Analysis of Secreted Proteins from Adipocytes: a Proteomics Approach

Analysis of secreted proteins from adipocytes: A proteomics approach

By Wei Tse Hsu

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science (Research)

School of Biotechnology and Biomolecular Sciences University of New South Wales

Supervisor: Professor Marc R. Wilkins Co-supervisor: Associate professor Mark Raftery Submitted: 2012

ORIGINALITY STATEMENT

ȼI hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.’

ȼI hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.'

AUTHENTICITY STATEMENT

ȼI certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.’

Acknowledgments

My thanks go to the many people at UNSW who helped and encouraged me during my time within their department. In particular Mark Raftery, my second supervisor, who provided an insightful view of my work. Also, Ben Herbert, Matt Padula, George Miklos and Graham Vesey, for their willing support in many aspects.

I also thank all the Marc Wilkins lab members. Without them the research would not have been possible, especially Ignatius Pang for all his advices on bioinformatic knowledge. I also thank the members of Marc Wilkins lab, especially Simone Li, Melissa Erce, Tim Couttas and Jason

Low for their guidance and suggestions.

I am particularly indebted to my family members in Taiwan and the best friend Dr. Jerry Wei in

University of Sydney and his family members for supporting and encouraging me to pursue this degree.

Finally, my biggest thanks go to my supervisor, Marc Wilkins, whose positive, informed, advices and encouraging nature has been an inspiration throughout.

Table of contents:

ABSTRACT ...... I LIST OF FIGURES ...... II LIST OF TABLES ...... V LIST OF ABBREVIATIONS ...... VI 1. INTRODUCTION ...... 1 1.1. ADIPOSE TISSUE AND OBESITY ...... 1 1.2. PROTEINS SECRETED FROM ADIPOCYTES OR ADIPOSE TISSUES ...... 2 1.3. EXTRACELLULAR AND INTRACELLULAR SIGNALING COMPONENTS OF ADIPOSE TISSUE OR ADIPOCYTES ...... 4 1.4. CLASSICAL SECRETORY AND NON-CLASSICAL SECRETORY PATHWAY ...... 4 1.5. DETECTION OF SECRETED PROTEINS...... 5 1.6 DETECTION OF PROTEINS USING MASS SPECTROMETRY ...... 6 1.7. QUANTITATIVE PROTEOME ANALYSIS ...... 8 1.7.1. Stable isotope labeling with amino acids in cell culture (SILAC) ...... 9 1.7.2. Isotope-coded affinity tags (ICAT) ...... 9 1.7.3. Isotope-coded amine-reactive reagents (iTRAQ) ...... 10 1.7.4. Selected or multiple reaction monitoring (SRM/MRM) technology ...... 11 1.7.5. The absolute quantification method (AQUA) ...... 12 1.7.6. Label-free quantitation ...... 15 1.8. PROTEOMICS AND CLINICAL DIAGNOSTICS ...... 15 1.9. HYPOTHESIS AND AIMS ...... 17 1.9.1. Hypothesis ...... 17 1.9.2. Aims ...... 18 2. MATERIALS AND METHODS ...... 19 2.1. SAMPLE PREPARATION ...... 19 2.1.1. Tissue harvest ...... 19 2.1.2. Cell isolation and culture ...... 19 2.2. PROTEIN IDENTIFICATION ...... 20 2.2.1. In solution digestion ...... 20 2.2.2. Nano-flow LC-MS ...... 21 2.3. ANALYSIS OF SECRETED PROTEINS ...... 22 2.3.1. Proteomic data analysis ...... 22 2.3.2. Identification of proteotypic peptides from secreted proteins ...... 23 2.3.3. Functional analysis of identification proteins ...... 23 3. RESULTS ...... 24 3.1. IDENTIFICATION OF PROTEINS FROM CONDITIONED MEDIA: CULTURED ADIPOCYTES 25 3.1.1. Subcellular localization of proteins ...... 27 3.1.2. Functional categorization of proteins ...... 29 3.1.3. Proteins from adipocytes culture: a summary ...... 34 3.2 IDENTIFICATION OF PROTEINS FROM CONDITIONED MEDIA: ADIPOSE TISSUE EXPLANTS . 38 3.2.1 Subcellular localization of proteins ...... 40 3.2.2. Functional categorization of proteins ...... 42 3.2.3 Proteins from tissue explants: a summary ...... 49 3.3. COMPARISON BETWEEN TWO TYPES OF CULTURE SYSTEMS ...... 52 3.4. QUANTITATIVE ANALYSIS OF CLASSICAL AND NON-CLASSICAL SECRETED PROTEINS 54 3.5. PROTEOTYPIC PEPTIDES DETECTION ...... 58 3.5.1. Characteristics of peptides identified in this study ...... 58

3.5.2. Theoretical proteotypic peptides ...... 60 3.5.3. Experimental proteotypic peptide detection ...... 63 3.5.4. Selected reaction monitoring (SRM) method prediction ...... 67 4. DISCUSSION ...... 71 4.1. GEL AND LABEL-FREE SECRETOMICS STRATEGIES IN SERUM-FREE ADIPOCYTES AND ADIPOSE EXPLANTS ...... 71 4.2. NEW PROTEINS DETECTED IN THIS STUDY ...... 80 4.3. CLINICAL APPLICATION OF OUR RESULTS ...... 83 5. REFERENCES ...... 86 6. APPENDICES ...... 94 6.1 THE PROTEIN EXPRESSION LEVELS OF ALL SECRETED AND NON-CLASSICAL SECRETED PROTEINS IN ADIPOCYTES CULTURE AND ADIPOSE TISSUE EXPLANTS...... 94 6.2 PERL SCRITP ...... 104 6.2.1 GO nodes ...... 104 6.2.2 GO edge ...... 105

Abstract

We have undertaken a systematic proteomic approach to identify classical and non-classical secreted proteins which are produced in adipocyte cultures and adipose tissue explants.

Serum-free primary culture was used, in association with tandem mass spectrometry, to characterize the secreted proteins during a time-course. In the 18 days of culture, a total of 281 proteins were found from adipocyte cultures or adipose tissue explants. Our method discovered

3 novel secreted proteins from adipocyte cultures: afamin (AFAM_RAT), seminal vesicle secretory protein 2 (SVS2_RAT) and xanthine dehydrogenase/oxidase (XDH_RAT). A further 9 novel secreted proteins were discovered from adipose tissue explants: chymase (CMA1_RAT), matrix Gla protein (MGP_RAT), biglycan (PGS1_RAT), serine protease inhibitor A3K

(SPA3K_RAT), serine protease inhibitor A3L (SPA3L_RAT), serine protease inhibitor A3M

(SPA3M_RAT), serine protease inhibitor A3N (SPA3N_RAT), SPARC-like protein 1

(SPRL1_RAT) and seminal vesicle secretory protein 2 (SVS2_RAT). Previous studies suggested that these proteins may play a role in tissue development in association with obesity or obesity-related conditions, giving them potential as biomarker of adipocyte-associated disease.

In this regard, we show how our data could be used to create a reliable SRM/MRM database that would be applicable to early diagnosis of adipocyte-associated disease.

~ I ~

List of Figures

FIGURE 1. COMPONENTS OF A FOURIER TRANSFORM MASS SPECTROMETER. THE LTQ-FT MASS SPECTROMETER IS AN INTEGRATED HYDRID MASS SPECTROMETER SYSTEM CONSISTING OF AN API ION SOURCE, LINEAR ION TRAP AND ION TRANSFER COMBINED WITH A FOURIER TRANSFORM ION CYCLOTRON RESONANCE AND SUPERCONDUCTING MAGNET [31]...... 7 FIGURE 2. SCHEMATIC DIAGRAM OF QUADRUPOLE ION TRAP AND FOURIER TRANSFORM ION CYCLOTRON RESONANCE[32]. (A) AT A GIVEN FIELD, IONS OF PROTEINS OR PEPTIDES ARE CONSTRAINED INTO STABLE TRAJECTORIES AND PASS DOWN THE QUADRUPOLE. ALL OTHER IONS DO NOT HAVE A STABLE TRAJECTORY THROUGH THE QUADRUPOLE MASS ANALYZER AND COLLIDE WITH THE QUADRUPOLE RODS OR ARE LOST BETWEEN THEM. (B) IONS OF PROTEINS OR PEPTIDES IN A FOURIER-TRANSFORM ION-CYCLOTRON RESONANCE (FT-ICR) OSCILLATE AROUND THE MAGNETIC FIELD AT FREQUENCIES THAT ARE RELATED TO THEIR M/Z SCALES. AS THE IONS OSCILLATE NEAR THE TOP AND BOTTOM METAL PLATES OF THE CUBIC TRAPPING CELL, THEY INDUCE AN ALTERNATING CURRENT THAT CAN BE MEASURED BY THE MASS ANALYZER AND THEN RELATED TO THEIR M/Z...... 8 FIGURE 3. PEPTIDE QUANTITATION USING SILAC. TWO POPULATIONS OF CELLS ARE GROWN IN THE SAME TYPE OF CULTURE MEDIUM. ONE OR MORE ESSENTIAL AMINO ACIDS ARE REPLACED BY HEAVY OR LIGHT ATOMS. THE “HEAVY” AND “LIGHT” SAMPLE ARE COMBINED AND LYSED. PROTEINS OF INTEREST ARE ANALYSED BY PROTEOLYTIC DIGESTION AND LC-MS/MS [37]...... 9 FIGURE 4. PEPTIDE QUANTITATION USING ICAT. SAMPLES ARE LABELED BY DIFFERENTIALLY ‘LIGHT’ OR ‘HEAVY’ ICAT TAG ON THE CYSTEINE RESIDUES OF PROTEINS. THE TWO PROTEIN GROUPS ARE THEN COMBINED, DIGESTED. THE ICAT-LABELED PEPTIDES ARE ENRICHED BY AFFINITY ISOLATION AND THOSE PEPTIDES ARE ANALYZED BY LC-MS/MS. THE IDENTIFICATION OF PROTEINS CAN BE OBTAINED BY DATABASE SEARCH. THE RATIOS OF THE ‘LIGHT’ AND ‘HEAVY’ ION PEAKS EXTRACTED FROM ION CHROMATOGRAMS CAN PROVIDE RELATIVE ABUNDANCE OF EACH IDENTIFIED PROTEIN [37]...... 10 FIGURE 5. PEPTIDE QUANTITATION USING ITRAQ. PROTEINS ARE EXTRACTED FROM DIFFERENT CONDITION OF BIOLOGY SAMPLES AND MIXED WELL. DIGESTED PROTEIN SAMPLES ARE LABELED WITH ISOBARIC TAGS (114, 115, 116 AND 117) ON THE N-TERMINUS OF ALL PEPTIDES. ALL TAGGED PEPTIDE GROUPS ARE MIXED TOGETHER, FOLLOWED BY LC-MS/MS ANALYSIS. AFTER FRAGMENTATION AND PEPTIDE IDENTIFICATION (RED RECTANGLE), THE PEPTIDE FRAGMENTS WITH ISOBARIC TAGS WILL DISPLAY AT THE LOW M/Z REGION OF THE MS/MS SPECTRUM AS THE REPORTER IONS (BLACK CIRCLE). THE RATIO OF REPORTER IONS EXTRACTED FROM THE MS/MS SPECTRUM CAN PROVIDE RELATIVE ABUNDANCE OF THE PEPTIDE IN EACH BIOLOGICAL SAMPLE [38]...... 11 FIGURE 6. PEPTIDE QUANTITATION USING LC-MS/MS IN THE SRM MODE. A: THE M/Z OF THE PRECURSOR IONS ARE SELECTED IN Q1, AND THE M/Z OF THE PRODUCT IONS ARE SELECTED IN Q3. B: SAMPLES ARE DIGESTED AND ISOTOPE-LABELED STANDARD PEPTIDES ARE ADDED. BOTH ENDOGENOUS AND ISOTOPE-LABELED PEPTIDES ARE THEN SELECTED FOR ANALYSIS USING A TRIPLE-QUADRUPOLE MASS SPECTROMETER. THE EXTRACTED ION CHROMATOGRAMS UNDER THE AUC CURVE ARE THEN USED FOR ABSOLUTE QUANTITATION OF THE PEPTIDE [39]...... 12 FIGURE 7. ABSOLUTE QUANTIFICATION OF PROTEINS STRATEGY. [40, 41] ABSOLUTE QUANTIFICATION INVOLVES TWO STAGES. STAGE 1 INVOLVES THE SELECTION AND STANDARD SYNTHESIS OF A PEPTIDE FROM THE PROTEIN OF INTEREST. DURING SYNTHESIS, A SINGLE AMINO ACID RESIDUE IS INCORPORATE WITH STABLE ISOTOPES. THESE PEPTIDE INTERNAL STANDARDS ARE ANALYZED BY MS/MS TO EXAMINE PEPTIDE FRAGMENTATION PATTERNS. THE NEXT STEP IS TO SET UP A SRM ANALYSIS. PROTEINS OF INTEREST ARE PROTEOLYZED AND MIXED WITH THE INTERNAL STANDARD PEPTIDE. AN LC–SRM ~ II ~

EXPERIMENT MEASURES THE ABUNDANCE OF A SPECIFIC FRAGMENT ION FROM BOTH THE NATIVE SAMPLE (ENDOGENOUS PEPTIDE) AND THE SYNTHESIZED SAMPLE (AQUA PEPTIDE) AS A FUNCTION OF REVERSE-PHASE CHROMATOGRAPHIC RETENTION TIME. THE ABSOLUTE QUANTIFICATION IS DETERMINED BY COMPARING THE ABUNDANCE OF THE KNOWN AQUA INTERNAL STANDARD PEPTIDE WITH THE NATIVE PEPTIDE (ENDOGENOUS PEPTIDE/AQUA PEPTIDE)...... 14 FIGURE 8. EXPERIMENTAL STRATEGY FOR INCUBATION AND HARVEST. ISOLATED ADIPOCYTES AND ADIPOSE TISSUES WERE CULTURED SEPARATELY IN SERUM-FREE DMEM MEDIUM AND EACH CONDITIONED MEDIUM SAMPLED AT 37°C WITH 5% CARBON DIOXIDE EVERY 24 HOURS. 1ML OF CONDITIONED CULTURE MEDIUM WAS COLLECTED EVERY 24 H FROM CELL CULTURES AND TISSUE EXPLANTS AND REPLACED WITH 1 ML FRESH DMEM F12 TO MAINTAIN A FINAL MEDIUM VOLUME OF 10 ML...... 24 FIGURE 9. SUBCELLULAR LOCATION CATEGORIES OF IDENTIFICATION PROTEINS FROM PRIMARY ADIPOCYTE CULTURE MEDIA. SUBCELLULAR LOCATION WERE DISTRIBUTED INTO THE FOLLOWING CATEGORIES: 38% FOR SECRETED OR EXTRACELLULAR MATRIX, 23% FOR CYTOPLASM AND 9% AS NUCLEUS PROTEINS, 2% FOR MITOCHONDRIA, 4% FOR ENDOPLASMIC RETICULUM (4%), 9% FOR CELL MEMBRANE AND 18% FOR UNKNOWN LOCATION...... 28 FIGURE 10. FUNCTIONAL CATEGORIES OF IDENTIFIED PROTEINS FROM PRIMARY ADIPOCYTE CULTURE MEDIA. THE DESCRIPTIVE TERMS WERE ACCORDING WITH GENE ONTOLOGY. 36% FOR SIGNAL PATHWAY, 32% FOR METABOLIC OR BIOSYNTHETIC PROCESS. 20% FOR CELL CYCLE INCLUDING CELL GROWTH, PROLIFERATION AND DIFFERENTIATION, 3% AND 9 % FOR IMMUNE SYSTEM AND OTHERS, RESPECTIVELY. ... 30 FIGURE 11. GENE ONTOLOGY ANALYSIS AND NETWORK REPRESENTATION OF PROTEINS FROM CULTURED ADIPOCYTE MEDIUM. CYTOSCAPE (VERSION 2.7) AND BINGO WERE USED TO PERFORM GENE ONTOLOGY ASSIGNMENTS AND DETERMINE SIGNIFICANTLY OVER-REPRESENTED GO CATEGORIES. THE FORCE DIRECTED LAYOUT WAS USED TO CALCULATE THIS LAYOUT (FORCE STRENGTHS: SPRING = 50, ORIGIN = 80, REPLULSION = 12)...... 31 FIGURE 12. SUBCELLULAR LOCATION CATEGORIES OF IDENTIFICATION PROTEINS FROM THE CULTURED MEDIAM OF ADIPOSE TISSUE EXPLANTS. SUBCELLULAR LOCATION WERE DISTRIBUTED INTO THE FOLLOWING CATEGORIES: 34% FOR SECRETED OR EXTRACELLULAR MATRIX, 31% FOR CYTOPLASM AND 6% AS NUCLEUS PROTEINS, 2% FOR MITOCHONDRIA, 7% FOR ENDOPLASMIC RETICULUM, 3% FOR CELL MEMBRANE AND 17% FOR UNKNOWN LOCATION...... 41 FIGURE 13. FUNCTIONAL CATEGORIES OF IDENTIFIED PROTEINS FROM THE CULTURE MEDIUM OF ADIPOSE TISSUE EXPLANTS. THE DESCRIPTIVE TERMS WERE ACCORDING WITH GENE ONTOLOGY. 38% FOR SIGNAL PATHWAY, 31% FOR METABOLIC OR BIOSYNTHETIC PROCESS. 23% FOR CELL CYCLE INCLUDING CELL GROWTH, PROLIFERATION AND DIFFERENTIATION, 3% AND 5 % FOR IMMUNE SYSTEM AND OTHERS, RESPECTIVELY...... 45 FIGURE 14. GENE ONTOLOGY ANALYSIS AND NETWORK REPRESENTATION OF PROTEINS FROM ADIPOSE TISSUE EXPLANTS MEDIUM. CYTOSCAPE (VERSION 2.7) AND BINGO WERE USED TO PERFORM GENE ONTOLOGY ASSIGNMENTS AND DETERMINE SIGNIFICANTLY OVER-REPRESENTED GO CATEGORIES. THE FORCE DIRECTED LAYOUT WAS USED TO CALCULATE THIS LAYOUT (FORCE STRENGTHS: SPRING = 50, ORIGIN = 80, REPLULSION = 12)...... 46 FIGURE 15. THE NUMBER OF IDENTIFIED PROTEINS COMPARED WITH ADIPOCYTES CULTURE AND TISSUE EXPLANTS. A TOTAL 281 PROTEINS WERE FOUND FROM ADIPOCYTES CULTURE OR ADIPOSE TISSUE EXPLANTS SYSTEM. TWENTY-NINE PROTEINS WERE SEEN ONLY IN THE MEDIA OF CULTURED ADIPOCYTES AND 156 PROTEINS ONLY IN ADIPOSE TISSUE EXPLANTS. NINETY-SIX SECRETED PROTEINS WERE FOUND IN ADIPOCYTE CULTURE AND ALSO IN ADIPOSE TISSUE EXPLANTS...... 53 FIGURE 16. EXAMPLE OF TYPE A PROTEIN EXPRESSION. THE PROTEIN QUANTIZATION WAS DECREASING DURING CULTURE. A) PROTEIN WAS DETECTED IN ADIPOCYTE CULTURE. B)&C) ~ III ~

PROTEINS WERE MEASURED IN ADIPOSE EXPLANTS...... 55 FIGURE 17. EXAMPLE OF TYPE B PROTEIN EXPRESSION. THE PROTEIN QUANTIZATION WAS INCREASING UNTIL REACH A PEAK THEN TREND GOING DOWN AT THE END OF CULTURE TIME. A) PROTEIN WAS DETECTED IN ADIPOCYTE CULTURE. B) PROTEIN WAS MEASURED IN ADIPOSE EXPLANTS...... 55 FIGURE 18. EXAMPLE OF TYPE C PROTEIN EXPRESSION. THE LEVELS OF PROTEIN QUANTIZATION WERE OBTAINED ONLY AT SPECIAL TIME POINTS DURING CULTURE. A) PROTEIN WAS DETECTED IN ADIPOCYTE CULTURE. B) PROTEIN WAS MEASURED IN ADIPOSE EXPLANTS...... 56 FIGURE 19. EXAMPLE OF TYPE D PROTEIN EXPRESSION. THE LEVELS OF PROTEIN QUANTIZATION WERE INCREASED AND DECREASED AS CYCLICAL DURING CULTURE. A) PROTEINS WERE DETECTED IN ADIPOCYTE CULTURE. B) PROTEINS WERE MEASURED IN ADIPOSE EXPLANTS...... 56 FIGURE 20. EXAMPLE OF TYPE E PROTEIN EXPRESSION. THE LEVELS OF PROTEIN QUANTIZATION WERE INCREASED DURING CULTURE. A) PROTEINS WERE DETECTED IN ADIPOCYTE CULTURE. B) PROTEINS WERE MEASURED IN ADIPOSE EXPLANTS...... 57 FIGURE 21. THE PROTEIN EXPRESSION OF LEPTIN AND ADIPONECTIN IN ADIPOCYTE CULTURE AND ADIPOSE TISSUE EXPLANT. A) ADIPONECTIN WAS PRESENT BUT LEPTIN WAS NOT DETECTED IN ADIPOCYTE CULTURE SYSTEMS. B) LEPTIN EXPRESSED AS TYPE C AND ADIPONECTIN EXPRESSED AS TYPE B IN ADIPOSE TISSUE EXPLANTS...... 58 FIGURE 22. PROPERTIES OF IDENTIFIED MS/MS SPECTRA. MASS AND CHARGE DISTRIBUTIONS OF THE IDENTIFIED PEPTIDES. AROUND 69% OF THE TRYPTIC PEPTIDE PRECURSORS WERE DOUBLY CHARGED, AND 28% WERE TRIPLY CHARGED. ONLY 3% WERE QUADRUPLY OR HIGHER CHARGED. ANY SINGLY CHARGED IONS WERE EXCLUDED ...... 59 FIGURE 23. DISTRIBUTION OF IDENTIFIED PEPTIDE BY MASS. ABOUT 5% OF PEPTIDES WERE BELOW 800DA. 18% OF PEPTIDES MASS WERE 1100 OR 1200 DA AND 70 % OF PEPTIDES WERE BETWEEN 900 TO 1700 DA. 7% OF PEPTIDE MASS BETWEEN 1800 TO 1900 WERE DETECTABLE...... 59 FIGURE 24. DISTRIBUTION OF THE NUMBER OF AMINO ACID COMPONENTS OF PROTEOTYPIC PEPTIDES. THE NUMBER OF AMINO ACID WERE OBTAINED FROM THEORETICAL AND EXPERIMENTAL PROTEOTYPIC PEPTIDE ANALYSIS. IT IS ALSO ANALYSIS THE MATCHED PEPTIDES OF AMINO ACID COMPONENTS FROM THEORETICAL AND EXPERIMENTAL PROCESS...... 61 FIGURE 25. AN EXAMPLE OF PROTEOTYPIC PEPTIDES OF LEPTIN AND ADIPONECTIN. THE RED, BLUE AND GREEN SEQUENCES WERE THE NUMBER OF PEPTIDES OBSERVED DURING A PROTEOMICS EXPERIMENT. THE RED PEPTIDE SEQUENCES WERE ONLY DETECTED ONCE BUT THE BLUE, GREEN AND BROWN SEQUENCES WERE OBSERVED MORE THAN ONCE DURING THE EXPERIMENTAL PROCESS...... 64 FIGURE 26. THE MASS SPECTRUM OF ADIPONECTIN AND LEPTIN. A) THE FRAGMENTATION PATTERN FOR THE AMINO ACID SEQUENCE VTVPNVPIR OF ADIPONECTIN WITH THE CORRESPONDING B AND Y ION SERIES. B) THE FRAGMENTATION PATTERN FOR THE AMINO ACID SEQUENCE AVLFTYDQYQEK OF ADIPONECTIN WITH THE CORRESPONDING B AND Y ION SERIES.C) THE FRAGMENTATION PATTERN FOR THE AMINO ACID SEQUENCE IFYNQQNHYDGSTGK OF ADIPONECTIN WITH THE CORRESPONDING B AND Y ION SERIES...... 69 FIGURE 27. THE MASS SPECTRUM OF ADIPONECTIN AND LEPTIN. D) THE FRAGMENTATION PATTERN FOR THE AMINO ACID SEQUENCE INDISHTQSVSAR OF LEPTIN WITH THE CORRESPONDING B AND Y ION SERIES. E) THE FRAGMENTATION PATTERN FOR THE AMINO ACID SEQUENCE VTGLDFIPGLHPILSLSK OF LEPTIN WITH THE CORRESPONDING B AND Y ION SERIES...... 70 FIGURE 28. THE PROTEIN EXPRESSION OF INSULIN-LIKE GROWTH FACTOR-BINDING PROTEIN 3 AND ADIPONECTIN. WHILE THE LEVEL OF INSULIN-LIKE GROWTH FACTOR-BINDING PROTEIN 3 SLIGHT INCREASED FROM DAY 9, THE LEVEL OF ADIPONECTIN WAS SLIGHT DECREASING FROM DAY 9 TO DAY 11...... 78

~ IV ~

List of Tables

TABLE 1. THE FALSE DISCOVERY RATE (FDR) FOR SECRETED PROTEINS IDENTIFIED FROM PRIMARY ADIPOCYTE CULTURES OVER AN 18 DAY TIME COURSE. FALSE DISCOVERY RATE = FALSE POSITIVE / FALSE POSITIVE + TRUE POSITIVE * 100...... 26 TABLE 2. THE LIST OF INDIVIDUAL NODES...... 33 TABLE 3. PROTEIN FROM CULTURED ADIPOCYTES. THE RESULTS FROM SECRETION PROTEIN PREDICTION SOFTWARE AND FUNCTIONAL KNOWLEDGE FROM THE SWISS-PROT DATABASE IS ALSO SHOWN, TO INDICATE THE LIKELIHOOD OF PROTEINS BEING ACTIVELY SECRETED. THOSE NOT REPORTED PREVIOUSLY TO BE SECRETED FROM ADIPOCYTES BUT APPEAR SECRETED BY CLASSICAL MEANS, ARE AFAMIN, SEMINAL VESICLE SECRETORY PROTEIN 2 AND XANTHINE DEHYDROGENASE/OXIDASE (SHOWN IN BOLD)...... 35 TABLE 4. THE FALSE DISCOVERY RATE (FDR) FOR SECRETED PROTEINS IDENTIFIED FROM ADIPOSE EXPLANTS CULTURES OVER AN 18 DAY TIME COURSE. FALSE DISCOVERY RATE = FALSE POSITIVE / FALSE POSITIVE + TRUE POSITIVE * 100...... 39 TABLE 5. THE LIST OF INDIVIDUAL NODES...... 47 TABLE 6. NOVEL SECRETED PROTEINS IN THE CONTEXT FROM ADIPOSE TISSUE EXPLANTS. SECRETION PROTEIN PREDICTION SOFTWARE AND KNOWN FUNCTIONAL KNOWLEDGE FROM SWISS-PROT DATABASE SUGGEST THESE ARE SECRETED BY CLASSICAL MEANS. THESE HAVE NOT BEEN REPORTED PREVIOUSLY IN ASSOCIATION WITH ADIPOCYTES. .... 51 TABLE 7. THE EXAMPLE OF THEORETICAL PROTEOTYPIC PEPTIDES OF LEPTIN. LEPIN IS A 19KDA, 167 AMINO ACID SEQUENCE. DIGESTION WITH TRYPSIN WITH CLEAVAGE SITES OF K AND R. THE THRESHOLD SCORES WAS GREATER THAN 0.95 AND EXPERIMENT TYPE WAS PAGE WITH MALDI...... 62 TABLE 8. AN EXAMPLE OF PROTEOTYPIC PEPTIDES. LEPTIN WAS DETECTED BY LTQ-FT ULTRA INSTRUMENT. THE PROTEIN IDENTIFICATION PROBABILITY WAS MORE THAN 99% AND THE HIGHEST SEQUENCE COVERAGE WAS 34.1%. THE ‘GLQKPESLDGVLEASLYSTEVVALSR’ AND ‘INDISHTQSVSAR’ SEQUENCES WERE THE MOST COMMON SEQUENCES IN THE DETECTABLE PEPTIDES FROM FT-ICR MS/MS AT TISSUE EXPLANTS SYSTEM ONLY...... 65 TABLE 9. AN EXAMPLE OF PROTEOTYPIC PEPTIDES. ADIPONECTIN WAS DETECTED BY LTQ-FT ULTRA INSTRUMENT. THE PROTEIN IDENTIFICATION PROBABILITY WAS MORE THAN 99% AND THE HIGHEST SEQUENCE COVERAGE WAS 14.6%. THE ‘AVLFTYDQYQEK’,’ IFYNQQNHYDGSTGK’ AND ‘VTVPNVPIR’ SEQUENCES WERE THE MOST COMMON SEQUENCES IN THE DETECTABLE PEPTIDES FROM FT-ICR MS/MS AND TWO CULTURE SYSTEMS...... 65 TABLE 10. THE MATCHED PEPTIDES OF LEPTIN FROM THEORETICAL AND EXPERIMENTAL PEPTIDES ANALYSIS. ‘INDISHTQSVSARQ’ AND ‘VTGLDFIPGLHPILSLSKM’ ARE IN ACCORDANCE WITH THE ALL PEPTIDE ANALYSIS...... 67 TABLE 11. THE LIST OF PRODUCTION IONS OF LEPTIN AND ADIPONECTIN. THE TWO GROUPS OF PEPTIDES WERE IN ACCORDANCE WITH EXPERIMENTAL PROTEOTYPIC PEPTIDES. THE PRODUCTION IONS (Y-IONS) WERE OBTAINED FROM SKYLINE SOFTWARE...... 68

~ V ~

List of Abbreviations

AQUA: absolute quantification method, BMI: body mass index, CID: collision induced dissociation, DMEM: Dulbecco's modified Eagle's /Ham's medium, ECM :extracellular matrix,

ELISA: enzyme-linked immunosorbent assay, ER: endoplasmic reticulum, ESI: Electrospray ionization, FDR: False Discovery Rate, FGF: fibroblast growth factor, FT-ICR: linear ion trap coupled to a Fourier transform ion cyclotron resonance, GO: Gene Ontology, HSP70: heat shock protein 70, HSP90: heat shock protein, ICAT: Isotope-Coded Affinity Tag, IL: interleukin, iTRAQ: amine-reactive isobaric Tags, JAK-STAT: Janus kinase/signal transducers and activators of transcription pathways, KL: Kullbach-Leibler, KS: Kolmogorov-Smirnov distance,

LC-MS/MS: Liquid chromatography-mass spectrometry, LDL: low density lipoproteins, m/z: mass-to-charge, MALDI: Matrix-assisted laser desorption/ionization, MAPKKK:

Mitogen-Activated Protein Kinase (MAPK) pathways, MS/MS: tandem mass spectrum, MS1: first mass analyzer, MS2: second mass analyzer, MSCs: mesenchymal stem cells, MUDPIT: multidimensional protein identification technology, NEFAs: Nonesterified fatty acids, PAI-1: plasminogen activator inhibitor 1, PBS: phosphate- buffered saline, PPAR γ: peroxisome proliferative activated receptor gamma, Q1: 1st quadrupole, SDS-PAGE: sodium dodecyl sulfate polyacrylamide gel electrophoresis, SELDI-TOF: Surface enhanced laser desorption and ionization time-of-flight mass spectrometry, SILAC: stable isotope labeling with amino acids in cell culture, SP: signal peptide, SPARC: Secreted Protein, Acidic and Rich in Cysteine,

SRM/MRM: Selected or multiple reaction monitoring, TCA: trichloroacetic acid, TNF-α: tumor necrosis factor-α, WAT: white adipose tissue

~ VI ~

Introduction

1. Introduction

1.1. Adipose tissue and obesity

Adipose tissue is a pivotal endocrine organ. It is an innervated tissue and contributes to the production of hormones, growth factors, adipokines, and other biomolecules. It is the primary site for the production of protein factors or signals peptides which are involved in energy balance, lipid uptake and transport, immune response and inflammation, cell differentiation and proliferation, and remodeling of the extracellular matrix. However, other cell types including macrophages, endothelial cells, fibroblasts and leucocytes are a source of secreted factors that also exert action on fatty tissue [1, 2].

Overweight and obesity are global problems that are on the increase in both developed and developing countries. In 2006, the World Health Organization (WHO) predicted that by 2015, approximately 2.3 billion adults will be overweight and more than 700 million will be obese [3].

Obesity is defined by an increased mass of white adipose tissue and is associated with cardiovascular disease, type 2 diabetes and several cancers [4]. The increased mass of white adipose tissue (WAT) is caused by an increase in size of adipocytes (hypertrophy) and increase the number of adipocytes (hyperplasia).

The hypertrophy and hyperplasia of adipocytes has been extensively studied. It is generally described as a two-step process [5]. The first step involves in the generation of committed pre-adipocytes from mesenchymal stem cells (MSCs). The second step involves the terminal differentiation of the pre-adipocytes into mature functional adipocytes. During the process, various transcription factors, extracellular matrix (ECM) components and hormones are influenced by each other. Adipokines, being cell-cell signaling proteins produced or secreted by adipocytes or adipose tissue [6], are central to this. Adiponectin and leptin are two well-known adipokins and correlate with obesity and adipocyte size and number. On the other hand,

~ 1 ~

Introduction

peroxisome proliferative activated receptor (PPAR) gamma is a transcription factor that can affect leptin and adiponectin production and secretion [7]. However, the molecular components that mediate the traffic of the adipokines are still poorly understood.

1.2. Proteins secreted from adipocytes or adipose tissues

The term “secretome” is referred to as the complex set of molecules, peptide or proteins secreted from living cells, tissue or organism through different secretion mechanisms [8].

Secreted proteins constitute an important class of molecules in cell signaling, communication and migration. They also play crucial roles in physiological or pathological processes including immune defence, blood coagulation, lipid, cholesterol metabolism and cancer angiogenesis, differentiation, invasion and metastasis [9]. Recent studies suggested that TNF-α in human adipose tissue serves as an important regulator of adipocyte sizes and numbers in healthy subjects [10]. Additional well-characterized factors known to play vital role in the regulation of body energy balance is the hormone leptin [11]. Circulating leptin levels are linked to adipocytes size. For example, large adipocytes produce more leptin than small adipocytes.

Secreted leptin can bind to its receptor in the hypothalamus to decrease food intake and increase energy expenditure. Despite this, in the obese state, leptin availability to the brain decreases at high leptin concentrations due to a decrease in leptin transport across the blood-brain barrier. Its ability to initiate cell signaling within the brain also decreases [12].

Other proteins or peptides are secreted by non-classical secretion pathways. These include the fibroblast growth factor (FGF) 1 and FGF2, interleukin (IL) 1α and IL1β, macrophage migration inhibitory factor, morphogen epimorphin, HSP70 and HSP90, S100 proteins, sphingosine kinase 1, thioredoxin, chromatin-associated proteins HMBG1 and Engrailed 2, secretory transglutaminase, annexins I and II, galectins [13].

~ 2 ~

Introduction

Most of these proteins are induced by cellular stress conditions such as heat shock, hypoxia, serum starvation and oxidized low density lipoproteins (LDL) treatment; they are found in numerous cell and tissue types with various functions ascribed to them [13, 14]. As these proteins are widespread, it is not always straightforward to determine the specific contribution of white adipose tissue (WAT) to their circulating levels or the manner they may increase in association with adipocyte sizes and numbers.

Approximately 10% of the human genome is predicted to encode secreted proteins [8]. Several studies have reported the use of genomics strategies to analyze whole-genome gene expression from obese or lean subjects that encode for secreted proteins [2]. But, only 21% and 36% of encoded secreted proteins were found from human and mouse as adipose tissue-specific genes, respectively [4]. Using alternate methods to characterize the novel adipocyte or adipose tissue secretome will compliment these studies and provide additional information regarding the endocrine function of adipose tissue in different species.

In a previous review [15], it was noted that most proteomic studies in obesity or obesity induced diseases were focused on the study of many organs in the body with special emphasis on adipose tissue and adipocyte primary cultures or cell lines. Useful protein databases were made from liver, muscle, WAT and pancreatic islets of lean mice or 3T3-L1 adipocyte cell line models.

The protein profiling in obesity models has been done in ob/ob knock-out mice. Proteomics of adipogenesis has been studied in 3T3-L1 adipocytes, in adipocyte derived from human mesenchymal stem cell line (hMSC) and in adipocyte derived from subcutaneous human adipose tissue [16].

Numerous studies have analysed the adipocyte secretome during adipogenesis or in different culture conditions using SDS-PAGE coupled to LC-MS/MS, SELDI-TOF, isotopic labeling and human cytokine antibody array [17]. It is clear that the physiological roles of secreted proteins from adipocytes or adipose tissue were signaling, extracellular matrix, immune function and ~ 3 ~

Introduction

degradation. But these identified factors cannot be fully understood without examining the complex interactions. A recent publication reported that disulfide bonds of resistin and adiponectin are of critical important for redox potential of secretory pathways [4]. Furthermore, little was known about a potential peripheral interaction of insulin, leptin, adiponectin and other secreted protein in adipose tissue or adipocytes [8, 18, 19]. The secretion of extracellular matrix components is actively regulated, not only during adipocyte differentiation, but also in the mature adipocyte by stimuli such as insulin and rosiglitazone (a PPAR gamma inhibitor) [16,

20].

1.3. Extracellular and intracellular signaling components of adipose

tissue or adipocytes

During the last decade there have been numerous reports indicating that fat tissue is an endocrine organ. It produces hormones, growth factors, adipokines and other molecules that may affect normal organ development, metastasis, growth or progression in the microenvironment, the insoluble extracellular matrix (ECM) or cell-cell interactions [21].

Specific interactions occur between a variety of extracellular proteins and fat tissues. Numerous biological processes are affected and appear to be involved in different physiological and pathophysiological conditions, such as cell development, immune reactions, and neoplastic transformation and metastasis. For example, galectin-3 is a high abundance, multifunctional protein that is intracellular and extracellular. It can be involved in the intercellular regulation of

Wnt signaling, cell differentiation, proliferation and apoptosis. The extracellular functions of galectin-3 include the modulation of cell adhesion or effects on immune or cancer cells [14].

1.4. Classical secretory and non-classical secretory pathway

Proteins trafficking and localization are essential in all living organisms. Secretion, one aspect of trafficking, involves the movement of proteins to the cell surface via the classical, non-classical or exosomal pathways. Classical protein secretion typically requires proteins to ~ 4 ~

Introduction

have a N-terminal hydrophobic signal peptide (SP) to traverse the endoplasmic reticulum (ER) and Golgi apparatus. Recent studies [22], however, have showed that some proteins lack a signal peptide or ER/Golgi-dependent post-translational modifications but can be exported to the extracellular space. This secretion pathway is defined as non-classical or unconventional.

Three distinct types of non-classical export can be distinguished. Firstly, protein export involves migration into intracellular vesicles, which are probably endosomal subcompartments. Secondly, proteins can reach the extracellular space by direct translocation across the plasma membrane or by a membrane flip-flop mechanism. Thirdly, the non-classical protein secretion pathway is known as membrane blebbing, which involves bulges on the outer surface of the cell [13, 22,

23].

Many software tools have been developed for predicting secreted proteins via the two different secretory pathways. SignalP 3.0 [24], SecretomeP 2.0 [23, 25] and TMHMM [26] were chosen in this study. SignalP 3.0 use Neural Networks (NN) and Hidden Markov Models (HMM) to predict classical secreted proteins depending on the analysis of protein N-terminal signal sequences [27]. SecretomeP 2.0 is based on known information such as protein functional roles or Gene Ontology classes to predict non-classical secretory pathways [22, 24]. TMHMM can be used to predict transmembrane domains by analyzing full-length protein sequences. Presence or absence of transmembrane domains can be used to increase the accuracy of predictions.

1.5. Detection of secreted proteins

DNA microarray and enzyme-linked immunosorbent assays are commonly used to investigate complex diseases. DNA microarrays permit a comprehensive understanding of the multiple genes involved in the mechanisms behind both physiologic and pathologic conditions [28].

However, the existing challenges of polymorphism detection, alternative mRNA splicing remain for this tool. The high-sensitive enzyme-linked immunosorbent assay (ELISA) and Western blot

~ 5 ~

Introduction

are widely used to monitor protein expression or concentrations [29] but this methodology is time consuming to develop and requires suitably species-specific antibodies.

1.6 Detection of proteins using mass spectrometry

Mass spectrometry-based proteomics has become an important and useful tool for identification and characterization of proteins. The mass-spectrometry-based proteomics experiment involves several steps [30]. Proteins are enzymatically digested to peptide mixtures and these peptides are then introduced into a high-performance liquid chromatography system. After separation, they are eluted into an ion source though a fused silica column or needle carrying an electrical potential. This causes the solution to spray. The spray contains droplets that encompass the sample and can be desolvated by applying heat; this generates positive or negative ions.

The LTQ-FT is a hybrid mass spectrometer with a linear ion trap coupled to a Fourier transform ion cyclotron resonance (FT-ICR) analyzer (figure 1). It has sufficient resolving power and mass accuracy to efficiently analyze protein and peptide ions.

~ 6 ~

Introduction

Figure 1. Components of a Fourier transform mass spectrometer. The LTQ-FT mass spectrometer is an integrated hydrid mass spectrometer system consisting of an API ion source, linear ion trap and ion transfer combined with a Fourier transform ion cyclotron resonance and superconducting magnet [31].

The quadrupole ion trap mass analyzer (figure 2a) is like a "mass filter" and the mass-to-charge (m/z) ratios of the peptide fragments are measured. At a given electric field, ions of proteins or peptides are constrained into stable trajectories and pass down the quadrupole. All other ions do not have a stable trajectory through the quadrupole mass analyzer and collide with the quadrupole rods or are lost between them.

Within the Fourier transform (FT) mass spectrometer, ions contained within a strong magnetic field will describe a circular motion in ion cyclotron resonance (ICR) (figure 2b); these 2 combined may be used to perform tandem mass spectrometry (MS/MS). As the ions oscillate near the top and bottom metal plates of the cell, they induce an alternating current that can be measured. In addition, Fourier transform can convert the time domain signal into a frequency domain signal which corresponds to m/z ratio.

~ 7 ~

Introduction

So with an LTQ-FT, it is possible to fragment large peptides or even protein ions in the ion trap and detect them with high resolution and accuracy.

Figure 2. Schematic diagram of quadrupole ion trap and Fourier transform ion cyclotron resonance[32]. (a) At a given field, ions of proteins or peptides are constrained into stable trajectories and pass down the quadrupole. All other ions do not have a stable trajectory through the quadrupole mass analyzer and collide with the quadrupole rods or are lost between them. (b) Ions of proteins or peptides in a Fourier-transform ion-cyclotron resonance (FT-ICR) oscillate around the magnetic field at frequencies that are related to their m/z scales. As the ions oscillate near the top and bottom metal plates of the cubic trapping cell, they induce an alternating current that can be measured by the mass analyzer and then related to their m/z.

Finally, the mass spectra are matched against computer-predicted spectra from the theoretical digestion of proteins in databases, resulting in peptide or protein identification. An example of a computer program used for this is Mascot.

1.7. Quantitative proteome analysis

Quantitative proteome analysis plays an important role in discovering biomarkers and their changes in concentration for disease treatment and diagnosis [33-38]. Quantitative proteomic strategies have been developed to quantify protein abundances and associated changes in a relative or absolute fashion. Relative protein quantitation can be achieved by incorporating stable isotope labeling into the samples to be analyzed. Approaches include stable isotope labeling with amino acids in cell culture (SILAC), isotope-coded affinity tags (ICAT) and amine-reactive isobaric Tags (iTRAQ).

~ 8 ~

Introduction

1.7.1. Stable isotope labeling with amino acids in cell culture (SILAC)

The SILAC strategy involves metabolic incorporation of isotopically heavy or light amino acids into proteins, as shown in figure 3 [37]. Two populations of cells are grown in the culture medium. One or more essential amino acids are replaced by those containing 13C, 2H, 15N or 18O labeled heavy or light atoms. The “heavy” and “light” sample are combined and lysed. Purified proteins of interest are submitted to proteolytic digestion and LC-MS/MS analysis.

Figure 3. Peptide quantitation using SILAC. Two populations of cells are grown in the same type of culture medium. One or more essential amino acids are replaced by heavy or light atoms. The “heavy” and “light” sample are combined and lysed. Proteins of interest are analysed by proteolytic digestion and LC-MS/MS [37].

1.7.2. Isotope-coded affinity tags (ICAT)

ICAT is a method to covalently bind cysteine residues of proteins or peptides with ‘light’ and

‘heavy’ ICAT tags and then enrich ICAT tag-containing peptides by affinity isolation. After

LC-MS/MS analysis and pair-wise comparison of the two proteome sample, the relative protein abundance can be calculated [37]. Figure 4 is indicates that samples from tissue, fluid or cell lines are labeled by differentially ‘light’ or ‘heavy’ ICAT tag in the cysteine residues of proteins.

The two protein samples are then combined and digested. The ICAT-labeled peptides are enriched by affinity isolation and those peptides are analyzed by LC-MS/MS. The identification of proteins can be obtained by database search. The ratios of the ‘light’ and ‘heavy’ ion peaks extracted from ion chromatograms can provide relative abundance of each identified protein.

~ 9 ~

Introduction

1.7.3. Isotope-coded amine-reactive reagents (iTRAQ) iTRAQ is also a kind of chemical modification labeling means for relative and absolute protein quantitation of samples. The character of iTRAQ is isobaric multiplexing tagging for the analysis of multiple samples. The figure 5 [38] shows that proteins are extracted from different biological samples. Digested protein samples are labeled with isobaric tags (114, 115, 116 and

117) at free amines on the N-terminus and lysine residues of all peptides. All tagged peptide groups are mixed together and followed by LC-MS/MS analysis. After fragmentation, the peptide fragments with isobaric tags are displayed at the low m/z region of the MS/MS spectrum as reporter ions. The ratio of reporter ions extracted from the MS/MS spectrum can provide relative abundance of the peptide in each biological sample.

Figure 4. Peptide quantitation using ICAT. Samples are labeled by differentially ‘light’ or ‘heavy’ ICAT tag on the cysteine residues of proteins. The two protein groups are then combined, digested. The ICAT-labeled peptides are enriched by affinity isolation and those peptides are analyzed by LC-MS/MS. The identification of proteins can be obtained by database search. The ratios of the ‘light’ and ‘heavy’ ion peaks extracted from ion chromatograms can provide relative abundance of each identified protein [37].

~ 10 ~

Introduction

Figure 5. Peptide quantitation using iTRAQ. Proteins are extracted from different condition of biology samples and mixed well. Digested protein samples are labeled with isobaric tags (114, 115, 116 and 117) on the N-terminus of all peptides. All tagged peptide groups are mixed together, followed by LC-MS/MS analysis. After fragmentation and peptide identification (red rectangle), the peptide fragments with isobaric tags will display at the low m/z region of the MS/MS spectrum as the reporter ions (black circle). The ratio of reporter ions extracted from the MS/MS spectrum can provide relative abundance of the peptide in each biological sample [38].

1.7.4. Selected or multiple reaction monitoring (SRM/MRM) technology

SRM/MRM are assays that use mass spectrometers with two mass filters or a triple-quadrupole analyzer (figure 6a) [39]. This allows for the peptide analysis to be selectively detected.

Samples from tissue, fluid or cell lines are digested and known amounts of isotope-labeled standard peptides are added. The precursor ion of unique peptide sequence is selected in the first quadrupole (Q1). Following fragmentation of the selected peptide ion by collision induced dissociation (CID) in Q2, the product ion is detected in the second phase, third quadrupole (Q3).

Quantization of proteins using this method is based on the detection of one or more peptides from target protein and known amounts of isotope-labeled standard peptide under the AUC curve extracted from ion chromatograms for each transition monitored (figure 6b) [39].

Additionally, hundreds of peptides can be monitored on a chromatographic time scale in a single

~ 11 ~

Introduction

mass run so SRM measurements are easily multiplexed (MRM).

Figure 6. Peptide quantitation using LC-MS/MS in the SRM mode. a: The m/z of the precursor ions are selected in Q1, and the m/z of the product ions are selected in Q3. b: Samples are digested and isotope-labeled standard peptides are added. Both endogenous and isotope-labeled peptides are then selected for analysis using a triple-quadrupole mass spectrometer. The extracted ion chromatograms under the AUC curve are then used for absolute quantitation of the peptide [39].

1.7.5. The absolute quantification method (AQUA)

The absolute quantification (AQUA) method involves the synthesis of stable isotope labeled peptides as known quantities and those peptides are added into the proteome sample to be analyzed. The figure 7 [40, 41] indicates the two strategies of AQUA approach. Stage 1 involves the selection and standard synthesis of a peptide from the protein of interest. During synthesis, a single amino acid residue is incorporated with stable isotopes. These peptide internal standards are analyzed by MS/MS to examine peptide fragmentation patterns. The next step is to set up a

SRM analysis. Proteins of interest are digested and mixed with the internal standard peptide. An

LC–SRM experiment measures the abundance of a specific fragment ion at a chromatographic retention time from both the native sample (endogenous peptide) and the synthesized sample

~ 12 ~

Introduction

(AQUA peptide). The absolute quantification is determined by comparing the abundance of the known internal standard peptide with the native peptide (endogenous peptide/ AQUA peptide).

~ 13 ~

Introduction

Figure 7. Absolute quantification of proteins strategy. [40, 41] Absolute quantification involves two stages. Stage 1 involves the selection and standard synthesis of a peptide from the protein of interest. During synthesis, a single amino acid residue is incorporate with stable isotopes. These peptide internal standards are analyzed by MS/MS to examine peptide fragmentation patterns. The next step is to set up a SRM analysis. Proteins of interest are proteolyzed and mixed with the internal standard peptide. An LC–SRM experiment measures the abundance of a specific fragment ion from both the native sample (endogenous peptide) and the synthesized sample (AQUA peptide) as a function of reverse-phase chromatographic retention time. The absolute quantification is determined by comparing the abundance of the known AQUA internal standard peptide with the native peptide (endogenous peptide/AQUA peptide).

~ 14 ~

Introduction

1.7.6. Label-free quantitation

A label-free approach has been increasingly employed for quantitative proteomic analysis of a large number of proteome samples without applying any stable isotope labeling. Proteome samples are digested using proteolytic enzymes and the resulting peptides are analyzed using

LC-MS/MS. After peptide ion peaks are aligned, the standard statistical tests could be applied on the raw or transformed counts to analyze the protein expression data depending on weighted peptide match scoring, normalization of the number of peptide matches and peptide sequence length [35, 42].

For each quantitative approach described, high-resolution mass spectrometers such as

FTICR-MS are expected to achieve better quantitation and higher confidence of protein identification. Nevertheless, chemical labeling is not 100% efficient and does not evenly represent all proteins or peptides. Chemical labeling also requires high cost, time, superior experimental skills for labeling with stable isotopes and offers a limited dynamic range of quantitation. Label-free protein quantitation methods overcome some of these limitations and therefore offer alternatives to stable isotope labeling methods.

1.8. Proteomics and clinical diagnostics

Obesity can be defined as an excess of body fat. A body mass index (BMI) is determined by weight (in kilograms) divided by height squared (in meters). A BMI of 25–29 kg/m2 will classify a patient as overweight and a higher BMI (>=30 kg/m2) will classify as obese [15].

Another way to define obesity is in terms of percent total body fat. This can be measured by several methods including skin-fold thickness, bioelectrical impedance and underwater weighing. In terms of percent body weight, obesity occurs when individuals have greater than

25% fat in men and greater than 35% in women [43]. However, these measurements are not necessarily convenient and can be costly.

~ 15 ~

Introduction

A term, metabolic syndrome, clusters together the conditions of dyslipidemia, hypertension and hyperglycemia. A diagnosis of this term can be made in relation to obesity [44]. Our understanding of the relation between obesity and metabolic risk factors is growing rapidly. This understanding is based on molecules known to be released in abnormal amounts. Each of these has been implicated in the causation of one or another of the metabolic risk factors. The following is a list of the factors most implicated in the development of metabolic syndrome:

Nonesterified fatty acids (NEFAs), Inflammatory cytokines, PAI-1, Adiponectin, Leptin and

Resistin [43, 45].

To profile metabolic risk factors or other secreted proteins with diagnostic information in obesity, it is necessary to screen multiple samples [46] for the presence of proteins of interest in a selective, parallel and absolute quantitative fashion using highly sensitive and specific

SRM/MRM technology [46, 47].

A selective SRM or MRM assay can also be established for the absolute quantification of the protein of interest. The necessary information about individual target proteins required for establishing the SRM or MRM assay can be retrieved either by searching the UniPep (or

PeptideAtlas) databases, or bioinformatically estimated using a suite of software tools [48, 49].

For each protein to be measured at least one peptide which is unique for the selected protein, called proteotypic peptide, has to be chosen [50]. Therefore, an effort is required to setup a database of targeted proteomics assays to detect and quantify proteins. For SRM/MRM-based mass spectrometric [51, 52] detection of peptides can provide a useful alternative to the immunoassays for example ELISA to develop and have limits with respect to multiplexing [46,

53].

~ 16 ~

Introduction

1.9. Hypothesis and Aims

1.9.1. Hypothesis

Cell numbers and cell types in adipose tissue are maintained by a combination of reciprocal and autonomic signaling involving adipocytes and mesenchyme-derived adult adipose stem/progenitor cells. This involves a balance of growth factors, growth inhibitors, signal enhancement and feedback inhibition. This balance can be shifted to recruit stem cells and thus increase adipocyte numbers. The molecular definition of adipose tissue as a complex multi-cell and multi-signalling system is essential to understand this recruitment process.

~ 17 ~

Introduction

1.9.2. Aims

To identify the peptides, proteins involved in signaling within and between adipocytes and adipose tissue. This will be done by:

1) To undertake a comprehensive proteomic analysis of proteins secreted from adipocytes in normal white adipose tissue

2) To understand if proteins secreted from isolated adipocytes are different to proteins secreted from tissue explants

3) To discover proteotypic peptides in proteins secreted from adipocytes that can be used to accurately monitor the quantity of proteins in clinical tests.

~ 18 ~

Materials and methods

2. Materials and methods

2.1. Sample preparation

2.1.1. Tissue harvest

Female Sprague-Dawley rats (Animal Resource Centre WA, Australia) were housed with chow at 20±2°C and maintained on a 12:12 h light/dark cycle until the day of tissue harvest. All procedures performed were in accordance with the guidelines described in the University of

New South Wales Animal Care and Ethics Committee (ACEC, ACEC Number: 08/13B) and under a protocol approved by Australian Code of Practice for the Care and Use of Animals for

Scientific Purposes, National Health and Medical Research, Council (6th edition, 1997).

Immediately after euthanasia, adipose tissue from the abdominal areas was surgically obtained using sterile techniques, placed in a petri dish containing 1X phosphate- buffered saline ([PBS]

Sigma, St. Louis, MO) and rapidly transported to the laboratory. Approximately 10g of fat tissue was obtained per harvest.

2.1.2. Cell isolation and culture

2.1.2.1. Isolation of adipocytes

Isolation methods were modified from those described by Tapp et al. [54] 20 to 30g of abdominal adipose tissue was placed in a 200ml sterile flask, finely minced and incubated with

50mg of type I collagenase (Sigma, St. Louis, MO) per 100ml Dulbecco’s modified Eagle’s

/Ham’s medium with 14.5g/L D-glucose, 25mM HEPES, 100U/ml penicillin. ([DMEM F12]

Gibco, Carlsbad, CA) at 37°C in a water-bath shaker for 2 h at 100 rpm. Three grams of undigested tissue was removed by filtering and used to create adipose tissue explants. Floating adipocytes were separated from stromal vascular fractions by centrifugation at 400 x g for 5 min at room temperature.

~ 19 ~

Materials and methods

2.1.2.2. Primary adipocytes culture

Isolated adipocytes were cultured in T75 tissue culture flasks (Primera, Falcon, BD Biosciences,

San Jose, CA) with 10ml serum-free DMEM F12. Cells were maintained in a humidified tissue culture incubator at 37°C with 5% carbon dioxide.

2.1.2.3. Adipose tissue explants

Three grams of undigested tissues were cultured in 60-mm2 round plastic tissue culture dishes

(Primera, Falcon, BD Biosciences, San Jose, CA) with 10ml serum-free DMEM F12. Cells were maintained in a humidified tissue culture incubator at 37°C with 5% carbon dioxide.

2.1.2.4. Harvest of secreted proteins

One milliliter of conditioned culture medium was collected every 24 h from cell cultures and tissue explants. This was replaced with 1 ml fresh DMEM F12 to maintain a final medium volume at 10 ml. The proteins in conditioned media were then concentrated. Ten volumes of cold 10% trichloroacetic acid ( [TCA] Sigma, St. Louis, MO) was added to the conditioned cultured medium and incubated at 4°C for 30 min. Proteins were pelleted by centrifugation at

4,000 g, at 4°C for 15 min. Supernatants were removed and pellets washed twice with cold acetone (Sigma, St. Louis, MO). The resulting pellet was air dried and finally resuspended in

15μl of 100 mM Tris-HCl buffer, pH=7.

2.2. Protein identification

2.2.1. In solution digestion

Proteins were digested with 40ng porcine trypsin (modified sequencing grade; Promega,

Madison, WI) in 50mM ammonium bicarbonate overnight at 37 °C. The resulting tryptic peptides were concentrated to dryness in a vacuum centrifuge, and then resuspended in 2% acetonitrile, 0.05% heptafluorobutyric acid buffer.

~ 20 ~

Materials and methods

2.2.2. Nano-flow LC-MS

Peptides from the digestion of secreted proteins were analysed using an LTQ-FT Ultra mass spectrometer (Thermo Electron, Bremen, Germany).Peptides were separated by nano-LC using an Ultimate 3000 HPLC and autosampler system (Dionex, Amsterdam, The Netherlands). One microlitre samples were concentrated and desalted onto a micro C-18 pre-column (500 μm × 2 mm, Microchem Bioresources, Auburn, CA) with 2% acetonitrile/0.05% heptafluorobutyric acid (v/v) at 15μL/min. After a 4 min wash, the pre-column was automatically switched (Valco

10 port valve, Houston, TX) into line with a fritless nano-column manufactured according to[55]. Peptides were eluted using a linear gradient of 2% acetonitrile/0.1% formic acid to 36%

acetonitrile/0.1% formic acid at ׽300 nL/ min over 50 min. The pre-column was connected via a fused silica capillary (25 cm, 25μL) to a low volume tee (Upchurch Scientific) where a high voltage (1800 V) was applied and the column tip positioned approximately 5 mm from the heated capillary (T = 200°C) of an LTQ-FT Ultra mass spectrometer (Thermo Electron, Bremen,

Germany).

Positive ions were generated by electrospray and the instrument operated in data dependent acquisition mode. A survey scan of m/z 350–1750 was acquired in the FT ICR cell (resolution

=100,000 at m/z 400). Up to seven of the most abundant ions (>2000 counts) with charge states of +2 or +3 were sequentially isolated and fragmented within the linear ion trap using collisionally induced dissociation. Peaks selected for MS/MS were dynamically excluded for

60s. Peak lists were generated using ‘MASCOT Daemon/extract_msn’ (Matrix Science, London,

England) using default parameters and submitted to the database search program MASCOT

(Version 2.2, Matrix Science) [56].

~ 21 ~

Materials and methods

2.3. Analysis of secreted proteins

2.3.1. Proteomic data analysis

All MS/MS data were matched against the UniProt database (version 15.10, consisting of

10,208,308 entries) using the MASCOT search engine (version 2.2, Matrix Science, London,

UK). The search parameters were ±4.0 ppm for parent mass ions and ±0.6 Da for MS/MS fragment spectra, allowing up to one missed cleavage. Variable modification of Met (O) and

Cysteine-carboxyamidomethyl were considered. Peptide identifications were accepted if ion scores indicated identity or extensive homology (significance threshold p<0.01). Protein scores were derived from ion scores as a non-probabilistic basis for ranking protein hits.

A “decoy database” [57, 58] was prepared by reversing the sequence of each entry and appending this database to the forward database. Initial peptide filtering was used to determine estimated 5% false discovery rates. Second, if proteins supported by only 1 spectral count were present but were seen consistently in a time course, they were also included.

2.3.1.1. Spectral counting

The MASCOT search results files (.dat) were imported into Scaffold 2.0™. Peptide identifications were accepted if they had > 90.0% probability, as specified by the

PeptideProphetTM algorithm [59]. Protein identifications were accepted if they could be established at greater than 90.0% probability and contained at least 1 identified peptide. Protein probabilities were assigned by the PeptideProphetTM algorithm.

Spectral counts were calculated by Scaffold. Normalized, quantitative values were normalized to the un-weighted spectral counts from the protein with the protein probabilities.

~ 22 ~

Materials and methods

2.3.2. Identification of proteotypic peptides from secreted proteins

All sequences of secreted proteins were submitted to the PeptideSieve program (version 0.51, released 5/2008) to generate a list of theoretical proteotypic peptides. This approach considers each peptide’s physico-chemical properties and tryptic enzyme digestion sites [48]. The parameters used for PeptideSieve were p-values less than 0.05, all experimental designs considered and allowing up to one missed cleavage. At the same time, the experimental data from all LC-MS-MS analyses were examined in Scaffold to find peptides consistently observed in replicate analyses of proteins secreted from adipocytes.

2.3.3. Functional analysis of identification proteins

To explore the localization and function of all secreted proteins, all of the identified and quantified proteins were analysed with SignalP 3.0 (http://www.cbs.dtu.dk/services/SignalP)

[24], SecretomeP 2.0 (http://www.cbs.dtu.dk/services/SecretomeP/) [23, 25], TMHMM

(http://www.cbs.dtu.dk/services/TMHMM/) [60], Gene Ontology (GO) annotation

(http://www.geneontology.org), GOA (version 74 of rat released) (http://www.ebi.ac.uk/GOA/ ),

Cytoscape (version 2.7) [61] and BiNGO (version 2.4.4.) [62]. Other functional annotations were examined in the Swiss-Prot/TrEMBL database (http://www.expasy.org), the human plasma protein dataset [63] and PubMed database (http://www.ncbi.nlm.nih.gov). Interactions of proteins were retrieved from the STRING database, version 8.3 (http://string-db.org/) [64].

Perl scripts were created to retrieve data from search results. Perl was used to reorganize the data form to an Excel file including protein name, gene name, function, location, molecular weight, pI, node and edge. This was used for 3-D- network construction with the GEOMI tool

[65].

~ 23 ~

Results

3. Results

To explore the proteins secreted from adipocytes and adipose tissue explants, isolated adipocytes and adipose tissues were harvested from wild Sprague-Dawley rats. Isolated adipocytes and adipose tissues were cultured separately in serum-free DMEM medium at 37°C with 5% carbon dioxide. One milliliter of conditioned medium was sampled 24 hours, for 18 days. The proteins were pelleted by acetone/ TCA precipitation and digested as described under

“Materials and Methods”. The strategy for incubation and harvest is shown in the figure 8, below. All samples were injected into FT-ICR mass spectrometry for qualitative and semi-quantitative and analysis.

Figure 8. Experimental strategy for incubation and harvest. Isolated adipocytes and adipose tissues were cultured separately in serum-free DMEM medium and each conditioned medium sampled at 37ƱC with 5% carbon dioxide every 24 hours. 1ml of conditioned culture medium was collected every 24 h from cell cultures and tissue explants and replaced with 1 ml fresh DMEM F12 to maintain a final medium volume of 10 ml.

~ 24 ~

Results

3.1. Identification of proteins from conditioned media: cultured

adipocytes

Proteins in conditioned media from culture adipocytes, over the 18 day time course, were identified by FT-ICR mass spectrometry and bioinformatics tools. The number of proteins identified on each day was variable, ranging from 119 (day 3) to 6 (day 13). In days 1 to 18, there were approximately 95 proteins that were seen in most samples. These proteins are detailed in supplementary table 1and table 6.1. Because of low quality of data obtained from day 11, it was not considered for further analyses, below.

To assess the accuracy of the identification process, the peptide false discovery rate (FDR) was calculated for each day, using the target decoy method [57]. This analysis showed that the number of false positive matches for each day was very low, ranging from 0 to 11 (Table 1).

This led to a false discovery rate of less than 5% in all cases.

~ 25 ~

Results

Table 1. The False Discovery Rate (FDR) for secreted proteins identified from primary adipocyte cultures over an 18 day time course. False Discovery Rate = False Positive / False Positive + True Positive * 100. Cell Peptide matches above homology or identity threshold (TP+FP) Decoy (false positive matches, FP) True positive matches (TP) Protein identified Peptide FDR Day 1 287 6 281 112 2.09% Day 2 132 5 127 48 3.79% Day 3 345 11 334 119 3.195 Day 4 138 6 132 53 4.35% Day 5 124 6 118 47 4.84% Day 6 291 4 287 69 1.37% Day 7 425 11 414 104 2.59% Day 8 188 5 183 44 2.66% Day 9 91 1 90 23 1.10% Day 10 100 0 100 24 0% Day 11 76 1 75 21 1.32% Day 12 33 0 33 12 0.00% Day 13 15 0 15 6 0.00% Day 14 94 2 92 16 2.13% Day 15 32 0 32 12 0.00% Day 16 28 0 28 10 0.00% Day 17 29 0 29 11 0.00% Day 18 86 0 86 15 0.00%

~ 26 ~

Results

3.1.1. Subcellular localization of proteins

Proteins which are found in conditioned media are known to be a mix of proteins which are actively secreted from the cell, along with proteins that arise from damage to cells or from cell death. To understand which types of proteins were present in the conditioned media, we analysed all 125 non-redundant identified proteins secreted by cultured adipocytes with a series of databases and tools. These included the Swiss-Prot database for subcellular location and tools

SecretomeP 2.0, SignalP 3.0 and TMHMM.

Subcellular location was determined for each protein, based on the Swiss-Prot database

‘subcellular location’ annotation. These results are presented in Figure 9 Thirty-eight percent of proteins were classified as secreted or extracellular matrix. Twenty-three percent and 6 percent of proteins were identified as cytoplasmic and nuclear, respectively. A small number of proteins were from the mitochondria (2%) or endoplasmic reticulum (4%). Nine percent of cell membrane proteins were detected and 18 percent of identified proteins were of unknown location.

Proteins were then analysed to understand if they are secreted through a classical or non-classical pathway. Sixty-two proteins (50%) were predicted by SignalP [24, 27] to contain a signal peptide suggesting they are secreted by classical means.

The SecretomeP 2.0, non-classically secreted proteins prediction tool was then applied to distinguish non-classical secreted proteins from cellular proteins. This considers protein amino acid composition, secondary structure and disordered regions. Twenty-four proteins were predicted to be non-classically secreted proteins, with no signal peptides present but NN-scores higher than 0.5. Scores between 0.5 and 1.0, the maximum possible score, are considered to be strong candidates for non-classical secretion by SecretomeP 2.0.

~ 27 ~

Results

To investigate if proteins have membrane-spanning regions, protein sequences were analysed with the TMHMM tool. Twelve proteins were predicted to have transmembrane domains, however all of these were predicted by TMHMM to have transmembrane domains which corresponded to N-terminal signal peptides. Interestingly, cell surface glycoprotein MUC18 was found with a transmembrane domain and it was known from above, to be cell membrane-associated. The above results may be expected that many glycoprotein targeted to cell surfaces have the signal peptide domains that allow them to pass-though membranes.

Unkown 18% Secreted, extracellular matrix. Mitochondrion 38% 2% Endoplasmic reticulum lumen, Peroxisome, 4% Cell membrane 9%

Nucleus 6%

Cytoplasm 23%

Figure 9. Subcellular location categories of identification proteins from primary adipocyte culture media. Subcellular location were distributed into the following categories: 38% for secreted or extracellular matrix, 23% for cytoplasm and 9% as nucleus proteins, 2% for mitochondria, 4% for endoplasmic reticulum (4%), 9% for cell membrane and 18% for unknown location.

~ 28 ~

Results

3.1.2. Functional categorization of proteins

The 125 identified proteins identified from conditioned media associated with isolated adipocytes were then investigated to understand their function. This was done based on the information extracted from the Swiss-Prot database and the Gene Ontology. Classification of proteins by molecular function revealed that most identified proteins were involved in protein binding such as metal ions binding, receptor binding and hormone binding. For example, adiponectin, transthyretin, talin and fibrinogen proteins were associated with hormone or receptor binding associated with extracellular signal pathways.

The result of biology process classification was presented in the figure 10. Thirty-six percent of proteins were involved in extracellular signal pathways. Proteins of this type included adiponectin which is known as a typical adipokine and can be regulated during fat cell differentiation. Platelet glycoprotein 4 can be stimulated by growth hormone or insulin.

Thirty-two percent were related with metabolic or biosynthetic processes. Examples of these include ATP synthase subunit (ADP biosynthetic process and ATP metabolic process) and sulfhydryl oxidase (protein thiol-disulfide exchange). Twenty percent of proteins were classified into cell cycle, growth, proliferation or differentiation. These included proteins such as collagen alpha-2(IV) (angiogenesis) and xanthine dehydrogenase/oxidase (bone resorption). Nine and three percent of proteins were functionally ‘other’ or associated with the immune system, respectively.

Gene Ontology (GO) slim is a framework consisting of controlled vocabularies describing three aspects of gene product functions: (I) molecular function, (II) biological process and (III) cellular component. The GO terms were represented as nodes in the figure 11 and were arranged hierarchically from general to specific. Here functional annotations of protein were combined with GO biological process, molecular functions and cellular component.

~ 29 ~

Results

Cytoscape (version 2.7) and BiNGO (version 2.4.4.) were used to perform gene ontology assignments and determine significantly over-represented GO categories. Analysis was performed using the default BiNGO Rattus norvegicus annotation containing 256,737

(12/25/2010; http://www.geneontology.org/) members and either the GO Slim (GOA; version

1.2). Statistical significance was determined by hypergeomtric analysis followed by Benjamini and Hochberg false discovery rate (FDR) correction (P<0.01). Of all 120 genes, 119 genes clustered and 3 genes were found to be unannotated. The GO slim network was built in the

3D-network visualization platform, GEOMI.

Cell cycle, growth, proliferation, differentiation Signal pathway 20% 36% Immune system 3%

Metabolic, Other: transport, biosynthetic behavior process 9% 32%

Figure 10. Functional categories of identified proteins from primary adipocyte culture media. The descriptive terms were according with Gene Ontology. 36% for signal pathway, 32% for metabolic or biosynthetic process. 20% for cell cycle including cell growth, proliferation and differentiation, 3% and 9 % for immune system and others, respectively.

~ 30 ~

Results

Figure 11. Gene Ontology analysis and network representation of proteins from cultured adipocyte medium. Cytoscape (version 2.7) and BiNGO were used to perform gene ontology assignments and determine significantly over-represented GO categories. The force directed layout was used to calculate this layout (force strengths: Spring = 50, Origin = 80, Replulsion = 12).

One limitation of functional classification is that, in most cases, proteins map to more than one functional group. Pie-graph representations, such as that shown in figure 10, cannot represent the multiple functions of many proteins. To better understand the functional inter-relationships of proteins secreted by adipocytes, a network-based representation was built (Figure 11 and

Table 2). The network shows three different connected components. These relate to the three parts of the gene ontology - molecular process, biological function and cellular compartment.

Within each of the three networks shown, each node (dot) is one functional group of proteins.

Connections between nodes show proteins that are common between two functional categories.

Results

ontology contained proteins significantly over-represented in response to stimulus (N51, p=6.8x10-9) as well as regulation of biological process (N50, p =6.87x10-9). The cellular component ontology contained significantly over-represented in extracellular space (N12, p=6.1x10-31) and cytoplasm (N17, p=5.22x10-14) which is a part of N13. The precise details of which proteins are involved in multiple processes can be seen in the table (supplementary data).

Moreover, through BiNGO analysis, classification of non-classical secreted proteins was possible for those not documented as secreted from Swiss-Prot but predicted by SecretomeP 2.0.

Annexin A2, for example, is significantly present in the extracellular space (p=6.10x10-31) and extracellular structure organization (p=6.69x10-5). Calmodulin is over-represented in the extracellular space. Carbonic anhydrase 2, sulfhydryl oxidase 1 and superoxide dismutase

[Cu-Zn] are involved in extracellular regions (p=9.82x10-38) or the extracellular space. Platelet glycoprotein 4 and 78 kDa glucose regulated protein were located on the cell surface

(p=1.29x10-6). Membrane primary amine oxidase can be represented in cell surface, extracellular structure organization and extracellular space.

~ 32 ~

Results

Table 2. The list of individual nodes. Node GO-ID p-value Description Node GO-ID p-value Description N0 3674 1.39E-01 molecular function N28 9055 2.98E-01 electron carrier activity N1 3676 1 nucleic acid binding N29 9056 2.35E-07 catabolic process N2 3824 1.12E-01 catalytic activity N30 9058 2.64E-02 biosynthetic process N3 4871 1 signal transducer activity N31 9986 1.29E-06 cell surface N4 4872 1 receptor activity N32 9987 3.25E-07 cellular process N5 5198 2.29E-03 structural molecule activity N33 15075 4.19E-01 ion transmembrane transporter activity N6 5215 2.10E-02 transporter activity N34 15267 9.01E-01 channel activity N7 5488 1.39E-08 binding N35 16020 4.57E-01 membrane N8 5515 4.12E-17 protein binding N36 16209 3.14E-04 antioxidant activity N9 5575 3.89E-06 cellular component N37 16301 9.68E-01 kinase activity N10 5576 9.82E-38 extracellular region N38 16491 3.13E-05 oxidoreductase activity N11 5578 2.13E-14 Proteinaceous extracellular matrix N39 16740 9.85E-01 transferase activity N12 5615 6.10E-31 extracellular space N40 16787 5.20E-01 hydrolase activity N13 5622 8.25E-05 intracellular N41 16829 3.41E-05 lyase activity N14 5623 2.82E-01 cell N42 16853 1.11E-02 isomerase activity N15 5634 5.88E-01 nucleus N43 30154 5.37E-05 cell differentiation N16 5694 7.41E-01 chromosome N44 30234 3.12E-06 enzyme regulator activity N17 5737 5.22E-14 cytoplasm N45 30528 9.44E-01 transcription regulator activity N18 6139 9.79E-03 nucleobase, nucleoside, nucleotide, nucleic acid metabolic N46 32501 5.77E-08 multicellular organismal process process N19 6519 3.23E-01 cellular amino acid and derivative metabolic process N47 43062 6.69E-05 extracellular structure organization N20 6810 6.77E-07 transport N48 43170 8.16E-01 macromolecule metabolic process N21 6928 3.64E-01 cellular component movement N49 46903 5.21E-03 secretion N22 7154 5.46E-02 cell communication N50 50789 6.87E-09 regulation of biological process N23 7275 3.93E-09 multicellular organismal development N51 50896 6.80E-09 response to stimulus N24 7610 3.46E-02 behavior N52 51704 1.94E-05 multi-organism process N25 8150 2.38E-03 biological process N26 8152 8.94E-05 metabolic process N27 8219 4.87E-02 cell death

~ 33 ~

Results

Finally, the list of all identified proteins were cross-checked with the human plasma proteome dataset, complied by Anderson et al [66]. Of the 125 proteins identified across all time points,

27 proteins were found to match homologs in human plasma proteome dataset. This suggests that there was some plasma contamination in the cultured adipocytes and in the analysis, but that the majority of proteins were secreted from or associated with adipocytes themselves.

3.1.3. Proteins from adipocytes culture: a summary

In summary, there were 34 identified proteins which were found as secreted proteins involved in extracellular signal pathways and 40 secreted proteins that function in metabolic processes.

Eleven non-classical secreted proteins were also found in this study.

The total of the identified secretome was compared to adipocyte proteomics literature describing the currently known secretome from human visceral and subcutaneous, rodent adipose tissue

[67-75] as well as Simpson-Golabi-Behmel syndrome (SGBS) cell lines [76]. Forty-six novel secreted proteins were found in this study, as compared previous analyses. Thirty-three of them were of known localization from Swiss-Prot.

Of the 46 proteins, only 3 of these were likely to have been actively secreted. This was determined by use of secretion protein prediction software and known functional knowledge from the Swiss-Prot database. These 3 proteins are afamin, seminal vesicle secretory protein 2 and xanthine dehydrogenase/oxidase, as shown in the Table 3, below.

~ 34 ~

Results

Table 3. Protein from cultured adipocytes. The results from secretion protein prediction software and functional knowledge from the Swiss-Prot database is also shown, to indicate the likelihood of proteins being actively secreted. Those not reported previously to be secreted from adipocytes but appear secreted by classical means, are afamin, seminal vesicle secretory protein 2 and xanthine dehydrogenase/oxidase (shown in bold). Accession Number Identified Proteins Location SignalP-NN SignalP-HMM Non-classical TMHMM unknown predictions predictions protein secretion in predictions literature as secreted by adipocytes AFAM_RAT Afamin Secreted. Y Y Y N novel AOC3_RAT Membrane primary amine oxidase Membrane protein. Y Y Y Y novel AT PA_RAT ATP synthase subunit alpha, mitochondrial Mitochondrion inner Y N Y N novel membrane. CAH1_RAT Carbonic anhydrase 1 Unknown N N Y N novel CAH2_RAT Carbonic anhydrase 2 Cytoplasm. N N Y N novel CALM_RAT Calmodulin Cytoplasm. N N Y N novel CD36_RAT Platelet glycoprotein 4 Membrane protein. Y N Y Y novel CES3_RAT Carboxylesterase 3 ER. Y Y Y N novel CLOS_CLOHI Clostripain Unknown Y Y Y N novel DLDH_RAT Dihydrolipoyl dehydrogenase, Mitochondrion matrix. N N Y N novel mitochondrial EST2_RAT Liver carboxylesterase 1 ER. Y Y Y N novel FAAA_RAT Fumarylacetoacetase Unknownġ N N Y N novel ~ 35 ~

Results

Accession Number Identified Proteins Location SignalP-NN SignalP-HMM Non-classical TMHMM unknown predictions predictions protein secretion in predictions literature as secreted by adipocytes FABP4_RAT Fatty acid-binding protein, adipocyte Cytoplasm. Nucleus. N N Y N novel FRIH_RAT Ferritin heavy chain Unknownġ N N Y N novel G6PI_RAT Glucose-6-phosphate isomerase Secreted. Cytoplasm. N N ġ N novel GPDA_RAT Glycerol-3-phosphate dehydrogenase Cytoplasm. N N Y N novel [NAD+], cytoplasmic GRIFN_MOUSE Grifin Unknownġ N Y Y N novel GUAD_RAT Guanine deaminase Unknownġ Y Y ġ N novel LDHB_RAT L-lactate dehydrogenase B chain Cytoplasm. N N Y N novel MYP0_RAT Myelin P0 protein Membrane protein. Y Y Y Y novel PLMN_RAT Plasminogen Secreted. Y Y ġ N novel PRDX1_RAT Peroxiredoxin-1 Cytoplasm. N N Y N novel Melanosome. PTRF_RAT Polymerase I and transcript release factor Cell membrane. ER. N N Y N novel Cytoplasm. Mitochondrion. S10A6_RAT Protein S100-A6 Nucleus. Cytoplasm. N N Y N novel Cell membrane. ~ 36 ~

Results

Accession Number Identified Proteins Location SignalP-NN SignalP-HMM Non-classical TMHMM unknown predictions predictions protein secretion in predictions literature as secreted by adipocytes Cytoplasm. S10AB_RAT Protein S100-A11 Cytoplasm. Nucleus. N N Y N novel SODC_RAT Superoxide dismutase [Cu-Zn] Cytoplasm. N N Y N novel SPA3K_RAT Serine protease inhibitor A3K Secreted. Y Y Y N novel SPA3N_RAT Serine protease inhibitor A3N Secreted. Y Y Y N novel SVS2_RAT Seminal vesicle secretory protein 2 Unknownġ Y Y ġ N novel SVS5_RAT Seminal vesicle secretory protein 5 Secreted, extracellular Y Y Y N novel space. UBIQ_RAT Ubiquitin Cytoplasm. Nucleus. N N Y N novel XDH_RAT Xanthine dehydrogenase/oxidase Peroxisome. N N ġ N novel Cytoplasm. Secreted.

~ 37 ~

Results

3.2 Identification of proteins from conditioned media: adipose tissue

explants

Proteins in conditioned media from adipose tissue explants, over the 18 day time course, were identified by FT-ICR mass spectrometry and bioinformatics tools. The range of proteins identified on each day was variable, ranging from 209 (day 12) to 57 (day 7). In days 1 to 18, there were approximately 251 proteins that were seen with consistency. All proteins are detailed in supplementary table 1and table 6.1. Because of low quality of data obtained from days 7th , 12th , and 17th ,they were not considered for further analyses, below.

This led to a false discovery rate of less than 5% in all cases, with the exception of Day 2.

~ 38 ~

Results

Table 4. The False Discovery Rate (FDR) for secreted proteins identified from adipose explants cultures over an 18 day time course. False Discovery Rate = False positive / False Positive + True Positive * 100. Tissue Peptide matches above homology or identity threshold Decoy (false positive matches, True positive matches (TP) Proteins Peptide FDR (TP+FP) FP) identified Day 1 335 7 328 100 2.09% Day 2 603 45 558 147 7.46% Day 3 478 9 469 145 1.88% Day 4 564 11 553 143 1.95% Day 5 550 13 537 152 2.36% Day 6 592 6 586 164 1.01% Day 7 169 5 164 57 2.96% Day 8 606 15 591 150 2.48% Day 9 643 12 631 173 1.87% Day 10 582 14 568 167 2.41% Day 11 716 15 701 184 2.09% Day 12 245 3 242 86 1.22% Day 13 802 17 785 209 2.12% Day 14 768 17 751 205 2.21% Day 15 611 14 597 166 2.29% Day 16 441 10 431 127 2.27% Day 17 161 6 155 60 3.73% Day 18 697 13 684 196 1.87%

~ 39 ~

Results

3.2.1 Subcellular localization of proteins

Proteins which are found in conditioned media are known to be a mix of those which are actively secreted from the cell, along with proteins that arise from damage to cells or from cell death. To understand which types of proteins were present in the conditioned media, we analysed all 258 identified proteins from adipose tissue explants model with a series of databases and tools. These included an analysis of information in Swiss-Prot, and use of the

SecretomeP 2.0, SignalP 3.0 and TMHMM prediction tools.

Subcellular location was determined for each protein, based on the Swiss-Prot “subcellular location” annotation. These results were presented in Figure 12. Thirty-four percent of proteins were classified as secreted or extracellular matrix. Thirty-one percent and six percent of proteins were identified as cytoplasmic and nuclear, respectively. A small number of proteins were from the mitochondria (2%) or endoplasmic reticulum (7%). Three percent of proteins were from the cell membrane. The remaining 17% of identified proteins were of unknown subcellular location.

Proteins were then analysed to understand if they are secreted through a classical or non-classical pathway. Eighty-six proteins were predicted by SignalP to contain a signal peptide suggesting they are secreted by classical means.

The SecretomeP 2.0 tool, for non-classically secreted protein prediction, was then applied to distinguish non-classical secreted proteins from cellular proteins by amino acid composition, secondary structure and disordered regions. Sixty proteins were predicted to be non-classically secreted proteins, with no signal peptides containing but NN-scores higher than 0.5. Scores between 0.5 and 1.0, the maximum possible score, are considered to be strong candidates for non-classical secretion by SecretomeP 2.0.

~ 40 ~

Results

To investigate if proteins have membrane-spanning regions, protein sequences were analysed with the TMHMM tool. Twenty-two proteins were predicted to have transmembrane domains, however 17 of these were predicted by TMHMM only to have transmembrane domains which corresponded to N-term signal peptides. Interestingly, 5 proteins were found with transmembrane domains but only dystroglycan was known from above, to be secreted. The above results may be expected as proteins targeted to membranes have the same signal sequence or domains as those for secretion.

Taking subcellular location results and classical and non-classical pathway analyses together, 87 proteins with signal peptide accorded with the subcellular location “secreted” from Swiss-Prot.

Of the remainder, there were 17 proteins with signal peptides which were located at endoplasmic reticulum, lysosome, Golgi membrane, and mitochondrion.

Mitochondrion Unkown 2% 17% Secreted, Endoplasmic extracellular matrix. reticulum lumen, 34% Peroxisome, Lysosome 7%

Cell membrane 3%

Nucleus 6%

Cytoplasm 31%

Figure 12. Subcellular location categories of identification proteins from the cultured mediam of adipose tissue explants. Subcellular location were distributed into the following categories: 34% for secreted or extracellular matrix, 31% for cytoplasm and 6% as nucleus proteins, 2% for mitochondria, 7% for endoplasmic reticulum, 3% for cell membrane and 17% for unknown location. ~ 41 ~

Results

3.2.2. Functional categorization of proteins

The 258 proteins identified from conditioned media were then investigated to understand their functional categorization. This was done based on the information extracted from the Swiss-Prot database and the Gene Ontology. Classification of proteins by molecular function revealed that most identified proteins were involved in protein binding such as metal ions binding, peptide binding, lipid binding and protein complex binding. However, some proteins like protein

S100-B, alpha-2-HS-glycoprotein, fibrinogen, laminin and talin were associated with receptor binding in signal pathways.

The result of biology process classification was summarized in the figure 13. Thirty-eight percent of proteins were involved in signal pathways, including extracellular sginalling. Proteins of this type included adiponectin and leptin which were known as typical adipokines and can be regulated by MAPKKK, JAK-STAT pathways or during cell differentiation. Complement factor

D was documented to be associated with Notch signaling pathway whereas Cathepsin B, collagen, SPARC and metalloproteinase inhibitor 1 can be stimulated by hormones. Thirty-one percent were related with metabolic or biosynthetic processes. Examples of these include

L-lactate dehydrogenase (lactate metabolic process), apolipoprotein A-I and IV (lipoprotein metabolic process ), and fatty acid-binding protein (fatty acid catabolic process).

Twenty-three percent of proteins were classified into cell cycle, growth, proliferation or differentiation. These included proteins such as gelsolinġ(apoptosis), matrix Gla protein (cell differentiation) and pigment epithelium-derived factor (cell proliferation). Five and three percent of proteins were functionally ‘other’ or associated with the immune system, respectively.

Gene Ontology (GO) slim is a framework consisting of controlled vocabularies describing three aspects of gene product functions: (I) molecular function, (II) biological process and (III)

~ 42 ~

Results

cellular component. The GO terms were represented as nodes in the figure 14 and were arranged hierarchically from general to a specific. Here functional annotations of protein were defined by combined with GO biological process, molecular functions and cellular component.

Cytoscape (version 2.7) and BiNGO (version 2.4.4.)were used to perform gene ontology assignments and determine significantly over-represented GO categories. Analysis was performed using the default BiNGO Rattus norvegicus annotation containing 256,737

(12/25/2010; http://www.geneontology.org/) members and either the GO Slim (GOA; version 74 of rat release). Statistical significance was determined by hypergeomtric analysis followed by

Benjamini and Hochberg false discovery rate (FDR) correction (P<0.01). Of all 253 genes,

247 genes were selected to be clustered and 6 genes were found to be unannotated. The GO slim network was built by using 3D-network visualization platform, GEOMI.

One limitation of functional classification is that, in most cases, proteins map to more than one functional group. Pie-graph representations, such as that shown above, cannot represent the multiple functions of many proteins. To better understand the functional inter-relationships of proteins secreted by adipocytes, a network-based representation was built (Figure 14 and Table

5). The network shows three different connected components. These relate to the three parts of the gene ontology - molecular process, biological function and cellular compartment. Within each of the three networks shown, each node (dot) is one functional group of proteins.

Connections between nodes show proteins that are common between two functional categories.

The networks reveal that many proteins have more than one function, are part of more than one biological process or are present in more than one part of the cell. Identified proteins can be significantly classified by P-values (p<0.01). For example, leptin of molecular function was assigned to binding (N8, p=3.5x10-20) which is a part of N9 (protein binding) and N1 (nucleic acid binding). The biology process ontology contained proteins significantly over-represented in response to cell differentiation (N45, p=3.59x10-11) as well as cell communication (N23, p

~ 43 ~

Results

=1.16x10-2). The cellular component ontology contained significantly over-represented in extracellular space (N13, p=3.99x10-49) and a part of cellular component N10 (p=2.71x10-10).

The precise details of which proteins are involved in multiple processes can be seen in the table

(supplementary data).

BiNGO analysis allows complete classification of non-classical secreted proteins which were not described as secreted from Swiss-Prot but were predicted by SecretomeP 2.0. Galectin-3 and annexin A2 were significantly represented the extracellular regions (p=2.56x10-60) and extracellular structure organization (p=9.1x10-13). Purine nucleoside phosphorylase, thioredoxin and calmodulin were over-represented only in the extracellular regions (p=2.56x10-60). Arginase

1, superoxide dismutase [Cu-Zn], translationally controlled tumor protein, pigment epithelium derived factor, delta aminolevulinic acid dehydratase and carbonic anhydrase 2 were involved in the extracellular regions (p=9.82x10-38) or extracellular space (p=3.99x10-49). Protein disulfide isomerase, heat shock protein beta 1 and 78 kDa glucose regulated protein were located on the cell surface (p=4.09x10-7). Membrane primary amine oxidase, phosphatidylethanolamine binding protein 1 and cathepsin B were represented the cell surface, extracellular region and extracellular space. Cathepsin L1 can be present in the cell surface or extracellular region.

~ 44 ~

Results

Cell cycle, growth, proliferation, differentiation Signal pathway 23% 38%

Immune system 3%

Metabolic, biosynthetic Other: transport, process behavior 31% 5%

Figure 13. Functional categories of identified proteins from the culture medium of adipose tissue explants. The descriptive terms were according with gene ontology. 38% for signal pathway, 31% for metabolic or biosynthetic process. 23% for cell cycle including cell growth, proliferation and differentiation, 3% and 5 % for immune system and others, respectively.

~ 45 ~

Results

Figure 14. Gene Ontology analysis and network representation of proteins from adipose tissue explants medium. Cytoscape (version 2.7) and BiNGO were used to perform gene ontology assignments and determine significantly over-represented GO categories. The force directed layout was used to calculate this layout (force strengths: Spring = 50, Origin = 80, Replulsion = 12).

~ 46 ~

Results

Table 5. The list of individual nodes. Node GO-ID p-value Description Node GO-ID p-value Description N0 3674 2.05E-05 molecular_function N29 9055 1.90E-01 electron carrier activity N1 3676 1.00E+00 nucleic acid binding N30 9056 3.25E-12 catabolic process N2 3774 5.63E-01 motor activity N31 9058 2.02E-02 biosynthetic process N3 3824 7.41E-05 catalytic activity N32 9986 4.09E-07 cell surface N4 4871 1 signal transducer activity N33 9987 1.49E-14 cellular process N5 4872 1 receptor activity N34 15075 9.94E-01 ion transmembrane transporter activity N6 5198 2.37E-04 structural molecule activity N35 15267 9.92E-01 channel activity N7 5215 7.77E-01 transporter activity N36 16020 9.96E-01 membrane N8 5488 3.50E-20 binding N37 16209 2.23E-12 antioxidant activity N9 5515 7.99E-29 protein binding N38 16301 9.95E-01 kinase activity N10 5575 2.71E-10 cellular_component N39 16491 3.47E-08 oxidoreductase activity N11 5576 2.56E-60 extracellular region N40 16740 9.95E-01 transferase activity N12 5578 2.14E-21 proteinaceous extracellular matrix N41 16787 2.28E-03 hydrolase activity N13 5615 3.99E-49 extracellular space N42 16829 1.11E-04 lyase activity N14 5622 9.19E-14 intracellular N43 16853 7.99E-05 isomerase activity N15 5623 3.42E-02 cell N44 16874 9.88E-01 ligase activity N16 5634 3.06E-04 nucleus N45 30154 3.59E-11 cell differentiation N17 5694 4.60E-01 chromosome N46 30234 2.47E-11 enzyme regulator activity N18 5737 2.07E-30 cytoplasm N47 30528 9.95E-01 transcription regulator activity N19 6139 3.85E-01 nucleobase, nucleoside, nucleotide and nucleic acid metabolic N48 32501 3.61E-14 multicellular organismal process ~ 47 ~

Results

Node GO-ID p-value Description Node GO-ID p-value Description process N20 6519 3.08E-03 cellular amino acid and derivative metabolic process N49 43062 9.10E-13 extracellular structure organization N21 6810 7.45E-03 transport N50 43170 4.44E-03 macromolecule metabolic process N22 6928 1.43E-02 cellular component movement N51 46903 4.72E-02 secretion N23 7154 1.16E-02 cell communication N52 50789 2.95E-16 regulation of biological process N24 7275 1.01E-18 multicellular organismal development N53 50896 1.75E-17 response to stimulus N25 7610 1.64E-04 behavior N54 51704 6.71E-10 multi-organism process N26 8150 2.09E-08 biological_process N27 8152 1.31E-11 metabolic process N28 8219 3.63E-05 cell death

~ 48 ~

Results

Finally, the lists of identified proteins were cross checked to compare with the human plasma proteome dataset, complied by Anderson et al [66]. Of the 258 proteins identified across all time points, homologs to only 33 proteins were found in the plasma proteome dataset. This suggests that there was some plasma contamination in the explants and in the analysis, but that the vast majority of proteins were from the tissue explants itself.

3.2.3 Proteins from tissue explants: a summary

In summary, there were 78 identified proteins which were found as secreted proteins involved in extracellular signal pathways and 94 secreted proteins functioning in metabolic processes.

Twenty-four non-classical secreted proteins lacking localization information were also found in this study.

The total of the identified secretome was compared to adipocyte proteomics literature describing the currently known secretome from human visceral and subcutaneous, rodent adipose tissue [67-75] as well as SGBS cell lines [76]. Eighty-one novel secreted proteins were found in this study. Twenty of them were found with known localization from Swiss-Prot.

Of the 81 proteins, only 9 of these were likely to have been actively secreted. This was determined by use of secretion protein prediction software and known functional knowledge

~ 49 ~

Results

from the Swiss-Prot database. These 10 proteins are shown in the table 6, below.

~ 50 ~

Results

Table 6. Novel secreted proteins in the context from adipose tissue explants. Secretion protein prediction software and known functional knowledge from Swiss-Prot database suggest these are secreted by classical means. These have not been reported previously in association with adipocytes. Accession Identified Proteins Location SignalP-NN SignalP-HMM Non-classical TMHMM unknown in Number predictions predictions protein secretion literature as predictions secreted by adipose tissue CMA1_RAT Chymase Secreted. Cytoplasm. Y Y Y N novel MGP_RAT Matrix Gla protein Secreted. Y Y Y N novel PGS1_RAT Biglycan Secreted, extracellular space. Y Y Y N novel SPA3K_RAT Serine protease inhibitor A3K Secreted. Y Y Y N novel SPA3L_RAT Serine protease inhibitor A3L Secreted. Y Y Y N novel SPA3M_RAT Serine protease inhibitor A3M Secreted. Y Y Y N novel SPA3N_RAT Serine protease inhibitor A3N Secreted. Y Y Y N novel SPRL1_RAT SPARC-like protein 1 Secreted, extracellular space. Y Y Y N novel SVS2_RAT Seminal vesicle secretory protein 2 Unknownġ Y Y ġ N novel

~ 51 ~

Results

3.3. Comparison between two types of culture systems

A total 281 proteins were found from cultured adipocytes or adipose tissue explants.

Twenty-nine proteins were seen only in the media of cultured adipocytes and 156 proteins only in adipose tissue explants. Ninety-nine of classical and non-classical secreted proteins were found from adipose tissue explants and sixteen of classical or non-classical secreted proteins were found in adipocytes culture model. Together, this suggests that tissue explants are a more powerful system in which to study proteins secreted from adipocytes.

Plasma or other cytoplasmic proteins contamination is a key issue in the study of secreted protein [77]. Therefore, we filtered our results to enrich for secreted proteins. Of the 29 candidate proteins were found in adipocyte culture system, 6 proteins were plasma proteins

(complement C3, fibrinogen beta and gamma chain, retinol-binding protein 4, fibrillin-1 and collagen) and 5 proteins were confirmed as secreted proteins (afamin, vitamin D-binding protein, seminal vesicle secretory protein 5, xanthine dehydrogenase/oxidase and sulfhydryl oxidase 1).

Others were located at the cell membrane, cell cytoplasm or nucleus except ER. On the other hand, of the 156 candidate proteins unique to the tissue explants, 20 plasma proteins and 30 secreted proteins matched those reported by previous literature or Swiss-Prot database. In addition, 7 ER proteins and 1 low concentration of adipokine (leptin) was found from adipose tissue explants.

~ 52 ~

Results

Figure 15. The number of identified proteins compared with adipocytes culture and tissue explants. A total 281 proteins were found from adipocytes culture or adipose tissue explants system. Twenty-nine proteins were seen only in the media of cultured adipocytes and 156 proteins only in adipose tissue explants. Ninety-six secreted proteins were found in adipocyte culture and also in adipose tissue explants.

~ 53 ~

Results

3.4. Quantitative analysis of classical and non-classical secreted

proteins

An aim of this project was to observe the relative changes of the same protein across same experimental conditions over time and to compare with relative change reported in previous studies [78] . Semi-quantitative MS based on spectral counting can be used to compare the relative abundance between different proteins. This is done by normalizing spectral counts by either the protein molecular weight or by the number of observable tryptic peptides [79]. The spectral count for each protein between different time points in this study was compared by use of Scaffold. It should be noted that, due to the lack of biological replicates here, that the trends reported may require further confirmation. However, we believe they highlight the types of expression profiles one may expect to observe in these culture systems.

To compare protein levels quantitatively in the two secretomes (adipocytes culture and adipose tissue explants), the spectral count was normalized per day. Days where there were low numbers of proteins identified were removed from this analysis, to avoid bias in results. These included days 9 and 11 to 18 in adipocyte cultre and days 7, 12 and 17 in the cultured explants.

Based on considerations of quantitative changes, the identified proteins were classified into 5 classes. Type A response is characterized by decreased levels of protein quantization.

~ 54 ~

Results

Figure 16. Example of type A protein expression. The protein quantization was decreasing during culture. a) protein was detected in adipocyte culture. b)&c) proteins were measured in adipose explants.

Type B response exhibits increased levels of protein quantization at the beginning phase until reach a peak and trend go down until the end of culture time.

Figure 17. Example of type B protein expression. The protein quantization was increasing until reach a peak then trend going down at the end of culture time. a) protein was detected in adipocyte culture. b) protein was measured in adipose explants.

~ 55 ~

Results

Type C response pattern exhibits that the levels of protein quantization were only obtained at special time points during culture.

Figure 18. Example of type C protein expression. The levels of protein quantization were obtained only at special time points during culture. a) protein was detected in adipocyte culture. b) protein was measured in adipose explants.

Type D response present characterized as cyclical during culture by increased and decreased levels of protein quantization.

Figure 19. Example of type D protein expression. The levels of protein quantization were increased and decreased as cyclical during culture. a) proteins were detected in adipocyte culture. b) proteins were measured in adipose explants.

~ 56 ~

Results

Type E response is characterized by increased levels of protein quantization.

Figure 20. Example of type E protein expression. The levels of protein quantization were increased during culture. a) proteins were detected in adipocyte culture. b) proteins were measured in adipose explants.

From adipocytes culture, eight and seven proteins were found and grouped into type A and type

B protein expression, respectively. 112 found proteins were classified as type C expression. The number of type C protein expression was 30 and only 2 proteins were involved in type E. On the other hand, from adipose tissue explants, there were 5 and 4 identified proteins which were involved in type A and type E expression, respectively. 14 identified proteins were found as type

B expression and type D expression contained 113 identified proteins. The large number of type

C proteins expression was127.

The protein expression levels of all secreted and non-classical secreted proteins were measured in two serum-free culture systems (appendix 6.1). Some proteins showed different protein expression levels in the two culture systems, for example, leptin and adiponectin (figure 21b). In adipocytes culture system, adiponectin was expressed as type C but leptin was undetectable. In adipose tissue explants, leptin was expressed as type C and adiponectin express as type B.

~ 57 ~

Results

Figure 21. The protein expression of leptin and adiponectin in adipocyte culture and adipose tissue explant. a) adiponectin was present but leptin was not detected in adipocyte culture systems. b) leptin expressed as type C and adiponectin expressed as type B in adipose tissue explants.

3.5. Proteotypic peptides detection

There has been increasing interest in the development of assays for the measurement of proteins and peptides secreted into the extracellular environment. Many of these can ultimately be detected in the circulation, or plasma. To investigate whether our study can inform future studies of this type, we evaluated whether there were proteotypic peptides (those seen consistently for a protein of interest) in our samples (supplementary table 2).

3.5.1. Characteristics of peptides identified in this study

Proteins were digested to peptides by trypsin and the mass of peptides were measured in the mass spectrometer. The peptides were fragmented to characteristic ion series in the MS/MS spectra. The Mascot search program was used to identify peptides whose calculated precursor masses fall into a suitable window around the measured mass. The Scaffold software can statistically analyse all data from the mass spectrometric analyses.

The charge distribution of identified peptides was shown below. About 69% of the tryptic peptide precursors were doubly charged, and 28% were triply charged. Only 3% were quadruply or higher charged. Any singly charged ions were excluded (figure 22).

~ 58 ~

Results

Figure 22. Properties of identified MS/MS spectra. Mass and charge distributions of the identified peptides. Around 69% of the tryptic peptide precursors were doubly charged, and 28% were triply charged. Only 3% were quadruply or higher charged. Any singly charged ions were excluded

Figure 23. Distribution of identified peptide by mass. About 5% of peptides were below 800Da. 18% of peptides mass were 1100 or 1200 Da and 70 % of peptides were between 900 to 1700 Da. 7% of peptide mass between 1800 to 1900 were detectable.

~ 59 ~

Results

The mass distribution can also be summarised. About 5% of peptides were below 800Da. 18% of peptides mass were 1100 or 1200 Da and 70 % of peptides were between 900 to 1700 Da. 7% of peptide mass between 1800 to 1900 were detectable (figure 23). More than 95% of all identified peptides have a tandem spectrum that contains a partial sequence of at least seven amino acids. Longer amino acid sequences may be desired for homology search, however a large proportion of peptides are of a size that makes them possible candidates for proteotypic peptides.

3.5.2. Theoretical proteotypic peptides

To detect the theoretical proteotypic peptides of identified secreted proteins, the PeptideSieve

(version 0.51, released 5/2008) was used. The parameters for PeptideSieve were P-values greater than 0.95, all experimental design selected and allowing up to one missed cleavage. The chemical property of amino acids were the default for that software. Table 7 shows an example of theoretical proteotypic peptides of leptin and adiponectin. All others are shown in supplementary table 3. A total of 6 peptides were suggested as good candidate proteotypic peptides for leptin. Nine good candidate proteotypic peptides were suggested for adiponectin.

The adiponectin fragment “GETGDVGMTGAEGPR”, was predicted as good candidate proteotypic peptides for gel-based and liquid-base mass spectrometry analysis. Moreover, the fragment “GTCAGWMAGIPGHPGHNGTPGR”, was suggested as useful for ICAT-labeling mass spectrometry analysis.

According to the statistical methods used in this software (Kolmogorov-Smirnov (KS) distance and the Kullbach-Leibler (KL) distance) [80], the PAGE-ESI and PAGE-MALDI share the same four of top five properties and should give a similar outcome. Yet, in the MUDPIT-ESI and

ICAT-ESI, the true-positive predictive value can be affected by charge issue. However, Mallick et al. [80] demonstrated the software has 90% positive predictive value from four mass spectrometers (MALDI, ESI, MUDPIT-ESI and ICAT-ESI) with 65 to 80 % coverage of

~ 60 ~

Results

proteotypic peptides tested.

In an analysis of amino acids from our experiment and theoretical proteotypic peptides (figure

24), cysteine was absent due to a lack of reduction and alkylation used in this experiment. The proportion of experimental peptides containing proline and glycine was lower than expected; for proline, this is probably because it can affect tryptic cleavage if adjacent to lysine or arginine.

12.00%

10.00%

8.00%

6.00%

4.00%

2.00%

0.00% FSTNKEYVQMCLAWPHDIRG

Matched PeptideSieve Experiment

Figure 24. Distribution of the number of amino acid components of proteotypic peptides. The number of amino acid were obtained from theoretical and experimental proteotypic peptide analysis. It is also analysis the matched peptides of amino acid components from theoretical and experimental process.

~ 61 ~

Results

Table 7. The example of theoretical proteotypic peptides of leptin. Lepin is a 19KDa, 167 amino acid sequence. Digestion with trypsin with cleavage sites of K and R. The threshold scores was greater than 0.95 and experiment type was PAGE with MALDI. Accession Peptide sequences Experiment designer Threshold score ADIPO_MOUSE GTCAGWMAGIPGHPGHNGTPGR PAGE_MALDI 0.992201 ADIPO_MOUSE GTCAGWMAGIPGHPGHNGTPGR ICAT_ESI 0.990833 ADIPO_MOUSE GDAGLLGPK PAGE_MALDI 0.965702 ADIPO_MOUSE GETGDVGMTGAEGPR PAGE_MALDI 0.970575 ADIPO_MOUSE GETGDVGMTGAEGPR MUDPIT_ESI 0.981542 ADIPO_MOUSE GEPGEAAYMYRS PAGE_MALDI 0.994055 ADIPO_MOUSE SAFSVGLETRV PAGE_MALDI 0.992096 ADIPO_MOUSE VTVPNVPIRF PAGE_MALDI 0.994523 ADIPO_MOUSE IFYNQQNHYDGSTGKF PAGE_MALDI 0.995445 ADIPO_MOUSE FYCNIPGLYYFSYHITVYMK PAGE_MALDI 0.98729 ADIPO_MOUSE AVLFTYDQYQEKN PAGE_MALDI 0.995036 LEP_RAT MCWRPLCRF PAGE_MALDI 0.997159 LEP_RAT FLWLWSYLSYVQAVPIHKV PAGE_MALDI 0.98878 LEP_RAT INDISHTQSVSARQ PAGE_MALDI 0.984587 LEP_RAT VTGLDFIPGLHPILSLSKM PAGE_MALDI 0.9942 LEP_RAT DLLHLLAFSKS PAGE_MALDI 0.99552 LEP_RAT LQGSLQDILQQLDLSPEC* PAGE_MALDI 0.99572

~ 62 ~

Results

3.5.3. Experimental proteotypic peptide detection

Having shown that proteotypic peptides could be predicted for many of the proteins analysed here, we then investigated whether peptides were repeatedly seen in our experimental analyses. A requirement was the identification of proteins/peptides by Scaffold 2.0™

(Proteome Software, Portland, release 2.0). Peptide identifications were accepted if they could be established at greater than 90.0% probability as specified by the PeptideProphetTM algorithm. To compile a proteotypic peptides list of identified proteins from a dynamic time course, proteins sharing common peptide sequences were considered unique when 1 unique peptide was found for that protein and observed at least two different time points. Only the highest score for proteins with similar peptide sequences was recorded.

Figure 25 presents an example of the proteotypic peptides of leptin and adiponectin. The red, blue, green and brown sequences show the number of peptides observed during a proteomics experiment. The red peptide sequences were only detected once but the blue, green and brown sequences were observed more than once during the experimental process.

In addition, the table 9 shows that the adiponectin peptide fragments, ‘AVLFTYDQYQEK’,’

IFYNQQNHYDGSTGK’ and ‘VTVPNVPIR can be reproduced not only from adipocytes cultured system but also from adipose tissue explants.

~ 63 ~

Results

Figure 25. An example of proteotypic peptides of leptin and adiponectin. The red, blue and green sequences were the number of peptides observed during a proteomics experiment. The red peptide sequences were only detected once but the blue, green and brown sequences were observed more than once during the experimental process.

~ 64 ~

Results

Table 8. An example of proteotypic peptides. Leptin was detected by LTQ-FT Ultra instrument. The protein identification probability was more than 99% and the highest sequence coverage was 34.1%. The ‘GLQKPESLDGVLEASLYSTEVVALSR’ and ‘INDISHTQSVSAR’ sequences were the most common sequences in the detectable peptides from FT-ICR MS/MS at tissue explants system only. Sample Accession numbers Identification probability Sequence Peptide sequence coverage

Tissue/Day9 LEP_RAT 100.00% 34.10% GLQKPESLDGVLEASLYSTEVVALSR Tissue/Day9 LEP_RAT 100.00% 34.10% INDISHTQSVSAR Tissue/Day9 LEP_RAT 100.00% 34.10% VTGLDFIPGLHPILSLSK Tissue/Day8 LEP_RAT 99.80% 23.40% GLQKPESLDGVLEASLYSTEVVALSR Tissue/Day8 LEP_RAT 99.80% 23.40% INDISHTQSVSAR Tissue/Day10 LEP_RAT 99.10% 7.78% INDISHTQSVSAR

Table 9. An example of proteotypic peptides. adiponectin was detected by LTQ-FT Ultra instrument. The protein identification probability was more than 99% and the highest sequence coverage was 14.6%. The ‘AVLFTYDQYQEK’,’ IFYNQQNHYDGSTGK’ and ‘VTVPNVPIR’ sequences were the most common sequences in the detectable peptides from FT-ICR MS/MS and two culture systems. Sample Accession numbers Identification probability Sequence Peptide sequence

coverage

Cell/Day7 ADIPO_MOUSE 99.80% 10.90% AVLFTYDQYQEK

Tissue/Day14 ADIPO_MOUSE 100.00% 14.60% AVLFTYDQYQEK

TissueDay18 ADIPO_MOUSE 100.00% 14.60% AVLFTYDQYQEK

TissueDay5 ADIPO_MOUSE 99.30% 4.86% AVLFTYDQYQEK

Cell/Day7 ADIPO_MOUSE 99.80% 10.90% IFYNQQNHYDGSTGK

Cell/Day1 ADIPO_MOUSE 99.80% 9.72% IFYNQQNHYDGSTGK

~ 65 ~

Results

Sample Accession numbers Identification probability Sequence Peptide sequence

coverage

Cell/Day3 ADIPO_MOUSE 99.80% 9.72% IFYNQQNHYDGSTGK

Tissue/Day14 ADIPO_MOUSE 100.00% 14.60% IFYNQQNHYDGSTGK

Tissue/Day18 ADIPO_MOUSE 100.00% 14.60% IFYNQQNHYDGSTGK

Cell/Day1 ADIPO_MOUSE 99.80% 9.72% VTVPNVPIR

Cell/Day3 ADIPO_MOUSE 99.80% 9.72% VTVPNVPIR

Tissue/Day14 ADIPO_MOUSE 100.00% 14.60% VTVPNVPIR

Tissue/Day18 ADIPO_MOUSE 100.00% 14.60% VTVPNVPIR

~ 66 ~

Results

Comparing theoretical and experimental proteotypic peptides of leptin and adiponectin, 5 good candidates (table 10) are likely to be useful as standard markers for leptin and adiponectin in mass spectrometry analysis.

Table 10. The matched peptides of leptin from theoretical and experimental peptides analysis. ‘INDISHTQSVSARQ’ and ‘VTGLDFIPGLHPILSLSKM’ are in accordance with the all peptide analysis. Accession number Peptide sequences ADIPO_MOUSE AVLFTYDQYQEKN ADIPO_MOUSE IFYNQQNHYDGSTGKF ADIPO_MOUSE VTVPNVPIRF LEP_RAT INDISHTQSVSARQ LEP_RAT VTGLDFIPGLHPILSLSKM

3.5.4. Selected reaction monitoring (SRM) method prediction

The software “Skyline” can be used to analyse experimental proteomics data to produce transition ion lists for selected reaction monitoring [51]. Identified secreted protein peptide sequences can be built as

FASTA files and used in the Skyline software.

The peptide settings for Skyline analysis were Trypsin enzyme digestion, no modification experimental design selected and allowing up to one missed cleavage. The transition settings were

Thermo mass spectrometer, 50 to 1,800 mass range, 2+ precursor charge, y-type ion and picking up 3 production ions.

The result of this analysis showed 3 production ions from the experimental proteotypic peptides

(supplementary table 4). The tryptic peptides fragments and y-type ions were determined as shown in table 11 and figure 26 and 27 from leptin and adiponectin. The peptide fragments of leptin were found with m/z 714 and m/z 954, y7,y8, y9, y10 and y11 ions, corresponding to the amino acid sequences

INDISHTQSVSAR and VTGLDFIPGLHPILSLSK. The peptide fragments of adiponectin were found with m/z 497, m/z 752, m/z 886, y3 and y5 to y11 ions, corresponding to the amino acid sequences

~ 67 ~

Results

VTVPNVPIR, AVLFTYDQYQEK and IFYNQQNHYDGSTGK. The y-ions were detected by

FT-ICR mass spectrometry and those ions were matched with the results form Skyline software. All y-ions predicted by Skyline were obtained in FT-ICR mass spectrometry experiments.

Table 11. The list of production ions of leptin and adiponectin. The two groups of peptides were in accordance with experimental proteotypic peptides. The production ions (y-ions) were obtained from Skyline software.

Accession numbers Peptide sequence Precursor m/z Precursor m/h Y-ion

LEP_RAT INDISHTQSVSAR 714.365504 748.394792 y7 LEP_RAT INDISHTQSVSAR 714.365504 885.453704 y8 LEP_RAT INDISHTQSVSAR 714.365504 972.485733 y9 LEP_RAT VTGLDFIPGLHPILSLSK 954.053669 1064.646256 y10 LEP_RAT VTGLDFIPGLHPILSLSK 954.053669 1161.69902 y11 LEP_RAT VTGLDFIPGLHPILSLSK 954.053669 757.481817 y7 LEP_RAT VTGLDFIPGLHPILSLSK 954.053669 1007.624792 y9 ADIPO_MOUSE VTVPNVPIR 497.805834 385.25578 y3 ADIPO_MOUSE VTVPNVPIR 497.805834 598.367121 y5 ADIPO_MOUSE VTVPNVPIR 497.805834 695.419885 y6 ADIPO_MOUSE VTVPNVPIR 497.805834 794.488299 y7 ADIPO_MOUSE AVLFTYDQYQEK 752.869556 810.362824 y6 ADIPO_MOUSE AVLFTYDQYQEK 752.869556 973.426152 y7 ADIPO_MOUSE IFYNQQNHYDGSTGK 886.405357 978.427549 y9 ADIPO_MOUSE AVLFTYDQYQEK 752.869556 1074.473831 y8 ADIPO_MOUSE IFYNQQNHYDGSTGK 886.405357 1106.486127 y10 ADIPO_MOUSE IFYNQQNHYDGSTGK 886.405357 1234.544704 y11

~ 68 ~

Results

Figure 26. The mass spectrum of adiponectin and leptin. a) The fragmentation pattern for the amino acid sequence VTVPNVPIR of adiponectin with the corresponding b and y ion series. b) The fragmentation pattern for the amino acid sequence AVLFTYDQYQEK of adiponectin with the corresponding b and y ion series.c) The fragmentation pattern for the amino acid sequence IFYNQQNHYDGSTGK of adiponectin with the corresponding b and y ion series.

~ 69 ~

Results

Figure 27. The mass spectrum of adiponectin and leptin. d) The fragmentation pattern for the amino acid sequence INDISHTQSVSAR of leptin with the corresponding b and y ion series. e) The fragmentation pattern for the amino acid sequence VTGLDFIPGLHPILSLSK of leptin with the corresponding b and y ion series.

~ 70 ~

Discussion

4. Discussion

4.1. Gel and label-free secretomics strategies in serum-free adipocytes and

adipose explants

Adipocytes are known to send cell to cell signals between metabolic organs like brain, skeletal muscle and liver. Adipose tissue also contains connective tissue matrix, nerve tissue, stromal vascular cells, and immune cells. Adipose tissue not only responds to afferent signals from traditional hormone systems and the central nervous system but also expresses and secretes factors with important endocrine functions that link obesity, cardiovascular and cancers. Moreover, adipose tissue can store glucose and lipid which is related to the metabolic syndrome.

The vast majority of proteomics investigations have centered on the study of adipose tissue from lean, obese models or molecular and physiological biology of adipocyte primary cultures or cell lines by inducing adipogenesis with adipogenic differentiation medium with a mixture of insulin, triiodothyronine, cortisol and a PPAR γ agonist. These cell lines display a differentiation capacity up to 90% and can retain this capacity over at least 30 generations [76]. However, these cell lines have been transformed or originate from adipose tissue tumor [81] and can have differences in the adipokine expression, regulation and response as compared to normal tissue. In addition, major adipokines have been shown to be produced by adipocytes, also other cell types that are involved in

~ 71 ~

Discussion

adipose tissue produce adipokines or influence production of adipokines by adipocytes [74]. It is important to study the secretion factors of adipocytes or adipose tissue to understand the fundamental mechanisms of lipid metabolism and the development of obesity and obesity-related disease. Since adipose tissue is a complex organ with several cell types, the analysis of the different cellular subtypes will give a clearer picture of the molecular network. Adipocyte culture studies are mainly limited to plasma and other body fluid. Therefore, the studies of adipose tissue explants are necessary.

Study design is an important subject. A major issue in characterizing the adipose or the adipocyte secretome is that protein composition of the culture media is highly dependent on the way the culture is performed. Approaches should reduce experimental and biological complexity. Zvonic et al. [69] have suggested that leakage or cell death my account for any other proteins detected in conditioned medium. Therefore, we established a serum-free culture protocol with dilution that minimizes contamination of secretome and to avoid proteins/peptides produced by external serum medium in order to obtain a high quality sample of secretome analysis.

Under this protocol, no proteins are added by the extracellular culture medium. Choi et al [82] demonstrated that α2 macroglobulin was derived from extracellular culture medium in preadipocyte culture with 10% fetal calf serum but from our results point out that the housekeeping proteins, α1 and

~ 72 ~

Discussion

β2 macroglobulin proteins had been de novo synthesized and identified. Serum-free culture protocol also maintains the cell at one stage and this allows us to use a pre-obese static phase for study.

Plasma protein contamination is a key issue for secreted protein analysis. The majority of plasma proteins are produced by the liver and are found throughput the mammalian body, including the intracellular space [63]. The other proteins in plasma result from tissue leakage and cell break down.

In order to address this issue, several methodologies have been proposed including sample preparation

(conditioned medium collection), sample concentration (phenol extraction following by TCA precipitation) and sample separation (chromatography) [77]. In this study, we used serum-free media and a dilution process to reduce any plasma protein contamination from tissues or cells. With fresh medium added in each time point, the concentration of plasma protein was diluted, making it possible for low abundance secreted proteins to be identified by MS analysis. Once the proteins were detected as secreted, bioinformatic tools were applied to improve the accuracy of classification results [27]. The validity of our procedure was demonstrated by the detection of several known adipocyte-derived secreted proteins: complement C3, adiponectin, leptin, resistin, plasminogen activator inhibitor-1, collagen type VI α3 and adipsin, collagens, matrix metalloproteinase-2, metalloproteinase inhibitor 2, haptoglobin and a housekeeping protein β2 macroglobulin.

~ 73 ~

Discussion

Some proteins were identified as adipocyte secreted factors that have previously been reported as being secreted by other cell systems. However, numbers of those proteins were revealed in our culture system. For example, galectin 1from both culture systems is reported here. Moreover, galectin 3, a superfamily of galectin and procollagen C endopeptidase enhancer 1can be detected from adipose tissue explants and Stress-70 protein can only be detected from adipocyte culture. Galectin is expressed in a wide variety of cells and tissue. Those suggest that secretion profiles were determined indicating a dynamic environment including an actively remodeling ECM and several secreted protein involved in growth regulation.

In the lists of protein we identified, many related to signal pathway and cell-cell communication.

Furthermore, our results include low abundance cell membrane proteins and a large number of secreted and extracellular matrix proteins. Finally, we also identified the number of novel classical and non-calssical secreted proteins from adipocyte and adipose tissue explants that have not previously been associated with adipogenesis, which include regulatory factors and cell-to-cell communicators.

The differential regulation of these proteins during adipogenesis suggests that they might participate in the regulation of adipocyte adipogenic differentiation and body metabolism.

~ 74 ~

Discussion

Our findings cover a broad range of the secretome. A total 281 proteins were found from adipocytes culture and adipose tissue explants system. Many of these concur with those of previous reports. In total, 36% of adipocyte secretome consists of uniquely identified protein. 22% of adipocyte secretome proteins had been identified in the analysis of human plasma proteins which can cross-talk between different organs or cell types via blood vessel. 60% of the proteins identified in the adipocyte secretome had been identified in proteomics analysis. Furthermore, 31% of adipose tissue explants secretome consists of uniquely identified protein. 12% of adipose tissue explants secretome proteins had been identified in the analysis of human plasma proteins. 68% of the proteins identified in the adipose tissue explants secretome had been identified in proteomics analysis.

Our analysis detected low abundance regulators like leptin. Leptin was absent in some other survey

(e.g. Asachi et al. [83]), probably due to low quality detection tools. However, we did not observe interleukin 6 or tumor necrosis factor-α. This may have been due to low abundance in the medium of adipose tissue explants or adipocytes culture and occur below the detection level of our system.

Our findings indicate that some cellular proteins might be shedded by non-classical secretion pathway.

Heat shock protein 70, adipocyte fatty acid-binding protein and annexin A2 have been demonstrated as non-classical secreted proteins and their pathway were involved in the formation of lipid rafts[84],

~ 75 ~

Discussion

pro-oxidative conditions [85] or in cancer signaling pathway [86]. Other proteins include cytochrome b5 and Heme oxygenase 1 from adipose tissue explants. SecretomeP analysis of those proteins was positive. TMHMM suggested presence of a transmembrane helix. Indeed, those proteins have been reported by Swiss-Prot classification as a microsome or endoplasmic reticulum proteins. However, cytochrome b5 has been described as a secreted protein participates in a variety of metabolic reactions, including cytochrome P450-mediated hydroxylation reactions and drug metabolism, cholesterol anabolism, fatty acid elongation and desaturation [87]. Miyake et al. [88] documented that heme oxygenase 1 plays an important role as a secreted protein in angiogenesis in both normal and cancerous cells.

In this project, it is very difficult to avoid the lipid droplet contamination. However, recent study demonstrate that lipid droplets contain abundant copies of histones H2A, H2Av and H2B and have revealed a novel role as maternally supplied proteins for intracellular organelles [89]. Histone H2A can be found in both culture systems and Histone H2Bcan be only found in adipose tissue explants.

Bioinformatic tools were used to analyse and point out they are as non-classical secreted proteins.

We also sought to examine any changes of the secreted proteins of lean adipose explants and adipocyte primary cultures under serum-free culture conditions using a label-free proteomic approach. Our

~ 76 ~

Discussion

findings covered a broad dynamic range of the secretome which might indicate the inhibition and enhancement of proteins expression during the culture.

A semi-quantitative analysis was undertaken of secreted proteins using spectral counting. Although this approach is less accurate than absolute quantitation, the replicates provided by the adjacent time points could classify the proteins into 5 groups. This suggests that increased levels of protein during culture could be dependent upon an inducing signal. When the inducing signal is removed, the level of protein diminishes to its basal level. The second reason of increased levels of protein is transient even in the continued presence of the regulatory signal. After the regulatory signal has terminated and the cell has been allowed to recover, a second transient response to a subsequent regulatory signal may be observed. The third reason of increased levels of protein is the regulatory signal. The increased levels persist indefinitely even after termination of the signal. Due to the experimental design, one of the reasons of decreased of protein levels is dilution. The other reason of protein decreased is protein degradation.

The interaction between leptin and adiponectin is present as ping-pong reaction. The figure 21b indicated that while increase the level of adiponection at day 1 to day 4, the level of leptin will decrease from around 2 at day 1 to zero at day 4. When the level of adiponectin is higher than 10, the

~ 77 ~

Discussion

level of leptin is almost absent. The interaction is complex. There are other secreted factors that can also affect the interaction of leptin and adiponectin such as insulin-like growth factor binding protein-3

[78]. Insulin-like growth factor binding protein 3 can inhibit the adiponectin expression. Figure 28 shows the dynamic change between adiponectin and insulin-like growth factor-binding protein 3.

While the level of insulin-like growth factor-binding protein 3 slight increased from day 9, the level of adiponectin was slight decreasing from day 9 to day 11.

18 16 14 Adiponectin 12 10 8 Insulin-like growth factor- 6 binding protein 3 4 Normalised 2 0 1 2 3 4 5 6 8 9 10111314151618 Time (Day)

Figure 28. The protein expression of insulin-like growth factor-binding protein 3 and adiponectin. While the level of insulin-like growth factor-binding protein 3 slight increased from day 9, the level of adiponectin was slight decreasing from day 9 to day 11.

The use of proteomics has broadened our knowledge of adipogenesis and adipocyte biology.

Furthermore, the application of FT-ICR MS-based technology is making striking progress in the identification of novel and low-abundance proteins secreted by adipose tissue and adipocytes. These

~ 78 ~

Discussion

findings may be immensely useful in relation to the crucial role of adipokines in signaling, modulation of metabolism and energy homeostasis across different organs in the body.

Adipose tissue which contains adipocytes, macrophages, precursor adipocyte cells and endothelial cells can lead to identification of obesity-related proteins as the physiological situation is better than primary adipocytes. Moreover, in order to understand the signal between adipose-drived stem cell and preadipocyte, the adipose tissue is good module to analyse and discover what secreted proteins may be produced by stem cells. Finally, only little amount of adipose tissue can produce enough amount of secreted protein which can be detected by FT-ICR.

Early insight into the adipose tissue secretome will contribute to a better understanding of its role in energy metabolism and related diseases and may lead to the discovery of unknown peptides/proteins involved in regulation of energy metabolism and new targets for obesity therapy.

The current study has characterized the shared proteomics feature of adipose tissue explants and adipocytes culture. Because gender is known to influence the metabolic characteristics of adipose tissue [90], further proteomics analysis will need to focus on comparison with two group (male and female) from rat.

~ 79 ~

Discussion

4.2. New proteins detected in this study

Our method confirmed 3 novel secreted proteins from adipocyte cultures (afamin, seminal vesicle secretory protein 2 and xanthine dehydrogenase/oxidase) and 10 novel secreted proteins from adipose tissue explants such as ceruloplasmin, chymase, matrix gla protein, biglycan, serine protease inhibitor superfamily (A3K, A3L, A3M and A3N), and SPARC-like protein.

Afamin is a 60 KDa glycoprotein and nexin belongs to the albumin gene family. According to

Swiss-Prot, afamin is usually related to protein transport. Recently studies demonstrated that afamin is a part of Vitamin E-Binding and a part of lipoprotein protein binding [91-93]. Afamin has been patented as a diagnostic marker for metabolic syndrome disease such as obesity or insulin resistance

[94] or as a biomarker for several type of cancers [91]. However, its molecular association with obesity is still unclear.

Seminal vesicle secretory protein 2 is a 40KDa copulatory plug protein. Its physiological function can be classified as reproduction [95]. Ci et al. [96] have recently shown that the mRNA of seminal vesicle secretory proteins was highly expressed in adipose tissue from prostate transcriptome and also verified that the prostate has unique roles in polyamine biosynthesis and transport as well as immunity. We hypothesize that seminal vesicle secretory protein 2 may be as a secreted factor to stimulate the

~ 80 ~

Discussion

prostate to secrete small molecules, protein and nutrients that may contribute to the development of prostate disease states such as benign hyperplasia and cancer. Xanthine dehydrogenase/oxidase is a

133 KDa enzyme for oxidation of purine metabolites to uric acid. It has also been reported as a novel regulator of adipogenesis [97] associated with the regulation of fat accretion. However, Cheung et al.

[97] observed the xanthine dehydrogenase/oxidase as a regulator of nuclear-hormone receptor activity.

The function of this enzyme in the extracellular space is still unknown.

Recently research has been confirmed that the high ceruloplasmin expression was significantly associated with higher BMI through the 2-DE analysis and ELISA assay [98]. Yet, the mechanism by which ceruloplasmin is related to obesity has not been identified. It is possible that ceruloplasmin is connected with inflammatory pathway and linked to obesity.

The function of chymase is to impair the function of phospholipid transfer protein in reverse cholesterol transport [99]. It is also suggested that chymase may play a role in reproductive processes, in lipid and lipoprotein metabolism in obesity and in neurodegenerative disease [100].

Matrix gla protein contains the modified amino acid, carboxyglutamic acid (Gla) which is produced in a posttranslational carboxylation reaction requiring vitamin K as a cofactor and involved in

~ 81 ~

Discussion

extracellular matrix calcification [101, 102]. Studies suggest that matrix gla protein is regulated in a very controlled manner during tissue development.

Biglycan is a small leucine-rich proteoglycan present in bone and cartilage extracellular matrix. It can be co-localize with apolipoprotein B in adipose tissue and may have a major role in atherosclerosis development [103-105].

SPARC-like, a member of the SPARC family of extracellular matrix proteins is expressed in the human central nervous system. Functional assays suggest that SPARC-like protein may serve as an antagonist of cell adhesion. The effects of SPARC-like protein on cell adhesion are of particular interest because the SPARC-like protein is down-regulated in many types of cancer cells and may assist as a negative regulator of cell growth and proliferation.

Less obesity related studies have studied the serine protease inhibitor superfamily (A3K, A3L, A3M and A3N), yet they are all involved in complement system and inflammation [106, 107]. Nevertheless, from the ArrayExpression, these superfamily genes are over-expression on brain, liver, digestion system component and adipose tissue derived stem cell [108]. The data set also show these genes may be involved in breast cancer [109, 110].

~ 82 ~

Discussion

4.3. Clinical application of our results

The ELISA test is commonly used for the purpose of clinical diagnosis. For example, Allred et al. [111] have reported a modified ELISA method to measure the quantitative of CD36 in adipose tissue. In addition, a method for diagnosing the metabolic syndrome [94] has been patented. In common, the quantitative level of target protein for clinical diagnosis is dependence on the sample source. Therefore, the body fluids like blood, plasma, serum, cerebrospinal fluid, sperm fluid, follicular fluid are generally used. The other issue is reference value from healthy individual or individual have not having the symptom of disease on clinical applications.

This project has made inroads into the development of a quick, reliable and more quantitative method for measuring clinical sample content that would be applicable to a large numbers of samples. The approach we describe here is to create a high-throughput MS-based platform for analysis of secreted proteins. The automated, nano-loading method could be applied to large numbers of samples and would allow a robust comparison between assays. It should be noted, however, that replicate analyses would always be required.

Secondly, we also point out that most of the frequently observed peptides showed potential as candidate proteotypic peptides of target protein. From a homology search, these peptides are able to be

~ 83 ~

Discussion

used on cross-species protein identification. This overcomes the species cross-reactivity issues of antibodies used in animal-based obesity studies. In addition, the candidate proteotypic peptides can replace the production of monoclonals, which are required for Western blotting and ELISA tests.

Thirdly, with the technology of mass spectrometer instrument development, the sensitivity and accuracy are all increasing. Target proteins, including post translation modification proteins can be found with or without any treatment (top-down or bottom up) by proteomics approach and correlate with their right molecular weight. This can overcome the post translation modification effect and right molecular weight issues associated with western blotting or ELISA.

Fourthly, we also provide a SRM datasheet from mass spectrometry analysis combine with bioinformatic tool analysis. In order to avoid that a signal precursor ion to a product ion is not sufficiently specific to define a unique peptide [112], we combined the experimental proteotypic peptides and each 3 of theoretical y-ions of target peptide. Our SRM analysis datasheet shows the fragmentation patterns for the identified amino acid sequences match with the high-confidence, correspond y-ion series. With this datasheet, the y-ions can be set for specific m/z of target fragment and used for Q3 detection.

~ 84 ~

Discussion

Finally, FT-ICR is suitable for label-free proteomics experiment without adding any internal standards in peptides mixtures for calibration [113]. For the clinical application, label-free quantification approach can be used for absolute or relative quantitation and can reduce normalization difficultly from both data sets. In addition, our datasheet also supply a candidate proteotypic peptide for synthesizing internal standards peptide for absolute protein quantitation. We are aware that some issues still need to be addressed to increase the confidence of large-scale quantitative proteomics such as biological variation. In order to increase the number confidence from identified MS/MS spectra, robust statistical models would need to be applied. This requires substantial additional work on the development of statistical method [114].

In conclusion, this thesis has expanded our knowledge about the proteins which are secreted by adipocytes. It showed that tissue explants secrete a large number of proteins than isolated adipocytes and are thus useful for this type of study. Many of the secreted proteins have potential as biomarkers, and this thesis suggested some novel ways that these could be monitored. This has potential for early diagnosis of adipocyte-associated disease and could also be useful to monitor any therapeutic strategy.

~ 85 ~

References

5. References

1. Henegar C, Tordjman J, Achard V, Lacasa D, Cremer I, Guerre-Millo M, Poitou C, Basdevant A, Stich V, Viguerie N, et al: Adipose tissue transcriptomic signature highlights the pathological relevance of extracellular matrix in human obesity. Genome Biol 2008, 9:R14. 2. Mutch DM, Rouault C, Keophiphath M, Lacasa D, Clement K: Using gene expression to predict the secretome of differentiating human preadipocytes. Int J Obes (Lond) 2009, 33:354-363. 3. obesity and overweight [http://www.who.int/mediacentre/factsheets/fs311/en/index.html] 4. Wang P, Mariman E, Renes J, Keijer J: The secretory function of adipocytes in the physiology of white adipose tissue. J Cell Physiol 2008, 216:3-13. 5. Billon N, Monteiro MC, Dani C: Developmental origin of adipocytes: new insights into a pending question. Biol Cell 2008, 100:563-575. 6. Antuna-Puente B, Feve B, Fellahi S, Bastard JP: Adipokines: the missing link between insulin resistance and obesity. Diabetes Metab 2008, 34:2-11. 7. Xie L, O'Reilly CP, Chapes SK, Mora S: Adiponectin and leptin are secreted through distinct trafficking pathways in adipocytes. Biochim Biophys Acta 2008, 1782:99-108. 8. Pavlou MP, Diamandis EP: The cancer cell secretome: A good source for discovering biomarkers? J Proteomics 2010. 9. Poulos SP, Hausman DB, Hausman GJ: The development and endocrine functions of adipose tissue. Mol Cell Endocrinol 2010, 323:20-34. 10. Arner P, Spalding KL: Fat cell turnover in humans. Biochem Biophys Res Commun 2010, 396:101-104. 11. Lang K, Ratke J: Leptin and Adiponectin: new players in the field of tumor cell and leukocyte migration. Cell Commun Signal 2009, 7:27. 12. Galic S, Oakhill JS, Steinberg GR: Adipose tissue as an endocrine organ. Mol Cell Endocrinol 2009. 13. Graziani I, Doyle A, Sterling S, Kirov A, Tarantini F, Landriscina M, Kumar TK, Neivandt D, Prudovsky I: Protein folding does not prevent the nonclassical export of FGF1 and S100A13. Biochem Biophys Res Commun 2009, 381:350-354. 14. Dumic J, Dabelic S, Flogel M: Galectin-3: an open-ended story. Biochim Biophys Acta 2006, 1760:616-635. 15. Barcelo-Batllori S, Gomis R: Proteomics in obesity research. Proteomics Clinical Applications 2009, 3:263-278. 16. Rosen ED, MacDougald OA: Adipocyte differentiation from the inside out. Nature reviews Molecular cell biology 2006, 7:885-896. ~ 86 ~

References

17. Siwy J, Vlahou A, Zimmerli LU, Zurbig P, Schiffer E: Clinical proteomics: current techniques and potential applications in the elderly. Maturitas 2011, 68:233-244. 18. Piersma SR, Fiedler U, Span S, Lingnau A, Pham TV, Hoffmann S, Kubbutat MH, Jimenez CR: Workflow comparison for label-free, quantitative secretome proteomics for cancer biomarker discovery: method evaluation, differential analysis, and verification in serum. J Proteome Res 2010, 9:1913-1922. 19. Kos K, Harte AL, James S, Snead DR, O'Hare JP, McTernan PG, Kumar S: Secretion of neuropeptide Y in human adipose tissue and its role in maintenance of adipose tissue mass. Am J Physiol Endocrinol Metab 2007, 293:E1335-1340. 20. Ali AA, Weinstein RS, Stewart SA, Parfitt AM, Manolagas SC, Jilka RL: Rosiglitazone causes bone loss in mice by suppressing osteoblast differentiation and bone formation. Endocrinology 2005, 146:1226-1235. 21. Celis JE, Moreira JM, Cabezon T, Gromov P, Friis E, Rank F, Gromova I: Identification of extracellular and intracellular signaling components of the mammary adipose tissue and its interstitial fluid in high risk breast cancer patients: toward dissecting the molecular circuitry of epithelial-adipocyte stromal cell interactions. Mol Cell Proteomics 2005, 4:492-522. 22. Nickel W: The mystery of nonclassical protein secretion. A current view on cargo proteins and potential export routes. Eur J Biochem 2003, 270:2109-2119. 23. Bendtsen JD, Kiemer L, Fausboll A, Brunak S: Non-classical protein secretion in bacteria. BMC Microbiol 2005, 5:58. 24. Bendtsen JD, Nielsen H, von Heijne G, Brunak S: Improved prediction of signal peptides: SignalP 3.0. J Mol Biol 2004, 340:783-795. 25. Bendtsen JD, Jensen LJ, Blom N, Von Heijne G, Brunak S: Feature-based prediction of non-classical and leaderless protein secretion. Protein Eng Des Sel 2004, 17:349-356. 26. !!! INVALID CITATION !!! 27. Klee EW, Ellis LB: Evaluating eukaryotic secreted protein prediction. BMC Bioinformatics 2005, 6:256. 28. Sun G: Application of DNA microarrays in the study of human obesity and type 2 diabetes. OMICS 2007, 11:25-40. 29. Kondoh M, Fukada K, Shimada T, Kitamura Y, Yasueda H, Enomoto T: [Highly sensitive enzyme-linked immunosorbent assay for the quantification of allergens from Dermatophagoides, DER p 1 and DER f 1]. Arerugi 2010, 59:109-116. 30. Cravatt BF, Simon GM, Yates JR, 3rd: The biological impact of mass-spectrometry-based proteomics. Nature 2007, 450:991-1000. 31. LC-MS Analysis of Peptides [http://www.apl.ku.edu/lcms.htm] 32. Glish GL, Vachet RW: The basics of mass spectrometry in the twenty-first century. Nat ~ 87 ~

References

Rev Drug Discov 2003, 2:140-150. 33. Choi S, Kim J, Yea K, Suh PG, Ryu SH: Targeted label-free quantitative analysis of secretory proteins from adipocytes in response to oxidative stress. Anal Biochem 2010. 34. Zhou H, Xiao Y, Li R, Hong S, Li S, Wang L, Zeng R, Liao K: Quantitative analysis of secretome from adipocytes regulated by insulin. Acta Biochim Biophys Sin (Shanghai) 2009, 41:910-921. 35. Stevenson SE, Chu Y, Ozias-Akins P, Thelen JJ: Validation of gel-free, label-free quantitative proteomics approaches: applications for seed allergen profiling. J Proteomics 2009, 72:555-566. 36. Kline KG, Finney GL, Wu CC: Quantitative strategies to fuel the merger of discovery and hypothesis-driven shotgun proteomics. Brief Funct Genomic Proteomic 2009, 8:114-125. 37. Gingras AC, Aebersold R, Raught B: Advances in protein complex analysis using mass spectrometry. J Physiol 2005, 563:11-21. 38. Rix U, Superti-Furga G: Target profiling of small molecules by chemical proteomics. Nature chemical biology 2009, 5:616-624. 39. Lange V, Picotti P, Domon B, Aebersold R: Selected reaction monitoring for quantitative proteomics: a tutorial. Molecular systems biology 2008, 4:222. 40. Gerber SA, Rush J, Stemman O, Kirschner MW, Gygi SP: Absolute quantification of proteins and phosphoproteins from cell lysates by tandem MS. Proc Natl Acad Sci U S A 2003, 100:6940-6945. 41. Kettenbach AN, Rush J, Gerber SA: Absolute quantification of protein and post-translational modification abundance with stable isotope-labeled synthetic peptides. Nature protocols 2011, 6:175-186. 42. America AH, Cordewener JH: Comparative LC-MS: a landscape of peaks and valleys. Proteomics 2008, 8:731-749. 43. Grundy SM: Obesity, metabolic syndrome, and cardiovascular disease. J Clin Endocrinol Metab 2004, 89:2595-2600. 44. Grundy SM, Brewer HB, Jr., Cleeman JI, Smith SC, Jr., Lenfant C: Definition of metabolic syndrome: Report of the National Heart, Lung, and Blood Institute/American Heart Association conference on scientific issues related to definition. Circulation 2004, 109:433-438. 45. Grundy SM: What is the contribution of obesity to the metabolic syndrome? Endocrinol Metab Clin North Am 2004, 33:267-282, table of contents. 46. Ahrens CH, Brunner E, Basler K: Quantitative proteomics: a central technology for systems biology. J Proteomics 2010, 73:820-827. 47. Schiess R, Wollscheid B, Aebersold R: Targeted proteomic strategy for clinical biomarker discovery. Mol Oncol 2009, 3:33-44. ~ 88 ~

References

48. Mallick P, Schirle M, Chen SS, Flory MR, Lee H, Martin D, Ranish J, Raught B, Schmitt R, Werner T, et al: Computational prediction of proteotypic peptides for quantitative proteomics. Nat Biotechnol 2007, 25:125-131. 49. Kuster B, Schirle M, Mallick P, Aebersold R: Scoring proteomes with proteotypic peptide probes. Nat Rev Mol Cell Biol 2005, 6:577-583. 50. Cox J, Hubner NC, Mann M: How much peptide sequence information is contained in ion trap tandem mass spectra? J Am Soc Mass Spectrom 2008, 19:1813-1820. 51. MacLean B, Tomazela DM, Shulman N, Chambers M, Finney GL, Frewen B, Kern R, Tabb DL, Liebler DC, MacCoss MJ: Skyline: an open source document editor for creating and analyzing targeted proteomics experiments. Bioinformatics 2010, 26:966-968. 52. Deutsch EW, Lam H, Aebersold R: PeptideAtlas: a resource for target selection for emerging targeted proteomics workflows. EMBO Rep 2008, 9:429-434. 53. Turtoi A, Mazzucchelli GD, De Pauw E: Isotope coded protein label quantification of serum proteins--comparison with the label-free LC-MS and validation using the MRM approach. Talanta 2010, 80:1487-1495. 54. Tapp H, Deepe R, Ingram JA, Kuremsky M, Hanley EN, Jr., Gruber HE: Adipose-derived mesenchymal stem cells from the sand rat: transforming growth factor beta and 3D co-culture with human disc cells stimulate proteoglycan and collagen type I rich extracellular matrix. Arthritis Res Ther 2008, 10:R89. 55. Couttas TA, Raftery MJ, Bernardini G, Wilkins MR: Immonium ion scanning for the discovery of post-translational modifications and its application to histones. J Proteome Res 2008, 7:2632-2641. 56. Ly L, Wasinger VC: Peptide enrichment and protein fractionation using selective electrophoresis. Proteomics 2008, 8:4197-4208. 57. Elias JE, Gygi SP: Target-decoy search strategy for mass spectrometry-based proteomics. Methods Mol Biol 2010, 604:55-71. 58. Elias JE, Gygi SP: Target-decoy search strategy for increased confidence in large-scale protein identifications by mass spectrometry. Nat Methods 2007, 4:207-214. 59. Keller A, Nesvizhskii AI, Kolker E, Aebersold R: Empirical Statistical Model To Estimate the Accuracy of Peptide Identifications Made by MS/MS and Database Search. Analytical Chemistry 2002, 74:5383-5392. 60. Krogh A, Larsson B, von Heijne G, Sonnhammer EL: Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. Journal of molecular biology 2001, 305:567-580. 61. Kohl M, Wiese S, Warscheid B: Cytoscape: software for visualization and analysis of biological networks. Methods Mol Biol 2011, 696:291-303. 62. Maere S, Heymans K, Kuiper M: BiNGO: a Cytoscape plugin to assess overrepresentation ~ 89 ~

References

of gene ontology categories in biological networks. Bioinformatics 2005, 21:3448-3449. 63. Anderson NL, Anderson NG: The human plasma proteome: history, character, and diagnostic prospects. Mol Cell Proteomics 2002, 1:845-867. 64. Jensen LJ, Kuhn M, Stark M, Chaffron S, Creevey C, Muller J, Doerks T, Julien P, Roth A, Simonovic M, et al: STRING 8--a global view on proteins and their functional interactions in 630 organisms. Nucleic Acids Res 2009, 37:D412-416. 65. Ho E, Webber R, Wilkins MR: Interactive three-dimensional visualization and contextual analysis of protein interaction networks. J Proteome Res 2008, 7:104-112. 66. Anderson NL, Polanski M, Pieper R, Gatlin T, Tirumalai RS, Conrads TP, Veenstra TD, Adkins JN, Pounds JG, Fagan R, Lobley A: The human plasma proteome: a nonredundant list developed by combination of four separate sources. Mol Cell Proteomics 2004, 3:311-326. 67. Kratchmarova I, Kalume DE, Blagoev B, Scherer PE, Podtelejnikov AV, Molina H, Bickel PE, Andersen JS, Fernandez MM, Bunkenborg J, et al: A proteomic approach for identification of secreted proteins during the differentiation of 3T3-L1 preadipocytes to adipocytes. Mol Cell Proteomics 2002, 1:213-222. 68. Molina H, Yang Y, Ruch T, Kim JW, Mortensen P, Otto T, Nalli A, Tang QQ, Lane MD, Chaerkady R, Pandey A: Temporal profiling of the adipocyte proteome during differentiation using a five-plex SILAC based strategy. J Proteome Res 2009, 8:48-58. 69. Zvonic S, Lefevre M, Kilroy G, Floyd ZE, DeLany JP, Kheterpal I, Gravois A, Dow R, White A, Wu X, Gimble JM: Secretome of primary cultures of human adipose-derived stem cells: modulation of serpins by adipogenesis. Mol Cell Proteomics 2007, 6:18-28. 70. Chen X, Cushman SW, Pannell LK, Hess S: Quantitative proteomic analysis of the secretory proteins from rat adipose cells using a 2D liquid chromatography-MS/MS approach. J Proteome Res 2005, 4:570-577. 71. Wang P, Mariman E, Keijer J, Bouwman F, Noben JP, Robben J, Renes J: Profiling of the secreted proteins during 3T3-L1 adipocyte differentiation leads to the identification of novel adipokines. Cell Mol Life Sci 2004, 61:2405-2417. 72. Aoki N, Jin-no S, Nakagawa Y, Asai N, Arakawa E, Tamura N, Tamura T, Matsuda T: Identification and characterization of microvesicles secreted by 3T3-L1 adipocytes: redox- and hormone-dependent induction of milk fat globule-epidermal growth factor 8-associated microvesicles. Endocrinology 2007, 148:3850-3862. 73. Lim JM, Sherling D, Teo CF, Hausman DB, Lin D, Wells L: Defining the regulated secreted proteome of rodent adipocytes upon the induction of insulin resistance. J Proteome Res 2008, 7:1251-1263. 74. Alvarez-Llamas G, Szalowska E, de Vries MP, Weening D, Landman K, Hoek A, Wolffenbuttel BH, Roelofsen H, Vonk RJ: Characterization of the human visceral adipose ~ 90 ~

References

tissue secretome. Mol Cell Proteomics 2007, 6:589-600. 75. Chiellini C, Cochet O, Negroni L, Samson M, Poggi M, Ailhaud G, Alessi MC, Dani C, Amri EZ: Characterization of human mesenchymal stem cell secretome at early steps of adipocyte and osteoblast differentiation. BMC Mol Biol 2008, 9:26. 76. Rosenow A, Arrey TN, Bouwman FG, Noben JP, Wabitsch M, Mariman EC, Karas M, Renes J: Identification of novel human adipocyte secreted proteins by using SGBS cells. J Proteome Res 2010, 9:5389-5401. 77. Makridakis M, Vlahou A: Secretome proteomics for discovery of cancer biomarkers. J Proteomics 2010. 78. Kim HS, Ali O, Shim M, Lee KW, Vuguin P, Muzumdar R, Barzilai N, Cohen P: Insulin-like growth factor binding protein-3 induces insulin resistance in adipocytes in vitro and in rats in vivo. Pediatr Res 2007, 61:159-164. 79. Kito K, Ito T: Mass spectrometry-based approaches toward absolute quantitative proteomics. Current genomics 2008, 9:263-274. 80. Blonder J, Veenstra TD: Computational prediction of proteotypic peptides. Expert Rev Proteomics 2007, 4:351-354. 81. Fischer-Posovszky P, Newell FS, Wabitsch M, Tornqvist HE: Human SGBS cells - a unique tool for studies of human fat cell biology. Obes Facts 2008, 1:184-189. 82. Choi KL, Wang Y, Tse CA, Lam KS, Cooper GJ, Xu A: Proteomic analysis of adipocyte differentiation: Evidence that alpha2 macroglobulin is involved in the adipose conversion of 3T3 L1 preadipocytes. Proteomics 2004, 4:1840-1848. 83. Adachi J, Kumar C, Zhang Y, Mann M: In-depth analysis of the adipocyte proteome by mass spectrometry and bioinformatics. Mol Cell Proteomics 2007, 6:1257-1273. 84. Zhan R, Leng X, Liu X, Wang X, Gong J, Yan L, Wang L, Wang Y, Qian LJ: Heat shock protein 70 is secreted from endothelial cells by a non-classical pathway involving exosomes. Biochemical and biophysical research communications 2009, 387:229-233. 85. Grimsrud PA, Picklo MJ, Sr., Griffin TJ, Bernlohr DA: Carbonylation of adipose proteins in obesity and insulin resistance: identification of adipocyte fatty acid-binding protein as a cellular target of 4-hydroxynonenal. Mol Cell Proteomics 2007, 6:624-637. 86. Danielsen EM, van Deurs B, Hansen GH: "Nonclassical" secretion of annexin A2 to the lumenal side of the enterocyte brush border membrane. Biochemistry 2003, 42:14670-14676. 87. Mirzaei SA, Yazdi MT, Sepehrizadeh Z: Secretory expression and purification of a soluble NADH cytochrome b5 reductase enzyme from Mucor racemosus in Pichia pastoris based on codon usage adaptation. Biotechnol Lett 2010, 32:1705-1711. 88. Miyake M, Fujimoto K, Anai S, Ohnishi S, Kuwada M, Nakai Y, Inoue T, Matsumura Y, Tomioka A, Ikeda T, et al: Heme oxygenase-1 promotes angiogenesis in urothelial ~ 91 ~

References

carcinoma of the urinary bladder. Oncol Rep 2011, 25:653-660. 89. Brasaemle DL, Hansen JC: Developmental biology: holding pattern for histones. Curr Biol 2006, 16:R918-920. 90. DeLany JP, Floyd ZE, Zvonic S, Smith A, Gravois A, Reiners E, Wu X, Kilroy G, Lefevre M, Gimble JM: Proteomic analysis of primary cultures of human adipose-derived stem cells: modulation by Adipogenesis. Mol Cell Proteomics 2005, 4:731-740. 91. Dieplinger H, Ankerst DP, Burges A, Lenhard M, Lingenhel A, Fineder L, Buchner H, Stieber P: Afamin and apolipoprotein A-IV: novel protein markers for ovarian cancer. Cancer Epidemiol Biomarkers Prev 2009, 18:1127-1133. 92. Jerkovic L, Voegele AF, Chwatal S, Kronenberg F, Radcliffe CM, Wormald MR, Lobentanz EM, Ezeh B, Eller P, Dejori N, et al: Afamin is a novel human vitamin E-binding glycoprotein characterization and in vitro expression. J Proteome Res 2005, 4:889-899. 93. Jackson D, Craven RA, Hutson RC, Graze I, Lueth P, Tonge RP, Hartley JL, Nickson JA, Rayner SJ, Johnston C, et al: Proteomic profiling identifies afamin as a potential biomarker for ovarian cancer. Clin Cancer Res 2007, 13:7370-7379. 94. Dieplinger H, Kronenberg, Florian, Willeit, Johann, Kiechl, Stefan, Engel, Wolfgang: Method for Diagnosing the Metabolic Syndrome (MS) In United States Patent Application 20100260771; 10/14/2010 95. Argraves WS: Members of the low density lipoprotein receptor family control diverse physiological processes. Frontiers in bioscience : a journal and virtual library 2001, 6:D406-416. 96. Ci M, Mayumi Y, Andre B, Pascal B, Lin G, Yasukazu T, Fernand L, St-Amand J: Prostate-specific genes and their regulation by dihydrotestosterone. The Prostate 2008, 68:241-254. 97. Cheung KJ, Tzameli I, Pissios P, Rovira I, Gavrilova O, Ohtsubo T, Chen Z, Finkel T, Flier JS, Friedman JM: Xanthine oxidoreductase is a regulator of adipogenesis and PPARgamma activity. Cell metabolism 2007, 5:115-128. 98. Kim OY, Shin MJ, Moon J, Chung JH: Plasma ceruloplasmin as a biomarker for obesity: A proteomic approach. Clinical biochemistry 2011. 99. Albers JJ, Cheung MC: Emerging roles for phospholipid transfer protein in lipid and lipoprotein metabolism. Current opinion in lipidology 2004, 15:255-260. 100. Favari E, Gomaraschi M, Zanotti I, Bernini F, Lee-Rueckert M, Kovanen PT, Sirtori CR, Franceschini G, Calabresi L: A unique protease-sensitive high density lipoprotein particle containing the apolipoprotein A-I(Milano) dimer effectively promotes ATP-binding Cassette A1-mediated cell cholesterol efflux. J Biol Chem 2007, 282:5125-5132. 101. Gopalakrishnan R, Ouyang H, Somerman MJ, McCauley LK, Franceschi RT: Matrix gamma-carboxyglutamic acid protein is a key regulator of PTH-mediated inhibition of ~ 92 ~

References

mineralization in MC3T3-E1 osteoblast-like cells. Endocrinology 2001, 142:4379-4388. 102. van Beek EA, Bakker AH, Kruyt PM, Hofker MH, Saris WH, Keijer J: Intra- and interindividual variation in gene expression in human adipose tissue. Pflugers Archiv : European journal of physiology 2007, 453:851-861. 103. Mangat R, Su J, Scott PG, Russell JC, Vine DF, Proctor SD: Chylomicron and apoB48 metabolism in the JCR:LA corpulent rat, a model for the metabolic syndrome. Biochemical Society transactions 2007, 35:477-481. 104. Knippenberg M, Helder MN, Zandieh Doulabi B, Wuisman PI, Klein-Nulend J: Osteogenesis versus chondrogenesis by BMP-2 and BMP-7 in adipose stem cells. Biochemical and biophysical research communications 2006, 342:902-908. 105. King VL, Hatch NW, Chan HW, de Beer MC, de Beer FC, Tannock LR: A murine model of obesity with accelerated atherosclerosis. Obesity 2010, 18:35-41. 106. Hambrock HO, Nitsche DP, Hansen U, Bruckner P, Paulsson M, Maurer P, Hartmann U: SC1/hevin. An extracellular calcium-modulated protein that binds collagen I. The Journal of biological chemistry 2003, 278:11351-11358. 107. Kos K, Wilding JP: SPARC: a key player in the pathologies associated with obesity and diabetes. Nature reviews Endocrinology 2010, 6:225-235. 108. Shahdadfar A, Fronsdal K, Haug T, Reinholt FP, Brinchmann JE: In vitro expansion of human mesenchymal stem cells: choice of serum is a determinant of cell proliferation, differentiation, gene expression, and transcriptome stability. Stem Cells 2005, 23:1357-1366. 109. Lukk M, Kapushesky M, Nikkila J, Parkinson H, Goncalves A, Huber W, Ukkonen E, Brazma A: A global map of human gene expression. Nature biotechnology 2010, 28:322-324. 110. Chan MM, Lu X, Merchant FM, Iglehart JD, Miron PL: Gene expression profiling of NMU-induced rat mammary tumors: cross species comparison with human breast cancer. Carcinogenesis 2005, 26:1343-1353. 111. Allred CC, Krennmayr T, Koutsari C, Zhou L, Ali AH, Jensen MD: A novel ELISA for measuring CD36 protein in human adipose tissue. J Lipid Res 2011, 52:408-415. 112. Duncan MW, Yergey AL, Patterson SD: Quantifying proteins by mass spectrometry: the selectivity of SRM is only part of the problem. Proteomics 2009, 9:1124-1127. 113. Bruce JE, Anderson GA, Brands MD, Pasa-Tolic L, Smith RD: Obtaining more accurate Fourier transform ion cyclotron resonance mass measurements without internal standards using multiply charged ions. J Am Soc Mass Spectrom 2000, 11:416-421. 114. Nesvizhskii AI: A survey of computational methods and error rate estimation procedures for peptide and protein identification in shotgun proteomics. Journal of Proteomics 2010, 73:2092-2123.

~ 93 ~

Appendices

6. Appendices 6.1 The protein expression levels of all secreted and non-classical secreted proteins in adipocytes culture and adipose tissue explants. ġ Protein expression Accession Number Identified Proteins adipocytes cultured adipose tissue explants A1AT_RAT Alpha-1-antiproteinase D D A1I3_RAT Alpha-1-inhibitor 3 D D A1M_RAT Alpha-1-macroglobulin A D ACT2_MOLOC Actin, muscle-type C D ADIPO_MOUSE Adiponectin C B AEBP1_RAT Adipocyte enhancer-binding protein 1 B D AFAM_RAT Afamin D D ALBU_RAT Serum albumin C D ANXA2_RAT Annexin A2 D A AOC3_RAT Membrane primary amine oxidase N\A D APOA1_RAT Apolipoprotein A-I C N\A APOA4_RAT Apolipoprotein A-IV C D APOE_HUMAN Apolipoprotein E D B APOE_RAT Apolipoprotein E D D ARGI1_RAT Arginase-1 C D ASPG_RAT N(4)-(beta-N-acetylglucosaminyl)-L-asparaginase C D ATPA_RAT ATP synthase subunit alpha liver isoform, mitochondrial (Fragment) C D ~ 94 ~

Appendices

ġ Protein expression Accession Number Identified Proteins adipocytes cultured adipose tissue explants ATPD_RAT ATP synthase subunit delta, mitochondrial D D B2MG_RAT Beta-2-microglobulin D D BAP31_MOUSE B-cell receptor-associated protein 31 C D C1RA_MOUSE Complement C1r-A subcomponent A D CAH1_RAT Carbonic anhydrase 1 C D CAH2_RAT Carbonic anhydrase 2 A D CALM_RAT Calmodulin C D CALR_RAT Calreticulin N\A D CATB_RAT Cathepsin B C D CATL1_RAT Cathepsin L1 D N\A CCL2_RAT C-C motif chemokine 2 C E CD36_RAT Platelet glycoprotein 4 C B CERU_MOUSE Ceruloplasmin D D CERU_RAT Ceruloplasmin C D CES3_RAT Carboxylesterase 3 N\A B CFAD_RAT Complement factor D C D CH10_RAT 10 kDa heat shock protein, mitochondrial N\A D CH3L1_RAT Chitinase-3-like protein 1 C D CLOS_CLOHI Clostripain D C CLUS_RAT Clusterin D D

~ 95 ~

Appendices

ġ Protein expression Accession Number Identified Proteins adipocytes cultured adipose tissue explants CMA1_RAT Chymase N\A C CO1A1_BOVIN Collagen alpha-1(I) chain D D CO1A1_RAT Collagen alpha-1(I) chain C D CO1A2_RAT Collagen alpha-2(I) chain C D CO3_CAVPO Complement C3 C D CO3_RAT Complement C3 D D CO3A1_RAT Collagen alpha-1(III) chain D D CO4_RAT Complement C4 N\A D CO4A2_MOUSE Collagen alpha-2(IV) chain D D CO6A1_MOUSE Collagen alpha-1(VI) chain D D CO6A3_HUMAN Collagen alpha-3(VI) chain C D COEA1_MOUSE Collagen alpha-1(XIV) chain D D COF1_RAT Cofilin-1 N\A D COFA1_MOUSE Collagen alpha-1(XV) chain D D COTL1_MOUSE Coactosin-like protein N\A D CRP_RAT C-reactive protein D A CSPG2_RAT Versican core protein (Fragments) C D CYB5_RAT Cytochrome b5 C D CYTB_RAT Cystatin-B C D CYTC_RAT Cystatin-C B D

~ 96 ~

Appendices

ġ Protein expression Accession Number Identified Proteins adipocytes cultured adipose tissue explants DAG1_MOUSE Dystroglycan C D DLDH_RAT Dihydrolipoyl dehydrogenase, mitochondrial C D DPP2_RAT Dipeptidyl-peptidase 2 C D ENOA_RAT Alpha-enolase C D ENOG_RAT Gamma-enolase C D EST2_RAT Liver carboxylesterase 1 N\A D F10A1_RAT Hsc70-interacting protein C D FAAA_RAT Fumarylacetoacetase C D FABP4_MOUSE Fatty acid-binding protein, adipocyte C D FABP4_RAT Fatty acid-binding protein, adipocyte N\A C FBN1_HUMAN Fibrillin-1 D C FETUA_RAT Alpha-2-HS-glycoprotein C D FETUB_RAT Fetuin-B N\A D FIBA_RAT Fibrinogen alpha chain N\A D FIBB_RAT Fibrinogen beta chain C D FIBG_RAT Fibrinogen gamma chain N\A D FINC_RAT Fibronectin N\A D FRIH_RAT Ferritin heavy chain C D FSTL1_RAT Follistatin-related protein 1 C D G3P_RAT Glyceraldehyde-3-phosphate dehydrogenase C D

~ 97 ~

Appendices

ġ Protein expression Accession Number Identified Proteins adipocytes cultured adipose tissue explants GELS_RAT Gelsolin N\A D GLO2_RAT Hydroxyacylglutathione hydrolase N\A D GLU2B_MOUSE Glucosidase 2 subunit beta A C GPDA_RAT Glycerol-3-phosphate dehydrogenase [NAD+], cytoplasmic N\A C GRIFN_MOUSE Grifin N\A D GRP75_RAT Stress-70 protein, mitochondrial C C GRP78_RAT 78 kDa glucose-regulated protein C C GUAD_RAT Guanine deaminase C D H2A1C_RAT Histone H2A type 1-H N\A D H31_RAT Histone H3.1 C N\A HA12_RAT RT1 class I histocompatibility antigen, AA alpha chain C D HBA_RAT Hemoglobin subunit alpha-1/2 C C HEM2_RAT Delta-aminolevulinic acid dehydratase N\A D HEMO_RAT Hemopexin C C HMOX1_RAT Heme oxygenase 1 N\A C HPT_RAT Haptoglobin C C HSPB1_RAT Heat shock protein beta-1 C C IBP3_RAT Insulin-like growth factor-binding protein 3 C D IBP4_RAT Insulin-like growth factor-binding protein 4 N\A D IBP7_MOUSE Insulin-like growth factor-binding protein 7 C C

~ 98 ~

Appendices

ġ Protein expression Accession Number Identified Proteins adipocytes cultured adipose tissue explants IC1_RAT Plasma protease C1 inhibitor C C IGG2A_RAT Ig gamma-2A chain C region N\A D IGG2B_RAT Ig gamma-2B chain C region N\A E IGHG1_RAT Ig gamma-1 chain C region N\A E ITIH3_RAT Inter-alpha-trypsin inhibitor heavy chain H3 N\A C KACB_RAT Ig kappa chain C region, B allele C C LAMA2_MOUSE Laminin subunit alpha-2 N\A D LAMA4_MOUSE Laminin subunit alpha-4 C A LAMB1_MOUSE Laminin subunit beta-1 N\A B LAMB2_RAT Laminin subunit beta-2 C N\A LAMC1_MOUSE Laminin subunit gamma-1 N\A E LDHA_RAT L-lactate dehydrogenase A chain C C LDHB_RAT L-lactate dehydrogenase B chain C C LEG3_RAT Galectin-3 N\A D LEP_RAT Leptin N\A C LGMN_RAT Legumain C C LUM_RAT Lumican N\A D LYOX_RAT Protein-lysine 6-oxidase N\A C LYSC1_RAT Lysozyme C-1 N\A C MGP_RAT Matrix Gla protein N\A C

~ 99 ~

Appendices

ġ Protein expression Accession Number Identified Proteins adipocytes cultured adipose tissue explants MIF_RAT Macrophage migration inhibitory factor N\A C MMP2_RAT 72 kDa type IV collagenase C C MOES_RAT Moesin D C MUC18_RAT Cell surface glycoprotein MUC18 N\A D MUG1_RAT Murinoglobulin-1 N\A C MYP0_RAT Myelin P0 protein C C NID1_MOUSE Nidogen-1 (Fragment) C D NID1_RAT Nidogen-1 N\A C NID2_MOUSE Nidogen-2 C C NUCB1_RAT Nucleobindin-1 N\A C OSTP_RAT Osteopontin C C PAI1_RAT Plasminogen N\A C PCOC1_RAT Procollagen C-endopeptidase enhancer 1 C N\A PDCD5_MOUSE Programmed cell death protein 5 N\A C PDIA1_RAT Protein disulfide-isomerase A3 N\A N\A PDIA3_RAT Protein disulfide-isomerase A6 N\A C PDIA6_RAT Protein disulfide-isomerase A C PEBP1_RAT Phosphatidylethanolamine-binding protein 1 N\A C PEDF_HUMAN Pigment epithelium-derived factor N\A C PEDF_MOUSE Pigment epithelium-derived factor N\A C

~ 100 ~

Appendices

ġ Protein expression Accession Number Identified Proteins adipocytes cultured adipose tissue explants PGBM_HUMAN Basement membrane-specific heparan sulfate proteoglycan core protein N\A N\A PGBM_MOUSE Basement membrane-specific heparan sulfate proteoglycan core protein B B PGCP_RAT Plasma glutamate carboxypeptidase N\A C PGS1_RAT Biglycan N\A D PGS2_RAT Decorin C C PLMN_RAT Plasminogen activator inhibitor 1 C C PLSL_MOUSE Plastin-2 N\A C PNPH_RAT Purine nucleoside phosphorylase D C POSTN_MOUSE Periostin N\A C PRDBP_RAT Protein kinase C delta-binding protein N\A C PRDX1_RAT Peroxiredoxin-1 N\A C PRDX2_RAT Peroxiredoxin-2 N\A C PRDX5_RAT Peroxiredoxin-5, mitochondrial N\A C PROF1_RAT Profilin-1 N\A B PTMA_RAT Prothymosin alpha N\A C PTMS_RAT Parathymosin C N\A PTRF_RAT Polymerase I and transcript release factor C C PXDN_MOUSE Peroxidasin homolog N\A C QSOX1_RAT Sulfhydryl oxidase 1 N\A C RET4_RAT Retinol-binding protein 4 C N\A

~ 101 ~

Appendices

ġ Protein expression Accession Number Identified Proteins adipocytes cultured adipose tissue explants RL12_RAT 60S ribosomal protein L12 N\A C RLA2_RAT 60S acidic ribosomal protein P2 C C S10A4_RAT Protein S100-A4 N\A C S10A6_RAT Protein S100-A6 C N\A S10AB_RAT Protein S100-A11 C C SDPR_RAT Serum deprivation-response protein N\A C SODC_RAT Superoxide dismutase [Cu-Zn] C N\A SODE_RAT Extracellular superoxide dismutase [Cu-Zn] N\A C SODM_RAT Superoxide dismutase [Mn], mitochondrial N\A C SPA3K_RAT Serine protease inhibitor A3K N\A C SPA3L_RAT Serine protease inhibitor A3L N\A C SPA3M_RAT Serine protease inhibitor A3M (Fragment) C C SPA3N_RAT Serine protease inhibitor A3N N\A C SPRC_RAT SPARC C C SPRL1_RAT SPARC-like protein 1 N\A C SVS2_RAT Seminal vesicle secretory protein 2 N\A C SVS5_RAT Seminal vesicle secretory protein 5 N\A C TAGL2_HUMAN Transgelin-2 C N\A TCTP_RAT Translationally-controlled tumor protein N\A C THIO_RAT Thioredoxin domain-containing protein 5 N\A C

~ 102 ~

Appendices

ġ Protein expression Accession Number Identified Proteins adipocytes cultured adipose tissue explants TIMP1_RAT Metalloproteinase inhibitor 1 N\A C TPM1_RAT Tropomyosin alpha-1 chain N\A C TRFE_RAT Serotransferrin C N\A TSP1_MOUSE Thrombospondin-1 N\A N\A TTHY_RAT Transthyretin N\A C TXND5_MOUSE Thioredoxin N\A C UBIQ_RAT Ubiquitin C N\A VAT1_MOUSE Synaptic vesicle membrane protein VAT-1 homolog C N\A VIME_RAT Vimentin N\A C VTDB_RAT Vitamin D-binding protein N\A C

~ 103 ~

Appendices

6.2 Perl scritp

6.2.1 GO nodes

#/usr/bin/perl print "\r\ \r\ \r\ \r\ ",; open FILE, "520.txt"; $index = 0; %Name_index =(); while (){ chomp; my@array= ($Name,$ID,$color)=split(/\t/,$_); # Change this if more data has been integrated $Name_index{$Name}'N'.$index; print "\r\ \r\ \r\ \r\ \r\ ",; print set_Property("Name","$Name"); print set_Property("ID","$ID");

print "\r\n"; $index++; } close(FILE); sub set_Property { my($key,$value)=@_; return ("\n"); }

~ 104 ~

Appendices

6.2.2 GO edge #/usr/bin/perl # parse dataset into Edge entries in .xwg open FILE, "526.txt"; while() { chomp; my@array= ($Genename,$sport)=split(/\t/,$_); # Change this if more data has been integrated #@line = split(/\t/,$_); #$prot1 = $line[0]."\t".$line[1]; #$prot2 = $line[2]."\t".$line[3]; print "\r\ \r\ \r\ \r\ \r\ "; } print "\r\ "; sub set_Property { my($key,$value)=my@array; return ("\n"); }

~ 105 ~