acquired MS/MS spectra were searched against SwissProt 2010\_09 (519348 sequences) database using Mascot Server (version 2.2) with trypsin specificity and allowing for up to one missed cleavage. iTRAQ at lysine residue and the N termini of the peptides and carbamidomethylation of cysteines were considered as fixed modifications whereas oxidations of methionine and iTRAQ at tyrosine residues were set as possible variable modifications. Mass tolerance was set to 50 ppm for precursor and to 0.1 Da for fragment ions, respectively. Low molecular mass reporter ions were used to relatively quantify the

Recently, we have developed protocols for lysis, protein extraction and *in-solution* digestion of UEs for MudPIT application to quantitative proteomics (Raj et al., 2011a). For the solubilization of exosomal membrane proteins the use of an acid cleavable detergent was found to be particularly useful. In a preliminary study four exosomal protein samples were prepared in parallel (Table 2) according to single- (sample 4) and double-cushion protocols (sample 1) from a pooled urine sample of 20 healthy donors (male, age group 25-45 years). Effects of age (sample 2) and sample pooling (sample 3) on the protein expression were also

1 25-45 20 Double-cushion iTRAQ-114 2 50-70 20 Double-cushion iTRAQ-115 3 43 1 Double-cushion iTRAQ-116 4 25-45 20 Single-cushion iTRAQ-117

Table 2. Samples analysed by *in-solution* digestion based MudPIT proteomics and iTRAQ

The 4-plex iTRAQ method (Ross et al., 2004) based on covalent labeling of the N-terminus and side-chain amines of peptides with four tags of varying mass was used for the protein

Protein samples were denatured, reduced, alkylated, enzymatically digested by trypsin and then labeled according to the manufacturer's protocol (iTRAQ reagent kit, Applied Biosystems). After iTRAQ labeling equal amounts of each sample (100 µg) were mixed, vacuum dried, detergent was acid cleaved and the resulting sample was desalted. The purified sample was then separated by two-dimensional HPLC. For strong cation-exchange (SCX) chromatography, in the first dimension, the following conditions were used: 95% solvent A (20% acetonitrile, 0.05% formic acid) and 5% solvent B (20% acetonitrile, 0.05% formic acid, 500mM KCl) for 3 min, solvent B ramped up to 90% in 40 min and maintained at 100% for 7 min. 47 fractions were collected between 0-55 min. Fractions were further separated in the second dimension on a reversed phase monolithic nano column using the following conditions: 95% solvent C (2% acetonitrile, 0.1% formic acid) and 5% solvent D (98% acetonitrile, 0.1% formic acid) for 5 min, ramp to 50% solvent D in 90 min and in 6 sec to 98% solvent D for 10 min. Eluting peptides were analyzed online by a QTOF type of tandem mass spectrometer (Qstar Elite) in an information dependent acquisition mode which facilitates both the protein identification and the multiplex quantitative analysis of the four samples. Tandem mass spectra were extracted and peak lists were generated by Analyst QS 2.0 software using the default parameters. Peak lists containing all acquired MS/MS spectra were searched against SwissProt 2010\_09 (519348 sequences) database using Mascot Server (version 2.2) with trypsin specificity and allowing for up to one missed cleavage. iTRAQ at lysine residue and the N termini of the peptides and carbamidomethylation of cysteines were considered as fixed modifications whereas oxidations of methionine and iTRAQ at tyrosine residues were set as possible variable modifications. Mass tolerance was set to 50 ppm for precursor and to 0.1 Da for fragment ions, respectively. Low molecular mass reporter ions were used to relatively quantify the

**method Label** 

**5. Quantitative proteomics of urinary exosomes for protein biomarker** 

**Sample Age (years) Number of samples Exosome preparation** 

**discovery** 

labeling.

quantitation (Figure 2.).

monitored in the same experiment.

peptides and the proteins from which they originate by Mascot iTRAQ 4-plex quantification method. Proteins which were quantified with a minimum of two unique peptides and p<0.05 significance threshold using MudPIT scoring have been considered.

More than hundred proteins were quantified in the iTRAQ analysis. Table 3. shows the weighted median ratios of the first 25 proteins ranked by Mascot protein score. Expression level of the major proteins isolated and purified by the double-cushion method are different from those purified with the single-cushion protocol (Table 3., ratio 117/114). In particular, cytoskeletal proteins (cubulin, megalin, actin, cofillin, moesin, tubulin, etc.) seem to be less abundant in the sample. They may be due to heterogeneous constituents of the cytoskeleton filaments present in urine which co-purify with the UEs in traditional methods. On the other hand a marked enrichment was observed in proteins which are related to the VPS4 complex of ESCRT machinery (IST1, VPS4A, and VPS4B), its associated proteins (CHM2A, CHMP5,


Table 3. The weighted median ratios of the 25 top-ranking proteins in the MudPIT based 4 plex iTRAQ experiment. 114, 115, 116 and 117 indicate sample-labeling by iTRAQ according to Table 2.

Urinary Exosomes for Protein Biomarker Research 61

Babst, M. (2011). MVB vesicle formation: ESCRT-dependent, ESCRT-independent and everything in between. *Current Opinion in Cell Biology,* In Press, Corrected Proof. Camussi, G., Deregibus, M. C., Bruno, S., Cantaluppi, V. & Biancone, L. (2011).

Cheruvanky, A., Zhou, H., Pisitkun, T., Kopp, J. B., Knepper, M. A., Yuen, P. S. T. & Star, R.

Denzer, K., Kleijmeer, M. J., Heijnen, H. F., Stoorvogel, W. & Geuze, H. J. (2000). Exosome:

Fernandez-Llama, P., Khositseth, S., Gonzales, P. A., Star, R. A., Pisitkun, T. & Knepper, M.

Gan, X. & Gould, S. J. (2011). Identification of an inhibitory budding signal that blocks the

Gonzales, P., Pisitkun, T. & Knepper, M. A. (2008). Urinary exosomes: is there a future?

Gonzales, P. A., Pisitkun, T., Hoffert, J. D., Tchapyjnikov, D., Star, R. A., Kleta, R., Wang, N.

Gutwein, P., Schramme, A., Abdel-Bakky, M., Doberstein, K., Hauser, I., Ludwig, A.,

Hogan, M. C., Manganelli, L., Woollard, J. R., Masyuk, A. I., Masyuk, T. V., Tammachote, R.,

Keller, S., Ridinger, J., Rupp, A.-K., Janssen, J. & Altevogt, P. (2011). Body fluid derived

Keller, S., Rupp, C., Stoeck, A., Runz, S., Fogel, M., Lugert, S., Hager, H. D., Abdel-Bakky, M.

Keller, S., Sanderson, M. P., Stoeck, A. & Altevogt, P. (2006). Exosomes: From biogenesis and secretion to biological function. *Immunology Letters,* 107 (2), 102-108. Kobayashi, K. & Fukuoka, S. (2001). Conditions for Solubilization of Tamm-Horsfall

urine and amniotic fluid. *Kidney International,* 72 (9), 1095-1102.

Exosomes/microvesicles as a mechanism of cell-to-cell communication. *Kidney* 

A. (2007). Rapid isolation of urinary exosomal biomarkers using a nanomembrane ultrafiltration concentrator. *American Journal of Physiololgy Renal Physiology,* 292 (5),

from internal vesicle of the multivesicular body to intercellular signaling device.

A. (2010). Tamm-Horsfall protein and urinary exosome isolation. *Kidney* 

release of HIV particles and exosome/microvesicle proteins. *Molecular Biology of the* 

S. & Knepper, M. A. (2009). Large-Scale Proteomics and Phosphoproteomics of Urinary Exosomes. *Journal of the American Society of Nephrology,* 20 (2), 363-379. Gruenberg, J. & Stenmark, H. (2004). The biogenesis of multivesicular endosomes. *Nature* 

Altevogt, P., Gauer, S., Hillmann, A., Weide, T., Jespersen, C., Eberhardt, W. & Pfeilschifter, J. (2010). ADAM10 is expressed in human podocytes and found in urinary vesicles of patients with glomerular kidney diseases. *Journal of Biomedical* 

Huang, B. Q., Leontovich, A. A., Beito, T. G., Madden, B. J., Charlesworth, M. C., Torres, V. E., LaRusso, N. F., Harris, P. C. & Ward, C. J. (2009). Characterization of PKD Protein-Positive Exosome-Like Vesicles. *Journal of the American Society of* 

exosomes as a novel template for clinical diagnostics. *Journal of Translational* 

S., Gutwein, P. & Altevogt, P. (2007). CD24 is a marker of exosomes secreted into

Protein/Uromodulin in Human Urine and Establishment of a Sensitive and

**8. References** 

*International,* 78 (9), 838-848.

*International,* 77 (8), 736-742.

*Cell,* 22 (6), 817-830.

*Science,* 17 (1), 3.

*Medicine,* 9 (1), 86.

*Nephrology,* 20 (2), 278-288.

*Journal of Cell Science,* 113 (19), 3365-3374.

*Nephrology Dialysis Transplantation,* 23 (6), 1799-1801.

*Reviews Molecular Cell Biology,* 5 (4), 317-323.

F1657-1661.

CHM4B) and proteins involved in the ubiquitination process (RS27A). The most abundant protein according to SDS-PAGE and MudPIT analyses is aminopeptidase (AMPN) known to reflect a periodicity in renal tubular function. Other proteins like AQP1, NEP, DPEP1 and DPP4 also related to renal function were identified among the most abundant proteins. Based on statistical analysis of the data, more than a 2-fold change was considered to be significant. Data obtained confirms that the double-cushion method efficiently removes the major urinary protein contamination characteristic of the current purification methods (more than a 20-fold change). In different single-cushion preparations (data not shown) the relative protein quantities vary considerably respect to that of uromodulin (i.e. mean of the fold changes of all quantified proteins unless uromodulin/uromodulin fold change). This can be explained by the poor reproducibility and it causes considerable complications in protein quantification and normalization. Comparing the two different age-groups we analysed, no significant difference in the expression was found in the 25 top-ranking exosomal proteins (Table 3., ratio 115/114). The individual sample, on the other hand shows few characteristic differences when compared with the pooled samples (116/114). In our study, the expression levels of PSCA and THY1 and ANX11 were found to be significantly altered respect to the age-matched control group. For a protein biomarker discovery platform which employs urinary exosomes as biomarker source, it is highly advisable to use a pooled control sample with a high number and clinically well defined individual samples.

#### **6. Conclusions**

Given the non-invasive nature of urine sample collection and the evolving biological significance of secreted membrane vesicles, unbiased quantitative analysis of biomolecules isolated from urinary exosomes is a step forward in clinical biomarker research. Recently we have set-up a multiplex quantitative approach for the analysis of protein contents of purified urinary exosomes (Figure 2.). This includes protocols for i.) the removal of major urinary exosome contaminations, ii.) the separation of urinary membrane vesicles of different sizes iii.) vesicle lysis and protein solubilization and, iv) the quantitative proteomics based urinary exosomal biomarker research. The novel isolation/purification procedure efficiently removes the major urinary exosomal contaminations and separates exosomes from other membrane vesicles. Thus it provides a good basis for the development of optimized methods for protein biomarker research. Quantitative MudPIT analysis performed on biological, analytical and technical replicates shows excellent reproducibility. No significant expression difference was found among normal healthy subjects grouped by age. Preliminary data suggests a superior performance in single sample biomarker analysis design over a pooling design. All together, these results suggest a prolific future of urinary exosomes in clinical proteomics of different diseases involving the renal and urinary tract.

#### **7. Acknowledgment**

The authors are grateful for the financial contribution of Italian Society of Nephrology granted by "Ricercando 2011" for the project entitled "Identification of reliable urinary biomarkers of Diabetic Nephropathy by means of powerful and complementary proteomic strategies". We also thank Rosarita Tatè and Michele Cermola (IGB, CNR) for the TEM analysis.

#### **8. References**

60 Proteomics – Human Diseases and Protein Functions

CHM4B) and proteins involved in the ubiquitination process (RS27A). The most abundant protein according to SDS-PAGE and MudPIT analyses is aminopeptidase (AMPN) known to reflect a periodicity in renal tubular function. Other proteins like AQP1, NEP, DPEP1 and DPP4 also related to renal function were identified among the most abundant proteins. Based on statistical analysis of the data, more than a 2-fold change was considered to be significant. Data obtained confirms that the double-cushion method efficiently removes the major urinary protein contamination characteristic of the current purification methods (more than a 20-fold change). In different single-cushion preparations (data not shown) the relative protein quantities vary considerably respect to that of uromodulin (i.e. mean of the fold changes of all quantified proteins unless uromodulin/uromodulin fold change). This can be explained by the poor reproducibility and it causes considerable complications in protein quantification and normalization. Comparing the two different age-groups we analysed, no significant difference in the expression was found in the 25 top-ranking exosomal proteins (Table 3., ratio 115/114). The individual sample, on the other hand shows few characteristic differences when compared with the pooled samples (116/114). In our study, the expression levels of PSCA and THY1 and ANX11 were found to be significantly altered respect to the age-matched control group. For a protein biomarker discovery platform which employs urinary exosomes as biomarker source, it is highly advisable to use a pooled control sample with a

Given the non-invasive nature of urine sample collection and the evolving biological significance of secreted membrane vesicles, unbiased quantitative analysis of biomolecules isolated from urinary exosomes is a step forward in clinical biomarker research. Recently we have set-up a multiplex quantitative approach for the analysis of protein contents of purified urinary exosomes (Figure 2.). This includes protocols for i.) the removal of major urinary exosome contaminations, ii.) the separation of urinary membrane vesicles of different sizes iii.) vesicle lysis and protein solubilization and, iv) the quantitative proteomics based urinary exosomal biomarker research. The novel isolation/purification procedure efficiently removes the major urinary exosomal contaminations and separates exosomes from other membrane vesicles. Thus it provides a good basis for the development of optimized methods for protein biomarker research. Quantitative MudPIT analysis performed on biological, analytical and technical replicates shows excellent reproducibility. No significant expression difference was found among normal healthy subjects grouped by age. Preliminary data suggests a superior performance in single sample biomarker analysis design over a pooling design. All together, these results suggest a prolific future of urinary exosomes in clinical proteomics

The authors are grateful for the financial contribution of Italian Society of Nephrology granted by "Ricercando 2011" for the project entitled "Identification of reliable urinary biomarkers of Diabetic Nephropathy by means of powerful and complementary proteomic strategies". We

also thank Rosarita Tatè and Michele Cermola (IGB, CNR) for the TEM analysis.

high number and clinically well defined individual samples.

of different diseases involving the renal and urinary tract.

**6. Conclusions** 

**7. Acknowledgment** 


Urinary Exosomes for Protein Biomarker Research 63

Rood, I. M., Deegens, J. K. J., Merchant, M. L., Tamboer, W. P. M., Wilkey, D. W., Wetzels, J.

Ross, P. L., Huang, Y. N., Marchese, J. N., Williamson, B., Parker, K., Hattan, S., Khainovski,

Schaeffer, C., Santambrogio, S., Perucca, S., Casari, G. & Rampoldi, L. (2009). Analysis of

Shen, B., Wu, N., Yang, J.-M. & Gould, S. J. (2011). Protein targeting to

Simpson, R. J., Lim, J. W. E., Moritz, R. L. & Mathivanan, S. (2009). Exosomes: proteomic insights and diagnostic potential. *Expert Review of Proteomics,* 6 (3), 267-283. Smalley, D. M., Sheman, N. E., Nelson, K. & Theodorescu, D. (2008). Isolation and

Sonoda, H., Yokota-Ikeda, N., Oshikawa, S., Kanno, Y., Yoshinaga, K., Uchida, K., Ueda, Y.,

Taylor, D. D. L., KY, US), Gercel-taylor, Cicek (Louisville, KY, US) (2010). Exosome-

Thery, C., Zitvogel, L. & Amigorena, S. (2002). Exosomes: composition, biogenesis and

Trajkovic, K., Hsu, C., Chiantia, S., Rajendran, L., Wenzel, D., Wieland, F., Schwille, P.,

Utleg, A. G., Yi, E. C., Xie, T., Shannon, P., White, J. T., Goodlett, D. R., Hood, L. & Lin, B. (2003). Proteomic analysis of human prostasomes. *The Prostate,* 56 (2), 150-161. Valadi, H., Ekstrom, K., Bossios, A., Sjostrand, M., Lee, J. J. & Lotvall, J. O. (2007). Exosome-

van Niel, G., Porto-Carreiro, I., Simoes, S. & Raposo, G. (2006). Exosomes: A Common Pathway for a Specialized Function. *Journal of Biochemistry,* 140 (1), 13-21. Welton, J. L., Khanna, S., Giles, P. J., Brennan, P., Brewis, I. A., Staffurth, J., Mason, M. D. &

Williams, R. L. & Urbe, S. (2007). The emerging shape of the ESCRT machinery. *Nature* 

Reagents. *Molecular & Cellular Proteomics,* 3 (12), 1154-1169.

*Journal of Physiology - Renal Physiology,* 297 (4), F1006-F1016.

into Multivesicular Endosomes. *Science,* 319 (5867), 1244-1247.

exchange between cells. *Nature Cell Biology,* 9 (6), 654-659.

*Journal of Proteome Research,* 7 (5), 2088-2096.

ResearchFoundation Inc. (Louisville, KY, US).

*Reviwes Molecular Cell Biology,* 8 (5), 355-368.

function. *Nature reviews,* 2, 569 - 579.

*Cellular Proteomics*.

816.

(2), 589-599.

*Chemistry*.

F. M. & Klein, J. B. (2010). Comparison of three methods for isolation of urinary microvesicles to identify biomarkers of nephrotic syndrome. *Kidney Int,* 78 (8), 810-

N., Pillai, S., Dey, S., Daniels, S., Purkayastha, S., Juhasz, P., Martin, S., Bartlet-Jones, M., He, F., Jacobson, A. & Pappin, D. J. (2004). Multiplexed Protein Quantitation in Saccharomyces cerevisiae Using Amine-reactive Isobaric Tagging

Uromodulin Polymerization Provides New Insights into the Mechanisms Regulating ZP Domain-mediated Protein Assembly. *Molecular Biology of the Cell,* 20

exosomes/microvesicles by plasma membrane anchors. *Journal of Biological* 

Identification of Potential Urinary Microparticle Biomarkers of Bladder Cancer.

Kimiya, K., Uezono, S., Ueda, A., Ito, K. & Ikeda, M. (2009). Decreased abundance of urinary exosomal aquaporin-1 in renal ischemia-reperfusion injury. *American* 

associated micro RNA as a dignostic marker. United States: University of Lousville

Brugger, B. & Simons, M. (2008). Ceramide Triggers Budding of Exosome Vesicles

mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic

Clayton, A. (2010). Proteomic analysis of bladder cancer exosomes. *Molecular &* 

Accurate Enzyme-Linked Immunosorbent Assay (ELISA) Method. *Archives of Biochemistry and Biophysics,* 388 (1), 113-120.


Lee, T., D'Asti, E., Magnus, N., Al-Nedawi, K., Meehan, B. & Rak, J. (2011). Microvesicles as

Li, J. Y., Yong, T. Y., Michael, M. Z. & Gleadle, J. M. (2010). Review: The role of microRNAs

Li, X.-B., Zhang, Z.-R., Schluesener, H. J. & Xu, S.-Q. (2006). Role of exosomes in immune

Li, Y., Zhang, Y., Qiu, F. & Qiu, Z. (2011). Proteomic identification of exosomal LRG1: A potential urinary biomarker for detecting NSCLC. *Electrophoresis,* 32, 1-8. Marsh, M. & van Meer, G. (2008). No ESCRTs for Exosomes. *Science,* 319 (5867), 1191-1192. Mathivanan, S., Lim, J. W. E., Tauro, B. J., Ji, H., Moritz, R. L. & Simpson, R. J. (2010).

Mathivanan, S. & Simpson, R. J. (2009). ExoCarta: A compendium of exosomal proteins and

Merchant, M. L., Powell, D. W., Wilkey, D. W., Cummins, T. D., Deegens, J. K., Rood, I. M.,

Miranda, K. C., Bond, D. T., McKee, M., Skog, J., Paunescu, T. G., Da Silva, N., Brown, D. &

potential biomarkers for renal disease. *Kidney International,* 78 (2), 191-199. Mitchell, P., Welton, J., Staffurth, J., Court, J., Mason, M., Tabi, Z. & Clayton, A. (2009). Can

Nilsson, J., Skog, J., Nordstrand, A., Baranov, V., Mincheva-Nilsson, L., Breakefield, X. O. &

biomarkers for prostate cancer. *British Journal of Cancer,* 100 (10), 1603-1607. Pisitkun, T., Shen, R.-F. & Knepper, M. A. (2004). Identification and proteomic profiling of

Porter, K. R. & Tamm, I. (1955). Direct visualization of a mucoprotein component of urine.

Raj, D. A. A., Capasso, G., Fiume, I. & Pocsfalvi, G. (2011a). A multiplex quantitative

Raj, D. A. A., Capasso, G., Fiume, I. & Pocsfalvi, G. (2011b). Procedura di isolamento e di

Pigmented Melanocytic Cells. *The Journal of Cell Biology,* 152 (4), 809-824.

regulation. *Journal of Cellular and Molecular Medicine,* 10 (2), 364-375.

*Biochemistry and Biophysics,* 388 (1), 113-120.

in kidney disease. *Nephrology,* 15 (6), 599-608.

RNA. *Proteomics,* 9 (21), 4997-5000.

*Applications,* 4 (1), 84-96.

*Translational Medicine,* 7 (1), 4.

*United States of America,* 101 (36), 13368-13373.

*Journal of Biological Chemistry,* 212 (1), 135-140.

*International,* accepted, manuscript ID: KI-09-11-1553.R1.

cellular 'debris'. *Seminars in Immunopathology*, 1-13.

Signature. *Molecular & Cellular Proteomics,* 9 (2), 197-208.

Accurate Enzyme-Linked Immunosorbent Assay (ELISA) Method. *Archives of* 

mediators of intercellular communication in cancer—the emerging science of

Proteomics Analysis of A33 Immunoaffinity-purified Exosomes Released from the Human Colon Tumor Cell Line LIM1215 Reveals a Tissue-specific Protein

McAfee, K. J., Fleischer, C., Klein, E. & Klein, J. B. (2010). Microfiltration isolation of human urinary exosomes for characterization by MS. *Proteomics – Clinical* 

Russo, L. M. (2010). Nucleic acids within urinary exosomes/microvesicles are

urinary exosomes act as treatment response markers in prostate cancer? *Journal of* 

Widmark, A. (2009). Prostate cancer-derived urine exosomes: a novel approach to

exosomes in human urine. *Proceedings of the National Academy of Sciences of the* 

proteomics strategy for protein biomarker studies in urinary exosomes *Kidney* 

purificazione degli esosomi urinari per la ricerca di biomarcatori proteici. In U. I. B. M. patent deposited in 14 March 2011 (ed.), *Ufficio Italiano Brevetti e Marchi,* Italy. Raposo, G. a., Tenza, D., Murphy, D. M., Berson, J. F. & Marks, M. S. (2001). Distinct Protein

Sorting and Localization to Premelanosomes, Melanosomes, and Lysosomes in


**4** 

*Canada* 

**Circadian Proteomics and Its Unique Advantage** 

Current statistics from the World Health Organization, Heart and Stroke Foundation of Canada, and the American Heart Association show that cardiovascular disease remains a leading cause of death (Heart and Stroke Foundation of Canada [HSFO], 2011; Roger et al., 2011; World Health Organization [WHO], 2011). Innovative and integrative approaches aimed at understanding and treating heart disease are needed. In this chapter we introduce a novel field of investigation called cardiovascular circadian proteomics. This approach is based on the application of high throughput proteomic technologies for discovery of molecular processes in cardiovascular tissues over the 24-hour day/night cycles. Time is a crucial but frequently overlooked factor affecting our physiology in health and disease. Circadian cardiovascular proteomics offers considerable promise to advance our understanding of heart disease (indeed disease in general), and opens new avenues for

**2. The circadian system and its importance to cardiovascular physiology** 

night and in the early morning (Guo & Stein, 2003; Imai et al., 1990).

Life on earth is subject to a 24-hour day/night (circadian or diurnal) cycle. Circadian systems have evolved to allow physiological and behavioural processes to be synchronous with this cycle – mammals are adapted to sleep either during the day or at night. Circadian clocks allow us to entrain to environmental cues and hence anticipate the differing physiologic and behavioural demands of daily events. In mammals, the system is organized as a hierarchy with multiple oscillators, as has been well reviewed (Hastings et al., 2003; Rajaratnam & Arendt, 2001; Reppert & Weaver, 2001, 2002). At the top is the hypothalamic suprachiasmatic nucleus (SCN), a brain region functioning as a master rhythmic regulator (Figure 1A). The SCN integrates light information received from the eyes to coordinate clocks throughout the body. There are additional circadian regulatory systems such as the food entrainable clock (Storch & Weitz, 2009) which we are just now beginning to understand, but are beyond the scope of this review. We observe the output of the entrained clocks as daily physiologic rhythms, many of which are crucial to the cardiovascular system, such as the cyclic variation in heart rate (HR) and blood pressure (BP). Daily HR and BP follow the diurnal variation of our autonomic nervous system and increase around waketime to help sustain cardiac output, then decrease during the period of vagal dominance at

**1. Introduction** 

treatment of patients clinically.

**for Discovery of Biomarkers of Heart Disease** 

Peter S. Podobed, Gordon M. Kirby and Tami A. Martino

*Cardiovascular Research Group, Biomedical Sciences,* 

*University of Guelph, Ontario* 


### **Circadian Proteomics and Its Unique Advantage for Discovery of Biomarkers of Heart Disease**

Peter S. Podobed, Gordon M. Kirby and Tami A. Martino *Cardiovascular Research Group, Biomedical Sciences, University of Guelph, Ontario Canada* 

#### **1. Introduction**

64 Proteomics – Human Diseases and Protein Functions

Zhou, H., Cheruvanky, A., Hu, X., Matsumoto, T., Hiramatsu, N., Cho, M. E., Berger, A.,

Zhou, H., Pisitkun, T., Aponte, A., Yuen, P. S. T., Hoffert, J. D., Yasuda, H., Hu, X., Chawla,

Zhou, H., Yuen, P. S. T., Pisitkun, T., Gonzales, P. A., Yasuda, H., Dear, J. W., Gross, P.,

621.

*International,* 70 (10), 1847-1857.

*International,* 69 (8), 1471-1476.

Leelahavanichkul, A., Doi, K., Chawla, L. S., Illei, G. G., Kopp, J. B., Balow, J. E., Austin, H. A., III, Yuen, P. S. T. & Star, R. A. (2008). Urinary exosomal transcription factors, a new class of biomarkers for renal disease. *Kidney International,* 74 (5), 613-

L., Shen, R. F., Knepper, M. A. & Star, R. A. (2006a). Exosomal Fetuin-A identified by proteomics: A novel urinary biomarker for detecting acute kidney injury. *Kidney* 

Knepper, M. A. & Star, R. A. (2006b). Collection, storage, preservation, and normalization of human urinary exosomes for biomarker discovery. *Kidney* 

> Current statistics from the World Health Organization, Heart and Stroke Foundation of Canada, and the American Heart Association show that cardiovascular disease remains a leading cause of death (Heart and Stroke Foundation of Canada [HSFO], 2011; Roger et al., 2011; World Health Organization [WHO], 2011). Innovative and integrative approaches aimed at understanding and treating heart disease are needed. In this chapter we introduce a novel field of investigation called cardiovascular circadian proteomics. This approach is based on the application of high throughput proteomic technologies for discovery of molecular processes in cardiovascular tissues over the 24-hour day/night cycles. Time is a crucial but frequently overlooked factor affecting our physiology in health and disease. Circadian cardiovascular proteomics offers considerable promise to advance our understanding of heart disease (indeed disease in general), and opens new avenues for treatment of patients clinically.

#### **2. The circadian system and its importance to cardiovascular physiology**

Life on earth is subject to a 24-hour day/night (circadian or diurnal) cycle. Circadian systems have evolved to allow physiological and behavioural processes to be synchronous with this cycle – mammals are adapted to sleep either during the day or at night. Circadian clocks allow us to entrain to environmental cues and hence anticipate the differing physiologic and behavioural demands of daily events. In mammals, the system is organized as a hierarchy with multiple oscillators, as has been well reviewed (Hastings et al., 2003; Rajaratnam & Arendt, 2001; Reppert & Weaver, 2001, 2002). At the top is the hypothalamic suprachiasmatic nucleus (SCN), a brain region functioning as a master rhythmic regulator (Figure 1A). The SCN integrates light information received from the eyes to coordinate clocks throughout the body. There are additional circadian regulatory systems such as the food entrainable clock (Storch & Weitz, 2009) which we are just now beginning to understand, but are beyond the scope of this review. We observe the output of the entrained clocks as daily physiologic rhythms, many of which are crucial to the cardiovascular system, such as the cyclic variation in heart rate (HR) and blood pressure (BP). Daily HR and BP follow the diurnal variation of our autonomic nervous system and increase around waketime to help sustain cardiac output, then decrease during the period of vagal dominance at night and in the early morning (Guo & Stein, 2003; Imai et al., 1990).

Circadian Proteomics and Its Unique Advantage for Discovery of Biomarkers of Heart Disease 67

A) Light activates dedicated retinal receptors and signals the hypothalamic suprachiasmatic nucleus (SCN). The SCN is a master body clock, and orchestrates physiologic and molecular rhythms in peripheral organs including the heart. Rhythms relevant to the heart include the daily cycling of heart rate and blood pressure, also timing of onset of acute cardiac events such as myocardial infarction. B) The molecular clock mechanism is dependent upon oscillating levels of proteins that interact via a 24-hour autoregulatory feedback. On the positive arm BMAL1 and CLOCK combine as heterodimers and bind to E-box elements upstream in the coding regions of other core clock elements, PERIOD and CRYPTOCHROME. These are phosphorylated by CASEIN KINASE 1 EPSILON (or DELTA), form heterodimers, then translocate to the nucleus. There, they bind to the same E-box elements thus negatively regulating their own expression. The positive loop also initiates production of RETINOIC ACID-RELATED ORPHAN NUCLEAR RECEPTOR ALPHA resulting in its inhibition, and completion

of the 24-hour cycle. CAPITAL=Proteins; *italic*=mRNA.

Fig. 1. The Circadian System.

Circadian rhythms also underlie the timing of onset of adverse cardiovascular events. The incidence of myocardial infarction in humans peaks in the morning (~6:00A.M.-12:00 noon) (Cohen et al., 1997; Goldberg et al., 1990; Muller et al., 1985). A similar pattern is observed in the incidences of sudden cardiac death (Muller et al., 1989; Willich et al., 1987), ventricular tachyarrhythmia (Eksik et al., 2007; Tofler et al., 1995), and rupture of aortic aneurysms (Manfredini et al., 2004; Mehta et al., 2002; Sumiyoshi et al., 2002). Precursor risk factors such as vasomotor tone, platelet aggregability, and other factors involved in thrombosis or thrombolysis also exhibit daily rhythms (Andrews et al., 1996; Angleton et al., 1989; Decousus et al., 1985; Maemura et al., 2000; Otto et al., 2004). Recent studies link timing of onset of adverse cardiac events with a circadian clock mechanism, reviewed in (Durgan & Young, 2010; Martino & Sole, 2009; Sole & Martino, 2009).

Proteins involved in the molecular clock mechanism have been identified in the last 10+ years, and daily oscillations of this mechanism in peripheral tissues, including the myocardium, are believed to be primarily combination of self-sustaining cycling along with neural/hormonal cues from the SCN. It is illustrated in Figure 1B and described in many excellent reviews (Hastings et al., 2003; Reddy et al., 2005; Roenneberg & Merrow, 2005). Though we focus here on protein cycling, it is important to note that there are posttranslational rhythms as well, such as phosphorylation. Though beyond the scope of this review the reader is directed to several examples (Akashi et al., 2002; Lowrey et al., 2000; Iitaka et al., 2005; Yin et al., 2006).

In the following sections, we describe how application of the circadian concepts, in combination with state-of-the-art proteomics, provides significant new opportunities for understanding disease physiology, for biomarker discovery, and helping patients clinically.

#### **3. Discovery of the circadian heart proteome**

Our laboratory is focussed on discovering the circadian heart proteome both in normal tissue and in disease. Initial studies are done using murine models of cardiovascular disease, and later are translated clinically. Here we describe the circadian proteome in normal C57Bl/6 mouse heart, and in our well-established murine model of heart disease termed pressure-overload induced cardiac hypertrophy by **T**ransverse **A**ortic **C**onstriction (TAC). To induce heart disease, eight week old male mice were entrained to a 12:12 light (L): dark (D) cycle and administered TAC surgery where a ligature was placed distal to the third bifurcation of aorta (Figure 2A). In sham operated animals the surgical procedure was identical, but the ligature was not tightened. (Figure 2A). For proteomic studies, heart tissues were collected one week later (as the heart remodels), at six timepoints 4 hours apart over the 24-hour L:D cycle (Figure 2B). The proteome was analyzed by two-dimensional difference in gel electrophoresis (2D-DIGE) and mass spectrometry (MS). Figure 2B illustrates the experimental workflow design. The technical details are described below.

#### **3.1 Protein purification and labelling**

The cytoplasmic soluble proteome was purified from either TAC or sham left ventricular heart tissue. Cardiac tissue was immersed in 600 l ice-cold cell lysis buffer (10 mM Tris pH 8, 8 M Urea, 4% w/v 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), and protease inhibitors). The lysis buffer helps to solubilize, denature, and

A) Light activates dedicated retinal receptors and signals the hypothalamic suprachiasmatic nucleus (SCN). The SCN is a master body clock, and orchestrates physiologic and molecular rhythms in peripheral organs including the heart. Rhythms relevant to the heart include the daily cycling of heart rate and blood pressure, also timing of onset of acute cardiac events such as myocardial infarction. B) The molecular clock mechanism is dependent upon oscillating levels of proteins that interact via a 24-hour autoregulatory feedback. On the positive arm BMAL1 and CLOCK combine as heterodimers and bind to E-box elements upstream in the coding regions of other core clock elements, PERIOD and CRYPTOCHROME. These are phosphorylated by CASEIN KINASE 1 EPSILON (or DELTA), form heterodimers, then translocate to the nucleus. There, they bind to the same E-box elements thus negatively regulating their own expression. The positive loop also initiates production of RETINOIC ACID-RELATED ORPHAN NUCLEAR RECEPTOR ALPHA resulting in its inhibition, and completion of the 24-hour cycle. CAPITAL=Proteins; *italic*=mRNA.

Fig. 1. The Circadian System.

66 Proteomics – Human Diseases and Protein Functions

Circadian rhythms also underlie the timing of onset of adverse cardiovascular events. The incidence of myocardial infarction in humans peaks in the morning (~6:00A.M.-12:00 noon) (Cohen et al., 1997; Goldberg et al., 1990; Muller et al., 1985). A similar pattern is observed in the incidences of sudden cardiac death (Muller et al., 1989; Willich et al., 1987), ventricular tachyarrhythmia (Eksik et al., 2007; Tofler et al., 1995), and rupture of aortic aneurysms (Manfredini et al., 2004; Mehta et al., 2002; Sumiyoshi et al., 2002). Precursor risk factors such as vasomotor tone, platelet aggregability, and other factors involved in thrombosis or thrombolysis also exhibit daily rhythms (Andrews et al., 1996; Angleton et al., 1989; Decousus et al., 1985; Maemura et al., 2000; Otto et al., 2004). Recent studies link timing of onset of adverse cardiac events with a circadian clock mechanism, reviewed in (Durgan &

Proteins involved in the molecular clock mechanism have been identified in the last 10+ years, and daily oscillations of this mechanism in peripheral tissues, including the myocardium, are believed to be primarily combination of self-sustaining cycling along with neural/hormonal cues from the SCN. It is illustrated in Figure 1B and described in many excellent reviews (Hastings et al., 2003; Reddy et al., 2005; Roenneberg & Merrow, 2005). Though we focus here on protein cycling, it is important to note that there are posttranslational rhythms as well, such as phosphorylation. Though beyond the scope of this review the reader is directed to several examples (Akashi et al., 2002; Lowrey et al., 2000;

In the following sections, we describe how application of the circadian concepts, in combination with state-of-the-art proteomics, provides significant new opportunities for understanding disease physiology, for biomarker discovery, and helping patients

Our laboratory is focussed on discovering the circadian heart proteome both in normal tissue and in disease. Initial studies are done using murine models of cardiovascular disease, and later are translated clinically. Here we describe the circadian proteome in normal C57Bl/6 mouse heart, and in our well-established murine model of heart disease termed pressure-overload induced cardiac hypertrophy by **T**ransverse **A**ortic **C**onstriction (TAC). To induce heart disease, eight week old male mice were entrained to a 12:12 light (L): dark (D) cycle and administered TAC surgery where a ligature was placed distal to the third bifurcation of aorta (Figure 2A). In sham operated animals the surgical procedure was identical, but the ligature was not tightened. (Figure 2A). For proteomic studies, heart tissues were collected one week later (as the heart remodels), at six timepoints 4 hours apart over the 24-hour L:D cycle (Figure 2B). The proteome was analyzed by two-dimensional difference in gel electrophoresis (2D-DIGE) and mass spectrometry (MS). Figure 2B illustrates the experimental workflow design. The technical details are

The cytoplasmic soluble proteome was purified from either TAC or sham left ventricular heart tissue. Cardiac tissue was immersed in 600 l ice-cold cell lysis buffer (10 mM Tris pH 8, 8 M Urea, 4% w/v 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), and protease inhibitors). The lysis buffer helps to solubilize, denature, and

Young, 2010; Martino & Sole, 2009; Sole & Martino, 2009).

**3. Discovery of the circadian heart proteome** 

Iitaka et al., 2005; Yin et al., 2006).

clinically.

described below.

**3.1 Protein purification and labelling** 

Circadian Proteomics and Its Unique Advantage for Discovery of Biomarkers of Heart Disease 69

A) Heart disease model. Pressure overload cardiac hypertrophy is induced by transverse aortic

created. Proteins are excised and identified by MS. Results are validated by Western blot.

Fig. 2. Experimental design.

constriction (TAC). A ligature is placed distal to the third bifurcation of aorta. Sham animals undergo the same procedure except the ligature is not tightened. B) Two-dimensional difference in gel electrophoresis (2D-DIGE) and mass spectrometry (MS) approach to characterize the circadian cardiovascular proteome in health and disease. Heart tissue is collected from TAC disease and sham control animals at 6 timepoints over the 24-hour light/dark cycle. The cytoplasmic proteome from TAC vs. sham hearts is labelled with Cy3 and Cy5 dye, respectively. An internal control consists of pooled samples labelled with Cy2. After 2D-DIGE protein expression is analyzed with DeCyder. SYPRO Ruby gel for picking spots of interest is

disaggregate proteins. Cell disruption was carried out using a Potter-Elvehjem tissue grinder. Following centrifugation the supernatant was collected and proteins were quantified by Bradford assay.

Prior to protein separation, 50 g of each sample was labelled with CyDyes (Cy3/Cy5), and an internal control was pooled from both lysates and labelled with Cy2 (Figure 2B). To account for any bias due to preferential binding of CyDyes, a dye swap approach was also done in a separate experiment, so that the samples were alternatively labelled with the reciprocal dye. CyDyes form a covalent bond between their NHS ester reactive group and epsilon amino group of protein's lysine residues. The labelling reaction does not cause a significant change in isolectric point because lysine carries an intrinsic +1 charge at neutral or acidic pH, which is replaced by CyDye's +1 charge. CyDyes were added so that there was a stoichiometric excess of proteins and thus only 1-5% of lysines were labelled. The reaction was carried out for 30 minutes in the dark, and then quenched with 10 mM lysine.

#### **3.2 First dimension isoelectric focussing**

Proteins were separated in the first dimension by isoelectric focusing (IEF) based on our standard operating protocol (Hobson et al., 2007). In the experiment shown here (Figure 3) we used nonlinear 13 cm immobilized pH gradient (IPG) strips pH 3 – 10 (GE Healthcare). The CyDye labelled samples were combined and mixed with rehydration buffer (8 M urea, 2% w/v CHAPS, Dithiothreitol (DTT), 0.5% v/v carrier pharmalytes). Then, samples in rehydration buffer were applied at the bottom of the IPG strip holder. The strip was put on top of the sample solution and covered with paraffin oil to prevent crystallization of urea, water loss and carbon dioxide dissolving at the alkaline end of the strip. IEF was done using an Ettan IPGphore unit. The initial step was active rehydration, allowing proteins to slowly enter the strip along its whole length under the influence of a small current (30 V for 10 h). The next series of steps were 500 V step and hold for 2 h, 1000 V gradient for 1 h, 8000 V gradient for 2.5 h, 8000 V step and hold for 16000 V h, 500 V step and hold for 2 h. The result was that proteins migrated to a position in accordance with their isoelectric point.

#### **3.3 Second dimension electrophoresis**

For the second dimension, the sample containing IPG strips were equilibrated with sodium dodecyl sulfate (SDS) running buffer (75 mM Tris (pH 8.8), 6 M Urea, 30% v/v Glycerol and 2% w/v SDS). DTT was added to reduce disulfide bonds, and iodoacetamide (IAA) to alkylate thiol groups what prevented disulfide bonds from reforming. After equilibration the IPG strip was applied directly on top of a large format 16 X 14 cm 12% acrylamide gel. Agarose was poured on top of the strip to prevent air from getting underneath and to hold strip in place. Molecular markers were applied 4 cm from the positive end of the strip. The proteins were then separated vertically according to their molecular weight by SDSpolyacrylamide gel electrophoresis (PAGE) on a DALT 6 electrophoresis unit at 20 0C, 6 W/ gel for 18 h.

#### **3.4 Protein detection and bioinformatics analysis**

After 2D-DIGE, the relative abundance of proteins from the TAC heart disease versus sham protein lysates was detected using a high resolution fluorescent scanner Typhoon 9410 (GE Healthcare). The excitation/emission wavelengths for the CyDyes were as follows: Cy2, 480 nm/530 nm (blue), Cy3, 540 nm/590 nm (green) and Cy5, 620 nm/680 nm (red). Images

A) Heart disease model. Pressure overload cardiac hypertrophy is induced by transverse aortic constriction (TAC). A ligature is placed distal to the third bifurcation of aorta. Sham animals undergo the same procedure except the ligature is not tightened. B) Two-dimensional difference in gel electrophoresis (2D-DIGE) and mass spectrometry (MS) approach to characterize the circadian cardiovascular proteome in health and disease. Heart tissue is collected from TAC disease and sham control animals at 6 timepoints over the 24-hour light/dark cycle. The cytoplasmic proteome from TAC vs. sham hearts is labelled with Cy3 and Cy5 dye, respectively. An internal control consists of pooled samples labelled with Cy2. After 2D-DIGE protein expression is analyzed with DeCyder. SYPRO Ruby gel for picking spots of interest is created. Proteins are excised and identified by MS. Results are validated by Western blot.

Fig. 2. Experimental design.

68 Proteomics – Human Diseases and Protein Functions

disaggregate proteins. Cell disruption was carried out using a Potter-Elvehjem tissue grinder. Following centrifugation the supernatant was collected and proteins were

Prior to protein separation, 50 g of each sample was labelled with CyDyes (Cy3/Cy5), and an internal control was pooled from both lysates and labelled with Cy2 (Figure 2B). To account for any bias due to preferential binding of CyDyes, a dye swap approach was also done in a separate experiment, so that the samples were alternatively labelled with the reciprocal dye. CyDyes form a covalent bond between their NHS ester reactive group and epsilon amino group of protein's lysine residues. The labelling reaction does not cause a significant change in isolectric point because lysine carries an intrinsic +1 charge at neutral or acidic pH, which is replaced by CyDye's +1 charge. CyDyes were added so that there was a stoichiometric excess of proteins and thus only 1-5% of lysines were labelled. The reaction

Proteins were separated in the first dimension by isoelectric focusing (IEF) based on our standard operating protocol (Hobson et al., 2007). In the experiment shown here (Figure 3) we used nonlinear 13 cm immobilized pH gradient (IPG) strips pH 3 – 10 (GE Healthcare). The CyDye labelled samples were combined and mixed with rehydration buffer (8 M urea, 2% w/v CHAPS, Dithiothreitol (DTT), 0.5% v/v carrier pharmalytes). Then, samples in rehydration buffer were applied at the bottom of the IPG strip holder. The strip was put on top of the sample solution and covered with paraffin oil to prevent crystallization of urea, water loss and carbon dioxide dissolving at the alkaline end of the strip. IEF was done using an Ettan IPGphore unit. The initial step was active rehydration, allowing proteins to slowly enter the strip along its whole length under the influence of a small current (30 V for 10 h). The next series of steps were 500 V step and hold for 2 h, 1000 V gradient for 1 h, 8000 V gradient for 2.5 h, 8000 V step and hold for 16000 V h, 500 V step and hold for 2 h. The result

was carried out for 30 minutes in the dark, and then quenched with 10 mM lysine.

was that proteins migrated to a position in accordance with their isoelectric point.

For the second dimension, the sample containing IPG strips were equilibrated with sodium dodecyl sulfate (SDS) running buffer (75 mM Tris (pH 8.8), 6 M Urea, 30% v/v Glycerol and 2% w/v SDS). DTT was added to reduce disulfide bonds, and iodoacetamide (IAA) to alkylate thiol groups what prevented disulfide bonds from reforming. After equilibration the IPG strip was applied directly on top of a large format 16 X 14 cm 12% acrylamide gel. Agarose was poured on top of the strip to prevent air from getting underneath and to hold strip in place. Molecular markers were applied 4 cm from the positive end of the strip. The proteins were then separated vertically according to their molecular weight by SDSpolyacrylamide gel electrophoresis (PAGE) on a DALT 6 electrophoresis unit at 20 0C,

After 2D-DIGE, the relative abundance of proteins from the TAC heart disease versus sham protein lysates was detected using a high resolution fluorescent scanner Typhoon 9410 (GE Healthcare). The excitation/emission wavelengths for the CyDyes were as follows: Cy2, 480 nm/530 nm (blue), Cy3, 540 nm/590 nm (green) and Cy5, 620 nm/680 nm (red). Images

quantified by Bradford assay.

**3.2 First dimension isoelectric focussing** 

**3.3 Second dimension electrophoresis** 

**3.4 Protein detection and bioinformatics analysis** 

6 W/ gel for 18 h.

Circadian Proteomics and Its Unique Advantage for Discovery of Biomarkers of Heart Disease 71

A) 2D-DIGE image. The gel on the left had proteins purified from TAC (heart disease) labelled with Cy3 and those from sham heart labelled with Cy5. Gel on the right shows alternatively labelled samples, where TAC proteins are labelled with Cy5 and sham with Cy3. Location of the proteins of interest is shown with arrows. B) DeCyder computer analysis of three identified protein spots # 596, 615 and 1012. Based on standardized abundance these protein spots increased in TAC vs. sham. Expression changes

are shown by 3D and Graph views.

Fig. 3. Circadian Cardiovascular Proteomics.

were visualized with Image Quant TL software. An overlay image of Cy3 and Cy5 scans is illustrated in Figure 3A. Proteins that had higher abundance in TAC heart disease vs. sham appeared as a red spot. Lower levels in TAC heart disease vs. sham appeared green. Equal amounts of protein in TAC heart disease vs. sham were yellow.

Statistical analysis and quantification of protein expression was achieved by DeCyder software (GE Healthcare). In the first step of DeCyder workflow, difference in gel analysis (DIA) used a codetection algorithm to detect and match differently labelled samples (Cy2, Cy3 and Cy5) within the same gel. It was necessary to define a specific area of interest within a gel and then manually confirm that the program detected all the spots within that area. Spot intensity corresponding to protein abundance was quantified after background subtraction and normalization. The second step termed biological variance analysis (BVA) simultaneously analyzed multiple DIGE gels by matching all the spots to a master gel, defined by user. Protein expression was compared between gels and statistically verified. Based on spot intensity, DeCyder constructed 3D views of relative protein abundance. The standardized volume of the peaks was used to calculate fold change.

In this study, and as shown in Figure 3B, we found three spots (# 596, 615, and 1012) that were upregulated in TAC heart disease compared to sham at our sleep-wake transition timepoint (ZT23). Spot # 596 had a 1.18 fold change, spot # 615 had a 1.48 fold change, and spot # 1012 had 1.77 fold change. To identify the proteins of interest, a pick gel was created containing 300 g of unlabelled protein. Glass plates were treated with Bind-Silane (-methacryloxypropyltrimethoxysilane) so that the gel was covalently attached to the surface. After electrophoresis, proteins were fixed in 10% methanol and 7% acetic acid solution for 2 hours, then stained with SYPRO Ruby for 18 hours and destained in 40% methanol/10% acetic acid solution for 2 hours. The gel was scanned with a Typhoon 9410 scanner at 532 nm (670 BP 30 emission filter). Images were uploaded to DeCyder and matched to the previous DIGE gel. Unique coordinates for each spot were created using position markers. After manual confirmation, an Ettan Spot picker (GE Healthcare) was used to excise proteins from the gel.

#### **3.5 Trypsin digestion and mass spectrometry**

To determine the identification of the proteins in gel spots # 596, 615 and 1012, we used an In-Gel Tryptic Digestion Kit (Thermo Scientific) followed by mass spectrometry (MS). Trypsin is a serine protease that cleaves the peptide bonds at the carboxyl side of lysine and arginine amino acids. Following digestion, tryptic peptide fragments were separated by nano-LC which consisted of a trap column (300μm ID) and an analytical column (75 μm ID) packed with 5μm, 300Å Zorbax SB C18 beads. A linear binary gradient was used where solvent A was 98% H2O:2% CH3CN and 0.1% (v/v) formic acid and solvent B was 2% H2O : 98% CH3CN and 0.1% (v/v) formic acid. Peptides were eluted over a 2 - 95% solvent B gradient for 100 min at a rate of 300 nL/min. The eluant from the nano-LC was coupled to a hybrid triple/quadrupole linear ion trap mass spectrometer (QTRAP 4000, ABSciex) through a nano-spray ionization source equipped with a 15 μm ID emittor tip. Our preferred database for searching was Mascot (http://www.matrixscience.com). This database contains information on more than 10 million proteins, with inherent redundancies built in since peptides can correspond to more than one protein. Each of the spots generated a list of possible protein candidates, and Mascot score and E value helped identify the correct match. The molecular weight and pI of the identified protein was comparable to the expected molecular weight and pI of the corresponding spot from the 2-DE gel.

were visualized with Image Quant TL software. An overlay image of Cy3 and Cy5 scans is illustrated in Figure 3A. Proteins that had higher abundance in TAC heart disease vs. sham appeared as a red spot. Lower levels in TAC heart disease vs. sham appeared green. Equal

Statistical analysis and quantification of protein expression was achieved by DeCyder software (GE Healthcare). In the first step of DeCyder workflow, difference in gel analysis (DIA) used a codetection algorithm to detect and match differently labelled samples (Cy2, Cy3 and Cy5) within the same gel. It was necessary to define a specific area of interest within a gel and then manually confirm that the program detected all the spots within that area. Spot intensity corresponding to protein abundance was quantified after background subtraction and normalization. The second step termed biological variance analysis (BVA) simultaneously analyzed multiple DIGE gels by matching all the spots to a master gel, defined by user. Protein expression was compared between gels and statistically verified. Based on spot intensity, DeCyder constructed 3D views of relative protein abundance. The

In this study, and as shown in Figure 3B, we found three spots (# 596, 615, and 1012) that were upregulated in TAC heart disease compared to sham at our sleep-wake transition timepoint (ZT23). Spot # 596 had a 1.18 fold change, spot # 615 had a 1.48 fold change, and spot # 1012 had 1.77 fold change. To identify the proteins of interest, a pick gel was created containing 300 g of unlabelled protein. Glass plates were treated with Bind-Silane (-methacryloxypropyltrimethoxysilane) so that the gel was covalently attached to the surface. After electrophoresis, proteins were fixed in 10% methanol and 7% acetic acid solution for 2 hours, then stained with SYPRO Ruby for 18 hours and destained in 40% methanol/10% acetic acid solution for 2 hours. The gel was scanned with a Typhoon 9410 scanner at 532 nm (670 BP 30 emission filter). Images were uploaded to DeCyder and matched to the previous DIGE gel. Unique coordinates for each spot were created using position markers. After manual confirmation, an Ettan Spot picker (GE Healthcare) was

To determine the identification of the proteins in gel spots # 596, 615 and 1012, we used an In-Gel Tryptic Digestion Kit (Thermo Scientific) followed by mass spectrometry (MS). Trypsin is a serine protease that cleaves the peptide bonds at the carboxyl side of lysine and arginine amino acids. Following digestion, tryptic peptide fragments were separated by nano-LC which consisted of a trap column (300μm ID) and an analytical column (75 μm ID) packed with 5μm, 300Å Zorbax SB C18 beads. A linear binary gradient was used where solvent A was 98% H2O:2% CH3CN and 0.1% (v/v) formic acid and solvent B was 2% H2O : 98% CH3CN and 0.1% (v/v) formic acid. Peptides were eluted over a 2 - 95% solvent B gradient for 100 min at a rate of 300 nL/min. The eluant from the nano-LC was coupled to a hybrid triple/quadrupole linear ion trap mass spectrometer (QTRAP 4000, ABSciex) through a nano-spray ionization source equipped with a 15 μm ID emittor tip. Our preferred database for searching was Mascot (http://www.matrixscience.com). This database contains information on more than 10 million proteins, with inherent redundancies built in since peptides can correspond to more than one protein. Each of the spots generated a list of possible protein candidates, and Mascot score and E value helped identify the correct match. The molecular weight and pI of the identified protein was comparable to the

expected molecular weight and pI of the corresponding spot from the 2-DE gel.

amounts of protein in TAC heart disease vs. sham were yellow.

standardized volume of the peaks was used to calculate fold change.

used to excise proteins from the gel.

**3.5 Trypsin digestion and mass spectrometry** 

A) 2D-DIGE image. The gel on the left had proteins purified from TAC (heart disease) labelled with Cy3 and those from sham heart labelled with Cy5. Gel on the right shows alternatively labelled samples, where TAC proteins are labelled with Cy5 and sham with Cy3. Location of the proteins of interest is shown with arrows. B) DeCyder computer analysis of three identified protein spots # 596, 615 and 1012. Based on standardized abundance these protein spots increased in TAC vs. sham. Expression changes are shown by 3D and Graph views.

Fig. 3. Circadian Cardiovascular Proteomics.

Circadian Proteomics and Its Unique Advantage for Discovery of Biomarkers of Heart Disease 73

**Function Reference** 

Urea cycle (Reddy et al.,

Urea cycle

**Dark period (rodent wake time)** (Moller,

Protein folding

membrane-binding

Mitochondrial electron transport

> Chaperone, signal transduction

Vesicle budding from the endoplasmic reticulum

> Activation of the complement system

Biosynthesis of serine from carbohydrates

Protein metabolism

2006)

Sparre et al. 2007)

**expression pattern** 

Arginosuccinate synthetase 1 (ASS1)

Carbamoyl phosphate synthetase 1 (CPS1)

Arginase 1 (ARG1) Ketohexokinase (KHK)

1 (SDH1)

Enolase 1 (ENO1)

Alpha enolase

Peptidyl-prolyl cis-trans isomerase A

Citrate synthase Triosephosphate Isomerise Ubiquinolcytochrome c reductase core protein I

Chaperonin containing TCP1

ER protein ERp29 precursor(Erp31) Transitional endoplasmic reticulum ATPase

> Complement C3 Precursor

Phosphoserine phosphatise

Contrapsin-like protease inhibitor 1 and 3 precursor

Succinate dehydrogenase Fructose metabolism

Aldolase 2 (ALDO2) Glycolysis

Aconitase 2 (ACO2) Kreb's cycle

Gamma enolase Glycolysis pathway Vimentin Morphogenesis Creatine kinase, B chain Energy transduction Guanine deaminase Purine catabolism

Annexin A2 and A5 Ca2+-regulated

**Light period (rodent sleep time)** Malate Dehydrogenase Krebs cycle

RNA recognition motif Processing of pre-mRNAs

**Tissue Methods Proteins with circadian** 

Mice were entrained under 12:12 L:D and then transferred to 12:12 D:D (constant darkness). Liver tissue was collected every 4-h over 24-h starting at CT 0 (n = 3/ time point)

2D-DIGE MALDI-TOF MS LC MS/MS

Rats were entrained under 12:12 L:D. Pineal glands were collected at ZT6 (6 hours after lights on) and ZT18 (6 hours after lights off) (n = 8/ time point)

2D-PAGE MALDI-TOF MS

Liver

CD1 mice

Pineal gland

Wistar rats

#### **3.6 Role of proteins identified in TAC heart disease**

Our laboratory is interested in circadian cardiovascular proteomics to better understand molecular processes underlying heart disease and clinical treatments. The TAC upregulated spots 519, 615 and 1012 were identified by MS as Succinyl-CoA:3-ketoacid-coenzyme A transferase 1, mitochondrial (SCOT), Desmin (DESM), and PDZ and LIM domain protein 1 (PDLIM1) respectively. Three examples from ZT23 (one hour before lights on, murine sleeptime) were shown here; other proteins were identified from different times of day or night. Figure 4 shows a representative mass spectrum for one of the identified proteins, SCOT (spot #519), as well as the number and list of tryptic hits, Mascot score, E-value, and protein sequence. Functionally, SCOT is a mitochondrial matrix protein and is the key rate limiting enzyme for ketone body metabolism (Fukao et al., 2004; Orii et al., 2008). That it exhibits increased expression around sleep time suggests that it plays a role in changing cardiac energy sources. The second identified protein, DESMIN (spot # 615) is crucial for muscle structure and function, as reviewed in (Paulin & Li, 2004). The third protein, PDLIM1 (spot #1012) is a cytoskeletal protein involved in cardiac contractility and cardioprotection, as reviewed in (Arias-Loza et al., 2008; Johnsen et al., 2009).

#### **3.7 Validation: Western blotting**

Candidate circadian biomarkers identified by 2D-DIGE and MS are validated by Western blot protein expression analyses using an independent set of cytoplasmic soluble proteins from TAC and sham hearts. Proteins (20 g) are separated according to their mass by 12% SDS-PAGE and transferred to polyvinylidene fluoride membrane (Bio-Rad) using a semidry transfer apparatus (Bio-Rad). Membranes are blocked for 2 hours at room temperature with 5% non-fat dry milk in TBS-T 0.05 % (20 mM Tris Base, 137mM NaCl, 0.05 % Tween 20, pH 7.6) and incubated overnight at 4 0C with primary antibodies against SCOT, DESM or PDLIM1. The antibodies are diluted according to the manufacturer's instructions. Anti-actin antibody (1:40000, Milipore) is used as a loading control. Immunoreactive protein bands are visualized with horseradish peroxidise-conjugated secondary antibodies (1:5000, Sigma) and ECLplus reagent (GE Healthcare). Blots are scanned using Storm 860 molecular imager (GE Healthcare) and protein expression is quantified by Image J software (NIH). Expression levels of SCOT, DESM and PDLIM1 in TAC samples at ZT23 are compared to sham, thus independently validating 2D-DIGE approach. Validating the candidate proteins/biomarkers in human samples will increase their significance and is key to translational applications.

#### **4. Circadian proteomes in other body organs**

Our time-of-day circadian approach led to discovery of SCOT, DESMIN, and PDLIM which were upregulated in TAC cardiac hypertrophy at sleep time. Other groups have also met with success investigating circadian proteomics in different tissues and clinical paradigms. A summary of identified circadian proteomes is shown in Table 1. The first identification of a circadian proteome was reported by Reddy and colleagues (Reddy et al., 2006), using a 2D-DIGE/MS approach to study murine liver. Mice were entrained to 12:12 L:D cycles, then placed in constant darkness (12:12 D:D) and sacrificed at 6 consecutive time-points 4 hours apart. It was observed that 135 (21%) out of 642 detected protein spots rhythmically cycled over the 24-hour period. Many of the newly identified circadian proteins were key rate limiting enzymes including ketohexokinase, succinate dehydrogenase 1, aldolase 2, enolase 1/aconitase 2, carbamoyl phosphate synthetase 1, CPS1, arginosuccinate synthetase 1, and arginase 1.

Our laboratory is interested in circadian cardiovascular proteomics to better understand molecular processes underlying heart disease and clinical treatments. The TAC upregulated spots 519, 615 and 1012 were identified by MS as Succinyl-CoA:3-ketoacid-coenzyme A transferase 1, mitochondrial (SCOT), Desmin (DESM), and PDZ and LIM domain protein 1 (PDLIM1) respectively. Three examples from ZT23 (one hour before lights on, murine sleeptime) were shown here; other proteins were identified from different times of day or night. Figure 4 shows a representative mass spectrum for one of the identified proteins, SCOT (spot #519), as well as the number and list of tryptic hits, Mascot score, E-value, and protein sequence. Functionally, SCOT is a mitochondrial matrix protein and is the key rate limiting enzyme for ketone body metabolism (Fukao et al., 2004; Orii et al., 2008). That it exhibits increased expression around sleep time suggests that it plays a role in changing cardiac energy sources. The second identified protein, DESMIN (spot # 615) is crucial for muscle structure and function, as reviewed in (Paulin & Li, 2004). The third protein, PDLIM1 (spot #1012) is a cytoskeletal protein involved in cardiac contractility and cardioprotection, as

Candidate circadian biomarkers identified by 2D-DIGE and MS are validated by Western blot protein expression analyses using an independent set of cytoplasmic soluble proteins from TAC and sham hearts. Proteins (20 g) are separated according to their mass by 12% SDS-PAGE and transferred to polyvinylidene fluoride membrane (Bio-Rad) using a semidry transfer apparatus (Bio-Rad). Membranes are blocked for 2 hours at room temperature with 5% non-fat dry milk in TBS-T 0.05 % (20 mM Tris Base, 137mM NaCl, 0.05 % Tween 20, pH 7.6) and incubated overnight at 4 0C with primary antibodies against SCOT, DESM or PDLIM1. The antibodies are diluted according to the manufacturer's instructions. Anti-actin antibody (1:40000, Milipore) is used as a loading control. Immunoreactive protein bands are visualized with horseradish peroxidise-conjugated secondary antibodies (1:5000, Sigma) and ECLplus reagent (GE Healthcare). Blots are scanned using Storm 860 molecular imager (GE Healthcare) and protein expression is quantified by Image J software (NIH). Expression levels of SCOT, DESM and PDLIM1 in TAC samples at ZT23 are compared to sham, thus independently validating 2D-DIGE approach. Validating the candidate proteins/biomarkers in human samples will increase their significance and is key to translational applications.

Our time-of-day circadian approach led to discovery of SCOT, DESMIN, and PDLIM which were upregulated in TAC cardiac hypertrophy at sleep time. Other groups have also met with success investigating circadian proteomics in different tissues and clinical paradigms. A summary of identified circadian proteomes is shown in Table 1. The first identification of a circadian proteome was reported by Reddy and colleagues (Reddy et al., 2006), using a 2D-DIGE/MS approach to study murine liver. Mice were entrained to 12:12 L:D cycles, then placed in constant darkness (12:12 D:D) and sacrificed at 6 consecutive time-points 4 hours apart. It was observed that 135 (21%) out of 642 detected protein spots rhythmically cycled over the 24-hour period. Many of the newly identified circadian proteins were key rate limiting enzymes including ketohexokinase, succinate dehydrogenase 1, aldolase 2, enolase 1/aconitase 2,

carbamoyl phosphate synthetase 1, CPS1, arginosuccinate synthetase 1, and arginase 1.

**3.6 Role of proteins identified in TAC heart disease** 

reviewed in (Arias-Loza et al., 2008; Johnsen et al., 2009).

**4. Circadian proteomes in other body organs** 

**3.7 Validation: Western blotting** 


Circadian Proteomics and Its Unique Advantage for Discovery of Biomarkers of Heart Disease 75

The circadian proteome in the rat pineal gland was identified using 2D-PAGE, silver stain and tandem mass spectrometry (MS/MS) (Moller et al., 2007). In this study, rats were entrained to 12:12 L:D cycle. Pineal glands were collected at two timepoints: ZT06 which is 6 hours after lights on (light phase, rats asleep) and ZT18 which is 6 hours after lights off (dark phase, rats awake). A total of 1737 pineal gland proteins were detected, 35 showed greater abundance during sleep time and 25 during wake time. Proteins upregulated during wake time were involved in glycolysis (alpha- and gamma-enolase), morphogenesis (vimentin), energy transduction (Creatine kinase), purine catabolism (guanine deaminase), protein folding (peptidyl-prolyl cis–trans isomerase A), and Ca2+-dependent membrane binding (annexin A2, annexin A5). Proteins with increased abundance during sleep mapped to the Krebs cycle (malate dehydrogenase, citrate synthase, triosephosphate isomerise), mitochondrial electron transport (ubiquinol-cytochrome c reductase core protein 1), RNA binding and processing (RNA recognition motif), protein folding (chaperone-containing TCP1, ER protein ERp29 precursor), cell transport (transitional ER ATPase), complement (C3 precursor), and metabolism (phosphoserine phosphotase, contrapsin-like protease

Cytidine Sarcosine Carnitine Valine Tryptophan 4-Guanidinobutyrate Isoleucine 3-Methylhistidine Leucine Proline Guanidoacetate 1-Methylnicotinamide Citrulline Creatinine Glycine Methionine sulfoxide a-Aminoadipate Methionine Phenylalanine N,N-Dimethylglycine Thr 13C Threonine Ornithine Hydroxyproline Creatine Corticosterone lysophosphatidylcoline

**Function Reference** 

Trimethylamine N-oxide (Minami et Glutamine 2-Aminobutyrate al., 2009)

**expression pattern** 

**Tissue Methods Proteins with circadian** 

Mice entrained under 12:12 L:D cycles, then transferred to either 12:12L:D or 12:12 D:D. Blood collected in fasting animals over 24-h starting at ZT 4 or CT 4

> LC-MS CE-MS

Table 1. Circadian proteomes.

inhibitor 1 and 3 precursors).

Blood metabolome

C57BL/6 and CBA/N mice



Table 1. Circadian proteomes.

74 Proteomics – Human Diseases and Protein Functions

**Function Reference** 

et al. 2007)

al., 2008)

Ref Martino same one as below

(Martino, Tata et al.

Vesicle transport (Tsuji, Hirota

Calcium binding

Photoreceptor adaptation, gating of photic input

RNA-binding

phototransduction, morphological changes

Protein degradation

Angiotensin I (Hatcher et Arginine vasopressin al., 2008) b

**expression pattern** 

N-ethylmaleimidesensitive fusion protein

Charged Multivesicular body protein 4b Reticulocalbin-2 precursor

Calbindin D28

Heterogeneous ribonucleoprotein A/B

Proteasome subunit alpha type 1

precursor

Apolipoprotein E precursor Apolipoprotein J

Complement C3 precursor

T-complex 1 subunit delta Protein folding,

Leukotriene A4 hydrolase Metabolism

Proenkephalin 219–229 Galanin Neurokinin-B Neurotensin POMC, melanotropin PEN Big LEN Little SAAS Somatostatin-14b proSomatostatin 89-100 Substance P Thymosin -4

Cytochrome C (CYCS) Apoptosis (Martino et

novo proteins cycle in the blood, and that the daily rhythmic variation changes in heart disease

and retinol (vitamin A)

Lipid and cholesterol regulation

Activation of compliment system

Fingerprinting assay Proof of concept that de-

Transthyretin Transports thyroxin (T4)

Plasminogen Plasmin formation

Apolipoprotein A1 2007)

**Tissue Methods Proteins with circadian** 

Mice were entrained under 12:12 L:D and transferred to 12:12 D:D. Retinal tissue was collected every 6-h over 24-h starting at CT 2 (n = 5/ time point)

2D-PAGE MALDI-TOF MS

Rats were entrained under 12:12 L:D and transferred to 12:12 D:D. Samples were taken every 4-h over 24-h starting at CT 0

LC MALDI-TOF MS LC MS/MS

Hamsters entrained under 14:10 L:D cycle.

> SDS-PAGE LC MS/MS

Mice entrained to 12:12 L:D cycle SELDI

Mice entrained under 12:12 L:D cycle. Samples collected every 4 hours, starting at ZT 23 (n = 3/ time point)

> SDS-PAGE LC MS/MS

Retina

C57BL/6 mice

SCN releasate

Long-Evans/BluGill rats

Urine

*Tau* mutant hamsters

> Blood C57Bl/6

> > Blood

C57BL/6 mice

The circadian proteome in the rat pineal gland was identified using 2D-PAGE, silver stain and tandem mass spectrometry (MS/MS) (Moller et al., 2007). In this study, rats were entrained to 12:12 L:D cycle. Pineal glands were collected at two timepoints: ZT06 which is 6 hours after lights on (light phase, rats asleep) and ZT18 which is 6 hours after lights off (dark phase, rats awake). A total of 1737 pineal gland proteins were detected, 35 showed greater abundance during sleep time and 25 during wake time. Proteins upregulated during wake time were involved in glycolysis (alpha- and gamma-enolase), morphogenesis (vimentin), energy transduction (Creatine kinase), purine catabolism (guanine deaminase), protein folding (peptidyl-prolyl cis–trans isomerase A), and Ca2+-dependent membrane binding (annexin A2, annexin A5). Proteins with increased abundance during sleep mapped to the Krebs cycle (malate dehydrogenase, citrate synthase, triosephosphate isomerise), mitochondrial electron transport (ubiquinol-cytochrome c reductase core protein 1), RNA binding and processing (RNA recognition motif), protein folding (chaperone-containing TCP1, ER protein ERp29 precursor), cell transport (transitional ER ATPase), complement (C3 precursor), and metabolism (phosphoserine phosphotase, contrapsin-like protease inhibitor 1 and 3 precursors).

Circadian Proteomics and Its Unique Advantage for Discovery of Biomarkers of Heart Disease 77

represents just chemical waste generated from muscle metabolism and is only an indirect marker of renal function. Here we identified a direct indicator of early and ongoing tissue damage, by investigating circadian proteome expression. This marker is easily obtained by collecting urine samples, thus making it potentially ideal for point-of-care or routine diagnostics. Additional examples of circadian-based biomarkers, including many

Circadian proteomics can be used for discovery of de-novo blood biomarkers of heart disease. Established cardiovascular factors (e.g. hypertension, smoking, diabetes) do not fully explain the risk for heart disease, and there is substantial interest in the development of new biomarkers to identify persons at risk and who may be targeted for preventative measures. In those with heart disease, biomarkers are in demand to help track disease progression and aid physicians in better treating their patients. Although a number of strategies are currently used to identify biomarkers, there have been very few clinical advances, and thus new approaches such as circadian proteomics

As an example, we recently investigated diurnal proteome cycling in murine blood (Martino et al., 2007). Blood was the preferred tissue for biomedical investigation because of its ease of accessibility and minimal invasiveness for sampling. As shown in Table 1, the first approach was proof of concept and used surface-enhanced laser desorption ionization (SELDI) MS. Only proteins retained on ion exchange solid phase chromatographic surfaces (or chips) were examined; substances with other biochemical properties remain open to future investigation. Expression profiles were collected over a wide mass range; those of lower molecular weight (1-10 kDa range) classically contained bioactive peptides, while those in the midrange (10-50 kDa) and larger (>50 kDa) reflected peptides/proteins involved in cell structural and functional processes. With SELDI MS we essentially created a fingerprint of blood protein expression over the 24-hour time, from which one could then quantify, statistically analyze and graph data to visualize daily

Since one of the drawbacks of SELDI MS was a relative inability to further identify the fingerprint proteins, the diurnal blood proteome was also characterized following prefractionation on column chromatography with an ion exchange resin (Martino et al., 2007). As shown in Table 1, proteins comprising effluent, salt elution, and bead retentive fractions were visualized by SDS-PAGE and silver stain. Protein bands that exhibited cyclic variation over 24 hours were excised, trypsin digested, and then injected by electro spray ionization (ESI) into a LCQ DECA XP ion trap. Proteins were identified by comparison searching molecular mass against murine databases. Many of the identified blood proteins with diurnal expression linked to cycles in physiology were those released from liver, such as transthyretin, apolipoprotein A1 precursor, apolipoprotein E precursor, apolipoprotein J, plasminogen and complement C3 precursor. Ultimately, comparing daily rhythms in sera from healthy individuals vs. heart disease patients would allow for the creation of new

Another study examining diurnal biomarkers in blood measured 24-hour profiles of small chemical substances (peptides, amino acids, hormones) (Minami et al., 2009). As shown in Table 1, blood was drawn every 4 hours over the 24-hour period from CBA mice maintained

fundamental to human health and disease, are described below.

**6. Circadian biomarkers in blood** 

are warranted.

protein rhythms.

biomarker profiles and discovery platforms.

The circadian proteome of the mouse retina was characterized by 2D-PAGE and Coomassie Brilliant stain (Tsuji et al., 2007). C57Bl/6 mice were entrained at 12:12 L:D cycle, then placed in constant darkness (12:12 D:D) and sacrificed on the fourth day at 4 timepoints: CT2, CT8, CT14, and CT20. CT2 corresponded to subjective dawn and CT12 to subjective dusk. A total of 415 protein spots were detected and 11 exhibited circadian rhythms. The cycling proteins were important in vesicular transport (N-ethylmaleimide-sensitive fusion protein, charged multivesicular body protein 4b), calcium-binding (calbindin D28, reticulocalbin-2 precursor), RNA-binding (heterogeneous ribonucleoprotein A/B), protein folding (T-complex 1 delta), metabolism (leukotriene A4 hydrolase), and protein degradation (proteasome subunit alpha1).

The circadian peptidome released from rat hypothalamic SCN was characterized using a gel-free approach (Hatcher et al., 2008). Long-Evans/BluGill rats were entrained to 12:12 L:D and then transferred to constant darkness. Animals were sacrificed during subjective day-time (CT 0-12) and brain slices containing SCN and optic nerve were prepared. SCN releasates were obtained from extracellular fluid with pipettes containing solid-phase extraction material or from the SCN itself with micrometer-sized beads with solid-phase extraction material. Samples of releasate were collected at the beginning and end of CT 0–4, 4–8, 8–12, 12–16, 16–20, and 20–24 intervals and analyzed with MALDI TOF MS. Identified peptides were independently verified with LC-MS/MS. Arginine vasopressin showed a robust circadian rhythm with a peak during sleep time. Other identified peptides were angiotensin I, arginine vasopressinb, proenkephalin, galanin, neurokinin-B, neurotensin, melanotropin , PEN, big LEN, little SAAS, somatostatin-14b, proSomatostatin, substance P and thymosin -4. Interestingly, this study also detected previously unknown peptides, which prompted further investigation of their biological role.

#### **5. The circadian proteome helps to understand disease and identify new biomarkers**

Disturbances of the circadian rhythm, such as might occur in humans in shift work or sleep disorders, can affect many physiologic processes (e.g. circadian rhythms of heart rate, body temperature, sympathetic nervous activity, chemical, inflammatory, and metabolic processes). The proteome should characteristically change as well, revealing de novo biomarkers. We investigated this using a +/tau hamster model bearing a mutation in the core circadian clockwork protein casein kinase 1 epsilon (Martino et al., 2008). The circadian rhythm disruption in these animals etiologically caused heart and kidney disease. The animals exhibited profound proteinuria. To detect proteomic changes, urine samples collected from +/tau vs. controls were analyzed on 10-20% tricine gels stained with Coomassie dye. Bands were excised from the gel, trypsin digested, subjected to MS/MS on a LCQ DECA XP ion trap, and analyzed using Sequest. Protein identification was validated by Western blot. As shown in Table 1, a ~15 kDa protein band, appearing only in the urine of the +/tau animals, was identified by MS as cytochrome c, a biomarker of cellular apoptosis. Apoptosis was confirmed in the renal tissues of +/tau mutants by terminal uridine deoxynucleotidyl transferase dUTP nick end labelling (TUNEL) staining. Thus by using circadian approaches, a novel biomarker of renal disease was discovered. This was a particularly exciting discovery as there are notably very few biomarkers of developing renal disease. For example, a usual biomarker for kidney disease is creatinine, however, this

represents just chemical waste generated from muscle metabolism and is only an indirect marker of renal function. Here we identified a direct indicator of early and ongoing tissue damage, by investigating circadian proteome expression. This marker is easily obtained by collecting urine samples, thus making it potentially ideal for point-of-care or routine diagnostics. Additional examples of circadian-based biomarkers, including many fundamental to human health and disease, are described below.

### **6. Circadian biomarkers in blood**

76 Proteomics – Human Diseases and Protein Functions

The circadian proteome of the mouse retina was characterized by 2D-PAGE and Coomassie Brilliant stain (Tsuji et al., 2007). C57Bl/6 mice were entrained at 12:12 L:D cycle, then placed in constant darkness (12:12 D:D) and sacrificed on the fourth day at 4 timepoints: CT2, CT8, CT14, and CT20. CT2 corresponded to subjective dawn and CT12 to subjective dusk. A total of 415 protein spots were detected and 11 exhibited circadian rhythms. The cycling proteins were important in vesicular transport (N-ethylmaleimide-sensitive fusion protein, charged multivesicular body protein 4b), calcium-binding (calbindin D28, reticulocalbin-2 precursor), RNA-binding (heterogeneous ribonucleoprotein A/B), protein folding (T-complex 1 delta), metabolism (leukotriene A4 hydrolase), and protein degradation

The circadian peptidome released from rat hypothalamic SCN was characterized using a gel-free approach (Hatcher et al., 2008). Long-Evans/BluGill rats were entrained to 12:12 L:D and then transferred to constant darkness. Animals were sacrificed during subjective day-time (CT 0-12) and brain slices containing SCN and optic nerve were prepared. SCN releasates were obtained from extracellular fluid with pipettes containing solid-phase extraction material or from the SCN itself with micrometer-sized beads with solid-phase extraction material. Samples of releasate were collected at the beginning and end of CT 0–4, 4–8, 8–12, 12–16, 16–20, and 20–24 intervals and analyzed with MALDI TOF MS. Identified peptides were independently verified with LC-MS/MS. Arginine vasopressin showed a robust circadian rhythm with a peak during sleep time. Other identified peptides were angiotensin I, arginine vasopressinb, proenkephalin, galanin, neurokinin-B, neurotensin, melanotropin , PEN, big LEN, little SAAS, somatostatin-14b, proSomatostatin, substance P and thymosin -4. Interestingly, this study also detected previously unknown peptides,

**5. The circadian proteome helps to understand disease and identify new** 

Disturbances of the circadian rhythm, such as might occur in humans in shift work or sleep disorders, can affect many physiologic processes (e.g. circadian rhythms of heart rate, body temperature, sympathetic nervous activity, chemical, inflammatory, and metabolic processes). The proteome should characteristically change as well, revealing de novo biomarkers. We investigated this using a +/tau hamster model bearing a mutation in the core circadian clockwork protein casein kinase 1 epsilon (Martino et al., 2008). The circadian rhythm disruption in these animals etiologically caused heart and kidney disease. The animals exhibited profound proteinuria. To detect proteomic changes, urine samples collected from +/tau vs. controls were analyzed on 10-20% tricine gels stained with Coomassie dye. Bands were excised from the gel, trypsin digested, subjected to MS/MS on a LCQ DECA XP ion trap, and analyzed using Sequest. Protein identification was validated by Western blot. As shown in Table 1, a ~15 kDa protein band, appearing only in the urine of the +/tau animals, was identified by MS as cytochrome c, a biomarker of cellular apoptosis. Apoptosis was confirmed in the renal tissues of +/tau mutants by terminal uridine deoxynucleotidyl transferase dUTP nick end labelling (TUNEL) staining. Thus by using circadian approaches, a novel biomarker of renal disease was discovered. This was a particularly exciting discovery as there are notably very few biomarkers of developing renal disease. For example, a usual biomarker for kidney disease is creatinine, however, this

which prompted further investigation of their biological role.

(proteasome subunit alpha1).

**biomarkers** 

Circadian proteomics can be used for discovery of de-novo blood biomarkers of heart disease. Established cardiovascular factors (e.g. hypertension, smoking, diabetes) do not fully explain the risk for heart disease, and there is substantial interest in the development of new biomarkers to identify persons at risk and who may be targeted for preventative measures. In those with heart disease, biomarkers are in demand to help track disease progression and aid physicians in better treating their patients. Although a number of strategies are currently used to identify biomarkers, there have been very few clinical advances, and thus new approaches such as circadian proteomics are warranted.

As an example, we recently investigated diurnal proteome cycling in murine blood (Martino et al., 2007). Blood was the preferred tissue for biomedical investigation because of its ease of accessibility and minimal invasiveness for sampling. As shown in Table 1, the first approach was proof of concept and used surface-enhanced laser desorption ionization (SELDI) MS. Only proteins retained on ion exchange solid phase chromatographic surfaces (or chips) were examined; substances with other biochemical properties remain open to future investigation. Expression profiles were collected over a wide mass range; those of lower molecular weight (1-10 kDa range) classically contained bioactive peptides, while those in the midrange (10-50 kDa) and larger (>50 kDa) reflected peptides/proteins involved in cell structural and functional processes. With SELDI MS we essentially created a fingerprint of blood protein expression over the 24-hour time, from which one could then quantify, statistically analyze and graph data to visualize daily protein rhythms.

Since one of the drawbacks of SELDI MS was a relative inability to further identify the fingerprint proteins, the diurnal blood proteome was also characterized following prefractionation on column chromatography with an ion exchange resin (Martino et al., 2007). As shown in Table 1, proteins comprising effluent, salt elution, and bead retentive fractions were visualized by SDS-PAGE and silver stain. Protein bands that exhibited cyclic variation over 24 hours were excised, trypsin digested, and then injected by electro spray ionization (ESI) into a LCQ DECA XP ion trap. Proteins were identified by comparison searching molecular mass against murine databases. Many of the identified blood proteins with diurnal expression linked to cycles in physiology were those released from liver, such as transthyretin, apolipoprotein A1 precursor, apolipoprotein E precursor, apolipoprotein J, plasminogen and complement C3 precursor. Ultimately, comparing daily rhythms in sera from healthy individuals vs. heart disease patients would allow for the creation of new biomarker profiles and discovery platforms.

Another study examining diurnal biomarkers in blood measured 24-hour profiles of small chemical substances (peptides, amino acids, hormones) (Minami et al., 2009). As shown in Table 1, blood was drawn every 4 hours over the 24-hour period from CBA mice maintained

Circadian Proteomics and Its Unique Advantage for Discovery of Biomarkers of Heart Disease 79

under L:D or D:D conditions and LC-MS analysis was performed that led to the detection of 176 negative and 142 positive ion peaks. Importantly, in transgenic murine model of clock disruption (Cry1-/- and Cry2-/-), expression of these metabolites was altered, suggesting that they were controlled by the clock mechanism. The authors demonstrated that their metabolite timetable could be used to accurately determine body time in mice with different genetic background, age, sex and feeding regime. The overall purpose of doing this was to create maps of body time, which could be applied clinically to optimize understanding of disease or determine the best times for administering drugs and

Long before protein cycling was even discovered, it was already known that some key neuroendocrine hormones important to the heart had daily rhythmic patterns of expression. Cycling of many of these hormones drives the protein rhythms we observe in normal heart and other tissues, and the changes that occur in disease. Further details are described below

The first example is melatonin, and its expression in humans is elevated in the dark but suppressed by light (thus it is sometimes called the "hormone of darkness") (Brzezinski, 1997). It has a cardioprotective effect through antioxidant and antiinflammatory activity, reviewed by (Tengattini et al., 2008). It is a useful circadian biomarker of heart disease. Nocturnal melatonin levels were low in patients with acute myocardial infarction (MI) vs. healthy controls within the first 24 hours after hospital admission (Dominguez-Rodriguez et al., 2002). Also, patients with ST-segment MI who developed adverse events during follow-up had significantly lower nocturnal melatonin levels than patients without the events (Dominguez-Rodriguez et al., 2006). In another study it was revealed that nocturnal melatonin levels were reduced in patients with cardiac syndrome X vs. controls (Altun et al., 2002). Second, the renin-angiotensinaldosterone system (RAAS) also exhibits circadian rhythm. Aldosterone produced by the adrenal cortex regulates Na+ and K+ homeostasis. Aldosterone plasma levels peak during sleep time (11:00 P.M. – 7:00 A.M.) (Charloux et al., 1999). As a biomarker of heart disease, it has been reported that while normal subjects showed an 81% decrease during the wake period, there was only a 40% decrease in low renin hypertensive patients (Grim et al., 1974). Similarly, plasma renin activity (PRA) cycles in healthy subjects with peak activity during sleep (Charloux et al., 1999). Patients with hypertensive heart disease exhibited a greater increase in PRA during sleep-time as compared to normotensive patients (Tuck et

Finally, growth hormone secretion by the anterior pituitary gland is pulsatile and has a circadian rhythm with the highest peak occurring at midnight (Surya et al., 2006). Growth hormone is involved in regulation of cardiac metabolism and contractility, reviewed by (Volterrani et al., 2000). In some heart failure patients (New York Heart Association classes I-III) there is a loss of circadian rhythm and overall reduction of growth hormone secretion

**7. Circadian rhythms in neuro/hormone biomarkers** 

therapies.

al., 1985).

(Duncan et al., 2003).

and in Figure 5 and Table 2.

**7.1 Sleep-time hormones** 

A) Representative mass spectrum of the SCOT peptide: RGGHVNLTMLGAMQVSKY. B) A total of 24 peptides were identified, corresponding to 13 unique sequences. The Mascot score is 150, E-value is 4.0 ×10-21. C) Identified peptides and their corresponding match to the SCOT protein sequence.

Fig. 4. MS based identification of Succinyl-CoA:3-ketoacid-coenzyme A transferase 1, mitochondrial (SCOT).

under L:D or D:D conditions and LC-MS analysis was performed that led to the detection of 176 negative and 142 positive ion peaks. Importantly, in transgenic murine model of clock disruption (Cry1-/- and Cry2-/-), expression of these metabolites was altered, suggesting that they were controlled by the clock mechanism. The authors demonstrated that their metabolite timetable could be used to accurately determine body time in mice with different genetic background, age, sex and feeding regime. The overall purpose of doing this was to create maps of body time, which could be applied clinically to optimize understanding of disease or determine the best times for administering drugs and therapies.

#### **7. Circadian rhythms in neuro/hormone biomarkers**

Long before protein cycling was even discovered, it was already known that some key neuroendocrine hormones important to the heart had daily rhythmic patterns of expression. Cycling of many of these hormones drives the protein rhythms we observe in normal heart and other tissues, and the changes that occur in disease. Further details are described below and in Figure 5 and Table 2.

#### **7.1 Sleep-time hormones**

78 Proteomics – Human Diseases and Protein Functions

A) Representative mass spectrum of the SCOT peptide: RGGHVNLTMLGAMQVSKY. B) A total of 24 peptides were identified, corresponding to 13 unique sequences. The Mascot score is 150, E-value is 4.0 ×10-21. C) Identified peptides and their corresponding match to the SCOT protein sequence. Fig. 4. MS based identification of Succinyl-CoA:3-ketoacid-coenzyme A transferase 1,

mitochondrial (SCOT).

The first example is melatonin, and its expression in humans is elevated in the dark but suppressed by light (thus it is sometimes called the "hormone of darkness") (Brzezinski, 1997). It has a cardioprotective effect through antioxidant and antiinflammatory activity, reviewed by (Tengattini et al., 2008). It is a useful circadian biomarker of heart disease. Nocturnal melatonin levels were low in patients with acute myocardial infarction (MI) vs. healthy controls within the first 24 hours after hospital admission (Dominguez-Rodriguez et al., 2002). Also, patients with ST-segment MI who developed adverse events during follow-up had significantly lower nocturnal melatonin levels than patients without the events (Dominguez-Rodriguez et al., 2006). In another study it was revealed that nocturnal melatonin levels were reduced in patients with cardiac syndrome X vs. controls (Altun et al., 2002). Second, the renin-angiotensinaldosterone system (RAAS) also exhibits circadian rhythm. Aldosterone produced by the adrenal cortex regulates Na+ and K+ homeostasis. Aldosterone plasma levels peak during sleep time (11:00 P.M. – 7:00 A.M.) (Charloux et al., 1999). As a biomarker of heart disease, it has been reported that while normal subjects showed an 81% decrease during the wake period, there was only a 40% decrease in low renin hypertensive patients (Grim et al., 1974). Similarly, plasma renin activity (PRA) cycles in healthy subjects with peak activity during sleep (Charloux et al., 1999). Patients with hypertensive heart disease exhibited a greater increase in PRA during sleep-time as compared to normotensive patients (Tuck et al., 1985).

Finally, growth hormone secretion by the anterior pituitary gland is pulsatile and has a circadian rhythm with the highest peak occurring at midnight (Surya et al., 2006). Growth hormone is involved in regulation of cardiac metabolism and contractility, reviewed by (Volterrani et al., 2000). In some heart failure patients (New York Heart Association classes I-III) there is a loss of circadian rhythm and overall reduction of growth hormone secretion (Duncan et al., 2003).

Circadian Proteomics and Its Unique Advantage for Discovery of Biomarkers of Heart Disease 81

**In healthy humans** 

Nocturnal peak in blood (in the dark)

Secretion is suppressed by light and activated in the dark

Peak serum levels during sleep time

Significantly higher plasma activity during sleep time

Nocturnal plasma peak

Exhibits pulsatile expression

**Hormone Cycling References** 

**\* Used in Figure 5 § Heart disease # Additional references** 

**\*** (Brzezinski, 1997) **#** (Claustrat et al., 1986) **#** (Follenius et al., 1995) **§** (Dominguez-Rodriguez et al., 2006) **§** (Dominguez-Rodriguez et al., 2002) **§** (Altun et al., 2002)

**\*** (Charloux et al., 1999) **#** (Katz et al., 1975) **#** (Lightman et al., 1981) **§** (Grim et al., 1974)

**\*** (Charloux et al., 1999) **#** (Katz et al., 1975) **§** (Tuck et al., 1985)

**\*** (Surya et al., 2006) **#** (Takahashi et al., 1968) **#** (Hartman et al., 1991) **§** (Duncan et al., 2003)

**Altered cycling in specified cardiovascular disease** 

Nocturnal decrease observed within the first 24h period in patients with myocardial infarct (MI)

Low levels of melatonin correlate with adverse events during follow up in ST-segment MI patients

Nocturnal decrease in patients with cardiac syndrome X vs. healthy controls

Does not decrease normally during the day in low renin hypertensive patients vs. healthy controls

Significantly upregulated only at night in some patients with essential hypertension

Overall decrease and loss of circadian rhythm in some patients with chronic heart failure (NY Heart Association classes I-III)

**Hormone Selective** 

**Melatonin** Antioxidant,

**Aldosterone** Regulates

**Plasma Renin Activity** 

> **Growth Hormone**

Na+/K+ homeostasis, blood pressure

Regulates blood pressure

Affects metabolism and contractility

**function(s) relevant to the cardiovascular system** 

circadian entrainment

Left; Melatonin, aldosterone, plasma renin activity, growth hormone peak during sleep time. Right; cortisol, epinephrine, vasoactive intestinal peptide peak during wake time. X-axis: white bar = light period, black bar = dark. Rhythms illustrated here are based on the references listed in Table 2.

Fig. 5. Dirunal cycling of hormones in blood.

Left; Melatonin, aldosterone, plasma renin activity, growth hormone peak during sleep time. Right; cortisol, epinephrine, vasoactive intestinal peptide peak during wake time. X-axis: white bar = light period, black bar = dark. Rhythms illustrated here are based on the references listed in Table 2.

Fig. 5. Dirunal cycling of hormones in blood.


Circadian Proteomics and Its Unique Advantage for Discovery of Biomarkers of Heart Disease 83

Discovery of the circadian cardiovascular proteome and it's endogenous drivers provides a new understanding of cardiovascular health and disease, where time is a new paradigm of functional significance. Proteomic biomarkers can be easily implemented with ELISA-type point-of-care diagnostic platforms that could be routinely applied in physician offices, or even potentially in the consumer's home. One of the practical applications of cardiovascular circadian proteomics is chronotherapy, which requires easily accessible markers of body time for optimizing the timing of drug treatments. For example, we recently demonstrated that the efficacy of treatment with the angiotensin converting enzyme inhibitor (ACEi) captopril exhibits a diurnal pattern, consistent with the diurnal variation in ACE expression (Martino et al., 2011). ACEi are common medications given to cardiovascular patients with hypertension, after a myocardial infarction or with heart failure. We found that drug administration at sleep-time improved heart function, but wake-time did not differ from placebo. This approach can be used in other diseases such as cancer (Hrushesky & Bjarnason, 1993; Innominato et al., 2010) and neuroendocrine disorders (Chung et al., 2011)

Circadian cardiovascular proteomics is an important new area of research that provides an excellent opportunity to elucidate molecular processes that underlie our health and disease across the 24-hour light/dark period. At this early stage, temporal analysis of the proteome in cardiovascular tissues (i.e. heart or blood) of experimental animal models reveals remarkable 24-hour variations in protein abundances. Diurnal protein profiles differ remarkably between health and disease. Characterization of these proteins is the key to understanding normal body physiology as well as providing new diagnostic capabilities, and new approaches to treatment by aiding in the design of personalized therapeutics.

Supported by a Grant from the Heart and Stroke Foundation (NA6466 to T.A. Martino).

Akashi, M., Tsuchiya, Y., Yoshino, T., & Nishida, E. (2002). Control of intracellular dynamics

Akerstedt, T., & Froberg, J. E. (1979). Sleep and stressor exposure in relation to circadian

Alonso-Fernandez, A., Garcia-Rio, F., Racionero, M. A., Pino, J. M., Ortuno, F., Martinez, I.,

mechanisms. *Chest,* Vol. 127, No. 1, pp. 15-22, issn 0012-3692

of mammalian period proteins by casein kinase I epsilon (CKIepsilon) and CKIdelta in cultured cells. *Mol Cell Biol,* Vol. 22, No. 6, pp. 1693-703, issn 0270-7306

rhythms in catecholamine excretion. *Biol Psychol,* Vol. 8, No. 1, pp. 69-80, issn 0301-

& Villamor, J. (2005). Cardiac rhythm disturbances and ST-segment depression episodes in patients with obstructive sleep apnea-hypopnea syndrome and its

**8. Clinical translation** 

as well.

**9. Conclusion** 

**10. Acknowledgments** 

**11. References** 

0511


Table 2. Circadian rhythms in neuro/hormones in humans

#### **7.2 Wake time hormones**

Cortisol exhibits diurnal variation with serum levels peaking early in the morning around wake time and troughing late in the evening (Charloux et al., 1999; Lightman et al., 1981). It may also be a useful biomarker or mediator of disease as plasma cortisol peaks earlier at 4:00 A.M. in some low renin hypertensive patients, as compared to 8:00 A.M. in normal controls (Grim et al., 1974). Epinephrine stimulates sympathetic activity including in the cardiovascular system. It exhibits a robust endogenous circadian rhythm that peaks during wake time (Linsell et al., 1985), and this expression changes in obstructive sleep apnea heart failure patients, in which there is increased nocturnal sympathetic tone (Alonso-Fernandez et al., 2005; Bradley & Floras, 2009). Lastly, vasoactive intestinal peptide affects vasodilation, heart rate and force of contraction, reviewed by (Henning & Sawmiller, 2001). It has a circadian rhythm with a peak before sleep time (20:00) and a trough later at night (0:00) (Cugini et al., 1991). The rhythm of this peptide is lost in patients with orthotopic heart transplant (Cugini et al., 1993).

#### **8. Clinical translation**

82 Proteomics – Human Diseases and Protein Functions

**In healthy humans** 

Serum levels peak around wake-time

Plasma levels peak during wake time

Plasma peak before sleep (20:00), trough at night (0:00)

Cortisol exhibits diurnal variation with serum levels peaking early in the morning around wake time and troughing late in the evening (Charloux et al., 1999; Lightman et al., 1981). It may also be a useful biomarker or mediator of disease as plasma cortisol peaks earlier at 4:00 A.M. in some low renin hypertensive patients, as compared to 8:00 A.M. in normal controls (Grim et al., 1974). Epinephrine stimulates sympathetic activity including in the cardiovascular system. It exhibits a robust endogenous circadian rhythm that peaks during wake time (Linsell et al., 1985), and this expression changes in obstructive sleep apnea heart failure patients, in which there is increased nocturnal sympathetic tone (Alonso-Fernandez et al., 2005; Bradley & Floras, 2009). Lastly, vasoactive intestinal peptide affects vasodilation, heart rate and force of contraction, reviewed by (Henning & Sawmiller, 2001). It has a circadian rhythm with a peak before sleep time (20:00) and a trough later at night (0:00) (Cugini et al., 1991). The rhythm of this peptide is lost in patients with orthotopic heart

**Hormone Cycling References** 

**\* Used in Figure 5 § Heart disease # Additional references** 

**\*** (Charloux et al., 1999) **\*** (Lightman et al., 1981) **#** (Lockinger et al., 2004) **#** (Cugini et al., 1991) **§** (Grim et al., 1974)

**\*** (Linsell et al., 1985) **#** (Scheer et al., 2009) **#** (Akerstedt & Froberg, 1979) **§** (Alonso-Fernandez et al., 2005)

**\*** (Cugini et al., 1991) **#** (Cugini et al., 1992) **§** (Cugini et al., 1993)

**Altered cycling in specified cardiovascular disease** 

Peaks at 4:00 A.M. in some low renin hypertensive patients

Nocturnal increase in some patients with obstructive sleep apnea vs. healthy controls

Circadian rhythm is lost in patients with orthotopic heart transplant

**Hormone Selective** 

**Cortisol** Regulates blood

**Epinephrine** Increases heart

**Vasoactive intestinal peptide** 

**7.2 Wake time hormones** 

transplant (Cugini et al., 1993).

**function(s) relevant to the cardiovascular system** 

sugar levels, increases blood pressure

rate and blood pressure

Mediates sympathetic response

Regulates vasodilation, heart rate, force of contraction

Table 2. Circadian rhythms in neuro/hormones in humans

Discovery of the circadian cardiovascular proteome and it's endogenous drivers provides a new understanding of cardiovascular health and disease, where time is a new paradigm of functional significance. Proteomic biomarkers can be easily implemented with ELISA-type point-of-care diagnostic platforms that could be routinely applied in physician offices, or even potentially in the consumer's home. One of the practical applications of cardiovascular circadian proteomics is chronotherapy, which requires easily accessible markers of body time for optimizing the timing of drug treatments. For example, we recently demonstrated that the efficacy of treatment with the angiotensin converting enzyme inhibitor (ACEi) captopril exhibits a diurnal pattern, consistent with the diurnal variation in ACE expression (Martino et al., 2011). ACEi are common medications given to cardiovascular patients with hypertension, after a myocardial infarction or with heart failure. We found that drug administration at sleep-time improved heart function, but wake-time did not differ from placebo. This approach can be used in other diseases such as cancer (Hrushesky & Bjarnason, 1993; Innominato et al., 2010) and neuroendocrine disorders (Chung et al., 2011) as well.

#### **9. Conclusion**

Circadian cardiovascular proteomics is an important new area of research that provides an excellent opportunity to elucidate molecular processes that underlie our health and disease across the 24-hour light/dark period. At this early stage, temporal analysis of the proteome in cardiovascular tissues (i.e. heart or blood) of experimental animal models reveals remarkable 24-hour variations in protein abundances. Diurnal protein profiles differ remarkably between health and disease. Characterization of these proteins is the key to understanding normal body physiology as well as providing new diagnostic capabilities, and new approaches to treatment by aiding in the design of personalized therapeutics.

#### **10. Acknowledgments**

Supported by a Grant from the Heart and Stroke Foundation (NA6466 to T.A. Martino).

#### **11. References**


Circadian Proteomics and Its Unique Advantage for Discovery of Biomarkers of Heart Disease 85

Dominguez-Rodriguez, A., Abreu-Gonzalez, P., Garcia-Gonzalez, M., & Reiter, R. J. (2006).

Dominguez-Rodriguez, A., Abreu-Gonzalez, P., Garcia, M. J., Sanchez, J., Marrero, F., & de

myocardial infarction. *J Pineal Res,* Vol. 33, No. 4, pp. 248-52, issn 0742-3098 Duncan, B., Moyna, N. M., Heller, G. V., McGill, C., Katten, D., Finta, L., Velusamy, M.*, et al.*

Durgan, D. J., & Young, M. E. (2010). The cardiomyocyte circadian clock: emerging roles in health and disease. *Circ Res,* Vol. 106, No. 4, pp. 647-58, issn 1524-4571 Eksik, A., Akyol, A., Norgaz, T., Aksu, H., Erdinler, I., Cakmak, N., Alper, A. T.*, et al.* (2007).

Follenius, M., Weibel, L., & Brandenberger, G. (1995). Distinct modes of melatonin secretion in normal men. *J Pineal Res,* Vol. 18, No. 3, pp. 135-40, issn 0742-3098 Fukao, T., Lopaschuk, G. D., & Mitchell, G. A. (2004). Pathways and control of ketone body

Goldberg, R. J., Brady, P., Muller, J. E., Chen, Z. Y., de Groot, M., Zonneveld, P., & Dalen, J.

Grim, C., Winnacker, J., Peters, T., & Gilbert, G. (1974). Low renin, "normal" aldosterone and

hormone. *J Clin Endocrinol Metab,* Vol. 39, No. 2, pp. 247-56, issn 0021-972X Guo, Y. F., & Stein, P. K. (2003). Circadian rhythm in the cardiovascular system: chronocardiology. *Am Heart J,* Vol. 145, No. 5, pp. 779-86, issn 1097-6744 Hartman, M. L., Faria, A. C., Vance, M. L., Johnson, M. L., Thorner, M. O., & Veldhuis, J. D.

Hastings, M. H., Reddy, A. B., & Maywood, E. S. (2003). A clockwork web: circadian timing

Hatcher, N. G., Atkins, N., Jr., Annangudi, S. P., Forbes, A. J., Kelleher, N. L., Gillette, M. U.,

Hobson, D. J., Rupa, P., Diaz, G. J., Zhang, H., Yang, M., Mine, Y., Turner, P. V.*, et al.* (2007).

Hrushesky, W. J., & Bjarnason, G. A. (1993). Circadian cancer therapy. *J Clin Oncol,* Vol. 11,

*Proc Natl Acad Sci U S A,* Vol. 105, No. 34, pp. 12527-32, issn 1091-6490 Henning, R. J., & Sawmiller, D. R. (2001). Vasoactive intestinal peptide: cardiovascular

effects. *Cardiovasc Res,* Vol. 49, No. 1, pp. 27-37, issn 0008-6363

*Am J Physiol,* Vol. 260, No. 1 Pt 1, pp. E101-10, issn 0002-9513

*Acids,* Vol. 70, No. 3, pp. 243-51, issn 0952-3278

Vol. 66, No. 2, pp. 140-4, issn 0002-9149

issn 0002-9149

52, issn 0277-0008

416, issn 1234-1010

649-61, issn 1471-003X

2372-80, issn 0278-6915

No. 7, pp. 1403-17, issn 0732-183X

Prognostic value of nocturnal melatonin levels as a novel marker in patients with ST-segment elevation myocardial infarction. *Am J Cardiol,* Vol. 97, No. 8, pp. 1162-4,

Armas-Trujillo, D. (2002). Decreased nocturnal melatonin levels during acute

(2003). A 24-hour comparison of serum growth hormone concentrations in patients with heart failure versus healthy controls. *Pharmacotherapy,* Vol. 23, No. 2, pp. 147-

Circadian pattern of spontaneous ventricular tachyarrhythmias in patients with implantable cardioverter defibrillators. *Med Sci Monit,* Vol. 13, No. 9, pp. CR412-

metabolism: on the fringe of lipid biochemistry. *Prostaglandins Leukot Essent Fatty* 

E. (1990). Time of onset of symptoms of acute myocardial infarction. *Am J Cardiol,*

hypertension: circadian rhythm of renin, aldosterone, cortisol and growth

(1991). Temporal structure of in vivo growth hormone secretory events in humans.

in brain and periphery, in health and disease. *Nat Rev Neurosci,* Vol. 4, No. 8, pp.

& Sweedler, J. V. (2008). Mass spectrometry-based discovery of circadian peptides.

Proteomic analysis of ovomucoid hypersensitivity in mice by two-dimensional difference gel electrophoresis (2D-DIGE). *Food Chem Toxicol,* Vol. 45, No. 12, pp.


Altun, A., Yaprak, M., Aktoz, M., Vardar, A., Betul, U. A., & Ozbay, G. (2002). Impaired

Andrews, N. P., Gralnick, H. R., Merryman, P., Vail, M., & Quyyumi, A. A. (1996).

Antos, C. L., McKinsey, T. A., Frey, N., Kutschke, W., McAnally, J., Shelton, J. M.,

Arias-Loza, P. A., Jazbutyte, V., & Pelzer, T. (2008). Genetic and pharmacologic strategies to

Charloux, A., Gronfier, C., Lonsdorfer-Wolf, E., Piquard, F., & Brandenberger, G. (1999).

Chung, S., Son, G. H., & Kim, K. (2011). Circadian rhythm of adrenal glucocorticoid: its

Claustrat, B., Brun, J., Garry, P., Roussel, B., & Sassolas, G. (1986). A once-repeated study of

Cohen, M. C., Rohtla, K. M., Lavery, C. E., Muller, J. E., & Mittleman, M. A. (1997). Meta-

Cugini, P., Lucia, P., Di Palma, L., Re, M., Canova, R., Gasbarrone, L., & Cianetti, A. (1992).

Cugini, P., Lucia, P., Di Palma, L., Re, M., Leone, G., Battisti, P., Canova, R.*, et al.* (1991).

Cugini, P., Lucia, P., Scibilia, G., Di Palma, L., Cioli, A. R., Cianetti, A., Gasbarrone, L.*, et al.*

Decousus, H. A., Croze, M., Levi, F. A., Jaubert, J. G., Perpoint, B. M., De Bonadona, J. F.,

men. *J Pineal Res,* Vol. 3, No. 4, pp. 301-10, issn 0742-3098

death. *Am J Cardiol,* Vol. 79, No. 11, pp. 1512-6, issn 0002-9149

consequences. *Lancet,* Vol. 373, No. 9657, pp. 82-93, issn 1474-547X Brzezinski, A. (1997). Melatonin in humans. *N Engl J Med,* Vol. 336, No. 3, pp. 186-95

Vol. 327, No. 2, pp. 143-5, issn 0304-3940

No. 1 Pt 1, pp. E43-9, issn 0002-9513

Vol. 2, No. 2, pp. 113-8, issn 0959-9851

*Regul Pept,* Vol. 34, No. 3, pp. 141-8, issn 0167-0115

101-6, issn 0009-7322

issn 0027-8424

91, issn 0006-3002

363-5, issn 0007-0769

0267-0623

nocturnal synthesis of melatonin in patients with cardiac syndrome X. *Neurosci Lett,*

Mechanisms underlying the morning increase in platelet aggregation: a flow cytometry study. *J Am Coll Cardiol,* Vol. 28, No. 7, pp. 1789-95, issn 0735-1097 Angleton, P., Chandler, W. L., & Schmer, G. (1989). Diurnal variation of tissue-type

plasminogen activator and its rapid inhibitor (PAI-1). *Circulation,* Vol. 79, No. 1, pp.

Richardson, J. A.*, et al.* (2002). Activated glycogen synthase-3 beta suppresses cardiac hypertrophy in vivo. *Proc Natl Acad Sci U S A,* Vol. 99, No. 2, pp. 907-12,

determine the function of estrogen receptor alpha and estrogen receptor beta in cardiovascular system. *Gend Med,* Vol. 5 Suppl A, pp. S34-45, issn 1550-8579 Bradley, T. D., & Floras, J. S. (2009). Obstructive sleep apnoea and its cardiovascular

Aldosterone release during the sleep-wake cycle in humans. *Am J Physiol,* Vol. 276,

regulation and clinical implications. *Biochim Biophys Acta,* Vol. 1812, No. 5, pp. 581-

nocturnal plasma melatonin patterns and sleep recordings in six normal young

analysis of the morning excess of acute myocardial infarction and sudden cardiac

The circadian rhythm of atrial natriuretic peptide, vasoactive intestinal peptide, beta-endorphin and cortisol in healthy young and elderly subjects. *Clin Auton Res,*

Vasoactive intestinal peptide fluctuates in human blood with a circadian rhythm.

(1993). Lack of circadian rhythm of plasma concentrations of vasoactive intestinal peptide in patients with orthotopic heart transplants. *Br Heart J,* Vol. 70, No. 4, pp.

Reinberg, A.*, et al.* (1985). Circadian changes in anticoagulant effect of heparin infused at a constant rate. *Br Med J (Clin Res Ed),* Vol. 290, No. 6465, pp. 341-4, issn


Circadian Proteomics and Its Unique Advantage for Discovery of Biomarkers of Heart Disease 87

Martino, T. A., Tata, N., Simpson, J. A., Vanderlaan, R., Dawood, F., Kabir, M. G., Khaper,

hypertrophy. *J Am Coll Cardiol,* Vol. 57, No. 20, pp. 2020-8, issn 1558-3597 Mehta, R. H., Manfredini, R., Hassan, F., Sechtem, U., Bossone, E., Oh, J. K., Cooper, J. V.*, et* 

Minami, Y., Kasukawa, T., Kakazu, Y., Iigo, M., Sugimoto, M., Ikeda, S., Yasui, A.*, et al.*

Moller, M., Sparre, T., Bache, N., Roepstorff, P., & Vorum, H. (2007). Proteomic analysis of

Muller, J. E., Tofler, G. H., & Stone, P. H. (1989). Circadian variation and triggers of onset of acute cardiovascular disease. *Circulation,* Vol. 79, No. 4, pp. 733-43, issn 0009-7322 Muller, J. E., Stone, P. H., Turi, Z. G., Rutherford, J. D., Czeisler, C. A., Parker, C., Poole, W.

Otto, M. E., Svatikova, A., Barretto, R. B., Santos, S., Hoffmann, M., Khandheria, B., &

Paulin, D., & Li, Z. (2004). Desmin: a major intermediate filament protein essential for the structural integrity and function of muscle. *Exp Cell Res,* Vol. 301, No. 1, pp. 1-7, Rajaratnam, S. M., & Arendt, J. (2001). Health in a 24-h society. *Lancet,* Vol. 358, No. 9286, pp.

Ralph, M. R., & Menaker, M. (1988). A mutation of the circadian system in golden hamsters.

Reddy, A. B., Wong, G. K., O'Neill, J., Maywood, E. S., & Hastings, M. H. (2005). Circadian

Reddy, A. B., Karp, N. A., Maywood, E. S., Sage, E. A., Deery, M., O'Neill, J. S., Wong, G. K.*,* 

Reppert, S. M., & Weaver, D. R. (2001). Molecular analysis of mammalian circadian rhythms.

Reppert, S. M., & Weaver, D. R. (2002). Coordination of circadian timing in mammals.

Roenneberg, T., & Merrow, M. (2005). Circadian clocks - the fall and rise of physiology. *Nat* 

Roger, V. L., Go, A. S., Lloyd-Jones, D. M., Adams, R. J., Berry, J. D., Brown, T. M.,

Carnethon, M. R.*, et al.* (2011). Heart disease and stroke statistics--2011 update: a

clocks: neural and peripheral pacemakers that impact upon the cell division cycle.

*et al.* (2006). Circadian orchestration of the hepatic proteome. *Curr Biol,* Vol. 16, No.

infarction. *N Engl J Med,* Vol. 313, No. 21, pp. 1315-22, issn 0028-4793 Orii, K. E., Fukao, T., Song, X. Q., Mitchell, G. A., & Kondo, N. (2008). Liver-specific

*Tohoku J Exp Med,* Vol. 215, No. 3, pp. 227-36, issn 1349-3329

humans. *Circulation,* Vol. 109, No. 21, pp. 2507-10,

*Science,* Vol. 241, No. 4870, pp. 1225-7, issn 0036-8075

*Mutat Res,* Vol. 574, No. 1-2, pp. 76-91, issn 0027-5107

*Annu Rev Physiol,* Vol. 63, pp. 647-76, issn 0066-4278

*Nature,* Vol. 418, No. 6901, pp. 935-41, issn 0028-0836

*Rev Mol Cell Biol,* Vol. 6, No. 12, pp. 965-71, issn 1471-0072

*Physiol,* Vol. 294, No. 5, pp. R1675-83, issn 0363-6119

*Sci U S A,* Vol. 106, No. 24, pp. 9890-5, issn 1091-6490

No. 9, pp. 1110-5, issn 1524-4539

12, pp. 2009-18, issn 1615-9853

999-1005, issn 0140-6736

11, pp. 1107-15, issn 0960-9822

cardiovascular and renal disease in hamsters. *Am J Physiol Regul Integr Comp* 

N.*, et al.* (2011). The primary benefits of angiotensin-converting enzyme inhibition on cardiac remodeling occur during sleep time in murine pressure overload

*al.* (2002). Chronobiological patterns of acute aortic dissection. *Circulation,* Vol. 106,

(2009). Measurement of internal body time by blood metabolomics. *Proc Natl Acad* 

day-night variations in protein levels in the rat pineal gland. *Proteomics,* Vol. 7, No.

K.*, et al.* (1985). Circadian variation in the frequency of onset of acute myocardial

silencing of the human gene encoding succinyl-CoA: 3-ketoacid CoA transferase.

Somers, V. (2004). Early morning attenuation of endothelial function in healthy


Heart and Stroke Foundation of Canada [HSFO]. Heart disease statistics. 7.27 2011,

<http://www.heartandstroke.com/site/c.ikIQLcMWJtE/b.3483991/k.34A8/Statis

Iitaka, C., Miyazaki, K., Akaike, T., & Ishida, N. (2005). A role for glycogen synthase kinase-

Imai, Y., Abe, K., Munakata, M., Sakuma, H., Hashimoto, J., Imai, K., Sekino, H.*, et al.* (1990).

Innominato, P. F., Levi, F. A., & Bjarnason, G. A. (2010). Chronotherapy and the molecular

Johnsen, S. A., Gungor, C., Prenzel, T., Riethdorf, S., Riethdorf, L., Taniguchi-Ishigaki, N.,

Linsell, C. R., Lightman, S. L., Mullen, P. E., Brown, M. J., & Causon, R. C. (1985). Circadian

Lockinger, A., Koberle, D., Konig, P. S., Saria, A., Herold, M., Cornelissen, G., & Halberg, F.

Maemura, K., de la Monte, S. M., Chin, M. T., Layne, M. D., Hsieh, C. M., Yet, S. F., Perrella,

Manfredini, R., Boari, B., Gallerani, M., Salmi, R., Bossone, E., Distante, A., Eagle, K. A.*, et al.*

Martino, T. A., & Sole, M. J. (2009). Molecular time: an often overlooked dimension to cardiovascular disease. *Circ Res,* Vol. 105, No. 11, pp. 1047-61, issn 1524-4571 Martino, T. A., Tata, N., Bjarnason, G. A., Straume, M., & Sole, M. J. (2007). Diurnal protein

*Regul Integr Comp Physiol,* Vol. 293, No. 3, pp. R1430-7, issn 0363-6119 Martino, T. A., Oudit, G. Y., Herzenberg, A. M., Tata, N., Koletar, M. M., Kabir, G. M.,

and circadian patterns. *Peptides,* Vol. 25, No. 4, pp. 533-42, issn 0196-9781 Lowrey, P. L., Shimomura, K., Antoch, M. P., Yamazaki, S., Zemenides, P. D., Ralph, M. R.,

*J Hypertens Suppl,* Vol. 8, No. 7, pp. S125-32, issn 0952-1178

3beta in the mammalian circadian clock. *J Biol Chem,* Vol. 280, No. 33, pp. 29397-

Circadian blood pressure variations under different pathophysiological conditions.

clock: Clinical implications in oncology. *Adv Drug Deliv Rev,* Vol. 62, No. 9-10, pp.

Rau, T.*, et al.* (2009). Regulation of estrogen-dependent transcription by the LIM cofactors CLIM and RLIM in breast cancer. *Cancer Res,* Vol. 69, No. 1, pp. 128-36 Katz, F. H., Romfh, P., & Smith, J. A. (1975). Diurnal variation of plasma aldosterone, cortisol and renin activity in supine man. *J Clin Endocrinol Metab,* Vol. 40, No. 1, pp. 125-34 Lightman, S. L., James, V. H., Linsell, C., Mullen, P. E., Peart, W. S., & Sever, P. S. (1981).

Studies of diurnal changes in plasma renin activity, and plasma noradrenaline, aldosterone and cortisol concentrations in man. *Clin Endocrinol (Oxf),* Vol. 14, No. 3,

rhythms of epinephrine and norepinephrine in man. *J Clin Endocrinol Metab,* Vol.

(2004). Neuropeptide chronomics in clinically healthy young adults: circaoctohoran

Menaker, M.*, et al.* (2000). Positional syntenic cloning and functional characterization of the mammalian circadian mutation tau. *Science,* Vol. 288, No.

M. A.*, et al.* (2000). CLIF, a novel cycle-like factor, regulates the circadian oscillation of plasminogen activator inhibitor-1 gene expression. *J Biol Chem,* Vol. 275, No. 47,

(2004). Chronobiology of rupture and dissection of aortic aneurysms. *J Vasc Surg,*

expression in blood revealed by high throughput mass spectrometry proteomics and implications for translational medicine and body time of day. *Am J Physiol* 

Belsham, D. D.*, et al.* (2008). Circadian rhythm disorganization produces profound

Available from:

tics.htm#heartdisease>

979-1001, issn 1872-8294

pp. 213-23, issn 0300-0664

60, No. 6, pp. 1210-5, issn 0021-972X

5465, pp. 483-92, issn 0036-8075

pp. 36847-51, issn 0021-9258

Vol. 40, No. 2, pp. 382-8, issn 0741-5214

402, issn 0021-9258

cardiovascular and renal disease in hamsters. *Am J Physiol Regul Integr Comp Physiol,* Vol. 294, No. 5, pp. R1675-83, issn 0363-6119


**5** 

Linda Papa

*USA* 

**Exploring the Role of Biomarkers** 

**for the Diagnosis and Management** 

There are an estimated 10 million people affected annually by traumatic brain injury (TBI) across the globe.1 In the United States, TBI is a major cause of death and disability2 with about 52,000 annual deaths and 5.3 million Americans impaired by its effects. TBI is a contributing factor to over 30% of all injury-related deaths in the United States and it has been referred to as the silent epidemic of our time. 3, 4 European TBI prevalence data is not consistently reported by each country but it has been estimated that 1.6 million head-injured patients are hospitalized annually in Europe with an incidence rate of about 235 per 100,000. There is an average mortality rate of about 15 per 100,000 and a case fatality rate of about 11 per 100. The TBI severity ratio of hospitalized patients is about 22:1.5:1 for mild vs. moderate vs. severe cases, respectively.5 According to the World Health Organization, TBI will

surpass many diseases as the major cause of death and disability by the year 2020.1

events. These events are mediated by many molecular and cellular processes.

Brain injuries can be focal, diffuse or a combination of focal and diffuse. The degree of brain injury depends on the primary mechanism/magnitude of injury, secondary insults and the patient's genetic and molecular response. Following the initial injury, cellular responses and neurochemical and metabolic cascades contribute to secondary injury.6, 7 Focal brain injuries include contusions, brain lacerations, and hemorrhage leading to the formation of hematoma in the extradural, subarachnoid, subdural, or intracerebral compartments within the head. Traumatic brain injury represents a spectrum of injury severity. The number, types, and location of lesions as well as the magnitude of overlapping injuries across this spectrum of injury severity are still not clearly described and are challenging to classify. There are two aspects to injury caused by TBI - the damage caused by the initial impact or insult, and that which may subsequently evolve over the ensuing hours and days referred to as secondary insults. Secondary insults can be mediated through physiologic events which decrease supply of oxygen and energy to the brain tissue or through a cascade of cytotoxic

**1. Introduction** 

**of Traumatic Brain Injury Patients** 

*Graduate Medical Education, Orlando Health* 

*Department of Emergency Medicine University of Florida, College of Medicine Florida State University, College of Medicine University of Central Florida, College of Medicine Orlando Regional Medical Center, Orlando, Florida* 

report from the American Heart Association. *Circulation,* Vol. 123, No. 4, pp. e18 e209, issn 1524-4539


### **Exploring the Role of Biomarkers for the Diagnosis and Management of Traumatic Brain Injury Patients**

Linda Papa

*Graduate Medical Education, Orlando Health Department of Emergency Medicine University of Florida, College of Medicine Florida State University, College of Medicine University of Central Florida, College of Medicine Orlando Regional Medical Center, Orlando, Florida USA* 

#### **1. Introduction**

88 Proteomics – Human Diseases and Protein Functions

Scheer, F. A., Hilton, M. F., Mantzoros, C. S., & Shea, S. A. (2009). Adverse metabolic and

Sole, M. J., & Martino, T. A. (2009). Diurnal physiology: core principles with application to

Sumiyoshi, M., Kojima, S., Arima, M., Suwa, S., Nakazato, Y., Sakurai, H., Kanoh, T.*, et al.*

Surya, S., Symons, K., Rothman, E., & Barkan, A. L. (2006). Complex rhythmicity of growth hormone secretion in humans. *Pituitary,* Vol. 9, No. 2, pp. 121-5, issn 1386-341X Takahashi, Y., Kipnis, D. M., & Daughaday, W. H. (1968). Growth hormone secretion during

Tengattini, S., Reiter, R. J., Tan, D. X., Terron, M. P., Rodella, L. F., & Rezzani, R. (2008).

Tofler, G. H., Gebara, O. C., Mittleman, M. A., Taylor, P., Siegel, W., Venditti, F. J., Jr.,

Tsuji, T., Hirota, T., Takemori, N., Komori, N., Yoshitane, H., Fukuda, M., Matsumoto, H.*, et* 

Tuck, M. L., Stern, N., & Sowers, J. R. (1985). Enhanced 24-hour norepinephrine and renin

Volterrani, M., Manelli, F., Cicoira, M., Lorusso, R., & Giustina, A. (2000). Role of growth

World Health Organization [WHO]. 2011. Cardiovascular Diseases. 7.27 2011, Available

Willich, S. N., Levy, D., Rocco, M. B., Tofler, G. H., Stone, P. H., & Muller, J. E. (1987).

from: <http://www.who.int/cardiovascular\_diseases/en/>

Investigators. *Circulation,* Vol. 92, No. 5, pp. 1203-8, issn 0009-7322

and failure. *J Appl Physiol,* Vol. 107, No. 4, pp. 1318-27, issn 1522-1601 Storch, K. F., & Weitz, C. J. (2009). Daily rhythms of food-anticipatory behavioral activity do

dissection. *Am J Cardiol,* Vol. 89, No. 5, pp. 619-23, issn 0002-9149

sleep. *J Clin Invest,* Vol. 47, No. 9, pp. 2079-90, issn 0021-9738

e209, issn 1524-4539

pp. 6808-13, issn 1091-6490

pp. 16-25, issn 0742-3098

3500-8, issn 1615-9853

pp. 711-9, issn 0012-6667

5763, pp. 1002-5, issn 1095-9203

Vol. 106, No. 11, pp. 4453-8, issn 1091-6490

report from the American Heart Association. *Circulation,* Vol. 123, No. 4, pp. e18-

cardiovascular consequences of circadian misalignment. *Proc Natl Acad Sci U S A,*

the pathogenesis, diagnosis, prevention, and treatment of myocardial hypertrophy

not require the known circadian clock. *Proc Natl Acad Sci U S A,* Vol. 106, No. 16,

(2002). Circadian, weekly, and seasonal variation at the onset of acute aortic

Cardiovascular diseases: protective effects of melatonin. *J Pineal Res,* Vol. 44, No. 1,

Rasmussen, C. A.*, et al.* (1995). Morning peak in ventricular tachyarrhythmias detected by time of implantable cardioverter/defibrillator therapy. The CPI

*al.* (2007). Circadian proteomics of the mouse retina. *Proteomics,* Vol. 7, No. 19, pp.

secretion in young patients with essential hypertension: relation with the circadian pattern of arterial blood pressure. *Am J Cardiol,* Vol. 55, No. 1, pp. 112-5, issn 2-9149

hormone in chronic heart failure. Therapeutic implications. *Drugs,* Vol. 60, No. 4,

Circadian variation in the incidence of sudden cardiac death in the Framingham Heart Study population. *Am J Cardiol,* Vol. 60, No. 10, pp. 801-6, issn 0002-9149 Yin, L., Wang, J., Klein, P. S., & Lazar, M. A. (2006). Nuclear receptor Rev-erbalpha is a

critical lithium-sensitive component of the circadian clock. *Science,* Vol. 311, No.

There are an estimated 10 million people affected annually by traumatic brain injury (TBI) across the globe.1 In the United States, TBI is a major cause of death and disability2 with about 52,000 annual deaths and 5.3 million Americans impaired by its effects. TBI is a contributing factor to over 30% of all injury-related deaths in the United States and it has been referred to as the silent epidemic of our time. 3, 4 European TBI prevalence data is not consistently reported by each country but it has been estimated that 1.6 million head-injured patients are hospitalized annually in Europe with an incidence rate of about 235 per 100,000. There is an average mortality rate of about 15 per 100,000 and a case fatality rate of about 11 per 100. The TBI severity ratio of hospitalized patients is about 22:1.5:1 for mild vs. moderate vs. severe cases, respectively.5 According to the World Health Organization, TBI will surpass many diseases as the major cause of death and disability by the year 2020.1

Brain injuries can be focal, diffuse or a combination of focal and diffuse. The degree of brain injury depends on the primary mechanism/magnitude of injury, secondary insults and the patient's genetic and molecular response. Following the initial injury, cellular responses and neurochemical and metabolic cascades contribute to secondary injury.6, 7 Focal brain injuries include contusions, brain lacerations, and hemorrhage leading to the formation of hematoma in the extradural, subarachnoid, subdural, or intracerebral compartments within the head. Traumatic brain injury represents a spectrum of injury severity. The number, types, and location of lesions as well as the magnitude of overlapping injuries across this spectrum of injury severity are still not clearly described and are challenging to classify.

There are two aspects to injury caused by TBI - the damage caused by the initial impact or insult, and that which may subsequently evolve over the ensuing hours and days referred to as secondary insults. Secondary insults can be mediated through physiologic events which decrease supply of oxygen and energy to the brain tissue or through a cascade of cytotoxic events. These events are mediated by many molecular and cellular processes.

Exploring the Role of

timeframe.

risks and costs of human clinical trials.

Biomarkers for the Diagnosis and Management of Traumatic Brain Injury Patients 91

reference standard for TBI is also more difficult to define than say cardiac ischemia. There is no early gold standard for stratification of patients by severity. Currently, diagnosis of TBI depends on a variety of measures including neurological examination and neuroimaging. Neuroimaging techniques such as CT scanning and MRI are used to provide objective information. However, CT scanning has low sensitivity to diffuse brain damage and confers exposure to radiation. MRI can provide information on the extent of diffuse injuries but its widespread application is restricted by cost, the limited availability of MRI in many centers, and the difficulty of performing it in physiologically unstable patients. Additionally, its role

While increasing CT use has reduced hospital admissions,27 it has also raised concern over unnecessary exposure to ionizing radiation.28-32 Although the calculation of projected cancer risk is still controversial, some studies suggest that CT scans of the head may be among the largest contributors to radiation exposure due to the frequency with which they are performed.33 There is significant consensus that efforts should be made to prevent

There have been a number of cerebrospinal fluid (CSF) and serum biomarkers evaluated in TBI animal models and in humans. However, many of these candidate biomarkers have failed to exhibit adequate sensitivity and specificity for brain injury, and they have added minimal diagnostic and prognostic information. As a result many are skeptical about the potential of neurotrauma biomarkers to influence future clinical management and clinical trials. This reservation is based on a handful of biomarkers studied using compromised research designs and without the advantage of advancements made in the field of proteomics. Even though the application of proteomics in brain injury is still in its infancy36, 37, neuroproteomics is penetrating the field of neurotrauma and brings great potential for improvements in research and patient care. As this technology advances and integrates other technologies such as bioinformatics and neuroimaging, characterization of CNS proteins will occur quickly and many more potential markers will be validated in a shorter

Another important challenge in validating biomarkers for TBI will be that traditional outcome measures used to measure injury severity are, in and of themselves, limited. This is true for all severities of injury, and is particularly germane to the less severe injuries where neuroimaging, such as computed tomography (CT), may not demonstrate any obvious pathology. Traditionally, TBI has been separated into three very broad categories: mild, moderate and severe. Unfortunately, this classification scheme fails to capture the spectrum of TBI and the different types of injuries associated with it. The difficulty in classifying injury severity is one which has made clinical trials in the field of TBI challenging. Therapeutic clinical trials for TBI have met with negative results at a cost of over \$200 million.38, 39 These failures have been attributed to a multitude of factors but particularly to the heterogeneity of TBI which makes classification of the different injury types problematic. This heterogeneity, together with the lack of early definitive measures of severity opens the door for using biomarkers as early prognostic indicators. Potentially, biomarkers could provide early outcome measure for clinical trial obtainable much more reliably and economically than conventional neurological assessments, thereby significantly reducing the

in the clinical management of TBI patients acutely has not been established.25, 26

unnecessary radiation exposure while maintaining quality of care.28, 29, 34, 35

**4. Challenges to the clinical application of biomarkers** 

#### **2. The importance of mild and moderate TBI**

Research in the field of TBI has long been dominated by research on severe brain injury. However, of the estimated 1.8 million people in the United States who sustain a TBI each year, over 90% will have either a "moderate" (GCS 9-12) or "mild" (GCS 13-15) injury; far outnumbering severe injuries.2, 8, 9 Moderate TBI comprises over 10% of all TBI and mild TBI over 80%.8 The majority of these patients will present to emergency departments (ED's) around the country for assessment and treatment.10 The direct medical costs for treatment of TBI in the United States have been estimated at more than \$4 billion annually.11 If the costs of lost productivity that result from TBI are added to this then the overall estimated cost is closer to \$56.3 billion. Moreover, mild TBI is significantly underdiagnosed and the likely societal burden is therefore even greater.12 Mild and moderate TBI are often difficult to assess and distinguish clinically during the first hours after injury because neurological examinations are of restricted value. The distinction between mild, moderate and severe TBI is initially based on a GCS score and this may be influenced by factors such as perfusion and intoxication from drugs or alcohol, sedative medications, and other distracting injuries.

The term "mild TBI" is actually a misnomer. Individuals who incur a TBI and have an initial GCS score of 13-15 are acutely at risk for intracranial bleeding and diffuse axonal injury.13 Additionally, a significant proportion is at risk for impairment of physical, cognitive, and psychosocial functioning.14-18 Although some patients with mild TBI may be admitted to the hospital overnight, the vast majority are treated and released from emergency departments with basic discharge instructions. Most receive little guidance with respect to follow-up care. This group of TBI patients represents the greatest challenges to accurate diagnosis and outcome prediction. With perhaps no overt signs of acute head injury and a lack of clinical tools to detect the subtle cognitive deficits the patient is considered "unimpaired" and is discharged home and typically left to deal with persisting neurocognitive deficits on their own.19 Accordingly, a significant minority has incomplete recoveries and has outcomes disproportionately worse than would have been predicted by the objective facts of the injury.19, 20 The lack of clinical tools to detect the deficits that affect daily function leads to a state of frustration for patients and families that arises out of a failure to understand the nature of the difficulties encountered daily. Treatment protocols for mild TBI have only slowly begun to emerge and are still experimental. The injury is often seen as "not severe" and subsequently therapies have not been aggressively sought for these individuals. Unfortunately, despite the better understanding of the anatomical, cellular and molecular mechanisms of TBI, these advances have not yet yielded significant improvements in treatment. Among the potential barriers to treatment are the heterogeneity of traumatic brain injury, difficulty with stratification of patients by injury severity and lack of early markers of injury.21-24

#### **3. The problem with current assessment of TBI**

Prognostic tools for risk stratification of TBI patients are limited in the early stages of injury in the emergency setting for all severities of TBI. Unlike other organ-based diseases where rapid diagnosis employing biomarkers from blood tests are clinically essential to guide diagnosis and treatment, such as for myocardial ischemia or kidney and liver dysfunction, there are no rapid, definitive diagnostic tests for traumatic brain injury. Moreover, the

Research in the field of TBI has long been dominated by research on severe brain injury. However, of the estimated 1.8 million people in the United States who sustain a TBI each year, over 90% will have either a "moderate" (GCS 9-12) or "mild" (GCS 13-15) injury; far outnumbering severe injuries.2, 8, 9 Moderate TBI comprises over 10% of all TBI and mild TBI over 80%.8 The majority of these patients will present to emergency departments (ED's) around the country for assessment and treatment.10 The direct medical costs for treatment of TBI in the United States have been estimated at more than \$4 billion annually.11 If the costs of lost productivity that result from TBI are added to this then the overall estimated cost is closer to \$56.3 billion. Moreover, mild TBI is significantly underdiagnosed and the likely societal burden is therefore even greater.12 Mild and moderate TBI are often difficult to assess and distinguish clinically during the first hours after injury because neurological examinations are of restricted value. The distinction between mild, moderate and severe TBI is initially based on a GCS score and this may be influenced by factors such as perfusion and intoxication from drugs or alcohol, sedative medications,

The term "mild TBI" is actually a misnomer. Individuals who incur a TBI and have an initial GCS score of 13-15 are acutely at risk for intracranial bleeding and diffuse axonal injury.13 Additionally, a significant proportion is at risk for impairment of physical, cognitive, and psychosocial functioning.14-18 Although some patients with mild TBI may be admitted to the hospital overnight, the vast majority are treated and released from emergency departments with basic discharge instructions. Most receive little guidance with respect to follow-up care. This group of TBI patients represents the greatest challenges to accurate diagnosis and outcome prediction. With perhaps no overt signs of acute head injury and a lack of clinical tools to detect the subtle cognitive deficits the patient is considered "unimpaired" and is discharged home and typically left to deal with persisting neurocognitive deficits on their own.19 Accordingly, a significant minority has incomplete recoveries and has outcomes disproportionately worse than would have been predicted by the objective facts of the injury.19, 20 The lack of clinical tools to detect the deficits that affect daily function leads to a state of frustration for patients and families that arises out of a failure to understand the nature of the difficulties encountered daily. Treatment protocols for mild TBI have only slowly begun to emerge and are still experimental. The injury is often seen as "not severe" and subsequently therapies have not been aggressively sought for these individuals. Unfortunately, despite the better understanding of the anatomical, cellular and molecular mechanisms of TBI, these advances have not yet yielded significant improvements in treatment. Among the potential barriers to treatment are the heterogeneity of traumatic brain injury, difficulty with stratification of patients by injury severity and lack of early

Prognostic tools for risk stratification of TBI patients are limited in the early stages of injury in the emergency setting for all severities of TBI. Unlike other organ-based diseases where rapid diagnosis employing biomarkers from blood tests are clinically essential to guide diagnosis and treatment, such as for myocardial ischemia or kidney and liver dysfunction, there are no rapid, definitive diagnostic tests for traumatic brain injury. Moreover, the

**2. The importance of mild and moderate TBI** 

and other distracting injuries.

markers of injury.21-24

**3. The problem with current assessment of TBI** 

reference standard for TBI is also more difficult to define than say cardiac ischemia. There is no early gold standard for stratification of patients by severity. Currently, diagnosis of TBI depends on a variety of measures including neurological examination and neuroimaging. Neuroimaging techniques such as CT scanning and MRI are used to provide objective information. However, CT scanning has low sensitivity to diffuse brain damage and confers exposure to radiation. MRI can provide information on the extent of diffuse injuries but its widespread application is restricted by cost, the limited availability of MRI in many centers, and the difficulty of performing it in physiologically unstable patients. Additionally, its role in the clinical management of TBI patients acutely has not been established.25, 26

While increasing CT use has reduced hospital admissions,27 it has also raised concern over unnecessary exposure to ionizing radiation.28-32 Although the calculation of projected cancer risk is still controversial, some studies suggest that CT scans of the head may be among the largest contributors to radiation exposure due to the frequency with which they are performed.33 There is significant consensus that efforts should be made to prevent unnecessary radiation exposure while maintaining quality of care.28, 29, 34, 35

#### **4. Challenges to the clinical application of biomarkers**

There have been a number of cerebrospinal fluid (CSF) and serum biomarkers evaluated in TBI animal models and in humans. However, many of these candidate biomarkers have failed to exhibit adequate sensitivity and specificity for brain injury, and they have added minimal diagnostic and prognostic information. As a result many are skeptical about the potential of neurotrauma biomarkers to influence future clinical management and clinical trials. This reservation is based on a handful of biomarkers studied using compromised research designs and without the advantage of advancements made in the field of proteomics. Even though the application of proteomics in brain injury is still in its infancy36, 37, neuroproteomics is penetrating the field of neurotrauma and brings great potential for improvements in research and patient care. As this technology advances and integrates other technologies such as bioinformatics and neuroimaging, characterization of CNS proteins will occur quickly and many more potential markers will be validated in a shorter timeframe.

Another important challenge in validating biomarkers for TBI will be that traditional outcome measures used to measure injury severity are, in and of themselves, limited. This is true for all severities of injury, and is particularly germane to the less severe injuries where neuroimaging, such as computed tomography (CT), may not demonstrate any obvious pathology. Traditionally, TBI has been separated into three very broad categories: mild, moderate and severe. Unfortunately, this classification scheme fails to capture the spectrum of TBI and the different types of injuries associated with it. The difficulty in classifying injury severity is one which has made clinical trials in the field of TBI challenging. Therapeutic clinical trials for TBI have met with negative results at a cost of over \$200 million.38, 39 These failures have been attributed to a multitude of factors but particularly to the heterogeneity of TBI which makes classification of the different injury types problematic. This heterogeneity, together with the lack of early definitive measures of severity opens the door for using biomarkers as early prognostic indicators. Potentially, biomarkers could provide early outcome measure for clinical trial obtainable much more reliably and economically than conventional neurological assessments, thereby significantly reducing the risks and costs of human clinical trials.

Exploring the Role of

**6. Status of biomarker research** 

from peripheral tissues.88-90

setting of hemolysis.96

Biomarkers for the Diagnosis and Management of Traumatic Brain Injury Patients 93

Brophy et al.43 an immunoblotting technique employing sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) was used to measure alpha-spectrin. Quantitative evaluation of intact αII-spectrin and its breakdown products (SBDP150, SBDP145 and SBDP120) was performed via computer-assisted densitometric scanning. An example of the ELISA technique is taken from a study published in Critical Care Medicine in 2010 by Papa et al.44 that measured Ubiquitin C-terminal hydrolase. In this study samples were measured using a standard UCH-L1 sandwich ELISA where reaction wells were coated with capture antibody and detection antibody was added to wells. The

Although there are a number of biochemical markers that have been investigated in TBI, our discussion will include the most current and widely studied ones. The most extensively studied among these are glial protein S-100 beta(β) 45-55, neuron-specific enolase (NSE)56-63, and myelin basic protein (MBP)41, 59, 64-66 Although some of these published studies suggest

S100β is the major low affinity calcium binding protein in astrocytes 76 and it is considered a marker of astrocyte injury or death. It can also be found in non-neural cells such as adipocytes, chondrocytes, and melanoma cells.77 Elevated serum levels have been associated with increased incidence of post concussive syndrome and impaired cognition.78, 79 Other studies have reported that serum levels of S-100β are associated with MRI abnormalities and with neuropsychological examination disturbances after mild TBI.80, 81 A number of studies have found significant correlations between elevated serum levels of S-100β and CT abnormalities.82-84 It has been suggested that adding the measurement of S-100B concentration to clinical decision tools for mild TBI patients could potentially reduce the number of CT scans by 30%.84 Other investigators have failed to detect associations between S-100β with CT abnormalities.67, 85, 86 87 The vast majority of these clinical studies have employed ELISA to measure levels of S100B. Although S-100β continues to be actively investigated and remains promising as an adjunctive marker, its utility as a biochemical diagnostic remains controversial. Some studies have observed high serum S-100β levels in trauma patients without head injuries suggesting that it lacks CNS specificity and is released

Neuron specific enolase is one of the five isozymes of the gycolytic enzyme enolase found in central and peripheral neurons and it has been shown be elevated following cell injury.91 It has a molecular weight of 78 kDa and a biological half-life of 48 hours.92 This protein is passively released into the extracellular space only under pathological conditions during cell destruction. Several reports on serum NSE measurements of mild TBI have been published.59, 62, 91, 93 Most of these studies employed an enzyme immunoassay for NSE detection. Many of these studies either contained inadequate control groups or concluded that serum NSE had limited utility as a marker of neuronal damage. Early levels of NSE and MBP concentrations have been correlated with outcome in children, particularly those under 4 years of age.64, 65, 94, 95 A limitation of NSE is the occurrence of false positive results in the

A supposedly cleaved form of tau, c-tau, has also been investigated as a potential biomarker of CNS injury. Tau is preferentially localized in the axon and tau lesions are apparently related to axonal disruption.97, 98 CSF levels of c-tau were significantly elevated in TBI

wells were developed with substrate solution and read with a spectrophotometer.

that these biomarkers correlate with degree of injury; conflicting results exist.67-75

The release of substances and potential biomarkers after an injury is not a static process. Understanding the biokinetic properties of a biomarker will be essential to understanding the release pattern and "optimum" time for measurement. Clinicians and researchers will have to keep in mind that different injury types (for instance, mass lesions versus diffuse injuries) may demonstrate different kinetic parameters and, thus, may produce different quantities of a marker with different peaks and rates of decay. Moreover, secondary insults may also contribute to secondary elevations in a marker, altering its sensitivity and specificity at different time-points.

For markers measured in serum, the level of a biomarker may also reflect the extent of blood brain barrier disruption. Furthermore, extracranial sources of the biomarker may limit its specificity by creating false positives, thus compromising its clinical utility. For instance, the release of a potential CNS marker into the serum from other traumatized tissues or organs would hamper its clinical value in the setting of polytrauma. Another possible situation in which false positive marker values could occur is in the presence of a pre-existing disease state that may alter the metabolism or clearance of the marker, as with kidney or liver disease. Such factors need to be carefully assessed in rigorously designed clinical studies. Future studies should also ensure that adequate control groups are selected for comparison.

Ongoing studies by our group are currently being conducted to more fully elucidate the relationships between novel biomarkers and severity of injury and clinical outcomes in all severities of TBI patients. Before clinical application neurochemical markers will have to be rigorously evaluated and the above mentioned challenges taken into consideration.

#### **5. Proteomic techniques in neurobiomarker discovery**

Two dimensional gel electrophoresis (2D GE) and mass spectrometry has classically been the gold standard for protein identification. It is an excellent technique for discovering a multitude of proteins and is widely used. However, it requires specialized training and technical expertise. Some of the disadvantages include sample to sample variation, the inability to detect certain classes or sizes of proteins, and the need for many samples and controls.40

There are also non–gel-based mass spectrometry methods for identifying proteins that use high-resolution chromatography to separate complex mixtures of proteins prior to mass spectrometry. Typically the technique uses capillary chromatography for sensitivity and high-resolution mass spectrometry for identification of proteins. There is no need for twodimensional gel electrophoresis for initial separation and it can analyze a wider range of proteins. However, the technique requires significant expertise and the cost of the materials and equipment to run this technique is much higher.40

Newer proteomic techniques are employing antibody-based methods such as high throughput immunoblotting and antibody panels and/or arrays (ELISA's). Antibodies are significantly more specific and selective than traditional techniques and allow the detection of proteins amid complex high-protein content biofluids such as serum or plasma.41 Methods of amplifying the signal are under development so that only very small samples will be required for analysis. The drawback of this technique is its reliance on the sensitivity and specificity of the antibodies, and the inability to identify a wide range of proteins because the protein of interest must be pre-selected.

Examples of these techniques will be taken from studies conducted by our group. In two studies published in the Journal of Neurotrauma in 2007 by Pineda et al.42 and in 2009 by Brophy et al.43 an immunoblotting technique employing sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) was used to measure alpha-spectrin. Quantitative evaluation of intact αII-spectrin and its breakdown products (SBDP150, SBDP145 and SBDP120) was performed via computer-assisted densitometric scanning.

An example of the ELISA technique is taken from a study published in Critical Care Medicine in 2010 by Papa et al.44 that measured Ubiquitin C-terminal hydrolase. In this study samples were measured using a standard UCH-L1 sandwich ELISA where reaction wells were coated with capture antibody and detection antibody was added to wells. The wells were developed with substrate solution and read with a spectrophotometer.

#### **6. Status of biomarker research**

92 Proteomics – Human Diseases and Protein Functions

The release of substances and potential biomarkers after an injury is not a static process. Understanding the biokinetic properties of a biomarker will be essential to understanding the release pattern and "optimum" time for measurement. Clinicians and researchers will have to keep in mind that different injury types (for instance, mass lesions versus diffuse injuries) may demonstrate different kinetic parameters and, thus, may produce different quantities of a marker with different peaks and rates of decay. Moreover, secondary insults may also contribute to secondary elevations in a marker, altering its sensitivity and

For markers measured in serum, the level of a biomarker may also reflect the extent of blood brain barrier disruption. Furthermore, extracranial sources of the biomarker may limit its specificity by creating false positives, thus compromising its clinical utility. For instance, the release of a potential CNS marker into the serum from other traumatized tissues or organs would hamper its clinical value in the setting of polytrauma. Another possible situation in which false positive marker values could occur is in the presence of a pre-existing disease state that may alter the metabolism or clearance of the marker, as with kidney or liver disease. Such factors need to be carefully assessed in rigorously designed clinical studies. Future studies should also ensure that adequate control groups are selected for comparison. Ongoing studies by our group are currently being conducted to more fully elucidate the relationships between novel biomarkers and severity of injury and clinical outcomes in all severities of TBI patients. Before clinical application neurochemical markers will have to be

rigorously evaluated and the above mentioned challenges taken into consideration.

Two dimensional gel electrophoresis (2D GE) and mass spectrometry has classically been the gold standard for protein identification. It is an excellent technique for discovering a multitude of proteins and is widely used. However, it requires specialized training and technical expertise. Some of the disadvantages include sample to sample variation, the inability to detect certain classes or sizes of proteins, and the need for many samples and

There are also non–gel-based mass spectrometry methods for identifying proteins that use high-resolution chromatography to separate complex mixtures of proteins prior to mass spectrometry. Typically the technique uses capillary chromatography for sensitivity and high-resolution mass spectrometry for identification of proteins. There is no need for twodimensional gel electrophoresis for initial separation and it can analyze a wider range of proteins. However, the technique requires significant expertise and the cost of the materials

Newer proteomic techniques are employing antibody-based methods such as high throughput immunoblotting and antibody panels and/or arrays (ELISA's). Antibodies are significantly more specific and selective than traditional techniques and allow the detection of proteins amid complex high-protein content biofluids such as serum or plasma.41 Methods of amplifying the signal are under development so that only very small samples will be required for analysis. The drawback of this technique is its reliance on the sensitivity and specificity of the antibodies, and the inability to identify a wide range of proteins

Examples of these techniques will be taken from studies conducted by our group. In two studies published in the Journal of Neurotrauma in 2007 by Pineda et al.42 and in 2009 by

**5. Proteomic techniques in neurobiomarker discovery** 

and equipment to run this technique is much higher.40

because the protein of interest must be pre-selected.

specificity at different time-points.

controls.40

Although there are a number of biochemical markers that have been investigated in TBI, our discussion will include the most current and widely studied ones. The most extensively studied among these are glial protein S-100 beta(β) 45-55, neuron-specific enolase (NSE)56-63, and myelin basic protein (MBP)41, 59, 64-66 Although some of these published studies suggest that these biomarkers correlate with degree of injury; conflicting results exist.67-75

S100β is the major low affinity calcium binding protein in astrocytes 76 and it is considered a marker of astrocyte injury or death. It can also be found in non-neural cells such as adipocytes, chondrocytes, and melanoma cells.77 Elevated serum levels have been associated with increased incidence of post concussive syndrome and impaired cognition.78, 79 Other studies have reported that serum levels of S-100β are associated with MRI abnormalities and with neuropsychological examination disturbances after mild TBI.80, 81 A number of studies have found significant correlations between elevated serum levels of S-100β and CT abnormalities.82-84 It has been suggested that adding the measurement of S-100B concentration to clinical decision tools for mild TBI patients could potentially reduce the number of CT scans by 30%.84 Other investigators have failed to detect associations between S-100β with CT abnormalities.67, 85, 86 87 The vast majority of these clinical studies have employed ELISA to measure levels of S100B. Although S-100β continues to be actively investigated and remains promising as an adjunctive marker, its utility as a biochemical diagnostic remains controversial. Some studies have observed high serum S-100β levels in trauma patients without head injuries suggesting that it lacks CNS specificity and is released from peripheral tissues.88-90

Neuron specific enolase is one of the five isozymes of the gycolytic enzyme enolase found in central and peripheral neurons and it has been shown be elevated following cell injury.91 It has a molecular weight of 78 kDa and a biological half-life of 48 hours.92 This protein is passively released into the extracellular space only under pathological conditions during cell destruction. Several reports on serum NSE measurements of mild TBI have been published.59, 62, 91, 93 Most of these studies employed an enzyme immunoassay for NSE detection. Many of these studies either contained inadequate control groups or concluded that serum NSE had limited utility as a marker of neuronal damage. Early levels of NSE and MBP concentrations have been correlated with outcome in children, particularly those under 4 years of age.64, 65, 94, 95 A limitation of NSE is the occurrence of false positive results in the setting of hemolysis.96

A supposedly cleaved form of tau, c-tau, has also been investigated as a potential biomarker of CNS injury. Tau is preferentially localized in the axon and tau lesions are apparently related to axonal disruption.97, 98 CSF levels of c-tau were significantly elevated in TBI

Exploring the Role of

from studies in severe TBI to mild and moderate TBI.

Fig. 1. The neuroanatomical locations of the above mentioned biomarkers.

outcome; 8) be easily measured by widely available, simple techniques

Research in the field TBI biomarkers has increased exponentially over the last 20 years with most of the publications on the topic occurring in the last 10 years.134 During the course of our work in the development of TBI biomarkers, it has become evident that there are a number of key features that a clinically useful biomarker should possess.135 An "ideal biomarker" would: 1) demonstrate a high sensitivity and specificity for brain injury; 2) stratify patients by severity of injury; 3) have a rapid appearance in accessible biological fluid; 4) provide information on injury mechanisms; 5) have well defined biokinetc properties; 6) monitor progress of disease and response to treatment; 7) predict functional

**7. Attributes of an ideal biomarker for TBI** 

Biomarkers for the Diagnosis and Management of Traumatic Brain Injury Patients 95

neurodegenerative disorders124, 125 and stroke126 to severe traumatic brain injury.127-131 Recently, Vos et al. described serum GFAP profile in severe and moderate TBI (GCS <12).54 In a recent study by our group, GFAP was systematically assessed in human serum following mild and moderate TBI. We confirmed that the GFAP levels were significantly elevated in this population using ELISA analysis, including those with mild TBI. GFAP was able to discriminate TBI patients from uninjured controls. Additionally, serum levels were able to distinguish orthopedic and motor vehicle controls form TBI patients. GFAP was detectable in serum within a few hours of injury and was associated with measures of injury severity including the GCS score and CT lesions.132, 133 The present work extends findings

patients compared to control patients and these levels correlated with clinical outcome.99, 100 Though levels of c-tau were also elevated in plasma from patients with severe TBI, there was no correlation between plasma levels and clinical outcome.101 A major limitation of all of these biomarkers is the lack of specificity for defining neuropathological cascades.

Alpha-II-spectrin (280 kDa) is the major structural component of the cortical membrane cytoskeleton and is particularly abundant in axons and presynaptic terminals.102, 103 It is also a major substrate for both calpain and caspase-3 cysteine proteases.104, 105 A hallmark feature of apoptosis and necrosis is an early cleavage of several cellular proteins by activated caspases and calpains. A signature of caspase-3 and calpain-2 activation is cleavage of several common proteins such as cytoskeletal αII-spectrin.106 In a rat model, mean levels of both ipsilateral cortex (IC) and cerebral spinal fluid (CSF) spectrin breakdown products (SBDP) at 2, 6, and 24 h after two levels of controlled cortical impact (1.0 mm and 1.6 mm of cortical deformation) were significantly elevated by injury using immunoblotting.107 Using the same proteomic Western blot technique, levels of spectrin breakdown products (SBDP's) have been reported in CSF from adults with severe TBI and they have shown a significant relationship with severity of injury and clinical outcome.42, 108-113 Following a TBI the axonally enriched cytoskeletal protein α-II-spectrin is proteolyzed by calpain and caspase-3 to signature breakdown products (SBDPs). Calpain and caspase-3 mediated SBDP levels in CSF have shown to be significantly increased in TBI patients at several time points after injury, compared to control subjects. The time course of calpain mediated SBDP150 and SBDP145 (markers of necrosis) differs from that of caspase-3 mediated SBDP120 (marker of apoptosis). Average SBDP values measured early after injury correlated with severity of injury, CT scan findings and outcome at 6 months post injury.43

A promising candidate biomarker for TBI currently under investigation is Ubiquitin Cterminal Hydrolase-L1 (UCH-L1). UCH-L1 was previously used as a histological marker for neurons due to its high abundance and specific expression in neurons.114 This protein is involved in the addition and removal of ubiquitin from proteins that are destined for metabolism.115 It has an important role in the removal of excessive, oxidized or misfolded proteins during both normal and pathological conditions in neurons.116 In initial studies, UCH-L1 was identified as a protein with a two-fold increase in abundance in the injured cortex 48 hours after controlled cortical impact in a rat model of TBI.117 Subsequently, a UCH-L1 sandwich enzyme-linked immunosorbent assay quantitatively showed that CSF and serum UCH-L1 levels in rats were significantly elevated as early as 2 hours following both traumatic and ischemic injury.118 Clinical studies in humans with severe TBI confirmed, using ELISA analysis, that the UCH-L1 protein was significantly elevated in human CSF44, 119 and was detectable very early after injury and remained significantly elevated for 168 hours post-injury.44 Further studies in severe TBI patients have revealed a very good correlation between CSF and serum levels.120 Most recently, UCH-L1 was detected in the serum of mild and moderate TBI (MMTBI) patients within an hour of injury.121 Serum levels of UCH-L1 discriminated MMTBI patients from uninjured and non-head injured trauma controls and were also able to distinguish mild TBI (concussion patients) from these controls. Most notable was that levels were significantly higher in those with intracranial lesions on CT than those without lesions.121

Glial Fibrillary Acidic Protein (GFAP) is a monomeric intermediate protein found in astroglial skeleton that was first isolated by Eng et al. in 1971.122 GFAP is found in white and gray brain matter and is strongly upregulated during astrogliosis.123 Current evidence indicates that serum GFAP might be a useful marker for various types of brain damage from

patients compared to control patients and these levels correlated with clinical outcome.99, 100 Though levels of c-tau were also elevated in plasma from patients with severe TBI, there was no correlation between plasma levels and clinical outcome.101 A major limitation of all

Alpha-II-spectrin (280 kDa) is the major structural component of the cortical membrane cytoskeleton and is particularly abundant in axons and presynaptic terminals.102, 103 It is also a major substrate for both calpain and caspase-3 cysteine proteases.104, 105 A hallmark feature of apoptosis and necrosis is an early cleavage of several cellular proteins by activated caspases and calpains. A signature of caspase-3 and calpain-2 activation is cleavage of several common proteins such as cytoskeletal αII-spectrin.106 In a rat model, mean levels of both ipsilateral cortex (IC) and cerebral spinal fluid (CSF) spectrin breakdown products (SBDP) at 2, 6, and 24 h after two levels of controlled cortical impact (1.0 mm and 1.6 mm of cortical deformation) were significantly elevated by injury using immunoblotting.107 Using the same proteomic Western blot technique, levels of spectrin breakdown products (SBDP's) have been reported in CSF from adults with severe TBI and they have shown a significant relationship with severity of injury and clinical outcome.42, 108-113 Following a TBI the axonally enriched cytoskeletal protein α-II-spectrin is proteolyzed by calpain and caspase-3 to signature breakdown products (SBDPs). Calpain and caspase-3 mediated SBDP levels in CSF have shown to be significantly increased in TBI patients at several time points after injury, compared to control subjects. The time course of calpain mediated SBDP150 and SBDP145 (markers of necrosis) differs from that of caspase-3 mediated SBDP120 (marker of apoptosis). Average SBDP values measured early after injury correlated with severity of

A promising candidate biomarker for TBI currently under investigation is Ubiquitin Cterminal Hydrolase-L1 (UCH-L1). UCH-L1 was previously used as a histological marker for neurons due to its high abundance and specific expression in neurons.114 This protein is involved in the addition and removal of ubiquitin from proteins that are destined for metabolism.115 It has an important role in the removal of excessive, oxidized or misfolded proteins during both normal and pathological conditions in neurons.116 In initial studies, UCH-L1 was identified as a protein with a two-fold increase in abundance in the injured cortex 48 hours after controlled cortical impact in a rat model of TBI.117 Subsequently, a UCH-L1 sandwich enzyme-linked immunosorbent assay quantitatively showed that CSF and serum UCH-L1 levels in rats were significantly elevated as early as 2 hours following both traumatic and ischemic injury.118 Clinical studies in humans with severe TBI confirmed, using ELISA analysis, that the UCH-L1 protein was significantly elevated in human CSF44, 119 and was detectable very early after injury and remained significantly elevated for 168 hours post-injury.44 Further studies in severe TBI patients have revealed a very good correlation between CSF and serum levels.120 Most recently, UCH-L1 was detected in the serum of mild and moderate TBI (MMTBI) patients within an hour of injury.121 Serum levels of UCH-L1 discriminated MMTBI patients from uninjured and non-head injured trauma controls and were also able to distinguish mild TBI (concussion patients) from these controls. Most notable was that levels were significantly higher in those with intracranial

Glial Fibrillary Acidic Protein (GFAP) is a monomeric intermediate protein found in astroglial skeleton that was first isolated by Eng et al. in 1971.122 GFAP is found in white and gray brain matter and is strongly upregulated during astrogliosis.123 Current evidence indicates that serum GFAP might be a useful marker for various types of brain damage from

of these biomarkers is the lack of specificity for defining neuropathological cascades.

injury, CT scan findings and outcome at 6 months post injury.43

lesions on CT than those without lesions.121

neurodegenerative disorders124, 125 and stroke126 to severe traumatic brain injury.127-131 Recently, Vos et al. described serum GFAP profile in severe and moderate TBI (GCS <12).54 In a recent study by our group, GFAP was systematically assessed in human serum following mild and moderate TBI. We confirmed that the GFAP levels were significantly elevated in this population using ELISA analysis, including those with mild TBI. GFAP was able to discriminate TBI patients from uninjured controls. Additionally, serum levels were able to distinguish orthopedic and motor vehicle controls form TBI patients. GFAP was detectable in serum within a few hours of injury and was associated with measures of injury severity including the GCS score and CT lesions.132, 133 The present work extends findings from studies in severe TBI to mild and moderate TBI.

Fig. 1. The neuroanatomical locations of the above mentioned biomarkers.

#### **7. Attributes of an ideal biomarker for TBI**

Research in the field TBI biomarkers has increased exponentially over the last 20 years with most of the publications on the topic occurring in the last 10 years.134 During the course of our work in the development of TBI biomarkers, it has become evident that there are a number of key features that a clinically useful biomarker should possess.135 An "ideal biomarker" would: 1) demonstrate a high sensitivity and specificity for brain injury; 2) stratify patients by severity of injury; 3) have a rapid appearance in accessible biological fluid; 4) provide information on injury mechanisms; 5) have well defined biokinetc properties; 6) monitor progress of disease and response to treatment; 7) predict functional outcome; 8) be easily measured by widely available, simple techniques

Exploring the Role of

monitoring secondary insults.

**9. Conclusion** 

**10. References** 

353.

distinguishing different pathoanatomic processes of injury.

improve outcome in patients suffering from these injuries.

Biomarkers for the Diagnosis and Management of Traumatic Brain Injury Patients 97

surgery. Biomarker measurements could potentially relate the effects of interventions on molecular and cellular pathways to clinical responses. In doing so, biomarkers would provide an avenue for researchers and clinicians to gain a mechanistic understanding of the

Intoxicated, unconscious, sedated, or polytraumatized patients suspected of having a TBI pose a particular challenge to emergency and trauma physicians. Biomarkers could expedite the evaluation of such patients by providing information on the degree of brain injury prior to neuroimaging. Biomarkers in this setting could also help determine the need for early

There are potential military applications as well. Serum biomarkers could help diagnose and/or triage brain injured military servicemen and women. TBI is a leading cause of combat casualty with an estimated 15-20% of all injuries sustained in 20th century conflicts being to the head.142-144 America's armed forces are sustaining attacks by rocket-propelled grenades, improvised explosive devices, and land mines almost daily in the recent conflicts in Iraq and Afghanistan.145 It has been suggested that over 50% of injuries sustained in combat are the result of such explosive munitions including bombs, grenades, land mines, missiles, and mortar/artillery shells. Neuroimaging techniques to diagnose brain injury acutely and other monitoring tools that assess secondary insults are not immediately available in combat zones and such casualties have to be evacuated. Triage and management of brain injured soldiers could be significantly improved if first responders had a quick and simple means of objectively assessing severity of brain injury and of

There is a unique opportunity to use the insight offered by biochemical markers to shed light on the complexities of the injury process. Accordingly, certain markers could be used as indicators of damage to a particular cell type or cellular process or may be indicative of a particular type of injury. Neuroanatomically, that could include evidence of, say, primary axonal damage versus glial damage. With such heterogeneity the solution may not lie with a single biomarker but more with a complementary panel of markers that may prove useful in

The exploration and validation of biomarkers for TBI using advances in proteomics, neuroimaging, genomics, and bioinformatics must continue. Biomarkers of TBI measured through a simple blood test have the potential to significantly improve the management of TBI patients by providing timely information on the pathophysiology of injury; improving stratification of patients by injury severity; monitoring of secondary insults and injury progression; monitoring response to treatment; and predicting functional outcome. Biomarkers could provide major opportunities for the conduct of clinical research including confirmation of injury mechanism(s) and drug target identification. Ultimately the goal is

[1] Hyder AA, Wunderlich CA, Puvanachandra P, Gururaj G, Kobusingye OC. The impact

of traumatic brain injuries: a global perspective. *NeuroRehabilitation.* 2007;22(5):341-

differences in clinical response that may be influenced by uncontrolled variables.

neurosurgical consultation or transfer to facilities with neurosurgical capabilities.

Clinical researchers have developed methodological standards for developing clinical decision tools in order to ensure the validity of study results.136, 137 As TBI biomarker research transitions from the bench to the bedside there are a number of important methodological issues that researchers will have to consider as they design their clinical protocols. Since TBI biomarkers are being designed for clinical management, the outcome or diagnosis being examined will need to be clearly defined and clinically important. In order to ensure external validity and the generalizability of the results, study patients will have to be selected without bias and represent a wide spectrum of clinical and demographic characteristics. When interpreting the data, clinical variables that potentially affect outcome will require careful consideration in the analysis. Multivariate statistical and bioinformatics models will also further improve classification of patients and help reduce systematic bias.138 Another essential consideration will be the examination of biokinetic properties and temporal profiles of the biomarkers as well as systematic comparisons to controls.

#### **8. The potential clinical role of biomarkers**

Biochemical markers could help with clinical decision making by elucidating injury severity, injury mechanism(s), and monitoring progression of injury. Temporal profiles of changes in biomarkers could guide timing of diagnosis and treatment. Biomarkers could have a role in management decisions regarding patients at high risk of repeated injury. Accurate identification of these patients could facilitate development of guidelines for return to duty, work or sports activities and also provide opportunities for counseling of patients suffering from these deficits. Repeated mild TBI occurring within a short period (i.e. hours, days, or weeks) can be catastrophic or fatal, a phenomenon termed "second impact syndrome."139, 140 Acute CT or MRI abnormalities are not usually found after these injuries, but levels of some neurotransmitters remain elevated, and a hypermetabolic state may persist in the brain for several days after the initial injury.141 During this time the brain appears to be particularly vulnerable to additional TBI, which may result in severe sequelae, including greatly increased cerebral edema and even death.139, 140

Biomarkers could serve as prognostic indicators by providing information for patients and their families about the expected course of recovery. It opens the door to the initiation of early therapies. Identifying at-risk patients with less apparent TBI or differentiating injury pathology in those with more severe intracranial processes would be tremendously valuable in the management of these patients. For example, in a patient with a normal CT scan or MRI, a biomarker that could predict worsening neurological status or long-term disability would have great clinical utility.

There have been a large number of clinical trials studying potential therapies for traumatic brain injury (TBI) that have resulted in negative findings. Biomarkers measurable in blood would have important applications in clinical research of these injuries. Biomarkers could provide clinical trial outcome measures that are cost-effective and more readily available than conventional neurological assessments, thereby significantly reducing the risks and costs of human clinical trials. Biomarkers that represent highly sensitive and specific indicators of disease pathways have been used as substitutes for outcomes in clinical trials when evidence indicates that they predict clinical risk or benefit.

Lack of quickly accessible pathophysiologic information during the post-injury course has made pharmacologic intervention problematic. Biomarkers could provide more timely information on disease progression and the effects of interventions such as drugs and surgery. Biomarker measurements could potentially relate the effects of interventions on molecular and cellular pathways to clinical responses. In doing so, biomarkers would provide an avenue for researchers and clinicians to gain a mechanistic understanding of the differences in clinical response that may be influenced by uncontrolled variables.

Intoxicated, unconscious, sedated, or polytraumatized patients suspected of having a TBI pose a particular challenge to emergency and trauma physicians. Biomarkers could expedite the evaluation of such patients by providing information on the degree of brain injury prior to neuroimaging. Biomarkers in this setting could also help determine the need for early neurosurgical consultation or transfer to facilities with neurosurgical capabilities.

There are potential military applications as well. Serum biomarkers could help diagnose and/or triage brain injured military servicemen and women. TBI is a leading cause of combat casualty with an estimated 15-20% of all injuries sustained in 20th century conflicts being to the head.142-144 America's armed forces are sustaining attacks by rocket-propelled grenades, improvised explosive devices, and land mines almost daily in the recent conflicts in Iraq and Afghanistan.145 It has been suggested that over 50% of injuries sustained in combat are the result of such explosive munitions including bombs, grenades, land mines, missiles, and mortar/artillery shells. Neuroimaging techniques to diagnose brain injury acutely and other monitoring tools that assess secondary insults are not immediately available in combat zones and such casualties have to be evacuated. Triage and management of brain injured soldiers could be significantly improved if first responders had a quick and simple means of objectively assessing severity of brain injury and of monitoring secondary insults.

There is a unique opportunity to use the insight offered by biochemical markers to shed light on the complexities of the injury process. Accordingly, certain markers could be used as indicators of damage to a particular cell type or cellular process or may be indicative of a particular type of injury. Neuroanatomically, that could include evidence of, say, primary axonal damage versus glial damage. With such heterogeneity the solution may not lie with a single biomarker but more with a complementary panel of markers that may prove useful in distinguishing different pathoanatomic processes of injury.

#### **9. Conclusion**

96 Proteomics – Human Diseases and Protein Functions

Clinical researchers have developed methodological standards for developing clinical decision tools in order to ensure the validity of study results.136, 137 As TBI biomarker research transitions from the bench to the bedside there are a number of important methodological issues that researchers will have to consider as they design their clinical protocols. Since TBI biomarkers are being designed for clinical management, the outcome or diagnosis being examined will need to be clearly defined and clinically important. In order to ensure external validity and the generalizability of the results, study patients will have to be selected without bias and represent a wide spectrum of clinical and demographic characteristics. When interpreting the data, clinical variables that potentially affect outcome will require careful consideration in the analysis. Multivariate statistical and bioinformatics models will also further improve classification of patients and help reduce systematic bias.138 Another essential consideration will be the examination of biokinetic properties and

temporal profiles of the biomarkers as well as systematic comparisons to controls.

Biochemical markers could help with clinical decision making by elucidating injury severity, injury mechanism(s), and monitoring progression of injury. Temporal profiles of changes in biomarkers could guide timing of diagnosis and treatment. Biomarkers could have a role in management decisions regarding patients at high risk of repeated injury. Accurate identification of these patients could facilitate development of guidelines for return to duty, work or sports activities and also provide opportunities for counseling of patients suffering from these deficits. Repeated mild TBI occurring within a short period (i.e. hours, days, or weeks) can be catastrophic or fatal, a phenomenon termed "second impact syndrome."139, 140 Acute CT or MRI abnormalities are not usually found after these injuries, but levels of some neurotransmitters remain elevated, and a hypermetabolic state may persist in the brain for several days after the initial injury.141 During this time the brain appears to be particularly vulnerable to additional TBI, which may result in severe sequelae, including greatly

Biomarkers could serve as prognostic indicators by providing information for patients and their families about the expected course of recovery. It opens the door to the initiation of early therapies. Identifying at-risk patients with less apparent TBI or differentiating injury pathology in those with more severe intracranial processes would be tremendously valuable in the management of these patients. For example, in a patient with a normal CT scan or MRI, a biomarker that could predict worsening neurological status or long-term disability

There have been a large number of clinical trials studying potential therapies for traumatic brain injury (TBI) that have resulted in negative findings. Biomarkers measurable in blood would have important applications in clinical research of these injuries. Biomarkers could provide clinical trial outcome measures that are cost-effective and more readily available than conventional neurological assessments, thereby significantly reducing the risks and costs of human clinical trials. Biomarkers that represent highly sensitive and specific indicators of disease pathways have been used as substitutes for outcomes in clinical trials

Lack of quickly accessible pathophysiologic information during the post-injury course has made pharmacologic intervention problematic. Biomarkers could provide more timely information on disease progression and the effects of interventions such as drugs and

**8. The potential clinical role of biomarkers** 

increased cerebral edema and even death.139, 140

when evidence indicates that they predict clinical risk or benefit.

would have great clinical utility.

The exploration and validation of biomarkers for TBI using advances in proteomics, neuroimaging, genomics, and bioinformatics must continue. Biomarkers of TBI measured through a simple blood test have the potential to significantly improve the management of TBI patients by providing timely information on the pathophysiology of injury; improving stratification of patients by injury severity; monitoring of secondary insults and injury progression; monitoring response to treatment; and predicting functional outcome. Biomarkers could provide major opportunities for the conduct of clinical research including confirmation of injury mechanism(s) and drug target identification. Ultimately the goal is improve outcome in patients suffering from these injuries.

#### **10. References**

[1] Hyder AA, Wunderlich CA, Puvanachandra P, Gururaj G, Kobusingye OC. The impact of traumatic brain injuries: a global perspective. *NeuroRehabilitation.* 2007;22(5):341- 353.

Exploring the Role of

738.

1182.

May 2002;19(5):503-557.

*Brain Injury.* 2000;14:851-857.

*Med.* Dec 2008;52(6):714-748.

2008;81(965):362-378.

487; discussion 487-489.

*Med.* Dec 14 2009;169(22):2071-2077.

*Emerg Med.* May 2008;9(2):120-122.

beginning. *Nat Neurosci.* May 2004;7(5):440-445.

synaptic proteins. *J Biol Chem.* Feb 18 2005;280(7):5972-5982.

injury. *Neurological Research.* 2001;Mar-Apr(23(2-3)):190-192.

Biomarkers for the Diagnosis and Management of Traumatic Brain Injury Patients 99

[21] Narayan RK, Michel ME, Ansell B, et al. Clinical trials in head injury. *J Neurotrauma.* 

[22] Saatman KE, Duhaime AC, Bullock R, Maas AI, Valadka A, Manley GT. Classification

[23] Doppenberg EM, Choi SC, Bullock R. Clinical trials in traumatic brain injury: lessons for

[24] Maas AI, Hukkelhoven CW, Marshall LF, Steyerberg EW. Prediction of outcome in

[25] Kesler ea. APECT, MR and quantitative MR imaging: correlates with neuropsycholgical.

[26] Jagoda AS, Bazarian JJ, Bruns JJ, Jr., et al. Clinical policy: neuroimaging and

[27] Wardlaw JM, Keir SL, Seymour J, et al. What is the best imaging strategy for acute

[28] Brenner DJ, Hall EJ. Computed tomography--an increasing source of radiation

[29] Fayngersh V, Passero M. Estimating radiation risk from computed tomography

[30] Hall EJ, Brenner DJ. Cancer risks from diagnostic radiology. *Br J Radiol.* May

[31] Heilbrun ME, Chew FS, Tansavatdi KR, Tooze JA. The role of negative CT of the

[33] Berrington de Gonzalez A, Mahesh M, Kim KP, et al. Projected cancer risks from

[34] Schwartz DT. Counter-Point: Are We Really Ordering Too Many CT Scans? *West J* 

[35] Stiell IG, Wells GA, Vandemheen K, et al. Variation in ED use of computed tomography for patients with minor head injury. *Ann Emerg Med.* Jul 1997;30(1):14-22. [36] Choudhary J, Grant SG. Proteomics in postgenomic neuroscience: the end of the

[37] Collins MO, Yu L, Coba MP, et al. Proteomic analysis of in vivo phosphorylated

[38] Choi SC, Bullock R. Design and statistical issues in multicenter trials of severe head

[39] Doppenberg EM, Choi SC, Bullock R. Clinical trials in traumatic brain injury. What can we learn from previous studies? *Ann N Y Acad Sci.* Oct 15 1997;825:305-322.

department after blunt trauma. *J Am Coll Radiol.* Nov 2005;2(11):889-895. [32] Livingston DH, Loder PA, Koziol J, Hunt CD. The use of CT scanning to triage patients

abdomen and pelvis in the decision to admit adults from the emergency

requiring admission following minimal head injury. *J Trauma.* Apr 1991;31(4):483-

computed tomographic scans performed in the United States in 2007. *Arch Intern* 

the future. *J Neurosurg Anesthesiol.* Jan 2004;16(1):87-94.

stroke? *Health Technol Assess.* Jan 2004;8(1):iii, ix-x, 1-180.

exposure. *N Engl J Med.* Nov 29 2007;357(22):2277-2284.

scanning. *Lung.* May-Jun 2009;187(3):143-148.

of traumatic brain injury for targeted therapies. *J Neurotrauma.* Jul 2008;25(7):719-

traumatic brain injury with computed tomographic characteristics: a comparison between the computed tomographic classification and combinations of computed tomographic predictors. *Neurosurgery.* Dec 2005;57(6):1173-1182; discussion 1173-

decisionmaking in adult mild traumatic brain injury in the acute setting. *Ann Emerg* 


[2] Consensus conference. Rehabilitation of persons with traumatic brain injury. NIH

[3] Faul M, Xu L, Wald MM, Coronado VG. Traumatic brain injury in the United States:

[4] Hoffman SW, Shesko K, Harrison CR. Enhanced neurorehabilitation techniques in the DVBIC Assisted Living Pilot Project. *NeuroRehabilitation.*26(3):257-269. [5] Tagliaferri F, Compagnone C, Korsic M, Servadei F, Kraus J. A systematic review of

[6] Graham DI, Adams JH, Nicoll JA, Maxwell WL, Gennarelli TA. The nature, distribution and causes of traumatic brain injury. *Brain Pathol.* Oct 1995;5(4):397-406. [7] Graham DI, Horsburgh K, Nicoll JA, Teasdale GM. Apolipoprotein E and the response

[8] Yealy DM, Hogan DE. Imaging after head trauma. Who needs what? *Emerg Med Clin* 

[9] Vollmer DG, Dacey RG, Jr. The management of mild and moderate head injuries.

[10] Langlois JA, Rutland-Brown W, Thomas KE. *Traumatic Brain Injury in the United States:* 

[12] Thurman DJ. Epidemiology and Economics of Head Trauma. *Head Trauma: Basic Preclinical and Clinical Directions*. New York: Wiley-Liss; 2001:327-347. [13] Stein SC, Fabbri A, Servadei F, Glick HA. A critical comparison of clinical decision

injury in adolescents and adults. *Ann Emerg Med.* Feb 2009;53(2):180-188. [14] Millis SR, Rosenthal M, Novack TA, et al. Long-term neuropsychological outcome after traumatic brain injury. *J Head Trauma Rehabil.* Aug 2001;16(4):343-355. [15] Alves W, Macciocchi S, Barth JT. Postconcussive Symptoms After Uncomplicated Mild

[16] Rimel RW, Giordani B, Barth JT, Boll TJ, Jane JA. Disability caused by minor head

[17] Alexander MP. Mild traumatic brain injury: pathophysiology, natural history, and

[18] Barth JT, Macciocchi SN, Giordani B, Rimel R, Jane JA, Boll TJ. Neuropsychological sequelae of minor head injury. *Neurosurgery.* Nov 1983;13(5):529-533. [19] Kennedy RE, Livingston L, Marwitz JH, Gueck S, Kreutzer JS, Sander AM. Complicated

[20] Kennedy JE, Lumpkin RJ, Grissom JR. A survey of mild traumatic brain injury

mild traumatic brain injury on the inpatient rehabilitation unit: a multicenter

treatment in the emergency room and primary care medical clinics. *Mil Med.* Jun

[11] TBI State Demonstration Grants. *J Head Trauma Rehabil.* Feb 2000;15(1):750-760.

Head Injury. *J Head Trauma Rehabil.* 1993 1993;8(3):48-59.

clinical management. *Neurology.* 1995;45(7):1253-1260.

analysis. *J Head Trauma Rehabil.* May-Jun 2006;21(3):260-271.

injury. *Neurosurgery.* Sep 1981;9(3):221-228.

*Emergency Department Visits, Hospitalizations, and Deaths.* Atlanta: Division of Injury and Disability Outcomes and Programs. National Center for Injury Prevention and

instruments for computed tomographic scanning in mild closed traumatic brain

of the brain to injury. *Acta Neurochir Suppl.* 1999;73:89-92.

Injury. *Jama.* 1999;282(10):974-983.

*North Am.* Nov 1991;9(4):707-717.

Control. CDC; October 2004.

2006;171(6):516-521.

*Neurosurg Clin N Am.* Apr 1991;2(2):437-455.

268; discussion 268.

*Control.* March 2010 ed. Atlanta, GA; 2010.

Consensus Development Panel on Rehabilitation of Persons With Traumatic Brain

emergency department visits, hospitalizations, and deaths. In: Services USDoHaH, ed. *Centers for Disease Control and Prevention, National Center for Injury Prevention and* 

brain injury epidemiology in Europe. *Acta Neurochir (Wien).* Mar 2006;148(3):255-


Exploring the Role of

Biomarkers for the Diagnosis and Management of Traumatic Brain Injury Patients 101

[59] Yamazaki Y, Yada K, Morii S, Kitahara T, Ohwada T. Diagnostic significance of serum

[60] de Kruijk JR, Leffers P, Menheere PP, Meerhoff S, Twijnstra A. S-100B and neuron-

[61] Raabe A, Grolms C, Seifert V. Serum markers of brain damage and outcome prediction in patients after severe head injury. *Br J Neurosurg.* Feb 1999;13(1):56-59. [62] Ross SA, Cunningham RT, Johnston CF, Rowlands BJ. Neuron-specific enolase as an aid to outcome prediction in head injury. *Br J Neurosurg.* Oct 1996;10(5):471-476. [63] Naeimi ZS, Weinhofer A, Sarahrudi K, Heinz T, Vecsei V. Predictive value of S-100B

[64] Berger RP, Adelson PD, Pierce MC, Dulani T, Cassidy LD, Kochanek PM. Serum

[65] Berger RP, Beers SR, Richichi R, Wiesman D, Adelson PD. Serum biomarker

[66] Beers SR, Berger RP, Adelson PD. Neurocognitive outcome and serum biomarkers in

[67] Piazza O, Storti MP, Cotena S, et al. S100B is not a reliable prognostic index in

[68] Martens P. Serum neuron-specific enolase as a prognostic marker for irreversible brain damage in comatose cardiac arrest surviviors. *Acad Emerg Med.* 1996;3:126-131. [69] Rainey T, Lesko M, Sacho R, Lecky F, Childs C. Predicting outcome after severe

[70] Bazarian JJ, Zemlan FP, Mookerjee S, Stigbrand T. Serum S-100B and cleaved-tau are

[71] Watt SE, Shores EA, Baguley IJ, Dorsch N, Fearnside MR. Protein S-100 and

[72] Morochovic R, Racz O, Kitka M, et al. Serum S100B protein in early management of patients after mild traumatic brain injury. *Eur J Neurol.* Oct 2009;16(10):1112-1117. [73] Dirnagl U CI, and Moskowitz MA. Pathology of ischaemic stroke: an integrated view.

[74] Laskowitz ea. Serum Markers of Cerebral Ischemia. *Journal of Stroke and Cerebrovascular* 

[75] Roine ea. Neurological outcome after out-of-hospital cardiac arrest. Prediction by

cerebrospinal fluid enzyme analysis. *Arch Neurol.* 1989;46:753-756.

injury. *Surg Neurol.* Mar 1995;43(3):267-270; discussion 270-261.

with health controls. *Acta Neurol Scand.* Mar 2001;103(3):175-179.

clinical use. *Brain Inj.* May 2006;20(5):463-468.

2005;103(1 Suppl):61-68.

Dec 2007;24(12):1793-1801.

Jun 2006;20(7):759-765.

Sep 2006;20(10):1007-1017.

*TINS.* 1999;22(9):391-397.

*Diseases.* 1998;7(4 (July-August)):234-241.

*Neurotrauma.* Jan 2007;24(1):97-105.

paediatric TBI. *Pediatr Neurosurg.* 2007;43(4):258-264.

(24h) time-point. *Resuscitation.* Mar 2009;80(3):341-345.

neuron-specific enolase and myelin basic protein assay in patients with acute head

specific enolase in serum of mild traumatic brain injury patients. A comparison

protein and neuron specific-enolase as markers of traumatic brain damage in

neuron-specific enolase, S100B, and myelin basic protein concentrations after inflicted and noninflicted traumatic brain injury in children. *J Neurosurg.* Jul

concentrations and outcome after pediatric traumatic brain injury. *J Neurotrauma.* 

inflicted versus non-inflicted traumatic brain injury in young children. *J* 

traumatic brain injury using the serum S100B biomarker: results using a single

poor predictors of long-term outcome after mild traumatic brain injury. *Brain Inj.* 

neuropsychological functioning following severe traumatic brain injury. *Brain Inj.* 


[40] Denslow N, Michel ME, Temple MD, Hsu CY, Saatman K, Hayes RL. Application of

[41] Wang KK, Ottens AK, Liu MC, et al. Proteomic identification of biomarkers of traumatic brain injury. *Expert Rev Proteomics.* Aug 2005;2(4):603-614. [42] Pineda JA, Lewis SB, Valadka AB, et al. Clinical significance of alphaII-spectrin

[43] Brophy GM, Pineda JA, Papa L, et al. alphaII-Spectrin breakdown product

[44] Papa L, Akinyi L, Liu MC, et al. Ubiquitin C-terminal hydrolase is a novel biomarker in humans for severe traumatic brain injury. *Crit Care Med.* Jan 2010;38(1):138-144. [45] MisslerU. S-100 protein and neuron-specific enolase concentrations in blood as

[46] Ytrebo LM NG, Korvald C, et al. Renal elimination of protein S-100beta in picgs with

[47] Jonsson H JP, Hoglund P, Alling C, Blomquist S. The elimination of S-100b and renal function after cardiac surgery. *J Cardiothorac Vasc Aneth.* 2000;14:698-701. [48] Usui A KK, Abe T, Murase M, Tanaka M, Takeuchi E. S-100ao protein in blood and

[49] Raabe A, Grolms C, Seifert V. Serum markers of brain damage and outcome prediction in patients after severe head injury. *Br J Neurosurg.* 1999;13(1):56-59. [50] Haimoto HH, S; Kato, K. Differential distribution of immunoreactive S100-a and S100-b proteins in normal nonnervous human tissues. *Lab Invest.* 1987;57:489-498. [51] Woertgen C, Rothoerl RD, Holzschuh M, Metz C, Brawanski A. Comparison of serial S-

[52] Romner B, Ingebrigtsen T, Kongstad P, Borgesen SE. Traumatic brain damage: serum S-

[53] Korfias S, Stranjalis G, Boviatsis E, et al. Serum S-100B protein monitoring in patients with severe traumatic brain injury. *Intensive Care Med.* Feb 2007;33(2):255-260. [54] Vos PE, Jacobs B, Andriessen TM, et al. GFAP and S100B are biomarkers of traumatic

[55] Berger RP, Pierce MC, Wisniewski SR, Adelson PD, Kochanek PM. Serum S100B

[56] BW M. A soluble protein characteristic of the nervous system. *Biochem Biophys Res* 

[57] Donato R. Functional roles of S100 proteins, calcium-binding proteins of the EF-hand

100 and NSE serum measurements after severe head injury. *Acta Neurochir (Wien).* 

100 protein measurements related to neuroradiological findings. *J Neurotrauma.* 

brain injury: an observational cohort study. *Neurology.* Nov 16 2010;75(20):1786-

concentrations are increased after closed head injury in children: a preliminary

acute encephalopathy. *Scand J Clin Lab Invest.* 2001;61:217-225.

urine during open-heart surgery. *Clin Chem.* 1989;35:1942-1944.

2003;20(5):401-407.

1997;28:1956-1960.

479.

*Neurotrauma.* Feb 2007;24(2):354-366.

1997;139(12):1161-1164; discussion 1165.

study. *J Neurotrauma.* Nov 2002;19(11):1405-1409.

type. *Biochim Biophys Acta.* 1999;1450:191-231.

[58] Cooper E. Neuron-specific enolase. *Int J Biol Markers.* 1994(4):205-210.

Aug 2000;17(8):641-647.

*Commun.* 1965;19:739-744.

1793.

proteomics technology to the field of neurotrauma. *J Neurotrauma.* May

breakdown products in cerebrospinal fluid after severe traumatic brain injury. *J* 

cerebrospinal fluid exposure metrics suggest differences in cellular injury mechanisms after severe traumatic brain injury. *J Neurotrauma.* Apr 2009;26(4):471-

indicators of infarction volume and prognosis in acute ischemic stroke. *Stroke.* 


Exploring the Role of

Biomarkers for the Diagnosis and Management of Traumatic Brain Injury Patients 103

[93] Ergun R, Bostanci U, Akdemir G, et al. Prognostic value of serum neuron-specific

[94] Varma S, Janesko KL, Wisniewski SR, et al. F2-isoprostane and neuron-specific enolase

[95] Bandyopadhyay S, Hennes H, Gorelick MH, Wells RG, Walsh-Kelly CM. Serum

[97] Kosik KS, Finch EA. MAP2 and tau segregate into dendritic and axonal domains after

[98] Higuchi M, Lee VM, Trojanowski JQ. Tau and axonopathy in neurodegenerative

[99] Shaw GJ, Jauch EC, Zemlan FP. Serum cleaved tau protein levels and clinical outcome in adult patients with closed head injury. *Ann Emerg Med.* Mar 2002;39(3):254-257. [100] Zemlan FP, Jauch EC, Mulchahey JJ, et al. C-tau biomarker of neuronal damage in

[101] Chatfield DA, Zemlan FP, Day DJ, Menon DK. Discordant temporal patterns of

[102] Goodman SR, Zimmer WE, Clark MB, Zagon IS, Barker JE, Bloom ML. Brain spectrin:

[103] Riederer BM, Zagon IS, Goodman SR. Brain spectrin(240/235) and brain

[104] Wang KK, Posmantur R, Nath R, et al. Simultaneous degradation of alphaII- and

[105] McGinn MJ, Kelley BJ, Akinyi L, et al. Biochemical, structural, and biomarker evidence

[106] Pike BR, Flint J, Dave JR, et al. Accumulation of calpain and caspase-3 proteolytic

[108] Cardali S, Maugeri R. Detection of alphaII-spectrin and breakdown products in humans after severe traumatic brain injury. *J Neurosurg Sci.* Jun 2006;50(2):25-31. [109] Papa L, D'Avella D, Aguennouz M, et al. Detection of Alpha-II Spectrin And

of cultured rat cerebrum. *J Neurosci.* Oct 1987;7(10):3142-3153.

disorders. *Neuromolecular Med.* 2002;2(2):131-150.

*Neurosurg.* Oct 2002;16(5):471-476.

1998;273(35):22490-22497.

2004;21(10):1443-1456.

*Acad Emerg Med.* May 2004;11(5).

clinical outcome. *Brain Res.* Aug 23 2002;947(1):131-139.

of mice and men. *Brain Res Bull.* 1995;36(6):593-606.

mammalian neural cells. *J Cell Biol.* Jun 1986;102(6):2088-2097.

by contusion. *J Neuropathol Exp Neurol.* Mar 2009;68(3):241-249.

closed traumatic brain injury. *Acad Emerg Med.* Aug 2005;12(8):732-738. [96] Johnsson P, Blomquist S, Luhrs C, et al. Neuron-specific enolase increases in plasma

in cerebrospinal fluid after severe traumatic brain injury in infants and children. *J* 

neuron-specific enolase as a predictor of short-term outcome in children with

during and immediately after extracorporeal circulation. *Ann Thorac Surg.* Mar

the elaboration of morphologically distinct neurites: an immunocytochemical study

severe brain injured patients: association with elevated intracranial pressure and

S100beta and cleaved tau protein elevation after head injury: a pilot study. *Br J* 

spectrin(240/235E): two distinct spectrin subtypes with different locations within

betaII-spectrin by caspase 3 (CPP32) in apoptotic cells. *J Biol Chem.* Aug 28

for calpain-mediated cytoskeletal change after diffuse brain injury uncomplicated

fragments of brain-derived alphaII-spectrin in cerebral spinal fluid after middle cerebral artery occlusion in rats. *J Cereb Blood Flow Metab.* Jan 2004;24(1):98-106. [107] Ringger NC, O'Steen BE, Brabham JG, et al. A novel marker for traumatic brain injury:

CSF alphaII-spectrin breakdown product levels. *J Neurotrauma.* Oct

Breakdown Products In Humans After Severe Traumatic Brain Injury [abstract].

enolase levels after head injury. *Neurol Res.* Jul 1998;20(5):418-420.

*Neurotrauma.* Aug 2003;20(8):781-786.

2000;69(3):750-754.


[76] Xiong H, Liang WL, Wu XR. [Pathophysiological alterations in cultured astrocytes

[77] Zimmer DB, Cornwall EH, Landar A, Song W. The S100 protein family: history,

[78] Ingebrigtsen T, Romner B. Management of minor head injuries in hospitals in Norway.

[79] Waterloo K, Ingebrigtsen T, Romner B. Neuropsychological function in patients with

[80] Ingebrigtsen T, Romner B. Serial S-100 protein serum measurements related to early

[81] Ingebrigtsen T, Waterloo K, Jacobsen EA, Langbakk B, Romner B. Traumatic brain

[82] Ingebrigtsen T, Romner B, Marup-Jensen S, et al. The clinical value of serum S-100

[83] Muller K, Townend W, Biasca N, et al. S100B serum level predicts computed tomography findings after minor head injury. *J Trauma.* Jun 2007;62(6):1452-1456. [84] Biberthaler P, Linsenmeier U, Pfeifer KJ, et al. Serum S-100B concentration provides

[85] Phillips JP, Jones HM, Hitchcock R, Adama N, Thompson RJ. Radioimmunoassay of

[86] Rothoerl RD, Woertgen C, Holzschuh M, Metz C, Brawanski A. S-100 serum levels after

[87] Bechtel K, Frasure S, Marshall C, Dziura J, Simpson C. Relationship of serum S100B

[88] Rothoerl RD, Woertgen C. High serum S100B levels for trauma patients without head injuries. *Neurosurgery.* Dec 2001;49(6):1490-1491; author reply 1492-1493. [89] Romner B, Ingebrigtsen T. High serum S100B levels for trauma patients without head

[90] Anderson RE, Hansson LO, Nilsson O, Dijlai-Merzoug R, Settergen G. High serum

[91] Skogseid IM, Nordby HK, Urdal P, Paus E, Lilleaas F. Increased serum creatine kinase

[92] Schmechel D, Marangos PJ, Brightman M. Neurone-specific enolase is a molecular

injuries. *Neurosurgery.* Dec 2001;49(6):1490; author reply 1492-1493.

minor and major head injury. *J Trauma.* Oct 1998;45(4):765-767.

function, and expression. *Brain Res Bull.* 1995;37(4):417-429.

*Acta Neurol Scand.* Jan 1997;95(1):51-55.

1999;45(3):468-475; discussion 475-466.

*Brain Inj.* Dec 2000;14(12):1047-1055.

20 1980;281(6243):777-779.

2009;124(4):e697-704.

2001;49(5):1272-1273.

1978;276(5690):834-836.

*Neurochir (Wien).* 1992;115(3-4):106-111.

*(Wien).* 1997;139(1):26-31; discussion 31-22.

221.

1996;85(5):945-948.

exposed to hypoxia/reoxygenation]. *Sheng Li Ke Xue Jin Zhan.* Jul 2000;31(3):217-

increased serum levels of protein S-100 after minor head injury. *Acta Neurochir* 

magnetic resonance imaging after minor head injury. Case report. *J Neurosurg.* Nov

damage in minor head injury: relation of serum S-100 protein measurements to magnetic resonance imaging and neurobehavioral outcome. *Neurosurgery.* Sep

protein measurements in minor head injury: a Scandinavian multicentre study.

additional information fot the indication of computed tomography in patients after minor head injury: a prospective multicenter study. *Shock.* May 2006;25(5):446-453.

serum creatine kinase BB as index of brain damage after head injury. *Br Med J.* Sep

levels and intracranial injury in children with closed head trauma. *Pediatrics.* Oct

S100B levels for trauma patients without head injuries. *Neurosurgery.* 

BB and neuron specific enolase following head injury indicates brain damage. *Acta* 

marker for peripheral and central neuroendocrine cells. *Nature.* Dec 21-28


Exploring the Role of

2677.

1012.

2002;326(1-2):151-154.

*Med.* Nov 7 2011.

2008;2(8):937-945.

Feb 2007;24(2):232-238.

1998;17(1):45-60.

1999;13(2):193-209.

*Care.* Apr 2008;14(2):135-141.

*Acad Emerg Med.* May 2008;15(5):Suppl.

Biomarkers for the Diagnosis and Management of Traumatic Brain Injury Patients 105

[127] Missler U, Wiesmann M, Wittmann G, Magerkurth O, Hagenstrom H. Measurement of

[128] Pelinka LE, Kroepfl A, Leixnering M, Buchinger W, Raabe A, Redl H. GFAP versus

[129] Pelinka LE, Kroepfl A, Schmidhammer R, et al. Glial fibrillary acidic protein in serum

[130] van Geel WJ, de Reus HP, Nijzing H, Verbeek MM, Vos PE, Lamers KJ. Measurement

[131] Nylen K, Ost M, Csajbok LZ, et al. Increased serum-GFAP in patients with severe

[132] Papa L, Akinyi L, Demery J, et al. Levels of Serum GFAP Are Associated With Severity

[133] Papa L, Lewis LM, Falk JL, et al. Elevated Levels of Serum Glial Fibrillary Acidic

[134] Kochanek PM, Berger RP, Bayr H, Wagner AK, Jenkins LW, Clark RS. Biomarkers of

[135] Papa L, Robinson G, Oli M, et al. Use of Biomarkers for Diagnosis and Management of

[136] Stiell IG, Wells GA. Methodologic standards for the development of clinical decision rules in emergency medicine. *Ann Emerg Med.* Apr 1999;33(4):437-447. [137] Laupacis A, Sekar N, Stiell IG. Clinical prediction rules. A review and suggested modifications of methodological standards. *Jama.* Feb 12 1997;277(6):488-494. [138] Maas AI, Marmarou A, Murray GD, Teasdale SG, Steyerberg EW. Prognosis and

[139] Cantu RC. Return to play guidelines after a head injury. *Clin Sports Med.* Jan

[140] Erlanger DM, Kutner KC, Barth JT, Barnes R. Neuropsychology of sports-related head

[141] McCrory PR, Berkovic SF. Second impact syndrome. *Neurology.* Mar 1998;50(3):677-683. [142] Carey ME. Analysis of wounds incurred by U.S. Army Seventh Corps personnel

10, 1991. *J Trauma.* Mar 1996;40(3 Suppl):S165-169.

clinical results. *Clin Chem.* Jan 1999;45(1):138-141.

outcome. *J Neurotrauma.* Nov 2004;21(11):1553-1561.

of protein S-100B and glial fibrillary acidic protein. *Stroke.* Nov 2000;31(11):2670-

glial fibrillary acidic protein in human blood: analytical method and preliminary

S100B in serum after traumatic brain injury: relationship to brain damage and

after traumatic brain injury and multiple trauma. *J Trauma.* Nov 2004;57(5):1006-

of glial fibrillary acidic protein in blood: an analytical method. *Clin Chim Acta.* Dec

traumatic brain injury is related to outcome. *J Neurol Sci.* Jan 15 2006;240(1-2):85-91.

Of Injury In Patients With Mild And Moderate Traumatic Brain Injury [abstract].

Protein Breakdown Products in Mild and Moderate Traumatic Brain Injury Are Associated With Intracranial Lesions and Neurosurgical Intervention. *Ann Emerg* 

primary and evolving damage in traumatic and ischemic brain injury: diagnosis, prognosis, probing mechanisms, and therapeutic decision making. *Curr Opin Crit* 

Traumatic Brain Injury Patients. *Expert Opinion on Medical Diagnostics.* 

clinical trial design in traumatic brain injury: the IMPACT study. *J Neurotrauma.* 

injury: Dementia Pugilistica to Post Concussion Syndrome. *Clin Neuropsychol.* May

treated in Corps hospitals during Operation Desert Storm, February 20 to March


[110] Papa L, Lewis SB, Heaton S, et al. Predicting Early Outcome Using Alpha-II Spectrin

[111] Papa L, Pineda J, Wang KKW, et al. Levels of Alpha-II Spectrin Breakdown Products in

[112] Farkas O, Polgar B, Szekeres-Bartho J, Doczi T, Povlishock JT, Buki A. Spectrin

[113] Mondello S, Robicsek SA, Gabrielli A, et al. alphaII-spectrin breakdown products

[114] Jackson P, Thompson RJ. The demonstration of new human brain-specific proteins by

[115] Tongaonkar P, Chen L, Lambertson D, Ko B, Madura K. Evidence for an interaction

[116] Gong B, Leznik E. The role of ubiquitin C-terminal hydrolase L1 in neurodegenerative

[117] Kobeissy FH, Ottens AK, Zhang Z, et al. Novel differential neuroproteomics analysis of traumatic brain injury in rats. *Mol Cell Proteomics.* Oct 2006;5(10):1887-1898. [118] Liu MC, Akinyi L, Scharf D, et al. Ubiquitin C-terminal hydrolase-L1 as a biomarker

[119] Siman R, Toraskar N, Dang A, et al. A panel of neuron-enriched proteins as markers for traumatic brain injury in humans. *J Neurotrauma.* Nov 2009;26(11):1867-1877. [120] Brophy G, Mondello S, Papa L, et al. Biokinetic Analysis of Ubiquitin C-Terminal

[121] Papa L, Lewis LM, Falk JL, et al. Serum levels of UCH-L1 distinguishes mild and

[122] Eng LF, Vanderhaeghen JJ, Bignami A, Gerstl B. An acidic protein isolated from

[123] Duchen LW. General pathology of neurons and neuroglia. In: Adams JA, Corsellis

[124] Baydas G, Nedzvetskii VS, Tuzcu M, Yasar A, Kirichenko SV. Increase of glial fibrillary

[125] Mouser PE, Head E, Ha KH, Rohn TT. Caspase-mediated cleavage of glial fibrillary

[126] Herrmann M, Vos P, Wunderlich MT, de Bruijn CH, Lamers KJ. Release of glial tissue-

observations. *Acta Neurochir (Wien).* Aug 2005;147(8):855-861.

disorders. *Drug News Perspect.* Jul-Aug 2007;20(6):365-370.

fibrous astrocytes. *Brain Res.* May 7 1971;28(2):351-354.

vitamin E. *Eur J Pharmacol.* Feb 21 2003;462(1-3):67-71.

*Pathol.* Mar 2006;168(3):936-946.

*Acad Emerg Med.* May 2006;13(5 (Suppl 1)).

*Emerg Med.* May 2005;12(5 (Suppl 1)).

*Neurotrauma.* Jul 2010;27(7):1203-1213.

Mar 1981;49(3):429-438.

2000;20(13):4691-4698.

*Neurotrauma.* Feb 10.

1984:1-52.

732.

Breakdown Products In Human CSF After Severe Traumatic Brain Injury [abstract].

Human CSF and Outcome After Severe Traumatic Brain Injury [abstract]. *Acad* 

breakdown products in the cerebrospinal fluid in severe head injury--preliminary

(SBDPs): diagnosis and outcome in severe traumatic brain injury patients. *J* 

high-resolution two-dimensional polyacrylamide gel electrophoresis. *J Neurol Sci.* 

between ubiquitin-conjugating enzymes and the 26S proteasome. *Mol Cell Biol.* Jul

for ischemic and traumatic brain injury in rats. *Eur J Neurosci.* Feb 2010;31(4):722-

Hydrolase-L1 (Uch-L1) in Severe Traumatic Brain Injury Patient Biofluids. *J* 

moderate traumatic brain injury from trauma controls and is associated with lesions on computed tomography [abstract]. *J Neurotrauma.* 2011;28(July):A1-A134.

JAN, Duchen LW, eds. *Greenfield's Neuropathology*. London: Edward Arnold;

acidic protein and S-100B in hippocampus and cortex of diabetic rats: effects of

acidic protein within degenerating astrocytes of the Alzheimer's disease brain. *Am J* 

specific proteins after acute stroke: A comparative analysis of serum concentrations

of protein S-100B and glial fibrillary acidic protein. *Stroke.* Nov 2000;31(11):2670- 2677.


**Part 2** 

**Proteomic Analysis of Protein Functions** 


## **Part 2**

### **Proteomic Analysis of Protein Functions**

106 Proteomics – Human Diseases and Protein Functions

[143] Sapsford W. Penetrating brain injury in military conflict: does it merit more research? *J* 

[144] Okie S. Traumatic brain injury in the war zone. *N Engl J Med.* May 19

[145] Warden D. Blast Injury. http://www.dvbic.org/cms.php?p=Blast\_injury]. Accessed

*R Army Med Corps.* Mar 2003;149(1):5-14.

2005;352(20):2043-2047.

April 9, 2008.

**6** 

**Comparative Proteomics:** 

Kenneth Ka Ho Lee1,2 et al.\*

*1Hong Kong 2China* 

**An Approach to Elucidating the** 

*Ministry of Education, JiNan University, Guangzhou* 

**Function of a Novel Gene Called BRE** 

*1Stem Cell and Regeneration Thematic Research Programme, School of Biomedical Sciences, Chinese University of Hong Kong, Shatin, N.T. 2Joint JUN-CUHK Key Laboratories for Regenerative Medicine,* 

Proteomics was developed in the early 1990s to allow proteins expressed by cells and tissues to be systematically studied (Celis et al., 1999; Arrell et al., 2001). The word proteome was coined by Marc Wilkins et al (Wilkins et al, 1996) from the words "protein and genome". It is therefore defined as protein equivalent of the genome. Generally, unique spectrum of proteins is only synthesized by specific cell types, for example amylase is secreted by the parotid gland, insulin by the pancreas and thyroxin by thyroid follicles. Protein synthesis is a complicated process formed by the different combination and length of the 20 unique amino acids found in our body (Arnstein, 1965). For example, following the transcription of genes encoded in the DNA, the mRNAs translocate into the cytoplasm where they are translated into a specific type of protein by the ribosomes (Lengyel, 1966). This is then followed by post-translational modification of the peptide chain to configure the protein so that it becomes biologically active. Post-translational modifications of proteins involve glycosylation, alkylation, methylation and sulfation (Blundell et al., 1993, Fleischer, 1983). The co- and post-translational modifications allow the protein to be transported and secreted during cellular homeostasis (Finnerty et al., 1979; Mao et al., 2011). In this chapter, we have described the comparative 2-dimensional electrophoresis (2-DE) proteomics workflow for protein identification by mass spectometry. Comparative proteomics was used

*1Stem Cell and Regeneration Thematic Research Programme, School of Biomedical Sciences, Chinese University of* 

*2Joint JUN-CUHK Key Laboratories for Regenerative Medicine, Ministry of Education, JiNan University,* 

*3Department of Chemical Pathology, Sir Y.K. Pao Centre for Cancer, Prince of Wales Hospital, Chinese* 

Mei KuenTang1, John Yeuk-Hon Chan2, Yiu Loon Chui3, Elve Chen1, Yao Yao1,

Olivia Miu Yung Ngan4 and Henry Siu Sum Lee1

*University of Hong Kong, Shatin, N.T., Hong Kong* 

*4Department of Biology, University of Michigan, Ann Arbor, USA* 

*Hong Kong, Shatin, N.T., Hong Kong* 

*Guangzhou, China* 

**1. Introduction** 

 \*

### **Comparative Proteomics: An Approach to Elucidating the Function of a Novel Gene Called BRE**

Kenneth Ka Ho Lee1,2 et al.\*

*1Stem Cell and Regeneration Thematic Research Programme, School of Biomedical Sciences, Chinese University of Hong Kong, Shatin, N.T. 2Joint JUN-CUHK Key Laboratories for Regenerative Medicine, Ministry of Education, JiNan University, Guangzhou 1Hong Kong 2China* 

#### **1. Introduction**

Proteomics was developed in the early 1990s to allow proteins expressed by cells and tissues to be systematically studied (Celis et al., 1999; Arrell et al., 2001). The word proteome was coined by Marc Wilkins et al (Wilkins et al, 1996) from the words "protein and genome". It is therefore defined as protein equivalent of the genome. Generally, unique spectrum of proteins is only synthesized by specific cell types, for example amylase is secreted by the parotid gland, insulin by the pancreas and thyroxin by thyroid follicles. Protein synthesis is a complicated process formed by the different combination and length of the 20 unique amino acids found in our body (Arnstein, 1965). For example, following the transcription of genes encoded in the DNA, the mRNAs translocate into the cytoplasm where they are translated into a specific type of protein by the ribosomes (Lengyel, 1966). This is then followed by post-translational modification of the peptide chain to configure the protein so that it becomes biologically active. Post-translational modifications of proteins involve glycosylation, alkylation, methylation and sulfation (Blundell et al., 1993, Fleischer, 1983). The co- and post-translational modifications allow the protein to be transported and secreted during cellular homeostasis (Finnerty et al., 1979; Mao et al., 2011). In this chapter, we have described the comparative 2-dimensional electrophoresis (2-DE) proteomics workflow for protein identification by mass spectometry. Comparative proteomics was used

Olivia Miu Yung Ngan4 and Henry Siu Sum Lee1

<sup>\*</sup> Mei KuenTang1, John Yeuk-Hon Chan2, Yiu Loon Chui3, Elve Chen1, Yao Yao1,

*<sup>1</sup>Stem Cell and Regeneration Thematic Research Programme, School of Biomedical Sciences, Chinese University of Hong Kong, Shatin, N.T., Hong Kong* 

*<sup>2</sup>Joint JUN-CUHK Key Laboratories for Regenerative Medicine, Ministry of Education, JiNan University, Guangzhou, China* 

*<sup>3</sup>Department of Chemical Pathology, Sir Y.K. Pao Centre for Cancer, Prince of Wales Hospital, Chinese University of Hong Kong, Shatin, N.T., Hong Kong* 

*<sup>4</sup>Department of Biology, University of Michigan, Ann Arbor, USA* 

Comparative Proteomics:

protein identification.

et al., 1988; Görg et al., 2009).

**2.2 New era in studying the protein profile** 

An Approach to Elucidating the Function of a Novel Gene Called BRE 111

now even easier to conduct proteomic studies. Protein databases are essential tools that allow the matching and identification of peptides from peak spectrums obtained from MS studies. In particular, the Protein Prospector (Chalkley et al., 2005) and Mascot (Perkins et al., 1999) databases are user-friendly and contain many years of interpreted MS data for

Protein chemistry has now shifted to studying the proteome which permits a better understanding of interaction between cells, hormones with cells and bioactive molecules with cells. Profiling of protein mixtures is still difficult, despite recent development in using a partial enzyme digestion strategy and advancement in instrumentation - such as electrospray ionization tandem (triple quadrupole) and mass spectrometry (ESI-MS/MS) (Ceglarek et al. 2009), quadrupole ion trap MS (Schwartz and Jardine, 1996) and Matrix-assisted laser desorption/ionization-time of flight mass spectrophotometer (Maldi-TOF MS) (Hillenkamp et al., 1991; Andersen et al., 1996). Studying the proteome also depends on the use of two dimensional electrophoresis (2-DE) (O'Farrell, 1975). This technique allows complex mixture of proteins found in cells to be separated into individual protein spots by isoelectrical focusing (IEF) and sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Proteinase inhibitors are always added to protein lysates freshly prepared from cells or tissues to prevent protein degradation. Contaminants such as phospholipids, nucleic acid and ionic molecules are also present and can be removed by gel filtration, dialysis and protein precipitation. Although O'Farell improved the IEF procedure, he used non-equilibrium pH gradient electrophoresis which cannot be reproducible from batch to batch - as the pH gradient is difficult to maintain during IEF. However, Bjellqvist et al. (1982) developed the immobilized pH gradients (IPG) method which replaced the use of the carrier-ampholyte. Development of the IPG strip was a milestone in proteomics and is now widely used in resolving individual proteins from complex protein mixtures (Weiss and Görg, 2009). In the IPG strip, proteins migrate under a high electrical field (up to 5000V) but always stop at same pI point. If several protein spots co- exist within the same pI, then a wider range of IPG strip could be flexibly used. SDS-PAGE is used to separate the protein spots according to their molecular weight. The limitation with this method is that it can only resolve proteins ranging from 120 kDa to 10 kDa. The protein spots resolved in the 2-DE gel need to be stained before it can be analyzed. Gels are most commonly stained with coomassie blue because it is inexpensive, and compatible for MS analysis. However, the sensitivity of this staining method is limited and cannot stain-up protein spots lower than 30 g. Silver staining is also another method widely used for revealing the resolved protein spots in the gel and only need 1 g of protein. Fluorescent dyes (CyDyes) have now been developed to label protein samples for Difference Gel Electrophoresis (DIGE). The DIGE technique is very sensitive, with protein detection range down to 125 pg per spot, giving it high precision in terms of protein quantification and use in comparative proteomics (Conrotto and Souchelnytskyi, 2008; Larbi and Jefferies, 2009). 2- DE/MS is now a well-established technique for large-scale protein expression studies. However, there are drawbacks with the method which hold it back from being developed for clinical diagnosis. Drawbacks such as the high abundance of plasma and albumin present in biofluids which interfere with detection of lower abundant proteins. Resolving hydrophobic, very acidic and basic proteins is also a major deficiency with the 2-DE/MS technique (Altland

to identify proteins that were differentially expressed in the tissues after treatment with various small molecules and siRNAs.

#### **2. Proteomics research and applications**

Protein properties are diverse and complex. They are dynamically influenced by physiological change in their environment, such as hormones, factors present in inflammatory response and enzymes activated by the presence of drugs. Proteomics is founded on three basic procedures: (1) the isolation and separation of proteins from cells and tissues, (2) the identification of the proteins by mass spectrometry and (3) the resolution of analyzed protein peptides by bioinformatics. Advancement in proteomic technologies has allowed researchers to investigate the proteome of many diverse biological systems – allowing breakthroughs to be made in biomedical and biological sciences. Proteomics has also enabled the identification of important biomarkers of many human diseases and allows the discovery of novel targets for drugs. In this section, we will to discuss how proteomic technologies have been applied in biomedical sciences research and the limitations encountered.

#### **2.1 History of protein research**

Swedish biochemist Pehr Victor Edman first developed the technique called Edman Degradation which allowed the amino acid sequence in peptides to be elucidated (Edman, 1950). Determination of the protein structure could be performed under micro scale. Pehr Victor Edman also developed an instrument, the protein sequenator, which allowed the amino acids sequence to be determined following Edman degradation reaction (Edman and Begg, 1967). This sequenator was commercialized by the company Beckman. The discovery popularized the studying of protein chemistry. However, there are several disadvantages associated with this method. Firstly, the technique can only accurately determine amino acid sequences up to 50-60 residuals after using Edman reagent, phenyl isothiocyanate for degradation. Secondly, the peptide N-terminal, with NH2-group, has to react with the Edman reagent. Thirdly, sequencing can only work on a single pure peptide and not a protein mixture. Finally, only the primary peptide structure can be determined but not information on the secondary structure, such as the position of disulfide bridge. Nevertheless, it has the advantage that only small quantity (10-100 pico-moles) of peptide is needed for the Edman reaction and can be performed directly from PVDF membranes. For its time, it was a pioneering and sophisticated method for studying protein chemistry, allowing the important amino acid sequence of hormones to be discovered (Niall et al., 1969 and Birr and Frank, 1975).

In the early 1970s, mass spectrometry was used to try and resolve all the peptide sequences derived from a protein mixture (Lucas et al., 1969; Morris et al, 1971). This early work has now developed leaps and bounds and protein mixtures can routinely be analyzed by computer aided high resolution mass spectrometry (MS). Consequently, John Fenn was awarded the 2002 Nobel Prize for his work in developing the electrospray ionization for mass spectrometry which provided a new platform for protein research (Fenn et al., 1989, 2002). The electrospray ionization mass spectrometer can rapidly, accurately and sensitively analyze peptide sequences from recombinant proteins, large biomolecules, protein mixture and body fluids (Chowdhury et al., 1990; Andersen et al., 1996; Bergquist et al., 2002). The parallel development of protein databases, search engines and new softwares has made it now even easier to conduct proteomic studies. Protein databases are essential tools that allow the matching and identification of peptides from peak spectrums obtained from MS studies. In particular, the Protein Prospector (Chalkley et al., 2005) and Mascot (Perkins et al., 1999) databases are user-friendly and contain many years of interpreted MS data for protein identification.

#### **2.2 New era in studying the protein profile**

110 Proteomics – Human Diseases and Protein Functions

to identify proteins that were differentially expressed in the tissues after treatment with

Protein properties are diverse and complex. They are dynamically influenced by physiological change in their environment, such as hormones, factors present in inflammatory response and enzymes activated by the presence of drugs. Proteomics is founded on three basic procedures: (1) the isolation and separation of proteins from cells and tissues, (2) the identification of the proteins by mass spectrometry and (3) the resolution of analyzed protein peptides by bioinformatics. Advancement in proteomic technologies has allowed researchers to investigate the proteome of many diverse biological systems – allowing breakthroughs to be made in biomedical and biological sciences. Proteomics has also enabled the identification of important biomarkers of many human diseases and allows the discovery of novel targets for drugs. In this section, we will to discuss how proteomic technologies have been applied in biomedical sciences research and the limitations

Swedish biochemist Pehr Victor Edman first developed the technique called Edman Degradation which allowed the amino acid sequence in peptides to be elucidated (Edman, 1950). Determination of the protein structure could be performed under micro scale. Pehr Victor Edman also developed an instrument, the protein sequenator, which allowed the amino acids sequence to be determined following Edman degradation reaction (Edman and Begg, 1967). This sequenator was commercialized by the company Beckman. The discovery popularized the studying of protein chemistry. However, there are several disadvantages associated with this method. Firstly, the technique can only accurately determine amino acid sequences up to 50-60 residuals after using Edman reagent, phenyl isothiocyanate for degradation. Secondly, the peptide N-terminal, with NH2-group, has to react with the Edman reagent. Thirdly, sequencing can only work on a single pure peptide and not a protein mixture. Finally, only the primary peptide structure can be determined but not information on the secondary structure, such as the position of disulfide bridge. Nevertheless, it has the advantage that only small quantity (10-100 pico-moles) of peptide is needed for the Edman reaction and can be performed directly from PVDF membranes. For its time, it was a pioneering and sophisticated method for studying protein chemistry, allowing the important amino acid sequence of hormones to be discovered (Niall et al., 1969

In the early 1970s, mass spectrometry was used to try and resolve all the peptide sequences derived from a protein mixture (Lucas et al., 1969; Morris et al, 1971). This early work has now developed leaps and bounds and protein mixtures can routinely be analyzed by computer aided high resolution mass spectrometry (MS). Consequently, John Fenn was awarded the 2002 Nobel Prize for his work in developing the electrospray ionization for mass spectrometry which provided a new platform for protein research (Fenn et al., 1989, 2002). The electrospray ionization mass spectrometer can rapidly, accurately and sensitively analyze peptide sequences from recombinant proteins, large biomolecules, protein mixture and body fluids (Chowdhury et al., 1990; Andersen et al., 1996; Bergquist et al., 2002). The parallel development of protein databases, search engines and new softwares has made it

various small molecules and siRNAs.

encountered.

**2.1 History of protein research** 

and Birr and Frank, 1975).

**2. Proteomics research and applications** 

Protein chemistry has now shifted to studying the proteome which permits a better understanding of interaction between cells, hormones with cells and bioactive molecules with cells. Profiling of protein mixtures is still difficult, despite recent development in using a partial enzyme digestion strategy and advancement in instrumentation - such as electrospray ionization tandem (triple quadrupole) and mass spectrometry (ESI-MS/MS) (Ceglarek et al. 2009), quadrupole ion trap MS (Schwartz and Jardine, 1996) and Matrix-assisted laser desorption/ionization-time of flight mass spectrophotometer (Maldi-TOF MS) (Hillenkamp et al., 1991; Andersen et al., 1996). Studying the proteome also depends on the use of two dimensional electrophoresis (2-DE) (O'Farrell, 1975). This technique allows complex mixture of proteins found in cells to be separated into individual protein spots by isoelectrical focusing (IEF) and sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Proteinase inhibitors are always added to protein lysates freshly prepared from cells or tissues to prevent protein degradation. Contaminants such as phospholipids, nucleic acid and ionic molecules are also present and can be removed by gel filtration, dialysis and protein precipitation. Although O'Farell improved the IEF procedure, he used non-equilibrium pH gradient electrophoresis which cannot be reproducible from batch to batch - as the pH gradient is difficult to maintain during IEF. However, Bjellqvist et al. (1982) developed the immobilized pH gradients (IPG) method which replaced the use of the carrier-ampholyte. Development of the IPG strip was a milestone in proteomics and is now widely used in resolving individual proteins from complex protein mixtures (Weiss and Görg, 2009). In the IPG strip, proteins migrate under a high electrical field (up to 5000V) but always stop at same pI point. If several protein spots co- exist within the same pI, then a wider range of IPG strip could be flexibly used. SDS-PAGE is used to separate the protein spots according to their molecular weight. The limitation with this method is that it can only resolve proteins ranging from 120 kDa to 10 kDa. The protein spots resolved in the 2-DE gel need to be stained before it can be analyzed. Gels are most commonly stained with coomassie blue because it is inexpensive, and compatible for MS analysis. However, the sensitivity of this staining method is limited and cannot stain-up protein spots lower than 30 g. Silver staining is also another method widely used for revealing the resolved protein spots in the gel and only need 1 g of protein. Fluorescent dyes (CyDyes) have now been developed to label protein samples for Difference Gel Electrophoresis (DIGE). The DIGE technique is very sensitive, with protein detection range down to 125 pg per spot, giving it high precision in terms of protein quantification and use in comparative proteomics (Conrotto and Souchelnytskyi, 2008; Larbi and Jefferies, 2009). 2- DE/MS is now a well-established technique for large-scale protein expression studies. However, there are drawbacks with the method which hold it back from being developed for clinical diagnosis. Drawbacks such as the high abundance of plasma and albumin present in biofluids which interfere with detection of lower abundant proteins. Resolving hydrophobic, very acidic and basic proteins is also a major deficiency with the 2-DE/MS technique (Altland et al., 1988; Görg et al., 2009).

Comparative Proteomics:

An Approach to Elucidating the Function of a Novel Gene Called BRE 113

Fig. 1. The principle and workflow involved in comparative proteomics.

gene is highly Expressed in Brain and Reproductive organs and that is why we named it, BRE. The gene is down-regulated after treatment of cells with DNA damaging agents such as ultraviolet light (UV), 4-nitroquinoline-1-oxide and all-trans retinoic acid (Li et al., 1995). The BRE gene encodes a 1.7-1.9 kb mRNA which give rise to a protein with 383 amino acid residues and a molecular weight of 44 kDa. Using the yeast two-hybrid assay, it was reported that BRE interacts with the juxtamembrane (JM) region of p55-TNFR, but has no affinity for the p75-TNFR, Fas or p75-NGFR of the TNFR family (Gu et al., 1998). Meanwhile, over-expression of BRE in the human 293 embryonic kidney cells that was treated with TNF-α could inhibit the activation of the transcriptional factor NF-B (Gu et al., 1998). Since NF-B is known to induce the survival pathway associated with TNF receptor, it is likely that BRE can modulate the cell death process. The expression of the BRE gene has been investigated in various biological models including adrenal glands (Miao et al., 2001), testis (Miao et al., 2005) and hepatocellular carcinoma cells (Chan et al., 2008), but the function of BRE has still not been clarified - the protein structure of BRE do not have identifiable functional domain. It has been suggested the BRE contained 2 ubiquitinconjugating enzyme family-like regions (Hu et al., 2011). However, these regions lacked the critical Cys residues required for ubiquitination but retain the ability to bind ubiquitin. The multifunctional nature of BRE and the lack of positive identifiable functional domains on BRE, make it an ideal candidate for study using proteomics. We therefore used comparative proteomics to examine the function of this novel gene in different cell types and also in vivo.

### **3. Breakthroughs in proteomics**

Proteins are separated according to their isoelectrical points and molecular weights by 2-DE. In addition, their m/s ratio and peptide sequences in MS can resolve up to 2,000-4,000 single protein spots at a time (Görg et al., 2004). Moreover, proteins that cause abbreviated changes in normal tissues may be identified and use as potential biomarkers in medical diagnosis. This is especially important in oncology where early detection of the cancer could be properly treated and not metastasize. In the last decade, advancement in 2-DE, mass spectrometry and bioinformatics has allowed potential cancer biomarkers to be identified in serum and biofluids in the blood (Voss, et al., 2001; Gioia et al., 2011), colon (McKerrow et al., 2000), breast (Sauter et al., 2002; Lau et al., 2007; Galvão et al., 2011), ovaries (Zhang et al., 2004; Tung et al., 2008) and prostate (Ornstein et al., 2004; Ornstein and Tyson, 2006). However, the 2-DE technique still has its limitation – where proteins with extreme isoelectric point and molecular mass are not resolvable and identified. Also, it is very difficult to resolve membrane proteins and non-water soluble proteins by 2-DE. Another approach is to use non-gel based proteomic techniques (for example, ionic exchange affinity, reverse-phase and liquid chromatography) followed by MS/MS provide a novel platform for identifying proteins and therefore it can resolve the disadvantage of 2-DE technique. Now, the development of laser capture micro-dissection and MALDI-MS has allowed proteomics to be performed on a specific cell population isolated from heterogeneous tissues (Marko-Varga, 2003). It is possible to surgically isolate cancer tissue from normal tissues in histological sections of biopsies for proteomic analysis. This will accelerate the discovery of cancer biomarkers as the laser capture micro-dissection will remove "background noise" generated by normal tissues.

#### **4. Comparative proteomics**

Comparative Proteomics is the identification of the differentially expressed proteins from comparison of two or more 2-DE protein profiles, for example, isolated from cells that were treated and untreated with a drug. This method allows proteins that are differentially expressed to be identified and quantified. It is a very powerful technique for identifying the molecular targets of drugs and understanding the function of novel genes. The comparative proteomic technique is schematically summarized in Figure 1. Basically, it involves image analysis of 2-DE by matching different sets of gels together; identifying and isolating of proteins which are differentially expressed; mass spectrometry and bioinformatics. The proteome of a wide variety of biological systems can be investigated that includes cells, tissues, organs, fractionated cell lysates, and immuno-precipitated cell lysates. Since the technique only requires micrograms of materials to create a complex protein profile, the proteomes of bacteria, yeast and insect have also been investigated (Chen and Snyder, 2010; Han et al., 2011; Novak et al., 2011; Sirot et al., 2011).

#### **5. As an example of the usefulness of comparative proteomics in identifying gene function, a gene called BRE which has anti-apoptotic properties, was analyzed**

We have been interested in genes that are responsive to DNA damage (Li et al., 1995; Dong et al., 2003), and identified a novel human gene which we named BRE in this context. The

Proteins are separated according to their isoelectrical points and molecular weights by 2-DE. In addition, their m/s ratio and peptide sequences in MS can resolve up to 2,000-4,000 single protein spots at a time (Görg et al., 2004). Moreover, proteins that cause abbreviated changes in normal tissues may be identified and use as potential biomarkers in medical diagnosis. This is especially important in oncology where early detection of the cancer could be properly treated and not metastasize. In the last decade, advancement in 2-DE, mass spectrometry and bioinformatics has allowed potential cancer biomarkers to be identified in serum and biofluids in the blood (Voss, et al., 2001; Gioia et al., 2011), colon (McKerrow et al., 2000), breast (Sauter et al., 2002; Lau et al., 2007; Galvão et al., 2011), ovaries (Zhang et al., 2004; Tung et al., 2008) and prostate (Ornstein et al., 2004; Ornstein and Tyson, 2006). However, the 2-DE technique still has its limitation – where proteins with extreme isoelectric point and molecular mass are not resolvable and identified. Also, it is very difficult to resolve membrane proteins and non-water soluble proteins by 2-DE. Another approach is to use non-gel based proteomic techniques (for example, ionic exchange affinity, reverse-phase and liquid chromatography) followed by MS/MS provide a novel platform for identifying proteins and therefore it can resolve the disadvantage of 2-DE technique. Now, the development of laser capture micro-dissection and MALDI-MS has allowed proteomics to be performed on a specific cell population isolated from heterogeneous tissues (Marko-Varga, 2003). It is possible to surgically isolate cancer tissue from normal tissues in histological sections of biopsies for proteomic analysis. This will accelerate the discovery of cancer biomarkers as the laser capture micro-dissection will remove

Comparative Proteomics is the identification of the differentially expressed proteins from comparison of two or more 2-DE protein profiles, for example, isolated from cells that were treated and untreated with a drug. This method allows proteins that are differentially expressed to be identified and quantified. It is a very powerful technique for identifying the molecular targets of drugs and understanding the function of novel genes. The comparative proteomic technique is schematically summarized in Figure 1. Basically, it involves image analysis of 2-DE by matching different sets of gels together; identifying and isolating of proteins which are differentially expressed; mass spectrometry and bioinformatics. The proteome of a wide variety of biological systems can be investigated that includes cells, tissues, organs, fractionated cell lysates, and immuno-precipitated cell lysates. Since the technique only requires micrograms of materials to create a complex protein profile, the proteomes of bacteria, yeast and insect have also been investigated (Chen and Snyder, 2010;

**5. As an example of the usefulness of comparative proteomics in identifying gene function, a gene called BRE which has anti-apoptotic properties, was** 

We have been interested in genes that are responsive to DNA damage (Li et al., 1995; Dong et al., 2003), and identified a novel human gene which we named BRE in this context. The

**3. Breakthroughs in proteomics** 

"background noise" generated by normal tissues.

Han et al., 2011; Novak et al., 2011; Sirot et al., 2011).

**4. Comparative proteomics** 

**analyzed** 

Fig. 1. The principle and workflow involved in comparative proteomics.

gene is highly Expressed in Brain and Reproductive organs and that is why we named it, BRE. The gene is down-regulated after treatment of cells with DNA damaging agents such as ultraviolet light (UV), 4-nitroquinoline-1-oxide and all-trans retinoic acid (Li et al., 1995). The BRE gene encodes a 1.7-1.9 kb mRNA which give rise to a protein with 383 amino acid residues and a molecular weight of 44 kDa. Using the yeast two-hybrid assay, it was reported that BRE interacts with the juxtamembrane (JM) region of p55-TNFR, but has no affinity for the p75-TNFR, Fas or p75-NGFR of the TNFR family (Gu et al., 1998). Meanwhile, over-expression of BRE in the human 293 embryonic kidney cells that was treated with TNF-α could inhibit the activation of the transcriptional factor NF-B (Gu et al., 1998). Since NF-B is known to induce the survival pathway associated with TNF receptor, it is likely that BRE can modulate the cell death process. The expression of the BRE gene has been investigated in various biological models including adrenal glands (Miao et al., 2001), testis (Miao et al., 2005) and hepatocellular carcinoma cells (Chan et al., 2008), but the function of BRE has still not been clarified - the protein structure of BRE do not have identifiable functional domain. It has been suggested the BRE contained 2 ubiquitinconjugating enzyme family-like regions (Hu et al., 2011). However, these regions lacked the critical Cys residues required for ubiquitination but retain the ability to bind ubiquitin. The multifunctional nature of BRE and the lack of positive identifiable functional domains on BRE, make it an ideal candidate for study using proteomics. We therefore used comparative proteomics to examine the function of this novel gene in different cell types and also in vivo.

Comparative Proteomics:

An Approach to Elucidating the Function of a Novel Gene Called BRE 115

Promega Corporation, United States), 1 μl of forward primer, 1 μl of reverse primer, 0.25 μl of Taq polymerase (Bio-firm, Hong Kong) and DEPC-treated water in a PCR microcentrifuge tube was placed into the thermal cycler for PCR amplification. All of the primers used in this study were manufactured and desalted by Invitrogen Corporation. The primers' sequences and the annealing temperature and duration shown in Table 1 were designed with Primer3

**Primers Sequences Annealing temp &** 

Reverse: 5'-TTCATGAGGTAGTCTGTCAGGTCC-3'

forward: 5'-TGAGACCTTCAACACCCCAG-3' reverse: 5'-TTCATGAGGTAGTCT GTCAGGTCC-3'

Reverse: 5'-TTCATGAGGTAGTCTGTCA-3'

Reverse: 5'-AATGACCTGGTCCTCCTAG-3'

Reverse: 5'-GTTACCCTCAGTGTCTTGGA-

Reverse: 5'-CAGTCTGCATAGGCACTTG-3'

Reverse: 5'-CTGGAGTCTTCCAGTGTGAT-3'

Reverse: 5'-CTCCATGTCGTCCCAGTTG GT-3'

Reverse: 5'-CACGTACTGCACCTTGTTGG-3'

Reverse: 5'- GGTAGGTGATGTTCCGAGAG-3'

Reverse: 5'- TAGGTCTGGTGAAGGTCCAT-3'

Reverse 5'-GACAGCTTCCCTGGTTAGTA-3'

Reverse: 5'-GTCTTGCAGCAGATCTCATC-3'

Reverse: 5'-TCCTTCCCGTACTTCTCCTT-3'

Reverse: 5'-GCTGCCATCCTGAGAGATAA -3'

Reverse: 5'-CTCCTCCCAAGTGGGTATCT-3'

Reverse: 5'-CTAAGTTGTTGCACCTCTCC-3'

Reverse: 5'-TACACACGGTGTTCTGTTTCTCC -3'

Reverse: 5'-AGGAGAACTGAGGGAAGTGT-3'

Forward: 5'-CCACATTCCCACATACCTTCTC-3' Reverse: 5'—GCCATTTCATTTCCATCCCATC-3'

*mouse β-actin* Forward: 5'-TGAGACCTTCAACACCCCAG-3' and

*mouse BRE* Forward: 5'-CTAGTCGCCGGTTACTGA-3'

*mouse Mdm4* Forward: 5'-CTCCAAGCAAGAGGTACTG-3'

*Mouse Akt-3* Forward: 5'- CTGGCACCAGAGGTATTAGA-3'

*Mouse 26S Proteasome* Forward: 5'-TGATCTGTAACCTGGCCTAC-3'

*mouse Prohibitin* Forward: 5'-TGAGTGATGACCTCACAGA-3

*mouse p53* Forward:5'-ACTCTCCTCCCCTCAATAAG-3'

*human BRE* Forward: 5'-ATCTTGCCTCCTGGAATCCT-3'

*human cyclin A* Forward: 5'-TCCTGTCTTCCATGTCAGTG-3'

*human p53* Forward: 5'-GCCTGACTCAGACTGACATT-3'

*mouse TUSC4* Forward: 5'-CTGGTATCC ATCCTCCAGTA-3'

*mouse DPF2* Forward: 5'-TCCTTGGCGAGC AATACTAC-3'

*mouse HSPA7* Forward: 5'-GCAGTCGGATATGAAGCACT-3'

*mouse HSPA2* Forward: 5'-GACGAATGTCAGGAGGTGAT-3'

Table 1. Primers used in the study.

*mouse ENO1* Forward: 5'-CTACGAGGCCCTCTAAGAACTCC-3'

*human β-actin* Forward: 5'-ATGGATGATGATATCGCCGCG-3'

*human Prohibitin* Forward: 5'- CGGAG AGGACTATGATGAGC-3'

*Human TNF-R1* Forward: 5'- ACCAAGTGCCACAAAGGAACC -3'

or

or

**duration** 

59 oC , 45s 55 oC , 60s

56 oC, 45s

55 oC, 60s

54 oC, 60s

56 oC, 60s

57 oC, 60s

54 oC, 45s

54 oC, 60s

55 oC, 45s

57 oC, 60s

57 oC, 60s

57 oC,60s

56 oC, 60s

54 oC, 60s

53 oC, 60s

58 oC, 60s

53 oC, 60s

58 oC, 60s

58 oC, 60s

#### **5.1 Materials and methods**

#### **5.1.1 Tissue cultures**

All of the cell cultures were maintained at 37oC and 5% CO2 in a humidified cultured chamber. C2C12 myoblasts (ATCC) and D122 Lewis lung carcinoma cells (gift from Lea Eisenbach) were cultured in DMEM medium supplemented with 10% FBS and penicillin/streptomycin. Two stably transfected cell lines were produced from D122 using a pcDNA3.1 expression vector. D122v3B harbor the empty vector, while D122a4 cells overexpress the full length *BRE* (Chan et al., 2005). D122v3B and D122a4 were maintained in DMEM plus 10% FBS and 400 mg/mL of G418 (Invitrogen), Immortalized human esophageal epithelial (SHEE) cell line and the malignantly transformed esophageal carcinoma cell line (SHEEC) were cultured in DMEM medium plus F-12 Nutrient Mixture (1:1) supplemented with 10% FBS (GibcoBRL) and penicillin/streptomycin (Shen et al., 2000). Chang cells (ATCC, CCL-13) were cultured in Minimum Essential Medium Eagle plus 10% FBS.

#### **5.1.2 Transgenic mice**

The transgenic mice were generated carrying the full-length BRE gene and the transthyretin (TTR) promoter. The TTR promoter is specifically expressed in hepatocytes in the liver (Ching et al, 2001). All mice were maintained in the Laboratory Animal Services Centre, Chinese University of Hong Kong. Ethical approval has been obtained from the animal ethics committee, Chinese University of Hong Kong before performing the animal experiments.

#### **5.1.3 Subcellular fractioning of soluble proteins**

SHEE and SHEEC cells were extracted in lysis buffer (8M Urea, 2M Thiourea, 2% CHAPS, 0.01% TBP, 0.01% NP-40) containing protease inhibitors (GE Healthcare). After extraction, the lysates were incubated on ice for 30 min and then centrifuged at 8000 rpm for 15 min to remove all cell debris. The fractions (cytosol, membrane, and nucleoplasm) were obtained using a ProteoExtract Subcellular Proteome Extraction Kit (Calbiochem) following instructions provided by the manufacturer. The total protein concentration for each fraction was determined using a Bio-Rad Protein Assay kit (Bio-Rad, Richmond).

#### **5.1.4 BRE gene silencing analysis**

Two BRE-specific siRNAs were designed corresponding to 5'- TCTGGCTGCACATCATTGA-3' (nucleotides 124–142, nucleotide position number 1 being the start of the initiation codon), and 5'-CTGGACTGGTGAATTTTCA-3' (nucleotides 491– 509). siRNA sequence 5'-AAGCCUCGAAAUAUCUCCU-dTT-3' with no known mRNA targets was used as a control.

#### **5.1.5 Semi-quantitative RT-PCR analysis**

The total RNA was isolated and purified by using TRIzol solution (Invitrogen Corporation, United States). 1 µg of the total RNA was used for reverse-transcription to synthesize the complementary DNA (cDNA) according to the procedures of ImProm-II™ Reverse Transcription System. cDNA was used as the template for PCR amplification. 20 μl of PCR mixture containing 1 μl of cDNA, 2.5 μl of PCR 10X buffer (Bio-firm, Hong Kong), 0.75 μl of magnesium chloride solution (25 mM, Bio-firm, Hong Kong), 1 μl of dNTP mix (10 mM,

All of the cell cultures were maintained at 37oC and 5% CO2 in a humidified cultured chamber. C2C12 myoblasts (ATCC) and D122 Lewis lung carcinoma cells (gift from Lea Eisenbach) were cultured in DMEM medium supplemented with 10% FBS and penicillin/streptomycin. Two stably transfected cell lines were produced from D122 using a pcDNA3.1 expression vector. D122v3B harbor the empty vector, while D122a4 cells overexpress the full length *BRE* (Chan et al., 2005). D122v3B and D122a4 were maintained in DMEM plus 10% FBS and 400 mg/mL of G418 (Invitrogen), Immortalized human esophageal epithelial (SHEE) cell line and the malignantly transformed esophageal carcinoma cell line (SHEEC) were cultured in DMEM medium plus F-12 Nutrient Mixture (1:1) supplemented with 10% FBS (GibcoBRL) and penicillin/streptomycin (Shen et al., 2000). Chang cells (ATCC, CCL-13) were cultured in Minimum Essential Medium Eagle plus

The transgenic mice were generated carrying the full-length BRE gene and the transthyretin (TTR) promoter. The TTR promoter is specifically expressed in hepatocytes in the liver (Ching et al, 2001). All mice were maintained in the Laboratory Animal Services Centre, Chinese University of Hong Kong. Ethical approval has been obtained from the animal ethics committee, Chinese University of Hong Kong before performing the animal

SHEE and SHEEC cells were extracted in lysis buffer (8M Urea, 2M Thiourea, 2% CHAPS, 0.01% TBP, 0.01% NP-40) containing protease inhibitors (GE Healthcare). After extraction, the lysates were incubated on ice for 30 min and then centrifuged at 8000 rpm for 15 min to remove all cell debris. The fractions (cytosol, membrane, and nucleoplasm) were obtained using a ProteoExtract Subcellular Proteome Extraction Kit (Calbiochem) following instructions provided by the manufacturer. The total protein concentration for each fraction

Two BRE-specific siRNAs were designed corresponding to 5'- TCTGGCTGCACATCATTGA-3' (nucleotides 124–142, nucleotide position number 1 being the start of the initiation codon), and 5'-CTGGACTGGTGAATTTTCA-3' (nucleotides 491– 509). siRNA sequence 5'-AAGCCUCGAAAUAUCUCCU-dTT-3' with no known mRNA

The total RNA was isolated and purified by using TRIzol solution (Invitrogen Corporation, United States). 1 µg of the total RNA was used for reverse-transcription to synthesize the complementary DNA (cDNA) according to the procedures of ImProm-II™ Reverse Transcription System. cDNA was used as the template for PCR amplification. 20 μl of PCR mixture containing 1 μl of cDNA, 2.5 μl of PCR 10X buffer (Bio-firm, Hong Kong), 0.75 μl of magnesium chloride solution (25 mM, Bio-firm, Hong Kong), 1 μl of dNTP mix (10 mM,

was determined using a Bio-Rad Protein Assay kit (Bio-Rad, Richmond).

**5.1 Materials and methods 5.1.1 Tissue cultures** 

10% FBS.

experiments.

**5.1.2 Transgenic mice** 

**5.1.3 Subcellular fractioning of soluble proteins** 

**5.1.4 BRE gene silencing analysis** 

**5.1.5 Semi-quantitative RT-PCR analysis** 

targets was used as a control.

Promega Corporation, United States), 1 μl of forward primer, 1 μl of reverse primer, 0.25 μl of Taq polymerase (Bio-firm, Hong Kong) and DEPC-treated water in a PCR microcentrifuge tube was placed into the thermal cycler for PCR amplification. All of the primers used in this study were manufactured and desalted by Invitrogen Corporation. The primers' sequences and the annealing temperature and duration shown in Table 1 were designed with Primer3


Table 1. Primers used in the study.

Comparative Proteomics:

An Approach to Elucidating the Function of a Novel Gene Called BRE 117

States), 50 μg/ml salmon sperm DNA and 50% formamide) for 2 hrs. The samples were then added and hybridized in 0.5 μg/ml of DIG-labeled antisense riboprobe. The sense probe was used as a negative control. The hybridization temperature was 55oC and the incubation time was 16 hrs. Following hybridization, the samples were washed in 2X SSC at 42oC for 20 mins with two changes, 0.1% SDS (w/v) in 0.2X SSC buffer for 15 min and then 0.2X SSC buffer for 10 mins. The alkaline phosphatase-conjugated digoxigenin antibody (1:50, Roche Applied Science, United States) was added to the specimens for 2 hrs and then washed in DPBS for 10 min with four changes. Nitroblue tetrazolium salt and 5-bromo-4 chloro-3-indolylphosphate (NBT/BCIP, Roche Applied Science, United States) were used as the color substrates. After color development, the sections were mounted in 50% glycerol

Chang liver cells were cultured in 8-well glass slide (Nalge Nunc international, Naperville) with Minimum Essential Medium Eagle plus 10% FBS. After 80% confluent, the cultures were transfected with Ctl-siRNA or BRE-siRNA respectively according to maufacturers' instructions. Forty-eight hours after transfection, BrdU was added into the cultures to a final concentration of 20 M and incubated at 37oC for 4 hrs. The treated cultures were then fixed with 2% paraformaldehyde for 24 hr. The fixed cultures were processed for immunohistochemistry by using mouse BrdU antibodies (1:1000, Sigma-Aldrich, United States). The BrdU positive and negative cells were counted and analysed by Spot Digital

The cell lysate for the first DE was performed on an IPGphor IEF system using 11-cm long IPG electrode strip with 4-7 pH gradient (Amersham Biosciences, United Kingdom) and an Ettan IPGphor Strip Holder (Amersham Biosciences, United Kingdom). 150 μg of protein was applied for each IPG strip. The total volume of protein sample and rehydration buffer (8M Urea, 2% CHAPS (w/v), 1% IPG buffer (v/v, Amersham Biosciences, United Kingdom), 40 mM DTT loaded onto the strip holder was 210 μl. 1ml of IPG Cover Fluid (Amersham Biosciences, United Kingdom) was applied to each strip so as to minimize evaporation and urea crystallization. The rehydration step was done under voltage and followed by a separation process. The electrophoresis condition for step 1 was 30 V for 13 hrs; step 2 was 500 V for 1 hr; step 3 was 2000 V for 1 hr and step 4 was 5000 V for 20 hrs.

**5.1.10 Second dimensional separation – Sodium dodecyl sulphate polyacrylamide-gel**  After first DE was completed, the IPG strips were removed from the strip holders. Each strip was then treated with 1% DTT in 6.5 ml of equilibration buffer (50 mM Tris, 6M of urea, 30% glycerol, 2% SDS, 0.1% bromophenol blue) for 30 min. The strips were further treated with 1% iodoacetamide (IAA, w/v, Sigma-Aldrich, United States) dissolved in the 6.5 ml of the same equilibration buffer. The strips were treated in the solution for 30 min. The equilibrated strips were then loaded on the 12% SDS-acrylamide separating gels. The 2-DE was performed in an ISO-DALT apparatus (Hoefer Scientific Instruments). Prestained protein molecular weight marker (Fermentas Life Science, Canada) with the range of 20 to

Camera & Carl Zeiss Microscope Axiophot 2 Integrated Biological Imaging System.

**5.1.9 First dimensional separation of samples – Isoelectric focusing** 

The program was stopped when the total volt-hours reached 40000.

120 kDa was used to determine the sizes of the proteins on the gel.

(v/v, USB, United States). The experiment was performed in triplicates.

**5.1.8 BrdU (Bromodeoxyuridine) labeling assay** 

software (version 0.4.0, Rozen and Skaletsky; http://frodo.wi.mit.edu). The PCR mixtures were reacted in a PTC-100 thermal cycler (MJ Research, Watertown, MA, USA) set under the following amplification conditions: initial denaturation at 95°C for 2 min, followed by a total of 35 cycles of denaturation at 95°C for 1 min, annealing at different temperature according to the primer' conditions as shown in Table 1 and extension at 72°C for 1 min. An additional 7 min extension step at 72ºC was performed at the end of the last cycle. After the electrophoresis, the PCR products were analyzed on a 1.5% agarose gel with ethidium bromide staining, the intensities of the PCR products were visualized and determined using the GelDoc-It imaging system (UVP, BioImaging System, USA). β-actin was used as a house keeping gene for internal control and normalization. The experiments were repeated three times.

#### **5.1.6 Western blot analysis**

Control and treated cells were lysed in 200 μl of lysis buffer (50 mM NaCl, 20 mM Tris, pH 7.6, 1% NP-40, 1 X protease inhibitor mixture) for 60 min. The lysates were cleared by centrifugation at 16 000×g at 4 oC for 10 min. Crude protein concentration was measured by using a protein assay kit (Bio-Rad). 30 to 50 μg of total protein lysate were resolved on 10 to 12% SDS-PAGE, with Rainbow molecular weight markers and electroblotted onto Hybond NC membranes (GE Healthcare). The blots were incubated with Akt-3 (1:100, sc-11521 Santa Cruz Biotechnology), Bre (1:500 to 1000, Chan et al,. 2008), mdmX (1:100, sc-14738, Santa Cruz Biotechnology), prohibitin (1:000, sc-18196, Santa Cruz Biotechnology),p53 (1:000, sc-6243, Santa Cruz Biotechnology) or -tubulin (1:1000 to 1500, Zymed Laboratories), tubulin (1:1500, Zymed Laboratories), cyclin A (1:1000, sc-11521, Santa Cruz Biotechnology), prohibitin (1:600, sc-18196, Santa Cruz Biotechnology), TNF-R1 (1:800, sc-8436, Santa Cruz Biotechnology), CDK2 (M2) (1:800, sc-163 Santa Cruz Biotechnology). Bound antibodies were detected using the appropriate horseradish peroxidase-conjugated secondary antibodies (Southern biotechnology), followed by development with an ECL Western blotting Detection kit (GE Healthcare). The blots were analyzed using Quantity One software (Bio-Rad) and the intensity of the bands produced for each antibody was normalized against the tubulin band (internal control) produced from each sample. Three replicates of each sample were studied.

#### **5.1.7 In situ hybridization**

All of the procedures performed were according to Lee et al. (2001). The liver samples were fixed in 4% paraformaldehyde (w/v, Sigma, United States) for 24 hrs. The fixed samples were washed in Dulbecco's Phosphate Buffered Saline (DPBS, Invitrogen Corporation, United States) for 15 min with three changes. The samples were then dehydrated, cleared and embedded in paraffin wax. Finally, the specimens were sectioned at 7 μm and mounted onto TESPA treated slides. The riboprobe was prepared from pGEM-T plasmid containing 1,205 bp encoding BRE sequence. The plasmid cDNA was linearized by EcoRI and in-vitro transcribed to generate digoxigenin (DIG)-labeled sense and antisense BRE riboprobe using a DIG RNA labeling kit (Roche Applied Science, United States). After dewaxing the paraffin sections, the specimens were rehydrated and equilibrated in DPBS for 10 min. The sections were digested with 10 μg/ml of proteinase K (Fermentas Life Science, Canada) for 7 min and post-fixed in 2% paraformaldehyde for 5 min. After washing in DPBS for 10 minutes twice, the samples were incubated in pre-hybridization buffer (2X SSC, 1X Denhardt's reagent, 5mM EDTA , 0.1% sodium dodecyl sulfate, 10X Dextran sulfate (Chemicon, United

software (version 0.4.0, Rozen and Skaletsky; http://frodo.wi.mit.edu). The PCR mixtures were reacted in a PTC-100 thermal cycler (MJ Research, Watertown, MA, USA) set under the following amplification conditions: initial denaturation at 95°C for 2 min, followed by a total of 35 cycles of denaturation at 95°C for 1 min, annealing at different temperature according to the primer' conditions as shown in Table 1 and extension at 72°C for 1 min. An additional 7 min extension step at 72ºC was performed at the end of the last cycle. After the electrophoresis, the PCR products were analyzed on a 1.5% agarose gel with ethidium bromide staining, the intensities of the PCR products were visualized and determined using the GelDoc-It imaging system (UVP, BioImaging System, USA). β-actin was used as a house keeping gene for internal

Control and treated cells were lysed in 200 μl of lysis buffer (50 mM NaCl, 20 mM Tris, pH 7.6, 1% NP-40, 1 X protease inhibitor mixture) for 60 min. The lysates were cleared by centrifugation at 16 000×g at 4 oC for 10 min. Crude protein concentration was measured by using a protein assay kit (Bio-Rad). 30 to 50 μg of total protein lysate were resolved on 10 to 12% SDS-PAGE, with Rainbow molecular weight markers and electroblotted onto Hybond NC membranes (GE Healthcare). The blots were incubated with Akt-3 (1:100, sc-11521 Santa Cruz Biotechnology), Bre (1:500 to 1000, Chan et al,. 2008), mdmX (1:100, sc-14738, Santa Cruz Biotechnology), prohibitin (1:000, sc-18196, Santa Cruz Biotechnology),p53 (1:000, sc-6243, Santa Cruz Biotechnology) or -tubulin (1:1000 to 1500, Zymed Laboratories), tubulin (1:1500, Zymed Laboratories), cyclin A (1:1000, sc-11521, Santa Cruz Biotechnology), prohibitin (1:600, sc-18196, Santa Cruz Biotechnology), TNF-R1 (1:800, sc-8436, Santa Cruz Biotechnology), CDK2 (M2) (1:800, sc-163 Santa Cruz Biotechnology). Bound antibodies were detected using the appropriate horseradish peroxidase-conjugated secondary antibodies (Southern biotechnology), followed by development with an ECL Western blotting Detection kit (GE Healthcare). The blots were analyzed using Quantity One software (Bio-Rad) and the intensity of the bands produced for each antibody was normalized against the tubulin band (internal control) produced from each sample. Three

All of the procedures performed were according to Lee et al. (2001). The liver samples were fixed in 4% paraformaldehyde (w/v, Sigma, United States) for 24 hrs. The fixed samples were washed in Dulbecco's Phosphate Buffered Saline (DPBS, Invitrogen Corporation, United States) for 15 min with three changes. The samples were then dehydrated, cleared and embedded in paraffin wax. Finally, the specimens were sectioned at 7 μm and mounted onto TESPA treated slides. The riboprobe was prepared from pGEM-T plasmid containing 1,205 bp encoding BRE sequence. The plasmid cDNA was linearized by EcoRI and in-vitro transcribed to generate digoxigenin (DIG)-labeled sense and antisense BRE riboprobe using a DIG RNA labeling kit (Roche Applied Science, United States). After dewaxing the paraffin sections, the specimens were rehydrated and equilibrated in DPBS for 10 min. The sections were digested with 10 μg/ml of proteinase K (Fermentas Life Science, Canada) for 7 min and post-fixed in 2% paraformaldehyde for 5 min. After washing in DPBS for 10 minutes twice, the samples were incubated in pre-hybridization buffer (2X SSC, 1X Denhardt's reagent, 5mM EDTA , 0.1% sodium dodecyl sulfate, 10X Dextran sulfate (Chemicon, United

control and normalization. The experiments were repeated three times.

**5.1.6 Western blot analysis** 

replicates of each sample were studied.

**5.1.7 In situ hybridization** 

States), 50 μg/ml salmon sperm DNA and 50% formamide) for 2 hrs. The samples were then added and hybridized in 0.5 μg/ml of DIG-labeled antisense riboprobe. The sense probe was used as a negative control. The hybridization temperature was 55oC and the incubation time was 16 hrs. Following hybridization, the samples were washed in 2X SSC at 42oC for 20 mins with two changes, 0.1% SDS (w/v) in 0.2X SSC buffer for 15 min and then 0.2X SSC buffer for 10 mins. The alkaline phosphatase-conjugated digoxigenin antibody (1:50, Roche Applied Science, United States) was added to the specimens for 2 hrs and then washed in DPBS for 10 min with four changes. Nitroblue tetrazolium salt and 5-bromo-4 chloro-3-indolylphosphate (NBT/BCIP, Roche Applied Science, United States) were used as the color substrates. After color development, the sections were mounted in 50% glycerol (v/v, USB, United States). The experiment was performed in triplicates.

#### **5.1.8 BrdU (Bromodeoxyuridine) labeling assay**

Chang liver cells were cultured in 8-well glass slide (Nalge Nunc international, Naperville) with Minimum Essential Medium Eagle plus 10% FBS. After 80% confluent, the cultures were transfected with Ctl-siRNA or BRE-siRNA respectively according to maufacturers' instructions. Forty-eight hours after transfection, BrdU was added into the cultures to a final concentration of 20 M and incubated at 37oC for 4 hrs. The treated cultures were then fixed with 2% paraformaldehyde for 24 hr. The fixed cultures were processed for immunohistochemistry by using mouse BrdU antibodies (1:1000, Sigma-Aldrich, United States). The BrdU positive and negative cells were counted and analysed by Spot Digital Camera & Carl Zeiss Microscope Axiophot 2 Integrated Biological Imaging System.

#### **5.1.9 First dimensional separation of samples – Isoelectric focusing**

The cell lysate for the first DE was performed on an IPGphor IEF system using 11-cm long IPG electrode strip with 4-7 pH gradient (Amersham Biosciences, United Kingdom) and an Ettan IPGphor Strip Holder (Amersham Biosciences, United Kingdom). 150 μg of protein was applied for each IPG strip. The total volume of protein sample and rehydration buffer (8M Urea, 2% CHAPS (w/v), 1% IPG buffer (v/v, Amersham Biosciences, United Kingdom), 40 mM DTT loaded onto the strip holder was 210 μl. 1ml of IPG Cover Fluid (Amersham Biosciences, United Kingdom) was applied to each strip so as to minimize evaporation and urea crystallization. The rehydration step was done under voltage and followed by a separation process. The electrophoresis condition for step 1 was 30 V for 13 hrs; step 2 was 500 V for 1 hr; step 3 was 2000 V for 1 hr and step 4 was 5000 V for 20 hrs. The program was stopped when the total volt-hours reached 40000.

#### **5.1.10 Second dimensional separation – Sodium dodecyl sulphate polyacrylamide-gel**

After first DE was completed, the IPG strips were removed from the strip holders. Each strip was then treated with 1% DTT in 6.5 ml of equilibration buffer (50 mM Tris, 6M of urea, 30% glycerol, 2% SDS, 0.1% bromophenol blue) for 30 min. The strips were further treated with 1% iodoacetamide (IAA, w/v, Sigma-Aldrich, United States) dissolved in the 6.5 ml of the same equilibration buffer. The strips were treated in the solution for 30 min. The equilibrated strips were then loaded on the 12% SDS-acrylamide separating gels. The 2-DE was performed in an ISO-DALT apparatus (Hoefer Scientific Instruments). Prestained protein molecular weight marker (Fermentas Life Science, Canada) with the range of 20 to 120 kDa was used to determine the sizes of the proteins on the gel.

Comparative Proteomics:

An Approach to Elucidating the Function of a Novel Gene Called BRE 119

Fig. 2. Representative 2-DE gel of protein extracts from C2C12 cells that had been

Fig. 3. Semiquantitative RT-PCR (A) and Western blots (B) analyses confirming the comparative proteomic results that silencing *BRE*, down-regulated p*rohibitin* and *26S proteasome regulatory subunit S14* expression, while *Akt-3* expression was up-regulated. *β-*

*actin* and α-tubulin serve as internal controls (Tang et al., 2006).

and MW in kDa (y-axis) (Tang et al., 2006).

transfected with CTL- or BRE-siRNAs. Four differentially expressed proteins were identified (Swiss-Prot accession number provided). Silencing BRE expression up-regulated protein spots Q9WUA6 and P16015, but P6778 and Q9Z2X2 were down-regulated. pI 4–7 (x-axis)

#### **5.1.11 Gel to gel matching**

The gels were stained and scanned by using a GS 800 Densitometer (Bio-Rad Laboratories, United States) and images were captured for further analysis. The protein spots on the gel were analyzed by the discovery series, PDQuest 2D Analysis Software (Bio-Rad Laboratories, United States) version 7.13 PC. The experiment was performed in triplicate.

#### **5.1.12 Protein identification by mass fingerprinting**

All protein spots of interest were isolated from the gel and processed for destaining. The gel pieces were first washed in MilliQ water, immersed in 200 μl of destaining solution (15 mM potassium ferricyanide and 50 mM sodium thiosulphate) and then incubated at room temperature until they turned into colorless. Each gel piece was then washed with 400 μl of MilliQ water for 15 min, three times. The destained gel pieces were equilibrated in 200 μl of 10 mM ammonium bicarbonate/50% acetonitrile each for about 15 min. The solution was discarded and the equilibrated gel pieces were dehydrated by incubating in 200 l of acetonitrile for 15 min. The solution was then poured off and the spots were dried in an incubator at 30ºC for 5 min. Fifteen μg/ml trypsin working solution in 40 mM ammonium bicarbonate/50% acetonitrile (v/v) was used for in-gel digestion. Twelve μl of the working solution was added to each gel sample. The samples were then incubated at 35ºC for 16 hrs. After trypsinization, 3 μl of extraction solution (50% acetonitrile (v/v) and 5% trifluoroacetic acid (Fluka Chemika, Switzerland) were added to each gel piece to stop the reaction. They are then centrifuged at 3,000 rpm for 2 min at room temperature. Three μl of reaction mixture from each sample was mixed with α-cyano-4-hydroxycinnamic acid matrix and then spotted onto a sample plate (Applied Biosystems, United States) for the MALDI-TOF mass spectroscopy. The mass spectrums generated were analyzed using the software Data Explorer Version 4.0.0.0 (Applied Biosystems, United States) and by mass fingerprinting search using the search engine provided by Protein prospector (http://prospector.ucsf.edu/ucsfhtml4.0/msfit.htm). To determine the significance of variance in the experiments, data were analyzed using the two-tailed, paired student's t-test. P<0.05 was considered to be statistically significant. All statistical analysis was performed using the SPSS software.

#### **5.2 Results and discussions of the comparative proteomic analysis of BRE 5.2.1 Comparative proteomic analysis reveals BRE regulates prohibitin and p53 expression**

BRE gene encodes a highly conserved stress-modulating protein. To gain further insight into the function of this gene, we used comparative proteomics to investigate the protein profiles of C2C12 and D122 cells resulting from small interfering RNA (siRNA)-mediated silencing as well as overexpression of BRE. It was found that silencing BRE expression in C2C12 cells would up-regulate Akt-3 and carbonic anhydrase III expression. In contrast, 26S proteasome regulatory subunit S14 and prohibitin expressions were down-regulated as shown in Figures 2 (2-DE gel) and 3 (semiquantitative RT-PCR and Western blot analyses). It has been reported that prohibitin is normally expressed in different cellular compartments involved in regulating cell proliferation, mitochondrial activities and protein processing (Mishra, 2010). Prohibitin can apparently directly interact with p53 in response to stress (Fusaro et al., 2003; Joshi et al., 2007). We established that cell proliferation was significantly increased after silencing BRE expression and this was accompanied by a reduction in p53 and

The gels were stained and scanned by using a GS 800 Densitometer (Bio-Rad Laboratories, United States) and images were captured for further analysis. The protein spots on the gel were analyzed by the discovery series, PDQuest 2D Analysis Software (Bio-Rad Laboratories, United States) version 7.13 PC. The experiment was performed in triplicate.

All protein spots of interest were isolated from the gel and processed for destaining. The gel pieces were first washed in MilliQ water, immersed in 200 μl of destaining solution (15 mM potassium ferricyanide and 50 mM sodium thiosulphate) and then incubated at room temperature until they turned into colorless. Each gel piece was then washed with 400 μl of MilliQ water for 15 min, three times. The destained gel pieces were equilibrated in 200 μl of 10 mM ammonium bicarbonate/50% acetonitrile each for about 15 min. The solution was discarded and the equilibrated gel pieces were dehydrated by incubating in 200 l of acetonitrile for 15 min. The solution was then poured off and the spots were dried in an incubator at 30ºC for 5 min. Fifteen μg/ml trypsin working solution in 40 mM ammonium bicarbonate/50% acetonitrile (v/v) was used for in-gel digestion. Twelve μl of the working solution was added to each gel sample. The samples were then incubated at 35ºC for 16 hrs. After trypsinization, 3 μl of extraction solution (50% acetonitrile (v/v) and 5% trifluoroacetic acid (Fluka Chemika, Switzerland) were added to each gel piece to stop the reaction. They are then centrifuged at 3,000 rpm for 2 min at room temperature. Three μl of reaction mixture from each sample was mixed with α-cyano-4-hydroxycinnamic acid matrix and then spotted onto a sample plate (Applied Biosystems, United States) for the MALDI-TOF mass spectroscopy. The mass spectrums generated were analyzed using the software Data Explorer Version 4.0.0.0 (Applied Biosystems, United States) and by mass fingerprinting search using the search engine provided by Protein prospector (http://prospector.ucsf.edu/ucsfhtml4.0/msfit.htm). To determine the significance of variance in the experiments, data were analyzed using the two-tailed, paired student's t-test. P<0.05 was considered to be statistically significant. All statistical analysis was performed

**5.2 Results and discussions of the comparative proteomic analysis of BRE 5.2.1 Comparative proteomic analysis reveals BRE regulates prohibitin and p53** 

BRE gene encodes a highly conserved stress-modulating protein. To gain further insight into the function of this gene, we used comparative proteomics to investigate the protein profiles of C2C12 and D122 cells resulting from small interfering RNA (siRNA)-mediated silencing as well as overexpression of BRE. It was found that silencing BRE expression in C2C12 cells would up-regulate Akt-3 and carbonic anhydrase III expression. In contrast, 26S proteasome regulatory subunit S14 and prohibitin expressions were down-regulated as shown in Figures 2 (2-DE gel) and 3 (semiquantitative RT-PCR and Western blot analyses). It has been reported that prohibitin is normally expressed in different cellular compartments involved in regulating cell proliferation, mitochondrial activities and protein processing (Mishra, 2010). Prohibitin can apparently directly interact with p53 in response to stress (Fusaro et al., 2003; Joshi et al., 2007). We established that cell proliferation was significantly increased after silencing BRE expression and this was accompanied by a reduction in p53 and

**5.1.11 Gel to gel matching** 

using the SPSS software.

**expression** 

**5.1.12 Protein identification by mass fingerprinting** 

Fig. 2. Representative 2-DE gel of protein extracts from C2C12 cells that had been transfected with CTL- or BRE-siRNAs. Four differentially expressed proteins were identified (Swiss-Prot accession number provided). Silencing BRE expression up-regulated protein spots Q9WUA6 and P16015, but P6778 and Q9Z2X2 were down-regulated. pI 4–7 (x-axis) and MW in kDa (y-axis) (Tang et al., 2006).

Fig. 3. Semiquantitative RT-PCR (A) and Western blots (B) analyses confirming the comparative proteomic results that silencing *BRE*, down-regulated p*rohibitin* and *26S proteasome regulatory subunit S14* expression, while *Akt-3* expression was up-regulated. *βactin* and α-tubulin serve as internal controls (Tang et al., 2006).

Comparative Proteomics:

An Approach to Elucidating the Function of a Novel Gene Called BRE 121

one of the components of BRCA1 A complex that is essential for tumor suppression (Harris and Khanna, 2011). BRE peptide has an ubiquitin E2 variant domain which has been determined to bind ubiquitin in co-immunoprecipitation experiments (Hu et al., 2011; Li et al., 2004). Coincidently, a 26S proteasome regulatory subunit S14 was one of the proteins found to be down-regulated by BRE over-expression. It is now known that the ubiquitinproteasome pathway plays an important role in regulating the proteolytic processes that occur during signal transduction, transcriptional regulation and cell-cycle progression (Clague and Urbé, 2010). In this context, we speculate that BRE participates in the ubiquitinproteasome pathway to regulate protein turnover within cells. In the 2-DE profiling of D122α4 cells, where BRE was stably overexpressed, we identified five proteins that were upregulated. They were granulin precursor, TNF receptor associated factor 6 (TRAF6), mitogen protein kinase 8, Mdm4 and baculoviral IAP repeat-containing protein 4 as shown in Figures 6 (2 DE gel) and 7 (semiquantitative RT-PCR and Western blot analyses).

Fig. 6. Representative 2-DE gel of protein extracts from D122v3B and D122αa4 cell lines. Five protein spots (O35618, P28798, Q07174, P70196 and Q60989) were up-regulated in D122αa4

cells (Swiss-Prot accession number provided) (Tang et al., 2006).

prohibitin expression. We also identified Akt-3 that was affected by BRE silencing which suggests BRE might be involved in the P13/AKT signaling pathway (Madhunapantula et al., 2009). We observed that cell proliferation was suppressed when BRE was overexpressed in the D122a4 cell line as shown in Figure 4. This was accompanied by an increase in p53 and prohibitin expression as shown in Figure 5. It has been reported that in the nucleus BRE is

Fig. 4. MTT assay of D122, D122v3B and D122αa4 cell lines. The chart shows *BRE*  overexpression in D122αa4 inhibited cell proliferation. Values = means +SEM, *P*, ≤0.01, \* D122αa4 significantly different from D122 and D122v3B (Tang et al., 2006).

Fig. 5. Semiquantitative RT-PCR (A) and Western blot (B) showing that D122a4 cells overexpressed *prohibitin*, *p53* and *mdm4*. β*-actin* and α-tubulin serve as internal controls (Tang et al., 2006) .

prohibitin expression. We also identified Akt-3 that was affected by BRE silencing which suggests BRE might be involved in the P13/AKT signaling pathway (Madhunapantula et al., 2009). We observed that cell proliferation was suppressed when BRE was overexpressed in the D122a4 cell line as shown in Figure 4. This was accompanied by an increase in p53 and prohibitin expression as shown in Figure 5. It has been reported that in the nucleus BRE is

Fig. 4. MTT assay of D122, D122v3B and D122αa4 cell lines. The chart shows *BRE*  overexpression in D122αa4 inhibited cell proliferation. Values = means +SEM, *P*, ≤0.01, \*

Fig. 5. Semiquantitative RT-PCR (A) and Western blot (B) showing that D122a4 cells overexpressed *prohibitin*, *p53* and *mdm4*. β*-actin* and α-tubulin serve as internal controls

(Tang et al., 2006) .

D122αa4 significantly different from D122 and D122v3B (Tang et al., 2006).

one of the components of BRCA1 A complex that is essential for tumor suppression (Harris and Khanna, 2011). BRE peptide has an ubiquitin E2 variant domain which has been determined to bind ubiquitin in co-immunoprecipitation experiments (Hu et al., 2011; Li et al., 2004). Coincidently, a 26S proteasome regulatory subunit S14 was one of the proteins found to be down-regulated by BRE over-expression. It is now known that the ubiquitinproteasome pathway plays an important role in regulating the proteolytic processes that occur during signal transduction, transcriptional regulation and cell-cycle progression (Clague and Urbé, 2010). In this context, we speculate that BRE participates in the ubiquitinproteasome pathway to regulate protein turnover within cells. In the 2-DE profiling of D122α4 cells, where BRE was stably overexpressed, we identified five proteins that were upregulated. They were granulin precursor, TNF receptor associated factor 6 (TRAF6), mitogen protein kinase 8, Mdm4 and baculoviral IAP repeat-containing protein 4 as shown in Figures 6 (2 DE gel) and 7 (semiquantitative RT-PCR and Western blot analyses).

Fig. 6. Representative 2-DE gel of protein extracts from D122v3B and D122αa4 cell lines. Five protein spots (O35618, P28798, Q07174, P70196 and Q60989) were up-regulated in D122αa4 cells (Swiss-Prot accession number provided) (Tang et al., 2006).

Comparative Proteomics:

An Approach to Elucidating the Function of a Novel Gene Called BRE 123

Fig. 8. Representative 2-DE gel of nucleic proteins extracted from SHEE and SHEEC cells. Ten silver-stained protein spots were found to be up-regulated in SHEEC cells (Chen et al., 2008).

Fig. 9. Representative 2-DE gel of nucleic proteins extracted from SHEE and SHEEC cells. Ten silver-stained protein spots were found to be down-regulated in SHEEC cells (Chen et al., 2008).

Fig. 7. Semiquantitative RT-PCR (A) and Western blot (B) showing that D122αa4 cells overexpressed *prohibitin*, *p53* and *mdm4*. β*-actin* and α-tubulin serve as internal controls (Tang et al., 2006).

Interestingly, TRAF6 is a unique member of the TRAF family of adaptor protein. It is associated with a diverse range of cellular responses to pathogens, growth factors or intracellular stress (Chung et al., 2007). Recent finding also showed that TRAF6 was involved in the RANK-TRAF6-NF-B pathways during osteoclastogenesis (Inoue et al., 2007). Overexpression of BRE in human 293 embryonic kidney cells has been reported to inhibit NF-B activation in response to TNFα (Gu et al., 1998). This finding suggests that BRE indirectly cross-talk with TRAF6 and NF-β, where it may play a central role in regulating cell proliferation, differentiation and survival. BRE may also mediate in posttranslational sumoylation, similar to the action of PML and MO25α proteins (Kretz-Remy and Tanguay, 1999). Our results established a crucial function for BRE in regulating key proteins of cellular stress-response and provided an explanation for the multifunctional nature of BRE.

#### **5.2.2 Comparative proteomic analysis reveals differentially expressed proteins regulated by a potential tumor promoter, BRE, in human esophageal carcinoma cells**

Esophageal cancer is one of the most common malignancies that cause high mortality. Esophageal carcinogenesis is a complex and cascading process that involve the interaction of many genes and proteins (Kuwano et al., 2005). In this study, we have used comparative proteomic approaches to identify proteins that maybe involved in esophageal carcinogenesis. Two dimensional electrophoresis (2-DE) and MALDI-TOF-MS analyses of esophageal carcinoma, SHEEC and control cells SHEE revealed 10 proteins that were up-regulated as shown in Figure 8 of the 2-DE. Additional 10 proteins were down-regulated as shown in Figure 9. Interestingly, BRE, prohibitin, cyclin A and p53

Fig. 7. Semiquantitative RT-PCR (A) and Western blot (B) showing that D122αa4 cells overexpressed *prohibitin*, *p53* and *mdm4*. β*-actin* and α-tubulin serve as internal controls

**5.2.2 Comparative proteomic analysis reveals differentially expressed proteins regulated by a potential tumor promoter, BRE, in human esophageal carcinoma cells**  Esophageal cancer is one of the most common malignancies that cause high mortality. Esophageal carcinogenesis is a complex and cascading process that involve the interaction of many genes and proteins (Kuwano et al., 2005). In this study, we have used comparative proteomic approaches to identify proteins that maybe involved in esophageal carcinogenesis. Two dimensional electrophoresis (2-DE) and MALDI-TOF-MS analyses of esophageal carcinoma, SHEEC and control cells SHEE revealed 10 proteins that were up-regulated as shown in Figure 8 of the 2-DE. Additional 10 proteins were down-regulated as shown in Figure 9. Interestingly, BRE, prohibitin, cyclin A and p53

Interestingly, TRAF6 is a unique member of the TRAF family of adaptor protein. It is associated with a diverse range of cellular responses to pathogens, growth factors or intracellular stress (Chung et al., 2007). Recent finding also showed that TRAF6 was involved in the RANK-TRAF6-NF-B pathways during osteoclastogenesis (Inoue et al., 2007). Overexpression of BRE in human 293 embryonic kidney cells has been reported to inhibit NF-B activation in response to TNFα (Gu et al., 1998). This finding suggests that BRE indirectly cross-talk with TRAF6 and NF-β, where it may play a central role in regulating cell proliferation, differentiation and survival. BRE may also mediate in posttranslational sumoylation, similar to the action of PML and MO25α proteins (Kretz-Remy and Tanguay, 1999). Our results established a crucial function for BRE in regulating key proteins of cellular stress-response and provided an explanation for the multifunctional

(Tang et al., 2006).

nature of BRE.

Fig. 8. Representative 2-DE gel of nucleic proteins extracted from SHEE and SHEEC cells. Ten silver-stained protein spots were found to be up-regulated in SHEEC cells (Chen et al., 2008).

Fig. 9. Representative 2-DE gel of nucleic proteins extracted from SHEE and SHEEC cells. Ten silver-stained protein spots were found to be down-regulated in SHEEC cells (Chen et al., 2008).

Comparative Proteomics:

(Chen et al., 2008).

An Approach to Elucidating the Function of a Novel Gene Called BRE 125

Fig. 11. Semiquantitative RT-PCR analysis of SHEE and SHEEC cells transfected with CTLand BRE-siRNAs. The results showed that our BRE construct can silence BRE expression, as well as suppressed prohibitin and cyclin A expressions. *β*-*actin* served as an internal control

**5.2.3 Livers over-expressing BRE transgene are under heightened state of stress-**

BRE is normally expressed at very low levels in the liver (Chan, et al., 2008). It binds to TNF-R1 and Fas, and modulates the actions of these cytokines (Li, et al., 2004; Chan et al., 2010). In this study, we demonstrated that BRE expression was rapidly induced when the liver was insulted with carbon tetrachloride (CCl4) or in human hepatocellular carcinoma (HCC) as shown in Figure 12. We produced transgenic mice that specifically over-expressed BRE in the liver to determine the effect of high levels of BRE in the liver. The livers of these transgenic mice were determined to be histologically normal. Because of the lack of a phenotype, we conducted comparative proteomics to determine whether there were any differences at the protein level (Figure 13). The 2-DE revealed four up-regulated protein spots and nine down-regulated protein spots as summarized in Table 2. It was established that several stress responsive proteins were up-regulated in the BRE-transgenic liver including: Alpha enolase (ENO 1), Heat shock-related 70 kDa protein 2 (HSPA2), Putative heat shock 70 kDa protein 7 (HSPA7), Zinc-finger protein Ubid 4 (DPF2) and Tumor suppressor candidate 4 G21 protein (TUSC4) as shown in Figure 14. Recently, it has been reported that HSPA7 is a biomarker for early detection of HCC (Park, 2011). In addition, we have silenced BRE expression in Chang liver cells and inversely demonstrated that it did not affect cell proliferation rate as confirmed by BrdU Labelling assay (Table 3). We have previously reported that BRE is not only expressed in the cytoplasm but also in the nuclei of HCC cells. BRE also accumulates in the nuclei of esophagus cancer SHEEC cells (Chen, et al., 2008). Since BRE is one of the components of BRCA1 A complex, it could be involved in DNA repair, as well as responding to environmental stress. We propose that the livers in our BRE transgenic mice were under a heighten state of stress response and this may explain

**response, as revealed by comparative proteomics** 

why the transgenic mice was more resistant to liver toxic drugs.

Fig. 10. Semiquantitative RT-PCR (A) and Western Blot (B) analyses of SHEE and SHEEC cells. The results confirmed the proteomic data that BRE, prohibitin and cyclin A were highly expressed in SHEEC cells. The SHEEC cells also expressed relatively higher levels of TNF-R1 but lower levels of p53, when compared with SHEE cells. *β-actin* and α-tubulin serve as internal controls (Chen et al., 2008).

expression were up-regulated in the cancer cells and this was confirmed by both semiquantitative RT-PCR and western blot analyses (Figure 10). Among these 20 differentially expressed proteins, BRE protein was identified as a potential tumor promoter. Furthermore, we have also determined p53 expression was down-regulated; whereas TNF-R1 expression was up-regulated in SHEEC cells (Figure 10). It has been reported that BRE can interact with the intracellular juxtamembrane domain TNF-R1 and inhibit the TNF-α induced activation of NF-B (Gu et al., 1998). Therefore, we propose that BRE plays an antiapoptotic role in SHEEC cells. To gain more insight into BRE's function, we silenced BRE expression in esophageal carcinoma cells using BRE-specific small interference RNA. It was found that silencing BRE expression corresponds to down-regulated prohibitin expression but up-regulated tumor-suppressor gene, p53 as shown in Figure 11. These findings contradicted the results with previous data (Tang, et al., 2006) that may due to multifunctional nature of BRE. Besides BRE, cyclin A and CDK2 expressions were suppressed in the SHEEC cells. Cyclin A is an important regulator of the cell cycle that rises in early S phase and falls in mid M phase (Parwaresch and Rudolph, 1996). Recent finding showed the cyclin A might be a prognostic marker in early breast cancer (Ahlin, et al. 2007). In summary, these results imply that BRE may be a survival factor and plays a proliferative role in esophageal carcinoma.

Fig. 10. Semiquantitative RT-PCR (A) and Western Blot (B) analyses of SHEE and SHEEC cells. The results confirmed the proteomic data that BRE, prohibitin and cyclin A were highly expressed in SHEEC cells. The SHEEC cells also expressed relatively higher levels of TNF-R1 but lower levels of p53, when compared with SHEE cells. *β-actin* and α-tubulin

expression were up-regulated in the cancer cells and this was confirmed by both semiquantitative RT-PCR and western blot analyses (Figure 10). Among these 20 differentially expressed proteins, BRE protein was identified as a potential tumor promoter. Furthermore, we have also determined p53 expression was down-regulated; whereas TNF-R1 expression was up-regulated in SHEEC cells (Figure 10). It has been reported that BRE can interact with the intracellular juxtamembrane domain TNF-R1 and inhibit the TNF-α induced activation of NF-B (Gu et al., 1998). Therefore, we propose that BRE plays an antiapoptotic role in SHEEC cells. To gain more insight into BRE's function, we silenced BRE expression in esophageal carcinoma cells using BRE-specific small interference RNA. It was found that silencing BRE expression corresponds to down-regulated prohibitin expression but up-regulated tumor-suppressor gene, p53 as shown in Figure 11. These findings contradicted the results with previous data (Tang, et al., 2006) that may due to multifunctional nature of BRE. Besides BRE, cyclin A and CDK2 expressions were suppressed in the SHEEC cells. Cyclin A is an important regulator of the cell cycle that rises in early S phase and falls in mid M phase (Parwaresch and Rudolph, 1996). Recent finding showed the cyclin A might be a prognostic marker in early breast cancer (Ahlin, et al. 2007). In summary, these results imply that BRE may be a survival factor and plays a proliferative

serve as internal controls (Chen et al., 2008).

role in esophageal carcinoma.

Fig. 11. Semiquantitative RT-PCR analysis of SHEE and SHEEC cells transfected with CTLand BRE-siRNAs. The results showed that our BRE construct can silence BRE expression, as well as suppressed prohibitin and cyclin A expressions. *β*-*actin* served as an internal control (Chen et al., 2008).

#### **5.2.3 Livers over-expressing BRE transgene are under heightened state of stressresponse, as revealed by comparative proteomics**

BRE is normally expressed at very low levels in the liver (Chan, et al., 2008). It binds to TNF-R1 and Fas, and modulates the actions of these cytokines (Li, et al., 2004; Chan et al., 2010). In this study, we demonstrated that BRE expression was rapidly induced when the liver was insulted with carbon tetrachloride (CCl4) or in human hepatocellular carcinoma (HCC) as shown in Figure 12. We produced transgenic mice that specifically over-expressed BRE in the liver to determine the effect of high levels of BRE in the liver. The livers of these transgenic mice were determined to be histologically normal. Because of the lack of a phenotype, we conducted comparative proteomics to determine whether there were any differences at the protein level (Figure 13). The 2-DE revealed four up-regulated protein spots and nine down-regulated protein spots as summarized in Table 2. It was established that several stress responsive proteins were up-regulated in the BRE-transgenic liver including: Alpha enolase (ENO 1), Heat shock-related 70 kDa protein 2 (HSPA2), Putative heat shock 70 kDa protein 7 (HSPA7), Zinc-finger protein Ubid 4 (DPF2) and Tumor suppressor candidate 4 G21 protein (TUSC4) as shown in Figure 14. Recently, it has been reported that HSPA7 is a biomarker for early detection of HCC (Park, 2011). In addition, we have silenced BRE expression in Chang liver cells and inversely demonstrated that it did not affect cell proliferation rate as confirmed by BrdU Labelling assay (Table 3). We have previously reported that BRE is not only expressed in the cytoplasm but also in the nuclei of HCC cells. BRE also accumulates in the nuclei of esophagus cancer SHEEC cells (Chen, et al., 2008). Since BRE is one of the components of BRCA1 A complex, it could be involved in DNA repair, as well as responding to environmental stress. We propose that the livers in our BRE transgenic mice were under a heighten state of stress response and this may explain why the transgenic mice was more resistant to liver toxic drugs.

Comparative Proteomics:

An Approach to Elucidating the Function of a Novel Gene Called BRE 127

Fig. 13. A representative 2-DE gel of BRE transgenic liver. Protein spots 1–15 were identified to be differentially expressed when compared with control gels. Protein spots 1–4 were downregulated in the transgenic (trans) liver, while protein spots 5–15 were upregulated in the wild type (wt) liver. These results were acquired from three independent liver samples

and 2-DE was correspondingly performed three times (Tang et al., 2009).

Fig. 12. In situ hybridization (A–D and G–I). BRE is normally expressed at very low levels in normal mouse liver (A). CCl4 insult induced increased BRE expression in the affected hepatocytes at 6 h (B) and 12h (C). Twenty-four hours after CCl4 insult, BRE expression declined. This was probably the result of the affected hepatocytes starting to die off (D). Immunohistological staining revealed that BRE expression was strongly induced in the affected hepatocytes by CCl4 (E, F). BRE expression remained low in the unaffected cells. We also examined BRE expression in HCC cells. BRE was expressed at low levels in nontumor human liver tissues (H). In HCC tissues, all the cells strongly expressed BRE (I). Sense control (G). Arrows, hepatocytes overexpressing BRE. C, liver central veins (Tang et al., 2009).

Fig. 12. In situ hybridization (A–D and G–I). BRE is normally expressed at very low levels in normal mouse liver (A). CCl4 insult induced increased BRE expression in the affected hepatocytes at 6 h (B) and 12h (C). Twenty-four hours after CCl4 insult, BRE expression declined. This was probably the result of the affected hepatocytes starting to die off (D). Immunohistological staining revealed that BRE expression was strongly induced in the affected hepatocytes by CCl4 (E, F). BRE expression remained low in the unaffected cells. We also examined BRE expression in HCC cells. BRE was expressed at low levels in nontumor human liver tissues (H). In HCC tissues, all the cells strongly expressed BRE (I). Sense control (G). Arrows, hepatocytes overexpressing BRE. C, liver central veins (Tang et al.,

2009).

Fig. 13. A representative 2-DE gel of BRE transgenic liver. Protein spots 1–15 were identified to be differentially expressed when compared with control gels. Protein spots 1–4 were downregulated in the transgenic (trans) liver, while protein spots 5–15 were upregulated in the wild type (wt) liver. These results were acquired from three independent liver samples and 2-DE was correspondingly performed three times (Tang et al., 2009).

Comparative Proteomics:

**6. Future perspective of proteomics** 

discovered and also lead to the discovery of new drugs.

off values? Histopathology. 51(4):491-8.

potential. Circ Res. 88(8):763-73.

2009).

**7. References** 

An Approach to Elucidating the Function of a Novel Gene Called BRE 129

Table 3. Effects of silencing BRE expression on Chang liver cell proliferation (Tang et al.,

Conventional "gel-based" electrophoresis and improved mass spectrometry have provided useful tools for revealing molecular changes in cells and tissues that otherwise maybe missed by morphological observation alone (Vercauteren et al., 2007). Nevertheless, the 2-DE protocol is still to be refined and improved so that 2-DE is more reproducible and sensitive. Therefore, it has still some distance to go before it can be adopted as a standard "diagnostic tool" in the 21st century (Colucci-D'Amato et al., 2011). The "shotgun" methodology has been used as a high-throughput screen to identify proteins that are differentially expressed in cells or tissues, as a result of some experimental procedure or changes in environmental condition (Lill, 2003; Zhu et al., 2010). Liu et al. (2011) recently described the SELDI-TOF-MS technology that could be used to screen and detect differentially expressed proteins in the serum of patients with cancer. Liquid chromatography interfaced plasma mass spectrometry has now been developed for absolute quantitation of proteins (Esteban-Fernández et al., 2011). Furthermore, latest development of computational tools for analyzing high-throughput 'shotgun' proteomic data also play a vital role in moving proteomic research forward (Dowsey et al., 2010). All of these improvements will allow proteomics to be rapidly developed as a practical, robust, accurate and inexpensive analytical tool for routine use in the clinical setting. The proteomics will also allow many novel disease biomarkers to be

[1] Ahlin C, Aaltonen K, Amini RM, Nevanlinna H, Fjällskog ML, Blomqvist C. (2007) Ki67

[2] Altland K, Becher P, Rossmann U, Bjellqvist B. (1988) Isoelectric focusing of basic proteins: the problem of oxidation of cysteines. Electrophoresis. (9):474-85. [3] Andersen JS, Svensson B, Roepstorff P. (1996) Electrospray ionization and matrix

in recombinant protein chemistry. Nat Biotechnol. 14(4):449-57. [4] Arnstein, HR (1965). Mechanism of protein biosynthesis. Br Med Bull. 21(3):217-22. [5] Arrell DK, Neverova I, Van Eyk JE. (2001) Cardiovascular proteomics: evolution and

and cyclin A as prognostic factors in early breast cancer. What are the optimal cut-

assisted laser desorption/ionization mass spectrometry: powerful analytical tools

Fig. 14. Semi-quantitative RT-PCR revealed that the proteins identified were differentially expressed in BRE transgenic livers were also correspondingly affected at the transcriptional level. \*p<0.05, \*\*p<0.01, denote significant difference in the staining intensity of wt and BRE transgenic PCR bands (Tang et al., 2009).


Table 2. Proteins that are differentially expressed in BRE transgenic liver (Tang et al., 2009).


Table 3. Effects of silencing BRE expression on Chang liver cell proliferation (Tang et al., 2009).

#### **6. Future perspective of proteomics**

Conventional "gel-based" electrophoresis and improved mass spectrometry have provided useful tools for revealing molecular changes in cells and tissues that otherwise maybe missed by morphological observation alone (Vercauteren et al., 2007). Nevertheless, the 2-DE protocol is still to be refined and improved so that 2-DE is more reproducible and sensitive. Therefore, it has still some distance to go before it can be adopted as a standard "diagnostic tool" in the 21st century (Colucci-D'Amato et al., 2011). The "shotgun" methodology has been used as a high-throughput screen to identify proteins that are differentially expressed in cells or tissues, as a result of some experimental procedure or changes in environmental condition (Lill, 2003; Zhu et al., 2010). Liu et al. (2011) recently described the SELDI-TOF-MS technology that could be used to screen and detect differentially expressed proteins in the serum of patients with cancer. Liquid chromatography interfaced plasma mass spectrometry has now been developed for absolute quantitation of proteins (Esteban-Fernández et al., 2011). Furthermore, latest development of computational tools for analyzing high-throughput 'shotgun' proteomic data also play a vital role in moving proteomic research forward (Dowsey et al., 2010). All of these improvements will allow proteomics to be rapidly developed as a practical, robust, accurate and inexpensive analytical tool for routine use in the clinical setting. The proteomics will also allow many novel disease biomarkers to be discovered and also lead to the discovery of new drugs.

#### **7. References**

128 Proteomics – Human Diseases and Protein Functions

Fig. 14. Semi-quantitative RT-PCR revealed that the proteins identified were differentially expressed in BRE transgenic livers were also correspondingly affected at the transcriptional level. \*p<0.05, \*\*p<0.01, denote significant difference in the staining intensity of wt and BRE

Table 2. Proteins that are differentially expressed in BRE transgenic liver (Tang et al., 2009).

transgenic PCR bands (Tang et al., 2009).


Comparative Proteomics:

Cell. 143(5):682-5.

10(23):4226-57.

Tech. 13(3):101-18.

Cytochem. 31(8):1033-40.

cells. Blood. 118(8):2211-21.

30 Suppl 1:S122-32.

signaling. J Biol Chem. 278(48):47853-61.

722.

An Approach to Elucidating the Function of a Novel Gene Called BRE 131

[19] Chowdhury SK, Katta V, Chait BT. (1990) An electrospray-ionization mass spectrometer

[20] Chung JY, Lu M, Yin Q, Lin SC, Wu H. (2007) Molecular basis for the unique specificity

[21] Clague MJ, Urbé S. (2010) Ubiquitin: same molecule, different degradation pathways.

[22] Colucci-D'Amato L, Farina A, Vissers JP, Chambery A. (2011) Quantitative

[23] Conrotto P, Souchelnytskyi S. (2008) Proteomic approaches in biological and medical

[24] Dong Y, Hakimi MA, Chen X, Kumaraswamy E, Cooch NS, Godwin AK, Shiekhattar R.

[26] Edman, P. (1950) Method for Determination of the Amino Acid Sequence in Peptides.

[31] Finnerty V, Johnson G. (1979) Post-Translational Modification as a Potential

[32] Fleischer, B. (1983) Mechanism of glycosylation in the Golgi apparatus. J Histochem

[33] Fusaro G, Dasgupta P, Rastogi S, Joshi B, Chellappan S. (2003) Prohibitin induces the

[34] Galvão ER, Martins LM, Ibiapina JO, Andrade HM, Monte SJ. (2011) Breast cancer proteomics: a review for clinicians. J Cancer Res Clin Oncol. 137(6):915-25. [35] Gioia R, Leroy C, Drullion C, Lagarde V, Etienne G, Dulucq S, Lippert E, Roche S,

[36] Görg A, Drews O, Lück C, Weiland F, Weiss W.(2009) 2-DE with IPGs. Electrophoresis.

[37] Görg A, Weiss W, Dunn MJ. (2004) Current two-dimensional electrophoresis

technology for proteomics. Proteomics. 4(12):3665-85.

Explanation of High Levels of Enzyme Polymorphism: Xanthine Dehydrogenase and Aldehyde Oxidase in DROSOPHILA MELANOGASTER. Genetics. 91(4):695-

transcriptional activity of p53 and is exported from the nucleus upon apoptotic

Mahon FX, Pasquet JM. (2011) Quantitative phosphoproteomics revealed interplay between Syk and Lyn in the resistance to nilotinib in chronic myeloid leukemia

[27] Edman, P. and Begg, G. (1967) A protein Sequenator. European J. Biochem. 1: 80-91. [28] Esteban-Fernández D, Scheler C, Linscheid MW. (2011) Absolute protein quantification by LC-ICP-MS using MeCAT peptide labeling. Anal Bioanal Chem. 401(2):657-66. [29] Fenn JB, Mann M, Meng CK, Wong SF, Whitehouse CM. (1989) Electrospray ionization for mass spectrometry of large biomolecules. Science. 246(4926):64-71. [30] Fenn JB. (2002) Electrospray ionization mass spectrometry: How it all began. J Biomol

neuroproteomics: classical and novel tools for studying neural differentiation and

(2003) Regulation of BRCC, a holoenzyme complex containing BRCA1 and BRCA2, by a signalosome-like subunit and its role in DNA repair. Mol Cell. 12(5):1087-99. [25] Dowsey AW, English JA, Lisacek F, Morris JS, Yang GZ, Dunn MJ. (2010) Image

analysis tools and emerging algorithms for expression proteomics. Proteomics.

with new features. Rapid Commun Mass Spectrom. 4(3):81-7.

sciences: principles and applications. Exp Oncol. 30(3):171-80.

of TRAF6. Adv Exp Med Biol. 597:122-30.

function. Stem Cell Rev. 7(1):77-93.

Acta Chem Scand 10: 283-293.


[6] Bergquist J, Palmblad M, Wetterhall M, Håkansson P, Markides KE. (2002) Peptide

[7] Birr C, Frank R. (1975) Control for uniformity of synthetic peptides: quantitative

[8] Bjellqvist B, Ek K, Righetti PG, Gianazza E, Görg A, Westermeier R, Postel W. (1982)

[11] Celis JE, Celis P, Ostergaard M, Basse B, Lauridsen JB, Ratz G, Rasmussen HH, Orntoft

[12] Chalkley RJ, Baker PR, Huang L, Hansen KC, Allen NP, Rexach M, Burlingame

[13] Chan BC, Ching AK, To KF, Leung JC, Chen S, Li Q, Lai PB, Tang NL, Shaw PC, Chan

[14] Chan BC, Li Q, Chow SK, Ching AK, Liew CT, Lim PL, Lee KK, Chan JY, Chui

[15] Chan JY, Li L, Miao J, Cai DQ, Lee KK, Chui YL.(2010) Differential expression of a novel

[16] Chen HB, Pan K, Tang MK, Chui YL, Chen L, Su ZJ, Shen ZY, Li EM, Xie W, Lee KK.

[17] Chen R, Snyder M. (2010) Yeast proteomics and protein microarrays. J Proteomics.

[18] Ching, A.K., Li, P.S., Li, Q., Chan, B.S., Chan, J.Y., Lim, P.L., Pang, J.C., Chui, Y.L.

synthetized by the Merrifield method. FEBS Lett. 55(1):68-71.

some applications. J Biochem Biophys Methods. 6(4):317-39. [9] Blundell TL, Johnson MS. (1993) Catching a common fold. Protein Sci. 2(6):877-83. [10] Ceglarek U, Kortz L, Leichtle A, Fiedler GM, Kratzsch J, Thiery J. (2009) Rapid

spectrometry. Clin Chim Acta.401(1-2):114-8.

datasets. Mol Cell Proteomics 4(8):1194-204.

Res. 59(12):3003-9.

27(9):1208-17.

73(11):2147-57.

Commun. 288, 535-545.

Commun 326(2):268-73.

signals. Mol Biol Rep. 37(1):363-8.

cells. Biochem Cell Biol. 86(4):302-11.

15.

mapping of proteins in human body fluids using electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. Mass Spectrom Rev. 21(1):2-

evaluation of the Edman degradation of support-bound human insulin B 22-30

Isoelectric focusing in immobilized pH gradients: principle, methodology and

quantification of steroid patterns in human serum by on-line solid phase extraction combined with liquid chromatography-triple quadrupole linear ion trap mass

TF, Hein B, Wolf H, Celis A.(1999) Proteomics and immunohistochemistry define some of the steps involved in the squamous differentiation of the bladder transitional epithelium: a novel strategy for identifying metaplastic lesions. Cancer

AL.(2005) Comprehensive analysis of a multidimensional liquid chromatography mass spectrometry dataset acquired on a quadrupole selecting, quadrupole collision cell, time-of-flight mass spectrometer: II. New developments in Protein Prospector allow for reliable and comprehensive automatic analysis of large

JY, James AE, Lai KN, Lim PL, Lee KK, Chui YL. (2008) BRE is an antiapoptotic protein in vivo and overexpressed in human hepatocellular carcinoma. Oncogene.

YL.(2005). BRE enhances in vivo growth of tumor cells. Biochem Biophys Res

gene BRE (TNFRSF1A modulator/BRCC45) in response to stress and biological

(2008) Comparative proteomic analysis reveals differentially expressed proteins regulated by a potential tumor promoter, BRE, in human esophageal carcinoma

(2001). Expression of human BRE in multiple isoforms. Biochem. Biophys. Res.


Comparative Proteomics:

5:19.

An Approach to Elucidating the Function of a Novel Gene Called BRE 133

[56] Madhunapantula SV, Robertson GP. (2009) The PTEN-AKT3 signaling cascade as a therapeutic target in melanoma. Pigment Cell Melanoma Res. 22(4):400-19. [57] Mao LM, Guo ML, Jin DZ, Fibuch EE, Choe ES, Wang JQ. (2011) Post-translational

[58] Marko-Varga G, Berglund M, Malmström J, Lindberg H, Fehniger TE. (2003) Targeting

[59] McKerrow JH, Bhargava V, Hansell E, Huling S, Kuwahara T, Matley M, Coussens L,

[60] Miao J, Chan KW, Chen GG, Chun SY, Xia NS, Chan JY, Panesar NS. (2005) Blocking

[61] Miao, J., Panesar, N.S., Chan, K.T., Lai, F.M., Xia, N., Wang, Y., Johnson, P.J. and Chan,

[62] Mishra S, Ande SR, Nyomba BL. (2010) The role of prohibitin in cell signaling. FEBS J.

[63] Morris, HR, Williams, DH, Ambler RP. (1971) Determination of the Sequences of

[64] Niall HD, Keutmann HT, Copp DH, Potts JT Jr. (1969) Amino acid sequence of salmon

[65] Novak A, Amit M, Ziv T, Segev H, Fishman B, Admon A, Itskovitz-Eldor J. (2011).

[66] O'Farrell PH.(1975) High resolution two-dimensional electrophoresis of proteins. J Biol

[67] Ornstein DK, Rayford W, Fusaro VA, Conrads TP, Ross SJ, Hitt BA, Wiggins WW,

[69] Park YN. (2011) Update on precursor and early lesions of hepatocellular carcinomas.

[70] Parwaresch R, Rudolph P. (1996) The Cell Cycle – Theory and Application to Cancer.

[71] Perkins DN, Pappin DJ, Creasy DM, Cottrell JS. (1999) Probability-based protein

[72] Sauter ER, Zhu W, Fan XJ, Wassell RP, Chervoneva I, Du Bois GC. (2002) Proteomic

identification by searching sequence databases using mass spectrometry data.

analysis of nipple aspirate fluid to detect biologic markers of breast cancer. Br J

ultimobranchial calcitonin. Proc Natl Acad Sci U S A. 64(2):771-8.

expression profiling. Electrophoresis. 24(21):3800-5.

hydroxysteroid dehydrogenase. J Endocrinol. 185(3):507-17.

carcinoma. Mol Med. 6(5):450-60.

Cytochem. 49, 491-500.

Stage. Stem Cell Rev. 2011 Jul 6.

biomarkers. Urol Oncol. 24(3):231-6.

Arch Pathol Lab Med. 135(6):704-15.

Electrophoresis. 20(18):3551-67.

Chem. 250(10):4007-21.

Onkologie 19:464-472.

Cancer. 86(9):1440-3.

277(19):3937-46.

125: 189-201.

modification biology of glutamate receptors and drug addiction. Front Neuroanat.

hepatocytes from liver tissue by laser capture microdissection and proteomics

Warren R. (2000) A functional proteomics screen of proteases in colorectal

BRE expression in Leydig cells inhibits steroidogenesis by down-regulating 3beta-

J.Y. (2001). Differential expression of a stress-modulating gene, BRE, in the adrenal gland, in adrenal neoplasia, and in abnormal adrenal tissues. J. Histochem.

Protein-Derived Peptides and Peptide Mixtures by Mass Spectrometry. Biochem. J.

Proteomics Profiling of Human Embryonic Stem Cells in the Early Differentiation

Veenstra TD, Liotta LA, Petricoin EF 3rd. (2004) Serum proteomic profiling can discriminate prostate cancer from benign prostates in men with total prostate specific antigen levels between 2.5 and 15.0 ng/ml. J Urol. 172(4 Pt 1):1302-5. [68] Ornstein DK, Tyson DR. (2006) Proteomics for the identification of new prostate cancer


[38] Gu C, Castellino A, Chan JY, Chao MV. (1998) BRE: a modulator of TNF-alpha action.

[39] Han MJ, Lee JW, Lee SY. (2011) Understanding and engineering of microbial cells based

[40] Harris JL, Khanna KK. (2011) BRCA1 A‐complex fine tunes repair functions of BRCA1.

[41] Hillenkamp F, Karas M, Beavis RC, Chait BT. (1991) Matrix-assisted laser

[42] Hu X, Kim JA, Castillo A, Huang M, Liu J, Wang B. (2011) NBA1/MERIT40 and BRE

[43] Inoue J, Gohda J, Akiyama T, Semba K. (2007) NF-kappaB activation in development

[44] Joshi B, Rastogi S, Morris M, Carastro LM, DeCook C, Seto E, Chellappan SP. (2007)

[45] Kretz-Remy C, Tanguay RM. (1999) SUMO/sentrin: protein modifiers regulating

[46] Kuwano H, Kato H, Miyazaki T, Fukuchi M, Masuda N, Nakajima M, Fukai Y, Sohda

[47] Larbi, NB and Jefferies, C. (2009) 2D-DIGE: Comparative Proteomics of Cellular

[48] Lau TY, O'Connor DP, Brennan DJ, Duffy MJ, Pennington SR, Gallagher WM.(2007) Breast cancer proteomics: clinical perspectives. Expert Opin Biol Ther. 7(2):209-19. [49] Lee KK, Leung AK, Tang MK, Cai DQ, Schneider C, Brancolini C, Chow PH (2001).

[51] Li Q, Ching AK, Chan BC, Chow SK, Lim PL, Ho TC, Ip WK, Wong CK, Lam CW, Lee

[53] Lill, J (2003) Proteomic Tools for Quantitation by Mass Spectrometry. Mass Spect Rev.

[54] Liu C, Pan C, Wang H, Yong L. (2011) Effect of surface-enhanced laser

[55] Lucas F, Barber M, Wolstenholme WA, Geddes AJ, Graham GN, Morris HR. (1969)

from the protein silk fibroin of Bombyx mori. Biochem J. 114(4):695-702.

[50] Lengyel P. (1966) Problems in protein biosynthesis. J Gen Physiol. 49(6):305-30.

retinoic acid. Biochem. Biophys. Res. Commun. 206, 764-774.

of laryngeal carcinoma. Tumour Biol. 2011 Aug 9.

BRCC36-containing complexes. J Biol Chem. 286(13):11734-45.

through E2F1 and p53 binding sites. Biochem J. 401(1):155-66.

important cellular functions. Biochem Cell Biol. 77(4):299-309.

and progression of cancer. Cancer Sci. 98(3):268-74.

Signalling Pathways. Methods Mol Biol. 517:105-32.

on proteomics and its conjunction with other omics studies. Proteomics. 11(4):721-

desorption/ionization mass spectrometry of biopolymers. Anal Chem.

interaction is required for the integrity of two distinct deubiquitinating enzyme

Differential regulation of human YY1 and caspase 7 promoters by prohibitin

M, Kimura H, Faried A. (2005) Genetic alterations in esophageal cancer. Surg

Functions of the growth arrest specific 1 gene in the development of the mouse

KK, Chan JY, Chui YL. (2004) A death receptor-associated anti-apoptotic protein, BRE, inhibits mitochondrial apoptotic pathway. J Biol Chem. 279(50):52106-16. [52] Li, L., Yoo, H., Becker, F.F., Ali-Osman, F. and Chan, J.Y. (1995). Identification of a

brain- and reproductive-organs-specific gene responsive to DNA damage and

desorption/ionization time-of-flight mass spectrometry on identifying biomarkers

Mass-spectrometric determination of the amino acid sequences in peptides isolated

FASEB J. 12(12):1101-8.

63(24):1193A-1203A.

Today. 35(1):7-18.

22: 182-194.

embryo. Dev Biol 234(1):188-203.

Aging (Albany NY). 3(5):461-3.

43.


**7** 

*Spain* 

**Proteomic Approaches to Unraveling** 

Jone Mitxelena1, Nerea Osinalde2, Jesus M. Arizmendi2,

Correct entry into and progression through the cell cycle require an intact RB/E2F pathway, and its deregulation is now considered a general hallmark of cancer (Nevins 2001). Pioneer work in the early 90's showed that E2F activity is controlled through the temporal association of E2F factors with Retinoblastoma (RB) tumor suppressor proteins (pRB, p107 and p130), also called pocket proteins (Bandara & La Thangue 1991; Chellappan et al. 1991). The additional finding that RB activity is regulated through phosphorylations by cyclin dependent kinases (CDKs) provided the groundwork for the current model of cell cycle control (Weinberg 1995). According to this model, non-phosphorylated RB binds to E2F in G0/G1, leading to the repression of E2F target genes. Subsequent phosphorylation of RB by CDKs in mid-to late G1 disrupts its association with E2F. As a result, free E2F triggers the expression of target genes necessary for entry into and progression through the cell cycle (Burkhart & Sage 2008). This pathway is thought to be disrupted in most human cancers, either by activation of positively acting components such as G1 cyclins and CDKs, or the inactivation of negatively acting components such as RB and cyclin kinase inhibitors (Nevins 2001). The predicted consequence of deficient RB-mediated regulation is that E2F activity is constantly unleashed from the inhibitory effects of RB (DeGregori & Johnson

Mammalian E2F is a family of related factors (E2F1-8), originally discovered for their pivotal role in transcriptional regulation of genes associated with DNA replication and G1/S progression (Attwooll et al. 2004; Trimarchi & Lees 2002). More recently, microarray expression profiling analyses and ChIP-chip analyses (chromatin immunoprecipitation coupled to microarray technology) in cells overexpressing individual E2Fs have revealed that the transactivation function of these factors exceeds beyond G1/S transition regulation. In fact, E2Fs regulate a wide spectrum of genes with diverse biological functions, including regulation of apoptosis, autophagy, mitosis, chromosome organization, macromolecule metabolism, or differentiation (Ma et al. 2002; Muller et al. 2001; Polager et al. 2008; Ren et al. 2002; Weinmann et al. 2002; Young et al. 2003). Thus, the role of E2F transcription factors

in cellular physiology is probably more complex than it was originally thought to be.

**1. Introduction** 

2006; Dimova & Dyson 2005; Iaquinta & Lees 2007).

*1Dept. Genetics, Physical Anthropology and Animal Physiology* 

**the RB/E2F Regulatory Pathway** 

Asier Fullaondo1 and Ana M. Zubiaga1

*2Dept. Biochemistry and Molecular Biology,* 

*University of the Basque Country, UPV/EHU, Bilbao* 

*University of the Basque Country, UPV/EHU, Bilbao* 


### **Proteomic Approaches to Unraveling the RB/E2F Regulatory Pathway**

Jone Mitxelena1, Nerea Osinalde2, Jesus M. Arizmendi2, Asier Fullaondo1 and Ana M. Zubiaga1 *1Dept. Genetics, Physical Anthropology and Animal Physiology University of the Basque Country, UPV/EHU, Bilbao 2Dept. Biochemistry and Molecular Biology, University of the Basque Country, UPV/EHU, Bilbao Spain* 

#### **1. Introduction**

134 Proteomics – Human Diseases and Protein Functions

[73] Schwartz JC, Jardine I. (1996) Quadrupole ion trap mass spectrometry. Methods

[74] Shen Z, Cen S, Shen J, et al. 2000. Study of immortalization and malignant

[75] Sirot LK, Hardstone MC, Helinski ME, Ribeiro JM, Kimura M, Deewatthanawong P,

[76] Tang, M.K., Wang, C.M., Shan, S.W., Chui, Y.L., Ching, A.K., Chow, P.H., Grotewold,

prohibitin and p53 expression and proliferation. Proteomics 6, 2376-2385. [77] Tang, MK, Liu, G, Hou, Z., Chui, YL, Chan, JYH, Lee, KKH (2009) Livers

revealed by comparative proteomics. Proteomics Clin. Appl. 3(12):1362-70. [78] Tung CS, Wong KK, Mok SC. (2008) Biomarker discovery in ovarian cancer. Womens

[79] Vercauteren FG, Arckens L, Quirion R. (2007) Applications and current challenges of

[80] Voss T, Ahorn H, Haberl P, Döhner H, Wilgenbus K. (2001) Correlation of clinical data

[81] Weiss W, Görg A. (2009) High-resolution two-dimensional electrophoresis. Methods

[82] Wilkins MR, Sanchez JC, Gooley AA, Appel RD, Humphery-Smith I, Hochstrasser DF,

[83] Zhang Z, Bast RC Jr, Yu Y, Li J, Sokoll LJ, Rai AJ, Rosenzweig JM, Cameron B, Wang

[84] Zhu, W, Smith, JW and Hung, CM (2010). Mass Spectrometry-Based Label-Free

Quantitative Proteomics. J. Biomed and Biotech 2010: 840518.

transformation of human embryonic esophageal epithelial cells induced by HPV18

Wolfner MF, Harrington LC. (2011) Towards a semen proteome of the dengue vector mosquito: protein identification and potential functions. PLoS Negl Trop

L., Chan, J.Y. and Lee, K.K. (2006). Comparative proteomic analysis reveals a function of the novel death receptor-associated protein BRE in the regulation of

overexpressing BRE transgene are under heightened state of stress-response, as

proteomic approaches, focusing on two-dimensional electrophoresis. Amino Acids.

with proteomics profiles in 24 patients with B-cell chronic lymphocytic leukemia.

Williams KL.(1996) Progress with proteome projects: why all proteins expressed by a genome should be identified and how to do it. Biotechnol Genet Eng Rev. 13:19-

YY, Meng XY, Berchuck A, Van Haaften-Day C, Hacker NF, de Bruijn HW, van der Zee AG, Jacobs IJ, Fung ET, Chan DW. (2004) Three biomarkers identified from serum proteomic analysis for the detection of early stage ovarian cancer. Cancer

Enzymol. 270:552-86.

Dis. 5(3):e989.

33(3):405-14.

50.

Health (Lond Engl). 4:27-40.

Int J Cancer. 91(2):180-6.

Mol Biol. 564:13-32.

Res. 64(16):5882-90.

E6E7. J Cancer Res Clin Oncol, 126(10):589-94.

Correct entry into and progression through the cell cycle require an intact RB/E2F pathway, and its deregulation is now considered a general hallmark of cancer (Nevins 2001). Pioneer work in the early 90's showed that E2F activity is controlled through the temporal association of E2F factors with Retinoblastoma (RB) tumor suppressor proteins (pRB, p107 and p130), also called pocket proteins (Bandara & La Thangue 1991; Chellappan et al. 1991). The additional finding that RB activity is regulated through phosphorylations by cyclin dependent kinases (CDKs) provided the groundwork for the current model of cell cycle control (Weinberg 1995). According to this model, non-phosphorylated RB binds to E2F in G0/G1, leading to the repression of E2F target genes. Subsequent phosphorylation of RB by CDKs in mid-to late G1 disrupts its association with E2F. As a result, free E2F triggers the expression of target genes necessary for entry into and progression through the cell cycle (Burkhart & Sage 2008). This pathway is thought to be disrupted in most human cancers, either by activation of positively acting components such as G1 cyclins and CDKs, or the inactivation of negatively acting components such as RB and cyclin kinase inhibitors (Nevins 2001). The predicted consequence of deficient RB-mediated regulation is that E2F activity is constantly unleashed from the inhibitory effects of RB (DeGregori & Johnson 2006; Dimova & Dyson 2005; Iaquinta & Lees 2007).

Mammalian E2F is a family of related factors (E2F1-8), originally discovered for their pivotal role in transcriptional regulation of genes associated with DNA replication and G1/S progression (Attwooll et al. 2004; Trimarchi & Lees 2002). More recently, microarray expression profiling analyses and ChIP-chip analyses (chromatin immunoprecipitation coupled to microarray technology) in cells overexpressing individual E2Fs have revealed that the transactivation function of these factors exceeds beyond G1/S transition regulation. In fact, E2Fs regulate a wide spectrum of genes with diverse biological functions, including regulation of apoptosis, autophagy, mitosis, chromosome organization, macromolecule metabolism, or differentiation (Ma et al. 2002; Muller et al. 2001; Polager et al. 2008; Ren et al. 2002; Weinmann et al. 2002; Young et al. 2003). Thus, the role of E2F transcription factors in cellular physiology is probably more complex than it was originally thought to be.

Proteomic Approaches to Unraveling the RB/E2F Regulatory Pathway 137

Fig. 1. A schematic diagram showing proteomic approaches to analyzing regulatory

the complexes have been removed from the cellular environment.

The interaction between the bait and target proteins brings into close proximity the DNA binding and activation domains of the transcription factor, resulting in the activation of a reporter gene (Fields & Song 1989). Given that the conditions applied in this methodology are not physiological, some of the detected interactions may not represent true interactions. Consequently, this experimental system is thought to yield a high false positive rate. Consequently, interactions detected by this system need to be further validated in an appropriate physiological system. Despite the mentioned drawbacks, it is also true that the yeast two-hybrid interaction screening provides a method to scrutinize protein-protein interactions within living cells, whereas other approaches measure protein interactions after

Interaction proteomics has been helpful in exploring the intricate macromolecular interactions established by RB and E2F for the regulation of gene expression. Work from many laboratories has shown that RB mediates transcriptional repression through the recruitment of a large number of co-repressors, resulting in an alteration of chromatin conformation that hinders transcription. Most RB/E2F co-repressors, including histone deacetylases (HDAC1-3), nucleosome remodeling proteins (BRG1), DNA methyl transferases (DNMT1) or RBP1 have been identified through hypothesis-driven classical biochemical approaches (Brehm et al. 1998; Luo et al. 1998; Magnaghi-Jaulin et al. 1998). Interestingly, HBP1 and CtIP/CtBP co-repressors were discovered by yeast two-hybrid screening analyses using the pocket protein p130 as the bait (Meloni et al. 1999a; Tevosian et al. 1997). HBP1, a tumor suppressor member of the HMG family of transcription factors,

signaling pathways.

Traditionally, the mammalian E2F family has been divided into "activators" (E2F1-3) and "repressors" (E2F4-8). However, recent *in vivo* data have questioned this oversimplified classification. Indeed, accumulating evidence suggests that most E2Fs can function both as activators as well as repressors, depending on the cellular context (Balciunaite et al. 2005; Iglesias et al. 2004; Infante et al. 2008; Lang et al. 2001; Lee et al. 2011; Ma et al. 2002; Morris et al. 2000; Muller et al. 2001; Polager et al. 2008; Young et al. 2003). Likewise, both oncogenic and tumor suppressor properties have been assigned to these factors (DeGregori & Johnson 2006; Johnson & DeGregori 2006). The mechanisms underlying this bimodal impact of individual E2Fs, and their implication in human cancer development remain to be elucidated. This is a particularly relevant point that needs to be addressed, since strategies based on E2F biology are being devised for the development of anticancer therapies (Kaelin, Jr. 2003). An additional level of complexity in the understanding of E2F function *in vivo*  derives from the considerable functional overlap existing among several E2F members (Chen et al. 2009a; DeGregori & Johnson 2006). Nonetheless, the characterization of mouse models lacking individual E2Fs has revealed that these factors play unique roles in development, tissue homeostasis and tumor formation (Chen et al. 2009a; DeGregori & Johnson 2006; Trimarchi & Lees 2002).

Functional specificity of individual E2F factors is thought to be established through the regulation of distinct sets of target genes. In fact, there is growing evidence that this specificity is achieved by interaction of E2Fs with other proteins or by post-translational modifications (PTMs) on E2Fs or E2F-containing complexes. Much of this evidence has been gathered through proteomic approaches such as yeast two-hybrid screening, twodimensional electrophoresis (2-DE) followed by mass spectrometry (MS) or shotgun proteomics (Figure 1). In this review, the application of proteomics in the study of RB/E2F regulatory pathway is summarized. Results derived from these experiments are expanding our current understanding of the RB/E2F biology in several important ways. They are piecing together the interactions within macromolecular complexes that regulate transcription of E2F target genes. Furthermore, they are helping define the mechanisms underlying RB/E2F–dependent control of cellular physiology and pathology.

#### **2. Identification of proteins that interact with E2Fs**

It has long been recognized that E2F activity is regulated through the association of E2F factors with specific protein partners. In fact, E2F1, the founder member of the family, was first identified as a sequence-specific DNA-binding activity that co-precipitated with RB (Chittenden et al. 1991). Recent development of non-hypothesis driven proteomic approaches has allowed a more extensive analysis of protein-protein interactions in the E2F field. Several methods have been successfully applied in the identification of RB/E2F interacting partners, in particular, yeast two-hybrid screenings and affinity purification coupled to MS.

#### **2.1 Genome-wide yeast two-hybrid interaction screening**

The yeast two-hybrid method is one of the most widely used methods for mapping proteinprotein interactions. In this method, the "bait" protein is typically expressed in yeast as a chimeric protein fused to the DNA-binding domain of a known transcription factor (usually *Gal4*). All other "target" proteins that the bait protein is going to be screened against are expressed within the cell fused to the activation domain of this same transcription factor.

Traditionally, the mammalian E2F family has been divided into "activators" (E2F1-3) and "repressors" (E2F4-8). However, recent *in vivo* data have questioned this oversimplified classification. Indeed, accumulating evidence suggests that most E2Fs can function both as activators as well as repressors, depending on the cellular context (Balciunaite et al. 2005; Iglesias et al. 2004; Infante et al. 2008; Lang et al. 2001; Lee et al. 2011; Ma et al. 2002; Morris et al. 2000; Muller et al. 2001; Polager et al. 2008; Young et al. 2003). Likewise, both oncogenic and tumor suppressor properties have been assigned to these factors (DeGregori & Johnson 2006; Johnson & DeGregori 2006). The mechanisms underlying this bimodal impact of individual E2Fs, and their implication in human cancer development remain to be elucidated. This is a particularly relevant point that needs to be addressed, since strategies based on E2F biology are being devised for the development of anticancer therapies (Kaelin, Jr. 2003). An additional level of complexity in the understanding of E2F function *in vivo*  derives from the considerable functional overlap existing among several E2F members (Chen et al. 2009a; DeGregori & Johnson 2006). Nonetheless, the characterization of mouse models lacking individual E2Fs has revealed that these factors play unique roles in development, tissue homeostasis and tumor formation (Chen et al. 2009a; DeGregori &

Functional specificity of individual E2F factors is thought to be established through the regulation of distinct sets of target genes. In fact, there is growing evidence that this specificity is achieved by interaction of E2Fs with other proteins or by post-translational modifications (PTMs) on E2Fs or E2F-containing complexes. Much of this evidence has been gathered through proteomic approaches such as yeast two-hybrid screening, twodimensional electrophoresis (2-DE) followed by mass spectrometry (MS) or shotgun proteomics (Figure 1). In this review, the application of proteomics in the study of RB/E2F regulatory pathway is summarized. Results derived from these experiments are expanding our current understanding of the RB/E2F biology in several important ways. They are piecing together the interactions within macromolecular complexes that regulate transcription of E2F target genes. Furthermore, they are helping define the mechanisms

It has long been recognized that E2F activity is regulated through the association of E2F factors with specific protein partners. In fact, E2F1, the founder member of the family, was first identified as a sequence-specific DNA-binding activity that co-precipitated with RB (Chittenden et al. 1991). Recent development of non-hypothesis driven proteomic approaches has allowed a more extensive analysis of protein-protein interactions in the E2F field. Several methods have been successfully applied in the identification of RB/E2F interacting partners, in particular, yeast two-hybrid screenings and affinity purification

The yeast two-hybrid method is one of the most widely used methods for mapping proteinprotein interactions. In this method, the "bait" protein is typically expressed in yeast as a chimeric protein fused to the DNA-binding domain of a known transcription factor (usually *Gal4*). All other "target" proteins that the bait protein is going to be screened against are expressed within the cell fused to the activation domain of this same transcription factor.

underlying RB/E2F–dependent control of cellular physiology and pathology.

**2. Identification of proteins that interact with E2Fs** 

**2.1 Genome-wide yeast two-hybrid interaction screening** 

Johnson 2006; Trimarchi & Lees 2002).

coupled to MS.

Fig. 1. A schematic diagram showing proteomic approaches to analyzing regulatory signaling pathways.

The interaction between the bait and target proteins brings into close proximity the DNA binding and activation domains of the transcription factor, resulting in the activation of a reporter gene (Fields & Song 1989). Given that the conditions applied in this methodology are not physiological, some of the detected interactions may not represent true interactions. Consequently, this experimental system is thought to yield a high false positive rate. Consequently, interactions detected by this system need to be further validated in an appropriate physiological system. Despite the mentioned drawbacks, it is also true that the yeast two-hybrid interaction screening provides a method to scrutinize protein-protein interactions within living cells, whereas other approaches measure protein interactions after the complexes have been removed from the cellular environment.

Interaction proteomics has been helpful in exploring the intricate macromolecular interactions established by RB and E2F for the regulation of gene expression. Work from many laboratories has shown that RB mediates transcriptional repression through the recruitment of a large number of co-repressors, resulting in an alteration of chromatin conformation that hinders transcription. Most RB/E2F co-repressors, including histone deacetylases (HDAC1-3), nucleosome remodeling proteins (BRG1), DNA methyl transferases (DNMT1) or RBP1 have been identified through hypothesis-driven classical biochemical approaches (Brehm et al. 1998; Luo et al. 1998; Magnaghi-Jaulin et al. 1998). Interestingly, HBP1 and CtIP/CtBP co-repressors were discovered by yeast two-hybrid screening analyses using the pocket protein p130 as the bait (Meloni et al. 1999a; Tevosian et al. 1997). HBP1, a tumor suppressor member of the HMG family of transcription factors,

Proteomic Approaches to Unraveling the RB/E2F Regulatory Pathway 139

JAB1 Promotes apoptosis Hallstron & Nevins

Activates transcription of TK and represses transcription of p14ARF. Promotes proliferation

B-MYB, MGA n.d. Giangrande et al. 2003

B-MYB n.d. Giangrande et al. 2003 ALIEN n.d. Escher et al. 2007 EAPP n.d. Novy et al. 2005

ALIEN n.d. Escher et al. 2007 EAPP n.d. Novy et al. 2005

MGA n.d. Giangrande et al. 2003

2006

Novy et al. 2005

2006

2006

n.d. Hallstron & Nevins

CDC6 Schlisio et al. 2002

CDC6 Schlisio et al. 2002

n.d. Giangrande et al. 2003

n.d. Hallstron & Nevins

gene of DNA Pol<sup>α</sup> Giangrande et al. 2003

**E2F member Interacting Protein Function Reference** 

RYBP Activates transcription of

RYBP Activates transcription of

TFE3 Activates the p68 subunit

E2F4 ALIEN n.d. Escher et al. 2007

E2F5 ALIEN n.d. Escher et al. 2007 E2F6 ALIEN n.d. Escher et al. 2007

Table 1. E2F-specific binding partners identified through proteomic approaches. n.d.: not

these interactions it has been proposed that transcriptional regulation of specific E2F target genes can only be achieved when relevant interacting proteins act jointly forming a functional complex on the promoter (Freedman et al. 2009). Therefore, the unique marked box of each individual E2F factor appears to play a pivotal role in determining the specificity of

Other proteins exhibit E2F-binding activity independently of the marked box, allowing the mapping of distinct functional domains within the E2Fs. A yeast two-hybrid interaction

EAPP

SKIP, CBP, TEF-5,

MCRS1, E1F1, FHOD1, FLNA, SUI1, A-FABP, MX1, RNF2, PRMT3, RANBP9

RXR-BP, WNK1, PKI-B, SKIP, RYBP, CBP, TEF-5, B-MYB, MGA, SPIB

HSP86, A-FABP, MX1, RNF2, PRMT3, RANBP9

E2F1

E2F2

E2F3

determined

interaction.

was found to function as a transcriptional repressor of N-MYC in association with RB in terminally differentiated muscle cells. This finding implies a role of this complex in the initiation and establishment of cell cycle arrest during differentiation (Tevosian et al. 1997). However, E2F proteins were not found in this repressor complex. By contrast, the complex formed by CtIP/CtBP and p130 also included E2F1, and provided an additional mechanism for RB/E2F-mediated repression (Meloni et al. 1999b). In agreement with the original findings, it has been recently shown that CtIP/CtBP plays a transcriptional co-repressor role in ZBRK1 expression. ZBRK1 is a zinc finger-containing transcriptional repressor that can modulate the expression of GADD45A to induce cell cycle arrest in response to DNA damage (Liao et al. 2010). It has been proposed that the contribution of RB to DNA damageinduced growth arrest may depend on the formation of this complex and loss of CtIP/CtBP– mediated repression could affect the cellular sensitivity to DNA damage. Conversely, CtIP/CtBP is able to activate the expression of a subset of E2F-target genes after its release from RB-imposed repression, implying that it can also function as an activator in other cellular contexts (Liu & Lee 2006).

Yeast two-hybrid screening has been particularly valuable to delve into the mechanistic basis for the functional specificity of E2F factors, particularly the so-called E2F "activators" (E2F1-3). This E2F subfamily exhibits a significant degree of functional redundancy among its members. However, E2F1 appears to be a stronger inducer of apoptosis than E2F2 and E2F3 (DeGregori et al. 1997; Hong et al. 2008; Kowalik et al. 1998; Lazzerini et al. 2005). The predominant role of E2F1 over the other E2F members in triggering apoptosis is thought to be conferred by unique protein partners that E2F1 associates with. The critical domain to specifically induce apoptosis has been shown to lie in the marked box of E2F1 (Hallstrom & Nevins 2003). Taking advantage of this knowledge, the E2F1 marked box has been used by Hallstrom and Nevins as the bait to screen for protein partners that could mediate E2F1 dependent apoptosis. JAB1 (c-JUN activating-binding protein) was identified as an E2F1 specific binding protein that functions synergistically with E2F1 to induce apoptosis coincident with an induction of p53 protein accumulation (Hallstrom & Nevins 2006). Interestingly, JAB1 association appears to regulate exclusively the apoptotic role of E2F1, as cell cycle entry is not affected by this E2F protein partner. In addition to JAB1, several more E2F1-interacting proteins were detected in this screen (Table 1), although their functional relevance in E2F1 function remains to be determined.

Remarkably, the E2F marked box has emerged as an important domain for mediating protein interactions that could dictate specificity of promoter recognition. For example, E2F2 and E2F3, but not E2F1 or E2F4, have been shown to interact specifically with RYBP (Ring-1 and YY1-binding protein) through their marked box. RYBP recruits these E2Fs to target promoters containing YY1 binding sites such as the CDC6 promoter. It has been proposed that the formation of an E2F2/3-RYBP-YY1 complex would facilitate the timely activation of CDC6 (Schlisio et al. 2002). An independent yeast two-hybrid screen with E2F3 as the bait discovered TFE3 (an E-box binding factor) as a protein that specifically interacts with E2F3. This association, which is dependent on the marked box of E2F3, facilitates transcriptional activation of the p68 subunit gene of DNA Polα (Giangrande et al. 2003). Furthermore, this screen also yielded several more proteins that bound specifically the marked box of E2F3 (Table 1). Some of these proteins, such as CBP, RYBP or MGA had previously been shown to interact with E2Fs (Morris et al. 2000; Ogawa et al. 2002; Schlisio et al. 2002; Trouche et al. 1996), providing a strong validation of the screen. By contrast, E2F1, E2F2 and E2F4 are unable to bind TFE3 or to activate transcription of p68. Based on the characterization of all

was found to function as a transcriptional repressor of N-MYC in association with RB in terminally differentiated muscle cells. This finding implies a role of this complex in the initiation and establishment of cell cycle arrest during differentiation (Tevosian et al. 1997). However, E2F proteins were not found in this repressor complex. By contrast, the complex formed by CtIP/CtBP and p130 also included E2F1, and provided an additional mechanism for RB/E2F-mediated repression (Meloni et al. 1999b). In agreement with the original findings, it has been recently shown that CtIP/CtBP plays a transcriptional co-repressor role in ZBRK1 expression. ZBRK1 is a zinc finger-containing transcriptional repressor that can modulate the expression of GADD45A to induce cell cycle arrest in response to DNA damage (Liao et al. 2010). It has been proposed that the contribution of RB to DNA damageinduced growth arrest may depend on the formation of this complex and loss of CtIP/CtBP– mediated repression could affect the cellular sensitivity to DNA damage. Conversely, CtIP/CtBP is able to activate the expression of a subset of E2F-target genes after its release from RB-imposed repression, implying that it can also function as an activator in other

Yeast two-hybrid screening has been particularly valuable to delve into the mechanistic basis for the functional specificity of E2F factors, particularly the so-called E2F "activators" (E2F1-3). This E2F subfamily exhibits a significant degree of functional redundancy among its members. However, E2F1 appears to be a stronger inducer of apoptosis than E2F2 and E2F3 (DeGregori et al. 1997; Hong et al. 2008; Kowalik et al. 1998; Lazzerini et al. 2005). The predominant role of E2F1 over the other E2F members in triggering apoptosis is thought to be conferred by unique protein partners that E2F1 associates with. The critical domain to specifically induce apoptosis has been shown to lie in the marked box of E2F1 (Hallstrom & Nevins 2003). Taking advantage of this knowledge, the E2F1 marked box has been used by Hallstrom and Nevins as the bait to screen for protein partners that could mediate E2F1 dependent apoptosis. JAB1 (c-JUN activating-binding protein) was identified as an E2F1 specific binding protein that functions synergistically with E2F1 to induce apoptosis coincident with an induction of p53 protein accumulation (Hallstrom & Nevins 2006). Interestingly, JAB1 association appears to regulate exclusively the apoptotic role of E2F1, as cell cycle entry is not affected by this E2F protein partner. In addition to JAB1, several more E2F1-interacting proteins were detected in this screen (Table 1), although their functional

Remarkably, the E2F marked box has emerged as an important domain for mediating protein interactions that could dictate specificity of promoter recognition. For example, E2F2 and E2F3, but not E2F1 or E2F4, have been shown to interact specifically with RYBP (Ring-1 and YY1-binding protein) through their marked box. RYBP recruits these E2Fs to target promoters containing YY1 binding sites such as the CDC6 promoter. It has been proposed that the formation of an E2F2/3-RYBP-YY1 complex would facilitate the timely activation of CDC6 (Schlisio et al. 2002). An independent yeast two-hybrid screen with E2F3 as the bait discovered TFE3 (an E-box binding factor) as a protein that specifically interacts with E2F3. This association, which is dependent on the marked box of E2F3, facilitates transcriptional activation of the p68 subunit gene of DNA Polα (Giangrande et al. 2003). Furthermore, this screen also yielded several more proteins that bound specifically the marked box of E2F3 (Table 1). Some of these proteins, such as CBP, RYBP or MGA had previously been shown to interact with E2Fs (Morris et al. 2000; Ogawa et al. 2002; Schlisio et al. 2002; Trouche et al. 1996), providing a strong validation of the screen. By contrast, E2F1, E2F2 and E2F4 are unable to bind TFE3 or to activate transcription of p68. Based on the characterization of all

cellular contexts (Liu & Lee 2006).

relevance in E2F1 function remains to be determined.


Table 1. E2F-specific binding partners identified through proteomic approaches. n.d.: not determined

these interactions it has been proposed that transcriptional regulation of specific E2F target genes can only be achieved when relevant interacting proteins act jointly forming a functional complex on the promoter (Freedman et al. 2009). Therefore, the unique marked box of each individual E2F factor appears to play a pivotal role in determining the specificity of interaction.

Other proteins exhibit E2F-binding activity independently of the marked box, allowing the mapping of distinct functional domains within the E2Fs. A yeast two-hybrid interaction

Proteomic Approaches to Unraveling the RB/E2F Regulatory Pathway 141

A wide range of RB/E2F partners have been identified to date by MS. Remarkably, the discovered proteins are mainly involved in transcriptional repression, which highlights the critical role of the RB/E2F network in the timely suppression of target gene expression. Combining immunological methods with an uncommon mass spectrometric technique named SELDI (surface-enhanced laser desorption/ionization)(Lehmann et al. 2005), corepressor ALIEN was identified in E2F1 containing endogenous protein complexes (Escher et al. 2007). Additional co-immunoprecipitation experiments revealed specific interactions of ALIEN with E2F2 through E2F6. ALIEN is able to repress E2F1 transcriptional activity when tethered to target promoters containing E2F binding sites (Tenbaum et al. 2007). The mechanism for this repressive activity remains undetermined. An RB-dependent process has been proposed, based on the finding that ALIEN interacts with pRB/p107 and HDAC (Escher et al. 2007). However, ALIEN-mediated co-repression is also evident in cells lacking

Affinity purification coupled to MS analysis has been particularly useful in the identification of native RB/E2F transcriptional repressor complexes (Table 2). They were firstly characterized in *Drosophila melanogaster* (Korenjak et al. 2004). Subsequently, homologous complexes were identified in other animal species, including *Caenorhabditis elegans* and human (Harrison et al. 2006; Litovchick et al. 2007), emphasizing the importance of this pathway. Before the application of proteomic methods, a wide assortment of chromatin-modifying and binding complexes had been implicated in RB-mediated repression (Frolov & Dyson 2004). However, it is unclear which of the many reported interactions are biologically meaningful, due to the nonphysiological methods employed for the screening. In an effort to characterize native RB/E2F repressor complexes in *Drosophila*, Brehm´s group took advantage of the relative simplicity of the *Drosophila* dRB/E2F network, consisting of two pocket proteins, RBF1 and RBF2, and two E2F proteins, dE2F1 and dE2F2 that heterodimerize with a common partner, dDP. By classical chromatography followed by MS analysis of the resulting fractions, they isolated an RBF multisubunit complex, termed dREAM (Korenjak et al. 2004), which contains RBF1/2, dE2F2, dDP, dMyb and dMyb-interacting proteins (Mip/TWIT, CAF1p55, Mip40 and Mip120). Interestingly, an independent analysis of Myb-associated proteins, involving affinity chromatography and DALPC mass spectrometry analysis (direct analysis of large protein complexes) resulted in the identification of a similar complex in *Drosophila* that was called

In agreement with a role in transcriptional repression, the identified complexes localize to transcriptionally silent sites on polytene chromosomes and mediate stable repression of a specific set of E2F targets that have sex- and differentiation-specific expression patterns. The mechanism by which these repressive complexes mediate the silencing of their target genes has been controversial, and enzymatic and non-enzymatic modes of repression have been suggested (Korenjak et al. 2004; Lewis et al. 2004). The finding that dREAM only binds deacetylated histone H4, characteristic of repressed chromatin, suggests that dREAM complexes perform their repressive function through binding unmodified nucleosomes and therefore protecting them from activating modifications (Korenjak et al. 2004). This hypothesis was further supported by the identification of histone deacetylase Rpd3, the HDAC1 homolog in *Drosophila*, associated with the repressor complex (Lewis et al. 2004). Strikingly, at least seven, and possibly all dREAM subunits are related to *C. elegans* synMuv class B genes (Fay & Han 2000). These proteins encompass the DRM complex, genetically resolved by Harrison and collaborators, which controls vulval differentiation in the worm (Harrison et al. 2006). The similarity between the worm DRM and the fly dREAM and Myb–

functional RB (Tenbaum et al. 2007).

Myb–MuvB complex (Lewis et al. 2004).

screen with the amino-terminal region of E2F1 comprising the NLS but lacking the marked box led to the cloning of EAPP (E2F-associated phosphoprotein) (Novy et al. 2005). EAPP appears to potentiate the proliferative functions of E2F1 in several ways. It enhances the transcription of growth-correlated E2F-target genes like thymidine kinase whereas it represses the expression of the tumor suppressor p14ARF, an E2F1-target gene that can mediate E2F-induced apoptosis. Of note, EAPP can also interact with E2F2 and E2F3a through their amino-terminal regions (Novy et al. 2005). Consequently, EAPP could mediate proliferation related redundant functions of the E2F "activators" through binding to their well-conserved N-terminal region.

E2F6 together with E2F7 and E2F8 have features that set them apart from other members of the E2F family. These factors lack RB-binding and transactivation domains, and pocket proteins are thought not to complex with E2F6-8 (Cartwright et al. 1998; Gaubatz et al. 1998; Trimarchi et al. 1998). Thus, there is a particular interest in identifying the proteins that interact with these non-classical E2F members and determining their mechanism for transcriptional regulation of target genes. Two independent yeast two-hybrid analyses have described distinct interactions between E2F6 and components of the Polycomb Group (PcG), known mammalian transcriptional repressors (Bunker & Kingston 1994). In the first study, E2F6 was shown to associate with members of the PRC1 complex, including some PcG proteins (RING1, MEL-18, MPH1, and BMI1) as well as RYBP (Trimarchi et al. 2001). The second screening identified the PcG protein EPC1 as a novel E2F6-binding protein, in complex with EZH2 and SIN3B (Attwooll et al. 2005). In both cases, the association of E2F6 with PcG proteins could account for an RB-independent mechanism for the recruitment of repressive complexes to E2F6 target promoters and the consequent transcriptional repression (Trimarchi et al. 2001). It should be mentioned that this work did not detect an interaction between RYBP and E2F2 or E2F3, as shown by Schlisio and co-workers (Schlisio et al. 2002). The basis for this discrepancy is not clear, but could reflect methodological differences. However, the common ability of E2F2, E2F3 and E2F6 to interact with RYBP, suggests that E2F2 and E2F3 may be able to repress a subset of target genes by RBindependent mechanisms.

#### **2.2 Affinity purification coupled to mass spectrometry**

Identification of protein-protein interactions is becoming increasingly easier since the extraordinary advances that have taken place in recent years in mass spectrometry (MS). Affinity purification coupled to mass spectrometry is currently the preferred method for screening protein-protein interactions, owing to the sensitivity, specificity and reliability of this approach (Gavin et al. 2002; Ho et al, 2002; Krogan et al. 2006; Sellers et al. 1998). In contrast to classical biochemical methods, MS-based proteomics is a discovery or explorative non-hypothesis driven science. However, it should be noted that the identification of protein complexes remains a significant challenge. This is because many interactions are transient and can be easily lost during sample preparation. Additionally, many proteins are expressed at low levels, and it remains difficult to purify them. In fact, isolating the protein complex is the most critical step in determining the success of a proteomic analysis. The most widely used method for protein complex isolation is antibody-based immunoprecitation (IP), although MS is also compatible with other affinity purification methods used to map protein-protein interactions, such as oligoprecipitation and tagged protein precipitation by affinity ("protein pulldown")(Gavin et al. 2002; Meng et al. 2006).

screen with the amino-terminal region of E2F1 comprising the NLS but lacking the marked box led to the cloning of EAPP (E2F-associated phosphoprotein) (Novy et al. 2005). EAPP appears to potentiate the proliferative functions of E2F1 in several ways. It enhances the transcription of growth-correlated E2F-target genes like thymidine kinase whereas it represses the expression of the tumor suppressor p14ARF, an E2F1-target gene that can mediate E2F-induced apoptosis. Of note, EAPP can also interact with E2F2 and E2F3a through their amino-terminal regions (Novy et al. 2005). Consequently, EAPP could mediate proliferation related redundant functions of the E2F "activators" through binding to their

E2F6 together with E2F7 and E2F8 have features that set them apart from other members of the E2F family. These factors lack RB-binding and transactivation domains, and pocket proteins are thought not to complex with E2F6-8 (Cartwright et al. 1998; Gaubatz et al. 1998; Trimarchi et al. 1998). Thus, there is a particular interest in identifying the proteins that interact with these non-classical E2F members and determining their mechanism for transcriptional regulation of target genes. Two independent yeast two-hybrid analyses have described distinct interactions between E2F6 and components of the Polycomb Group (PcG), known mammalian transcriptional repressors (Bunker & Kingston 1994). In the first study, E2F6 was shown to associate with members of the PRC1 complex, including some PcG proteins (RING1, MEL-18, MPH1, and BMI1) as well as RYBP (Trimarchi et al. 2001). The second screening identified the PcG protein EPC1 as a novel E2F6-binding protein, in complex with EZH2 and SIN3B (Attwooll et al. 2005). In both cases, the association of E2F6 with PcG proteins could account for an RB-independent mechanism for the recruitment of repressive complexes to E2F6 target promoters and the consequent transcriptional repression (Trimarchi et al. 2001). It should be mentioned that this work did not detect an interaction between RYBP and E2F2 or E2F3, as shown by Schlisio and co-workers (Schlisio et al. 2002). The basis for this discrepancy is not clear, but could reflect methodological differences. However, the common ability of E2F2, E2F3 and E2F6 to interact with RYBP, suggests that E2F2 and E2F3 may be able to repress a subset of target genes by RB-

Identification of protein-protein interactions is becoming increasingly easier since the extraordinary advances that have taken place in recent years in mass spectrometry (MS). Affinity purification coupled to mass spectrometry is currently the preferred method for screening protein-protein interactions, owing to the sensitivity, specificity and reliability of this approach (Gavin et al. 2002; Ho et al, 2002; Krogan et al. 2006; Sellers et al. 1998). In contrast to classical biochemical methods, MS-based proteomics is a discovery or explorative non-hypothesis driven science. However, it should be noted that the identification of protein complexes remains a significant challenge. This is because many interactions are transient and can be easily lost during sample preparation. Additionally, many proteins are expressed at low levels, and it remains difficult to purify them. In fact, isolating the protein complex is the most critical step in determining the success of a proteomic analysis. The most widely used method for protein complex isolation is antibody-based immunoprecitation (IP), although MS is also compatible with other affinity purification methods used to map protein-protein interactions, such as oligoprecipitation and tagged protein precipitation by affinity ("protein pull-

well-conserved N-terminal region.

independent mechanisms.

**2.2 Affinity purification coupled to mass spectrometry** 

down")(Gavin et al. 2002; Meng et al. 2006).

A wide range of RB/E2F partners have been identified to date by MS. Remarkably, the discovered proteins are mainly involved in transcriptional repression, which highlights the critical role of the RB/E2F network in the timely suppression of target gene expression. Combining immunological methods with an uncommon mass spectrometric technique named SELDI (surface-enhanced laser desorption/ionization)(Lehmann et al. 2005), corepressor ALIEN was identified in E2F1 containing endogenous protein complexes (Escher et al. 2007). Additional co-immunoprecipitation experiments revealed specific interactions of ALIEN with E2F2 through E2F6. ALIEN is able to repress E2F1 transcriptional activity when tethered to target promoters containing E2F binding sites (Tenbaum et al. 2007). The mechanism for this repressive activity remains undetermined. An RB-dependent process has been proposed, based on the finding that ALIEN interacts with pRB/p107 and HDAC (Escher et al. 2007). However, ALIEN-mediated co-repression is also evident in cells lacking functional RB (Tenbaum et al. 2007).

Affinity purification coupled to MS analysis has been particularly useful in the identification of native RB/E2F transcriptional repressor complexes (Table 2). They were firstly characterized in *Drosophila melanogaster* (Korenjak et al. 2004). Subsequently, homologous complexes were identified in other animal species, including *Caenorhabditis elegans* and human (Harrison et al. 2006; Litovchick et al. 2007), emphasizing the importance of this pathway. Before the application of proteomic methods, a wide assortment of chromatin-modifying and binding complexes had been implicated in RB-mediated repression (Frolov & Dyson 2004). However, it is unclear which of the many reported interactions are biologically meaningful, due to the nonphysiological methods employed for the screening. In an effort to characterize native RB/E2F repressor complexes in *Drosophila*, Brehm´s group took advantage of the relative simplicity of the *Drosophila* dRB/E2F network, consisting of two pocket proteins, RBF1 and RBF2, and two E2F proteins, dE2F1 and dE2F2 that heterodimerize with a common partner, dDP. By classical chromatography followed by MS analysis of the resulting fractions, they isolated an RBF multisubunit complex, termed dREAM (Korenjak et al. 2004), which contains RBF1/2, dE2F2, dDP, dMyb and dMyb-interacting proteins (Mip/TWIT, CAF1p55, Mip40 and Mip120). Interestingly, an independent analysis of Myb-associated proteins, involving affinity chromatography and DALPC mass spectrometry analysis (direct analysis of large protein complexes) resulted in the identification of a similar complex in *Drosophila* that was called Myb–MuvB complex (Lewis et al. 2004).

In agreement with a role in transcriptional repression, the identified complexes localize to transcriptionally silent sites on polytene chromosomes and mediate stable repression of a specific set of E2F targets that have sex- and differentiation-specific expression patterns. The mechanism by which these repressive complexes mediate the silencing of their target genes has been controversial, and enzymatic and non-enzymatic modes of repression have been suggested (Korenjak et al. 2004; Lewis et al. 2004). The finding that dREAM only binds deacetylated histone H4, characteristic of repressed chromatin, suggests that dREAM complexes perform their repressive function through binding unmodified nucleosomes and therefore protecting them from activating modifications (Korenjak et al. 2004). This hypothesis was further supported by the identification of histone deacetylase Rpd3, the HDAC1 homolog in *Drosophila*, associated with the repressor complex (Lewis et al. 2004). Strikingly, at least seven, and possibly all dREAM subunits are related to *C. elegans* synMuv class B genes (Fay & Han 2000). These proteins encompass the DRM complex, genetically resolved by Harrison and collaborators, which controls vulval differentiation in the worm (Harrison et al. 2006). The similarity between the worm DRM and the fly dREAM and Myb–

Proteomic Approaches to Unraveling the RB/E2F Regulatory Pathway 143

lacks MYB. In fact, MYB immunoprecipitates with LIN9, LIN37, and LIN54, but not with E2F or p130 in human cells, indicating the existence of two distinct transcriptional complexes incorporating either p130 or MYB. Indeed, several reports have demonstrated that the complex composition is dynamic and cell cycle phase-dependent. In quiescent cells LIN complex proteins associate with p130 and E2F4, and the complex binds to its target genes. In late G1/S phase, E2F4 and p130 dissociate and the LIN complex associates with MYB (Litovchick et al. 2007; Schmit et al. 2007), forming a new complex that is required for activation of G2/M genes. The dynamic interaction of the LIN complex with different DNA-binding proteins (E2F4 *vs*. MYB) supports a model in which DREAM is bound in G0 to E2F-regulated promoters via E2F4/p130 and in S-phase to G2/M promoters via MYB. Another finding that distinguishes the human complex from its fly and worm orthologs is that human DREAM does not appear to regulate target genes involved in development

Mass spectrometry has also been applied to elucidate the mechanism by which E2F6 regulates transcriptional repression in quiescent cells. E2F6 was shown to interact directly with MGA and MAX (Ogawa et al. 2002). These proteins are known to bind as heterodimers to E boxes such as the MYC binding elements, and to antagonize MYC function (Hurlin et al. 1999). Given that MYC and E2F share common functions such as mitotic responses, cell cycle stimulation and induction of apoptosis, it has been proposed that E2F- and MYCresponsive genes could be co-regulated by E2F6 (Ogawa et al. 2002). In the complex containing E2F6-MGA-MAX, several Polycomb group proteins (RING1/2, MBLR, hl(3)MBT-like protein, and YAF2) are also present, together with HP1γ, a methyltransferase related to gene silencing in euchromatic loci (Horsley et al. 1996). The recruitment of E2F6 to its target genes could form a "platform" that is required for nucleating PcG proteins. Subsequent recruitment of HP1γ to this platform would propagate chromatin inactivation, leading to entire repressed regions (Ogawa et al. 2002). It is remarkable that as many as three types of E2F6 repressor complexes containing different PcG proteins have been identified by proteomic methods (Table 2). The biological relevance for such diversity has not been clarified, although there is evidence that these complexes could be formed at different phases throughout the cell cycle. MAX and HP1γ were found associated with E2F target promoters in G0, but not following re-entry into the cell cycle, suggesting a role for this complex in gene repression in quiescent cells (Ogawa et al. 2002). By contrast, the PcG protein EZH2 only forms complexes with E2F6-EPC1 in proliferating cells, suggesting that this complex regulates genes required for cell cycle progression (Attwooll et al. 2005). Deciphering the protein interactome of upstream components of the RB/E2F pathway has also led to the characterization of new functions of the pathway. This is the case of a recently work published by Sicinski´s group in which cyclin D1 protein partners were characterized by immunoaffinity purification coupled to MS analysis (Jirawatnotai et al. 2011). Identification of cyclin D1 interactors revealed a network of DNA repair proteins, including RAD51, a key DNA recombinase that drives the homologous recombination process in response to DNA damage (Baumann & West 1998). Remarkably, the finding that Cyclin D1´s function in DNA repair appears to be independent of its kinase activity could have clinical applications. For instance, a large pool of RB-negative cancers, which do not require D-cyclins for proliferation, may still be benefited by therapeutic interventions targeting cyclin D1 in combination with radiation treatment. This work clearly shows how interacting-proteomics can fundamentally change our understanding of signaling networks

(Litovchick et al. 2007).


Table 2. Multiprotein complexes encompassing E2F factors identified by proteomic approaches.

MuvB complexes indicates that the DRM complex likely acts in transcriptional repression of E2F targets, and implies a remarkable conservation in the mechanism of pocket protein function across species.

An integration of proteomic, genomic, and bioinformatic approaches has allowed DeCaprio's group the identification and functional characterization of human DREAM, the homolog of the fly dREAM (Litovchick et al. 2007). To determine the composition of this complex the authors purified p130-associated proteins and applied a multidimensional protein identification method (MudPIT). This method couples biphasic and triphasic microcapillary columns to high performance liquid chromatography followed by tandem mass spectrometry analysis (Florens & Washburn 2006). The core components of the identified complex include p130, E2F4/5, DP1/2 and mammalian orthologs of synMuvB proteins LIN9, LIN37, LIN52, LIN54, and LIN53/RBBP4 (also known as LIN complex). Interestingly, the human DREAM repressor complex differs from the fly complex in that it

**E2F member Complex Function Reference** 

Transcriptional repression of developmentally regulated genes

Transcriptional repression of developmentally regulated genes

Repression of ZBRK1. RB-dependent growth arrest after DNA damage

Repression of E2F1 expression and cellular proliferation

Transcriptional repression of cell cycle genes in G0

Target gene repression

Target gene repression

Korenjak et al. 2004

Lewis et al. 2004

Meloni et al. 1999 Liao Ching-Chun, 2010

Escher et al. 2007 Tenbaum et al. 2007

Litovchick et al. 2007

2001

not determined Trimarchi et al.

in proliferating cells Attwoll et al. 2005

in quiescent cells Ogawa et al. 2002

dREAM: RBF1/2, dDP,dMyb, Mip/TWIT,CAF1p55, Mip40 , Mip120, Rpd3

Myb-MuvB: RBF1/2, dDP,dMyb, Mip40, Caf1p55, Mip130 , Mip120, dLin52, Rpd3, L(3)MBT

CtIP/CtBP, p130

ALIEN, pRB/p107, CDK2

DREAM: p130, LIN9, LIN37, LIN52, LIN54 LIN53/RBBP4

Complex: RYBP, RING1, MEL-18, MPH1, BMI1

Complex: EPC1, EZH2, SIN3B

Complex: MGA, MAX RING1/2, MBLR, h-l(3)MBT-like, YAF2, HP1γ

Table 2. Multiprotein complexes encompassing E2F factors identified by proteomic

MuvB complexes indicates that the DRM complex likely acts in transcriptional repression of E2F targets, and implies a remarkable conservation in the mechanism of pocket protein

An integration of proteomic, genomic, and bioinformatic approaches has allowed DeCaprio's group the identification and functional characterization of human DREAM, the homolog of the fly dREAM (Litovchick et al. 2007). To determine the composition of this complex the authors purified p130-associated proteins and applied a multidimensional protein identification method (MudPIT). This method couples biphasic and triphasic microcapillary columns to high performance liquid chromatography followed by tandem mass spectrometry analysis (Florens & Washburn 2006). The core components of the identified complex include p130, E2F4/5, DP1/2 and mammalian orthologs of synMuvB proteins LIN9, LIN37, LIN52, LIN54, and LIN53/RBBP4 (also known as LIN complex). Interestingly, the human DREAM repressor complex differs from the fly complex in that it

Drosophila E2F2

E2F1

E2F4/5

E2F6

function across species.

approaches.

lacks MYB. In fact, MYB immunoprecipitates with LIN9, LIN37, and LIN54, but not with E2F or p130 in human cells, indicating the existence of two distinct transcriptional complexes incorporating either p130 or MYB. Indeed, several reports have demonstrated that the complex composition is dynamic and cell cycle phase-dependent. In quiescent cells LIN complex proteins associate with p130 and E2F4, and the complex binds to its target genes. In late G1/S phase, E2F4 and p130 dissociate and the LIN complex associates with MYB (Litovchick et al. 2007; Schmit et al. 2007), forming a new complex that is required for activation of G2/M genes. The dynamic interaction of the LIN complex with different DNA-binding proteins (E2F4 *vs*. MYB) supports a model in which DREAM is bound in G0 to E2F-regulated promoters via E2F4/p130 and in S-phase to G2/M promoters via MYB. Another finding that distinguishes the human complex from its fly and worm orthologs is that human DREAM does not appear to regulate target genes involved in development (Litovchick et al. 2007).

Mass spectrometry has also been applied to elucidate the mechanism by which E2F6 regulates transcriptional repression in quiescent cells. E2F6 was shown to interact directly with MGA and MAX (Ogawa et al. 2002). These proteins are known to bind as heterodimers to E boxes such as the MYC binding elements, and to antagonize MYC function (Hurlin et al. 1999). Given that MYC and E2F share common functions such as mitotic responses, cell cycle stimulation and induction of apoptosis, it has been proposed that E2F- and MYCresponsive genes could be co-regulated by E2F6 (Ogawa et al. 2002). In the complex containing E2F6-MGA-MAX, several Polycomb group proteins (RING1/2, MBLR, hl(3)MBT-like protein, and YAF2) are also present, together with HP1γ, a methyltransferase related to gene silencing in euchromatic loci (Horsley et al. 1996). The recruitment of E2F6 to its target genes could form a "platform" that is required for nucleating PcG proteins. Subsequent recruitment of HP1γ to this platform would propagate chromatin inactivation, leading to entire repressed regions (Ogawa et al. 2002). It is remarkable that as many as three types of E2F6 repressor complexes containing different PcG proteins have been identified by proteomic methods (Table 2). The biological relevance for such diversity has not been clarified, although there is evidence that these complexes could be formed at different phases throughout the cell cycle. MAX and HP1γ were found associated with E2F target promoters in G0, but not following re-entry into the cell cycle, suggesting a role for this complex in gene repression in quiescent cells (Ogawa et al. 2002). By contrast, the PcG protein EZH2 only forms complexes with E2F6-EPC1 in proliferating cells, suggesting that this complex regulates genes required for cell cycle progression (Attwooll et al. 2005).

Deciphering the protein interactome of upstream components of the RB/E2F pathway has also led to the characterization of new functions of the pathway. This is the case of a recently work published by Sicinski´s group in which cyclin D1 protein partners were characterized by immunoaffinity purification coupled to MS analysis (Jirawatnotai et al. 2011). Identification of cyclin D1 interactors revealed a network of DNA repair proteins, including RAD51, a key DNA recombinase that drives the homologous recombination process in response to DNA damage (Baumann & West 1998). Remarkably, the finding that Cyclin D1´s function in DNA repair appears to be independent of its kinase activity could have clinical applications. For instance, a large pool of RB-negative cancers, which do not require D-cyclins for proliferation, may still be benefited by therapeutic interventions targeting cyclin D1 in combination with radiation treatment. This work clearly shows how interacting-proteomics can fundamentally change our understanding of signaling networks

Proteomic Approaches to Unraveling the RB/E2F Regulatory Pathway 145

hypothesis-driven, that is, the identification of the PTMs was, to an extent, a targeted approach. Moreover, each work was focused on individual proteins where a single or a few modifications were described at a time. Consequently, the information about PTMs affecting the biological function of E2Fs has been considerably limited. Indeed, these classical approaches have led to the identification and characterization of merely 7 phosphorylations on E2F1 over a period of 10 years, all of which are registered in the Phosphosite database (Hornbeck et al. 2004). Phosphorylation of E2F1 at Ser332 and Ser337 disrupts its interaction with RB to become transcriptionally active (Fagan et al. 1994). By contrast, phosphorylation at Ser375 increases its ability to interact with RB and, consequently, decreases its DNA binding capacity (Peeper et al. 1995). Phosphorylation of E2F1 residues Ser403 and Thr433 by TFIIH kinase in S-phase triggers rapid degradation of the transcription factor (Vandel & Kouzarides 1999). Finally, two residues on E2F1, Ser31 and Ser364, can be phosphorylated in response to DNA damage by ATM/ATR and Chk1 kinases, respectively. Both phosphorylations are involved in E2F1-mediated apoptosis (Lin et al. 2001; Stevens et al. 2003). Phosphorylation of E2F3a at Ser124 and its role in DNA-damage induced apoptosis

has also been described by classical biochemical approaches (Martinez et al. 2010).

Recent advances in mass spectrometry have largely overcome the disadvantages faced by classical biochemical approaches allowing the large-scale identification of phosphorylated peptides and the more challenging localization of them within the peptide sequence. In order to decipher the biological meaning of site-specific phosphorylation events in the regulation of cellular processes, quantitative phosphoproteomics studies have produced a considerable body of work during the last 5 years. As a result, thousands of new phosphorylations have been identified (Huttlin et al. 2010; Monetti et al. 2011), some of them

Due to the low stoichiometry of site-specific phosphorylation, refinement of phosphopeptide enrichment methodologies has been crucial for the fruitful characterization of this type of PTM by MS. Currently, there are several phosphopeptide enrichment methods available, such as strong cation exchange chromatography (SCX) (Beausoleil et al. 2004), hydrophilic interaction chromatography (HILIC) (Boersema et al. 2007), immobilized metal affinity chromatography (IMAC) (Li & Dass 1999) and titanium dioxide enrichment method (TiO2) (Larsen et al. 2005; Pinkse et al. 2004). They all manage to separate phosphorylated peptides from nonphosphorylated ones taking advantage of the properties that the negative charge of the phosphate groups confer to phosphopeptides. Antibody-based enrichment approaches have also been demonstrated to be very efficient and useful for phosphoprotein/phosphopeptide enrichment (Choudhary et al. 2009). For optimal analysis of the phosphoproteome these methods are usually combined in multi-stage enrichment strategies (Thingholm et al. 2008). In addition, implementation of quantitation methods has opened new perspectives in studies of complex and dynamic biological signaling networks. In the past decade, highly accurate techniques based on stable isotope labeling for protein quantitation by MS have been developed. Stable isotope labeling of amino acids in cell culture (SILAC) (Ong et al. 2002) and iTRAQ (Ross et al. 2004) are currently the most frequently used techniques in quantitative MS-

The first global *in vivo* SILAC-based quantitative phosphoproteomic study to be reported was performed by Mann's group. It combined SCX and TiO2 enrichment of phosphopeptides followed by LC-MS/MS analysis to quantify changes in phosphopeptide

**3.2 MS-based identification of phosphorylations** 

corresponding to E2F family members.

based phosphoproteomics.

and how combining multiple approaches is possible to unveil novel therapeutic strategies to different disorders such as oncogenic malignancies.

#### **3. Post-translational modifications on E2F transcription factors**

Post-translational modifications (PTMs) are gene non-template chemical modifications occurring at distinct amino acid side chains of proteins. PTMs can change the size, chargestate, structure or conformation of proteins. As a result, PTMs influence several aspects that directly or indirectly affect protein function. Sub-cellular localization (Mueller et al. 2009), protein half-life (Min et al. 2010) or binding affinity to other molecules such as nucleic acids, lipids or other proteins (Takasaki et al. 1999) are usually conditioned by these modifications. Consequently, PTMs play a key role in functional proteomics and are involved in the modulation of almost every single cellular process. Not surprisingly, malfunctions on these critical cellular processes have usually been related to several diseases (Ko et al. 2010; Song et al. 2010). More than 300 types of PTMs have been described to date and the number and variety of identified modifications is continuously increasing (Zhao & Jensen 2009). Some modifications such as phosphorylation, acetylation, glycosylation and methylation are very common and are found in almost every protein while others, for example biotinylation, are very rare. Given the high abundance, dynamism and diversity of PTMs, they likely constitute one of the most complex regulatory mechanisms in eukaryotic cells.

The E2F transcription factors are not an exception to such modifications, as they are known to be regulated by a variety of PTMs. In the mid 90's *in vitro* assays showed that ubiquitination occurring at the C-terminal end of E2F1 is determinant in mediating cell cycle-dependent degradation of this transcription factor by the ubiquitin/proteasome pathway (Hofmann et al. 1996). Additionally, two independent studies concluded that E2F1- 3 but not E2F4-6 can be acetylated by p300/CBP acetyltransferases. Further *in vitro* acetylation assays using deletion mutants of E2F1, restricted the site of acetylation to three conserved lysine residues located at the N-terminal region of its DNA binding domain. This modification could be reversed by HDAC1, and was found to influence the DNA-binding ability of E2F1 and consequently its transcriptional activity (Martinez-Balbas et al. 2000; Marzio et al. 2000). More recently, Set9 methyltransferase and LSD1 demethylase have been shown to regulate E2F1-mediated p53-independent cell death (Kontaki & Talianidis 2010; Xie et al. 2011). Interestingly, a complex cross-talk between different PTMs was found on E2F1. Precisely, Set9-catalyzed methylation at Lys185 inhibits acetylation and phosphorylation events involved in the stabilization of E2F1, while ubiquitination, known to favor protein degradation, is stimulated (Kontaki & Talianidis 2010). Although these diverse PTMs have been described to affect the biological functions of E2Fs, phosphorylation has been the most studied modification.

#### **3.1 Biochemical analysis of phosphorylations**

Phosphorylation of proteins is a highly dynamic process involved in the regulation of several essential cellular functions such as cell cycle regulation, differentiation and signal transduction. Given its biological relevance, the identification and kinetic studies of protein phosphorylation have become a major challenge in molecular biology. Pioneer works describing phosphorylation events on E2F were mainly based on results obtained from electrophoretic mobility shift assays (Peeper et al. 1995) and *in vitro* kinase assays (Lin et al. 2001; Vandel & Kouzarides 1999). A major common feature of these studies is that they were

and how combining multiple approaches is possible to unveil novel therapeutic strategies to

Post-translational modifications (PTMs) are gene non-template chemical modifications occurring at distinct amino acid side chains of proteins. PTMs can change the size, chargestate, structure or conformation of proteins. As a result, PTMs influence several aspects that directly or indirectly affect protein function. Sub-cellular localization (Mueller et al. 2009), protein half-life (Min et al. 2010) or binding affinity to other molecules such as nucleic acids, lipids or other proteins (Takasaki et al. 1999) are usually conditioned by these modifications. Consequently, PTMs play a key role in functional proteomics and are involved in the modulation of almost every single cellular process. Not surprisingly, malfunctions on these critical cellular processes have usually been related to several diseases (Ko et al. 2010; Song et al. 2010). More than 300 types of PTMs have been described to date and the number and variety of identified modifications is continuously increasing (Zhao & Jensen 2009). Some modifications such as phosphorylation, acetylation, glycosylation and methylation are very common and are found in almost every protein while others, for example biotinylation, are very rare. Given the high abundance, dynamism and diversity of PTMs, they likely

**3. Post-translational modifications on E2F transcription factors** 

constitute one of the most complex regulatory mechanisms in eukaryotic cells.

The E2F transcription factors are not an exception to such modifications, as they are known to be regulated by a variety of PTMs. In the mid 90's *in vitro* assays showed that ubiquitination occurring at the C-terminal end of E2F1 is determinant in mediating cell cycle-dependent degradation of this transcription factor by the ubiquitin/proteasome pathway (Hofmann et al. 1996). Additionally, two independent studies concluded that E2F1- 3 but not E2F4-6 can be acetylated by p300/CBP acetyltransferases. Further *in vitro* acetylation assays using deletion mutants of E2F1, restricted the site of acetylation to three conserved lysine residues located at the N-terminal region of its DNA binding domain. This modification could be reversed by HDAC1, and was found to influence the DNA-binding ability of E2F1 and consequently its transcriptional activity (Martinez-Balbas et al. 2000; Marzio et al. 2000). More recently, Set9 methyltransferase and LSD1 demethylase have been shown to regulate E2F1-mediated p53-independent cell death (Kontaki & Talianidis 2010; Xie et al. 2011). Interestingly, a complex cross-talk between different PTMs was found on E2F1. Precisely, Set9-catalyzed methylation at Lys185 inhibits acetylation and phosphorylation events involved in the stabilization of E2F1, while ubiquitination, known to favor protein degradation, is stimulated (Kontaki & Talianidis 2010). Although these diverse PTMs have been described to affect the biological functions of E2Fs, phosphorylation has

Phosphorylation of proteins is a highly dynamic process involved in the regulation of several essential cellular functions such as cell cycle regulation, differentiation and signal transduction. Given its biological relevance, the identification and kinetic studies of protein phosphorylation have become a major challenge in molecular biology. Pioneer works describing phosphorylation events on E2F were mainly based on results obtained from electrophoretic mobility shift assays (Peeper et al. 1995) and *in vitro* kinase assays (Lin et al. 2001; Vandel & Kouzarides 1999). A major common feature of these studies is that they were

different disorders such as oncogenic malignancies.

been the most studied modification.

**3.1 Biochemical analysis of phosphorylations** 

hypothesis-driven, that is, the identification of the PTMs was, to an extent, a targeted approach. Moreover, each work was focused on individual proteins where a single or a few modifications were described at a time. Consequently, the information about PTMs affecting the biological function of E2Fs has been considerably limited. Indeed, these classical approaches have led to the identification and characterization of merely 7 phosphorylations on E2F1 over a period of 10 years, all of which are registered in the Phosphosite database (Hornbeck et al. 2004). Phosphorylation of E2F1 at Ser332 and Ser337 disrupts its interaction with RB to become transcriptionally active (Fagan et al. 1994). By contrast, phosphorylation at Ser375 increases its ability to interact with RB and, consequently, decreases its DNA binding capacity (Peeper et al. 1995). Phosphorylation of E2F1 residues Ser403 and Thr433 by TFIIH kinase in S-phase triggers rapid degradation of the transcription factor (Vandel & Kouzarides 1999). Finally, two residues on E2F1, Ser31 and Ser364, can be phosphorylated in response to DNA damage by ATM/ATR and Chk1 kinases, respectively. Both phosphorylations are involved in E2F1-mediated apoptosis (Lin et al. 2001; Stevens et al. 2003). Phosphorylation of E2F3a at Ser124 and its role in DNA-damage induced apoptosis has also been described by classical biochemical approaches (Martinez et al. 2010).

#### **3.2 MS-based identification of phosphorylations**

Recent advances in mass spectrometry have largely overcome the disadvantages faced by classical biochemical approaches allowing the large-scale identification of phosphorylated peptides and the more challenging localization of them within the peptide sequence. In order to decipher the biological meaning of site-specific phosphorylation events in the regulation of cellular processes, quantitative phosphoproteomics studies have produced a considerable body of work during the last 5 years. As a result, thousands of new phosphorylations have been identified (Huttlin et al. 2010; Monetti et al. 2011), some of them corresponding to E2F family members.

Due to the low stoichiometry of site-specific phosphorylation, refinement of phosphopeptide enrichment methodologies has been crucial for the fruitful characterization of this type of PTM by MS. Currently, there are several phosphopeptide enrichment methods available, such as strong cation exchange chromatography (SCX) (Beausoleil et al. 2004), hydrophilic interaction chromatography (HILIC) (Boersema et al. 2007), immobilized metal affinity chromatography (IMAC) (Li & Dass 1999) and titanium dioxide enrichment method (TiO2) (Larsen et al. 2005; Pinkse et al. 2004). They all manage to separate phosphorylated peptides from nonphosphorylated ones taking advantage of the properties that the negative charge of the phosphate groups confer to phosphopeptides. Antibody-based enrichment approaches have also been demonstrated to be very efficient and useful for phosphoprotein/phosphopeptide enrichment (Choudhary et al. 2009). For optimal analysis of the phosphoproteome these methods are usually combined in multi-stage enrichment strategies (Thingholm et al. 2008). In addition, implementation of quantitation methods has opened new perspectives in studies of complex and dynamic biological signaling networks. In the past decade, highly accurate techniques based on stable isotope labeling for protein quantitation by MS have been developed. Stable isotope labeling of amino acids in cell culture (SILAC) (Ong et al. 2002) and iTRAQ (Ross et al. 2004) are currently the most frequently used techniques in quantitative MSbased phosphoproteomics.

The first global *in vivo* SILAC-based quantitative phosphoproteomic study to be reported was performed by Mann's group. It combined SCX and TiO2 enrichment of phosphopeptides followed by LC-MS/MS analysis to quantify changes in phosphopeptide

Proteomic Approaches to Unraveling the RB/E2F Regulatory Pathway 147

Fig. 2. Phosphorylations detected in human E2F family members. In black phosphorylations detected by classical biochemical methods; in red, phosphorylations detected by MS; in bold red, phosphorylations detected by both methods. S: Serine; T: Threonine; Y: Tyrosine.

Novel methodological approaches in SILAC-based quantitative phosphoproteomics include a double phosphopeptide enrichment using IMAC followed by HILIC. This kind of strategy has allowed the identification of pSer7 on E2F8 (Yao et al. 2011). Furthermore, quantitative data indicated that protein SUMOylation influences phosphorylation detected at residue Ser7, since inhibition of SUMOylation was found to increase the amount of pSer7. This new evidence of cross-talk between different PTMs underscores once more the complexity of

A large number of potential E2F targets have been discovered recently through the application of high throughput transcriptomic analyses. However, mRNA abundance poorly correlates with protein levels. This discrepancy can in part be explained by several post-transcriptional mechanisms such as alternative splicing of pre-mRNAs, microRNA mediated regulation, or selective degradation and post–translational modifications of proteins. Therefore, protein expression analyses should be carried out in addition to mRNA

PTMs and the relevance of the combinatorial PTM patterns.

expression analyses to achieve a full picture of E2F regulated pathways.

**4. Protein profiling to elucidate E2F function** 

levels in response to EGF treatment (Olsen et al. 2006). From the 6,600 phosphosites corresponding to 2,244 proteins that were detected in this study, the previously described E2F1 phosphoserine-375 (pSer375) (Peeper et al. 1995) was identified. This large-scale analysis showed that the phosphorylation state of E2F1 Ser375 is independent of EGF treatment, since the amount of phosphopeptide remained constant over time after the addition of the stimulus (Olsen et al. 2006).

Similar studies carried out by Dephoure and colleagues combined SILAC-based quantitation with SCX followed by TiO2 and IMAC to assess the quantitative atlas of mitotic phosphorylation on HeLa cells. MS analysis of enriched phosphopeptides revealed the presence of E2F1 pSer375 in G1 phase cell lysates. In addition, two new phosphorylations were identified on E2F2 and a single one on E2F7. E2F2 residues Ser133 and Tyr130 were shown to be phosphorylated during the M phase of the cell cycle while E2F7 Ser410 was found to be phosphorylated in G1 phase (Dephoure et al. 2008).

Additional quantitative phosphoproteomic studies have reported the identification of a large number of phosphoresidues on various members of the E2F family, some of which had not been reported before (Figure 2). To decipher the phosphorylation dynamics occurring during early differentiation of human embryonic stem cells, a SILAC-based quantitation combined with SCX and TiO2 followed by MS analysis was performed (Van et al. 2009). Of the 5,222 identified proteins, 1,399 were phosphorylated in 3,067 residues. Two phosphorylated residues, pSer16 and pThr14 precisely, corresponded to E2F4 and were first described in this work. Quantitative phosphoproteomic analysis of T cell receptor signaling revealed the existence of 2 new phosphorylations on E2F family members. Apart from the phosphorylations at Ser375 and Ser332, previously described to affect E2F1 binding affinity for RB (Fagan et al., 1994; Peeper et al., 1995), a new phosphorylated serine at position 340 was detected on E2F1. Moreover, phosphorylation on E2F2 Ser117 was also observed in this analysis (Mayya et al. 2009).

A quantitative phosphoproteomic analysis focused on dissecting the rapamycin-dependent activation of oncogenic cascades also identified several new phosphorylation sites on E2Fs (Chen et al. 2009b). Differentially SILAC labeled cells were cultured in the presence or absence of rapamycin, and stimulated with EGF. Both protein lysates were combined, digested and phosphopeptides were subjected to double IMAC purification followed by MudPIT LC-MS/MS analysis. Of the 6,175 phosphosites identified in this screening, 3 corresponded to E2F8. For the first time, E2F8 Ser355, Ser357 and Ser358 were found phosphorylated. Quantitative phosphoproteomics was also applied in mouse cells to unravel the signaling cascade initiated in response to the oncogenic mutant receptor tyrosine kinase FIt3. This kinase has been involved in the development of several hematopoietic malignancies (Stirewalt & Radich 2003). A total of 14,700 phosphosites were identified, of which 1 corresponded to E2F7 (pTyr416) and 5 to E2F8 (pSer20, pSer71, pSer102, pSer358 and pSer429) (Choudhary et al. 2009). The Tyr416 residue of mouse E2F7 is not conserved in humans. By contrast, 4 of the 5 novel murine E2F8 phosphosites identified in this screening are conserved in humans, suggesting that they may be biologically relevant (Choudhary et al. 2009). In fact, pSer71 of E2F8 has recently been detected in the phosphoproteome of human embryonic stem cells undergoing differentiation (Rigbolt et al. 2011). In this report, an additional phosphosite on E2F8, pSer316, was described. Recently, in an attempt to elucidate the mechanism that regulates the switch of the MuvB core from B-MYB to DREAM, phosphorylation of E2F4 at Ser384 was detected. This phosphorylation could contribute to entry of cells into quiescence (Litovchick et al. 2007).

levels in response to EGF treatment (Olsen et al. 2006). From the 6,600 phosphosites corresponding to 2,244 proteins that were detected in this study, the previously described E2F1 phosphoserine-375 (pSer375) (Peeper et al. 1995) was identified. This large-scale analysis showed that the phosphorylation state of E2F1 Ser375 is independent of EGF treatment, since the amount of phosphopeptide remained constant over time after the

Similar studies carried out by Dephoure and colleagues combined SILAC-based quantitation with SCX followed by TiO2 and IMAC to assess the quantitative atlas of mitotic phosphorylation on HeLa cells. MS analysis of enriched phosphopeptides revealed the presence of E2F1 pSer375 in G1 phase cell lysates. In addition, two new phosphorylations were identified on E2F2 and a single one on E2F7. E2F2 residues Ser133 and Tyr130 were shown to be phosphorylated during the M phase of the cell cycle while E2F7 Ser410 was

Additional quantitative phosphoproteomic studies have reported the identification of a large number of phosphoresidues on various members of the E2F family, some of which had not been reported before (Figure 2). To decipher the phosphorylation dynamics occurring during early differentiation of human embryonic stem cells, a SILAC-based quantitation combined with SCX and TiO2 followed by MS analysis was performed (Van et al. 2009). Of the 5,222 identified proteins, 1,399 were phosphorylated in 3,067 residues. Two phosphorylated residues, pSer16 and pThr14 precisely, corresponded to E2F4 and were first described in this work. Quantitative phosphoproteomic analysis of T cell receptor signaling revealed the existence of 2 new phosphorylations on E2F family members. Apart from the phosphorylations at Ser375 and Ser332, previously described to affect E2F1 binding affinity for RB (Fagan et al., 1994; Peeper et al., 1995), a new phosphorylated serine at position 340 was detected on E2F1. Moreover, phosphorylation on E2F2 Ser117 was also observed in this

A quantitative phosphoproteomic analysis focused on dissecting the rapamycin-dependent activation of oncogenic cascades also identified several new phosphorylation sites on E2Fs (Chen et al. 2009b). Differentially SILAC labeled cells were cultured in the presence or absence of rapamycin, and stimulated with EGF. Both protein lysates were combined, digested and phosphopeptides were subjected to double IMAC purification followed by MudPIT LC-MS/MS analysis. Of the 6,175 phosphosites identified in this screening, 3 corresponded to E2F8. For the first time, E2F8 Ser355, Ser357 and Ser358 were found phosphorylated. Quantitative phosphoproteomics was also applied in mouse cells to unravel the signaling cascade initiated in response to the oncogenic mutant receptor tyrosine kinase FIt3. This kinase has been involved in the development of several hematopoietic malignancies (Stirewalt & Radich 2003). A total of 14,700 phosphosites were identified, of which 1 corresponded to E2F7 (pTyr416) and 5 to E2F8 (pSer20, pSer71, pSer102, pSer358 and pSer429) (Choudhary et al. 2009). The Tyr416 residue of mouse E2F7 is not conserved in humans. By contrast, 4 of the 5 novel murine E2F8 phosphosites identified in this screening are conserved in humans, suggesting that they may be biologically relevant (Choudhary et al. 2009). In fact, pSer71 of E2F8 has recently been detected in the phosphoproteome of human embryonic stem cells undergoing differentiation (Rigbolt et al. 2011). In this report, an additional phosphosite on E2F8, pSer316, was described. Recently, in an attempt to elucidate the mechanism that regulates the switch of the MuvB core from B-MYB to DREAM, phosphorylation of E2F4 at Ser384 was detected. This phosphorylation

could contribute to entry of cells into quiescence (Litovchick et al. 2007).

addition of the stimulus (Olsen et al. 2006).

analysis (Mayya et al. 2009).

found to be phosphorylated in G1 phase (Dephoure et al. 2008).

Fig. 2. Phosphorylations detected in human E2F family members. In black phosphorylations detected by classical biochemical methods; in red, phosphorylations detected by MS; in bold red, phosphorylations detected by both methods. S: Serine; T: Threonine; Y: Tyrosine.

Novel methodological approaches in SILAC-based quantitative phosphoproteomics include a double phosphopeptide enrichment using IMAC followed by HILIC. This kind of strategy has allowed the identification of pSer7 on E2F8 (Yao et al. 2011). Furthermore, quantitative data indicated that protein SUMOylation influences phosphorylation detected at residue Ser7, since inhibition of SUMOylation was found to increase the amount of pSer7. This new evidence of cross-talk between different PTMs underscores once more the complexity of PTMs and the relevance of the combinatorial PTM patterns.

#### **4. Protein profiling to elucidate E2F function**

A large number of potential E2F targets have been discovered recently through the application of high throughput transcriptomic analyses. However, mRNA abundance poorly correlates with protein levels. This discrepancy can in part be explained by several post-transcriptional mechanisms such as alternative splicing of pre-mRNAs, microRNA mediated regulation, or selective degradation and post–translational modifications of proteins. Therefore, protein expression analyses should be carried out in addition to mRNA expression analyses to achieve a full picture of E2F regulated pathways.

Proteomic Approaches to Unraveling the RB/E2F Regulatory Pathway 149

type T lymphocyte expression profiles were gathered by 2-DE followed by MS, we identified a set of deregulated proteins involved in TCR-mediated signaling, cell survival and stress responses (Azkargorta et al. 2006). The aberrant expression of these proteins was linked to the hyperproliferative phenotype that characterizes E2F2-deficient cells (Iglesias et al. 2004; Infante et al. 2008; Murga et al. 2001; Zhu et al. 2001). Interestingly, comparative proteomics has also revealed novel pathways regulated by E2F2. We have recently found that mediators of the Aryl-hydrocarbon receptor (AHR) are aberrantly expressed in the proteome of proliferating E2F2-/- lymphocytes relative to wild-type counterparts (Azkargorta et al. 2010). Consequently, E2F2-/- cells exhibit an increase in their sensitivity to dioxin-triggered apoptosis. These results suggest that E2F2 modulates cellular sensitivity to

Remarkably, a comparison of proteome and transcriptome profiling results derived from the same cellular systems has shown a clear discrepancy in individual targets regulated by E2Fs (Azkargorta et al. 2006; Azkargorta et al. 2010; Infante et al. 2008; Li et al. 2006a; Liontos et al. 2009; Muller et al. 2001). These discrepancies may be explained by the different methods of assay/sample preparation, different detection sensitivity, alternative splicing, posttranscriptional regulation, PTMs, selective degradation of proteins, and the time discrepancy between mRNA and protein expression. Thus, proteomics and DNA-based

Transcriptomic analyses have demonstrated a commonly conserved RB/E2F-dependent "proliferation signature" in cancer cells, supporting a role for this pathway in regulation of cellular proliferation in normal as well as tumor cells (Whitfield et al. 2006). Given the prevalence of this signature in cancer, a similar finding would be expected at the protein level. However, most 2-DE/MS-based proteomic analyses of tumor cells performed to date have not displayed this pathway unequivocally, probably due to the limitations of the technology. An important drawback of traditional 2-DE-based proteomic methods is that they can only reveal the presence of highly abundant proteins within the cells, whereas proteins that function at low levels, such as certain transcription factors and checkpoint/regulatory proteins, are not easily identified if no pre-enrichment and/or prefractionation steps are applied (Gygi et al. 2000). In general, the number of proteins identified in these experiments is quite low, in the range of 100-300 proteins per experiment. This number is several orders of magnitude lower than the thousands of genes that can be interrogated by DNA microarray analyses. Consequently, cancer proteome profiles gathered

xenobiotic signals through the negative regulation of the AHR pathway.

technologies should be considered as complementary approaches.

to date with this technology should, in general, be considered as preliminary.

An RB/E2F dependent signature was appreciable in the proteome of cervical cancer cells with high-risk HPV infection analyzed by 2-DE and MS. Differentially expressed proteins in cervical cancer cells were identified and functionally classified as proteins involved in the regulation of the cell cycle, general genomic stability, telomerase activation and cell immortalization (Choi et al. 2005). A significant proportion of these genes (30%), particularly those present in the nuclear fraction, are well-known E2F transcriptional targets (PCNA, CDC25A, MCM3,4,8, CHRAC-1), most of which were upregulated. As exceptions, XRCC2 involved in DNA repair and CASP-2 involved in apoptosis induction, which were downregulated. These results are consistent with a model in which functional inactivation of RB by HPV oncoprotein E7 unleashes E2F and triggers aberrant E2F-dependent gene expression (McLaughlin-Drubin & Munger 2009). By contrast, a similar type of analysis

**4.1 The RB/E2F pathway in cancer protein profiling** 

Two-dimensional gel electrophoresis (2-DE) combined with protein identification by MS has been the most popular method in expression proteomics. In this global protein profiling approach, 2-DE is used to separate proteins within a complex mixture based on their pI and molecular weight. Subsequent staining and image analysis can detect differences in spot intensities, rendering a list of proteins that are differentially expressed between the conditions under comparison. The resulting spots are then excised and analyzed by MS (Görg et al. 2004). Global protein expression analyses of cells that either overexpress or are deficient for individual E2Fs have emerged as a useful methodology to elucidate the spectrum of E2F activity. Pützer's group analyzed protein expression in p53-deficient osteosarcoma cells (Saos-2) expressing E2F1 fused to the murine estrogen receptor ligandbinding domain (ER). This construction permitted conditional activation of E2F1 after addition of 4-hydroxytamoxifen (4-OHT) to the cultures (Stanelle et al. 2002). 2-DE analysis of E2F1 induced and non-induced Saos-2 cells' proteome led to the identification of 33 novel differentially regulated E2F1 target proteins by MALDI-MS (Li et al. 2006b), 15 upregulated and 18 dowregulated. Only eight of these differentially regulated proteins harbor E2Fconsensus sites in their promoter (Rabinovich et al. 2008): SRSF1, TUBB, GDI2, H2B.1, TCP1γ, HNRNPA2/B1, HNRNPK and MATR3. Nevertheless, since E2F1 overexpression was the only stimulus applied to the culture, most of the identified target genes are probably E2F1 targets. Functional analysis of E2F1-regulated target genes suggests that E2F1 plays predominantly a negative role in Saos-2 cellular proliferation. By altering the balance between anti- and pro-apoptotic Bcl-2 family members, E2F1 would render cells susceptible to both mitochondrial apoptosis and ER-stress related death signals (Li et al. 2006b). These results are consistent with the findings described by several groups, including ours, in mouse models lacking E2F1, and favor the hypothesis that E2F1 plays anti-proliferative and pro-apoptotic roles *in vivo* (Field et al. 1996; Garcia et al. 2000; Yamasaki et al. 1996).

A similar pro-apoptotic role has been attributed to E2F1 in proteomic analyses involving p53-wild-type osteosarcoma cells (U2OS) carrying an inducible E2F1 transgene. 2-DE followed by MS led to the identification of 76 proteins, 53 overexpressed and 23 suppressed after E2F1 induction (Liontos et al. 2009). Many of them appear to be potentially novel E2F1 regulated targets. As many as 63% and 77% of the overexpressed and suppressed proteins, respectively, harbor E2F1 responsive elements in their promoters. The spectrum of identified E2F1-regulated proteins included chaperones, metabolic enzymes, proteins associated with RNA processing, components of the protein degradation/turnover machinery, cytoskeletal and motor/contractile related proteins, regulatory and cell signaling molecules, transport carriers and channels, as well as putative oncogenes. A significant number of the identified E2F1 downstream targets are part of pro-apoptotic signaling cascades regulating ATM and/or p53. In agreement with an oncosuppressor function for E2F1 *in vivo*, the authors reported a positive correlation between E2F1 expression and DNA damage response and apoptosis in primary osteosarcoma tumors with wild-type p53. This picture is contrary to the general view that the increased levels of E2F activators, such as those observed in many cancer types, are mediating uncontrolled proliferation (Chen et al. 2009a). Whether deregulated expression of E2F proteins promotes or limits cancer progression has not been unequivocally established. Other factors, such as RB or p53 status may determine the final outcome of E2F-dependent activity in each particular context.

Protein profiling analyses performed by our group are also helping to unveil the physiological role of E2F2. A repressor function for E2F2 is emerging from these studies, in agreement with DNA microarray and functional data. When quiescent E2F2-/- and wild-

Two-dimensional gel electrophoresis (2-DE) combined with protein identification by MS has been the most popular method in expression proteomics. In this global protein profiling approach, 2-DE is used to separate proteins within a complex mixture based on their pI and molecular weight. Subsequent staining and image analysis can detect differences in spot intensities, rendering a list of proteins that are differentially expressed between the conditions under comparison. The resulting spots are then excised and analyzed by MS (Görg et al. 2004). Global protein expression analyses of cells that either overexpress or are deficient for individual E2Fs have emerged as a useful methodology to elucidate the spectrum of E2F activity. Pützer's group analyzed protein expression in p53-deficient osteosarcoma cells (Saos-2) expressing E2F1 fused to the murine estrogen receptor ligandbinding domain (ER). This construction permitted conditional activation of E2F1 after addition of 4-hydroxytamoxifen (4-OHT) to the cultures (Stanelle et al. 2002). 2-DE analysis of E2F1 induced and non-induced Saos-2 cells' proteome led to the identification of 33 novel differentially regulated E2F1 target proteins by MALDI-MS (Li et al. 2006b), 15 upregulated and 18 dowregulated. Only eight of these differentially regulated proteins harbor E2Fconsensus sites in their promoter (Rabinovich et al. 2008): SRSF1, TUBB, GDI2, H2B.1, TCP1γ, HNRNPA2/B1, HNRNPK and MATR3. Nevertheless, since E2F1 overexpression was the only stimulus applied to the culture, most of the identified target genes are probably E2F1 targets. Functional analysis of E2F1-regulated target genes suggests that E2F1 plays predominantly a negative role in Saos-2 cellular proliferation. By altering the balance between anti- and pro-apoptotic Bcl-2 family members, E2F1 would render cells susceptible to both mitochondrial apoptosis and ER-stress related death signals (Li et al. 2006b). These results are consistent with the findings described by several groups, including ours, in mouse models lacking E2F1, and favor the hypothesis that E2F1 plays anti-proliferative and

pro-apoptotic roles *in vivo* (Field et al. 1996; Garcia et al. 2000; Yamasaki et al. 1996).

A similar pro-apoptotic role has been attributed to E2F1 in proteomic analyses involving p53-wild-type osteosarcoma cells (U2OS) carrying an inducible E2F1 transgene. 2-DE followed by MS led to the identification of 76 proteins, 53 overexpressed and 23 suppressed after E2F1 induction (Liontos et al. 2009). Many of them appear to be potentially novel E2F1 regulated targets. As many as 63% and 77% of the overexpressed and suppressed proteins, respectively, harbor E2F1 responsive elements in their promoters. The spectrum of identified E2F1-regulated proteins included chaperones, metabolic enzymes, proteins associated with RNA processing, components of the protein degradation/turnover machinery, cytoskeletal and motor/contractile related proteins, regulatory and cell signaling molecules, transport carriers and channels, as well as putative oncogenes. A significant number of the identified E2F1 downstream targets are part of pro-apoptotic signaling cascades regulating ATM and/or p53. In agreement with an oncosuppressor function for E2F1 *in vivo*, the authors reported a positive correlation between E2F1 expression and DNA damage response and apoptosis in primary osteosarcoma tumors with wild-type p53. This picture is contrary to the general view that the increased levels of E2F activators, such as those observed in many cancer types, are mediating uncontrolled proliferation (Chen et al. 2009a). Whether deregulated expression of E2F proteins promotes or limits cancer progression has not been unequivocally established. Other factors, such as RB or p53 status may determine the final outcome of E2F-dependent activity in each particular context. Protein profiling analyses performed by our group are also helping to unveil the physiological role of E2F2. A repressor function for E2F2 is emerging from these studies, in agreement with DNA microarray and functional data. When quiescent E2F2-/- and wildtype T lymphocyte expression profiles were gathered by 2-DE followed by MS, we identified a set of deregulated proteins involved in TCR-mediated signaling, cell survival and stress responses (Azkargorta et al. 2006). The aberrant expression of these proteins was linked to the hyperproliferative phenotype that characterizes E2F2-deficient cells (Iglesias et al. 2004; Infante et al. 2008; Murga et al. 2001; Zhu et al. 2001). Interestingly, comparative proteomics has also revealed novel pathways regulated by E2F2. We have recently found that mediators of the Aryl-hydrocarbon receptor (AHR) are aberrantly expressed in the proteome of proliferating E2F2-/- lymphocytes relative to wild-type counterparts (Azkargorta et al. 2010). Consequently, E2F2-/- cells exhibit an increase in their sensitivity to dioxin-triggered apoptosis. These results suggest that E2F2 modulates cellular sensitivity to xenobiotic signals through the negative regulation of the AHR pathway.

Remarkably, a comparison of proteome and transcriptome profiling results derived from the same cellular systems has shown a clear discrepancy in individual targets regulated by E2Fs (Azkargorta et al. 2006; Azkargorta et al. 2010; Infante et al. 2008; Li et al. 2006a; Liontos et al. 2009; Muller et al. 2001). These discrepancies may be explained by the different methods of assay/sample preparation, different detection sensitivity, alternative splicing, posttranscriptional regulation, PTMs, selective degradation of proteins, and the time discrepancy between mRNA and protein expression. Thus, proteomics and DNA-based technologies should be considered as complementary approaches.

#### **4.1 The RB/E2F pathway in cancer protein profiling**

Transcriptomic analyses have demonstrated a commonly conserved RB/E2F-dependent "proliferation signature" in cancer cells, supporting a role for this pathway in regulation of cellular proliferation in normal as well as tumor cells (Whitfield et al. 2006). Given the prevalence of this signature in cancer, a similar finding would be expected at the protein level. However, most 2-DE/MS-based proteomic analyses of tumor cells performed to date have not displayed this pathway unequivocally, probably due to the limitations of the technology. An important drawback of traditional 2-DE-based proteomic methods is that they can only reveal the presence of highly abundant proteins within the cells, whereas proteins that function at low levels, such as certain transcription factors and checkpoint/regulatory proteins, are not easily identified if no pre-enrichment and/or prefractionation steps are applied (Gygi et al. 2000). In general, the number of proteins identified in these experiments is quite low, in the range of 100-300 proteins per experiment. This number is several orders of magnitude lower than the thousands of genes that can be interrogated by DNA microarray analyses. Consequently, cancer proteome profiles gathered to date with this technology should, in general, be considered as preliminary.

An RB/E2F dependent signature was appreciable in the proteome of cervical cancer cells with high-risk HPV infection analyzed by 2-DE and MS. Differentially expressed proteins in cervical cancer cells were identified and functionally classified as proteins involved in the regulation of the cell cycle, general genomic stability, telomerase activation and cell immortalization (Choi et al. 2005). A significant proportion of these genes (30%), particularly those present in the nuclear fraction, are well-known E2F transcriptional targets (PCNA, CDC25A, MCM3,4,8, CHRAC-1), most of which were upregulated. As exceptions, XRCC2 involved in DNA repair and CASP-2 involved in apoptosis induction, which were downregulated. These results are consistent with a model in which functional inactivation of RB by HPV oncoprotein E7 unleashes E2F and triggers aberrant E2F-dependent gene expression (McLaughlin-Drubin & Munger 2009). By contrast, a similar type of analysis

Proteomic Approaches to Unraveling the RB/E2F Regulatory Pathway 151

identifications. With regard to E2Fs, MS-based phosphorylation analyses have greatly increased the number of phosphorylated residues identified within this transcription factor family, and many more will probably be identified in the future. However, they will need to be validated and functionally characterized *in vivo* in order to decipher the influence of each individual PTM on E2F activity. As works focused on MS-based identification are gaining relevance, other types of PTMs such as acetylations and ubiquitinations, will probably be identified on E2Fs in a short period of time. Furthermore, multiple PTMs should be possible to analyze in the same system, producing direct data on their crosstalk at the global level in regulating E2F function. In addition, implementation of quantitative methods will allow studying the dynamics of site-specific PTMs on E2Fs, which was almost unachievable a decade ago. Defining their kinetics will be crucial to understand how PTMs influence on protein regulation and function. This explosion of information, still very descriptive, opens a

Several differential proteomic analyses have been performed to examine how the presence/absence of each individual E2F is reflected in the proteome. The strategy to accomplish this objective will surely move from the low scale approach that represents 2- DE/MS used so far, to gel-free high throughput strategies. This approach has been widely adopted in proteomics lately, owing to its superior performance compared to 2-DE technology. Advances in sample preparation, including novel fractionation methods, together with more powerful mass spectrometers have been proved to qualitatively and quantitatively assess changes in large scale protein analyses. Indeed, large-scale protein expression analysis of cancer cell proteomes by shotgun proteomics promises to be valuable for investigating mechanisms of cancer transformation (Chen & Yates, III 2007). Nonetheless, the limitations of global proteomic analyses in identifying and quantifying low abundant proteins such as transcription factors will push the emergence of hypothesis-

This work was supported by grants from the Spanish Ministry of Science and Innovation (SAF2009-12037 and Consolider-Ingenio Programme CSD2007-00017 to AMZ) and the Basque Government Department of Industry (ETORTEK-IE09-256 and SAIOTEK-S-PE10UN82 to AMZ and JMA). JM is supported by a Basque Government fellowship for graduate studies, and NO is supported by a UPV/EHU fellowship for graduate studies.

Attwooll, C., Lazzerini, D. E., & Helin, K. (2004). The E2F family: specific functions and

Attwooll, C., Oddi, S., Cartwright, P., Prosperini, E., Agger, K., Steensgaard, P., Wagener, C.,

Azkargorta, M., Arizmendi, J. M., Elortza, F., Alkorta, N., Zubiaga, A. M., & Fullaondo, A.

Sardet, C., Moroni, M. C., & Helin, K. (2005). A novel repressive E2F6 complex containing the polycomb group protein, EPC1, that interacts with EZH2 in a

(2006). Differential proteome profiles in E2F2-deficient T lymphocytes. *Proteomics,* 6

huge range of possible new studies in the field of RB/E2F network.

driven strategies based on targeted proteomic approaches.

overlapping interests. *EMBO J.,* 23 (24): 4709-4716.

proliferation-specific manner. *J.Biol.Chem.,* 280 (2): 1199-1208.

**6. Acknowledgments** 

**7. References** 

Suppl (1): S42-S50.

performed with retinoblastoma tumors, whereby both copies of the RB1 gene are inactivated, did not identify any E2F targets among the 27 differentially expressed proteins (Mallikarjuna et al. 2010). In this work, more aggressive tumors showed significantly higher expression of CRABP2, APOA1, PRDX6 and RCVRA, and lower expression of CRABP1. Differentially expressed proteins were shown to be involved in metabolic process, transport activity, response to oxidative stress, development, and cell signaling and transduction, reflecting the important role of these processes during retinoblastoma progression.

In general, 2-DE/MS-based proteomic analyses of cancer cells have barely yielded a handful of E2F-target genes among the differentially expressed proteins. For example, the comprehensive proteome profiles of mouse lung adenocarcinoma cell lines, have revealed 82 and 40 unique proteins significantly up- or down-regulated respectively in highly metastatic cells compared to low metastatic controls (Zhang et al. 2008). Several of the proteins are involved in proteasome, cell-cycle and cell-cell communication pathways. Among them, several E2F targets have previously been associated with cancer development and metastasis: KRT8, the main cytokeratin in lung cancer, MCM7, a highly informative biomarker for cervical cancer, or ANX4, overexpressed in pancreatic adenocarcinoma. Similarly, a comparative proteome analysis of human lung squamous carcinoma and paired normal bronchial epithelial tissues revealed some deregulated E2F target genes in a list of 68 proteins that were identified by MS (LTBP4, GNB1L, MDM2, IRS1)(Li et al. 2006a).

As 2-DE/MS methodology is being replaced by more powerful techniques, the number of proteins that can be identified in a given proteome is increasing rapidly, allowing for a more accurate definition of the pathways involved in cancer. A recent report has described the use of 1-D gel electrophoresis followed by tryptic in-gel digestion and chromatography coupled to MS. This shotgun proteomic approach revealed changes in the expression levels of 281 proteins in meningiomas compared to normal human arachnoidal cells (Saydam et al. 2010). Interestingly, a highly significant functional network involved in DNA replication, recombination and repair, and in cell cycle was exclusively expressed in cancer cells, arguing for a prevalent RB/E2F signature in the meningioma proteome. Similar results have been gathered recently with high-risk neuroblastoma proteomes analyzed by another type of shotgun proteomics approach. Quantitative global protein expression profiling performed using isotope-coded affinity tags (ICAT) followed by LC-MS/MS analysis identified a total of 1,461 proteins that were differentially expressed in neuroblastoma. Again, pathway analysis of these proteins showed enrichment in the RB/E2F regulated network (Chen et al. 2010). These results underscore the power of shotgun proteomics, and suggest that many novel insights regarding RB/E2F function will be unraveled by this approach.

#### **5. Future perspectives**

Much has been learned on RB/E2F-mediated regulation of gene transcription by proteomic approaches to date. The identification of E2F interacting proteins as well as transcriptional complexes encompassing RB and E2F has helped elucidating the regulation of this pathway. However, the current knowledge is still too limited to fully understand the complex RB/E2F-regulated network. Further work involving high throughput proteomic approaches should help elucidate the nature of the diverse macromolecular complexes that are thought to harbor RB/E2F, and the role of these complexes in different biological settings.

Development of modified peptide enrichment methods together with recent advances in MS has allowed large-scale analyses of PTMs, providing a large pool of new PTM identifications. With regard to E2Fs, MS-based phosphorylation analyses have greatly increased the number of phosphorylated residues identified within this transcription factor family, and many more will probably be identified in the future. However, they will need to be validated and functionally characterized *in vivo* in order to decipher the influence of each individual PTM on E2F activity. As works focused on MS-based identification are gaining relevance, other types of PTMs such as acetylations and ubiquitinations, will probably be identified on E2Fs in a short period of time. Furthermore, multiple PTMs should be possible to analyze in the same system, producing direct data on their crosstalk at the global level in regulating E2F function. In addition, implementation of quantitative methods will allow studying the dynamics of site-specific PTMs on E2Fs, which was almost unachievable a decade ago. Defining their kinetics will be crucial to understand how PTMs influence on protein regulation and function. This explosion of information, still very descriptive, opens a huge range of possible new studies in the field of RB/E2F network.

Several differential proteomic analyses have been performed to examine how the presence/absence of each individual E2F is reflected in the proteome. The strategy to accomplish this objective will surely move from the low scale approach that represents 2- DE/MS used so far, to gel-free high throughput strategies. This approach has been widely adopted in proteomics lately, owing to its superior performance compared to 2-DE technology. Advances in sample preparation, including novel fractionation methods, together with more powerful mass spectrometers have been proved to qualitatively and quantitatively assess changes in large scale protein analyses. Indeed, large-scale protein expression analysis of cancer cell proteomes by shotgun proteomics promises to be valuable for investigating mechanisms of cancer transformation (Chen & Yates, III 2007). Nonetheless, the limitations of global proteomic analyses in identifying and quantifying low abundant proteins such as transcription factors will push the emergence of hypothesisdriven strategies based on targeted proteomic approaches.

#### **6. Acknowledgments**

150 Proteomics – Human Diseases and Protein Functions

performed with retinoblastoma tumors, whereby both copies of the RB1 gene are inactivated, did not identify any E2F targets among the 27 differentially expressed proteins (Mallikarjuna et al. 2010). In this work, more aggressive tumors showed significantly higher expression of CRABP2, APOA1, PRDX6 and RCVRA, and lower expression of CRABP1. Differentially expressed proteins were shown to be involved in metabolic process, transport activity, response to oxidative stress, development, and cell signaling and transduction,

In general, 2-DE/MS-based proteomic analyses of cancer cells have barely yielded a handful of E2F-target genes among the differentially expressed proteins. For example, the comprehensive proteome profiles of mouse lung adenocarcinoma cell lines, have revealed 82 and 40 unique proteins significantly up- or down-regulated respectively in highly metastatic cells compared to low metastatic controls (Zhang et al. 2008). Several of the proteins are involved in proteasome, cell-cycle and cell-cell communication pathways. Among them, several E2F targets have previously been associated with cancer development and metastasis: KRT8, the main cytokeratin in lung cancer, MCM7, a highly informative biomarker for cervical cancer, or ANX4, overexpressed in pancreatic adenocarcinoma. Similarly, a comparative proteome analysis of human lung squamous carcinoma and paired normal bronchial epithelial tissues revealed some deregulated E2F target genes in a list of 68

reflecting the important role of these processes during retinoblastoma progression.

proteins that were identified by MS (LTBP4, GNB1L, MDM2, IRS1)(Li et al. 2006a).

regarding RB/E2F function will be unraveled by this approach.

**5. Future perspectives** 

As 2-DE/MS methodology is being replaced by more powerful techniques, the number of proteins that can be identified in a given proteome is increasing rapidly, allowing for a more accurate definition of the pathways involved in cancer. A recent report has described the use of 1-D gel electrophoresis followed by tryptic in-gel digestion and chromatography coupled to MS. This shotgun proteomic approach revealed changes in the expression levels of 281 proteins in meningiomas compared to normal human arachnoidal cells (Saydam et al. 2010). Interestingly, a highly significant functional network involved in DNA replication, recombination and repair, and in cell cycle was exclusively expressed in cancer cells, arguing for a prevalent RB/E2F signature in the meningioma proteome. Similar results have been gathered recently with high-risk neuroblastoma proteomes analyzed by another type of shotgun proteomics approach. Quantitative global protein expression profiling performed using isotope-coded affinity tags (ICAT) followed by LC-MS/MS analysis identified a total of 1,461 proteins that were differentially expressed in neuroblastoma. Again, pathway analysis of these proteins showed enrichment in the RB/E2F regulated network (Chen et al. 2010). These results underscore the power of shotgun proteomics, and suggest that many novel insights

Much has been learned on RB/E2F-mediated regulation of gene transcription by proteomic approaches to date. The identification of E2F interacting proteins as well as transcriptional complexes encompassing RB and E2F has helped elucidating the regulation of this pathway. However, the current knowledge is still too limited to fully understand the complex RB/E2F-regulated network. Further work involving high throughput proteomic approaches should help elucidate the nature of the diverse macromolecular complexes that are thought

Development of modified peptide enrichment methods together with recent advances in MS has allowed large-scale analyses of PTMs, providing a large pool of new PTM

to harbor RB/E2F, and the role of these complexes in different biological settings.

This work was supported by grants from the Spanish Ministry of Science and Innovation (SAF2009-12037 and Consolider-Ingenio Programme CSD2007-00017 to AMZ) and the Basque Government Department of Industry (ETORTEK-IE09-256 and SAIOTEK-S-PE10UN82 to AMZ and JMA). JM is supported by a Basque Government fellowship for graduate studies, and NO is supported by a UPV/EHU fellowship for graduate studies.

#### **7. References**


Proteomic Approaches to Unraveling the RB/E2F Regulatory Pathway 153

Choi, Y. P., Kang, S., Hong, S., Xie, X., & Cho, N. H. (2005). Proteomic analysis of progressive factors in uterine cervical cancer. *Proteomics,* 5 (6): 1481-1493. Choudhary, C., Olsen, J. V., Brandts, C., Cox, J., Reddy, P. N., Bohmer, F. D., Gerke, V.,

DeGregori, J., Leone, G., Miron, A., Jakoi, L., & Nevins, J. R. (1997). Distinct roles for E2F

DeGregori, J. & Johnson, D. G. (2006). Distinct and Overlapping Roles for E2F Family

Dephoure, N., Zhou, C., Villen, J., Beausoleil, S. A., Bakalarski, C. E., Elledge, S. J., & Gygi, S.

Dimova, D. K. & Dyson, N. J. (2005). The E2F transcriptional network: old acquaintances

Escher, N., Kob, R., Tenbaum, S. P., Eisold, M., Baniahmad, A., von, E. F., & Melle, C. (2007).

Fagan, R., Flint, K. J., & Jones, N. (1994). Phosphorylation of E2F-1 modulates its interaction

Fay, D. S. & Han, M. (2000). The synthetic multivulval genes of C. elegans: functional redundancy, Ras-antagonism, and cell fate determination. *Genesis.,* 26 (4): 279-284. Field, S. J., Tsai, F. Y., Kuo, F., Zubiaga, A. M., Kaelin, W. G., Jr., Livingston, D. M., Orkin, S.

Fields, S. & Song, O. (1989). A novel genetic system to detect protein-protein interactions.

Florens, L. & Washburn, M. P. (2006). Proteomic analysis by multidimensional protein

Freedman, J. A., Chang, J. T., Jakoi, L., & Nevins, J. R. (2009). A combinatorial mechanism for

Frolov, M. V. & Dyson, N. J. (2004). Molecular mechanisms of E2F-dependent activation and

Garcia, I., Murga, M., Vicario, A., Field, S. J., & Zubiaga, A. M. (2000). A role for E2F1 in the

Gaubatz, S., Wood, J. G., & Livingston, D. M. (1998). Unusual proliferation arrest and

Gavin, A. C., Bosche, M., Krause, R., Grandi, P., Marzioch, M., Bauer, A., Schultz, J., Rick, J.

outcomes. *Mol.Cell,* 36 (2): 326-339.

with new faces. *Oncogene,* 24 (17): 2810-2826.

suppress proliferation. *Cell,* 85 (4): 549-561.

*Proc.Natl.Acad.Sci.U.S.A,* 95 (16): 9190-9195.

identification technology. *Methods Mol.Biol.,* 328: 159-175.

pRB-mediated repression. *J.Cell Sci.,* 117 (Pt 11): 2173-2181.

*Nature,* 340 (6230): 245-246.

corepressor alien. *J.Proteome Res.,* 6 (3): 1158-1164.

7245-7250.

105 (31): 10762-10767.

78 (5): 799-811.

2873-2881.

91-98.

748.

Schmidt-Arras, D. E., Berdel, W. E., Muller-Tidow, C., Mann, M., & Serve, H. (2009). Mislocalized activation of oncogenic RTKs switches downstream signaling

proteins in cell growth control and apoptosis. *Proc.Natl.Acad.Sci.U.S.A,* 94 (14):

Members in Transcription, Proliferation and Apoptosis. *Curr.Mol.Med.,* 6 (7): 739-

P. (2008). A quantitative atlas of mitotic phosphorylation. *Proc.Natl.Acad.Sci.U.S.A,*

Various members of the E2F transcription factor family interact in vivo with the

with the retinoblastoma gene product and the adenoviral E4 19 kDa protein. *Cell,*

H., & Greenberg, M. E. (1996). E2F-1 functions in mice to promote apoptosis and

determining the specificity of E2F activation and repression. *Oncogene,* 28 (32):

induction of apoptosis during thymic negative selection. *Cell Growth Differ.,* 11 (2):

transcriptional control properties of a newly discovered E2F family member, E2F-6.

M., Michon, A. M., Cruciat, C. M., Remor, M., Hofert, C., Schelder, M., Brajenovic, M., Ruffner, H., Merino, A., Klein, K., Hudak, M., Dickson, D., Rudi, T., Gnau, V.,


Azkargorta, M., Fullaondo, A., Laresgoiti, U., Aloria, K., Infante, A., Arizmendi, J. M., &

Balciunaite, E., Spektor, A., Lents, N. H., Cam, H., Te, R. H., Scime, A., Rudnicki, M. A.,

Bandara, L. R. & La Thangue, N. B. (1991). Adenovirus E1a prevents the retinoblastoma

Baumann, P. & West, S. C. (1998). Role of the human RAD51 protein in homologous

Beausoleil, S. A., Jedrychowski, M., Schwartz, D., Elias, J. E., Villen, J., Li, J., Cohn, M. A.,

Boersema, P. J., Divecha, N., Heck, A. J., & Mohammed, S. (2007). Evaluation and

Brehm, A., Miska, E. A., McCance, D. J., Reid, J. L., Bannister, A. J., & Kouzarides, T. (1998).

Bunker, C. A. & Kingston, R. E. (1994). Transcriptional repression by Drosophila and

Burkhart, D. L. & Sage, J. (2008). Cellular mechanisms of tumour suppression by the

Cartwright, P., Muller, H., Wagener, C., Holm, K., & Helin, K. (1998). E2F-6: a novel member

Chellappan, S. P., Hiebert, S., Mudryj, M., Horowitz, J. M., & Nevins, J. R. (1991). The E2F transcription factor is a cellular target for the RB protein. *Cell,* 65 (6): 1053-1061. Chen, E. I. & Yates, J. R., III (2007). Cancer proteomics by quantitative shotgun proteomics.

Chen, H. Z., Tsai, S. Y., & Leone, G. (2009a). Emerging roles of E2Fs in cancer: an exit from

Chen, Q. R., Song, Y. K., Yu, L. R., Wei, J. S., Chung, J. Y., Hewitt, S. M., Veenstra, T. D., &

pathways related to high-risk neuroblastoma. *J.Proteome Res.,* 9 (1): 373-382. Chen, R. Q., Yang, Q. K., Lu, B. W., Yi, W., Cantin, G., Chen, Y. L., Fearns, C., Yates, J. R., III,

Chittenden, T., Livingston, D. M., & Kaelin, W. G., Jr. (1991). The T/E1A-binding domain of

Khan, J. (2010). Global genomic and proteomic analysis identifies biological

& Lee, J. D. (2009b). CDC25B mediates rapamycin-induced oncogenic responses in

the retinoblastoma product can interact selectively with a sequence-specific DNA-

retinoblastoma gene. *Nat.Rev.Cancer,* 8 (9): 671-682.

cell cycle control. *Nat.Rev.Cancer,* 9 (11): 785-797.

cancer cells. *Cancer Res.,* 69 (6): 2663-2668.

binding protein. *Cell,* 65 (6): 1073-1082.

phosphoproteins. *Proc.Natl.Acad.Sci.U.S.A,* 101 (33): 12130-12135.

2194.

251.

(6326): 494-497.

(3): 937-946.

611-623.

391 (6667): 597-601.

*Biol.,* 14 (3): 1721-1732.

*Mol.Oncol.,* 1 (2): 144-159.

Zubiaga, A. M. (2010). Differential proteomics analysis reveals a role for E2F2 in the regulation of the Ahr pathway in T lymphocytes. *Mol.Cell Proteomics,* 9 (10): 2184-

Young, R., & Dynlacht, B. D. (2005). Pocket protein complexes are recruited to distinct targets in quiescent and proliferating cells. *Mol.Cell Biol.,* 25 (18): 8166-8178.

gene product from complexing with a cellular transcription factor. *Nature,* 351

recombination and double-stranded-break repair. *Trends Biochem.Sci.,* 23 (7): 247-

Cantley, L. C., & Gygi, S. P. (2004). Large-scale characterization of HeLa cell nuclear

optimization of ZIC-HILIC-RP as an alternative MudPIT strategy. *J.Proteome Res.,* 6

Retinoblastoma protein recruits histone deacetylase to repress transcription. *Nature,*

mammalian Polycomb group proteins in transfected mammalian cells. *Mol.Cell* 

of the E2F family is an inhibitor of E2F-dependent transcription. *Oncogene,* 17 (5):


Proteomic Approaches to Unraveling the RB/E2F Regulatory Pathway 155

Huttlin, E. L., Jedrychowski, M. P., Elias, J. E., Goswami, T., Rad, R., Beausoleil, S. A., Villen,

Iaquinta, P. J. & Lees, J. A. (2007). Life and death decisions by the E2F transcription factors.

Iglesias, A., Murga, M., Laresgoiti, U., Skoudy, A., Bernales, I., Fullaondo, A., Moreno, B.,

Infante, A., Laresgoiti, U., Fernandez-Rueda, J., Fullaondo, A., Galan, J., Diaz-Uriarte, R.,

Jirawatnotai, S., Hu, Y., Michowski, W., Elias, J. E., Becks, L., Bienvenu, F., Zagozdzon, A.,

Johnson, D. G. & DeGregori, J. (2006). Putting the Oncogenic and Tumor Suppressive

Kaelin, W. G., Jr. (2003). E2F1 as a target: promoter-driven suicide and small molecule

Ko, M., Huang, Y., Jankowska, A. M., Pape, U. J., Tahiliani, M., Bandukwala, H. S., An, J.,

Kontaki, H. & Talianidis, I. (2010). Lysine methylation regulates E2F1-induced cell death.

Korenjak, M., Taylor-Harding, B., Binne, U. K., Satterlee, J. S., Stevaux, O., Aasland, R.,

Kowalik, T. F., DeGregori, J., Leone, G., Jakoi, L., & Nevins, J. R. (1998). E2F1-specific

Krogan, N. J., Cagney, G., Yu, H., Zhong, G., Guo, X., Ignatchenko, A., Li, J., Pu, S., Datta,

Lang, S. E., McMahon, S. B., Cole, M. D., & Hearing, P. (2001). E2F transcriptional activation requires TRRAP and GCN5 cofactors. *J.Biol.Chem.,* 276 (35): 32627-32634.

Lamperti, E. D., Koh, K. P., Ganetzky, R., Liu, X. S., Aravind, L., Agarwal, S., Maciejewski, J. P., & Rao, A. (2010). Impaired hydroxylation of 5-methylcytosine in

White-Cooper, H., Dyson, N., & Brehm, A. (2004). Native E2F/RBF complexes contain Myb-interacting proteins and repress transcription of developmentally

induction of apoptosis and p53 accumulation, which is blocked by Mdm2. *Cell* 

N., Tikuisis, A. P., Punna, T., Peregrin-Alvarez, J. M., Shales, M., Zhang, X., Davey, M., Robinson, M. D., Paccanaro, A., Bray, J. E., Sheung, A., Beattie, B., Richards, D. P., Canadien, V., Lalev, A., Mena, F., Wong, P., Starostine, A., Canete, M. M., Vlasblom, J., Wu, S., Orsi, C., Collins, S. R., Chandran, S., Haw, R., Rilstone, J. J., Gandi, K., Thompson, N. J., Musso, G., St, O. P., Ghanny, S., Lam, M. H., Butland, G., Altaf-Ul, A. M., Kanaya, S., Shilatifard, A., O'Shea, E., Weissman, J. S., Ingles, C. J., Hughes, T. R., Parkinson, J., Gerstein, M., Wodak, S. J., Emili, A., & Greenblatt, J. F. (2006). Global landscape of protein complexes in the yeast Saccharomyces

protein phosphorylation and expression. *Cell,* 143 (7): 1174-1189.

regulators to maintain quiescence. *Cell Cycle,* 7 (24): 3915-3927.

Activities of E2F into Context. *Curr.Mol.Med.,* 6 (7): 731-738.

myeloid cancers with mutant TET2. *Nature,* 468 (7325): 839-843.

modulators. *Cancer Biol.Ther.,* 2 (4 Suppl 1): S48-S54.

controlled E2F target genes. *Cell,* 119 (2): 181-193.

*Curr.Opin.Cell Biol.,* 19 (6): 649-657.

cancers. *Nature,* 474 (7350): 230-234.

*Mol.Cell,* 39 (1): 152-160.

*Growth Differ.,* 9 (2): 113-118.

cerevisiae. *Nature,* 440 (7084): 637-643.

1398-1407.

J., Haas, W., Sowa, M. E., & Gygi, S. P. (2010). A tissue-specific atlas of mouse

Lloreta, J., Field, S. J., Real, F. X., & Zubiaga, A. M. (2004). Diabetes and exocrine pancreatic insufficiency in E2F1/E2F2 double-mutant mice. *J.Clin.Invest.,* 113 (10):

Malumbres, M., Field, S. J., & Zubiaga, A. M. (2008). E2F2 represses cell cycle

Goswami, T., Wang, Y. E., Clark, A. B., Kunkel, T. A., van, H. T., Xia, B., Correll, M., Quackenbush, J., Livingston, D. M., Gygi, S. P., & Sicinski, P. (2011). A function for cyclin D1 in DNA repair uncovered by protein interactome analyses in human

Bauch, A., Bastuck, S., Huhse, B., Leutwein, C., Heurtier, M. A., Copley, R. R., Edelmann, A., Querfurth, E., Rybin, V., Drewes, G., Raida, M., Bouwmeester, T., Bork, P., Seraphin, B., Kuster, B., Neubauer, G., & Superti-Furga, G. (2002). Functional organization of the yeast proteome by systematic analysis of protein complexes. *Nature,* 415 (6868): 141-147.


Giangrande, P. H., Hallstrom, T. C., Tunyaplin, C., Calame, K., & Nevins, J. R. (2003).

Görg, A., Weiss, W., & Dunn, M. J. (2004). Current two-dimensional electrophoresis

Gygi, S. P., Rist, B., & Aebersold, R. (2000). Measuring gene expression by quantitative

Hallstrom, T. C. & Nevins, J. R. (2003). Specificity in the activation and control of

Hallstrom, T. C. & Nevins, J. R. (2006). Jab1 is a specificity factor for E2F1-induced

Harrison, M. M., Ceol, C. J., Lu, X., & Horvitz, H. R. (2006). Some C. elegans class B synthetic

from a NuRD-like complex. *Proc.Natl.Acad.Sci.U.S.A,* 103 (45): 16782-16787. Ho, Y., Gruhler, A., Heilbut, A., Bader, G.D., Moore, L., Adams, S.L., Millar, A., Paul Taylor,

Hofmann, F., Martelli, F., Livingston, D. M., & Wang, Z. (1996). The retinoblastoma gene

Hong, S., Paulson, Q. X., & Johnson, D. G. (2008). E2F1 and E2F3 activate ATM through distinct mechanisms to promote E1A-induced apoptosis. *Cell Cycle,* 7 (3): 391-400. Hornbeck, P. V., Chabra, I., Kornhauser, J. M., Skrzypek, E., & Zhang, B. (2004).

Horsley, D., Hutchings, A., Butcher, G. W., & Singh, P. B. (1996). M32, a murine homologue

Hurlin, P. J., Steingrimsson, E., Copeland, N. G., Jenkins, N. A., & Eisenman, R. N. (1999).

T-domain DNA-binding motif. *EMBO J.,* 18 (24): 7019-7028.

product protects E2F-1 from degradation by the ubiquitin-proteasome pathway.

PhosphoSite: A bioinformatics resource dedicated to physiological protein

of Drosophila heterochromatin protein 1 (HP1), localises to euchromatin within interphase nuclei and is largely excluded from constitutive heterochromatin.

Mga, a dual-specificity transcription factor that interacts with Max and contains a

complexes. *Nature,* 415 (6868): 141-147.

factor. *Mol.Cell Biol.,* 23 (11): 3707-3720.

apoptosis. *Genes Dev.,* 20 (5): 613-623.

10848-10853.

(6868): 180-183.

*Genes Dev.,* 10 (23): 2949-2959.

phosphorylation. *Proteomics,* 4 (6): 1551-1561.

*Cytogenet.Cell Genet.,* 73 (4): 308-311.

technology for proteomics. *Proteomics,* 4 (12): 3665-3685.

proteome analysis. *Curr.Opin.Biotechnol.,* 11 (4): 396-401.

Bauch, A., Bastuck, S., Huhse, B., Leutwein, C., Heurtier, M. A., Copley, R. R., Edelmann, A., Querfurth, E., Rybin, V., Drewes, G., Raida, M., Bouwmeester, T., Bork, P., Seraphin, B., Kuster, B., Neubauer, G., & Superti-Furga, G. (2002). Functional organization of the yeast proteome by systematic analysis of protein

Identification of E-box factor TFE3 as a functional partner for the E2F3 transcription

transcription factor E2F-dependent apoptosis. *Proc.Natl.Acad.Sci.U.S.A,* 100 (19):

multivulva proteins encode a conserved LIN-35 Rb-containing complex distinct

Bennett, K., Boutilier, K., Yang, L., Wolting, C., Donaldson, I., Schandorff, S., Shewnarane, J., Vo, M., Taggart, J., Goudreault, M., Muskat, B., Alfarano, C., Dewar, D., Lin, Z., Michalickova, K., Willems, A.R., Sassi, H., Nielsen, P.A., Rasmussen, K.J., Andersen, J.R., Johansen, L.E., Hansen, L.H., Jespersen, H., Podtelejnikov, A., Nielsen, E., Crawford, J., Poulsen, V., Sørensen, B.D., Matthiesen, J., Hendrickson, R.C., Gleeson, F., Pawson, T., Moran, M.F., Durocher, D., Mann, M., Hogue, C.W.V., Figeys, D., & Tyers, M. (2002). Systematic identification of protein complexes in *Saccharomyces cerevisiae* by mass spectrometry. *Nature*, 415


Proteomic Approaches to Unraveling the RB/E2F Regulatory Pathway 157

Magnaghi-Jaulin, L., Groisman, R., Naguibneva, I., Robin, P., Lorain, S., Le Villain, J. P.,

Mallikarjuna, K., Sundaram, C. S., Sharma, Y., Deepa, P. R., Khetan, V., Gopal, L., Biswas, J.,

Martinez, L. A., Goluszko, E., Chen, H. Z., Leone, G., Post, S., Lozano, G., Chen, Z., &

Martinez-Balbas, M. A., Bauer, U. M., Nielsen, S. J., Brehm, A., & Kouzarides, T. (2000).

Marzio, G., Wagener, C., Gutierrez, M. I., Cartwright, P., Helin, K., & Giacca, M. (2000). E2F

Mayya, V., Lundgren, D. H., Hwang, S. I., Rezaul, K., Wu, L., Eng, J. K., Rodionov, V., &

McLaughlin-Drubin, M. E. & Munger, K. (2009). The human papillomavirus E7 oncoprotein.

Meloni, A. R., Smith, E. J., & Nevins, J. R. (1999a). A mechanism for Rb/p130-mediated

Meloni, A. R., Smith, E. J., & Nevins, J. R. (1999b). A mechanism for Rb/p130-mediated

Meng, Z., Camalier, C. E., Lucas, D. A., Veenstra, T. D., Beck, G. R., Jr., & Conrads, T. P.

Min, S. W., Cho, S. H., Zhou, Y., Schroeder, S., Haroutunian, V., Seeley, W. W., Huang, E. J.,

Monetti, M., Nagaraj, N., Sharma, K., & Mann, M. (2011). Large-scale phosphosite quantification in tissues by a spike-in SILAC method. *Nat.Methods,* 8 (8): 655-658. Morris, L., Allen, K. E., & La Thangue, N. B. (2000). Regulation of E2F transcription by cyclin

Mueller, T., Breuer, P., Schmitt, I., Walter, J., Evert, B. O., & Wullner, U. (2009). CK2-

Muller, H., Bracken, A. P., Vernell, R., Moroni, M. C., Christians, F., Grassilli, E., Prosperini,

Regulation of E2F1 activity by acetylation. *EMBO J.,* 19 (4): 662-671.

605.

*Proteomics.Clin.Appl.,* 4 (4): 449-463.

*Mol.Cell Biol.,* 30 (2): 524-536.

275 (15): 10887-10892.

*Sci.Signal.,* 2 (84): ra46.

(8): 1931-1939.

239.

*Neuron,* 67 (6): 953-966.

3. *Hum.Mol.Genet.,* 18 (17): 3334-3343.

*Virology,* 384 (2): 335-344.

*Proc.Natl.Acad.Sci.U.S.A,* 96 (17): 9574-9579.

*Proc.Natl.Acad.Sci.U.S.A,* 96 (17): 9574-9579.

Troalen, F., Trouche, D., & Harel-Bellan, A. (1998). Retinoblastoma protein represses transcription by recruiting a histone deacetylase. *Nature,* 391 (6667): 601-

Sharma, T., & Krishnakumar, S. (2010). Comparative proteomic analysis of differentially expressed proteins in primary retinoblastoma tumors.

Chauchereau, A. (2010). E2F3 is a mediator of DNA damage-induced apoptosis.

family members are differentially regulated by reversible acetylation. *J.Biol.Chem.,*

Han, D. K. (2009). Quantitative phosphoproteomic analysis of T cell receptor signaling reveals system-wide modulation of protein-protein interactions.

transcription repression involving recruitment of the CtBP corepressor.

transcription repression involving recruitment of the CtBP corepressor.

(2006). Probing early growth response 1 interacting proteins at the active promoter in osteoblast cells using oligoprecipitation and mass spectrometry. *J.Proteome Res.,* 5

Shen, Y., Masliah, E., Mukherjee, C., Meyers, D., Cole, P. A., Ott, M., & Gan, L. (2010). Acetylation of tau inhibits its degradation and contributes to tauopathy.

E-Cdk2 kinase mediated through p300/CBP co-activators. *Nat.Cell Biol.,* 2 (4): 232-

dependent phosphorylation determines cellular localization and stability of ataxin-

E., Vigo, E., Oliner, J. D., & Helin, K. (2001). E2Fs regulate the expression of genes


Larsen, M. R., Thingholm, T. E., Jensen, O. N., Roepstorff, P., & Jorgensen, T. J. (2005).

Lee, B. K., Bhinge, A. A., & Iyer, V. R. (2011). Wide-ranging functions of E2F4 in

Lehmann, R., Melle, C., Escher, N., & von, E. F. (2005). Detection and identification of

Lewis, P. W., Beall, E. L., Fleischer, T. C., Georlette, D., Link, A. J., & Botchan, M. R. (2004).

Li, C., Xiao, Z., Chen, Z., Zhang, X., Li, J., Wu, X., Li, X., Yi, H., Li, M., Zhu, G., & Liang, S.

Li, S. & Dass, C. (1999). Iron(III)-immobilized metal ion affinity chromatography and mass

Li, Z., Kreutzer, M., Mikkat, S., Mise, N., Glocker, M. O., & Putzer, B. M. (2006b). Proteomic

Liao, C. C., Tsai, C. Y., Chang, W. C., Lee, W. H., & Wang, J. M. (2010). RB.E2F1 complex

Lin, W. C., Lin, F. T., & Nevins, J. R. (2001). Selective induction of E2F1 in response to DNA

Liontos, M., Niforou, K., Velimezi, G., Vougas, K., Evangelou, K., Apostolopoulou, K., Vrtel,

Litovchick, L., Sadasivam, S., Florens, L., Zhu, X., Swanson, S. K., Velmurugan, S., Chen, R.,

Liu, F. & Lee, W. H. (2006). CtIP activates its own and cyclin D1 promoters via the E2F/RB

Luo, R. X., Postigo, A. A., & Dean, D. C. (1998). Rb interacts with histone deacetylase to

Ma, Y., Croxton, R., Moorer, R. L., Jr., & Cress, W. D. (2002). Identification of novel E2F1 regulated genes by microarray. *Arch.Biochem.Biophys.,* 399 (2): 212-224.

pathway during G1/S progression. *Mol.Cell Biol.,* 26 (8): 3124-3134.

dependent genes in quiescence. *Mol.Cell,* 26 (4): 539-551.

repress transcription. *Cell,* 92 (4): 463-473.

regulation of cell survival and death. *Proteomics,* 6 (21): 5735-5745.

using titanium dioxide microcolumns. *Mol.Cell Proteomics,* 4 (7): 873-886. Lazzerini, D. E., Attwooll, C., Pasini, D., & Helin, K. (2005). Deregulated E2F activity induces

*Biol.,* 25 (7): 2660-2672.

(5): 1717-1721.

547-558.

1844.

376-391.

*Nucleic Acids Res.,* 39 (9): 3558-3573.

*Genes Dev.,* 18 (23): 2929-2940.

*J.Biol.Chem.,* 285 (43): 33134-33143.

phosphopeptides. *Anal.Biochem.,* 270 (1): 9-14.

Highly selective enrichment of phosphorylated peptides from peptide mixtures

hyperplasia and senescence-like features in the mouse pituitary gland. *Mol.Cell* 

transcriptional activation and repression revealed by genome-wide analysis.

protein interactions of S100 proteins by ProteinChip technology. *J.Proteome Res.,* 4

Identification of a Drosophila Myb-E2F2/RBF transcriptional repressor complex.

(2006a). Proteome analysis of human lung squamous carcinoma. *Proteomics,* 6 (2):

spectrometry for the purification and characterization of synthetic

analysis of the E2F1 response in p53-negative cancer cells: new aspects in the

mediates DNA damage responses through transcriptional regulation of ZBRK1.

damage, mediated by ATM-dependent phosphorylation. *Genes Dev.,* 15 (14): 1833-

R., Damalas, A., Kontovazenitis, P., Kotsinas, A., Zoumpourlis, V., Tsangaris, G. T., Kittas, C., Ginsberg, D., Halazonetis, T. D., Bartek, J., & Gorgoulis, V. G. (2009). Modulation of the E2F1-driven cancer cell fate by the DNA damage response machinery and potential novel E2F1 targets in osteosarcomas. *Am.J.Pathol.,* 175 (1):

Washburn, M. P., Liu, X. S., & Decaprio, J. A. (2007). Evolutionarily conserved multisubunit RBL2/p130 and E2F4 protein complex represses human cell cycle-


Proteomic Approaches to Unraveling the RB/E2F Regulatory Pathway 159

Schlisio, S., Halperin, T., Vidal, M., & Nevins, J. R. (2002). Interaction of YY1 with E2Fs,

Schmit, F., Korenjak, M., Mannefeld, M., Schmitt, K., Franke, C., von, E. B., Gagrica, S.,

Sellers, W. R., Novitch, B. G., Miyake, S., Heith, A., Otterson, G. A., Kaye, F. J., Lassar, A. B.,

Song, Y., Willer, J. R., Scherer, P. C., Panzer, J. A., Kugath, A., Skordalakes, E., Gregg, R. G.,

Stanelle, J., Stiewe, T., Theseling, C. C., Peter, M., & Putzer, B. M. (2002). Gene expression changes in response to E2F1 activation. *Nucleic Acids Res.,* 30 (8): 1859-1867. Stevens, C., Smith, L., & La Thangue, N. B. (2003). Chk2 activates E2F-1 in response to DNA

Stirewalt, D. L. & Radich, J. P. (2003). The role of FLT3 in haematopoietic malignancies.

Takasaki, A., Hayashi, N., Matsubara, M., Yamauchi, E., & Taniguchi, H. (1999).

in calmodulin-target protein interaction. *J.Biol.Chem.,* 274 (17): 11848-11853. Tenbaum, S. P., Papaioannou, M., Reeb, C. A., Goeman, F., Escher, N., Kob, R., von, E. F.,

Tevosian, S. G., Shih, H. H., Mendelson, K. G., Sheppard, K. A., Paulson, K. E., & Yee, A. S.

Thingholm, T. E., Jensen, O. N., Robinson, P. J., & Larsen, M. R. (2008). SIMAC (sequential

Trimarchi, J. M., Fairchild, B., Verona, R., Moberg, K., Andon, N., & Lees, J. A. (1998). E2F-6,

Trimarchi, J. M., Fairchild, B., Wen, J., & Lees, J. A. (2001). The E2F6 transcription factor is a

Trimarchi, J. M. & Lees, J. A. (2002). Sibling rivalry in the E2F family. *Nat.Rev.Mol.Cell Biol.,* 3

Identification of the calmodulin-binding domain of neuron-specific protein kinase C substrate protein CAP-22/NAP-22. Direct involvement of protein myristoylation

Melle, C., & Baniahmad, A. (2007). Alien inhibits E2F1 gene expression and cell

(1997). HBP1: a HMG box transcriptional repressor that is targeted by the

elution from IMAC), a phosphoproteomics strategy for the rapid separation of monophosphorylated from multiply phosphorylated peptides. *Mol.Cell Proteomics,*

a member of the E2F family that can behave as a transcriptional repressor.

component of the mammalian Bmi1-containing polycomb complex.

485-494.

21 (21): 5775-5786.

(10): e13743.

7 (4): 661-671.

(1): 11-20.

genes. *Cell Cycle,* 6 (15): 1903-1913.

damage. *Nat.Cell Biol.,* 5 (5): 401-409.

*Nat.Rev.Cancer,* 3 (9): 650-665.

suppress tumor cell growth. *Genes Dev.,* 12 (1): 95-106.

proliferation. *Biochim.Biophys.Acta,* 1773 (9): 1447-1454.

retinoblastoma family. *Genes Dev.,* 11 (3): 383-396.

*Proc.Natl.Acad.Sci.U.S.A,* 95 (6): 2850-2855.

*Proc.Natl.Acad.Sci.U.S.A,* 98 (4): 1519-1524.

Comparative protein profiling reveals minichromosome maintenance (MCM) proteins as novel potential tumor markers for meningiomas. *J.Proteome Res.,* 9 (1):

mediated by RYBP, provides a mechanism for specificity of E2F function. *EMBO J.,*

Hanel, F., Brehm, A., & Gaubatz, S. (2007). LINC, a human complex that is related to pRB-containing complexes in invertebrates regulates the expression of G2/M

& Kaelin, W. G., Jr. (1998). Stable binding to E2F is not required for the retinoblastoma protein to activate transcription, promote differentiation, and

Willer, G. B., & Balice-Gordon, R. J. (2010). Neural and synaptic defects in slytherin, a zebrafish model for human congenital disorders of glycosylation. *PLoS.One.,* 5

involved in differentiation, development, proliferation, and apoptosis. *Genes Dev.,* 15 (3): 267-285.


Murga, M., Fernandez-Capetillo, O., Field, S. J., Moreno, B., Borlado, L. R., Fujiwara, Y.,

Ogawa, H., Ishiguro, K., Gaubatz, S., Livingston, D. M., & Nakatani, Y. (2002). A complex

Olsen, J. V., Blagoev, B., Gnad, F., Macek, B., Kumar, C., Mortensen, P., & Mann, M. (2006).

Ong, S. E., Blagoev, B., Kratchmarova, I., Kristensen, D. B., Steen, H., Pandey, A., & Mann,

Peeper, D. S., Keblusek, P., Helin, K., Toebes, M., van der Eb, A. J., & Zantema, A. (1995).

Pinkse, M. W., Uitto, P. M., Hilhorst, M. J., Ooms, B., & Heck, A. J. (2004). Selective isolation

Polager, S., Ofir, M., & Ginsberg, D. (2008). E2F1 regulates autophagy and the transcription

Rabinovich, A., Jin, V. X., Rabinovich, R., Xu, X., & Farnham, P. J. (2008). E2F in vivo binding

Ren, B., Cam, H., Takahashi, Y., Volkert, T., Terragni, J., Young, R. A., & Dynlacht, B. D.

Rigbolt, K. T., Prokhorova, T. A., Akimov, V., Henningsen, J., Johansen, P. T., Kratchmarova,

Ross, P. L., Huang, Y. N., Marchese, J. N., Williamson, B., Parker, K., Hattan, S., Khainovski,

Saydam, O., Senol, O., Schaaij-Visser, T. B., Pham, T. V., Piersma, S. R., Stemmer-

Nevins, J. R. (2001). The Rb/E2F pathway and cancer. *Hum.Mol.Genet.,* 10 (7): 699-703. Novy, M., Pohn, R., Andorfer, P., Novy-Weiland, T., Galos, B., Schwarzmayr, L., &

dependent transcription. *Mol.Biol.Cell,* 16 (5): 2181-2190.

and promotes pRB-binding in vitro. *Oncogene,* 10 (1): 39-48.

of autophagy genes. *Oncogene,* 27 (35): 4860-4864.

G(2)/M checkpoints. *Genes Dev.,* 16 (2): 245-256.

cell differentiation. *Sci.Signal.,* 4 (164): rs3.

reagents. *Mol.Cell Proteomics,* 3 (12): 1154-1169.

15 (3): 267-285.

*Science,* 296 (5570): 1132-1136.

*Cell,* 127 (3): 635-648.

*Res.,* 18 (11): 1763-1777.

970.

3943.

involved in differentiation, development, proliferation, and apoptosis. *Genes Dev.,*

Balomenos, D., Vicario, A., Carrera, A. C., Orkin, S. H., Greenberg, M. E., & Zubiaga, A. M. (2001). Mutation of E2F2 in mice causes enhanced T lymphocyte proliferation, leading to the development of autoimmunity. *Immunity,* 15 (6): 959-

Rotheneder, H. (2005). EAPP, a novel E2F binding protein that modulates E2F-

with chromatin modifiers that occupies E2F- and Myc-responsive genes in G0 cells.

Global, in vivo, and site-specific phosphorylation dynamics in signaling networks.

M. (2002). Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics. *Mol.Cell Proteomics,* 1 (5): 376-386.

Phosphorylation of a specific cdk site in E2F-1 affects its electrophoretic mobility

at the femtomole level of phosphopeptides from proteolytic digests using 2D-NanoLC-ESI-MS/MS and titanium oxide precolumns. *Anal.Chem.,* 76 (14): 3935-

specificity: comparison of consensus versus nonconsensus binding sites. *Genome* 

(2002). E2F integrates cell cycle progression with DNA repair, replication, and

I., Kassem, M., Mann, M., Olsen, J. V., & Blagoev, B. (2011). System-wide temporal characterization of the proteome and phosphoproteome of human embryonic stem

N., Pillai, S., Dey, S., Daniels, S., Purkayastha, S., Juhasz, P., Martin, S., Bartlet-Jones, M., He, F., Jacobson, A., & Pappin, D. J. (2004). Multiplexed protein quantitation in Saccharomyces cerevisiae using amine-reactive isobaric tagging

Rachamimov, A. O., Wurdinger, T., Peerdeman, S. M., & Jimenez, C. R. (2010).

Comparative protein profiling reveals minichromosome maintenance (MCM) proteins as novel potential tumor markers for meningiomas. *J.Proteome Res.,* 9 (1): 485-494.


**8** 

*Italy* 

**F0F1 ATP Synthase:** 

**A Fascinating Challenge for Proteomics** 

Federica Dabbeni-Sala1, Amit Kumar Rai1 and Giovanna Lippe2

The aim of this review is to provide insight and encouragement into the development of new proteomic approaches aimed at analyzing the relationship between structure and function of ATP synthase in different organisms and under different metabolic conditions. Particular attention will be paid to the preparation of the sample for the mass spectrometry

F0F1ATP synthase is the terminal enzyme of the oxidative phosphorylation pathway (named complex V of the OXPHOS system) that is responsible for the majority of ATP synthesis in all living cells (Boyer, 1997). It is an exceptionally complicated protein complex, whose molecular mass varies from 540 to 585 kDa depending on the source, which is organized into a globular catalytic part (F1) and a membranous moiety (F0) linked by central and

The enzyme is present in bacterial plasma membranes and chloroplast thylakoids, where it contains 8 and 9 subunits, respectively (Borghese et al,. 1998; Richter et al,. 2000), and in mitochondria, where it is located in the inner membrane and consists of at least 15 and 17 different subunits in mammals and yeasts, respectively (Wittig and Schägger, 2008). The F1 sector always consists of five subunits α3β3γ1δ1ε1 and the α- and β-subunits are arranged alternatively, forming a hexagonal cylinder around the coiled-coil structure of the γ subunit. The F0 part, which is responsible for ion translocation across the membrane, has instead a variable composition. The simplest form is present in bacteria and consists of the subunits ac10-14. The c subunits form a ring structure with variable stoichiometry among species connected to F1 by the central stalk, constituted by the subunits γ and ε, latter being homologous to subunit δ of the mitochondrial enzyme (Vignais and Satre, 1984) and able to modulate the catalysis (Suzuki et al,. 2011). The a subunit associates with the c-ring peripherally and with the lateral stalk, which is formed by the homo-dimer of subunits b and by subunit δ, present in single copy and located at the top of F1(Walker and Dickson,

In mitochondria the additional subunits d, e, f, g, A6L, F6 and OSCP are associated to the complex, of which the subunits b, d, F6 and OSCP form the lateral stalk in single copies, OSCP homologous to prokaryotic subunit δ (Walker and Dickson, 2006). The so-called

(MS) analyses, which is also a critical step and requires a specific competence.

**2. The nano-motor enzyme F0F1ATP synthase** 

**1. Introduction** 

peripheral stalks.

2006)(Fig.1A).

*1Department of Pharmacology, University of Padova, Padova 2Department of Food Science, University of Udine, Udine* 


### **F0F1 ATP Synthase: A Fascinating Challenge for Proteomics**

Federica Dabbeni-Sala1, Amit Kumar Rai1 and Giovanna Lippe2 *1Department of Pharmacology, University of Padova, Padova 2Department of Food Science, University of Udine, Udine Italy* 

#### **1. Introduction**

160 Proteomics – Human Diseases and Protein Functions

Trouche, D., Cook, A., & Kouzarides, T. (1996). The CBP co-activator stimulates E2F1/DP1

Van, H. D., Munoz, J., Braam, S. R., Pinkse, M. W., Linding, R., Heck, A. J., Mummery, C. L.,

Vandel, L. & Kouzarides, T. (1999). Residues phosphorylated by TFIIH are required for E2F-

Weinberg, R. A. (1995). The retinoblastoma protein and cell cycle control. *Cell,* 81 (3): 323-

Weinmann, A. S., Yan, P. S., Oberley, M. J., Huang, T. H., & Farnham, P. J. (2002). Isolating

Whitfield, M. L., George, L. K., Grant, G. D., & Perou, C. M. (2006). Common markers of

Xie, Q., Bai, Y., Wu, J., Sun, Y., Wang, Y., Zhang, Y., Mei, P., & Yuan, Z. (2011). Methylation-

Yamasaki, L., Jacks, T., Bronson, R., Goillot, E., Harlow, E., & Dyson, N. J. (1996). Tumor induction and tissue atrophy in mice lacking E2F-1. *Cell,* 85 (4): 537-548. Yao, Q., Li, H., Liu, B. Q., Huang, X. Y., & Guo, L. (2011). SUMOylation-regulated Protein

Young, A. P., Nagarajan, R., & Longmore, G. D. (2003). Mechanisms of transcriptional

Zhang, K., Wrzesinski, K., Stephen, J. F., Larsen, P. M., Zhang, X., & Roepstorff, P. (2008).

Zhao, Y. & Jensen, O. N. (2009). Modification-specific proteomics: strategies for

Zhu, J. W., Field, S. J., Gore, L., Thompson, M., Yang, H., Fujiwara, Y., Cardiff, R. D.,

& Krijgsveld, J. (2009). Phosphorylation dynamics during early differentiation of

human transcription factor targets by coupling chromatin immunoprecipitation

mediated regulation of E2F1 in DNA damage-induced cell death.

Phosphorylation, Evidence from Quantitative Phosphoproteomics Analyses.

regulation by Rb-E2F segregate by biological pathway. *Oncogene,* 22 (46): 7209-7217.

Comparative proteome analysis of three mouse lung adenocarcinoma CMT cell lines with different metastatic potential by two-dimensional gel electrophoresis and

characterization of post-translational modifications using enrichment techniques.

Greenberg, M., Orkin, S. H., & DeGregori, J. (2001). E2F1 and E2F2 determine thresholds for antigen-induced T-cell proliferation and suppress tumorigenesis.

activity. *Nucleic Acids Res.,* 24 (21): 4139-4145.

proliferation. *Nat.Rev.Cancer,* 6 (2): 99-106.

*J.Recept.Signal.Transduct.Res.,* 31 (2): 139-146.

mass spectrometry. *Proteomics,* 8 (23-24): 4932-4945.

*J.Biol.Chem.,* 286 (31): 27342-27349.

*Proteomics,* 9 (20): 4632-4641.

*Mol.Cell Biol.,* 21 (24): 8547-8564.

330.

human embryonic stem cells. *Cell Stem Cell,* 5 (2): 214-226.

1 degradation during S-phase. *EMBO J.,* 18 (15): 4280-4291.

and CpG island microarray analysis. *Genes Dev.,* 16 (2): 235-244.

The aim of this review is to provide insight and encouragement into the development of new proteomic approaches aimed at analyzing the relationship between structure and function of ATP synthase in different organisms and under different metabolic conditions. Particular attention will be paid to the preparation of the sample for the mass spectrometry (MS) analyses, which is also a critical step and requires a specific competence.

#### **2. The nano-motor enzyme F0F1ATP synthase**

F0F1ATP synthase is the terminal enzyme of the oxidative phosphorylation pathway (named complex V of the OXPHOS system) that is responsible for the majority of ATP synthesis in all living cells (Boyer, 1997). It is an exceptionally complicated protein complex, whose molecular mass varies from 540 to 585 kDa depending on the source, which is organized into a globular catalytic part (F1) and a membranous moiety (F0) linked by central and peripheral stalks.

The enzyme is present in bacterial plasma membranes and chloroplast thylakoids, where it contains 8 and 9 subunits, respectively (Borghese et al,. 1998; Richter et al,. 2000), and in mitochondria, where it is located in the inner membrane and consists of at least 15 and 17 different subunits in mammals and yeasts, respectively (Wittig and Schägger, 2008). The F1 sector always consists of five subunits α3β3γ1δ1ε1 and the α- and β-subunits are arranged alternatively, forming a hexagonal cylinder around the coiled-coil structure of the γ subunit. The F0 part, which is responsible for ion translocation across the membrane, has instead a variable composition. The simplest form is present in bacteria and consists of the subunits ac10-14. The c subunits form a ring structure with variable stoichiometry among species connected to F1 by the central stalk, constituted by the subunits γ and ε, latter being homologous to subunit δ of the mitochondrial enzyme (Vignais and Satre, 1984) and able to modulate the catalysis (Suzuki et al,. 2011). The a subunit associates with the c-ring peripherally and with the lateral stalk, which is formed by the homo-dimer of subunits b and by subunit δ, present in single copy and located at the top of F1(Walker and Dickson, 2006)(Fig.1A).

In mitochondria the additional subunits d, e, f, g, A6L, F6 and OSCP are associated to the complex, of which the subunits b, d, F6 and OSCP form the lateral stalk in single copies, OSCP homologous to prokaryotic subunit δ (Walker and Dickson, 2006). The so-called

F0F1 ATP Synthase: A Fascinating Challenge for Proteomics 163

Stoichiometry Prokaryotes Eukaryotes

3 α α α α 3 β β β β 1 γ γ γ γ 1 ε δ δ δ 1 - ε ε ε

1 δ 5 OSCP OSCP 1 a 6 a a 1 - 8 A6L A6L 10-14 c 9 c c 1-2 b 4 b b 1 - d d d 1 - h F6 F6 1 - f f f 1-2 - e e e n.d\* - g g g

1 i - n.d\* - k - -

n.d\* Stf1 - -

fact.B -

AGP or DAPIT

MLQ or 6.8 PL

DAPIT

6.8 PL

1 - - Coupl.

Protein 0-1 - Inh1 IF1 IF1

Table 1. Subunit composition and nomenclature of the ATP synthases from prokaryotes (*Escherichia coli*) and Eukaryotes (*Saccharomyces cerevisiae*, *Bos taurus* and *Homo sapiens*).

direction of ATP hydrolysis, sustaining the formation of the proton gradient, when there is

First direct visualization of ATP-driven rotation of *Bacillus* F1 immobilized on the glass surface via the N-termini of its β-subunits was obtained more than 10 years ago (Noji et al,. 1997). These experiments showed that a fluorescent actin filament attached on the γ-subunit rotates uni-directionally, counterclockwise when viewed from membrane side, upon addition of ATP. A further technical sophistication used a sub-millisecond resolution camera to detect the rotation of gold beads attached to the γ subunit of the α3β3γ subcomplex along with fluorescence changes of an ATP hydrolysable analog. This technique allowed to display in real time the binding and release of nucleotides at the three catalytic sites simultaneously with the γ rotation (Adachi et al,. 2007). The rotation probe reports a pause, which corresponds to the period during which ATP binds to the empty catalytic site,

n.d\* - Stf2 -



*E.coli S. cerevisiae B. taurus H. sapiens* 

Homologous subunits

F1

F0

Species specific subunits

Inhibitor

Stabilizing factor for IF1

Associated proteins

\*n.d.- not determined

loss of membrane potential (Fig. 2).

(A) Bacterial (*E.coli*) enzyme is assembled with a stiochiometry of α (3), β (3), γ, δ, ε, b (2), a, c (10-14). The central stalk is composed of γ and ε subunits, the peripheral one is composed of subunits b and δ, the proton channel is formed by the subunits ac10-14.

(B) In the eukaryotic enzyme (*Bos taurus*) the central stalk is formed by subunits γ, δ and ε, the peripheral stalk is composed of subunits OSCP, b, d and F6 and the additional subunits A6L, e, f, g are associated to the proton channel ac10.

Fig. 1. Schematic representation of bacterial and mitochondrial F0F1 ATP synthase.

minor subunits e, f, g and A6L all span the membrane and, apart for subunit e (Bisetto et al, 2008), their exact stoichiometries are poorly defined (Fig. 1B). Other subunits are speciesspecific, such as subunit i and k in *Saccharomyces cerevisiae* (Wittig and Schägger, 2008) and coupling factor B in *Bos taurus* (Lee et al,. 2008).

Besides these subunits, in some mammals such as beef, rat and man two hydrophobic proteins namely MLQ/6.8-kDa proteolipid (Chen et al,. 2007; Meyer et al,. 2007), and AGP/DAPIT (Ohsakaya et al,. 2011) are associated to the F0 part when phospholipids are not extracted. All together, the mitochondrial membrane domain is constituted by approximately 30 trans-membrane α-helices (Carroll et al,. 2009). In addition, the mitochondrial complex can bind the inhibitor protein IF1, which reversibly binds to F1 with a 1:1 stoichiometry and fully inhibits the enzyme activity (Bason et al,. 2011; Harris and Das, 1991). In yeast, along with IF1, the enzyme can be regulated by two additional proteins, namely Stf1 and Stf2 (stabilizing factor 1 and 2) (Andrianaivomananjaona et al,. 2010). The subunit composition of the ATP synthase from prokaryotic and eukaryotic sources along with the subunit homology and the corresponding nomenclature is reported in Table 1. The molecular masses of the different subunits are reported in Table 2, based on ATP synthase from *Bos taurus* and *Bacillus pseudofirmus*.

Despite the differences in the complexity, functionally important subunits are conserved and in all sources the enzyme catalyses the synthesis of ATP by using the energy of the electrochemical gradient of protons (or less commonly of sodium) generated by the respiratory chain. F0F1 is an unusually efficient rotary motor that synthesizes ATP at rates exceeding 100 molecules per second (Senior, 2007). Protons traveling down the H+ gradient generate the rotation of the F0 c-ring, making the central stalk also rotates, which in turn drives ATP synthesis from ADP and Pi by forcing different conformations sequentially on each of the catalytic sites in the three F1 subunits β. The enzyme is able to work in the

(A) Bacterial (*E.coli*) enzyme is assembled with a stiochiometry of α (3), β (3), γ, δ, ε, b (2), a, c (10-14). The central stalk is composed of γ and ε subunits, the peripheral one is composed of subunits b and δ,

minor subunits e, f, g and A6L all span the membrane and, apart for subunit e (Bisetto et al, 2008), their exact stoichiometries are poorly defined (Fig. 1B). Other subunits are speciesspecific, such as subunit i and k in *Saccharomyces cerevisiae* (Wittig and Schägger, 2008) and

Besides these subunits, in some mammals such as beef, rat and man two hydrophobic proteins namely MLQ/6.8-kDa proteolipid (Chen et al,. 2007; Meyer et al,. 2007), and AGP/DAPIT (Ohsakaya et al,. 2011) are associated to the F0 part when phospholipids are not extracted. All together, the mitochondrial membrane domain is constituted by approximately 30 trans-membrane α-helices (Carroll et al,. 2009). In addition, the mitochondrial complex can bind the inhibitor protein IF1, which reversibly binds to F1 with a 1:1 stoichiometry and fully inhibits the enzyme activity (Bason et al,. 2011; Harris and Das, 1991). In yeast, along with IF1, the enzyme can be regulated by two additional proteins, namely Stf1 and Stf2 (stabilizing factor 1 and 2) (Andrianaivomananjaona et al,. 2010). The subunit composition of the ATP synthase from prokaryotic and eukaryotic sources along with the subunit homology and the corresponding nomenclature is reported in Table 1. The molecular masses of the different subunits are reported in Table 2, based on ATP synthase

Despite the differences in the complexity, functionally important subunits are conserved and in all sources the enzyme catalyses the synthesis of ATP by using the energy of the electrochemical gradient of protons (or less commonly of sodium) generated by the respiratory chain. F0F1 is an unusually efficient rotary motor that synthesizes ATP at rates exceeding 100 molecules per second (Senior, 2007). Protons traveling down the H+ gradient generate the rotation of the F0 c-ring, making the central stalk also rotates, which in turn drives ATP synthesis from ADP and Pi by forcing different conformations sequentially on each of the catalytic sites in the three F1 subunits β. The enzyme is able to work in the

(B) In the eukaryotic enzyme (*Bos taurus*) the central stalk is formed by subunits γ, δ and ε, the peripheral stalk is composed of subunits OSCP, b, d and F6 and the additional subunits A6L, e, f, g are

Fig. 1. Schematic representation of bacterial and mitochondrial F0F1 ATP synthase.

the proton channel is formed by the subunits ac10-14.

coupling factor B in *Bos taurus* (Lee et al,. 2008).

from *Bos taurus* and *Bacillus pseudofirmus*.

associated to the proton channel ac10.


\*n.d.- not determined

Table 1. Subunit composition and nomenclature of the ATP synthases from prokaryotes (*Escherichia coli*) and Eukaryotes (*Saccharomyces cerevisiae*, *Bos taurus* and *Homo sapiens*).

direction of ATP hydrolysis, sustaining the formation of the proton gradient, when there is loss of membrane potential (Fig. 2).

First direct visualization of ATP-driven rotation of *Bacillus* F1 immobilized on the glass surface via the N-termini of its β-subunits was obtained more than 10 years ago (Noji et al,. 1997). These experiments showed that a fluorescent actin filament attached on the γ-subunit rotates uni-directionally, counterclockwise when viewed from membrane side, upon addition of ATP. A further technical sophistication used a sub-millisecond resolution camera to detect the rotation of gold beads attached to the γ subunit of the α3β3γ subcomplex along with fluorescence changes of an ATP hydrolysable analog. This technique allowed to display in real time the binding and release of nucleotides at the three catalytic sites simultaneously with the γ rotation (Adachi et al,. 2007). The rotation probe reports a pause, which corresponds to the period during which ATP binds to the empty catalytic site,

F0F1 ATP Synthase: A Fascinating Challenge for Proteomics 165

Fig. 2. Schematic representation of ATP synthesis and hydrolysis by F0F1 ATP synthase (A) Proton powered rotation of c ring makes the central stalk turn with it, generating torque and conformational changes in the catalytic αβ domain to synthesize ATP from ADP and Pi. (B)

Table 3. Assembly factors of yeast and human F0F1 ATP Synthase. The assembly factors for

Atp22p) (Helfenbein et al,. 2003; Rak et al,. 2011). In mammalian cells only two factors are

The assembly process is best characterized in yeast, where recent *in organello* pulse-labeling and pulse-chase experiments have enabled to identify three different assembly intermediates and to demonstrate that the whole enzyme is formed by two separate

Eukaryotes Factors Mass (Da) Swiss-Prot Accession Yeast Fmc1p ≤18364 P40491 Yeast Atp10p 32093.91 P18496 Yeast ATP11p ≤36581 P32453 Yeast ATP12p ≤36554 P22135 Yeast ATP22p ≤79756 A6ZYV0 Yeast ATP23p 26890.37 P53722 Yeast Mdm38p 58610.83 Q08179 Yeast Aep3p 70310.03 Q12089 Yeast Oxa 1p 40000.09 P39952 Human ATPF2 32772 **Q8N5M1\***  Human ATPF1 36437 Q5TC12\*

The hydrolysis of ATP sustains proton flow in opposite direction.

F1 are in blue and F0 are in red (Wittig and Schägger, 2008).

known, which are orthologus to yeast F1 assembly factors (Table 3).

\* UNIPROT accession number


\*Accession numbers were obtained from UniProt

+Molecular masses are shown in Dalton

aAssuming the assigned stoichiometry for *Bos taurus* including the subunits e1 and g1 (excluding the amino-terminal modifications). If proteins MLQ and AGP are considered, the total protein mass increases to 596579 Da (Wittig and Schägger, 2008).

bAccording to the stoichiometry (α3β3γδεab2c13) for *Bacillus pseudofirmus* 

Table 2. Molecular masses of subunits of *Bos taurus* and *Bacillus pseudofirmus* F0F1 ATP synthase

and a 120° step rotation, constituted by two sub-steppings, whose duration is still debated, during which ATP hydrolysis and release occur. Single molecule technology studies have been applied also to the whole F0F1 complexes from *Propionigenium modestum* and *Escherichia coli.* Rotation was probed with probes attached to the c-ring in the immobilized F0F1 and, as expected, occurred in the opposite direction when c-ring rotation was driven by ATP or by proton-flow (Ueno et al,. 2005).

#### **2.1 ATP synthase biogenesis**

The mitochondrial F0F1 complex is composed of both nuclear and mitochondrial gene products. In yeast the three F0 core proteins a (Su6), A6L (Su8) and c (Su9) are encoded by mDNA, while in mammals only subunits a and A6L are encoded by mitochondrial genome. This arrangement highlights the complexity of enzyme assembly, which requires accessory factors, whose definition is still under investigation (Wittig and Schägger, 2008). Altogether 9 factors have been identified in yeast, but, until now, the role of only five of them has been defined. Three factors mediate the F1 formation (Atp11p, Atp12p and possibly Fmc1p) (Ackerman, 2002; Lefebvre-Legendre et al,. 2001) and two the F0 assembly (Atp10p and

α P19483 55263.39 AAG48361.1 54674.48 β P00829 51562.97 AAG48363.1 51752.14 γ P05631 30255.71 AAG48362.1 31835.49 δ P05630 15064.93 AAG48360.1 20534.63 ε P05632 5651.67 AAG48364.1 14327.68 a P00847 24787.91 AAG48358.1 26863.77 b P13619 24668.72 AAG48359.1 18510.16 c P32876 14223 AAC08039.1 6956.06 c P07926 15029 - c Q3ZC75 14693 - d P13620 18561.28 - e Q00361 8189.47 - f Q28851 10165.99 - g Q28852 11286.26 - - 8 or A6L P03929 7936.56 - - F6 P02721 8958.09 - - OSCP P13621 20929.75 - -

Accession Numbers\* Molecular mass+

Subunits *Bos taurus Bacillus pseudofirmus*

Total F0F1 583442a 540290.53b

aAssuming the assigned stoichiometry for *Bos taurus* including the subunits e1 and g1 (excluding the amino-terminal modifications). If proteins MLQ and AGP are considered, the total protein mass

Table 2. Molecular masses of subunits of *Bos taurus* and *Bacillus pseudofirmus* F0F1 ATP

and a 120° step rotation, constituted by two sub-steppings, whose duration is still debated, during which ATP hydrolysis and release occur. Single molecule technology studies have been applied also to the whole F0F1 complexes from *Propionigenium modestum* and *Escherichia coli.* Rotation was probed with probes attached to the c-ring in the immobilized F0F1 and, as expected, occurred in the opposite direction when c-ring rotation was driven by ATP or by

The mitochondrial F0F1 complex is composed of both nuclear and mitochondrial gene products. In yeast the three F0 core proteins a (Su6), A6L (Su8) and c (Su9) are encoded by mDNA, while in mammals only subunits a and A6L are encoded by mitochondrial genome. This arrangement highlights the complexity of enzyme assembly, which requires accessory factors, whose definition is still under investigation (Wittig and Schägger, 2008). Altogether 9 factors have been identified in yeast, but, until now, the role of only five of them has been defined. Three factors mediate the F1 formation (Atp11p, Atp12p and possibly Fmc1p) (Ackerman, 2002; Lefebvre-Legendre et al,. 2001) and two the F0 assembly (Atp10p and

Molecular mass+

Accession Numbers\*

\*Accession numbers were obtained from UniProt

increases to 596579 Da (Wittig and Schägger, 2008).

bAccording to the stoichiometry (α3β3γδεab2c13) for *Bacillus pseudofirmus* 

+Molecular masses are shown in Dalton

proton-flow (Ueno et al,. 2005).

**2.1 ATP synthase biogenesis** 

synthase

Fig. 2. Schematic representation of ATP synthesis and hydrolysis by F0F1 ATP synthase (A) Proton powered rotation of c ring makes the central stalk turn with it, generating torque and conformational changes in the catalytic αβ domain to synthesize ATP from ADP and Pi. (B) The hydrolysis of ATP sustains proton flow in opposite direction.


\* UNIPROT accession number

Table 3. Assembly factors of yeast and human F0F1 ATP Synthase. The assembly factors for F1 are in blue and F0 are in red (Wittig and Schägger, 2008).

Atp22p) (Helfenbein et al,. 2003; Rak et al,. 2011). In mammalian cells only two factors are known, which are orthologus to yeast F1 assembly factors (Table 3).

The assembly process is best characterized in yeast, where recent *in organello* pulse-labeling and pulse-chase experiments have enabled to identify three different assembly intermediates and to demonstrate that the whole enzyme is formed by two separate

F0F1 ATP Synthase: A Fascinating Challenge for Proteomics 167

Bauza et al,. 2011). These isoforms differ in their cleavable mitochondrial targeting peptides, whereas the mature peptides are identical. Considering that in mammals c genes determine the ATP synthase content, the existence of iso-genes would be advantageous for regulation of subunit c synthesis, and thus ATP synthase biogenesis, by multiple factors. It appears that

Biochemical evidence and electron microscopy studies recently demonstrated that within the inner mitochondrial membrane the enzyme is organized in dimers and oligomers, which possibly associate with other inner mitochondrial membrane proteins, e.g. with phosphate and adenine nucleotide carriers in the "phosphorylating assemblies"– the so called ATP synthasome (Chen et al,. 2004; Wittig and Schägger, 2008). Elucidation of dimer/oligomers structural properties and of their formation process is quite important. In fact, proposed roles of the ATP synthase oligomers are higher efficiency and higher stability. In this regard, we demonstrated that dimers have a greater specific activity than monomers (Bisetto et al,. 2007). In accordance, a recent numerical simulation indicated a significant increase in charge density in regions of high membrane curvature induced by ATP synthase dimerization, thus favoring effective ATP synthesis under proton-limiting conditions (Fig. 4) (Strauss et al,. 2008). Moreover, these oligomers appear to play a special role for mitochondrial morphology, being involved in cristae formation (Couoh-Cardel et al,. 2010; Paumard et al,.

The structural properties of dimers/oligomers were initially characterized in yeast where genetic approaches, cross-linking analyses and electron microscopy images established preferential interactions within the inner membrane (Thomas et al,. 2008) mainly through the subunits Su6 (Wittig et al,. 2008), Su4 (Spannagel et al,. 1998), e (Everard-Gigot et al,. 2005) and g (Bustos and Velours, 2005), which are conserved in mammals, and also through the F0 subunits h and i in yeast (Fronzes et al,. 2006). High-resolution images showed that both in yeast and mammals the dimers display angles between two F1-F1 ranging from 35° to 180°. Recent images of yeast dimers at 27 Å resolution showed that the dominant angle is

42°, suggesting that this is the most stable conformation (Couoh-Cardel et al., 2010).

We recently demonstrated by a structural proteomic approach that also in mammals the e subunit is essential for ATP synthase self-association in dimers and oligomers. Selective degradation by *in situ* limited proteolysis caused an alteration of the oligomeric distribution of ATP synthase by eliminating oligomers and reducing dimers in favor of monomers

A critical aspect of F0F1 dimerization is related to the role of IF1, which is still controversial (Wittig and Schägger, 2008). IF1 is well known to bind ATP synthase under energy deficiency, i.e. at low pH and membrane potential, when the enzyme hydrolyzes rather than synthesizes ATP. Therefore, IF1 is considered responsible for the beneficial down-regulation of F0F1 during ischemia in *in vitro* experimental models, but also *in vivo*, as demonstrated by our group in anaesthetized open-chest goat heart (Di Pancrazio et al,. 2004). Nevertheless, because isolated IF1 from bovine heart is present in dimeric form in solution where it has been shown to link two F1-subcomplexes (Cabezón et al,. 2003), it seemed conceivable that dimeric IF1 might also be able to link two F0F1 complexes in the inner mitochondrial membrane. However, in yeast deletion of IF1 and of the associated proteins Stf1 and Stf2 did not eliminate dimers and oligomers (Dienhart et al,. 2002), thus excluding an essential role of IF1 in ATP synthase self-association. On the other hand, a very recent study revealed that

much remains to be learned about this argument.

**2.2 ATP synthase self-association** 

2002).

(Bisetto et al., 2008).

pathways that converge to form the ATP synthase from their respective end-products. One pathway leads to the formation of F1, which was already known to assemble as an independent unit (Tzagoloff, 1969), and of the Su9-ring. These two sub-complexes subsequently combine to constitute the F1/Su9-ring end-product. The other pathway leads to the formation of the Su6/Su8/stator sub-complex, which, in addition to Su6 and Su8, contains the chaperone Atp10p and additional still undefined proteins of the lateral stalk (Rak et al., 2011) (Fig. 3).

Because in yeast the interaction between Su6 and Su8 is kinetically much more rapid, the entire process is regulated by the control of the Su6 and Su8 translation by F1 in order to obtain a balanced production of the different intermediates (Rak et al., 2011). Conversely, in mammals the amount of ATP synthase seems to be controlled by the availability of subunit c, as demonstrated in brown fat (Houstek et al,. 1995) and other tissues (Andersson et al,. 1997).

Independent of the mechanism, it has been proposed that the ATP synthase assembly may recapitulate some of the evolutionary events that gave rise to this enzyme. In fact, there is evidence that F1 evolved from an ATP-dependent helicase (Gomis-Rüth et al,. 2001), while the Su9 derived from an ion channel (Rak et al., 2011), so that their combination converted a passive channel into an active pump.

Fig. 3. Assembly of F0F1 ATP synthase in yeast. The scheme shows the two separate pathways, leading to two separate end-products, i.e. the F1/Su9-ring and Su6/Su8/stator sub-complexes that converge at the end to form the whole F0F1 complex (Rak et al., 2011).

So far there is little evidence for tissue-specific or developmentally regulated isoforms of ATP synthase subunits. In tobacco plants three isoforms of the F1 subunit β have been identified, of which only one is exclusively expressed in pollen (Lalanne et al,. 1998). In mammals, two tissue-specific isoforms of the F1 γ subunit**,** heart and liver type, were identified in bovine F0F1-ATP synthase (Matsuda et al,. 1993). These two isoforms are generated by alternative splicing and their Vmax and Km are identical (Matsuda et al,. 1994). In addition, three isoforms (P1, P2, and P3) of the F0 subunit c have been identified (VivesBauza et al,. 2011). These isoforms differ in their cleavable mitochondrial targeting peptides, whereas the mature peptides are identical. Considering that in mammals c genes determine the ATP synthase content, the existence of iso-genes would be advantageous for regulation of subunit c synthesis, and thus ATP synthase biogenesis, by multiple factors. It appears that much remains to be learned about this argument.

#### **2.2 ATP synthase self-association**

166 Proteomics – Human Diseases and Protein Functions

pathways that converge to form the ATP synthase from their respective end-products. One pathway leads to the formation of F1, which was already known to assemble as an independent unit (Tzagoloff, 1969), and of the Su9-ring. These two sub-complexes subsequently combine to constitute the F1/Su9-ring end-product. The other pathway leads to the formation of the Su6/Su8/stator sub-complex, which, in addition to Su6 and Su8, contains the chaperone Atp10p and additional still undefined proteins of the lateral stalk

Because in yeast the interaction between Su6 and Su8 is kinetically much more rapid, the entire process is regulated by the control of the Su6 and Su8 translation by F1 in order to obtain a balanced production of the different intermediates (Rak et al., 2011). Conversely, in mammals the amount of ATP synthase seems to be controlled by the availability of subunit c, as demonstrated in brown fat (Houstek et al,. 1995) and other tissues (Andersson et al,.

Independent of the mechanism, it has been proposed that the ATP synthase assembly may recapitulate some of the evolutionary events that gave rise to this enzyme. In fact, there is evidence that F1 evolved from an ATP-dependent helicase (Gomis-Rüth et al,. 2001), while the Su9 derived from an ion channel (Rak et al., 2011), so that their combination converted a

Fig. 3. Assembly of F0F1 ATP synthase in yeast. The scheme shows the two separate pathways, leading to two separate end-products, i.e. the F1/Su9-ring and Su6/Su8/stator sub-complexes that converge at the end to form the whole F0F1 complex (Rak et al., 2011). So far there is little evidence for tissue-specific or developmentally regulated isoforms of ATP synthase subunits. In tobacco plants three isoforms of the F1 subunit β have been identified, of which only one is exclusively expressed in pollen (Lalanne et al,. 1998). In mammals, two tissue-specific isoforms of the F1 γ subunit**,** heart and liver type, were identified in bovine F0F1-ATP synthase (Matsuda et al,. 1993). These two isoforms are generated by alternative splicing and their Vmax and Km are identical (Matsuda et al,. 1994). In addition, three isoforms (P1, P2, and P3) of the F0 subunit c have been identified (Vives-

(Rak et al., 2011) (Fig. 3).

passive channel into an active pump.

1997).

Biochemical evidence and electron microscopy studies recently demonstrated that within the inner mitochondrial membrane the enzyme is organized in dimers and oligomers, which possibly associate with other inner mitochondrial membrane proteins, e.g. with phosphate and adenine nucleotide carriers in the "phosphorylating assemblies"– the so called ATP synthasome (Chen et al,. 2004; Wittig and Schägger, 2008). Elucidation of dimer/oligomers structural properties and of their formation process is quite important. In fact, proposed roles of the ATP synthase oligomers are higher efficiency and higher stability. In this regard, we demonstrated that dimers have a greater specific activity than monomers (Bisetto et al,. 2007). In accordance, a recent numerical simulation indicated a significant increase in charge density in regions of high membrane curvature induced by ATP synthase dimerization, thus favoring effective ATP synthesis under proton-limiting conditions (Fig. 4) (Strauss et al,. 2008). Moreover, these oligomers appear to play a special role for mitochondrial morphology, being involved in cristae formation (Couoh-Cardel et al,. 2010; Paumard et al,. 2002).

The structural properties of dimers/oligomers were initially characterized in yeast where genetic approaches, cross-linking analyses and electron microscopy images established preferential interactions within the inner membrane (Thomas et al,. 2008) mainly through the subunits Su6 (Wittig et al,. 2008), Su4 (Spannagel et al,. 1998), e (Everard-Gigot et al,. 2005) and g (Bustos and Velours, 2005), which are conserved in mammals, and also through the F0 subunits h and i in yeast (Fronzes et al,. 2006). High-resolution images showed that both in yeast and mammals the dimers display angles between two F1-F1 ranging from 35° to 180°. Recent images of yeast dimers at 27 Å resolution showed that the dominant angle is 42°, suggesting that this is the most stable conformation (Couoh-Cardel et al., 2010).

We recently demonstrated by a structural proteomic approach that also in mammals the e subunit is essential for ATP synthase self-association in dimers and oligomers. Selective degradation by *in situ* limited proteolysis caused an alteration of the oligomeric distribution of ATP synthase by eliminating oligomers and reducing dimers in favor of monomers (Bisetto et al., 2008).

A critical aspect of F0F1 dimerization is related to the role of IF1, which is still controversial (Wittig and Schägger, 2008). IF1 is well known to bind ATP synthase under energy deficiency, i.e. at low pH and membrane potential, when the enzyme hydrolyzes rather than synthesizes ATP. Therefore, IF1 is considered responsible for the beneficial down-regulation of F0F1 during ischemia in *in vitro* experimental models, but also *in vivo*, as demonstrated by our group in anaesthetized open-chest goat heart (Di Pancrazio et al,. 2004). Nevertheless, because isolated IF1 from bovine heart is present in dimeric form in solution where it has been shown to link two F1-subcomplexes (Cabezón et al,. 2003), it seemed conceivable that dimeric IF1 might also be able to link two F0F1 complexes in the inner mitochondrial membrane. However, in yeast deletion of IF1 and of the associated proteins Stf1 and Stf2 did not eliminate dimers and oligomers (Dienhart et al,. 2002), thus excluding an essential role of IF1 in ATP synthase self-association. On the other hand, a very recent study revealed that

F0F1 ATP Synthase: A Fascinating Challenge for Proteomics 169

because mutated F0 can translocate protons from the cytosol to the mitochondrial matrix, thus sustaining membrane potential (Sgarbi et al,. 2006). The impaired ATP synthesis mainly affects brain tissue and at high mutation load, up to approximately 95%, the heteroplasmic ATP6 gene mutations manifest as neuropathy, ataxia, retinitis pigmentosa (NARP) or as fatal encephalopathy known as Leigh syndrome (Houstek et al., 2006). Examples of quantitative defects are those in which the cellular content of the enzyme is reduced to less than 30%. Apparently, these disorders are caused by different nuclear genetic defects that remain to be identified, but most of them display a uniform fatal phenotype with onset in newborns characterized by lactic acidosis and hypertrophic cardiomyopathy (Houstek et al., 2006). In both types of ATP synthase disorders, hyperpolarization due to decreased ATP synthesis promotes ROS production by the respiratory chain, an event that can contribute to the clinical phenotypes as suggested by the beneficial effect of antioxidants observed in NARP cells (Mattiazzi et al,. 2004). This finding is quite important, considering that, in spite of the considerable progress in understanding of the molecular mechanisms of ATP synthase disorders, the available therapeutic approaches are still extremely limited (Kucharczyk et al., 2009). It has been proposed that other secondary effects possibly involved in the pathogenic pathways of ATP synthase deficiency could be changes in mitochondrial cristae morphology, which is mediated by ATP synthase oligomerization (Couoh-Cardel et al., 2010; Paumard et al., 2002), and/or a concomitant impairment of an ectopic function of ATP synthase localized

Microarray analyses have been performed in an attempt to gain a more global view of the cellular consequences of ATP synthase deficiency. In fibroblast cell lines from 13 genetically heterogeneous patients, 1632 human genes involved in mitochondrial biology, cell cycle regulation, signal transduction and apoptosis have been analysed. Surprisingly, only minor changes in expression of ATP synthase related genes were shown. Moreover, the cellular gene expression phenotypes were different depending on the site (mtDNA vs nuclear DNA) and the severity (ATP synthase content) of the underlying defect, indicating the need for further investigation of these pathways in other ATP synthase disorders (Cížková et al,. 2008). As far as our knowledge is concerned, the proteomic profiles of ATP synthase-related

Other intriguing examples of ATP synthase-related diseases are Batten disease in man or ceroid lipofuscinosis in sheep. Both are storage diseases with abnormal accumulation of subunit c in lysosomes occurring in the brain and liver, respectively. MS and protein sequencing have shown that the stored protein is structurally identical to the normal

Up- and down-regulation of ATP synthase biogenesis has been observed under different pathophysiological conditions (Houstek et al., 2006). We developed polyspecific antibodies directed against the whole human mitochondrial subproteome by hyperimmunization of rabbits with purified skeletal muscle mitochondria, which allowed detection of upregulation of ATP synthase, in muscle biopsies from patients affected by MELAS (mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes) which are characterized by a drastic reduction of OXPHOS complex 1 (Loro et al,. 2009). MS approaches have also been applied to compare the ATP synthase expression levels *in vivo* and *in vitro*. Recent examples of ATP synthase down-regulation have been obtained by the mitochondrial proteome analyses in heart from type 2 diabetic patients (Heather and Clarke,

2010) and in pancreatic β-cells exposed to high glucose (Ahmed et al,. 2010).

on cell surface (see paragraph 4) (Kucharczyk et al., 2009).

diseases have not yet been reported.

mitochondrial subunit c (Chen et al,. 2004).

Fig. 4. Model of ATP synthase dimers. Dimers mainly interact through F0 sector and enforce a strong local curvature on the inner mitochondrial membrane, where an increased charge density favors effective ATP synthesis (Strauss et al., 2008).

also in yeast oligomers contain considerable amounts of IF1, raising the question of whether bound IF1 inhibits the oligomer activity (Couoh-Cardel et al., 2010).

In bovine heart mitochondria, we demonstrated that physical release of IF1 from the inner membrane did not markedly alter the amount of dimers separated by Blue Native electrophoresis (Tomasetig et al,. 2002), suggesting that F0F1 dimers could form independently from IF1. Nevertheless, in human HeLa cells IF1 overexpression increased ATP synthase dimers, as revealed by native electrophoresis, and this was paralleled by a higher ATP synthesis efficiency (Campanella et al,. 2009). However, it should be noted that in this paper the identification and quantification of ATP synthase dimers, as well as of their IF1 content, seems questionable due to the use of dodecylmaltoside as detergent, which is well known to alter the dimer/monomer ratio (Tomasetig et al., 2002).

#### **2.3 Disorders related to ATP synthase**

In spite of the fact that the assembly process of the mitochondrial ATP synthase is still poorly characterized in humans, its defects have been recognized as a cause of human diseases (Houstek et al,. 2006; Kucharczyk et al,. 2009). Alteration of ATP synthase biogenesis leading to mitochondrial ATP synthase deficiency may cause two types of defects: qualitative when the enzyme is structurally modified and does not function properly, and quantitative when it is present in insufficient amounts. Examples of qualitative defects are those caused by missense mutations in the mitochondriallyencoded subunit a gene. Eight point mutations and a two-nucleotide micro-deletion in the ATP6 gene have been identified, of which the most common and best studied is the T8993G mutation that leads to replacement of a highly conserved leucine by arginine (Kucharczyk et al,. 2009). These mutations prevent ATP synthesis but not ATP hydrolysis

Fig. 4. Model of ATP synthase dimers. Dimers mainly interact through F0 sector and enforce a strong local curvature on the inner mitochondrial membrane, where an increased charge

also in yeast oligomers contain considerable amounts of IF1, raising the question of whether

In bovine heart mitochondria, we demonstrated that physical release of IF1 from the inner membrane did not markedly alter the amount of dimers separated by Blue Native electrophoresis (Tomasetig et al,. 2002), suggesting that F0F1 dimers could form independently from IF1. Nevertheless, in human HeLa cells IF1 overexpression increased ATP synthase dimers, as revealed by native electrophoresis, and this was paralleled by a higher ATP synthesis efficiency (Campanella et al,. 2009). However, it should be noted that in this paper the identification and quantification of ATP synthase dimers, as well as of their IF1 content, seems questionable due to the use of dodecylmaltoside as detergent, which is

In spite of the fact that the assembly process of the mitochondrial ATP synthase is still poorly characterized in humans, its defects have been recognized as a cause of human diseases (Houstek et al,. 2006; Kucharczyk et al,. 2009). Alteration of ATP synthase biogenesis leading to mitochondrial ATP synthase deficiency may cause two types of defects: qualitative when the enzyme is structurally modified and does not function properly, and quantitative when it is present in insufficient amounts. Examples of qualitative defects are those caused by missense mutations in the mitochondriallyencoded subunit a gene. Eight point mutations and a two-nucleotide micro-deletion in the ATP6 gene have been identified, of which the most common and best studied is the T8993G mutation that leads to replacement of a highly conserved leucine by arginine (Kucharczyk et al,. 2009). These mutations prevent ATP synthesis but not ATP hydrolysis

density favors effective ATP synthesis (Strauss et al., 2008).

bound IF1 inhibits the oligomer activity (Couoh-Cardel et al., 2010).

well known to alter the dimer/monomer ratio (Tomasetig et al., 2002).

**2.3 Disorders related to ATP synthase** 

because mutated F0 can translocate protons from the cytosol to the mitochondrial matrix, thus sustaining membrane potential (Sgarbi et al,. 2006). The impaired ATP synthesis mainly affects brain tissue and at high mutation load, up to approximately 95%, the heteroplasmic ATP6 gene mutations manifest as neuropathy, ataxia, retinitis pigmentosa (NARP) or as fatal encephalopathy known as Leigh syndrome (Houstek et al., 2006). Examples of quantitative defects are those in which the cellular content of the enzyme is reduced to less than 30%. Apparently, these disorders are caused by different nuclear genetic defects that remain to be identified, but most of them display a uniform fatal phenotype with onset in newborns characterized by lactic acidosis and hypertrophic cardiomyopathy (Houstek et al., 2006). In both types of ATP synthase disorders, hyperpolarization due to decreased ATP synthesis promotes ROS production by the respiratory chain, an event that can contribute to the clinical phenotypes as suggested by the beneficial effect of antioxidants observed in NARP cells (Mattiazzi et al,. 2004). This finding is quite important, considering that, in spite of the considerable progress in understanding of the molecular mechanisms of ATP synthase disorders, the available therapeutic approaches are still extremely limited (Kucharczyk et al., 2009). It has been proposed that other secondary effects possibly involved in the pathogenic pathways of ATP synthase deficiency could be changes in mitochondrial cristae morphology, which is mediated by ATP synthase oligomerization (Couoh-Cardel et al., 2010; Paumard et al., 2002), and/or a concomitant impairment of an ectopic function of ATP synthase localized on cell surface (see paragraph 4) (Kucharczyk et al., 2009).

Microarray analyses have been performed in an attempt to gain a more global view of the cellular consequences of ATP synthase deficiency. In fibroblast cell lines from 13 genetically heterogeneous patients, 1632 human genes involved in mitochondrial biology, cell cycle regulation, signal transduction and apoptosis have been analysed. Surprisingly, only minor changes in expression of ATP synthase related genes were shown. Moreover, the cellular gene expression phenotypes were different depending on the site (mtDNA vs nuclear DNA) and the severity (ATP synthase content) of the underlying defect, indicating the need for further investigation of these pathways in other ATP synthase disorders (Cížková et al,. 2008). As far as our knowledge is concerned, the proteomic profiles of ATP synthase-related diseases have not yet been reported.

Other intriguing examples of ATP synthase-related diseases are Batten disease in man or ceroid lipofuscinosis in sheep. Both are storage diseases with abnormal accumulation of subunit c in lysosomes occurring in the brain and liver, respectively. MS and protein sequencing have shown that the stored protein is structurally identical to the normal mitochondrial subunit c (Chen et al,. 2004).

Up- and down-regulation of ATP synthase biogenesis has been observed under different pathophysiological conditions (Houstek et al., 2006). We developed polyspecific antibodies directed against the whole human mitochondrial subproteome by hyperimmunization of rabbits with purified skeletal muscle mitochondria, which allowed detection of upregulation of ATP synthase, in muscle biopsies from patients affected by MELAS (mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes) which are characterized by a drastic reduction of OXPHOS complex 1 (Loro et al,. 2009). MS approaches have also been applied to compare the ATP synthase expression levels *in vivo* and *in vitro*. Recent examples of ATP synthase down-regulation have been obtained by the mitochondrial proteome analyses in heart from type 2 diabetic patients (Heather and Clarke, 2010) and in pancreatic β-cells exposed to high glucose (Ahmed et al,. 2010).

F0F1 ATP Synthase: A Fascinating Challenge for Proteomics 171

specificity lack of the proteolytic cleavage and lability of phosphoester bonds during MS analysis. Different enrichment strategies for phosphorylated peptides or proteins, such as immunoaffinity chromatography (IMAC) or metal oxide affinity chromatography (MOAC), have been established to reduce sample complexity. Concurrently, attention has been paid to the LC-MS instrumentation to avoid loss of phosphorylated peptides within the analytical system. In addition, specific MS techniques have been developed for the identification and relative quantification of phosphorylation sites down to the femtomole range. Nevertheless,

Regarding ATP synthase, in a recent phospho-proteomic study an improved protocol called BEMAD enabled to identify in a cytosolic lysate from mouse brain Ser76 of the F1 subunit α as being phosphorylated (Vosseller et al,. 2005). This method involved differential isotopic labeling of O-phosphate-modified serine/ threonine residues through Michael addition with normal or deuterated dithiothreitol and enrichment of these peptides by thiol chromatography. Specificity of O-phosphate mapping was achieved by blocking of cysteine labeling by prior oxidation and by subsequent enzymatic dephosphorylation of Ophosphate-modified peptides. In a single mass spectrometry analysis along with αSer76 other 20 phosphorylation sites (5 previously reported) were identified and quantified. MS allowed to identify phosphorylated tyrosine and serine residues in the F1 subunits α and β from yeast, which was long considered a "zero background" organism for tyrosine phosphorylation (Krause-Buchholz et al,. 2006). A novel screening technique was applied in combination with Blue Native electrophoresis to separate the ATP synthase complex in native state and second dimension SDS-PAGE to resolve its subunit composition (see below). LA–ICP–MS (Laser ablation inductively coupled plasma mass spectrometry) was used to rapidly screen for the presence of phosphorus in the subunits using sulfur as the internal standard element for quantification. The subunits containing phosphorus were then identified by MALDI–FTICR–MS (matrix-assisted laser desorption/ionization Fourier transform ion cyclotron resonance mass spectrometry) as Tyr434, Ser 413 and Ser426 of the α

At variance from yeast, by combining Blue Native electrophoresis and second/third dimension SDS-PAGEs with LC-ESI/MS analysis we found that in bovine heart mitochondria only the F1 γ subunit contained one phosphorylated tyrosine (Di Pancrazio et al,. 2006). Moreover, the tyrosine residue was phosphorylated only in the monomeric form of ATP synthase and was present in low amount (about 6% of the total protein monomer), while the ATP synthase dimers were lacking. Interestingly, this finding suggested that the oligomerization process might be regulated through cell signaling (Di Pancrazio et al., 2006), but the pathway is still to be clarified. To obtain these results a novel procedure was developed because, due to the low percentage of the phosphorylated species, standard MS/MS analysis failed to detect phosphorylated peptides. The screening of the phosphorylated subunits was done by immunoblotting using anti-phosphotyrosine antibody after the third SDS-PAGE and after trypsin digestion the phosphorylated fragment of γ subunit was identified and quantified by a novel LC-ESI/MS method. This latter was based on the use of two different de-clustering potential values that allowed to obtain, with a single LC-ESI/MS run, the pattern of the phosphorylated and unphosphorylated species. These species were further analyzed by tracing back the origin of the HPO3-deprived forms

Several other studies have shown that phoshorylation can occur on different sites of ATP synthase in mammals (Højlund et al,. 2003; Ko et al,. 2002), yeast and plants (Struglics et

phospho-proteomics still remains far from being routine.

subunit and Tyr7 of the β subunit (Krause-Buchholz et al., 2006).

using tandem MS (Alverdi et al,. 2005).

#### **3. Mass spectrometry data of F0F1 ATP synthase**

The whole F0F1 complex, as well as individual subunits or sub-complexes, have been purified by classical approaches and their amino acid sequences determined almost entirely by direct protein sequence analysis (Walker et al,. 1991). The detailed molecular structures of F0F1 from several species have been studied intensively by different groups (Chen et al,. 2006; Cingolani and Duncan, 2011; Dautant et al,. 2010), and the importance of these studies is highlighted by the award of the Nobel Prize to John Walker in 1997. The structural analyses by X-ray diffraction required the identification of the subunit composition by measurement of accurate molecular masses by mass spectrometry, allowing the detection of posttranslational modifications. The exploration of the exposed regions by limited proteolysis had similar analytical requirements.

Mass spectrometry of F0F1 has been a methodological challenge due to the presence of both hydrophilic parts and hydrophobic subunits, which are difficult to detect by standard ionization techniques, and the fact that all subunits are bound non-covalently. More than 15 years ago, reverse phase liquid chromatography methods were applied to purify the hydrophilic subunits and also some membrane-bound subunits (b, d, F6, e, f, g and A6L) of the bovine heart enzyme, so as to allow their molecular masses to be measured by a mass spectrometer with electrospray ionization (Collinson et al,. 1994). Over the years, direct identification of all the hydrophilic and most of the hydrophobic subunits of ATP synthase from many sources has been obtained by MALDI- and ESI-MS/MS analyses of the tryptic peptides of individual bands or spots on native polyacrylamide or SDS gels. However, the most hydrophobic membrane proteins, such as subunit c and A6L, could not be detected by these approaches (Wittig et al,. 2010). All of the hydrophobic subunits have been identified by tandem mass spectrometry after optimization of their purification in organic solvents and by fragmenting the intact protein ions by collision induced dissociation (CID) with argon (Carroll et al,. 2007). Moreover, a procedure that allows to measure the molecular masses of all of the 17 subunits of F0F1 from bovine heart in a single experiment has been published, this approach is based on the use of a mobile phase during liquid chromatography separation, in which the hydrophilic and hydrophobic components remained soluble, linked directly via an electrospray interface to a triple quadrupole mass spectrometer operated in positive ion mass spectrometry. The method has been used to characterize the ATP synthase subunits from a variety of species and to follow the progress of mild trypsinolysis of the enzyme (Carroll et al., 2009).

#### **3.1 Phosphoproteome of ATP synthase**

The measurement of the mass of a protein allows the presence but not the location of any posttranslational modifications (PTMs) to be detected. Phosphorylation of serine, threonine and tyrosine residues is one of the most prominent PTMs and a key regulator of nearly all biological processes including mitochondrial oxidative phosphorylation (Hüttemann et al,. 2007). However, phosphorylation is often a sub-stoichiometric process and usually only a low percentage of a given protein is present in phosphorylated state at a given time, making its observation challenging. In the past, phosphorylation analysis was mostly done by radiolabeling with 32/33P (Bendt et al,. 2003; MacDonald et al,. 2002) combined with amino acid analysis. In the last decade, considerable effort has been devoted to improving the analysis of phospho-proteome by MS (Eyrich et al,. 2011; Gerber et al,. 2003). The main criticisms are ion suppression of phosphorylated species in a high background of non-phosphorylated ones,

The whole F0F1 complex, as well as individual subunits or sub-complexes, have been purified by classical approaches and their amino acid sequences determined almost entirely by direct protein sequence analysis (Walker et al,. 1991). The detailed molecular structures of F0F1 from several species have been studied intensively by different groups (Chen et al,. 2006; Cingolani and Duncan, 2011; Dautant et al,. 2010), and the importance of these studies is highlighted by the award of the Nobel Prize to John Walker in 1997. The structural analyses by X-ray diffraction required the identification of the subunit composition by measurement of accurate molecular masses by mass spectrometry, allowing the detection of posttranslational modifications. The exploration of the exposed regions by limited

Mass spectrometry of F0F1 has been a methodological challenge due to the presence of both hydrophilic parts and hydrophobic subunits, which are difficult to detect by standard ionization techniques, and the fact that all subunits are bound non-covalently. More than 15 years ago, reverse phase liquid chromatography methods were applied to purify the hydrophilic subunits and also some membrane-bound subunits (b, d, F6, e, f, g and A6L) of the bovine heart enzyme, so as to allow their molecular masses to be measured by a mass spectrometer with electrospray ionization (Collinson et al,. 1994). Over the years, direct identification of all the hydrophilic and most of the hydrophobic subunits of ATP synthase from many sources has been obtained by MALDI- and ESI-MS/MS analyses of the tryptic peptides of individual bands or spots on native polyacrylamide or SDS gels. However, the most hydrophobic membrane proteins, such as subunit c and A6L, could not be detected by these approaches (Wittig et al,. 2010). All of the hydrophobic subunits have been identified by tandem mass spectrometry after optimization of their purification in organic solvents and by fragmenting the intact protein ions by collision induced dissociation (CID) with argon (Carroll et al,. 2007). Moreover, a procedure that allows to measure the molecular masses of all of the 17 subunits of F0F1 from bovine heart in a single experiment has been published, this approach is based on the use of a mobile phase during liquid chromatography separation, in which the hydrophilic and hydrophobic components remained soluble, linked directly via an electrospray interface to a triple quadrupole mass spectrometer operated in positive ion mass spectrometry. The method has been used to characterize the ATP synthase subunits from a variety of species and to follow the progress

The measurement of the mass of a protein allows the presence but not the location of any posttranslational modifications (PTMs) to be detected. Phosphorylation of serine, threonine and tyrosine residues is one of the most prominent PTMs and a key regulator of nearly all biological processes including mitochondrial oxidative phosphorylation (Hüttemann et al,. 2007). However, phosphorylation is often a sub-stoichiometric process and usually only a low percentage of a given protein is present in phosphorylated state at a given time, making its observation challenging. In the past, phosphorylation analysis was mostly done by radiolabeling with 32/33P (Bendt et al,. 2003; MacDonald et al,. 2002) combined with amino acid analysis. In the last decade, considerable effort has been devoted to improving the analysis of phospho-proteome by MS (Eyrich et al,. 2011; Gerber et al,. 2003). The main criticisms are ion suppression of phosphorylated species in a high background of non-phosphorylated ones,

**3. Mass spectrometry data of F0F1 ATP synthase** 

proteolysis had similar analytical requirements.

of mild trypsinolysis of the enzyme (Carroll et al., 2009).

**3.1 Phosphoproteome of ATP synthase** 

specificity lack of the proteolytic cleavage and lability of phosphoester bonds during MS analysis. Different enrichment strategies for phosphorylated peptides or proteins, such as immunoaffinity chromatography (IMAC) or metal oxide affinity chromatography (MOAC), have been established to reduce sample complexity. Concurrently, attention has been paid to the LC-MS instrumentation to avoid loss of phosphorylated peptides within the analytical system. In addition, specific MS techniques have been developed for the identification and relative quantification of phosphorylation sites down to the femtomole range. Nevertheless, phospho-proteomics still remains far from being routine.

Regarding ATP synthase, in a recent phospho-proteomic study an improved protocol called BEMAD enabled to identify in a cytosolic lysate from mouse brain Ser76 of the F1 subunit α as being phosphorylated (Vosseller et al,. 2005). This method involved differential isotopic labeling of O-phosphate-modified serine/ threonine residues through Michael addition with normal or deuterated dithiothreitol and enrichment of these peptides by thiol chromatography. Specificity of O-phosphate mapping was achieved by blocking of cysteine labeling by prior oxidation and by subsequent enzymatic dephosphorylation of Ophosphate-modified peptides. In a single mass spectrometry analysis along with αSer76 other 20 phosphorylation sites (5 previously reported) were identified and quantified.

MS allowed to identify phosphorylated tyrosine and serine residues in the F1 subunits α and β from yeast, which was long considered a "zero background" organism for tyrosine phosphorylation (Krause-Buchholz et al,. 2006). A novel screening technique was applied in combination with Blue Native electrophoresis to separate the ATP synthase complex in native state and second dimension SDS-PAGE to resolve its subunit composition (see below). LA–ICP–MS (Laser ablation inductively coupled plasma mass spectrometry) was used to rapidly screen for the presence of phosphorus in the subunits using sulfur as the internal standard element for quantification. The subunits containing phosphorus were then identified by MALDI–FTICR–MS (matrix-assisted laser desorption/ionization Fourier transform ion cyclotron resonance mass spectrometry) as Tyr434, Ser 413 and Ser426 of the α subunit and Tyr7 of the β subunit (Krause-Buchholz et al., 2006).

At variance from yeast, by combining Blue Native electrophoresis and second/third dimension SDS-PAGEs with LC-ESI/MS analysis we found that in bovine heart mitochondria only the F1 γ subunit contained one phosphorylated tyrosine (Di Pancrazio et al,. 2006). Moreover, the tyrosine residue was phosphorylated only in the monomeric form of ATP synthase and was present in low amount (about 6% of the total protein monomer), while the ATP synthase dimers were lacking. Interestingly, this finding suggested that the oligomerization process might be regulated through cell signaling (Di Pancrazio et al., 2006), but the pathway is still to be clarified. To obtain these results a novel procedure was developed because, due to the low percentage of the phosphorylated species, standard MS/MS analysis failed to detect phosphorylated peptides. The screening of the phosphorylated subunits was done by immunoblotting using anti-phosphotyrosine antibody after the third SDS-PAGE and after trypsin digestion the phosphorylated fragment of γ subunit was identified and quantified by a novel LC-ESI/MS method. This latter was based on the use of two different de-clustering potential values that allowed to obtain, with a single LC-ESI/MS run, the pattern of the phosphorylated and unphosphorylated species. These species were further analyzed by tracing back the origin of the HPO3-deprived forms using tandem MS (Alverdi et al,. 2005).

Several other studies have shown that phoshorylation can occur on different sites of ATP synthase in mammals (Højlund et al,. 2003; Ko et al,. 2002), yeast and plants (Struglics et

F0F1 ATP Synthase: A Fascinating Challenge for Proteomics 173

Native mass spectrometry is a powerful technique to define the stoichiometry of protein complexes, but this can be defined also by quantifying the absolute amounts of the different subunits and therefore by calculating their molar ratio. An interesting approach is the AQUA workflow, which is a variation of isotope dilution MS techniques used for decades for quantification of small molecules and more recently successfully applied in the proteomics context to the absolute quantification of proteins and their modification states in whole cell lysates (Gerber et al., 2003). This approach is based on the addition of synthetic isotopically labeled reference peptides in known amounts to a protein sample in solution prior to tryptic digestion and LC-MS analysis. Being gel-free, this approach avoids errors due to incomplete peptide extraction from the gel or impaired protein digestion within the

Regarding ATP synthase, we recently applied the AQUA workflow to determine the stoichiometry of the F0 subunit e in bovine heart mitochondria (Bisetto et al., 2008). This subunit is involved in dimer/oligomer formation both in yeast (Fronzes et al., 2006) and mammals (Bisetto et al., 2008), but its stoichiometry was still unknown (Arakaki et al,. 2001; Hong and Pedersen, 2003). A critical point of the AQUA approach is the design of the heavy-labeled peptides, which must be unique to the proteins of interest and show high ionization efficiency (proteotypic peptides). In addition, such peptides should have a good fragmentation pattern with reliable matching of b- and γ- ion series, a preferable length between 7 and 15 amino acid residues and contain no chemically unstable residues (M,W, C or N-terminal Q or N) or unstable peptide bonds (e.g. D-P). Moreover, it is necessary to choose reference subunits, whose stoichiometry is already known, to validate the method. We have chosen the F0 subunit b and the F1 subunit γ as reference subunits which are present in single copies in the whole F0F1 complex, as defined by crystal structures. After having chosen among the proteotypic peptides characterized by LC-MS/MS three peptides one from subunits e, b and γ-, the corresponding isotopically labeled analogues were added in known amounts to the detergent extracts of bovine heart mitochondria prior to tryptic digestion. The absolute quantification of the three subunits was then achieved by comparison of the areas under the curve (AUC) of the extracted ion chromatograms of the endogenous and labeled peptides in LC-MS mode. Accuracy of the method was demonstrated by confirming the 1:1 stoichiometry of subunit γ and b and by the low coefficients of variation which were <12% for technical and biological replicates. In the samples analyzed, which contained extracts of mitochondria in resting state, subunit e was present in 1:1 molar ratio with respect to subunit b or γ, demonstrating that in F0F1 it is

Classical 2D IEF-SDS-PAGE do not resolve all the subunits of ATP synthase, because they are quite often small, hydrophobic and basic (pI>9). For this reason, in some studies MS analysis has been run on peptide mixtures obtained after in-solution trypsinization of sample extracts (Bisetto et al., 2008). Alternatively, a powerful approach for efficient separation from tissue homogenates, tissue biopsies and cell cultures of the whole complex of ATP synthase, as well as of assembly intermediates and supra-molecular structures is the native polyacrylamide gel electrophoresis after mild detergent extraction (Wittig et al,. 2006). Following native PAGE, proteins of interest can be extracted in native state and

**3.3 AQUA workflow** 

gel matrix.

contained as unique copy.

**4. Native electrophoresis of F0F1 ATP synthase** 

al,. 1998). Nevertheless, a comprehensive mapping is still lacking. The constant technological progress might soon enable to generate the quantitative and temporal phosphorylation pattern of the enzyme in all organisms and under different pathophysiological conditions, thus allowing understanding of the regulation of ATP synthase in light of cell signaling.

#### **3.2 Native mass spectrometry**

ESI-MS and the novel LILBID-MS (laser induced liquid bead ion desorption mass spectrometry) are two techniques that can be used under native conditions to determine the molecular mass of non-covalently assembled complexes up to the MDa-range with high accuracy. The techniques are complementary, LILBID being more tolerant than nESI to addition of detergents, which are necessary to solubilise membrane proteins such as ATP synthase. Very recently, LILBID-MS combined with Blue Native electrophoresis was successfully applied to compare the subunit composition of the whole F0F1 from *Bacillus pseudofirmus* and from bovine and human heart (Hoffmann et al,. 2010), but also to determine the subunit composition of other even larger membrane complexes such as the NADH-dehydrogenase (complex I of the OXPHOS system) from *Yarrowia lipolytica*  (Sokolova et al,. 2010). LILBID-MS can be applied in several modes, from soft laser desorption (yielding the intact macromolecular complexes) to medium to high laser intensity (which disassembles the protein complexes partially into sub-complexes and these latter into the single subunits). The analysis of the bacterial enzyme revealed at low laser intensity the masses of the F1 sub-complex, the F1 sub-complex lacking the δ-subunit and the F0 sub-complex, while at high laser energy the signal of the 8 subunits appeared. In the case of the mammalian enzymes, the spectra evidenced all of the 15 subunits, of which the masses agreed within ± 150 Da with theoretical masses (see Table 2). While functionally important subunits are conserved, others, such as the so-called minor F0 subunits, show differences in their masses among the species. Determination of their masses from many sources might help to clarify their structural and functional roles, which are still only partially known, and LILBID-MS certainly offers a novel and rapid way to obtain such results using very low material and in detergent solution.

LILBID-MS has also provided an accurate determination of stoichiometry (cn) of the subcomplex formed by the c-ring from the thermoalkaliphilic bacterium *Bacillus sp*. (Meier et al,. 2007). This result is particularly interesting, considering the difficulty to detect c subunit by classical MALDI-MS and LC-MS/MS (Bisetto et al., 2008; Wittig et al., 2010). Moreover, it represents an important MS application in cell bioenergetics, the number of c-subunits being in principle equals to the number of H+ transported across the membrane for every 360° rotation of the rotor in which three ATP molecules are synthesized in the three β subunits. Hence the H+/ATP ratio can be expressed by cn/3.

At variance from LILBID, nESI is ideal for soluble complexes and we recently applied this technique to determine the exact molecular mass of the non-covalent complex formed in solution by the ATP synthase inhibitor IF1 and Calmodulin, the archetypal EF-hand calcium sensor. Interestingly, nESI established a 1:1 stoichiometry between IF1 and Calmodulin, suggesting that binding to Calmodulin promotes the dissociation of the pre-existing dimeric form of IF1 (Cabezon et al,. 2001). Furthermore, native mass analysis was paralleled to the definition of the IF1-CaM complex topology by combining limited proteolysis and crosslinking data with MALDI-MS and LC-MS/MS analyses (Pagnozzi et al,. 2010).

#### **3.3 AQUA workflow**

172 Proteomics – Human Diseases and Protein Functions

al,. 1998). Nevertheless, a comprehensive mapping is still lacking. The constant technological progress might soon enable to generate the quantitative and temporal phosphorylation pattern of the enzyme in all organisms and under different pathophysiological conditions, thus allowing understanding of the regulation of ATP

ESI-MS and the novel LILBID-MS (laser induced liquid bead ion desorption mass spectrometry) are two techniques that can be used under native conditions to determine the molecular mass of non-covalently assembled complexes up to the MDa-range with high accuracy. The techniques are complementary, LILBID being more tolerant than nESI to addition of detergents, which are necessary to solubilise membrane proteins such as ATP synthase. Very recently, LILBID-MS combined with Blue Native electrophoresis was successfully applied to compare the subunit composition of the whole F0F1 from *Bacillus pseudofirmus* and from bovine and human heart (Hoffmann et al,. 2010), but also to determine the subunit composition of other even larger membrane complexes such as the NADH-dehydrogenase (complex I of the OXPHOS system) from *Yarrowia lipolytica*  (Sokolova et al,. 2010). LILBID-MS can be applied in several modes, from soft laser desorption (yielding the intact macromolecular complexes) to medium to high laser intensity (which disassembles the protein complexes partially into sub-complexes and these latter into the single subunits). The analysis of the bacterial enzyme revealed at low laser intensity the masses of the F1 sub-complex, the F1 sub-complex lacking the δ-subunit and the F0 sub-complex, while at high laser energy the signal of the 8 subunits appeared. In the case of the mammalian enzymes, the spectra evidenced all of the 15 subunits, of which the masses agreed within ± 150 Da with theoretical masses (see Table 2). While functionally important subunits are conserved, others, such as the so-called minor F0 subunits, show differences in their masses among the species. Determination of their masses from many sources might help to clarify their structural and functional roles, which are still only partially known, and LILBID-MS certainly offers a novel and rapid way to obtain such

LILBID-MS has also provided an accurate determination of stoichiometry (cn) of the subcomplex formed by the c-ring from the thermoalkaliphilic bacterium *Bacillus sp*. (Meier et al,. 2007). This result is particularly interesting, considering the difficulty to detect c subunit by classical MALDI-MS and LC-MS/MS (Bisetto et al., 2008; Wittig et al., 2010). Moreover, it represents an important MS application in cell bioenergetics, the number of c-subunits being in principle equals to the number of H+ transported across the membrane for every 360° rotation of the rotor in which three ATP molecules are synthesized in the three β subunits.

At variance from LILBID, nESI is ideal for soluble complexes and we recently applied this technique to determine the exact molecular mass of the non-covalent complex formed in solution by the ATP synthase inhibitor IF1 and Calmodulin, the archetypal EF-hand calcium sensor. Interestingly, nESI established a 1:1 stoichiometry between IF1 and Calmodulin, suggesting that binding to Calmodulin promotes the dissociation of the pre-existing dimeric form of IF1 (Cabezon et al,. 2001). Furthermore, native mass analysis was paralleled to the definition of the IF1-CaM complex topology by combining limited proteolysis and cross-

linking data with MALDI-MS and LC-MS/MS analyses (Pagnozzi et al,. 2010).

synthase in light of cell signaling.

**3.2 Native mass spectrometry** 

results using very low material and in detergent solution.

Hence the H+/ATP ratio can be expressed by cn/3.

Native mass spectrometry is a powerful technique to define the stoichiometry of protein complexes, but this can be defined also by quantifying the absolute amounts of the different subunits and therefore by calculating their molar ratio. An interesting approach is the AQUA workflow, which is a variation of isotope dilution MS techniques used for decades for quantification of small molecules and more recently successfully applied in the proteomics context to the absolute quantification of proteins and their modification states in whole cell lysates (Gerber et al., 2003). This approach is based on the addition of synthetic isotopically labeled reference peptides in known amounts to a protein sample in solution prior to tryptic digestion and LC-MS analysis. Being gel-free, this approach avoids errors due to incomplete peptide extraction from the gel or impaired protein digestion within the gel matrix.

Regarding ATP synthase, we recently applied the AQUA workflow to determine the stoichiometry of the F0 subunit e in bovine heart mitochondria (Bisetto et al., 2008). This subunit is involved in dimer/oligomer formation both in yeast (Fronzes et al., 2006) and mammals (Bisetto et al., 2008), but its stoichiometry was still unknown (Arakaki et al,. 2001; Hong and Pedersen, 2003). A critical point of the AQUA approach is the design of the heavy-labeled peptides, which must be unique to the proteins of interest and show high ionization efficiency (proteotypic peptides). In addition, such peptides should have a good fragmentation pattern with reliable matching of b- and γ- ion series, a preferable length between 7 and 15 amino acid residues and contain no chemically unstable residues (M,W, C or N-terminal Q or N) or unstable peptide bonds (e.g. D-P). Moreover, it is necessary to choose reference subunits, whose stoichiometry is already known, to validate the method. We have chosen the F0 subunit b and the F1 subunit γ as reference subunits which are present in single copies in the whole F0F1 complex, as defined by crystal structures. After having chosen among the proteotypic peptides characterized by LC-MS/MS three peptides one from subunits e, b and γ-, the corresponding isotopically labeled analogues were added in known amounts to the detergent extracts of bovine heart mitochondria prior to tryptic digestion. The absolute quantification of the three subunits was then achieved by comparison of the areas under the curve (AUC) of the extracted ion chromatograms of the endogenous and labeled peptides in LC-MS mode. Accuracy of the method was demonstrated by confirming the 1:1 stoichiometry of subunit γ and b and by the low coefficients of variation which were <12% for technical and biological replicates. In the samples analyzed, which contained extracts of mitochondria in resting state, subunit e was present in 1:1 molar ratio with respect to subunit b or γ, demonstrating that in F0F1 it is contained as unique copy.

#### **4. Native electrophoresis of F0F1 ATP synthase**

Classical 2D IEF-SDS-PAGE do not resolve all the subunits of ATP synthase, because they are quite often small, hydrophobic and basic (pI>9). For this reason, in some studies MS analysis has been run on peptide mixtures obtained after in-solution trypsinization of sample extracts (Bisetto et al., 2008). Alternatively, a powerful approach for efficient separation from tissue homogenates, tissue biopsies and cell cultures of the whole complex of ATP synthase, as well as of assembly intermediates and supra-molecular structures is the native polyacrylamide gel electrophoresis after mild detergent extraction (Wittig et al,. 2006). Following native PAGE, proteins of interest can be extracted in native state and

F0F1 ATP Synthase: A Fascinating Challenge for Proteomics 175

**BNE** *hr* **CNE** V2

I

V

III

IV

II

Fig. 6. BNE and *hr* CNE of DDM extracts of bovine heart mitochondria showing OXPHOS

respect to their native masses independently of their p*I.* Negative charge also helps in preventing protein aggregation as negative charges repel each other and this is the reason for its very good resolution. Furthermore, upon binding the previously detergentsolubilised membrane, proteins lose their hydrophobic character and become water soluble hence no further detergent is required in these gels minimising the risk of detergentdependent protein denaturation (Wittig et al., 2006). Besides Coomassie, the presence of Imidazole in anode and cathode buffer (BNE, CNE, *hr* CNE) helps in maintaining pH in the range 7.0-7.5, and incorporation of high concentration of 6-aminohexanoic acid (a zwitterionic substance) improves the solubilisation of membranes (Wittig and Schägger, 2008). Separation of proteins as blue bands helps in gel excision and recovery of blue stained native proteins by electroelution (Wittig et al., 2006) for further MS analysis, as recently

A rapid way to identify ATP synthase in BNE is to monitor ATP hydrolysis by in-gel activity staining, which was developed in our lab and is based on the formation of a white lead phosphate precipitate from phosphate (Pi) released during the reaction (Fig. 7) (Zerbetto et al,. 1997). The catalytic activity can be obtained by incubating the gels in glycine buffer supplemented with Mg-ATP in the presence of 0.2% Pb(NO3)2. The native staining of ATP synthase is reproducible and the white bands on gel can be easily quantified by densitometry (Bisetto et al., 2007). Moreover, the bands can be excised and easily destained in acetic acid solution giving a colourless protein complex ready for MS analysis. This method had been successfully applied by our group to analyse skeletal muscle and heart biopsies from patients

Detection limits of in-gel activity staining are in the microgram range of protein or micromolar phosphate and the resulting white bands are challenging for detection and documentation of low activity due to interference of Coomassie dye. Different strategies of optimization for activity staining of BNE have been applied by us (Bisetto et al., 2007) and

with oxidative phosphorylation enzyme deficiencies (Zerbetto et al., 1997)

complexes.

applied for LILBID MS (Hoffmann et al., 2010).

analyzed by MS or electroblotted for immunodetection or analyzed by in-gel catalytic activity assays. In addition, subunit composition of the complexes can be identified by various denaturating techniques: SDS-PAGE, doubled SDS-PAGEs, as we used to detect phosphotyrosine in monomer/dimers of ATP synthase (Di Pancrazio et al., 2006) and IEF/SDS PAGEs (Wittig et al., 2006), as shown in Fig. 5. These advantages make this approach superior for functional proteomic analyses. For this reason, a brief introduction to native electrophoresis will be presented.

Fig. 5. Applications of native electrophoresis for functional proteomic analyses.

Blue native electrophoresis, abbreviated as BNE, was developed to isolate native membrane proteins and complexes on micro-scale. It separates proteins in the mass range of 10 kDa to 10 MDa. It is a one step technique to isolate proteins from mitochondria, whole cell lysates and tissue homogenates (Wittig et al., 2006). BNE has also been used for the identification of protein-protein interactions, as we recently did to define the new interaction between ATP synthase and Cyclophilin D in mammalian mitochondria (Giorgio et al,. 2009). Besides BNE a similar native electrophoresis method was developed that is called Clear native electrophoresis (CNE) with its variant high resolution CNE (*hr* CNE). Fig. 6 depicts the resolution of the five OXPHOS complexes including ATP synthase with its dimeric form obtained by BNE and *hr* CNE. All these methods are quite similar and have been used for MS analyses, but differ in few aspects, which are discussed below, with their drawbacks and applications.

From a practical stand point BNE, CNE and hr CNE differ in the composition of cathode buffers as well as in the mechanisms by which proteins migrate in the gel. What makes BNE different from CNE and *hr* CNE is the incorporation of Coomassie Brilliant Blue G-250 both in the cathode buffer as well as in loading dye. Coomassie is an anionic dye, it binds to the proteins and imparts negative charge over their surface. In this way proteins migrate with

analyzed by MS or electroblotted for immunodetection or analyzed by in-gel catalytic activity assays. In addition, subunit composition of the complexes can be identified by various denaturating techniques: SDS-PAGE, doubled SDS-PAGEs, as we used to detect phosphotyrosine in monomer/dimers of ATP synthase (Di Pancrazio et al., 2006) and IEF/SDS PAGEs (Wittig et al., 2006), as shown in Fig. 5. These advantages make this approach superior for functional proteomic analyses. For this reason, a brief introduction to

Fig. 5. Applications of native electrophoresis for functional proteomic analyses.

Blue native electrophoresis, abbreviated as BNE, was developed to isolate native membrane proteins and complexes on micro-scale. It separates proteins in the mass range of 10 kDa to 10 MDa. It is a one step technique to isolate proteins from mitochondria, whole cell lysates and tissue homogenates (Wittig et al., 2006). BNE has also been used for the identification of protein-protein interactions, as we recently did to define the new interaction between ATP synthase and Cyclophilin D in mammalian mitochondria (Giorgio et al,. 2009). Besides BNE a similar native electrophoresis method was developed that is called Clear native electrophoresis (CNE) with its variant high resolution CNE (*hr* CNE). Fig. 6 depicts the resolution of the five OXPHOS complexes including ATP synthase with its dimeric form obtained by BNE and *hr* CNE. All these methods are quite similar and have been used for MS analyses, but differ in few aspects, which are discussed below, with their drawbacks and

From a practical stand point BNE, CNE and hr CNE differ in the composition of cathode buffers as well as in the mechanisms by which proteins migrate in the gel. What makes BNE different from CNE and *hr* CNE is the incorporation of Coomassie Brilliant Blue G-250 both in the cathode buffer as well as in loading dye. Coomassie is an anionic dye, it binds to the proteins and imparts negative charge over their surface. In this way proteins migrate with

native electrophoresis will be presented.

applications.

Fig. 6. BNE and *hr* CNE of DDM extracts of bovine heart mitochondria showing OXPHOS complexes.

respect to their native masses independently of their p*I.* Negative charge also helps in preventing protein aggregation as negative charges repel each other and this is the reason for its very good resolution. Furthermore, upon binding the previously detergentsolubilised membrane, proteins lose their hydrophobic character and become water soluble hence no further detergent is required in these gels minimising the risk of detergentdependent protein denaturation (Wittig et al., 2006). Besides Coomassie, the presence of Imidazole in anode and cathode buffer (BNE, CNE, *hr* CNE) helps in maintaining pH in the range 7.0-7.5, and incorporation of high concentration of 6-aminohexanoic acid (a zwitterionic substance) improves the solubilisation of membranes (Wittig and Schägger, 2008). Separation of proteins as blue bands helps in gel excision and recovery of blue stained native proteins by electroelution (Wittig et al., 2006) for further MS analysis, as recently applied for LILBID MS (Hoffmann et al., 2010).

A rapid way to identify ATP synthase in BNE is to monitor ATP hydrolysis by in-gel activity staining, which was developed in our lab and is based on the formation of a white lead phosphate precipitate from phosphate (Pi) released during the reaction (Fig. 7) (Zerbetto et al,. 1997). The catalytic activity can be obtained by incubating the gels in glycine buffer supplemented with Mg-ATP in the presence of 0.2% Pb(NO3)2. The native staining of ATP synthase is reproducible and the white bands on gel can be easily quantified by densitometry (Bisetto et al., 2007). Moreover, the bands can be excised and easily destained in acetic acid solution giving a colourless protein complex ready for MS analysis. This method had been successfully applied by our group to analyse skeletal muscle and heart biopsies from patients with oxidative phosphorylation enzyme deficiencies (Zerbetto et al., 1997)

Detection limits of in-gel activity staining are in the microgram range of protein or micromolar phosphate and the resulting white bands are challenging for detection and documentation of low activity due to interference of Coomassie dye. Different strategies of optimization for activity staining of BNE have been applied by us (Bisetto et al., 2007) and

F0F1 ATP Synthase: A Fascinating Challenge for Proteomics 177

stalk combine with the F1 sector and the c-ring independently from the mitochondrial encoded subunits a and A6L which bind at the late stage. Conversely it has been proposed that in yeast the subunits Su6/Su8, homologous to subunits a and A6L, first combine to the subunits of the lateral stalk, forming the sub-complex Su6/Su8/stator, which finally binds to the sub-complex F1/c-ring, as already shown in Fig. 3. Moreover, the finding that ρ<sup>0</sup> mitochondria still contain dimers/oligomers of ATP synthase although in lower amounts than control mitochondria, supported the idea that subunits a/A6L contribute, together with subunits e/g (Bisetto et al., 2008), to stabilize the supra-molecular structures, but they are not the most important interface as previously proposed in yeast (Wittig et al., 2008). Recently Yan et al., using rat brain mitochondria, demonstrated that a non-gradient highly porous BNE of 8% polyacrilamide is an efficient technique to resolve all OXPHOS complex along with other mitochondrial proteins, such as DLDH and Hsp60 polymer. Further, the gel strips can be even used to perform 2D BN-/SDS PAGE or the bands can be excised for

The choice of detergent for protein extraction is an integral part of a successful native electrophoresis. The principal effect of detergents during solubilisation is the breaking of lipid-lipid and lipid-protein interactions present on biomembranes. Competing with lipids for the occupation of the surface of integral hydrophobic proteins, they form mixed micelles containing detergent, lipids and proteins. The solubilisation effect is maximum when the detergent is used at a concentration equal or higher to the Critical Micelle Concentration (CMC), the concentration at which the detergent molecules form micelles. These are detergent self associating structures with hydrophobic ends facing inside and hydrophilic groups facing outside the aqueous phase. CMC is a characteristic of each detergent and

The anionic SDS is, in principle, not suitable for native electrophoresis as besides solubilising membranes, it dissociates and denatures the enzyme complex leading to loss of activity. Non-ionic detergents are uncharged and milder and hence can be used in membrane solubilisation to isolate mitochondrial complexes with varying degrees of association dependent on the kind of detergent used and its concentration. The most commonly used non-ionic detergents are Triton X-100, Digitonin and Dodecyl-β-Dmaltoside (DDM) which form micelles at very low concentration, thus avoiding protein denaturation (Wittig and Schägger, 2008). Regarding isolation of OXPHOS complexes, these non-ionic detergents behave very differently from each other. For example, Digitonin can be used in a very broad concentration range (from 0.5 to 8 g/g proteins), as compared to Triton and DDM which work in the range of 1-2 g/g proteins (Schägger and Pfeiffer, 2000). Besides this, Digitonin is the best candidate for the isolation of supercomplexes due to its milder nature (Reisinger and Eichacker, 2008). They can be used to extract OXPHOS complexes from bacteria, yeast and mammals, as well as from subcellular fractions or total membranes (Wittig et al., 2006). The quantity of detergent required to solubilise membrane proteins vary in the different cells/tissues and optimal solubilisation conditions for each membrane and

Recently, an interesting modification of classical native PAGE has been proposed. Klodmann et al. reported the treatment of the samples with low amounts of SDS before BNE. This allowed to destabilize the OXPHOS complexes in sub-complexes in a very defined and reproducible manner and to study their internal architectures. SDS was added

depends on pH, temperature and ionic strength (Reisinger and Eichacker, 2008).

MS peptide sequencing (Yan and Forster, 2009).

**4.1 Types of detergents and their use in native electrophoresis** 

each membrane complex should be experimentally investigated.

others (Suhai et al,. 2009; Wittig and Schägger, 2005). Alternatively, CNE can be used that was introduced to circumvent this disadvantage of BNE. In fact, it uses the same buffers and conditions for electrophoresis but the difference lies on the absence of Coomassie dye both in cathode buffer as well as in sample buffer. In this way, much higher activity staining has been obtained (Wittig et al,. 2007). However, due to the absence of Coomassie there is no negative charge shift and the movement of proteins totally depends on their intrinsic mass and p*I.* Therefore CNE suffers from a poor resolution and is limited to proteins having a p*I*<7 unlike BNE where even proteins having pI>10.5 migrate towards anode (Wittig and Schägger, 2008). Anyway, CNE offers advantage over BNE for isolation of supra-molecular structures, including ATP synthase dimers and oligomers, being the mildest technique to separate mitochondrial membrane proteins.

Fig. 7. In-gel ATPase activity staining of the different oligomeric forms of ATP synthase extracted by Digitonin (0.5 g/g protein) from mouse heart mitochondria and analyzed by BNE.

To preserve the advantages of both the techniques Wittig et al introduced *hr* CNE. In this technique cathode buffer is supplemented with a combination of colourless anionic and neutral detergents such as Triton X-100, Deoxycholate, or Dodecyl-β-D-maltoside (DDM). This leads to a charge shift over the surface of the proteins and helps them in migration with a resolution comparable with BNE. Also there is no interference in in-gel assays due to absence of Coomassie.

An interesting recent application of *hr* CNE is related to the molecular characterization of assembly intermediates of ATP synthase in mammals. In fact, using particularly mild detergent conditions and *hr* CNE, Wittig et al. were able to separate the assembly intermediate in human ρ0 cells, which lack the mitochondrial DNA encoding subunits a and A6L (Wittig et al., 2010). By analyzing its subunit composition by ESI-MS/MS after excising the native band from 1D *hr* CN-PAGE or by MALDI-MS (MS/MS) after 2D *hr* CN-/SDS-PAGE they established that this intermediate contains all the nuclear-encoded subunits. These results allowed to propose that in mammals, differently from yeast (Hoffmann et al., 2010), the assembly of the whole enzyme is a linear process and the subunits of the lateral

others (Suhai et al,. 2009; Wittig and Schägger, 2005). Alternatively, CNE can be used that was introduced to circumvent this disadvantage of BNE. In fact, it uses the same buffers and conditions for electrophoresis but the difference lies on the absence of Coomassie dye both in cathode buffer as well as in sample buffer. In this way, much higher activity staining has been obtained (Wittig et al,. 2007). However, due to the absence of Coomassie there is no negative charge shift and the movement of proteins totally depends on their intrinsic mass and p*I.* Therefore CNE suffers from a poor resolution and is limited to proteins having a p*I*<7 unlike BNE where even proteins having pI>10.5 migrate towards anode (Wittig and Schägger, 2008). Anyway, CNE offers advantage over BNE for isolation of supra-molecular structures, including ATP synthase dimers and oligomers, being the mildest technique to

Fig. 7. In-gel ATPase activity staining of the different oligomeric forms of ATP synthase extracted by Digitonin (0.5 g/g protein) from mouse heart mitochondria and analyzed by

To preserve the advantages of both the techniques Wittig et al introduced *hr* CNE. In this technique cathode buffer is supplemented with a combination of colourless anionic and neutral detergents such as Triton X-100, Deoxycholate, or Dodecyl-β-D-maltoside (DDM). This leads to a charge shift over the surface of the proteins and helps them in migration with a resolution comparable with BNE. Also there is no interference in in-gel assays due to

An interesting recent application of *hr* CNE is related to the molecular characterization of assembly intermediates of ATP synthase in mammals. In fact, using particularly mild detergent conditions and *hr* CNE, Wittig et al. were able to separate the assembly intermediate in human ρ0 cells, which lack the mitochondrial DNA encoding subunits a and A6L (Wittig et al., 2010). By analyzing its subunit composition by ESI-MS/MS after excising the native band from 1D *hr* CN-PAGE or by MALDI-MS (MS/MS) after 2D *hr* CN-/SDS-PAGE they established that this intermediate contains all the nuclear-encoded subunits. These results allowed to propose that in mammals, differently from yeast (Hoffmann et al., 2010), the assembly of the whole enzyme is a linear process and the subunits of the lateral

separate mitochondrial membrane proteins.

BNE.

absence of Coomassie.

stalk combine with the F1 sector and the c-ring independently from the mitochondrial encoded subunits a and A6L which bind at the late stage. Conversely it has been proposed that in yeast the subunits Su6/Su8, homologous to subunits a and A6L, first combine to the subunits of the lateral stalk, forming the sub-complex Su6/Su8/stator, which finally binds to the sub-complex F1/c-ring, as already shown in Fig. 3. Moreover, the finding that ρ<sup>0</sup> mitochondria still contain dimers/oligomers of ATP synthase although in lower amounts than control mitochondria, supported the idea that subunits a/A6L contribute, together with subunits e/g (Bisetto et al., 2008), to stabilize the supra-molecular structures, but they are not the most important interface as previously proposed in yeast (Wittig et al., 2008). Recently Yan et al., using rat brain mitochondria, demonstrated that a non-gradient highly porous BNE of 8% polyacrilamide is an efficient technique to resolve all OXPHOS complex along with other mitochondrial proteins, such as DLDH and Hsp60 polymer. Further, the gel strips can be even used to perform 2D BN-/SDS PAGE or the bands can be excised for MS peptide sequencing (Yan and Forster, 2009).

#### **4.1 Types of detergents and their use in native electrophoresis**

The choice of detergent for protein extraction is an integral part of a successful native electrophoresis. The principal effect of detergents during solubilisation is the breaking of lipid-lipid and lipid-protein interactions present on biomembranes. Competing with lipids for the occupation of the surface of integral hydrophobic proteins, they form mixed micelles containing detergent, lipids and proteins. The solubilisation effect is maximum when the detergent is used at a concentration equal or higher to the Critical Micelle Concentration (CMC), the concentration at which the detergent molecules form micelles. These are detergent self associating structures with hydrophobic ends facing inside and hydrophilic groups facing outside the aqueous phase. CMC is a characteristic of each detergent and depends on pH, temperature and ionic strength (Reisinger and Eichacker, 2008).

The anionic SDS is, in principle, not suitable for native electrophoresis as besides solubilising membranes, it dissociates and denatures the enzyme complex leading to loss of activity. Non-ionic detergents are uncharged and milder and hence can be used in membrane solubilisation to isolate mitochondrial complexes with varying degrees of association dependent on the kind of detergent used and its concentration. The most commonly used non-ionic detergents are Triton X-100, Digitonin and Dodecyl-β-Dmaltoside (DDM) which form micelles at very low concentration, thus avoiding protein denaturation (Wittig and Schägger, 2008). Regarding isolation of OXPHOS complexes, these non-ionic detergents behave very differently from each other. For example, Digitonin can be used in a very broad concentration range (from 0.5 to 8 g/g proteins), as compared to Triton and DDM which work in the range of 1-2 g/g proteins (Schägger and Pfeiffer, 2000). Besides this, Digitonin is the best candidate for the isolation of supercomplexes due to its milder nature (Reisinger and Eichacker, 2008). They can be used to extract OXPHOS complexes from bacteria, yeast and mammals, as well as from subcellular fractions or total membranes (Wittig et al., 2006). The quantity of detergent required to solubilise membrane proteins vary in the different cells/tissues and optimal solubilisation conditions for each membrane and each membrane complex should be experimentally investigated.

Recently, an interesting modification of classical native PAGE has been proposed. Klodmann et al. reported the treatment of the samples with low amounts of SDS before BNE. This allowed to destabilize the OXPHOS complexes in sub-complexes in a very defined and reproducible manner and to study their internal architectures. SDS was added

F0F1 ATP Synthase: A Fascinating Challenge for Proteomics 179

synthase. We recently separated by detergent extraction and BNE the whole ecto- F0F1 from plasma membranes of rat liver (Giorgio et al,. 2010) and we found that the low, but constant amounts of F0F1 complexes display a similar molecular weight to the monomeric form of the mitochondrial F0F1 ATP synthase, as evidenced by in-gel ATPase activity staining and immunoblotting. This suggests that the plasma membranes of normal liver do contain complete, functional F0F1 ATP synthase complexes, which display very similar subunit composition and assembly of the mitochondrial enzymes for which MS analysis is in

All together these studies support the view that ATP synthase is mainly located in lipid rafts of plasma membranes, is enzymatically active and functions as a cell-surface receptor involved in different biological effects depending on the cell types (Fig. 8). In hepatocytes ecto-F0F1 functions as high affinity ApoA1 receptor and regulates HDL metabolism, in endothelial cells it functions as angiostatin receptor thus mediating angiogenesis, in tumour cells it functions as a pH regulator and participates in tumor recognition by cytotoxic Vγ9/Vδ2 T lymphocytes (Vantourout et al., 2010). Various independent studies have reported that ecto-F0F1 can synthesize ATP from ADP and Pi extruding protons from cytoplasm in different cell types, such as endothelial cells or hepatocytes. The resulting ATP can triggers cation influx into the cells through ATP-gated ion channels (P2X purinoreceptors) or can bind to G-protein coupled receptors (P2Y purinoreceptors) activating a downstream signaling pathways (Mowery and Pizzo, 2010). However, whether or not ecto-F0F1 can synthesize ATP is still debated, as there are several conflicting reports. Conversely, it is widely accepted that ecto-F0F1 catalyses the hydrolysis of ATP, potentially affecting purinergic signaling (Fig. 8). As an example, in hepatocytes, ADP generated by

Fig. 8. The model depicts the orientation of ectopic F0F1 ATP synthase in eukaryotes with the F1 sector facing outside. In some cells the ATP synthesized by the enzyme leads to influx of cations such as Ca2+ through P2X receptors and in others the ADP generated by ATP

hydrolysis triggers signaling pathways through activation of G-protein coupled P2Y receptors.

progress in our laboratory.

to the mitochondrial Digitonin extracts from *Arabidopsis* in the range of 0.05-1.0% just before BNE loading. By combining with 2D BN-/ SDS or 2D BN/BN PAGE, the authors clearly demonstrated the variable effect of SDS on the OXPHOS complexes. As an example, at 0.2% SDS concentration, the ATP synthase complex dissociated in to F0 and F1. At 0.3% F0 even dissociated in to a sub-complex composed of c-ring (Klodmann et al,. 2011).

#### **5. Ectopic F0F1ATP synthase of mammalian cells**

The application of proteomic analyses to sub-cellular mammalian fractions other than mitochondria revealed the presence of mitochondrial membrane components in unexpected cellular locations, such as plasma membranes or nuclei. Some of these studies ascribed it to cross-contamination, due to the contiguity of the different membranes within the cell and to the high sensitivity of MS which identifies the proteins up to subfentomolar levels, but in others cases the parallel demonstration of such unusual locations obtained in cells and tissues by immunofluorescence and functional studies led the scientists to consider the proteomics results a mainstay to discover new scenarios in the intracellular traffic connections.

Regarding ATP synthase, many proteomics studies have recently reported that subunits of this complex, along with other OXPHOS complexes, are expressed in extra-mitochondrial membranes of different mammalian cell types – especially on the cell surface, but also in the endoplasmic reticulum and nuclear envelope (Panfoli et al,. 2011). In human apoptotic Tleukemia cells the presence of eight ATP synthase subunits was revealed in the nucleus fraction obtained by differential extraction and stable isotope labeling of cell culture followed by LC-MS/MS analysis and it has been ascribed to a dynamic recruitment of mitochondria into nuclear invaginations during apoptosis (Hwang et al,. 2006). In mouse brain a proteomics analysis of microsomal fraction obtained both by 2D-LC-MS/MS and shotgun LC-MS/MS found many subunits of the OXPHOS complexes, including ATP synthase, which were proposed to represent mitochondrial proteins with high turnover rates in the cell (Stevens Jr et al,. 2008). Conversely, ATP synthase is now considered a true resident on the plasma membranes. In fact, the identification by MS of ATP synthase subunits in plasma membrane preparations from different sources has been paralleled by the demonstration of the enzyme expression with the F1 sector facing outside (and for this reason the enzyme is named ecto-F0F1) obtained by cytometry, confocal microscopy and functional studies (Vantourout et al,. 2010). ATP synthase subunits were identified in plasma membranes isolated from cell culture, i.e. in hypoxia-adapted tumor cells where differential 16O/18O stable isotopic labeling and multidimensional LC-MS/MS revealed an increased expression of ATP synthase α subunit with respect to normoxia (Stockwin et al,. 2006), and in tissues that were characterized by high purity, i.e. obtained by combining subcellular fractionation with immunoisolation strategies so that no proteins from endoplasmic reticulum and nuclear envelope were detected (Zhang et al,. 2007). In other studies, the presence of ATP synthase subunits was found in the detergent-resistant fragments of plasma membranes, i.e. in the lipid rafts which are cholesterol and sphingolipid-rich microdomains involved in signal trasduction. In particular, in lipid rafts isolated from rat liver and subjected to an efficient *in solution* digestion followed by cRPLC/MS/MS four subunits both of F1 and F0 sector were identified (Bae et al,. 2004).

The major limit of the proteomic studies related to plasma membranes is that in no one the complete subunit composition of ecto-F0F1 has been recognized, leaving open the possibility of a different assembly of the ectopic enzyme with respect to the mitochondrial ATP

to the mitochondrial Digitonin extracts from *Arabidopsis* in the range of 0.05-1.0% just before BNE loading. By combining with 2D BN-/ SDS or 2D BN/BN PAGE, the authors clearly demonstrated the variable effect of SDS on the OXPHOS complexes. As an example, at 0.2% SDS concentration, the ATP synthase complex dissociated in to F0 and F1. At 0.3% F0 even

The application of proteomic analyses to sub-cellular mammalian fractions other than mitochondria revealed the presence of mitochondrial membrane components in unexpected cellular locations, such as plasma membranes or nuclei. Some of these studies ascribed it to cross-contamination, due to the contiguity of the different membranes within the cell and to the high sensitivity of MS which identifies the proteins up to subfentomolar levels, but in others cases the parallel demonstration of such unusual locations obtained in cells and tissues by immunofluorescence and functional studies led the scientists to consider the proteomics

Regarding ATP synthase, many proteomics studies have recently reported that subunits of this complex, along with other OXPHOS complexes, are expressed in extra-mitochondrial membranes of different mammalian cell types – especially on the cell surface, but also in the endoplasmic reticulum and nuclear envelope (Panfoli et al,. 2011). In human apoptotic Tleukemia cells the presence of eight ATP synthase subunits was revealed in the nucleus fraction obtained by differential extraction and stable isotope labeling of cell culture followed by LC-MS/MS analysis and it has been ascribed to a dynamic recruitment of mitochondria into nuclear invaginations during apoptosis (Hwang et al,. 2006). In mouse brain a proteomics analysis of microsomal fraction obtained both by 2D-LC-MS/MS and shotgun LC-MS/MS found many subunits of the OXPHOS complexes, including ATP synthase, which were proposed to represent mitochondrial proteins with high turnover rates in the cell (Stevens Jr et al,. 2008). Conversely, ATP synthase is now considered a true resident on the plasma membranes. In fact, the identification by MS of ATP synthase subunits in plasma membrane preparations from different sources has been paralleled by the demonstration of the enzyme expression with the F1 sector facing outside (and for this reason the enzyme is named ecto-F0F1) obtained by cytometry, confocal microscopy and functional studies (Vantourout et al,. 2010). ATP synthase subunits were identified in plasma membranes isolated from cell culture, i.e. in hypoxia-adapted tumor cells where differential 16O/18O stable isotopic labeling and multidimensional LC-MS/MS revealed an increased expression of ATP synthase α subunit with respect to normoxia (Stockwin et al,. 2006), and in tissues that were characterized by high purity, i.e. obtained by combining subcellular fractionation with immunoisolation strategies so that no proteins from endoplasmic reticulum and nuclear envelope were detected (Zhang et al,. 2007). In other studies, the presence of ATP synthase subunits was found in the detergent-resistant fragments of plasma membranes, i.e. in the lipid rafts which are cholesterol and sphingolipid-rich microdomains involved in signal trasduction. In particular, in lipid rafts isolated from rat liver and subjected to an efficient *in solution* digestion followed by cRPLC/MS/MS four subunits both of F1 and F0 sector were identified (Bae et al,. 2004). The major limit of the proteomic studies related to plasma membranes is that in no one the complete subunit composition of ecto-F0F1 has been recognized, leaving open the possibility of a different assembly of the ectopic enzyme with respect to the mitochondrial ATP

dissociated in to a sub-complex composed of c-ring (Klodmann et al,. 2011).

results a mainstay to discover new scenarios in the intracellular traffic connections.

**5. Ectopic F0F1ATP synthase of mammalian cells** 

synthase. We recently separated by detergent extraction and BNE the whole ecto- F0F1 from plasma membranes of rat liver (Giorgio et al,. 2010) and we found that the low, but constant amounts of F0F1 complexes display a similar molecular weight to the monomeric form of the mitochondrial F0F1 ATP synthase, as evidenced by in-gel ATPase activity staining and immunoblotting. This suggests that the plasma membranes of normal liver do contain complete, functional F0F1 ATP synthase complexes, which display very similar subunit composition and assembly of the mitochondrial enzymes for which MS analysis is in progress in our laboratory.

All together these studies support the view that ATP synthase is mainly located in lipid rafts of plasma membranes, is enzymatically active and functions as a cell-surface receptor involved in different biological effects depending on the cell types (Fig. 8). In hepatocytes ecto-F0F1 functions as high affinity ApoA1 receptor and regulates HDL metabolism, in endothelial cells it functions as angiostatin receptor thus mediating angiogenesis, in tumour cells it functions as a pH regulator and participates in tumor recognition by cytotoxic Vγ9/Vδ2 T lymphocytes (Vantourout et al., 2010). Various independent studies have reported that ecto-F0F1 can synthesize ATP from ADP and Pi extruding protons from cytoplasm in different cell types, such as endothelial cells or hepatocytes. The resulting ATP can triggers cation influx into the cells through ATP-gated ion channels (P2X purinoreceptors) or can bind to G-protein coupled receptors (P2Y purinoreceptors) activating a downstream signaling pathways (Mowery and Pizzo, 2010). However, whether or not ecto-F0F1 can synthesize ATP is still debated, as there are several conflicting reports. Conversely, it is widely accepted that ecto-F0F1 catalyses the hydrolysis of ATP, potentially affecting purinergic signaling (Fig. 8). As an example, in hepatocytes, ADP generated by

Fig. 8. The model depicts the orientation of ectopic F0F1 ATP synthase in eukaryotes with the F1 sector facing outside. In some cells the ATP synthesized by the enzyme leads to influx of cations such as Ca2+ through P2X receptors and in others the ADP generated by ATP hydrolysis triggers signaling pathways through activation of G-protein coupled P2Y receptors.

F0F1 ATP Synthase: A Fascinating Challenge for Proteomics 181

Ahmed, M.; Muhammed, S. J.; Kessler, B. & Salehi, A. (2010). Mitochondrial proteome

Andersson, U.; Houstek, J. & Cannon, B. (1997). ATP synthase subunit c expression:

Andrianaivomananjaona, T.; Moune-Dimala, M.; Herga, S.; David, V. & Haraux, F. (2010).

Arakaki, N.; Ueyama, Y.; Hirose, M.; Himeda, T.; Shibata, H.; Futaki, S.; Kitagawa, K. &

Bae, T. J.; Kim, M. S.; Kim, J. W.; Kim, B. W.; Choo, H. J.; Lee, J. W.; Kim, K. B.; Lee, C. S.;

Bason, J. V.; Runswick, M. J.; Fearnley, I. M. & Walker, J. E. (2011). Binding of the Inhibitor

Bendt, A. K.; Burkovski, A.; Schaffer, S.; Bott, M.; Farwick, M. & Hermann, T. (2003).

Bisetto, E.; Giorgio, V.; Di Pancrazio, F.; Mavelli, I. & Lippe, G. (2007). Characterization of

synthase. *Journal of bioenergetics and biomembranes*, Vol.40, No.4, pp. 257-267 Borghese, R.; Turina, P.; Lambertini, L. & Melandri, B. A. (1998). The atpIBEXF operon

Boyer, P. D. (1997). The ATP synthase-a splendid molecular machine. *Annual review of* 

Bustos, D. M. & Velours, J. (2005). The modification of the conserved GXXXG motif of the

detergents. *The Italian journal of biochemistry*, Vol.56, No.4, pp. 254-258 Bisetto, E.; Picotti, P.; Giorgio, V.; Alverdi, V.; Mavelli, I. & Lippe, G. (2008). Functional and

in the cell surface. *Proteomics*, Vol.4, No.11, pp. 3536-3548

pancreatic -cells exposed to high glucose. *Islets*, Vol.2, No.5, pp. 283-292 Alverdi, V.; Di Pancrazio, F.; Lippe, G.; Pucillo, C.; Casetta, B. & Esposito, G. (2005).

No.22, pp. 3343-3348

Vol.1807, No.2, pp. 197-204

Vol.3, No.8, pp. 1637-1646

No.5, pp. 385-388

29010

*biochemistry*, Vol.66, No.1, pp. 717-749

2, pp. 379-385

228

453

analysis reveals altered expression of voltage dependent anion channels in

Determination of protein phosphorylation sites by mass spectrometry: a novel electrospray based method. *Rapid communications in mass spectrometry*, Vol.19,

physiological regulation of the P1 and P2 genes. *Biochemical Journal*, Vol.323, No.Pt

How the N-terminal extremity of Saccharomyces cerevisiae IF1 interacts with ATP synthase: A kinetic approach. *Biochimica et Biophysica Acta (BBA)-Bioenergetics*,

Higuti, T. (2001). Stoichiometry of subunit e in rat liver mitochondrial H+-ATP synthase and membrane topology of its putative Ca2+-dependent regulatory region. *Biochimica et Biophysica Acta (BBA)-Bioenergetics*, Vol.1504, No.2-3, pp. 220-

Kim, J. H. & Chang, S. Y. (2004). Lipid raft proteome reveals ATP synthase complex

Protein IF1 to Bovine F1-ATPase. *Journal of Molecular Biology*, Vol.406, No.3, pp. 443-

Towards a phosphoproteome map of Corynebacterium glutamicum. *Proteomics*,

oligomeric forms from mammalian F0F1ATP synthase by BN-PAGE: the role of

stoichiometric analysis of subunit e in bovine heart mitochondrial F 0 F 1 ATP

coding for the F0 sector of the ATP synthase from the purple nonsulfur photosynthetic bacterium Rhodobacter capsulatus. *Archives of microbiology*, Vol.170,

membrane-spanning segment of subunit g destabilizes the supramolecular species of yeast ATP synthase. *Journal of Biological Chemistry*, Vol.280, No.32, pp. 29004-

ecto-F0F1 upon binding of ApoA1 activates the P2Y receptors resulting in HDL endocytosis and downstream the small GTPase RhoA and its effector ROCK I (Malaval et al,. 2009). The mechanism used by ATP synthase to reach the plasma membrane is still unknown (Vantourout et al., 2010). The hypothesis that the single ATP synthase subunits are routed to plasma membrane instead of the mitochondria seems unlikely, because different mRNA isoforms of ATP synthase subunits have not been found in mammals apart for the bovine subunit c (Vives-Bauza et al., 2011). The simpler explanation seems that once assembled into mitochondria, the whole complex reaches the cell surface *via* vescicular transport or fusion of mitochondrial membranes with plasma membranes (Vantourout et al., 2010). It is tempting to hypothesize that the new technology of imaging MS (MALDI MS profiling/imaging), which can acquire individual spectra from the surface of frozen tissue sections (Chaurand et al,. 2006), could give important answers regarding the trafficking of the enzyme to cell surface.

#### **6. Conclusions**

F0F1 ATP synthase is an intensely studied enzyme complex, for which single molecule studies have allowed to define the fascinating catalysis in great detail. In addition, highresolution molecular structures have been obtained mainly by X-ray crystal analyses. In spite of this tremendous progress, many aspects of ATP synthase physiology, such as biogenesis or super-complex formation, and its role in pathology are still unknown. The omni-comprehensive nature of proteomics, unlike the more reductionistic approaches of classical biochemistry and genetics, makes it the best candidate for revealing changes in the expression level of the whole complex and/or of its single subunits, but also to define the quantitative and temporal phosphorylation pattern of the enzyme in all organisms and under different physiopathological conditions, thus allowing the understanding of the ATP synthase regulation in a better way. In addition, the constant technological progress will enable to define the intriguing enzyme intracellular trafficking and its translocation to cell surface. In this context, native electrophoresis combined with MS techniques offers a powerful top-down approach for functional and structural analyses of such complicated enzyme using minimal amount of cell lysates or tissue homogenates and making this approach useful also for clinical investigation.

#### **7. Acknowledgements**

Authors acknowledge Prof. Paolo Bernardi, Dept. of Biomedical Sciences, University of Padova and Dr. Paola Picotti, Institute of Biochemistry, ETH, Zürich for critical reading of the manuscript.

#### **8. References**


ecto-F0F1 upon binding of ApoA1 activates the P2Y receptors resulting in HDL endocytosis and downstream the small GTPase RhoA and its effector ROCK I (Malaval et al,. 2009). The mechanism used by ATP synthase to reach the plasma membrane is still unknown (Vantourout et al., 2010). The hypothesis that the single ATP synthase subunits are routed to plasma membrane instead of the mitochondria seems unlikely, because different mRNA isoforms of ATP synthase subunits have not been found in mammals apart for the bovine subunit c (Vives-Bauza et al., 2011). The simpler explanation seems that once assembled into mitochondria, the whole complex reaches the cell surface *via* vescicular transport or fusion of mitochondrial membranes with plasma membranes (Vantourout et al., 2010). It is tempting to hypothesize that the new technology of imaging MS (MALDI MS profiling/imaging), which can acquire individual spectra from the surface of frozen tissue sections (Chaurand et al,. 2006), could give important answers regarding the trafficking of

F0F1 ATP synthase is an intensely studied enzyme complex, for which single molecule studies have allowed to define the fascinating catalysis in great detail. In addition, highresolution molecular structures have been obtained mainly by X-ray crystal analyses. In spite of this tremendous progress, many aspects of ATP synthase physiology, such as biogenesis or super-complex formation, and its role in pathology are still unknown. The omni-comprehensive nature of proteomics, unlike the more reductionistic approaches of classical biochemistry and genetics, makes it the best candidate for revealing changes in the expression level of the whole complex and/or of its single subunits, but also to define the quantitative and temporal phosphorylation pattern of the enzyme in all organisms and under different physiopathological conditions, thus allowing the understanding of the ATP synthase regulation in a better way. In addition, the constant technological progress will enable to define the intriguing enzyme intracellular trafficking and its translocation to cell surface. In this context, native electrophoresis combined with MS techniques offers a powerful top-down approach for functional and structural analyses of such complicated enzyme using minimal amount of cell lysates or tissue homogenates and making this

Authors acknowledge Prof. Paolo Bernardi, Dept. of Biomedical Sciences, University of Padova and Dr. Paola Picotti, Institute of Biochemistry, ETH, Zürich for critical reading of

Ackerman, S. H. (2002). Atp11p and Atp12p are chaperones for F1-ATPase biogenesis in

Adachi, K.; Oiwa, K.; Nishizaka, T.; Furuike, S.; Noji, H.; Itoh, H.; Yoshida, M. & Kinosita Jr,

molecule imaging and manipulation. *Cell*, Vol.130, No.2, pp. 309-321

mitochondria. *Biochimica et Biophysica Acta (BBA)-Bioenergetics*, Vol.1555, No.1-3, pp.

K. (2007). Coupling of rotation and catalysis in F1-ATPase revealed by single-

the enzyme to cell surface.

approach useful also for clinical investigation.

**7. Acknowledgements** 

101-105

the manuscript.

**8. References** 

**6. Conclusions** 


F0F1 ATP Synthase: A Fascinating Challenge for Proteomics 183

Dautant, A.; Velours, J. & Giraud, M. F. (2010). Crystal Structure of the Mg· ADP-inhibited

Di Pancrazio, F.; Bisetto, E.; Alverdi, V.; Mavelli, I.; Esposito, G. & Lippe, G. (2006).

Di Pancrazio, F.; Mavelli, I.; Isola, M.; Losano, G.; Pagliaro, P.; Harris, D. A. & Lippe, G.

Dienhart, M.; Pfeiffer, K.; Schägger, H. & Stuart, R. A. (2002). Formation of the yeast F1F0-

Everard-Gigot, V.; Dunn, C. D.; Dolan, B. M.; Brunner, S.; Jensen, R. E. & Stuart, R. A. (2005).

Eyrich, B.; Sickmann, A. & Zahedi, R. P. (2011). Catch me if you can: Mass spectrometry

Fronzes, R.; Weimann, T.; Vaillier, J.; Velours, J. & Brèthes, D. (2006). The peripheral stalk

Gerber, S. A.; Rush, J.; Stemman, O.; Kirschner, M. W. & Gygi, S. P. (2003). Absolute

Giorgio, V.; Bisetto, E.; Franca, R.; Harris, D. A.; Passamonti, S. & Lippe, G. (2010). The

Gomis-Rüth, F. X.; Moncalián, G.; Pérez-Luque, R.; González, A.; Cabezón, E.; de la Cruz, F.

Højlund, K.; Wrzesinski, K.; Larsen, P. M.; Fey, S. J.; Roepstorff, P.; Handberg, A.; Dela, F.;

Harris, D. A. & Das, A. M. (1991). Control of mitochondrial ATP synthesis in the heart.

inhibitor protein IF 1. *Journal of bioenergetics and biomembranes*, pp. 1-7 Giorgio, V.; Bisetto, E.; Soriano, M. E.; Dabbeni-Sala, F.; Basso, E.; Petronilli, V.; Forte, M. A.;

*Journal of Biological Chemistry*, Vol.277, No.42, pp. 39289-39295

*Eukaryotic cell*, Vol.4, No.2, pp. 346-355

subunits. *Biochemistry*, Vol.45, No.21, pp. 6715-6723

*Chemistry*, Vol.284, No.49, pp. 33982-33988

*Chemistry*, Vol.278, No.12, pp. 10436-10442

*Biochemical Journal*, Vol.280, No.Pt 3, pp. 561-573

and F1-ATPase. *Nature*, Vol.409, No.6820, pp. 637-641

No.38, pp. 29502-29510

No.3, pp. 921-926

No.1, pp. 52-62

pp. 554-570

No.12, pp. 6940-6945

State of the Yeast F1c10-ATP Synthase. *Journal of Biological Chemistry*, Vol.285,

Differential steady-state tyrosine phosphorylation of two oligomeric forms of mitochondrial F0F1ATPsynthase: A structural proteomic analysis. *Proteomics*, Vol.6,

(2004). In vitro and in vivo studies of F0F1ATP synthase regulation by inhibitor protein IF1 in goat heart. *Biochimica et Biophysica Acta (BBA)-Bioenergetics*, Vol.1659,

ATP synthase dimeric complex does not require the ATPase inhibitor protein, Inh1.

Functional analysis of subunit e of the F1Fo-ATP synthase of the yeast Saccharomyces cerevisiae: importance of the N-terminal membrane anchor region.

based phosphoproteomics and quantification strategies. *Proteomics*, Vol.11, No.4,

participates in the yeast ATP synthase dimerization independently of e and g

quantification of proteins and phosphoproteins from cell lysates by tandem MS. *Proceedings of the National Academy of Sciences of the United States of America*, Vol.100,

ectopic F O F 1 ATP synthase of rat liver is modulated in acute cholestasis by the

Bernardi, P. & Lippe, G. (2009). Cyclophilin D modulates mitochondrial F0F1-ATP synthase by interacting with the lateral stalk of the complex. *Journal of Biological* 

& Coll, M. (2001). The bacterial conjugation protein TrwB resembles ring helicases

Vinten, J. r.; McCormack, J. G.; Reynet, C. & Beck-Nielsen, H. (2003). Proteome Analysis Reveals Phosphorylation of ATP Synthase β-Subunit in Human Skeletal Muscle and Proteins with Potential Roles in Type 2 Diabetes. *Journal of Biological* 


Cabezón, E.; Montgomery, M. G.; Leslie, A. G. W. & Walker, J. E. (2003). The structure of

Cabezon, E.; Runswick, M. J.; Leslie, A. G. W. & Walker, J. E. (2001). The structure of bovine

Campanella, M.; Parker, N.; Tan, C. H.; Hall, A. M. & Duchen, M. R. (2009). IF1: setting the

Carroll, J.; Altman, M. C.; Fearnley, I. M. & Walker, J. E. (2007). Identification of membrane

Carroll, J.; Fearnley, I. M.; Wang, Q. & Walker, J. E. (2009). Measurement of the molecular

Chen, C.; Ko, Y.; Delannoy, M.; Ludtke, S. J.; Chiu, W. & Pedersen, P. L. (2004).

Chen, C.; Saxena, A. K.; Simcoke, W. N.; Garboczi, D. N.; Pedersen, P. L. & Ko, Y. H. (2006).

Chen, R.; Fearnley, I. M.; Palmer, D. N. & Walker, J. E. (2004). Lysine 43 is trimethylated in

Chen, R.; Runswick, M. J.; Carroll, J.; Fearnley, I. M. & Walker, J. E. (2007). Association of

Cingolani, G. & Duncan, T. M. (2011). Structure of the ATP synthase catalytic complex (F1)

Cížková, A.; Stránecký, V.; Ivánek, R.; Hartmannová, H.; Nosková, L.; Piherová, L.; Tesa

Collinson, I. R.; Fearnley, I. M.; Skehel, J. M.; Runswick, M. J. & Walker, J. E. (1994). ATP

Couoh-Cardel, S. J.; Uribe-Carvajal, S.; Wilkens, S. & García-Trejo, J. J. (2010). Structure of

mitochondria. *FEBS letters*, Vol.581, No.17, pp. 3145-3148

*Molecular Biology*, Vol.18, No.6, pp. 701-707

deficiency. *BMC genomics*, Vol.9, No.1, pp. 38

*Biochemical Journal*, Vol.303, No.Pt 2, pp. 639-645

a single experiment. *Analytical Biochemistry*, Vol.395, No.2, pp. 249-255 Chaurand, P.; Cornett, D. S. & Caprioli, R. M. (2006). Molecular imaging of thin mammalian

*Academy of Sciences*, Vol.104, No.36, pp. 14330-14335

*biology*, Vol.10, No.9, pp. 744-750

No.24, pp. 6990-6996

350

pp. 431-436

31761-31768

13777-13783

21883-21887

36447-36455

bovine F1-ATPase in complex with its regulatory protein IF1. *Nature structural* 

IF1, the regulatory subunit of mitochondrial F-ATPase. *The EMBO journal*, Vol.20,

pace of the F1Fo-ATP synthase. *Trends in biochemical sciences*, Vol.34, No.7, pp. 343-

proteins by tandem mass spectrometry of protein ions. *Proceedings of the National* 

masses of hydrophilic and hydrophobic subunits of ATP synthase and complex I in

tissue sections by mass spectrometry. *Current opinion in biotechnology*, Vol.17, No.4,

Mitochondrial ATP Synthasome. *Journal of Biological Chemistry*, Vol.279, No.30, pp.

Mitochondrial ATP Synthase. *Journal of Biological Chemistry*, Vol.281, No.19, pp.

subunit c from bovine mitochondrial ATP synthase and in storage bodies associated with Batten disease. *Journal of Biological Chemistry*, Vol.279, No.21, pp.

two proteolipids of unknown function with ATP synthase from bovine heart

from Escherichia coli in an autoinhibited conformation. *Nature Structural &* 

ová, M.; Hansíková, H.; Honzík, T. & Zeman, J. (2008). Development of a human mitochondrial oligonucleotide microarray (h-MitoArray) and gene expression analysis of fibroblast cell lines from 13 patients with isolated F1Fo ATP synthase

synthase from bovine heart mitochondria: identification by proteolysis of sites in F0 exposed by removal of F1 and the oligomycin-sensitivity conferral protein.

dimeric F1F0-ATP synthase. *Journal of Biological Chemistry*, Vol.285, No.47, pp.


F0F1 ATP Synthase: A Fascinating Challenge for Proteomics 185

Lalanne, E.; Mathieu, C.; Vedel, F. & De Paepe, R. (1998). Tissue-specific expression of genes

Lee, J. K.; Belogrudov, G. I. & Stroud, R. M. (2008). Crystal structure of bovine mitochondrial

Lefebvre-Legendre, L.; Vaillier, J.; Benabdelhak, H.; Velours, J.; Slonimski, P. P. & di Rago, J.

Loro, E.; Gianazza, E.; Cazzola, S.; Malena, A.; Wait, R.; Begum, S.; Brizio, C.; Dabbeni Sala,

MacDonald, J. A.; Mackey, A. J.; Pearson, W. R. & Haystead, T. A. J. (2002). A strategy for

Malaval, C.; Laffargue, M.; Barbaras, R.; Rolland, C.; Peres, C.; Champagne, E.; Perret, B.;

Matsuda, C.; Endo, H.; Hirata, H.; Morosawa, H.; Nakanishi, M. & Kagawa, Y. (1993).

Matsuda, C.; Muneyuki, E.; Endo, H.; Yoshida, M. & Kagawa, Y. (1994). Comparison of the

Mattiazzi, M.; Vijayvergiya, C.; Gajewski, C. D.; DeVivo, D. C.; Lenaz, G.; Wiedmann, M. &

Meier, T.; Morgner, N.; Matthies, D.; Pogoryelov, D.; Keis, S.; Cook, G. M.; Dimroth, P. &

Meyer, B.; Wittig, I.; Trifilieff, E.; Karas, M. & Schägger, H. (2007). Identification of two

Mowery, Y. M. & Pizzo, S. V. (2010). Cell Surface ATP Synthase: A Potential Target for Anti-

Noji, H.; Yasuda, R.; Yoshida, M. & Kinosita, K. (1997). Direct observation of the rotation of

*Plant molecular biology*, Vol.38, No.5, pp. 885-888

*Electrophoresis*, Vol.30, No.8, pp. 1329-1341

*& Cellular Proteomics*, Vol.1, No.4, pp. 314-322

*Cellular signalling*, Vol.21, No.1, pp. 120-127

subunit. *FEBS letters*, Vol.325, No.3, pp. 281-284

*research communications*, Vol.200, No.2, pp. 671-678

*Human molecular genetics*, Vol.13, No.8, pp. 869-879

F1-ATPase. *Nature*, Vol.386, No.6622, pp. 299-302

No.5, pp. 1181-1192

*Function*, pp. 139-159

Vol.6, No.10, pp. 1690-1699

*Journal of Biological Chemistry*, Vol.276, No.9, pp. 6789-6796

No.36, pp. 13379-13384

encoding isoforms of the mitochondrial ATPase subunit in Nicotiana sylvestris.

factor B at 0.96-Å resolution. *Proceedings of the National Academy of Sciences*, Vol.105,

P. (2001). Identification of a nuclear gene (FMC1) required for the assembly/stability of yeast mitochondrial F1-ATPase in heat stress conditions.

F. & Vergani, L. (2009). Development and characterization of polyspecific anti mitochondrion antibodies for proteomics studies on in toto tissue homogenates.

the rapid identification of phosphorylation sites in the phosphoproteome. *Molecular* 

Tercé, F.; Collet, X. & Martinez, L. O. (2009). RhoA/ROCK I signalling downstream of the P2Y13 ADP-receptor controls HDL endocytosis in human hepatocytes.

Tissue-specific isoforms of the bovine mitochondrial ATP synthase [gamma]-

ATPase Activities of Bovine Heart and Liver Mitochondrial ATP Synthases with Different Tissue-Specific [gamma] Subunit Isoforms. *Biochemical and biophysical* 

Manfredi, G. (2004). The mtDNA T8993G (NARP) mutation results in an impairment of oxidative phosphorylation that can be improved by antioxidants.

Brutschy, B. (2007). A tridecameric c ring of the adenosine triphosphate (ATP) synthase from the thermoalkaliphilic Bacillus sp. strain TA2. A1 facilitates ATP synthesis at low electrochemical proton potential. *Molecular microbiology*, Vol.65,

proteins associated with mammalian ATP synthase. *Molecular & Cellular Proteomics*,

Angiogenic Therapy. *Extracellular ATP and Adenosine as Regulators of Endothelial Cell* 


Heather, L. C. & Clarke, K. (2010). Metabolism, hypoxia and the diabetic heart. *Journal of* 

Helfenbein, K. G.; Ellis, T. P.; Dieckmann, C. L. & Tzagoloff, A. (2003). ATP22, a nuclear

Hoffmann, J.; Sokolova, L.; Preiss, L.; Hicks, D. B.; Krulwich, T. A.; Morgner, N.; Wittig, I.;

Hong, S. & Pedersen, P. L. (2003). Subunit e of mitochondrial ATP synthase: a bioinformatic

*Proteins: Structure, Function, and Bioinformatics*, Vol.51, No.2, pp. 155-161 Houstek, J.; Andersson, U.; Tvrdík, P.; Nedergaard, J. & Cannon, B. (1995). The expression of

Houstek, J.; Pícková, A.; Vojtísková, A.; Mrácek, T.; Pecina, P. & Jesina, P. (2006).

Hüttemann, M.; Lee, I.; Samavati, L.; Yu, H. & Doan, J. W. (2007). Regulation of

*Biophysica Acta (BBA)-Molecular Cell Research*, Vol.1773, No.12, pp. 1701-1720 Hwang, S. I.; Lundgren, D. H.; Mayya, V.; Rezaul, K.; Cowan, A. E.; Eng, J. K. & Han, D. K.

Klodmann, J.; Lewejohann, D. & Braun, H. P. (2011). Low SDS Blue native PAGE. *Proteomics*,

Ko, Y. H.; Pan, W.; Inoue, C. & Pedersen, P. L. (2002). Signal transduction to mitochondrial

Krause-Buchholz, U.; Becker, J. S.; Zoriy, M.; Pickhardt, C.; Przybylski, M. & Rodel, G.

Kucharczyk, R.; Salin, B. & Di Rago, J. P. (2009). Introducing the human Leigh syndrome

lysophosphatidic acid. *Mitochondrion*, Vol.1, No.4, pp. 339-348

*(BBA)-Molecular Cell Research*, Vol.1793, No.1, pp. 186-199

*Acta (BBA)-Bioenergetics*, Vol.1757, No.9-10, pp. 1400-1405

*& Cellular Proteomics*, Vol.5, No.6, pp. 1131-1145

gene required for expression of the F0 sector of mitochondrial ATPase in Saccharomyces cerevisiae. *Journal of Biological Chemistry*, Vol.278, No.22, pp. 19751-

Schägger, H.; Meier, T. & Brutschy, B. (2010). ATP synthases: cellular nanomotors characterized by LILBID mass spectrometry. *Phys. Chem. Chem. Phys.*, Vol.12,

analysis reveals a phosphopeptide binding motif supporting a multifunctional regulatory role and identifies a related human brain protein with the same motif.

subunit c correlates with and thus may limit the biosynthesis of the mitochondrial F0F1-ATPase in brown adipose tissue. *Journal of Biological Chemistry*, Vol.270, No.13,

Mitochondrial diseases and genetic defects of ATP synthase. *Biochimica et Biophysica* 

mitochondrial oxidative phosphorylation through cell signaling. *Biochimica et* 

(2006). Systematic characterization of nuclear proteome during apoptosis. *Molecular* 

ATP synthase: Evidence that PDGF-dependent phosphorylation of the [delta] subunit occurs in several cell lines, involves tyrosine, and is modulated by

(2006). Detection of phosphorylated subunits by combined LA-ICP-MS and MALDI-FTICR-MS analysis in yeast mitochondrial membrane complexes separated by blue native/SDS-PAGE. *International Journal of Mass Spectrometry*, Vol.248, No.1-

mutation T9176G into Saccharomyces cerevisiae mitochondrial DNA leads to severe defects in the incorporation of Atp6p into the ATP synthase and in the mitochondrial morphology. *Human molecular genetics*, Vol.18, No.15, pp. 2889 Kucharczyk, R.; Zick, M.; Bietenhader, M.; Rak, M.; Couplan, E.; Blondel, M.; Caubet, S. D. &

di Rago, J. P. (2009). Mitochondrial ATP synthase disorders: molecular mechanisms and the quest for curative therapeutic approaches. *Biochimica et Biophysica Acta* 

*Molecular and Cellular Cardiology*, Vol.50, No.4, pp. 529-540

19756

No.41, pp. 13375-13382

pp. 7689-7694

pp. 1-6

2, pp. 56-60


F0F1 ATP Synthase: A Fascinating Challenge for Proteomics 187

Struglics, A.; Fredlund, K. M.; Møller, I. M. & Allen, J. F. (1998). Two Subunits of the FoF1-

Suhai, T.; Heidrich, N. G.; Dencher, N. A. & Seelert, H. (2009). Highly sensitive detection of ATPase activity in native gels. *Electrophoresis*, Vol.30, No.20, pp. 3622-3625 Suzuki, T.; Wakabayashi, C.; Tanaka, K.; Feniouk, B. A. & Yoshida, M. (2011). Modulation of

Thomas, D.; Bron, P.; Weimann, T.; Dautant, A.; Giraud, M.; Paumard, P.; Salin, B.; Cavalier,

Tomasetig, L.; Di Pancrazio, F.; Harris, D. A.; Mavelli, I. & Lippe, G. (2002). Dimerization of

Tzagoloff, A. (1969). Assembly of the mitochondrial membrane system. II. Synthesis of the

Ueno, H.; Suzuki, T.; Kinosita, K. & Yoshida, M. (2005). ATP-driven stepwise rotation of

Vantourout, P.; Radojkovic, C.; Lichtenstein, L.; Pons, V.; Champagne, E. & Martinez, L. O.

Vives-Bauza, C.; Magrane, J.; Andreu, A. L. & Manfredi, G. (2011). Novel Role of ATPase

Vosseller, K.; Hansen, K. C.; Chalkley, R. J.; Trinidad, J. C.; Wells, L.; Hart, G. W. &

Walker, J. E. & Dickson, V. K. (2006). The peripheral stalk of the mitochondrial ATP

Walker, J. E.; Lutter, R.; Dupuis, A. & Runswick, M. J. (1991). Identification of the subunits of

Wittig, I. & Schägger, H. (2005). Advantages and limitations of clear native PAGE.

Wittig, I. & Schägger, H. (2008). Features and applications of blue-native and clear-native

Wittig, I. & Schägger, H. (2008). Structural organization of mitochondrial ATP synthase. *Biochimica et Biophysica Acta (BBA)-Bioenergetics*, Vol.1777, No.7-8, pp. 592-598

*biophysical research communications*, Vol.243, No.3, pp. 664-668

*Biological Chemistry*, Vol.286, No.19, pp. 16807-16813

ATP synthase. *Biology of the Cell*, Vol.100, pp. 591-601

*States of America*, Vol.102, No.5, pp. 1333-1338

*Biology of the Cell*, Vol.21, No.1, pp. 131-139

*Proteomics*, Vol.5, No.17, pp. 4338-4346

with dithiothreitol. *Proteomics*, Vol.5, No.2, pp. 388-398

electrophoresis. *Proteomics*, Vol.8, No.19, pp. 3974-3990

synthase. *BBA-Bioenergetics*, Vol.1757, No.5-6, pp. 286-296

No.2-3, pp. 133-141

No.1, pp. 33-70

5369-5378

Vol.244, No.18, pp. 5027-5033

ATPase Are Phosphorylated in the Inner Mitochondrial Membrane. *Biochemical and* 

nucleotide specificity of thermophilic FoF1-ATP synthase by -subunit. *Journal of* 

A.; Velours, J. & Brethes, D. (2008). Supramolecular organization of the yeast F1Fo-

F0F1ATP synthase from bovine heart is independent from the binding of the inhibitor protein IF1. *Biochimica et Biophysica Acta (BBA)-Bioenergetics*, Vol.1556,

mitochondrial adenosine triphosphatase. F1. *The Journal of biological chemistry*,

FoF1-ATP synthase. *Proceedings of the National Academy of Sciences of the United* 

(2010). Ecto-F1-ATPase: A moonlighting protein complex and an unexpected apoA-I receptor. *World Journal of Gastroenterology: WJG*, Vol.16, No.47, pp. 5925-5935 Vignais, P. V. & Satre, M. (1984). Recent developments on structural and functional aspects

of the F 1 sector of H+-linked ATPases. *Molecular and cellular biochemistry*, Vol.60,

Subunit C Targeting Peptides Beyond Mitochondrial Protein Import. *Molecular* 

Burlingame, A. L. (2005). Quantitative analysis of both protein expression and serine/threonine post translational modifications through stable isotope labeling

F1F0-ATPase from bovine heart mitochondria. *Biochemistry*, Vol.30, No.22, pp.


Ohsakaya, S.; Fujikawa, M.; Hisabori, T. & Yoshida, M. (2011). Knockdown of DAPIT

Pagnozzi, D.; Birolo, L.; Leo, G.; Contessi, S.; Lippe, G.; Pucci, P. & Mavelli, I. (2010).

Paumard, P.; Vaillier, J.; Coulary, B.; Schaeffer, J.; Soubannier, V.; Mueller, D. M.; Brèthes,

Reisinger, V. & Eichacker, L. A. (2008). Solubilization of membrane protein complexes for

Richter, M. L.; Hein, R. & Huchzermeyer, B. (2000). Important subunit interactions in the

Schägger, H. & Pfeiffer, K. (2000). Supercomplexes in the respiratory chains of yeast and mammalian mitochondria. *The EMBO journal*, Vol.19, No.8, pp. 1777-1783 Senior, A. E. (2007). ATP synthase: Motoring to the finish line. *Cell*, Vol.130, No.2, pp. 220-

Sgarbi, G.; Baracca, A.; Lenaz, G.; Valentino, L. M.; Carelli, V. & Solaini, G. (2006). Inefficient

Sokolova, L.; Wittig, I.; Barth, H. D.; Schägger, H.; Brutschy, B. & Brandt, U. (2010). Laser

Stevens Jr, S. M.; Duncan, R. S.; Koulen, P. & Prokai, L. (2008). Proteomic analysis of mouse

Stockwin, L. H.; Blonder, J.; Bumke, M. A.; Lucas, D. A.; Chan, K. C.; Conrads, T. P.; Issaq,

Strauss, M.; Hofhaus, G.; Schröder, R. R. & Kühlbrandt, W. (2008). Dimer ribbons of ATP

blue native PAGE. *Journal of Proteomics*, Vol.71, No.3, pp. 277-283

mtDNA. *Biochemical Journal*, Vol.395, No.Pt 3, pp. 493-500

*Biophysica Acta (BBA)-Biomembranes*, Vol.1414, No.1-2, pp. 260-264

20296

221

Vol.8, No.2, pp. 231-239

Vol.1458, No.2-3, pp. 326-342

*research*, Vol.7, No.3, pp. 1046-1054

*research*, Vol.5, No.11, pp. 2996-3007

No.7, pp. 1154-1160

synthase. *The EMBO journal*, Vol.30, pp. 920-930

(diabetes-associated protein in insulin-sensitive tissue) results in loss of ATP synthase in mitochondria. *Journal of Biological Chemistry*, Vol.286, No.23, pp. 20292-

Stoichiometry and Topology of the complex of the endogenous ATP Synthase Inhibitor Protein IF1 with Calmodulin. *Biochemistry*, Vol.49, No.35, pp. 7542-7552 Panfoli, I.; Ravera, S.; Bruschi, M.; Candiano, G. & Morelli, A. (2011). Proteomics unravels

the exportability of mitochondrial respiratory chains. *Expert Review of Proteomics*,

D.; Di Rago, J. P. & Velours, J. (2002). The ATP synthase is involved in generating mitochondrial cristae morphology. *The EMBO journal*, Vol.21, No.3, pp. 221-230 Rak, M.; Gokova, S. & Tzagoloff, A. (2011). Modular assembly of yeast mitochondrial ATP

chloroplast ATP synthase. *Biochimica et Biophysica Acta (BBA)-Bioenergetics*,

coupling between proton transport and ATP synthesis may be the pathogenic mechanism for NARP and Leigh syndrome resulting from the T8993G mutation in

induced liquid bead ion desorption MS of protein complexes from blue native gels, a sensitive top down proteomic approach. *Proteomics*, Vol.10, No.7, pp. 1401-1407 Spannagel, C.; Vaillier, J.; Arselin, G.; Graves, P. V.; Grandier-Vazeille, X. & Velours, J.

(1998). Evidence of a subunit 4 (subunit b) dimer in favor of the proximity of ATP synthase complexes in yeast inner mitochondrial membrane. *Biochimica et* 

brain microsomes: identification and bioinformatic characterization of endoplasmic reticulum proteins in the mammalian central nervous system. *Journal of proteome* 

H. J.; Veenstra, T. D.; Newton, D. L. & Rybak, S. M. (2006). Proteomic analysis of plasma membrane from hypoxia-adapted malignant melanoma. *Journal of proteome* 

synthase shape the inner mitochondrial membrane. *The EMBO journal*, Vol.27,


**9** 

*USA* 

Noriko Yokoyama

**Proteomic Analysis of Wnt-Dependent** 

*State University of New York at Stony Brook, Stony Brook, New York* 

**Dishevelled-Based Supermolecular Complexes** 

Wnt signaling is critical and indispensable for numerous cellular pathways including embryonic development and adult tissue homeostasis such as determination, proliferation, migration and differentiation (Clevers, 2006; Glass and Karsenty, 2006; Nusse, 2005; Wang and Wynshaw-Boris, 2004). There are three independent branches of the Wnt signaling cascade, Wnt canonical beta-catenin/Lef-Tcf-sensitive transcriptional response, the planar cell polarity (PCP) response, and the Wnt5a/Frizzled-2/Ca2+/cGMP response (He et al., 2004; Katanaev et al., 2005; Liu et al., 2001; Malbon, 2005; Wallingford and Habas, 2005; Willert et al., 2003). Dishevelleds (Dvls) are essential components in three major Wnt signaling pathways. Dvls function as scaffold protein bridging the receptors and distinct downstream signaling components. Through formation of dynamic multiprotein complexes,

In Wnt/beta-catenin signaling pathway, Wnts bind to members of the Frizzled (Fz) family of G protein-coupled receptors and to co-receptor low density lipoprotein receptor related protein LRP5 and LRP6 (LRP5/6). Binding of Wnt3a to these receptors facilitates a variety of intracellular events. Phosphorylation of LRP5/6 triggers the interaction of Fz-LRP5/6 complex with dynamic multiprotein complexes, including *adenomatous polyposis coli* (APC), Dishevelled (Dsh/Dvl), Axin, glycogen synthase kinase-3β (GSK3β) and casein kinase 1α (Hart, 1998; Kishida et al., 1998). Scaffold protein Dvl facilitates destruction of multiprotein complexes (Hart, 1998 ; Kishida et al., 1998; Malbon, 2005). As a results, stabilized β-catenin, translocates to nuclear and enhances Lef/Tcf-sensitive transcription

Fly Dsh and three isoforms of mammals Dvls (Dvl1-3) all share several prominent, highlyconserved domains: a Dsh homology domain called DIX; a conserved sequence element with homology to the postsynaptic density protein *P*SD-95, *D*iscs-large, and *Z*O-1, termed PDZ; and *D*isheveled, *E*gl-10, *P*leckstrin domain, termed DEP (Wharton, 2003). In addition, there are conserved sequence regions harboring basic amino acids residues and a prolinerich putative Src homology 3 (SH3) binding domain (Penton et al., 2002). Dvl3 knockout mouse is lethal, whereas Dvl1 or Dvl2 knockout mouse is viable (Hamblet et al., 2002; Etheridge et al., 2008; Lijam et al., 1997). The isoforms of mammals Dvl are not truly "redundant" with respect to function (Hamblet et al., 2002; Etheridge et al., 2008; Lee et al., 2008), although functions of each isoform of Dvls are not fully resolved. Previous studies reveal distinct localization of Dvl isoforms in totipotent mouse F9 embryonal

(Angers & Moon, 2009; Clevers, 2006; van Amerongen and Nusse, 2009).

**1. Introduction** 

signal is transduced to inside cells.


### **Proteomic Analysis of Wnt-Dependent Dishevelled-Based Supermolecular Complexes**

Noriko Yokoyama *State University of New York at Stony Brook, Stony Brook, New York USA* 

#### **1. Introduction**

188 Proteomics – Human Diseases and Protein Functions

Wittig, I.; Braun, H. P. & Schägger, H. (2006). Blue native PAGE. *Nature protocols*, Vol.1,

Wittig, I.; Carrozzo, R.; Santorelli, F. M. & Schagger, H. (2007). Functional assays in high-

Wittig, I.; Meyer, B.; Heide, H.; Steger, M.; Bleier, L.; Wumaier, Z.; Karas, M. & Schägger, H.

Wittig, I.; Velours, J.; Stuart, R. & Schägger, H. (2008). Characterization of domain interfaces

Yan, L. J. & Forster, M. J. (2009). Resolving mitochondrial protein complexes using

Zerbetto, E.; Vergani, L. & Dabbeni Sala, F. (1997). Quantification of muscle mitochondrial

Zhang, L.; Xi'e, W.; Peng, X.; Wei, Y.; Cao, R.; Liu, Z.; Xiong, J.; Ying, X.; Chen, P. & Liang, S.

polyacrylamide gels. *Electrophoresis*, Vol.18, No.11, pp. 2059-2064

and cell lines. *Electrophoresis*, Vol.28, No.21, pp. 3811-3820

resolution clear native gels to quantify mitochondrial complexes in human biopsies

(2010). Assembly and oligomerization of human ATP synthase lacking mitochondrial subunits a and A6L. *Biochimica et Biophysica Acta (BBA)-Bioenergetics*,

in monomeric and dimeric ATP synthase. *Molecular & Cellular Proteomics*, Vol.7,

nongradient blue native polyacrylamide gel electrophoresis. *Analytical Biochemistry*,

oxidative phosphorylation enzymes via histochemical staining of blue native

(2007). Immunoaffinity purification of plasma membrane with secondary antibody superparamagnetic beads for proteomic analysis. *Journal of proteome research*, Vol.6,

No.1, pp. 418-428

No.5, pp. 995-1004

No.1, pp. 34-43

Vol.389, No.2, pp. 143-149

Vol.1797, No.6-7, pp. 1004-1011

Wnt signaling is critical and indispensable for numerous cellular pathways including embryonic development and adult tissue homeostasis such as determination, proliferation, migration and differentiation (Clevers, 2006; Glass and Karsenty, 2006; Nusse, 2005; Wang and Wynshaw-Boris, 2004). There are three independent branches of the Wnt signaling cascade, Wnt canonical beta-catenin/Lef-Tcf-sensitive transcriptional response, the planar cell polarity (PCP) response, and the Wnt5a/Frizzled-2/Ca2+/cGMP response (He et al., 2004; Katanaev et al., 2005; Liu et al., 2001; Malbon, 2005; Wallingford and Habas, 2005; Willert et al., 2003). Dishevelleds (Dvls) are essential components in three major Wnt signaling pathways. Dvls function as scaffold protein bridging the receptors and distinct downstream signaling components. Through formation of dynamic multiprotein complexes, signal is transduced to inside cells.

In Wnt/beta-catenin signaling pathway, Wnts bind to members of the Frizzled (Fz) family of G protein-coupled receptors and to co-receptor low density lipoprotein receptor related protein LRP5 and LRP6 (LRP5/6). Binding of Wnt3a to these receptors facilitates a variety of intracellular events. Phosphorylation of LRP5/6 triggers the interaction of Fz-LRP5/6 complex with dynamic multiprotein complexes, including *adenomatous polyposis coli* (APC), Dishevelled (Dsh/Dvl), Axin, glycogen synthase kinase-3β (GSK3β) and casein kinase 1α (Hart, 1998; Kishida et al., 1998). Scaffold protein Dvl facilitates destruction of multiprotein complexes (Hart, 1998 ; Kishida et al., 1998; Malbon, 2005). As a results, stabilized β-catenin, translocates to nuclear and enhances Lef/Tcf-sensitive transcription (Angers & Moon, 2009; Clevers, 2006; van Amerongen and Nusse, 2009).

Fly Dsh and three isoforms of mammals Dvls (Dvl1-3) all share several prominent, highlyconserved domains: a Dsh homology domain called DIX; a conserved sequence element with homology to the postsynaptic density protein *P*SD-95, *D*iscs-large, and *Z*O-1, termed PDZ; and *D*isheveled, *E*gl-10, *P*leckstrin domain, termed DEP (Wharton, 2003). In addition, there are conserved sequence regions harboring basic amino acids residues and a prolinerich putative Src homology 3 (SH3) binding domain (Penton et al., 2002). Dvl3 knockout mouse is lethal, whereas Dvl1 or Dvl2 knockout mouse is viable (Hamblet et al., 2002; Etheridge et al., 2008; Lijam et al., 1997). The isoforms of mammals Dvl are not truly "redundant" with respect to function (Hamblet et al., 2002; Etheridge et al., 2008; Lee et al., 2008), although functions of each isoform of Dvls are not fully resolved. Previous studies reveal distinct localization of Dvl isoforms in totipotent mouse F9 embryonal

Proteomic Analysis of Wnt-Dependent Dishevelled-Based Supermolecular Complexes 191

dynamic character of the Dvls-based complexes is the key to understand Wnt signaling. For the first time, the assembly of supermolecular Dvl3-based complexes is shown in response to Wnt3a. Peak fractions (Dvl3-based supermolecular complexes) separated by SEC were subjected to LC-ESI-MS-MS. To avoid eliminating potential contamination of proteins/complexes, analysis was carried out at distinct time points. These approaches

Wnt signaling is spatially and temporally transduced through the assembly of dynamic multiprotein complexes. Dvls, scaffold proteins, provide dynamic protein platform including protein kinases, phosphatases, receptors, adaptor molecules, and other signaling molecules. The Wnt home page provides an invaluable frame of Dvls interacting proteins (**http://www.stanford.edu/~rnusse/wntwindow.html**), although real Dvls-based complexes seem much more complicated. Recent measurement of the mass of Dvl3-based complexes showed that the MW of the complexes is >2 MegaDa by SEC and ~35 MegaDa by *fcs*  (Yokoyama et al., 2010). Structure and functional analysis of these Dvls-based supermolecular complexes is critical to understanding Wnt signaling. Proteomics provide a portal to identify complex partners assembled into signalsomes. In this study, two distinct

**2.1 Analysis of Dvl2 interacting proteins using glutathione-S-transferase (GST) fusion** 

The pull-down technique is an invaluable tool for studying cellular pathways via proteinprotein interactions. GST fusion protein pull down experiments are one approach to identify interaction of between probe protein and unknown targets. GST fusion protein pulls down offers an important biological assay for direct protein-to-protein interactions. In this study, GST fusion proteins of Dvl2 domain are employed to discover novel Dvl2 interacting proteins. Identification of novel Dvls interacting proteins facilitates understanding the

GST fusion proteins of the conserved domains DIX, PDZ, and DEP and the putative SH3 binding domain of Dvl2 were immobilized on glutathione-derivatized agarose matrix. Immobilized GST-PDZ (aa 267-309), GST-DIX (aa 11-93), GST-DEP (aa 433-507), GSTputative SH3 binding containing region (aa 356-378) and GST itself (as a control) were incubated with cell lysates from F9 cells stimulated with or without Wnt 3a. The interacting proteins were pulled down and eluted from the beads. Eluted proteins were separated by immobilized pH gradient (IPG) strips (pH 3-10, first dimension separation) and subjected to second dimensional SDS-gel electrophoresis. Proteins were stained with SYPRO Ruby. Non specific proteins were eliminated by comparing gel patterns obtained with GST-Dvl2

To identify Wnt-dependent interacting proteins, spots according in response to Wnt3a stimulation were excised from the gel (fig. 2), digested with trypsin overnight at 37 oC and analyzed by liquid chromatography electrospray ionization mass spectrometry (LC-ESI-MS-MS, Applied Biosystems/MDS SCIEX) using a micro-column reverse phase HPLC interfaced to an LTQ ion trap mass spectrometer. Electrospray tandem mass spectrometry

identified both expected and also novel components.

approaches are employed.

regulatory mechanism of Wnt signaling.

domains with those obtained by GST itself (fig. 1).

**2.1.1 Glutathione-S-transferase (GST) fusion protein pulls down** 

**protein pull down** 

**2. Proteomic analysis of Dvls-based multiprotein complexes** 

teratocarcinoma (F9) cells. Dvl1 and Dvl2 are found in the cytosol-enriched fraction (80 %), the least amount of each of these isoforms is observed in the nuclear-enriched fraction (5 %). Dvl3 content in the cytosol-enriched fraction is reduced to 60 %, whereas content of Dvl3 in the nuclear-enriched fraction is several fold greater than that for either Dvl1 or Dvl2 (20 %). Abundance of Dvl3 and Dvl1 following Wnt3a stimulation is increased and Wnt3a stimulates dramatic Dvl3 trafficking to the plasma membrane. Thus, Dvl3 has a role in Wnt/β-catenin signaling that is distinct from those of the Dvl1 and Dvl2 (Yokoyama et al., 2007). Dvl3 has unique roles in Wnt/β-catenin signaling pathway (Lee et al., 2008; Yokoyama et al., 2007). Dvls are believed to be recruited Fz receptor through their PDZ domain. Many Dvls associating proteins also interact with Dvls via their PDZ domain (Wong et al., 2003). The DIX domain is critical for the ability of Dvl to recruit Axin and for the dynamic self-association of Dvls. Fluorescence microscopy in cells suggests that Dvls undergo dynamic oligomerization associated with activation of Wnt/β-catenin signaling by Wnt3a (Axelrod et al., 1998; Kishida et al., 1999; Rothbacher et al., 2000). Properties of dynamic polymerization correlate with the activation of Wnt/β-catenin signaling (Schwarz-Romond et al., 2007a; Schwarz-Romond et al., 2007b). Wnt induces LRP6 aggregation and phosphorylation in a Dvl-dependent manner. Furthermore, Dvl mutants that lack normal oligomerization also influence the formation of LRP6 complexes (Bilic et al., 2007).

Thus, Dvls scaffold many interacting proteins temporally/spatially to transduce Wnt signaling. To define a regulation of Wnt signaling, it is necessary to establish the regulatory mechanism of assembly of Dvls-based supermolecular complexes. Tight regulation of Wnt signaling by protein-protein interaction involves many post-translational modifications such as phosphorylation, ubiquitination, sumoylation and methylation. Phosphorylation is a crucial and central mechanism of regulating docking of signal molecules to Dvl scaffold. Phosphorylation of protein regulates protein activities, binding affinities, stability and its trafficking to distinct cellular compartments (Yokoyama and Malbon, 2007; Yokoyama et al., 2007; Yokoyama and Malbon, 2009).

To probing Dvls-based complexes, two different approaches were employed. First, our attention is to identify interacting proteins of Dvl2 (most abundant form of Dvl in F9 cells). Dvl2 interacting proteins have been established using glutathione-S-transferase (GST) fusion protein pull down. Interacting proteins were pulled down and eluted from the beads. Eluted proteins were separated by immobilized pH gradient (IPG) strips (pH 3-10, first dimension separation) and subjected to second dimensional SDS-gel electrophoresis. Samples were analyzed by liquid chromatography electrospray ionization mass spectrometry (LC-ESI-MS-MS). Novel Dvls interacting proteins were identified. In addition, immobilized Dvl2 domains pull down is able to find novel roles of Src in Wnt signaling. Identification of tyrosine phosphorylation sites on Dvl2 by Src family tyrosine kinases led to discover that Src family tyrosine kinases are a positive regulator of Wnt/β-catenin signaling (Yokoyama and Malbon, 2009). Second attempt is designed to probe Wnt-dependent assembly of Dvlsbased supermolecular complexes. Dishevelled-based "punctae" had been observed earlier by fluorescence microscopy. These "punctae" have been shown to be the assemblies of protein and the size and cellular distribution changed in response to Wnt stimulation. The physical evidence for the existence for these putative "aggregates" or "punctae" of Dvl3 based complexes was established using size-exclusion chromatography (SEC) technique, affinity pull-downs, proteomics, and fluorescent correlation microscopy (*fcs*). Dvl3-based complexes were interrogated physically *in vitro* by SEC analysis of cell extracts and *in vivo* by *fcs* analysis in live cells (Yokoyama et al., 2010). Establishment of physical nature and

teratocarcinoma (F9) cells. Dvl1 and Dvl2 are found in the cytosol-enriched fraction (80 %), the least amount of each of these isoforms is observed in the nuclear-enriched fraction (5 %). Dvl3 content in the cytosol-enriched fraction is reduced to 60 %, whereas content of Dvl3 in the nuclear-enriched fraction is several fold greater than that for either Dvl1 or Dvl2 (20 %). Abundance of Dvl3 and Dvl1 following Wnt3a stimulation is increased and Wnt3a stimulates dramatic Dvl3 trafficking to the plasma membrane. Thus, Dvl3 has a role in Wnt/β-catenin signaling that is distinct from those of the Dvl1 and Dvl2 (Yokoyama et al., 2007). Dvl3 has unique roles in Wnt/β-catenin signaling pathway (Lee et al., 2008; Yokoyama et al., 2007). Dvls are believed to be recruited Fz receptor through their PDZ domain. Many Dvls associating proteins also interact with Dvls via their PDZ domain (Wong et al., 2003). The DIX domain is critical for the ability of Dvl to recruit Axin and for the dynamic self-association of Dvls. Fluorescence microscopy in cells suggests that Dvls undergo dynamic oligomerization associated with activation of Wnt/β-catenin signaling by Wnt3a (Axelrod et al., 1998; Kishida et al., 1999; Rothbacher et al., 2000). Properties of dynamic polymerization correlate with the activation of Wnt/β-catenin signaling (Schwarz-Romond et al., 2007a; Schwarz-Romond et al., 2007b). Wnt induces LRP6 aggregation and phosphorylation in a Dvl-dependent manner. Furthermore, Dvl mutants that lack normal

oligomerization also influence the formation of LRP6 complexes (Bilic et al., 2007).

2007; Yokoyama and Malbon, 2009).

Thus, Dvls scaffold many interacting proteins temporally/spatially to transduce Wnt signaling. To define a regulation of Wnt signaling, it is necessary to establish the regulatory mechanism of assembly of Dvls-based supermolecular complexes. Tight regulation of Wnt signaling by protein-protein interaction involves many post-translational modifications such as phosphorylation, ubiquitination, sumoylation and methylation. Phosphorylation is a crucial and central mechanism of regulating docking of signal molecules to Dvl scaffold. Phosphorylation of protein regulates protein activities, binding affinities, stability and its trafficking to distinct cellular compartments (Yokoyama and Malbon, 2007; Yokoyama et al.,

To probing Dvls-based complexes, two different approaches were employed. First, our attention is to identify interacting proteins of Dvl2 (most abundant form of Dvl in F9 cells). Dvl2 interacting proteins have been established using glutathione-S-transferase (GST) fusion protein pull down. Interacting proteins were pulled down and eluted from the beads. Eluted proteins were separated by immobilized pH gradient (IPG) strips (pH 3-10, first dimension separation) and subjected to second dimensional SDS-gel electrophoresis. Samples were analyzed by liquid chromatography electrospray ionization mass spectrometry (LC-ESI-MS-MS). Novel Dvls interacting proteins were identified. In addition, immobilized Dvl2 domains pull down is able to find novel roles of Src in Wnt signaling. Identification of tyrosine phosphorylation sites on Dvl2 by Src family tyrosine kinases led to discover that Src family tyrosine kinases are a positive regulator of Wnt/β-catenin signaling (Yokoyama and Malbon, 2009). Second attempt is designed to probe Wnt-dependent assembly of Dvlsbased supermolecular complexes. Dishevelled-based "punctae" had been observed earlier by fluorescence microscopy. These "punctae" have been shown to be the assemblies of protein and the size and cellular distribution changed in response to Wnt stimulation. The physical evidence for the existence for these putative "aggregates" or "punctae" of Dvl3 based complexes was established using size-exclusion chromatography (SEC) technique, affinity pull-downs, proteomics, and fluorescent correlation microscopy (*fcs*). Dvl3-based complexes were interrogated physically *in vitro* by SEC analysis of cell extracts and *in vivo* by *fcs* analysis in live cells (Yokoyama et al., 2010). Establishment of physical nature and dynamic character of the Dvls-based complexes is the key to understand Wnt signaling. For the first time, the assembly of supermolecular Dvl3-based complexes is shown in response to Wnt3a. Peak fractions (Dvl3-based supermolecular complexes) separated by SEC were subjected to LC-ESI-MS-MS. To avoid eliminating potential contamination of proteins/complexes, analysis was carried out at distinct time points. These approaches identified both expected and also novel components.

#### **2. Proteomic analysis of Dvls-based multiprotein complexes**

Wnt signaling is spatially and temporally transduced through the assembly of dynamic multiprotein complexes. Dvls, scaffold proteins, provide dynamic protein platform including protein kinases, phosphatases, receptors, adaptor molecules, and other signaling molecules. The Wnt home page provides an invaluable frame of Dvls interacting proteins (**http://www.stanford.edu/~rnusse/wntwindow.html**), although real Dvls-based complexes seem much more complicated. Recent measurement of the mass of Dvl3-based complexes showed that the MW of the complexes is >2 MegaDa by SEC and ~35 MegaDa by *fcs*  (Yokoyama et al., 2010). Structure and functional analysis of these Dvls-based supermolecular complexes is critical to understanding Wnt signaling. Proteomics provide a portal to identify complex partners assembled into signalsomes. In this study, two distinct approaches are employed.

#### **2.1 Analysis of Dvl2 interacting proteins using glutathione-S-transferase (GST) fusion protein pull down**

The pull-down technique is an invaluable tool for studying cellular pathways via proteinprotein interactions. GST fusion protein pull down experiments are one approach to identify interaction of between probe protein and unknown targets. GST fusion protein pulls down offers an important biological assay for direct protein-to-protein interactions. In this study, GST fusion proteins of Dvl2 domain are employed to discover novel Dvl2 interacting proteins. Identification of novel Dvls interacting proteins facilitates understanding the regulatory mechanism of Wnt signaling.

#### **2.1.1 Glutathione-S-transferase (GST) fusion protein pulls down**

GST fusion proteins of the conserved domains DIX, PDZ, and DEP and the putative SH3 binding domain of Dvl2 were immobilized on glutathione-derivatized agarose matrix. Immobilized GST-PDZ (aa 267-309), GST-DIX (aa 11-93), GST-DEP (aa 433-507), GSTputative SH3 binding containing region (aa 356-378) and GST itself (as a control) were incubated with cell lysates from F9 cells stimulated with or without Wnt 3a. The interacting proteins were pulled down and eluted from the beads. Eluted proteins were separated by immobilized pH gradient (IPG) strips (pH 3-10, first dimension separation) and subjected to second dimensional SDS-gel electrophoresis. Proteins were stained with SYPRO Ruby. Non specific proteins were eliminated by comparing gel patterns obtained with GST-Dvl2 domains with those obtained by GST itself (fig. 1).

To identify Wnt-dependent interacting proteins, spots according in response to Wnt3a stimulation were excised from the gel (fig. 2), digested with trypsin overnight at 37 oC and analyzed by liquid chromatography electrospray ionization mass spectrometry (LC-ESI-MS-MS, Applied Biosystems/MDS SCIEX) using a micro-column reverse phase HPLC interfaced to an LTQ ion trap mass spectrometer. Electrospray tandem mass spectrometry

Proteomic Analysis of Wnt-Dependent Dishevelled-Based Supermolecular Complexes 193

peptide identified

Dvl1 6 14 6.9 RGS18 2 5 6.8 Dvl3 3 4 3.2

kinase 3 3 2.6 PLK4 2 4 1.45 Axin 2 2 2 2.42 PTPRR 1 1 0.8

Actin 8 18 21.9 Cyclin 2 2 5.36

Axin 2 1 1 1.42 Axin 1 1 1 1.0 B-Raf 1 1 0.91

Adenylate kinase 1 2 2 8.76

enzyme E2M 1 1 4.91 Cyclin 1 1 3.7

Tropomyosine α 8 12 22.6 Actin 12 46 31 HSP 70K 3 4 5.6 PARD-3 2 14 1.56 Desmoplakin-3 3 3 1.0 Junction plakoglobin 2 2 3.4

kinase 2 8 1.0 Cullin 3 1 1 1.3

demonstrating Dvls assemble Dvls-based complexes and thus provide a Dvls platform. Polo-like kinase 4 (PLK4), Regulator of G protein signaling (RGS) 18, Rho-associated protein kinase and Receptor-type protein tyrosine phosphatase R (PTPRR) were identified as novel

PDZ domain beads pulled down interacting proteins Axin1 and Axin2. Novel interacting proteins were Tropomyosine α, cyclin, PKC and casein kinase substrate in neuron protein 1 and B-Raf proto-oncogen serine/threonine kinase. Adenylate kinase 1, cyclin and ubiquitinconjugating enzyme E2M were identified as novel DEP domain interacting proteins. SH3

DIX Dvl2 7 38 8

PDZ Tropomyosine α 7 9 23.4

DEP Actin 6 14 18.1
