High-throughput protein analysis using mass spectrometry ...757226/FULLTEXT01.pdf ·...

High-throughput protein analysis usingmass spectrometry-based methods

Tove Boström

KTH Royal Institute of TechnologySchool of Biotechnology

Stockholm 2014

c© Tove Boström 2014

KTH Royal Institute of TechnologySchool of BiotechnologyDivision of Protein TechnologyAlbaNova University centerSE-106 91 StockholmSweden

Paper I c© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, WeinheimPaper II c© 2012 American Chemical SocietyPaper IV c© 2014 American Chemical SocietyPaper V c© 2014 The American Society for Biochemistry and

Molecular Biology, Inc.

ISBN 978-91-7595-292-5TRITA-BIO Report 2014:15ISSN 1654-2312

Printed by Universitetsservice US-AB 2014

"If you can dream it, you can do it."

- Walt Disney

Abstract

In the field of proteomics, proteins are analyzed and quantified in high numbers. Proteinanalysis is of great importance and can for example generate information regarding proteinfunction and involvement in disease. Different strategies for protein analysis and quan-tification have emerged, suitable for different applications. The focus of this thesis lies onprotein identification and quantification using different setups and method developmenthas a central role in all included papers.

The presented research can be divided into three parts. Part one describes the develop-ment of two different screening methods for His6-tagged recombinant protein fragments.In the first investigation, proteins were purified using immobilized metal ion affinity chro-matography in a 96-well plate format and in the second investigation this was downscaledto nanoliter-scale using the miniaturized sample preparation platform, integrated selectiveenrichment target (ISET). The aim of these investigations was to develop methods thatcould work as an initial screening step in high-throughput protein production projects,such as the Human Protein Atlas (HPA) project, for more efficient protein production andpurification. In the second part of the thesis, focus lies on quantitative proteomics. Proteinfragments were produced with incorporated heavy isotope-labeled amino acids and usedas internal standards in absolute protein quantification mass spectrometry experiments.The aim of this investigation was to compare the protein levels obtained using quanti-tative mass spectrometry to mRNA levels obtained by RNA sequencing. Expression of32 different proteins was studied in six different cell lines and a clear correlation betweenprotein and mRNA levels was observed when analyzing genes on an individual level. Thethird part of the thesis involves the antibodies generated within the HPA project. In thefirst investigation a method for validation of antibodies using protein immunoenrichmentcoupled to mass spectrometry was described. In a second study, a method was developedwhere antibodies were used to capture tryptic peptides from a digested cell lysate withspiked in heavy isotope-labeled protein fragments, enabling quantification of 20 proteinsin a multiplex format. Taken together, the presented research has expanded the pro-teomics toolbox in terms of available methods for protein analysis and quantification in ahigh-throughput format.

Keywords: Proteomics, mass spectrometry, affinity proteomics, immunoenrichment,immunoprecipitation, IMAC, screening, protein production, protein purification, ISET,quantification, SILAC, stable isotope standard, antibody validation

I

Sammanfattning

I det forskningsområde som kallas proteomik studeras proteiner i stor skala, det vill sägamånga proteiner parallellt. Genom att studera proteiner kan man till exempel få informa-tion om proteiners funktion och huruvida ett protein är involverat i en sjukdomsprocess.I denna avhandling har proteiner analyserats och kvantifierats på olika sätt och i allainkluderade artiklar har metodutveckling haft en central roll.

Forskningen som presenteras i denna avhandling kan delas in i tre delar. Den första delenbeskriver två nya screeningmetoder för His6-taggade rekombinanta proteinfragment. I ettförsta projekt renades proteinerna med immobiliserad metalljonsaffinitetskromatografi iett 96-brunnsformat och i ett andra projekt skalades detta ner till nanoliter-format genomatt använda provberedningsplattformen ISET. Målet med dessa projekt var att utvecklametoder som skulle kunna användas som ett initialt screeningsteg vid storskalig protein-produktion för att öka effektiviteten. I del två ligger fokus på kvantifiering av proteiner.Proteinfragment producerades med aminosyror med inkorporerade tunga isotoper och an-vändes sedan som interna standarder i absolut kvantifiering med masspektrometri. Måletvar här att jämföra proteinnivåer med motsvarande mRNA-nivåer, vilka bestämts medRNA-sekvensering. Koncentrationen av 32 proteiner bestämdes i sex olika cellinjer ochdet fanns en tydlig korrelation mellan protein och mRNA när generna studerades indi-viduellt. I den sista delen av avhandlingen användes antikroppar som tagits fram inomproteinatlas-projektet (HPA). I ett projekt utvecklades en metod för validering av an-tikroppar genom att först använda antikropparna för anrikning av målproteinerna frånett cellysat och sedan med masspektrometri bestämma proteinernas identitet. I ett andraprojekt användes antikropparna istället för att fånga ut peptider från ett trypsinklyvtcellysat. I en setup där tunga isotopinmärkta proteinfragment adderades till lysatet föreklyvning kunde 20 proteiner kvantifieras parallellt. Sammantaget har de presenteradeprojekten bidragit till att möjligheterna för parallell proteinanalys och kvantifiering harökat.

II

List of publications

This thesis is based on the five publications listed below. The publicationsare referred to by Roman numbers (I-V) and are included in the Appendixof the thesis.

Paper I Tegel H.*, Yderland L.*, Boström T., Eriksson C., Ukko-nen K., Vasala A., Neubauer P., Ottosson J. and Hober S.Parallel production and verification of protein products us-ing a novel high-throughput method. Biotechnol J (2011)vol. 6 (8) pp. 1018-25

Paper II Adler B.*, Boström T.*, Ekström S., Hober S. and Lau-rell T. Miniaturized and automated high-throughput veri-fication of proteins in the ISET platform with MALDI MS.Anal Chem (2012) vol. 84 (20) pp. 8663-9

Paper III Boström T., Danielsson F., Lundberg E., Tegel H., Jo-hansson H. J., Lehtiö J., Uhlén M., Hober S., and TakanenJ. O. Investigating the correlation of protein and mRNAlevels in human cell lines using quantitative proteomicsand transcriptomics. Submitted

Paper IV Boström T., Johansson H. J., Lehtiö J., Uhlén M. andHober S. Investigating the applicability of antibodies gen-erated within the Human Protein Atlas as capture agentsin immunoenrichment coupled to mass spectrometry. JProteome Res (2014) vol. 13 (10) pp. 4424-35

Paper V Edfors F.*, Boström T.*, Forsström B., Zeiler M., Jo-hansson H., Lundberg E., Hober S., Lehtiö J., Mann M.and Uhlén M. Immuno-proteomics using polyclonal anti-bodies and stable isotope labeled affinity-purified recom-binant proteins. Mol Cell Proteomics (2014) vol. 13 (6)pp. 1611-24

*Both authors contributed equally to the work.

III

Published work not included in the thesis

Boström T.*, Nilvebrant J.* and Hober S. Purification systems based onbacterial surface proteins. Protein Purification, R. Ahmad (Ed.), (2012)ISBN: 978-953-307-831-1, InTech

*Both authors contributed equally to the work.

IV

Contributions to the included papers

Paper I Performed all experimental work and data analysis to-gether with coauthor. Wrote the manuscript together withcoauthors.

Paper II Performed all experimental work together with coauthor.Wrote the manuscript together with coauthors.

Paper III Performed all experimental work except cell cultivationand performed all MS data analysis. Wrote the manuscripttogether with coauthors.

Paper IV Performed all experimental work and all MS data analysis.Wrote the manuscript together with coauthors.

Paper V Performed cell cultivation, MS sample preparation andall MS data analysis. Produced and quantified all heavyisotope-labeled PrESTs. Contributed to the writing of themanuscript.

V

Abbreviations

2D-GE two-dimensional gel electrophoresisAb antibodyABD albumin binding domainABP albumin binding proteinACM antibody colocalization microarrayAPEX absolute protein expressionAQUA absolute quantificationAUC area under curveBAC bacterial artificial chromosomeBCA bicinchoninic acidcAb capture antibodyCDR complementary determining regionCIMS context-independent motif specificCR cross reactivitydAb detection antibodyDARPin designed ankyrin repeat proteinDNA deoxyribonucleic acidDWP deep well plateEGTA ethylene glycol tetraacetic acidEIA enzyme immunoassayELISA enzyme-linked immunosorbent assayemPAI exponentially modified protein abundance indexESI electrospray ionizationEtEP equimolarity through equalizer peptideFab fragment antigen bindingFASP filter-aided sample preparationFc fragment crystallizableFT-ICR fourier transform ion cyclotron resonanceGFP green fluorescent proteinGPS global proteome surveyHis6 hexahistidineHPA human protein atlasIBAQ intensity-based absolute quantificationICAT isotope coded affinity tagIEF isoelectric focusingIg immunoglobulinIMAC immobilized metal ion affinity chromatographyiMALDI immuno-MALDIIRMA immunoradiometric assayISET integrated selective enrichment targetiTRAQ isobaric tags for relative and absolute quantification

VI

LC liquid chromatographyLFQ label free quantificationmAb monoclonal antibodyMALDI matrix-assisted laser desorption ionizationMRM multiple reaction monitoringmRNA messenger ribonucleic acidMS mass spectrometryMSIA mass spectrometric immunoassayMS/MS tandem mass spectrometrym/z mass over chargepAb polyclonal antibodyPCR polymerase chain reactionPEA proximity elongation assayPFL protein frequency librarypI isoelectric pointPLA proximity ligation assayPMF peptide mass fingerprintPrEST protein epitope signature tagPSAQ protein standard absolute quantificationPSM peptide spectrum matchPTM post-translational modificationQconCAT quantification concatamerQUICK quantitative immunoprecipitation combined with knockdownRIA radioimmunoassayRNA ribonucleic acidscfv single-chain fragment variableSDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresisSILAC stable isotope labeling by/with amino acids in cell cultureSISCAPA stable isotope standards and capture by anti-peptide antibodiesSMAC sequential multiplex analyte capturingSRM selected reaction monitoringTAP tandem affinity purificationTEV tobacco etch virusTMT tandem mass tagTOF time of flightTXP Triple X ProteomicsµTAS micro total analysis systemWB western blot

VII

Preface

When I started my journey as a PhD student in 2010, I was convinced thatI would find a cure for cancer. Well, maybe not a cure, but at least that Iwould on my dissertation day have been part of the development of a fab-ulous diagnostic platform that would revolutionize the medical field. Fiveyears seemed like such a long time - just think of all the things I could ac-complish! I soon realized however, that I might have overestimated thingsslightly. I came to the understanding that when doing research, five years isactually quite a short period. My enthusiasm was slowly replaced with dis-couragement and I started wondering if I would manage to achieve anythingat all. Eventually, things started to turn and I could experience the joy ofone good result that was easily enough to outweigh a dozen bad ones. TodayI know that every little step towards a cure for cancer, or any other goal onemay have, is of great importance and truly makes a difference. Therefore Ifeel extremely happy and proud that I have made a contribution to the fieldof proteomics through this thesis.

The new findings that are presented in this book are mainly the developmentof novel tools to study proteins in different ways. By taking advantage of thelarge resource of both antibodies and antigens available within the HumanProtein Atlas project, methods for protein identification, quantification andantibody validation have been developed. Present investigations are sum-marized in chapter 6. In order for the reader to get a better picture of theresearch as well as its impact on science, an overview of the field is presentedin chapters 1-5.

Many people have contributed to the research presented in this thesis andto you I am most grateful.

Tove BoströmStockholm, August 25th 2014

VIII

Contents

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ISammanfattning . . . . . . . . . . . . . . . . . . . . . . . . . . . . IIList of publications . . . . . . . . . . . . . . . . . . . . . . . . . . . IIIContributions to the included papers . . . . . . . . . . . . . . . . . VAbbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VIPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VIII

1 Proteins and proteomics 1What are proteins and why should we study them? . . . . . . 1The challenges of proteomics . . . . . . . . . . . . . . . . . . 4The proteome - complex but informative . . . . . . . . . . . . 6

2 Mass spectrometry-based proteomics 9History of mass spectrometry . . . . . . . . . . . . . . . . . . 9Mass spectrometry for protein analysis . . . . . . . . . . . . . 10Sample preparation for mass spectrometry . . . . . . . . . . . 15Data independent acquisition methods . . . . . . . . . . . . . 18Quantitative proteomics . . . . . . . . . . . . . . . . . . . . . 19Proteome coverage and sensitivity . . . . . . . . . . . . . . . . 29

3 Affinity proteomics 33History of the immunoassay . . . . . . . . . . . . . . . . . . . 33Antibodies as affinity reagents . . . . . . . . . . . . . . . . . . 34Alternative affinity reagents . . . . . . . . . . . . . . . . . . . 37The Human Protein Atlas . . . . . . . . . . . . . . . . . . . . 38Immunoassays . . . . . . . . . . . . . . . . . . . . . . . . . . . 42Cross-reactivity in immunoassays . . . . . . . . . . . . . . . . 46

IX

CONTENTS

4 Bridging affinity proteomics with mass spectrometry 49Benefits of combining immunoenrichment with mass spectrom-

etry . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49Immunoenrichment setups . . . . . . . . . . . . . . . . . . . . 50Immunoenrichment of protein or peptide groups . . . . . . . . 58

5 Miniaturization 61Miniaturization in proteomics . . . . . . . . . . . . . . . . . . 61The ISET platform . . . . . . . . . . . . . . . . . . . . . . . . 64

6 Present Investigation 67Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67Development of screening methods for recombinant protein

production (papers I and II) . . . . . . . . . . . . . . . 68Absolute MS-based protein quantification to study the corre-

lation between protein and mRNA levels (paper III) . 75Antibody validation using immunoenrichment coupled to mass

spectrometry (paper IV) . . . . . . . . . . . . . . . . . 79Protein quantification using immunoenrichment and mass spec-

trometry (paper V) . . . . . . . . . . . . . . . . . . . . 83Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . 88

7 Populärvetenskaplig sammanfattning 91

8 Acknowledgements 93

9 Bibliography 97

X

Chapter 1

Proteins and proteomics

What are proteins and why should we study them?

Proteins are everything and they are everywhere. In almost all biochemicalprocesses, you will find that proteins are among the key players. As FrancisCrick pointed out in 1958: "The most significant thing about proteins is thatthey can do almost anything" [1]. Proteins are commonly called the buildingblocks of life, as without them, there would be no life at all. They are crucialin practically every cellular process and a small error in a single protein canbe enough to cause a disease [2, 3].

Today we know quite a lot about proteins. We know what chemical sub-stances they consist of, we know how they are produced in a cell [4, 5], wecan determine what they look like [6] and their exact molecular weight [7].We can determine in what type of cells a protein is expressed, even in whatcellular compartment [8–10] and we can quantify the amounts of differentproteins in various cells and body fluids [11–13]. However, we have yet along way to go before we can fully understand the complex nature of pro-teins in how they function and interact with one another. The research areain which proteins are studied in large-scale (i.e. in high numbers) is calledproteomics.

The composition of a protein is determined by its genetic code, containedwithin the DNA in the cell. DNA, or deoxyribonucleic acid, was discovered

1


in the late 1860s, although it was then not known that this molecule wasthe carrier of genetic information. This was not confirmed until almost onehundred years later [14] and in 1953, Francis Crick and James Watson wereable to determine the now well-known alpha-helical structure of DNA, basedon X-ray analyses performed by Rosalind Franklin and Maurice Wilkins.Once having determined the structure of DNA, they could also propose asystem for DNA replication, that ensured that the DNA content would beidentical in the two daughter cells after cell division [15]. Crick could in1961 propose a sequence hypothesis, in which he stated that three bases ofDNA codes for one specific amino acid [16] and finally, he could also describethe complete central dogma of molecular biology that explains how DNA istranscribed to RNA, which is further translated to protein [1, 4].

Proteins were established as a unique class of molecules already in the late18th century [17], but it was in the following century that the composition ofproteins was revealed. The Dutch chemist Gerardus Johannes Mulder hadin 1838 observed that proteins were very large molecules with similar atomiccomposition, only differing in the amount of sulphur and phosphorus [18].Mulder was the first to use the name protein in a publication, however it wasthe Swedish chemist and clinician Jöns Jacob Berzelius who first proposedthe name [18, 19]. The name protein has its origin from the Greek wordπρωτα ("prota"), meaning "of primary importance". The 20 different pro-tein building blocks, amino acids, were discovered during a period of morethan 100 years between 1819 and 1936. In the beginning of the 20th centurythe structure of proteins was elucidated when the scientists Emil Fischerand Franz Hofmeister independently demonstrated how amino acids wereconnected by peptide bonds [20]. Frederick Sanger could in 1951 present thefirst complete amino acid sequence of a protein, namely the protein insulinof 110 amino acid residues [21, 22]. The amino acid sequence of a protein iscalled the primary structure (Figure 1.1). However, a chain of amino acidslinked together by peptide bonds, a polypeptide chain, will in most cases notbe very functional unless it also has obtained a correct fold. Locally formedstructures are denoted secondary structures and include mainly α-helices,β-sheets and loops [6]. Secondary structure patterns can be predicted by theprotein sequence and the hydrogen bonding between amino acids, as demon-strated by Linus Pauling in 1951 [23, 24]. Secondary structural elements

2


primary structure secondary structure

tertiary structure quaternary structure

Figure 1.1: The different levels of protein structure. Theamino acid sequence of a protein is called the primary struc-ture. The organization of the polypeptide chain into α-helices,β-sheets and loops makes up the secondary structure. Sec-ondary structural elements are organized into a tertiary struc-ture and several polypeptide chains can be combined to forma quaternary protein structure.

of the protein are further organized into a tertiary protein structure. Onemajor driving force of this process is to hide hydrophobic residues in the pro-tein core, hence shielding them from surrounding water molecules, a theorypresented by Irving Langmuir already in 1938 [25]. However it was not un-til 1959, when Walter Kauzmann gave the proposal attention that the idea

3


started to obtain acceptance [26]. Proteins consisting of several polypep-tide chains can be further organized into a quaternary structure. Althoughwe can quite accurately predict protein secondary structure based on aminoacid sequence using different algorithms [6], predicting the complete three-dimensional structure of a protein only by looking at the primary structureis an extreme challenge and is today not possible for the majority of pro-teins. Determining protein structure is, together with investigating proteinfunction, two of the key questions of the proteomic research field.

The challenges of proteomics

The entire set of genes of a cell or organism is called its genome and the large-scale study of genes and genomes is called genomics. A genome is constant,meaning that all cells within an organism contain the same set of genes.PCR technology has made it possible to efficiently multiply DNA for easieranalysis [27,28]. In addition, the development of next generation sequencingtechnology has enabled high-throughput sequencing to a reasonable cost [29].This has resulted in a rapid evolvement in the field of genomics and completegenomes of several organisms have currently been mapped. Among these isthe human genome, which was completed in 2001 [30,31].

All expressed proteins within a cell together make up the proteome, a termfirst coined by Marc Wilkins in the 1990s [32]. In contrast to the genome,protein expression can vary in different cells as well as in different cell stagesand upon different cell stimuli. The proteome is hence very dynamic anddepends on both cell type and time of analysis as well as external factors.In proteomics, the major technique used today to study proteins is massspectrometry (MS), where proteins and peptides are analyzed based on theirmass and charge [33, 34]. However, the first method for large-scale pro-tein analysis was two-dimensional gels (2D-GE), which was developed in the1970s. Using these gels hundreds or thousands of proteins could be sepa-rated and visualized, but analysis of the different proteins was challengingand mainly proteins of high abundance could be identified [35]. After thebreakthrough of MS for analysis of biomolecules in the 1990s, it has grownto become a widely used technique to study proteins due to its accuracy, sen-sitivity and possibility for high-throughput analysis [7]. Another branch of

4


proteomics is affinity proteomics, where affinity proteins such as antibodiesare used for protein identification and quantification. The great advantage ofaffinity proteomics is the high sensitivity of the assays, however the need todevelop affinity reagents towards all target proteins is a great bottleneck andthe assays are greatly dependent on the quality of the affinity reagents [36].Mass spectrometry-based proteomics and affinity proteomics will be furtherdiscussed in chapters 2 and 3.

The human genome contains around 20,000 genes, each coding for one or mul-tiple specific proteins [37]. Even though we know the amino acid sequenceof these proteins, this does not necessarily tell us much about protein func-tion. In eukaryotic organisms, each protein-coding gene is built up by exonsand introns (Figure 1.2). Only the exons are coding sequences and intronswill be removed during a process called alternative splicing that takes placeafter transcription [5]. The exons of a gene are then combined in differentways to create different protein variants, so called protein isoforms. Further-more, proteins can be modified in different ways with different functionalgroups, termed post-translational modifications (PTMs) [38, 39]. For exam-ple protein phosphorylation can activate different signaling cascades, addi-tion of ubiquitin to a protein marks it for degradation by the proteasome andglycosylation can affect cell-cell recognition. Other modifications can alterprotein-protein interactions and cellular localization [38]. Phosphorylationsites have been identified in more than 10,000 proteins [40], emphasizing thegreat importance of post-translational protein processing. Different isoformsand PTMs together increase the number of human protein versions tremen-dously [41]. In addition, the dynamic range of protein concentrations withina typical proteome is huge, especially in blood samples where concentrationsbetween high and low abundant proteins vary more than ten orders of mag-nitude [42]. To generate a full picture of a proteome is therefore extremelychallenging. Moreover, the proteome is very versatile and external changesduring cell cultivation and sample preparation or differences in handling ofblood samples can alter the composition of the proteome [43], making repro-ducibility of proteomic data difficult. All together, the proteome is extremelycomplex, which makes proteomics research a tough task.

5


DNA

mRNA

protein isoforms

protein isoforms with post-translational modi�cations

intron exon post-translational modi�cation

Figure 1.2: The complexity of the proteome compared tothe genome. From one protein-coding gene multiple proteinvariants can arise. Exons can be combined in different waysduring alternative splicing, and proteins can be modified withdifferent post-translational modifications, resulting in a largenumber of protein variants.

The proteome - complex but informative

There is a tremendous amount of information that can be extracted fromproteomic studies that we cannot learn from genomics. Transcriptomics, inwhich RNA is analyzed in a large scale, can act as a bridge between thesetwo areas of research [44, 45]. Levels of mRNA within a cell can tell usapproximately how many times a gene is transcribed. This can indicate atwhat level the corresponding protein is expressed and could therefore giveclues to elucidate protein function. It can be investigated at what time

6


point a certain gene is transcribed, for example if there is a difference intranscription throughout the cell cycle [46]. However, even though mRNAabundance can indicate the amount of protein within a cell, several studiesshow a rather weak global correlation between protein and mRNA levels [47].It has however been shown that analyzing mRNA levels of individual genesinstead of large gene sets can give a much better estimation of protein abun-dance, which is actually rather expected, since different genes are regulateddifferently [40]. Hence, a great amount of useful information regarding pro-teins can come from transcriptomics and transcript data is very useful toresearchers within the proteomics field.

Measuring protein abundance in different cells and tissues is an importantpart of proteomics research. Proteins that are expressed with similar abun-dance across different cells, expressed from so-called housekeeping genes, aremost likely involved in basic cellular functions whereas proteins of very dif-ferent abundance are more likely to have cell-specific functions [48, 49]. If aprotein is believed to be involved in a certain disease, a difference in con-centration of this protein in a sample for a diseased patient compared to ahealthy one would be expected. Proteins that can potentially determine thedisease state of a patient are denoted disease markers or biomarkers [50].Investigating the subcellular localization of proteins can also generate infor-mation about protein function as different subcellular compartments havespecific characteristics and carry out specialized functions [9]. Analysis ofprotein complex structures and interaction networks is also very informa-tive [51, 52] and mapping the PTMs of a protein can generate informationregarding for example protein activity state [39].

7


8

Chapter 2

Mass spectrometry-basedproteomics

History of mass spectrometry

MS is today the most widespread technique to study proteins, however ana-lysis of large biomolecules using this technology is a rather new application.Development of the first mass spectrometer-like instrument dates back to1912, when the British physicist Sir Joseph John Thompson constructed a"mass spectrograph" and managed to obtain mass spectra for O2, N2, CO,CO2 and COCl2. This instrument paved the way for the development of moreadvanced mass spectrometers. The possibility to use MS for the analysis oflarge biomolecules such as proteins was however not realized until severaldecades later. Ionization of the analytes is fundamental for the analysis andthe ionization techniques that were used at this time required the analyte tobe present in gas phase, limiting the analysis to small, volatile compounds.Proteins, which are large and polar molecules and hence extremely non-volatile, were much more problematic [53]. In the 1980s however, two newionization techniques were developed that made it possible to study largerbiomolecules such as peptides and proteins. Matrix-assisted laser desorp-tion ionization (MALDI) and electrospray ionization (ESI) were presentedroughly at the same time by Michael Karas and Franz Hillenkamp [54] andJohn Fenn [55], respectively. Since then, the MS technology has kept on de-

9

Mass spectrometry-based proteomics

veloping at a high speed, however MALDI and ESI remain the ion sources ofchoice for the analysis of biomolecules. In 2002, the Nobel prize was awardedto John Fenn and Koichi Tanaka for their work with ESI and laser desorptionionization.

Mass spectrometry for protein analysis

In MS, proteins are analyzed based on their mass and charge. The instru-mentation consists of three parts: an ionization source, a mass analyzer anda detector (Figure 2.1). Ionization is the first step of the process, where an-alyte ions are produced in gas phase. If the ionization is "soft", the analytestays intact, however harder ionization techniques exist where analyte frag-mentation occurs in the process. In the mass analyzer, the analyte ions areseparated based on their mass to charge ratio (m/z ) and the output from thedetector is a mass spectrum with intensity plotted against m/z. Commonmass analyzers for analysis of biomolecules are the time of flight (TOF),quadrupole, ion trap, fourier-transform ion cyclotron resonance (FT-ICR)and orbitrap mass analyzers [34].

ionization m/z separation fragmentation m/z separation detection

Figure 2.1: Overview of a mass spectrometer setup. Theworkflow includes ionization, m/z separation and detection.Peptide ions can be detected after a first m/z separation orfurther fragmented to enable peptide sequencing as indicatedby the grey box. Usually both modes (MS and MS/MS) areused simultaneously.

MS is a valuable tool in many aspects of protein analysis. After recombi-nant protein production and purification MS can be used for verification ofthe protein identity and the protein can then usually be measured in itsfull-length form [7]. The generated data is the molecular weight of the pro-tein, which in most cases is enough for a reliable identification. However,

10


modified protein variants can be difficult to map, especially if the proteinis produced in a mammalian host where multiple modifications is not un-common. Solubility can be a problem especially when dealing with larger,full-length proteins. Adding detergents to the sample is generally a commonstrategy to tackle this problem, however detergents are not compatible withMS instrumentation and will hamper the analysis.

The most common strategy for MS-based proteomics is "bottom up" pro-teomics, where a proteolytic enzyme is used to cleave a protein pool intopeptides before MS analysis [56] (Figure 2.2). Peptides are, compared tofull-length proteins, more easily ionized in the mass spectrometer leading toan increased sensitivity [57]. This also decreases the issue of protein solu-bility and hence makes it possible to identify insoluble membrane proteins,which is otherwise a very challenging protein class from a proteomics per-spective. In addition, the exact mass of a full-length protein is in manycases not known, due to for example complex modification patterns, as men-tioned above. When analyzing smaller amino acid sequences, this problemis decreased, as peptides from unmodified regions can be used to identify thecorresponding protein. Peptides generated from trypsin digestion (cleav-age after lysine and arginine) are of good size for MS in terms of accuratemass determination and easily deconvoluted charge states and this enzymeis today the standard option for protein digestion, even though combiningseveral proteolytic enzymes can to some extent increase the obtained se-quence coverage [58]. The obtained peptides can be analyzed directly ina mass spectrometer or after separation based on hydrophobicity using anon-line coupled reversed-phase column before injection. However, due todifferences in ionization efficiency not all peptides will be detectable in themass spectrometer and therefore complete sequence coverage will very rarelybe obtained.

Intact peptides can be identified in a process called peptide mass finger-printing (PMF). The molecular weights of peptides identified in the massspectrometer are then used to map the peptides to their corresponding full-length protein. For complex protein mixtures, peptide mass fingerprintinghas the drawback that two peptides of equal mass and charge cannot be dis-tinguished from one another. To get a more reliable identification, MS/MSor tandem MS is used, where not only peptide molecular weight, but also

11


m/z

m/z

intensity

intensity

protein mixture peptide mixture

digestion

MS analysis

MS/MS analysis

1

2

m/z separation m/z separationfragmentation

m/z separation

Figure 2.2: Bottom up proteomics. A proteolytic enzyme isused to digest proteins into peptides, which are injected andanalyzed in a mass spectrometer. (1) In a full MS scan, pep-tides are directly analyzed by m/z. (2) In tandem MS, theions of highest intensity are selected and fragmented to gen-erate smaller ion species. These are further separated by m/zand detected to generate a fragment ion spectrum. From thisspectrum, the peptide amino acid sequence can be determined.

amino acid sequence can be determined. Peptides are first separated in amass analyzer before one chosen peptide ion, the precursor ion, is fragmentedin a collision chamber, for example by collision of the ion with residual gas.After fragmentation, the resulting product or fragment ions are analyzed ina second mass analyzer and a fragment ion spectrum is recorded. The pro-cess is repeated throughout the entire chromatographic separation, with aset number of precursor ions selected per cycle [57]. This setup is described

12


as "tandem in space", with different mass analyzers connected in sequenceto perform the different analyses. In contrast, "tandem in time" can beperformed when using trapping instruments (ion trap, FT-ICR and orbi-trap), where one mass analyzer can perform both MS scan, fragmentationand MS/MS analysis [59]. Three different amino acid bonds can be cleavedduring peptide fragmentation: C(R)-C, C-N or N-C(R). From these cleav-ages, six different fragment ions can be produced, depending on whether thecharge is kept on the N- (a, b or c) or C-terminal side (x, y or z) of the pep-tide (Figure 2.3). The type of cleavage that occurs depends on the appliedfragmentation method.

In a perfect product ion spectrum, peaks representing each peptide fragmentwould be present. The difference in m/z between the peaks reveals whichamino acid has been cleaved off at each specific position. In reality, production spectra rarely contain all possible fragment ions and therefore, de novosequencing is difficult. Instead, product ion spectra are usually searchedagainst protein databases. Theoretical tryptic peptides are generated byin silico digestion of the proteins within the chosen database and used togenerate theoretical fragment ion spectra that are then compared to the ex-perimental data. The molecular weight of the precursor ion together witha mapped fragment ion spectrum is usually enough to determine the ex-act peptide identity. Peptide identifications are reported with a probabilityscore as a measure of the reliability of the identification and today, severaldifferent search engines are available for searching MS data, such as Mascot,X!Tandem and SEQUEST [60–64].

It is also possible to obtain sequence information directly from full-lengthproteins, in a process called "top down" proteomics [65, 66]. Intact proteinions are then fragmented in the mass spectrometer to generate detectablefragment ions. Advantages of this approach are the higher obtained se-quence coverage compared to bottom up proteomics and the increased po-tential to localize PTMs. In addition, the exclusion of the protein digestionstep is beneficial from a time and cost perspective. Using top down pro-teomics, sequence predictions have been made on proteins with molecularweight exceeding 200 kDa [67]. However, this approach also suffers fromseveral drawbacks. Due to the increased complexity of the fragment ionspectra for full-length protein ions, where product ions can have as many

13


CC

CC

CC

H

H

H

HHH

H

H O

O

ON NN

R1

R2

R3

O 1 2 3

CC

C

H

H

H

H O

NN

R1

R2

H

H1 CC

C

H

H

O

O

ON

R3

H

a2 x1

2 CC

C

H

H

H

H O

NN

R1

R2

H

C

O

C

C

H

H

O

ON

R3

H

H

H

b2 y1

CC

C

H

H

H

H O

NN

R1

R2

H

C

O

3 HN

H

H

C

C H

O

O

R3

H

c2 z1

Figure 2.3: Different fragmentation paths of a tripeptideion. (1) C(R)-C bond cleavage resulting in a and x ion series.(2) C-N amide bond cleavage resulting in b and y ion series.(3) N-C(R) bond cleavage resulting in c and z ion series.

different charges as the full-length protein, the strategy is limited to simpleprotein mixtures [66]. In order to reach a high enough resolution of the highcharge state ions expensive FT-ICR instrumentation is most commonly used,especially if PTMs are studied [68].

14


Sample preparation for mass spectrometry

To analyze proteins in cells and tissues, the first step is protein extractionby cell lysis. This can for example be performed by exposing the cells toa detergent-containing lysis buffer or by sonication. Alternatively, thesestrategies can be combined. After cell lysis, a chromatographic step can beapplied for protein purification or fractionation, or proteins can be directlydigested using a proteolytic enzyme.

Purification can be performed to enrich tagged proteins (for example pro-teins containing a His6-tag or FLAG-tag) or for enrichment of a certaintype of proteins (for example charged proteins or proteins with certain mod-ifications) [69] (Figure 2.4). Purification of tagged proteins can today beperformed in a streamlined format, where a large number of proteins canbe purified using a common protocol. For purification of native proteins,the process generally requires optimization making it more time-consuming.Antibodies or other affinity agents can also be used to enrich specific pro-teins [70], as will be discussed in chapters 3 and 4. A purification step willnot only remove interfering proteins, decreasing the risk of ion suppressionduring MS analysis, but can also increase the concentration of the protein inthe sample, which for low-abundant proteins is often crucial. Proteins can befractionated for example based on molecular weight by using size exclusionchromatography or using 2D-GE [32] in which, proteins are separated bothbased on molecular weight and isoelectric point (pI) in two dimensions. Fora proteome-wide experiment where the aim is to cover as large portion ofthe proteome as possible, sample fractionation before and/or after digestionwould be advisable. Dividing the highly complex protein or peptide sampleinto several fractions of lower complexity will generally result in a larger listof identified proteins, however fractionation also leads to longer analysis timeon the mass spectrometer. Extensive sample preparation workflows can alsolead to a substantial sample loss.

Protein digestion can be performed in several ways using proteolytic en-zymes. In solution digestion is a good alternative, especially for simpleprotein mixtures. However, if proteins are directly digested after cell ly-sis without an affinity purification or fractionation step, consideration has tobe taken to the tolerance of the proteolytic enzyme for the sample conditions.

15


protein digestion

protein extraction

MS analysis

A B C

A B

Figure 2.4: Strategies for MS sample preparation. Afterprotein extraction, proteins can either be subjected to affinitypurification (A), fractionation (B) or 2D-GE or SDS-PAGEseparation (C). After digestion affinity purification can be ap-plied to enrich certain peptides or peptides can be fractionatedbased on several different properties before MS analysis.

The commonly used filter aided sample preparation (FASP) method enablesefficient washing of cell lysates before digestion so that any detergents orsalts from the lysis buffer are efficiently removed [71]. This is achieved at

16


the cost of lower sample recovery. SDS-PAGE, where proteins are separatedbased on size, can be used as a combined fractionation and digestion plat-form. Bands from the gel can be cut out and digestion can be performeddirectly in the gel piece [72]. Analysis can be performed on fractions cov-ering the whole molecular weight range or a specific fraction if a particularprotein is of interest. In addition to fractionation, separation of proteinson a gel results in sample cleanup with the removal of salts and detergents.Even after an affinity purification step, there are usually several interferingproteins present in the sample, corresponding to sticky or high abundantproteins and in-gel digestion is therefore a good option to get rid of theseproteins. However, this setup is difficult to scale up and when dealing withmany samples, another method would therefore be recommended.

The resulting peptides can be further fractionated in different ways. Com-monly, two-dimensional fractionation is performed, where the second step ispeptide reversed phase separation coupled on-line to ESI-MS analysis. Thefractionation methods should be orthogonal, meaning that the separation isperformed based on different peptide properties, thereby increasing separa-tion efficiency. For example, isoelectric focusing (IEF) where peptides areseparated by isoelectric point [73], or ion exchange chromatography [74] canbe applied. This can be performed either in an off-line mode where fraction-ation is performed prior to LC-MS/MS analysis or in an on-line setup, suchas the MudPIT strategy where the analytical column is packed with twolayers of chromatographic material (ion exchange and reversed phase) [75].Peptides can also be enriched using specific antibodies [76], which can reducethe sample complexity significantly. If analysis of PTMs is desired, enrich-ment of modified peptides can be performed also on the peptide level usingfor example antibodies targeting a certain modification or with TiO2 for en-richment of phosphopeptides [77,78]. The last step prior to MS analysis is todesalt the sample, which can be done with Stop and Go Extraction (Stage)tips (C18 material loaded in a pipette tip) [79] or commercially availabledesalting platforms [80]. Alternatively, desalting can be performed using anon-line setup with a C18 column, a so called trap column. If analysis is per-formed on an ESI-MS instrument, peptides are usually separated on an LCcolumn on-line prior to injection into the mass spectrometer, as mentionedabove. For MALDI-MS analysis this is more troublesome and off-line LC

17


fractionation can instead be applied if the sample is too complex for directanalysis.

Data independent acquisition methods

In a standard MS/MS analysis, the instrument is operated in data-dependentmode, meaning that the ions of highest intensity are chosen for fragmenta-tion. This is a good strategy for a proteome-wide experiment, however incertain cases only analysis of a set of specific proteins is desired. The dataacquisition can then be performed in targeted mode where specific precursorions are chosen computationally beforehand and all other ions are excludedfrom the analysis, hence operating the instrument in data-independent mode(Figure 2.5). This has the advantage of increased sensitivity, as low abundantions are chosen for sequencing that would otherwise have been masked byhigher abundant ions [81,82]. In a method called multiple or selected reactionmonitoring (MRM or SRM) [83–85], a triple quadrupole mass spectrometeris used for efficient selection of both precursor and product ions to monitorspecific transitions. A triple quadrupole has three quadrupole analyzers insequence. In the first quadrupole, a specific precursor ion is selected, thision is fragmented in the second quadrupole and the third quadrupole selectsone or several product ions that are passed on to the detector [84]. Once anMRM assay is established, many samples can be analyzed rapidly with highsensitivity and reproducibility, however the assay development can be timeconsuming [86]. MRM is therefore commonly used in for example biomarkervalidation studies where relatively few, usually low abundant, proteins areanalyzed in large sample sets.

Data-independent acquisition can also be performed in an untargeted man-ner [87, 88]. One example is SWATH MS, where the m/z detection spaceis divided into 32 sets of 25 Da windows and the mass spectrometer con-stantly scans through these windows and fragments all ions in each win-dow [89].

18


time

intensity

Q1 Q2 Q3

Figure 2.5: Targeted MS using MRM. A peptide ion is se-lected in the first quadrupole (Q1) and thereafter fragmentedin the second quadrupole (Q2). In the third quadrupole (Q3),a fragment ion is selected and passed on to the detector.

Quantitative proteomics

In many cases, quantitative information is required in order to answer acertain biological question. Protein quantification using MS is not straight-forward due to differences in ionization efficiency between different peptides,meaning that two peptides of the same abundance in a sample can giverise to different intensities in the mass spectrometer. Therefore comparingintensities or peak areas between different peptide species will not gener-ate accurate information regarding abundance. In addition, ion suppressionand other matrix effects decreases the reproducibility making comparisonsbetween runs difficult. A common strategy is therefore to add an internalstandard possessing identical chemical properties as the target peptide tothe sample, to which one can compare the ion signal. This can for examplebe done by metabolic labeling or chemical tagging methods to either obtainrelative abundances between two samples or for absolute quantification todetermine absolute copy numbers or protein concentrations within a cell orsample. Although it is also possible to generate quantitative information bycomparing signals from two separate MS runs, the result will be of lower ac-curacy. Many different quantitative methods have been developed and theyall have advantages and drawbacks. One important aspect is at which stagethe samples are mixed, or the internal standard is spiked into the sample.Sample preparation prior to MS is usually quite extensive and large errors

19


can be introduced during these steps. A method that enables mixing ofthe samples at an early stage is therefore desirable to minimize these er-rors. Some of the existing methods for MS-based protein quantification arediscussed in the following sections.

Metabolic labeling strategies for quantitative proteomics

Labeling peptides and proteins with heavy isotope-labeled amino acids hasbecome a widespread strategy to enable accurate protein quantification.Consider a peptide that is present in a sample in a "light" unlabeled formand a "heavy" form with heavy isotope-labeled amino acids. The two vari-ants can be distinguished from one another due to the introduced mass shiftby the heavy isotopes but the isotope labeling will not affect the proper-ties of the peptide. Signal intensities of the two variants can therefore becompared to determine their relative abundance. Labeling of arginine andlysine residues is most common, as cleavage with trypsin will then ensurethat all tryptic peptides will contain at least one labeled amino acid. Severaldifferent variants of heavy isotope-labeled amino acids can be used, wherecarbon and/or nitrogen is labeled [90], however quantification using labeledversions of other amino acids or complete labeling of all amino acids can alsobe performed [91–93].

Stable isotope labeling of amino acids in cell culture (SILAC) is a widely usedmethod for relative protein quantification [91, 94] (Figure 2.6). In SILAC,cells are cultivated in medium containing heavy isotope-labeled amino acidsand hence, the stable isotopes are incorporated into the proteins by the ma-chinery of the cell. Since all expressed proteins are labeled, SILAC generatesrelative quantitative data on a proteome-wide scale. SILAC has for examplebeen used to analyze protein signaling pathways [92, 93, 95], to investigatecancer proteomic profiles [96,97] and to determine the cellular response upondrug treatment [98]. SILAC can be used to obtain relative quantitative databetween two [93], three [95] or even five samples, with differently labeledarginine variants [90]. Experiments with four or five different samples usingdeuterated amino acids is also possible, however this may lead to alterationsin LC retention time due to the deuterated amino acids, called the deuteriumisotope effect [99]. In addition, multiplexing will lead to an increased sample

20


light amino acids heavy amino acids

mix

digestion

protein extraction

m/z

intensity

MS analysis

MS spectrum

Figure 2.6: Workflow for relative quantification of two sam-ples using SILAC. One cell sample is cultivated using stan-dard isotope (light) amino acids and the other with heavyisotope-labeled (heavy) amino acids. The samples are mixed,digested and analyzed in a mass spectrometer. Intensity ra-tios between the peptide variants are used to determine therelative amounts of the peptides in the two samples.

complexity, hence making quantification more difficult.

A variant of the SILAC method, termed super SILAC or spike-in SILAC,has been developed where labeling of the sample with heavy isotopes canbe avoided. Instead, a super SILAC mix is used, containing lysates fromcell lines cultivated in media with heavy isotope-labeled amino acids. Thesuper SILAC mix is spiked into each sample and ratios between heavy and

21


light peptides are obtained. These ratios can then be used to determine therelative abundances of the proteins between the different samples [10, 100,101]. Since sample labeling is not necessary, patient samples such as tissuescan also be analyzed [102, 103]. However, the results are dependent on thechoice of cell lines for the super SILAC mix, since quantification of proteinsin tissue requires that these proteins are present also in the internal standardsample.

Chemical labeling strategies for quantitative proteomics

Other methods for protein quantification use chemical labeling of proteins orpeptides. Two commonly used methods generally used for relative quantifica-tion are isobaric tags for relative and absolute quantification (iTRAQ) [104]and tandem mass tags (TMT) [105]. Here, labeling is performed on the pep-tide level after protein digestion and peptides are labeled at primary amineresidues. Several different tags exist, enabling multiplex experiments, up to8-plex for iTRAQ [106] and 10-plex for TMT [107]. The tags are isobaric,meaning that they have the same mass and different tags can therefore notbe distinguished in an MS spectrum. All tags contain a reporter group, withslightly different mass for the different tags. During peptide fragmentation,reporter ions of a specific size are generated for each isobaric tag. Thesewill be visible in the product ion scan, where their relative intensities canbe used to determine the relative abundance of the peptide in the differentsamples.

One significantly cheaper option is dimethyl labeling, where formaldehydewith either hydrogen or deuterium atoms is used as a reagent to generatea mass shift of 4 Da between the labeled variants [108, 109]. The wholelabeling procedure takes less than five minutes. A drawback with this quan-tification strategy is however, as mentioned previously, the fact that deuter-ated peptides show a chromatographic retention time shift, which can affectquantification accuracy. Another strategy is labeling with 18O. This label-ing is performed during proteolysis by trypsin, where hydrolysis in H2

18Oresults in two 18O being incorporated into the carboxyl terminus of the tryp-tic peptide [110, 111]. The efficiency of the labeling however depends bothon peptide length and sequence, resulting in variable incorporation of 18O.

22


This generates peptides with either one or two labeled oxygen atoms, whichdecreases the accuracy of the method.

The first method based on isotopic labeling developed for quantitative MS-based proteomics was isotope coded affinity tags (ICAT). ICAT was pre-sented in 1999 and applies labeling on the protein level at cysteine residues[112]. The original tag contained either zero or eight deuterium residues,along with a biotin molecule for affinity purification of the labeled peptidesafter digestion. Hence, this system also suffered from the deuterium iso-tope effect and a modified version was therefore developed where carbonisotopes were used instead of deuterium atoms [113]. An advantage of ICATcompared to iTRAQ and TMT is that labeling is performed earlier in theprocess, leading to less error due to differences during sample preparation.Since only proteins containing a cysteine can be quantified with this methodICAT is not suitable for all applications, however the affinity enrichment oftagged peptides reduces the sample complexity, which can be an advantageif peptides of low abundance are of interest.

Label-free relative quantification

Methods not relying on protein or peptide labeling, termed label-free quan-tification, also exist, with the advantage of easier sample preparation andlower cost but at the expense of lower quantitative precision (Figure 2.7).Peptide spectrum matches (PSMs), peptide signal intensities or areas undercurves (AUCs) from two or more samples can be directly compared withoutprior mixing of the samples [114, 115]. However, when comparing signalsfrom different samples, the requirements on MS instrumentation regardingaccuracy and precision are increased.

In spectral counting, it is assumed that the number of identified PSMs fora certain protein is relative to its abundance [115,116]. This is a commonlyused strategy, however also controversial since protein quantification is basedon the number of spectra instead of actual data [117]. Advantages with thistechnique include the straightforward data analysis, which does not requirespecialized software. Since quantification is based on the number of tandemMS spectra mapped to a certain protein and the accuracy of the quantifi-cation is increased with more data, multiple sequencing of each peptide is

23


intensity

protein extraction

protein digestion

data analysis and comparison

MS analysis

m/z

intensity

m/z

intensity

time

intensity

m/z

intensitym/z

intensity

m/z

intensity

m/z

intensity

m/z

intensitym/z

intensity

time

Figure 2.7: Workflow for label-free quantification. Sam-ples are treated separately during the whole sample prepara-tion and analyzed separately in a mass spectrometer. Duringdata analysis, samples are compared either by peptide inten-sities, AUCs or number of PSMs.

24


favorable. However, sequencing the same peptide over and over will decreasethe overall proteome coverage, as the mass spectrometer can only sequencea certain number of peptides in a run. In addition, peptides with broaderchromatographic peaks will be identified more times than peptides with verynarrow elution profiles, leading to more PSMs and a higher estimation of pro-tein abundance. Lower abundant proteins will generally have a higher varia-tion in the number of PSMs compared to higher abundant proteins, makingthis strategy less accurate for proteins of low concentration [12,13].

Contrary to spectral counting, when comparing peptide signal intensities orAUCs for relative quantification, it is assumed that the peak area or inten-sity of a peptide is relative to its abundance in the sample [118, 119]. Thelinear relationship between signal intensity and peptide abundance makesthis approach more accurate for quantification of lower abundant proteinsthan spectral counting, where very few PSMs are mapped to a certainprotein [120]. However, necessary data processing such as feature detec-tion, normalization, noise reduction and accurate matching of MS peaksbetween runs makes data analysis more challenging compared to spectralcounting [117].

Absolute protein quantification by spike-in standards

Compared to relative quantification where the difference in abundance be-tween two or more samples is determined, absolute quantification generatesa specific protein copy number or protein concentration within a sample.Absolute quantification strategies generally make use of isotope-labeled stan-dards, of which the absolute concentration is known [121] (Figure 2.8). Syn-thetic heavy isotope-labeled peptides, so-called AQUA peptides (for Abso-lute QUAntification), can be spiked into a sample and the difference in signalintensity between the AQUA peptide and the corresponding endogenous pep-tide is thereafter used to determine the absolute peptide abundance in thesample. AQUA peptides are widely used and commercially available [122].Another type of standard is QconCAT (quantification concatamer), whichconsists of several concatenated peptides in sequence resulting in quantita-tive data from more than one peptide [123]. The standard is added to thesample before digestion, which reduces the error from incomplete proteoly-

25


sis. However, it is important that the QconCAT standard is digested withequal efficiency compared to the endogenous protein, which would otherwiselead to inaccurate quantitative data. Since QconCAT standards are usuallyexpressed in Escherichia coli, the production is both cheap and simple.

digestion

protein extraction

m/z

intensity

MS analysis

MS spectrum

addition of protein standard

addition of peptide standard

Figure 2.8: Workflow for absolute quantification using heavyisotope-labeled standards. Protein standards can be added tothe sample already after protein extraction, whereas peptidestandards are added after proteolytic digestion. After MSanalysis, heavy to light ratios are used to determine the abso-lute concentration of the peptide within the sample.

An optimal strategy for absolute quantification is to add a full-length heavyisotope-labeled protein as quantification standard. A full-length proteinstandard can be spiked in before proteolysis and will be digested with the

26


same efficiency as the endogenous version. Quantitative data from pep-tides over the whole protein sequence will be generated, resulting in a veryaccurate quantification. Methods using full-length proteins as internal stan-dards include absolute SILAC [124], protein standard absolute quantifica-tion (PSAQ) [125] and FlexiQuant [126]. Furthermore, production of theprotein standard in the same host as the sample with the endogenous pro-tein would be beneficial as this would ensure that the proteins will carrythe same modifications. However, the time-consuming and challenging pro-duction of full-length proteins, especially in mammalian hosts, hinders thelarge-scale applicability of these methods. A relatively new strategy, usingprotein fragments of 25-150 amino acids generated in a high-throughput for-mat could be a promising strategy for absolute quantification in large-scalestudies [127,128]. This strategy will be further described in chapter 3.

Absolute label-free quantification

Label-free approaches can also be used to generate absolute quantitativedata, however since peptide intensities and number of PSMs cannot directlybe used for absolute protein quantification, data normalization is needed.The normalization can be performed in different ways. In intensity-basedabsolute quantification (iBAQ), the total signal intensity of all peptides froma protein is normalized by the number of theoretical peptides for the pro-tein [129]. The high3 method uses a similar normalization approach, howeveronly the three peptides of highest intensity are used for quantification [130].It is assumed that the best ionizing peptides from different proteins shouldgenerate roughly the same intensities, wherefore these peptides should gen-erate more accurate data than including all peptides. The exponentiallymodified protein abundance index (emPAI) normalizes the number of iden-tified peptides by the number of theoretical peptides to determine absoluteprotein quantities [131]. In absolute protein expression (APEX), the num-ber of PSMs instead of number of identified peptides is used. This valueis normalized by the number of expected peptides, after first estimating adetection probability for each peptide based on experimental data [132]. Ina study comparing the performance of emPAI, APEX and T3PQ (similarprinciple as high3) using bovine Fetuin spiked into a yeast lysate, it wasobserved that the instensity-based method T3PQ showed the highest linear-

27


ity over the investigated dynamic range, whereas the emPAI signal reachedsaturation at a certain point. At a certain abundance level, no further pep-tides will be identified even though the abundance is increased, explainingthe lower linearity for this method [133]. Another study compared the em-PAI, APEX and iBAQ methods for proteome-wide quantification of proteinswithin an E. coli lysate. Similar correlations were determined when com-paring the methods to one another, however the iBAQ method showed thelowest variation between biological replicates [134].

Comparing quantification methods

As mentioned earlier, the time point at which the samples are mixed (relativequantification) or the quantification standard is added (absolute quantifica-tion) can have an impact on the final result, as it can be assumed thaterrors are introduced in every step of the sample preparation. In SILAC thesamples are mixed directly after cell lysis, whereas in iTRAQ, samples arehandled separately until after protein digestion and peptide labeling. There-fore, for experiments where extensive sample preparation is performed onthe protein level, SILAC would be a better alternative. Still, research hasbeen presented where iTRAQ and TMT generated quantitative data withbetter overall accuracy compared to a 14N/15N metabolic labeling strategy,even though mixing was performed later in the sample preparation work-flow [135]. In addition, iTRAQ enables multiplexing of up to eight samples,whereas SILAC is usually performed in 2- or 3-plex. However, comparisonsbetween different levels of iTRAQ and TMT multiplexing have showed thatincreased multiplexing lowers the sensitivity, leading to less identified pro-teins [136, 137]. In addition, not all samples can be analyzed with SILAC,as cells need to be grown in medium containing heavy isotope-labeled aminoacids. Super SILAC can however be used to solve this problem. Interest-ingly, it has recently been shown that dimethyl labeling can perform equallycompared to SILAC in terms of quantitative precision and could thereforebe an alternative to super SILAC for protein quantification in tissues [138].The same study also investigated TMT labeling and found that the quan-tification rate for this method was significantly higher than for SILAC anddimethyl labeling. Co-isolation during precursor isolation is however a prob-lem for methods where quantification is performed on the MS/MS level,

28


such as TMT and iTRAQ, and can affect the accuracy of the quantitativeresults. Using an MS3-based approach or only regarding PSMs with suffi-ciently low isolation interference has however been shown to decrease thisproblem [137,138].

In absolute quantification, AQUA peptides are added to the sample afterdigestion, whereas full-length proteins, protein fragments or QconCAT pro-teins can be spiked in at an earlier stage. However, if the standard is notidentical to the endogenous protein, as for protein fragments or QconCATstandards care needs to be taken to ensure that the digestion efficiency isnot altered. Moreover, even though full-length proteins and protein frag-ments are promising for absolute quantification, some peptides will generateinaccurate data due to differences in modification patterns if the standardprotein was not produced in the same host as the endogenous protein.

Label-free methods are beneficial regarding both time and cost, making thisapproach a good alternative for the analysis of large sample sets. However,since data between different MS runs is compared and both sample prepara-tion and MS acquisition can differ between samples, the accuracy is lower forthese methods. This means that label-based methods can detect smaller dif-ferences between samples than what is possible with label-free quantificationmethods [120,139]. If an accurate estimation of protein abundance is desired,a label-free approach should therefore not be the method of choice, althoughwhen a large difference in protein abundance between samples is expected,a label-free quantification is a fast and simple alternative. It should howeverbe noted that the natural biological variation between individuals can insome cases be relatively large, leading to high variability between biologicalreplicates even for a very accurate and precise quantitative method.

Proteome coverage and sensitivity

Even though proteome-wide analysis using MS can detect thousands of pro-teins in a single run, achieving full coverage is today still not feasible. Theyeast proteome contains around 6,000 proteins and of these more than 4,000have been identified in MS analyses [140–143]. A lot has happened regard-ing instrumentation in the past years that has improved the potential forlarger proteome coverage. In 2001, Washburn and coworkers managed to

29


identify almost 1,500 yeast proteins in 68 hours of analysis time [75], whileearlier this year, Hebert and coworkers identified almost 4,000 yeast pro-teins in a little more than one hour [143]. There are roughly 20,000 genesin the human genome, however it is unlikely that all of these are expressedsimultaneously within a cell. More than 10,000 proteins in total have beenidentified in human cell lines, indicating that the complexity of a humanproteome lies at least around this value [10, 58, 73, 144]. This has been ac-complished using quite different setups with MS instrumentation time ofroughly one day [10] to twelve days [58]. A typical MS experiment identifiesbetween 5,000-8,000 proteins [71, 97, 129], however the number of proteinsidentified in one experiment depends on multiple factors, such as samplepreparation and fractionation, LC setup, mass spectrometer instrumenta-tion and data analysis. Several projects have been initiated to make largeamounts of MS-data available to the research community. The Peptide At-las [145] aims to achieve complete annotation of genomes of different speciesby providing verified proteomics data. In 2013 the Peptide Atlas consistedof data corresponding to 12,644 proteins, i.e. above 60% of the numberof coding genes [146]. In addition, the recently published ProteomicsDB,contains searchable proteomics data based on almost 17,000 LC-MS/MS ex-periments from human cell lines, tissues and body fluids [40]. The dynamicrange of proteins in human cells has been shown to span around seven or-ders of magnitude [144] and proteins spanning this concentration range havebeen detected when using iBAQ label-free quantification to analyze proteinabundance in eleven human cell lines [10].

Targeted approaches such as MRM assays can be used to decrease or entirelyexclude the need for sample fractionation and significantly lower the requiredMS analysis time [81]. However, if low abundant proteins are analyzed, frac-tionation is advisable even when using MRM, in order to avoid interferingmolecules [84]. In targeted approaches, assay development is a major bottle-neck and therefore, this strategy is not suitable for whole-proteome analysis.In one study, Ebhardt and coworkers analyzed an unfractionated lysate froma U2-OS human cell line using a 35 min LC gradient. They managed to iden-tify more than 70% of the 52 targeted proteins with copy numbers per cellas low as 7,500 [147]. Even though this targeted approach could decreasethe analysis time significantly, proteins of similar copy numbers have been

30


detected also with shotgun proteomics using fractionation and longer gra-dients [127]. One major application for MRM is to analyze more complexsamples, such as plasma, for the detection of low abundant plasma pro-teins [82, 83, 148]. The protein abundance range in plasma spans ten ordersof magnitude with some serum proteins, e.g. serum albumin, being presentin concentrations up to 40 mg/mL [42]. These proteins often hinder theidentification of lower abundant proteins in undepleted plasma making theanalysis of plasma samples a great challenge. However, due to the largeamount of information residing within this sample type and the relative easeof sample collection, plasma is very attractive for diagnostic purposes [42].Plasma biomarker detection with the possibility to accurately diagnose pa-tients with a certain disease is of course very beneficial. However, regardingbiomarkers a yes or no answer will not be sufficient and quantitative assaysare of more use. Quantitative data has also been generated for proteins oflow attomole levels or ng/mL concentrations [81, 82, 85, 149, 150], althoughfor many low abundant plasma proteins this is still not sensitive enough andfurther method development is required [151]. However, with the rapid im-provement in mass spectrometry instrumentation regarding sensitivity, thisshould in the future not be impossible [152].

31


32

Chapter 3

Affinity proteomics

History of the immunoassay

Affinity proteomics, the second branch of proteomics, makes use of affinityreagents such as antibodies to detect and quantify target proteins. In an im-munoassay setup, analytes within liquid samples such as blood or cell lysatescan be detected by the specific recognition of an affinity molecule. RosalynYalow and Solomon Berson developed the first immunoassay in 1960 [153].It required a pure analyte labeled with a radioactive isotope atom and thedetection and quantification of the analyte within the sample was enabledthrough competitive binding between the labeled and unlabeled analyte toan antibody. The method was given the name radioimmunoassay (RIA)and Yalow was in 1977 awarded the Nobel Prize in physiology or medicinefor this work. In 1968, Laughton E.M. Miles and Charles Nicholas Halesintroduced the immunoradiometric assay (IRMA) [154, 155]. This methodresembles RIA, although the antibody instead of the analyte is labeled withthe radioactive isotope, introducing several improvements. Firstly, in RIA,a small decrease in radioactivity is detected against a relatively large back-ground signal, which lowers the sensitivity. Secondly, when labeling theanalyte it is possible that the interaction between analyte and antibody is af-fected, which could alter the equilibrium. A few years later, Peter Perlmannand Eva Engvall introduced the new important method enzyme-linked im-munosorbent assay (ELISA) [156]. Although similar to RIA and the IRMA,

33

Affinity proteomics

ELISA uses enzyme-coupled antibodies to generate a signal output. Usingenzymes instead of radioactive isotopes has the advantage of higher stability,making labeled antibodies active and useful for a longer period of time. Atthe same time, Anton Schuurs and Bauke van Weemen presented the en-zyme immunoassay (EIA) [157], where, in contrast to ELISA, the enzyme iscoupled to the analyte instead of the antibody. Today, ELISA is a widelyused method for analysis and quantification of different molecules in com-plex samples [158] and exists in different formats, such as direct ELISA,indirect ELISA and sandwich ELISA (Figure 3.1). In addition, several otherimmunoassay setups exist, of which some will be discussed later in this chap-ter.

A B C

Figure 3.1: Different ELISA setups. (A) Direct ELISAwhere a labeled primary antibody is used for detection, (B)indirect ELISA where a secondary labeled antibody is added tothe sample for signal output and (C) sandwich ELISA wheretwo target-specific antibodies are needed for target detection.

Antibodies as affinity reagents

Even though several different types of affinity molecules are used in researchand in the clinic, the predominant molecule by far is the antibody (Ab)or immunoglobulin (Ig) [159]. In nature, antibodies have a central role inour immune system, where they recognize and mark foreign objects suchas bacteria or viruses for destruction [160]. Antibodies are large (150 kDa)proteins consisting of four subunits: two identical light chains and two iden-tical heavy chains, linked to each other through disulfide bonds forming a

34

Affinity proteomics

Y-shaped molecule (Figure 3.2A). Different isotypes of heavy chains resultin the different Ig variants IgA, IgD, IgE, IgM and the most common IgG,all with different functions. Two different isotypes also exist for the lightchains; λ and κ, however they have no known functional divergence. Threeconstant domains (CH1, CH2 and CH3) and one variable domain (VH) makeup each heavy chain whereas the smaller light chains consist of one constant(CL) and one variable domain (VL). The variable domains are located at thetips of the Y-shaped structure, where the outer part makes up the antigen-binding site with three complementary determining regions (CDRs). CDRsof different antibodies differ both in sequence and size and give antibodiestheir different specificities. The fragment crystallizable (Fc) consists of thestem of the antibody (CH2 and CH3) and is a constant region not involvedin antigen binding. Instead, the Fc region binds to cell surface receptors ofimmune cells, in this way directing the immune system toward the invadingobject [6]. Antibodies are produced in B cells. In a process called somatic re-combination, different CDR gene segments are randomly combined resultingin a large diversity of antibodies with different specificities [160].

Generation of antibodies for research or diagnostic applications can be car-ried out in different ways, for example by immunization of a host animal,such as a rabbit or mouse, with a target antigen [160]. Antibodies canbe purified from the sera of the host animal, for example using a proteinA/G column for enrichment of all IgG molecules or preferably by using anantigen-coupled column, generating a polyclonal antibody mixture with IgGvariants targeting the protein of interest. Since these antibodies are gener-ated from different B cells their specificities will not be identical and differentantibodies will bind to different parts (epitopes) of the antigen. Polyclonalantibodies are not a renewable source and re-immunizing an animal withthe same antigen will generate a polyclonal mix of antibodies with similar,but not identical epitope specificity [161]. An alternative approach is to iso-late individual B cells and immortalize them by fusion with a tumor cell, atechnique that was invented by Georges J.F. Köhler and César Milstein in1975 [162]. The resulting hybridoma cells can grow and produce antibodiesindefinitely, secreting the product into the surrounding medium. Since allantibodies will be generated from the same B cell, these monoclonal anti-bodies will share the same epitope-specificity [163]. However, in order to

35

Affinity proteomics

find a monoclonal antibody with the right specificity, tedious screening isoften required, as many of the clones will not show good enough affinity orspecificity for the target antigen. It is also possible to mimic nature’s anti-body selection machinery and select high affinity binders in vitro using forexample phage display. Antibody selection using in vitro display systemsoffers a more controlled selection process, where the selection can continueuntil satisfactory binders have been obtained. However, smaller antibodyfragments rather than full-length antibodies are mainly generated using thisapproach [164].

CH1

CH2

CH3

VH

VL

CL

CDRs

Fc

Fab

CH1

VH

VL

CL

CDRs

Fab

antibody

F(ab’)2

CH1

VH

VL

CL

Fab

VH VL

ScFv

A B

Figure 3.2: (A) Schematic structure of an IgG molecule and(B) some common antibody derivatives.

Depending on what kind of antigen is used in the immunization, antibodies ofdifferent epitope specificity will be obtained. If a peptide is used as antigen,antibodies recognizing linear epitopes are most likely generated. However,since small peptides can adopt multiple structures [165], it is likely thatsome of these antibodies will not be able to bind the structured, full-lengthtarget protein. Therefore, larger peptides or protein fragments are in mostcases preferred. Larger antigens can generate antibodies targeting either lin-ear or structural epitopes and the probability that these antibodies will beable to detect structured target proteins is therefore increased. This typeof antibodies are in many cases functional in assays where native proteinsare to be detected, as for example in immunoenrichment experiments as will

36

Affinity proteomics

be discussed in chapter 4. However, in immunohistochemistry experimentsor western blot, where target proteins are denatured, antibodies targetingstructural epitopes might not be functional. Hence, it is important to con-sider in what applications the antibodies are to be used before choosing asuitable antigen for antibody generation.

Alternative affinity reagents

In certain applications, the large size of antibodies can be a drawback. An-tibodies have a long half-life, resulting from both size and effector functionsof the Fc fragment. This is disadvantageous for example in imaging applica-tions, where the molecule should be eliminated from the body shortly afterthe image is generated [166]. In addition, production using recombinanttechniques in bacterial hosts, which is both economical and efficient [167],is difficult for antibodies due to their large size. To overcome these obsta-cles, several smaller antibody derivatives have been developed, as well asnon-immunoglobulin based scaffolds [168, 169]. Common antibody deriva-tives include fragment antigen-binding (Fab) fragments, which consist ofthe light chain and VH and CH1 from the heavy chain (Figure 3.2B). Theenzyme papain cleaves the antibody just below CH1 and CL, thus generat-ing two Fab fragments. Treating the antibodies with pepsin instead, whichcleaves slightly below the cleavage site of papain, instead generates a F(ab’)2fragment, consisting of two Fab fragments connected by a disulfide bridge.Another antibody derivative is the single-chain fragment variable (scfv) frag-ment, which consists of the VH and VL domains, connected by a linker.These antibody fragments all lack the Fc part and consequently do not pos-sess any effector functions but only act as affinity molecules.

One non-immunoglobulin based scaffold is the Affibody molecule: a small,three helical bundle protein of 58 amino acids originating from one of the IgG-binding domains of staphylococcal protein A [170]. The original specificityof the protein is directed towards IgG, however by randomizing 13 positionssituated in two of the helices involved in IgG binding, large libraries can begenerated, from which proteins possessing new desired specificities can be se-lected. A similar protein is derived from an albumin-binding domain (ABD)of streptococcal protein G. This protein is similar in structure and size to

37

Affinity proteomics

the Affibody molecule, however its original affinity is towards albumin. Ran-domizing positions involved in albumin binding has generated ABD variantswith other specificities [171]. In addition, randomization has also been per-formed on the opposite side of the albumin-binding site, hence preservingthis affinity and generating new bi-specific binders [172]. Other alternativescaffolds include designed ankyrin repeat proteins (DARPins), which are de-rived from the naturally occurring ankyrin proteins and consist of severalrepetitive units of two antiparallel α-helices [173]. Adnectins have a β-sheetfold similar to the IgG-fold, with the strands connected by six loops that arerandomized for selection purposes [174]. Anticalins have a β-barrel structurewith four connecting loops targeted for randomization [175]. Aptamers areshort single-stranded DNA or RNA molecules that have emerged as alter-natives to protein affinity binders due to for example their high stability,straight-forward production and low immunogenicity [176].

The Human Protein Atlas

The Human Protein Atlas (HPA) project is a proteomics initiative withthe aim to map the human proteome with antibody-based methods in agene-centric manner and generate a resource of antibodies targeting all hu-man proteins [177]. The generated antibodies are thoroughly validated andare used to study protein expression in cells, healthy and cancer tissuesand plasma. Since the launch in 2003, polyclonal antibodies targeting over16,000 genes, corresponding to over 80% of the human genome, have beengenerated.

Antigen selection is the first step of the workflow, where protein fragments,termed protein epitope signature tags (PrESTs) of 25-150 amino acids are de-signed [178] (Figure 3.3). The antigen selection is based on sequence homol-ogy towards other human proteins and sequences with low similarity to othergenes are chosen to minimize the risk for cross-reactivity. In addition, trans-membrane regions and signal peptides are avoided. It would be desirableto choose protein sequences located at the protein surface, since this wouldincrease the success rate of the antibodies in assays where native proteins areanalyzed. However, since only a subset of the protein structures have cur-rently been solved, this is not a criterion in the antigen design process. After

38

Affinity proteomics

cloningantigen production

and puri�cation immunization antibody

puri�cation

western blot immunohistochemistry immuno�uorescenceprotein arrays

PrEST design

His6

ABP PrEST

Figure 3.3: Workflow of the HPA project. Antigen se-quences are chosen and cloned into an expression vector.Antigens are produced and characterized before rabbit immu-nization. Sera are purified to obtain a mixture of polyclonalPrEST-specific antibodies, which are thoroughly validated andused for protein analysis in immunohistochemistry and im-munofluorescence.

reverse transcription of the selected gene fragments from RNA pools andinsertion of the DNA fragment into an expression vector, the antigens arerecombinantly produced using E. coli as expression host. Antigens are pro-duced in fusion with a hexahistidine (His6) tag as a purification handle andalbumin binding protein (ABP) for increased solubility, in a high-throughputfashion [179]. With the present workflow, production of almost 300 proteinsper week in batches of 72 proteins is possible. The cell cultivation is per-formed in 1 L shake flasks and the protein purification is performed using afully automated liquid handling system. The purified proteins are validatedusing SDS-PAGE for purity estimation and MS for molecular weight verifi-cation with a success rate of over 80%. The following step is immunization ofrabbits to generate polyclonal antibodies. Antibody-containing sera are pu-

39

Affinity proteomics

rified in a two-step protocol in order to separate antibodies targeting the tag(His6ABP) from PrEST-specific antibodies. In the first step, tag-specificantibodies are retained on a His6ABP column after which PrEST-specificantibodies can be captured on a column with immobilized PrEST [180]. Pu-rified antibodies are thereafter validated using several methods. Microarrayswith 384 printed PrESTs are used to verify antibody specificity and westernblot analysis using samples from two human cell lines, two human tissues andhuman plasma is used to verify specific binding of the antibody to the full-length protein in a complex matrix [181,182]. The antibodies are thereafterused to study protein expression in 48 normal human tissues, 20 cancer tis-sues and 47 different human cell lines using immunohistochemistry [183,184].Subcellular localization is investigated using immunofluorescence and confo-cal microscopy analysis of three human cell lines [9,185]. Finally, all data ismade publicly available on the HPA website (www.proteinatlas.org).

High-throughput production of heavy isotope-labeled protein frag-ments for mass spectrometry-based protein quantification

As discussed in chapter 2, quantification using MS is feasible but not straight-forward. For absolute quantification, heavy isotope-labeled proteins or pep-tides can be used as internal standards. Due to difficulties in producinglarge numbers of heavy isotope-labeled full-length proteins, synthetic pep-tides with incorporated heavy isotopes are most commonly used, althoughfull-length proteins would be desirable due to the possibility to add the stan-dard at an earlier stage. Within the HPA project, PrEST protein fragmentsare used as antigens for antibody generation. The majority of all avail-able PrESTs generate at least one unique tryptic peptide and could there-fore be used as internal standards for MS-based absolute quantification. Intheir light form, PrESTs can be directly spiked into heavy isotope-labeledcell lysates to determine absolute protein copy numbers. However, sincemany interesting samples, such as plasma or tissue, cannot be labeled withheavy isotopes, heavy isotope-labeled standards are of great importance. Apipeline for the production of heavy isotope-labeled PrESTs has been setup [127], where a modified E. coli strain, auxotrophic for arginine and lysineis used for the PrEST production [186]. Lysine and arginine are supplied toa minimal culture medium in heavy isotope-labeled forms to generate heavy

40

Affinity proteomics

isotope-labeled protein fragments. Lysis and purification is performed withthe standard protocols used within the HPA project.

His6 ABP PrEST

His6 ABP OneStrep

combine and digest

heavy PrESTlight HisABPOneStrep

MS analysis

A

B

Sequence used for PrEST quanti!cation

Sequence used for quanti!cation of endogenous protein

heavy protein

light protein

Figure 3.4: Generation of heavy isotope-labeled PrESTstandards for MS-based absolute quantification purposes. (A)PrEST quantification is performed using the quantificationstandard His6ABPOneStrep. Peptides originating from theHis6ABP region are used for accurate PrEST quantification.Peptides originating from the PrEST sequence can thereafterbe used to quantify endogenous proteins in for example a celllysate. (B) Workflow for PrEST quantification. PrEST andquantification standard are mixed and digested. Heavy to lightratios from the His6ABP region are used to determine thePrEST concentration.

Within the ordinary pipeline, the concentration of purified PrESTs is mea-sured using bicinchoninic acid (BCA) assay, which is accurate enough forthe immunization purposes of the regular, light PrESTs. However, theconcentration of internal standards for absolute quantification needs to be

41

Affinity proteomics

very accurately determined in order to obtain reliable quantitative data.Amino acid analysis, where proteins are subjected to complete hydrolysisgenerating free amino acids that can be quantified one by one, gives anaccurate measure of protein abundance in a pure sample. However, thismethod is rather expensive making it disadvantageous for high-throughputpurposes. Instead, PrEST quantification is performed using an MS-basedsetup (Figure 3.4), similar to for example the EtEP strategy, describedby Holzmann and coworkers [187]. In EtEP, an N-terminal equalizer pep-tide (EP), added to the internal standard peptide during chemical synthe-sis, is used to determine the concentration of the peptide standard. ThePrEST-based method instead uses a quantification standard consisting ofthe His6ABP tag and a strep tag as a second purification handle for thispurpose. For quantification of heavy isotope-labeled PrESTs, an unlabeledversion of the quantification standard is used and vice versa. The quantifi-cation standard, termed His6ABPOneStrep, is purified in a two-step setupusing both immobilized metal ion affinity chromatography (IMAC) and pu-rification on a StrepTactin column. His6ABPOneStrep is further quantifiedusing amino acid analysis. When mixing unlabeled His6ABPOneStrep and aheavy isotope-labeled PrEST and digesting them with trypsin, peptides orig-inating from the His6ABP tag will be present in both unlabeled and heavyisotope-labeled forms. Since the absolute quantity of the light peptides orig-inating from His6ABPOneStrep is known, the obtained heavy to light ratioscan be used for determination of the absolute quantity of the heavy isotope-labeled PrEST. Accurately quantified PrESTs can then in turn be used asinternal standards in MS-based absolute quantification setups where pep-tides from the PrEST region are compared to peptides from the correspond-ing endogenous protein in for example a cell lysate. The PrESTs commonlyenable protein quantification using multiple peptides, although differencesin sequence length and location within the gene of interest make differentPrESTs more or less suitable for this application.

Immunoassays

The core of an immunoassay is the affinity binder, which could be an anti-body, an antibody fragment or an alternative scaffold molecule. Optimallythis affinity binder should bind its target with both high affinity and speci-

42

Affinity proteomics

ficity and the binding event needs to be detectable. The affinity moleculecan be directly coupled to a detectable label, such as a fluorescent dye or anenzyme. Other options are to use a secondary detection agent that carriesthe detectable label or instead label the antigen [11,36]. Labeling antibodieswith DNA strands is another strategy that enables increased assay sensitivitydue to the signal enhancement obtained after DNA amplification [188, 189].Compared to MS, affinity proteomics has the advantage of very high sensitiv-ity with detection levels reaching sub-femtomolar levels [190–192]. However,one major issue for all antibody-based methods is the specificity as the accu-racy of the immunoassay is entirely dependent on the quality of the affinityreagent. The total signal output is converted to an abundance approximationof the analyte, with the assumption that nothing within the sample but theanalyte is contributing to the signal intensity. In reality this is not always thecase as proteins are likely to interact nonspecifically with both each other,surrounding molecules and surfaces. Since it is impossible to distinguish asignal generated from an antibody binding specifically to its target from asignal generated from unspecific binding or cross reactivity, antibodies needto be thoroughly validated before they can be used as affinity reagents inimmunoassays. The issue of antibody cross-reactivity will be discussed laterin this chapter.

planar array bead array

Figure 3.5: Different microarray formats.

ELISA is the golden standard of immunoassays with low detection levels, awide dynamic range and good reproducibility. However, scaling up ELISAexperiments for analysis of a large number of antibodies requires large sam-ple and reagent volumes. Immunoassays can instead be scaled up into amultiplexed microarray format to enable fast analysis of a large number of

43

Affinity proteomics

analytes in small reaction volumes [193]. The concept of microarrays; a setupof orderly arranged spots, where each spot is a separate reaction chamber,was first described by Roger Ekins in 1989 [194]. Microarrays can be planar,where molecules are covalently attached to a glass or silicon slide, resem-bling a highly ordered system of ELISA assays. An alternative to planararrays is bead-based microarrays, where molecules are instead attached tothe surface of small micro-scale beads. This has the advantage of faster ki-netics compared to planar formats [193] (Figure 3.5). With the LuminexxMap technology, up to 500 different color-coded paramagnetic beads canbe analyzed in parallel [195]. Each bead contains a precise ratio of two flu-orophores, making each bead type unique and possible to identify in a flowcytometer setup.

Microarrays can be divided into analytical, functional and reversed phaseprotein microarrays (Figure 3.6). Analytical microarrays, also called proteincapture arrays, is the most common setup and can be used to verify thepresence of a protein in a sample, to compare protein abundance in differentsamples, for biomarker discovery and affinity measurements [196]. Here, well-characterized capture molecules such as antibodies are immobilized onto thesolid support and a sample containing the analyte is thereafter added to themicroarray. For detection, the sample can be labeled with for example biotin[197]. By subsequently adding a fluorophore-labeled streptavidin molecule,bound proteins are visualized. The small size of biotin minimizes the risk ofinterfering with antibody target binding and moreover, the tight interactionbetween biotin and streptavidin ensures that the majority of bound analytemolecules will be detected. In a sandwich setup, unlabeled sample containingthe analyte is added to the array with immobilized capture antibodies anddetection is enabled by addition of a labeled detection antibody [11]. Thesandwich assay has the advantage of increased specificity, as two antibodiesare required for a signal output. However, due to issues concerning cross-reactivity as will be discussed below, multiplexing can become problematic[198].

Functional protein microarrays, also called target protein arrays, can for ex-ample be used to study biochemical activities, protein interactions or fordetection of autoantibodies [199]. Here, purified proteins are attached tothe solid support, thus including the need to produce large numbers of re-

44

Affinity proteomics

analytical microarray

direct labeling

sandwich

functional microarray

reversed-phase microarray

Figure 3.6: Setup for analytical, functional and reversed-phase protein microarrays.

combinant proteins. A labeled sample is added to the array and capturedmolecules are subsequently detected.

Reversed-phase microarrays are a good option if scaling up the number ofsamples instead of number of target proteins is desired, as compared toanalytical and functional arrays. In this approach, proteins from a complexsample, for example a cell lysate, are covalently attached to the solid support.A detection antibody is added to the microarray to detect the presence ofa specific protein across many samples. This approach can for examplebe applied to investigate protein modifications or other protein alterations[196].

45

Affinity proteomics

Multiplexing an immunoassay setup is desirable but not completely straight-forward. Cross-reactivity becomes an issue when multiplexing sandwich as-says as will be discussed in the following section. In addition the dynamicrange becomes problematic if target analytes have very different abundanceand since assay parameters cannot be optimized for each protein, com-promises are necessary [200, 201]. Comparisons made between multiplexedimmunoassays and classical ELISA assays however show that the resultsgenerally correlate well [202–205], even though multiplexed assays cannotusually match standard single-analyte immunoassays in terms of sensitiv-ity [193,203].

Cross-reactivity in immunoassays

False positive signals in immunoassays can be divided into sample-drivenand reagent-driven cross-reactivity (CR). Sample-driven CR is the resultof molecules in the sample reacting nonspecifically to one another or thesolid support. This type of CR causes problems in direct detection assays,where all sample molecules are labeled and can generate a signal output(Figure 3.7). In a sandwich setup the sample-driven CR is decreased, sincethe specific binding of two separate antibodies is required for analyte de-tection [198]. However, scaling up multiplexed sandwich assays has provento be difficult due to severe cross-reactivity issues and the upper limit ofmultiplexing is around 50-plex [198,200]. When applying a mix of secondarydetection antibodies (dAbs), these antibodies can interact with each other,the capture antibodies (cAbs) printed on the solid support or any of themolecules within the applied sample captured on the array. In addition,molecules within the sample can interact with each other or nonspecificallywith the cAbs. Pla-Roca and colleagues termed these unspecific interactions"liability pairs" and divided them into five different categories: cAb-antigen,antigen-antigen, dAb-cAb, dAb-antigen and dAb-dAb (Figure 3.7). In to-tal, the number of liability interactions is 4N(N-1), where N is the numberof targets. This CR, termed reagent-driven CR, therefore increases rapidlywith increased multiplexing [206]. In a multiplexed sandwich format, dueto the addition of dAbs in a mixture, it is enough if one of the dAbs showscross reactivity to other species in the sample for the whole assay to beaffected [198].

46

Affinity proteomics

cAb-antigen antigen-antigen dAb-cAb dAb-antigen dAb-dAb

reagent-driven cross-reactivity

sample-driven cross-reactivity

Figure 3.7: Different types of CR in immunoassays.Sample-driven CR occurs from labeled proteins within the ap-plied sample binding nonspecifically to other proteins, antibod-ies or plastics. Reagent-driven CR may occur in multiplexedsandwich assays when a mixed pool of dAbs is added to thearray. Unspecific interactions are in the figure represented bythe orange molecules.

In direct labeling approaches, where only one antibody is needed, the ob-served CR will not depend on the level of multiplexing and therefore thissetup is advisable for highly multiplexed experiments. Applying harsh wash-ing conditions, which will hopefully break unspecific interactions while re-taining stronger, specific interactions, is a strategy to decrease the CR. Usingthe sample labeling approach, multiplexed assays have been developed whereover several hundred targets have been identified simultaneously with detec-tion limits in the ng/mL range [207]. Several strategies to evade the issueof reagent-driven CR have been proposed. In the antibody colocalizationmicroarray (ACM), detection antibodies are not mixed, but are instead sep-

47

Affinity proteomics

arately added to the specific spots on the microarray [206]. Another strategyis sequential multiplex analyte capturing (SMAC), where one sample is in-cubated with different batches of antibody-coupled beads in sequence. Thebeads are thereafter incubated with the corresponding dAb [208]. Proximityligation assay (PLA) and proximity elongation assay (PEA) are two otherexamples of methods that can decrease reagent-driven CR. In PLA and PEA,two affinity reagents carrying complementary single-stranded DNA barcodesare needed for detection and upon simultaneous binding to the target pro-tein the DNA strands hybridize after which, qPCR is used for quantifica-tion [209–211].

CR is only a problem as long as signals from nonspecifically binding ana-lytes cannot be distinguished from signals arising from specific interactionsbetween antibody and target. This is the case in immunoassays since theoutput is the total signal intensity from the microarray spot, well or bead.As discussed earlier, MS enables specific detection by amino acid sequencedetermination. The combination of immunoenrichment using antibodies andMS readout can therefore to some extent overcome the problems of CR sinceseparate signals are obtained for different molecules and hence, the specificityof the affinity reagent becomes less crucial. Merging the fields of MS-basedand affinity proteomics will be discussed in chapter 4.

48

Chapter 4

Bridging affinity proteomicswith mass spectrometry

Benefits of combining immunoenrichment with mass spec-trometry

In MS analysis, where highly accurate protein identification is possible, sen-sitivity can be a limiting factor in the analysis of low abundant proteins.High abundant proteins are often troublesome, as they tend to quench sig-nals from lower abundant molecules. Targeted MS approaches can diminishthis problem, however target peptides then need to be chosen beforehandand non-targeted peptides will never be detected. Depleting the samplefrom high abundant proteins before analysis is also a strategy. However,this can lead to the loss of interesting proteins that interact with the de-pleted species. Researchers within the MS field are constantly strugglingto increase the sensitivity, especially since it is often low abundant proteinsthat are interesting from a clinical perspective [212]. Affinity proteomics onthe other hand struggles with specificity issues, as discussed above. If theaffinity agent is cross-reactive, this can hamper the whole assay and carefulvalidation of the specificity of affinity reagents is therefore crucial in orderto generate accurate results.

In other words, MS, which generates very specific data, has trouble with

49

Bridging affinity proteomics with mass spectrometry

sensitivity, and affinity proteomics, where very high sensitivity has beenachieved, suffers from inaccuracies due to potential antibody cross reactiv-ity. Combining the two techniques in a setup where a molecule of interestis first captured from a sample using an affinity reagent and thereafter an-alyzed using MS for verification of the target identity, brings together theadvantages of both technologies and an assay with both high specificity andincreased sensitivity is obtained.

Immunoenrichment setups

Immunoenrichment can be performed using different setups, depending onthe application and the affinity reagent. An antibody targeting a structuralepitope should be used for immunoenrichment of a full-length, structuredprotein. For an anti-peptide antibody, immunoenrichment from a trypticdigest would be preferred, if the epitope does not contain an arginine orlysine residue. In this case a proteolytic enzyme with different specificitycould be used. For quantification purposes using isotope-labeled standards,the immunoenrichment should preferably be performed after addition of theinternal standard, since the recovery after the antibody capture step will notbe complete and depends on the antibody affinity. This is not a problemif the spike-in standard is present during the immunoenrichment since therelative portion of bound protein or peptide will be equal for the standardand the endogenous molecule. If the enrichment is instead performed beforethe spike-in, a portion of the target molecule will be lost at this stage andit will be more difficult to obtain an accurate estimation of protein abun-dance. The affinity of the capture reagent towards its target would in thiscase need to be determined beforehand to enable an approximation of thetotal amount of protein or peptide present in the sample from the start. If,however, a full-length protein is used as quantification standard, immunoen-richment can be performed either on protein or peptide level. Performingthe enrichment experiment on the protein level would possibly generate datafrom multiple peptides using only one capture reagent. To generate the sameamount of data from an experiment performed on the peptide level, differentanti-peptide antibodies would be needed, making immunoenrichment on theprotein level advantageous in this case. However, since AQUA peptides are,as mentioned in chapter 2, the most common strategy for absolute quantifica-

50


tion, immunoenrichment on the peptide level commonly applied for proteinquantification purposes.

Peptide enrichment coupled to mass spectrometry

Enrichment of tryptic peptides with subsequent MS readout has become pop-ular for quantification of proteins in complex samples, primarily plasma. Thestable isotope standards and capture by anti-peptide antibodies (SISCAPA)technology, developed by Anderson and coworkers, is the most widespreadstrategy for this purpose [76] (Figure 4.1). The SISCAPA protocol con-sists of four steps: (1) tryptic digestion of the sample, (2) addition of heavyisotope-labeled peptide standards, (3) immunoenrichment using anti-peptideantibodies and (4) peptide quantification using MS. The peptide capturestep can result in an enrichment factor of above 1000 [213, 214], increasingthe detection limits to the lower ng/mL range for targeted MRM assays inplasma [214–216] and saliva [217]. Detection limits as low as pg/mL havealso been reported, however large plasma volumes (1 mL) were used in thiscase [218]. The SISCAPA assay has been multiplexed to enable simultane-ous detection of up to 50 targets, with retained quantification accuracy ascompared to single-plex assays [219].

SISCAPA quantification is generally performed using liquid chromatographycoupled to MRM MS, however other setups have also been tried. For exam-ple SISCAPA enrichment and MALDI MS quantification showed promisingresults when six different peptides were quantified in human plasma withcoefficients of variation below 4% [220]. The simplicity, robustness and pos-sibility of high-throughput measurements using MALDI makes it a very at-tractive platform for analyzing clinical samples, however obtaining quantita-tive precision and reproducibility has proven to be more difficult when usingMALDI MS, compared to ESI [221]. SISCAPA enrichment has also beenperformed in combination with an LC-free MRM setup for peptide quan-tification, where the cycle-time was reduced 300-fold, although with slightlyhigher coefficients of variation [222].

Another method combining peptide immunoenrichment with protein quan-tification using MS is iMALDI (immuno-MALDI), developed by Borchersand colleagues [223–225] (Figure 4.1). The workflow is similar to SISCAPA

51


protein mixture peptide mixture

digestion

heavy isotope-labeled standard

SISCAPA work!ow

iMALDI work!ow

m/z

intensity

ESI-MS

MALDI-MS

m/z

intensity

MALDI-MS

Figure 4.1: Peptide enrichment coupled to MS quantifica-tion. Proteins are digested and heavy isotope-labeled peptidesare added to the digested sample before immunoenrichmentusing anti-peptide antibodies. Peptides are further analyzedusing ESI-MS (SISCAPA) or MALDI-MS (iMALDI).

in that peptides from a trypsin-digested sample are captured using anti-peptide antibodies and analyzed using MS. However, in iMALDI, as thename implies, analysis is performed using MALDI MS. The iMALDI proto-col is simple as beads containing bound peptides are directly spotted onto aMALDI target, hence eliminating the need for peptide elution. Studies usingthe iMALDI methodology have been focused towards disease diagnosis, forexample through the analysis of peptides from Francisella Tularensis for thediagnosis of tularemia [225]. The cancer biomarker EGFR has been analyzedby enrichment of EGFR peptides from various cancer cell lines and tumorbiopsies and EGFR could be detected in a sample corresponding to less thanten EGFR-expressing breast cancer cells [224]. Furthermore, iMALDI hasalso been applied in the diagnosis of hypertension by analysis of the peptidesangiotensin I and II in plasma [226–228]. One limitation of the iMALDI tech-nology is the inability to multiplex the analysis to more than 5-10 targets,as MALDI cannot be coupled on-line to a liquid chromatography system tofractionate peptides [220]. The iMALDI setup has however been used for

52


simultaneous capture and quantification of two peptides (angiotensin I andII) in plasma, with a limit of detection in the pg/mL range [227].

Apart from that these setups can be easily combined with quantificationusing AQUA peptides, one advantage of using peptide immunoenrichmentis that a peptide sample is generally easier to handle than a protein sam-ple, due to degradation, unfolding and solubility issues often associated withhandling of full-length proteins. The wide variation of physicochemical prop-erties of proteins also makes optimization of experimental conditions easierfor enrichment using peptide samples [229]. In addition, production of full-length proteins is quite troublesome compared to the expression of proteinfragments or generation of synthetic peptides [230], wherefore most avail-able antibodies are generated towards a peptide or protein fragment and aretherefore likely to be functional in peptide enrichment setups.

Protein enrichment coupled to mass spectrometry

Due to the reasons mentioned above, protein immunoenrichment can bedifficult, however the increased protein sequence coverage acquired using en-richment on the protein level, as opposed the peptide level, makes this anattractive alternative. In addition, performing the immunoenrichment priorto trypsin cleavage is advantageous from an economical perspective, espe-cially when samples of high protein content are analyzed, such as plasmaor serum. The sample complexity is after immunoenrichment significantlydecreased, lowering the required amount of costly enzyme to a fraction ofthe quantity needed to digest the original sample. Moreover, protein im-munoenrichment can generate information regarding protein complexes andinteraction networks, which is information that is completely lost when per-forming immunoenrichment at the peptide level.

The mass spectrometric immunoassay (MSIA) was developed already in 1995by Nelson and colleagues [231] and since then, several studies using thistechnology have been described [232–234]. MSIA differs from SISCAPA andiMALDI as full-length proteins instead of peptides are captured in the im-munoenrichment step (Figure 4.2). In MSIA, the sample is incubated withantibody-coated beads in a pipette-tip format and subsequently eluted withMALDI matrix prior to MALDI MS readout. In the first publication, two

53


toxins were analyzed in human blood. Here, full-length toxins were cap-tured using two different antibodies and were then analyzed using MALDIMS without enzymatic digestion. Further development of the system hasled to a fully automated setup [235] and the MSIA approach has also beenused with following trypsin digestion along with quantification using heavyisotope-labeled spike-in peptide standards and targeted MRM MS analy-sis [236, 237]. Several different setups have been developed using full-lengthinternal standards for protein quantification. For example, rat and pig pro-teins have been used as internal standards to determine the concentration ofthe corresponding human protein. The molecular weights of the homologsare not identical, enabling differentiation between the variants in the MSspectrum. For this setup to be functional, an antibody with cross-speciesspecificity is required [238, 239]. Unrelated proteins have also been used asstandards, however this setup requires separate antibodies for capture oftarget protein and internal standard [240, 241]. Limits of quantification inthe range of ng/mL have been achieved both with and without subsequenttrypsin digestion [236, 241]. As discussed above, immunoenrichment of full-length proteins has the advantage that protein isoforms and modified proteinvariants can be distinguished from one another and MSIA has been used toidentify different variants of for example the protein cystatin C [240] andtransthyretin [241].

Another application of protein immunoenrichment coupled to MS is to verifythe identity of captured proteins in bead-based immunoassays. After iden-tifying differences in signals between diseased and control plasma samples,it is crucial to verify that the difference in signal intensity originates fromdifferences in abundance of the target protein between the two samples andnot unspecific binding to antibodies or beads. This has been performed toverify the specific capture of C2, C8 [70] and carnosine dipeptidase 1 [242]in plasma.

A large amount of information can be obtained from immunoenrichment ofprotein complexes in combination with MS readout. Different setups can beapplied, the most straightforward being to use a target-specific antibody tocapture the protein complexes involving the target protein [243–245] (Fig-ure 4.3). Usually cell lines are used for this purpose and by applying avery mild lysis strategy, protein interactions may be maintained. However,

54


elution and digestion

m/z

intensity

time

intensity

MRM MS

MALDI MS

elution

Figure 4.2: Workflow for the MSIA method. Protein immu-noenrichment is performed in a pipette-tip format and elutedproteins are either digested and analyzed using MRM MS orproteins are directly analyzed using MALDI MS.

the fact that enrichment conditions cannot be standardized for all antibod-ies in combination with the limitation that validated antibodies do not ex-ist for all targets hinders this approach [246]. Instead, cells expressing atagged version of the target protein (the bait) can be used in combinationwith anti-tag antibodies to capture the components of protein complexes(the prey) [52, 246–248]. One common strategy is tagging proteins withthe tandem affinity purification (TAP) tag, which enables purification intwo steps, both under native conditions. The tag consists of a calmodulin-binding peptide and a Protein A IgG-binding domain separated by a TEVlinker [249, 250]. First, proteins are purified on an IgG column and boundproteins are eluted by TEV cleavage to leave the Protein A domain on thecolumn. The sample is thereafter subjected to a calmodulin column, fromwhich proteins are eluted by the addition of EGTA. Other common tags arethe FLAG tag and the GFP tag, that both bind to corresponding tag-specificantibodies [251–253]. Although these tagging approaches enable large-scaleanalysis and eliminate the problem with variation in antibody affinity andspecificity, there are some issues with tagging endogenous proteins. For ex-

55


ample, verifying that the tag does not affect the protein in terms of expres-sion, subcellular localization and interaction characteristics can be difficult.Altered protein expression could change the equilibrium within the cell andthereby alter the composition of the studied protein complex. In humancells, bacterial artificial chromosome (BAC) transgenomics is typically usedto generate clones with tagged bait proteins, which generates a stable ex-pression of the tagged protein [254].

In order to preserve weak interactions, very mild washing conditions are usu-ally applied, which also increases the amount of nonspecifically interactingproteins that remain bound to the antibodies or the beads. Consequentlyit is of great importance to include good controls in the analysis in orderto distinguish specific interaction partners from nonspecifically interactingproteins. Databases of contaminating proteins usually seen in immunoen-richment experiments, such as bead proteomes [255], protein frequency li-braries (PFL) [256] and the CRAPome [257] can be helpful in identifyingtrue interactors.

Furthermore, SILAC-based methods can be used to study protein interac-tions, where the bait protein is tagged in either the unlabeled or labeledsample. Specifically interacting proteins will be enriched only in the sam-ple and not in the control, resulting in high or low heavy to light ratios.Nonspecifically interacting proteins on the other hand will be enriched to asimilar extent in both sample and control and the heavy to light ratios forthese proteins will therefore be close to one [258, 259]. Mixing the samplesbefore immunoenrichment can however lead to exchange of dynamic inter-action partners between the two samples, which may lead to false-negativeresults with specific interaction partners generating ratios close to one. Tocircumvent this problem, immunoenrichment can instead be performed be-fore mixing the samples [260]. Experiments involving three different SILAClabeling states have also been performed to make it possible to distinguishbetween constitutive and signal-dependent interaction partners [261]. An-other example is QUICK (quantitative immunoprecipitation combined withknockdown), where SILAC labeling is combined with RNA interference. Ineither the labeled or unlabeled sample the target protein has been knockeddown. This leads to large differences in intensities for peptides originatingfrom the knocked down protein and its interaction partners, whereas non-

56


m/z

intensity

m/z

intensity

m/z

intensity

m/z

intensity

m/z

intensity

m/z

intensity

m/z

intensity

m/z

intensity

m/z

intensity

m/z

intensity

m/z

intensity

m/z

intensity

m/z

intensity

m/z

intensity

m/z

intensity

target-speci c Abs

label-free MS

tag-speci c Abs

label-free MStag-speci c Abs

SILAC labeling

Figure 4.3: Strategies for analysis of protein interactionsusing immunoenrichment coupled to MS. Immunoenrichmentusing (left) target-specific antibodies and label-free MS quan-tification, (middle) tag-specific antibodies and label-free MSquantification and (right) a SILAC-based approach where thetarget protein is tagged only in the light sample. In all casesspecific interaction partners (blue and pink) will be presentonly in the sample and not in the control, whereas nonspecifi-cally interacting proteins (grey) will be present in both sampleand control.

57


specific binders will show ratios close to one. In this strategy no tagging ofthe bait protein is performed and the method hence makes use of anti-targetinstead of anti-tag antibodies [243].

Immunoenrichment of protein or peptide groups

Proteins can be modified in a number of different ways and different mod-ifications can for example regulate protein activity or localization [39]. Inmany cases, the analysis of protein modifications can be far more informativethan determining the protein concentration. However, due to the generallylower abundance of modified peptides versus their unmodified counterparts,analysis and identification of PTMs is challenging. Antibodies are thereforecommonly used to enrich groups of peptides or proteins carrying a certainmodification, although the specificity of these antibodies is usually lowerthan for target-specific antibodies [262].

Immunoenrichment of phosphorylated peptides [77, 263–265] and proteins[95,264,266] is feasible with antibodies recognizing phospho-tyrosine residues,which have shown higher specificity than antibodies targeting phosphory-lated serine and threonine [267]. Ubiquitination can be studied with antibod-ies specific for mono- or poly-ubiquitinated proteins [268–271]. Upon trypticdigestion of ubiquitinated proteins, a glycine-glycine remnant remains at theubiquitination site and peptides carrying this modification can be capturedusing an antibody specific to this site [272,273]. Antibodies specific for differ-ent ubiquitin chain conformations have also been generated [274,275], whichcould generate a very detailed view of the ubiquitinated proteome. In addi-tion to phosphorylation and ubiquitination, antibodies have been used in im-munoenrichment setups to analyze for example lysine acetylation [276,277],methylation [278,279] and nitration [280,281].

Another possibility is to use antibodies specific for certain sequence mo-tifs, hence targeting multiple proteins. Wingren and coworkers developedthe global proteome survey (GPS), in which antibodies targeting short C-terminal motifs of four to six amino acids are used to enable a proteome-wide analysis using only a few hundred antibodies instead of one antibodyper target protein [282]. The antibodies are called context-independent mo-tif specific (CIMS) antibodies and are selected from a large (2x1010) scfv

58


library. A similar approach is the Triple X Proteomics (TXP) technology,developed by Poetz and colleagues in which, polyclonal affinity-purified an-tibodies are used to target short C- or N-terminal regions of three to fiveamino acids [283,284].

59


60

Chapter 5

Miniaturization

Miniaturization in proteomics

Miniaturization of experimental setups generating micro total analysis sys-tems (µTAS) or lab-on-a-chip devices, is becoming more and more populardue to advantages such as shorter analysis times, lower reagent consump-tion and the possibility for high-throughput and automated analysis [285].These platforms usually have dimensions in the micrometer range and handlesample volumes of a few microliters down to picoliters [286]. This is a rela-tively new field with the first analytical microsystem (a gas chromatograph)developed in 1979 by Stephen Terry [287].

When an object is down-scaled isomorphically, the length, area and volumeratios are altered, which has effects on certain physical properties. In thebook "Fundamentals of Microfabrication: The Science of Miniaturization",Marc Madou compares the downscaling of analytical assays to differences inproperties of small and large animals. For example, the heat loss of a livingcreature is proportional to its surface area and is therefore very differentfor an elephant compared to a small pygme shrew. The pygme shrew musteat constantly or otherwise freeze to death, a non-existing problem for thelarge elephant [288]. The same can be said about downscaling analyticalsystems. The larger surface area to volume ratio for miniaturized analysisplatforms affects for example evaporation characteristics [289]. In addition,in miniaturized systems, surface and interfacial tension have a larger impact

61

Miniaturization

as compared to macro scale setups, where gravity is the more dominantforce [290].

Scaling down a reaction chamber will in general not change the reactionkinetics, although a common strategy for protein and peptide separationis to use bead-based systems, where the locally high analyte concentrationon the beads can alter the kinetics of the reaction. The density of theimmobilized molecule can in such a setup reach higher levels than would bepossible in solution due to precipitation issues. Moreover, miniaturizationwill increase the reaction rate. This can be explained by the fact that moretime will be required for two molecules present in a large sample volume tocome in contact with one another than the time required in a much smallervolume. A reaction requiring 1 h in a standard microtiter plate well can ina miniaturized device with a 50 µm reaction chamber take a few seconds[285]. The diffusion of molecules in a macro-scale system such as a testtube or a well in a microtiter plate is negligible and mixing is performedby convection (hand-shaking or using a vortex) [291]. However, as diffusiontime is decreased with decreased size of the reaction chamber, mixing bydiffusion only is usually enough in a miniaturized system [292].

In addition, miniaturization reduces both reagent and sample volumes, andby that also waste volumes, resulting in a decreased cost per assay. If amolecule is present in a sample at a very low concentration, the volumeneeded to detect the molecule might be larger than what can be used in themicro-device. However, depending on the setup of the miniaturized device,a large sample volume can in many cases be used in combination with avery small sample elution volume using for example solid phase extraction,resulting in a large enrichment factor and increased analyte concentrationin the eluate [293]. Miniaturization also facilitates automation and integra-tion of analytical steps and further enables the use of portable, single-useplatforms that is very beneficial in the clinic [294]. Despite these many ad-vantages, miniaturized analysis systems have not yet become as widespreadin for example clinical applications as was anticipated a couple of years agoand Sackmann reported earlier this year than the majority of microfluidicsarticles are still published in engineering journals [290]. The difficulty tointegrate these methods into a standard laboratory setting and the troublefor non-experts to use the methods could partly explain this [295]. However,

62

Miniaturization

the field is still growing rapidly, which is promising for the future.

In MS-based proteomics, as discussed in chapter 2, proteolytic protein di-gestion is very often applied prior to MS detection. Digestion using a prote-olytic enzyme can be incorporated into a miniaturized analysis system. Theenzyme can then for example be immobilized onto beads after which, thesample is passed through the chromatographic resin or alternatively, the di-gestion reaction can be performed using a droplet-based system where eachdroplet acts as a unique reaction compartment. With these small volumes,the digestion time can be down-scaled from overnight incubation in a stan-dard setup in a test tube to a couple of minutes [296–299].

Miniaturized analytical devices have been developed that integrate differentsample preparation steps with mass spectrometry readout [293,296,300,301],one of them being the ISET platform, discussed below.

Figure 5.1: The ISET platform. An array of nanovials inthe ISET chip can be filled with chromatographic matrix forenrichment of target molecules. Molecules are eluted to thebackside of the chip that serves as a MALDI target.

63

Miniaturization

The ISET platform

The integrated selective enrichment target (ISET) platform integrates sam-ple enrichment, proteolytic digestion and MALDI MS readout in a singlechip. ISET sample preparation is bead-based and theoretically, any type ofbead (with the right dimensions) can be used in the ISET platform in orderto enrich for proteins or peptides of interest. The ISET device is a small(54x39 or 44x39 mm) silicon chip with 96 or 48 nanovials that can be filledwith the chromatographic matrix of choice [302]. At the bottom of eachnanovial there is a membrane with 3x3 outlet holes of 4,500 or 8,100 µm2.By attaching the device to a vacuum manifold, liquid can easily be pulledthrough the vials at a desired speed. On-bead protein digestion can be per-formed directly in the nanovials. Before elution the applied vacuum pressureis lowered, which will cause the elution liquid to form a droplet attached tothe backside of the chip. This side of the chip acts as a MALDI target andafter crystallization of eluted sample together with a MALDI matrix thechip can be directly inserted into a MALDI mass spectrometer for analysis(Figure 5.1).

The integrated format of the device minimizes sample loss, since no sampletransfers are needed between sample application and data generation. Atypical ISET workflow is shown in Figure 5.2. The simplicity of the deviceand the dimensions of the vials makes it possible to integrate the devicewith robotics and sample handling platforms. In addition, the possibility tochoose between a large collection of chromatographic resins makes the ISETplatform exceptionally versatile.

1. Load beads

2. Wash

3. Sample binding

4. Wash 5. Tryptic digestion 6. Elution 7. MALDI analysis

Figure 5.2: A typical workflow for the ISET platform.

64

Miniaturization

The ISET chip was first described in 2004 by Ekström and coworkers [303].Initially, the application of the chip was as a sample cleanup device whereC18 beads were used for peptide desalting [303–305] and the ISET plat-form showed superior performance compared to two other available peptidedesalting methods; ZipTips and MassPREP PROtarget [304]. Further de-velopment of the ISET chip has been performed to increase the conductivityof the chip [306] and to optimize the nanovial outlet design [302]. Proteindigestion has also been integrated into the ISET workflow [307] and so far,the ISET platform has been used for peptide desalting with reversed phasematrix [303–306], antibody-based enrichment to detect prostate-specific anti-gen from seminal plasma [307] and detection of thrombin in human serumby capture on aptamer-functionalized beads [308].

65

Miniaturization

66

Chapter 6

Present Investigation

Summary

I have now tried to summarize the field in which I have been working forthe past couple of years. I hope that it has become clear why studying ourproteins is so essential, both to reach a deeper understanding about how wefunction but also in order to understand disease progression and for the de-velopment of drugs for treatment of disease. This chapter will summarize theresearch that I have been working on. The main focus has been on develop-ment of high-throughput methods for protein analysis and quantification. Ihave worked both with production and analysis of recombinant proteins thatcan be used in proteomics research (papers I and II), but also methods forthe analysis of endogenous proteins in human cell lines (papers III-V).

The projects that I will present have all been performed within the HPAproject, described in chapter 3. This is a large-scale proteomics project withgreat resources both regarding instrumentation, but also human knowledgeand manpower, which for a graduate student of course is extremely valu-able. In the presented projects, both polyclonal antibodies generated withinthe project, but also monoclonal antibodies generated by the spin-off com-pany Atlas Antibodies have been used. We have also taken advantage ofthe antigens generated within the project. These protein epitope signaturetags (PrESTs) have been produced with heavy isotope-labeled amino acids,making it possible to use them as internal standards for MS-based absolute

67


quantification in different setups.

The different projects will be described in the following sections. In paper I,the high-throughput protein production setup used within the HPA projectwas scaled down to a 96-well plate format. The ISET plate, described inchapter 5, was in paper II used to further scale down the IMAC purifica-tion to a nL-scale format. PrEST antigens were in paper III produced withheavy isotope-labeled amino acids and used as internal standards for deter-mination of protein copy numbers in six different cell lines using MS. Thecopy numbers were further compared to mRNA levels determined by RNAsequencing. Paper IV describes how antibodies generated within the HPAproject and Atlas Antibodies were used in immunoenrichment coupled tomass spectrometry to investigate whether the antibodies can bind to theirnative target in a complex sample and also to determine their specificity. Theimmunoenrichment was in paper IV performed on the protein level, howeverin paper V, antibodies were used to capture tryptic peptides from a digestedcell lysate. In this paper, PrESTs were also used as internal standards, whichwere spiked into the sample directly after cell lysis to enable absolute proteinquantification.

Development of screening methods for recombinant proteinproduction (papers I and II)

Proteins are diverse molecules with varying properties, wherefore proteinproduction can be performed in very different ways. Optimization of theproduction parameters can be performed in order to increase the proteinyield. However, in large-scale protein production projects, where hundredsof proteins are produced every week [179], optimization is not possible inthe same way. Instead, a protocol that maximizes the overall success rate isdesired. Within the HPA project, the success rate over the antigen produc-tion step is above 80%, although since the protein verification is performedas the last step in the production pipeline, a large amount of time and effortis spent on proteins that in the end will not be of use. It would therefore bevery beneficial to be able to exclude these proteins at an early stage in theprocess.

One possible solution is to include a small-scale screening step before the

68


large-scale protein production [309–311]. In this way, non-expressing or im-pure proteins can be excluded at an early stage, although it is of greatimportance that the different setups generate corresponding outcomes. Inpaper I, a µL-scale method for protein production screening was developed.The results were subsequently compared to data obtained from the HPAprotein production pipeline. Protein concentration, purity and molecularweight verification were compared between the methods (Figure 6.1).

m/z

intensity

MALDI MSSDS-PAGE

IMAC

Cultivation fail fail

success

fail

SDS-PAGE

fail

success

fail

m/z

intensity

MALDI MSIMACCultivation

fail

A

B

Figure 6.1: Workflow for (A) the standard protein produc-tion and validation pipeline and (B) the small-scale screeningsetup.

E. coli cells expressing recombinant proteins were cultivated in 96-well platesusing only 150 µL of culture medium. To enable cultivation in this small for-mat, a fed-batch-like system for cell cultivation called the EnBase

R© culturetechnology was applied [312]. During cultivation in a glucose-rich medium,cells will initially grow very fast, resulting in an accumulation of acetate inthe medium, a process known as overflow metabolism. In a fed-batch sys-tem, glucose is constantly applied to the medium in a manageable amount,which leads to a more controlled cell growth and finally, to higher cell den-sities [313]. In the EnBase

R© system, glucose is supplied to the cells by

69


constant, enzymatic degradation of starch in the culture medium, whichresembles a fed-batch setup. One advantage with this system compared toconventional fed-batch cultivation is the simplicity, as starch and enzyme aresimply added at the start of the cultivation. This setup makes cultivationin small volumes possible, which is not applicable for standard fed-batchsystems. In paper I, the EnBase

R© technology was first compared to thestandard HPA cultivation method. Shake flasks with 100 mL tryptic soybroth medium were used for the standard cultivation and EnBase

R© culti-vation was performed in deep well plates (DWP) with 3 mL medium. Theadvantage of the EnBase

R© technology was apparent as 96% of the proteinsshowed increased yields when comparing the obtained amount of proteinafter purification in relation to cultivation volume. A 21 times maximumincrease in protein yield per mL culture volume was observed, with an av-erage of 3.3 times increased yield for the EnBase

R© cultivation (Figure 6.2).The significant increase in produced protein for the EnBase

R© system madeit possible to downscale the cultivation to 150 µL, enabling cultivation in a96-well plate format.

Protein purification was also successfully performed in a 96-well format.IMAC was used for this purpose, which is also the purification method usedin the standard HPA workflow, however the amount of matrix used wasdownscaled from 1 mL to 12.5 µL in the screening method (80 times). Thepurification was performed in a filter plate, which enabled efficient and fastpurification with the aid of a vacuum manifold.

In total, 96 PrESTs were analyzed in triplicate using the two methods. Inthe standard HPA workflow, validation after protein production and purifi-cation includes SDS-PAGE analysis for determination of protein purity andMS analysis of full-length proteins for molecular weight verification. Pro-teins can be failed in each of these steps. For the small-scale filter plateformat, samples were at first only analyzed using MS. This led to an agree-ment between the methods of 91% (87 out of 96 PrESTs). The majority ofthe proteins (n=82) were produced successfully in both setups, whereas afew proteins (n=5) showed no expression using either of the methods. Six ofthe proteins that did not have coinciding results were failed in the standardsetup due to impurities seen on the SDS-PAGE, however these contaminantswere not observed in the small-scale setup where no electrophoretic analy-

70


1 3 5 7 91

11

31

51

71

92

12

32

52

72

93

13

33

53

73

94

14

34

54

74

95

15

35

55

75

96

16

36

56

76

97

17

37

57

77

98

18

38

58

78

99

19

39

5

µg p

rote

in/m

L c

ulture

media

PrEST (1−96)

0

100

200

300

400●

●

EnBase DWP

shake flask

Figure 6.2: Comparison of standard protein production inshake flasks with EnBase

R© cultivation in DWPs. The amountof protein obtained per mL culture volume in the standardcultivation and the EnBase

R© DWP cultivation is shown forall 96 proteins.

sis was performed. When including this step in the small-scale setup, theagreement increased to 95% (91 out of 96 PrESTs). These highly coincidingresults verify that this small-scale screening method can be used to accu-rately predict the outcome from protein production and purification using alarge-scale setup. This could reduce both the time and cost spent on proteinsthat will not be possible to produce under standardized conditions.

In paper II, the protein purification setup was downscaled further and theaim was to develop a method for recombinant protein verification using theISET platform, discussed in chapter 5 (Figure 6.3). The integrated for-mat makes this platform very suitable for automation and in this paper, tworobotic systems made it possible to automate all steps until MALDI analysis.A method was developed in which the His6-tagged PrESTs present within abacterial lysate from the HPA production pipeline were first retained in theISET vials by Co2+ coated beads. This type of beads had not earlier been

71


IMAC

Cultivation

Digestion Spotting and crystallization m/z

intensity

MALDI MS

m/z

intensity

MALDI MS

ISET puri!cation and digestion

A

B

Figure 6.3: Workflow for (A) the ISET protein screeningsetup and (B) the standard HPA protein purification workflowwith the addition of a tryptic digestion step. In both setupsprotein production was performed in 1 L shake flasks.

used in combination with the ISET platform and hence this investigationshows the possibility to integrate the ISET platform with commonly usedlaboratory methods. All steps were performed on a vacuum manifold, whichenabled efficient liquid transfer through the vials. After washing to removeunbound proteins, PrESTs were digested directly in the vials of the ISETchip. In this step, the vacuum was turned off to prevent enzyme and peptidesfrom leaving the vials. Due to the small vial volumes, efficient digestion wasperformed in only 1 h. The resulting peptides could be eluted by additionof MALDI matrix and applying a low vacuum pressure, which enabled for-mation of droplets on the backside of the ISET chip. After crystallization ofthe droplets, the chip was inserted into a MALDI mass spectrometer and thepeptides were analyzed using both PMF and MS/MS. A total of 45 proteinswere included in the analysis and they were all analyzed in triplicate usingboth the new ISET protein verification workflow and the standard proteinproduction and purification setup used within the HPA project. To enablea more direct comparison between the two methods, a tryptic digestion stepwas performed on the samples from the standard HPA workflow.

PMF resulted in identification of 42 out of the 45 PrESTs for the ISET setup.The success rate for MS/MS analysis was slightly lower with 41 successfullyidentified proteins. Three PrESTs were not identified with either PMF orMS/MS using the ISET method. When determining the sequence coverage,

72


1 2 3 4 5 6 7 8 91

01

11

21

31

41

51

61

71

81

92

02

12

22

32

42

52

62

72

82

93

03

13

23

33

43

53

63

73

83

94

04

14

24

34

44

5

ide

ntifie

d u

niq

ue

pe

ptid

es

PrEST (1−45)

0

5

10

15

20●

●

ISET

standard setup

Figure 6.4: Comparison of number of identified unique pep-tides for the ISET screening method and the standard HPAsetup using PMF.

peptides covering the entire protein sequence were included. However, sincethe PrESTs all contain a common His6ABP tag, as discussed in chapter 3,detection of a unique peptide originating from the PrEST part was requiredfor verification (Figure 6.4). One of the unidentified PrESTs (PrEST 17)did not contain any tryptic peptides in the chosen mass range and therefore,identification of this protein was theoretically not possible. One of the re-maining two PrESTs (PrEST 40) showed very low expression, resulting in alow protein concentration, explaining why this protein could not be identi-fied. Results obtained using the ISET setup were compared to results fromthe standard HPA workflow. After in solution digestion and MALDI PMFand MS/MS, 39 out of 45 PrESTs were successfully identified. Even thoughfewer proteins were identified after protein purification and digestion usingthe standard method, higher average sequence coverage was obtained. PMFresulted in an average sequence coverage of 68% and 78% for the ISET andstandard methods respectively (Figure 6.5) and the corresponding numbersfor MS/MS analysis were 54% and 64%. The ISET method was also com-pared to the setup usually applied in the HPA project with MS analysis offull-length proteins. Here, 40 proteins were identified and hence, the ISETsetup could increase the success rate in this step.

73


1 2 3 4 5 6 7 8 91

01

11

21

31

41

51

61

71

81

92

02

12

22

32

42

52

62

72

82

93

03

13

23

33

43

53

63

73

83

94

04

14

24

34

44

5

sequ

ence c

ove

rag

e P

MF

(%

)

PrEST (1−45)

0

20

40

60

80

100

120●

●

ISET

standard setup

Figure 6.5: Comparison of degree of sequence coverage (%)for the ISET screening method and the standard HPA setupusing PMF.

Apart from comparing sequence coverage and success rate between the twomethods, an important aspect is the total time needed for analysis and es-pecially the required hands-on time. For analysis of 48 samples using theISET method, the time required is approximately 4 h excluding MS analysis.This includes only 30 min hands-on time, which is needed for preparationof the robotic systems. In comparison, the time needed for the standardmethod is approximately 25 h, where a large part (around 16 h) is due tothe overnight tryptic digestion. The difference in hands-on time is smallersince the standard method uses an automated liquid handling system forprotein purification, however approximately 1-1.5 hours is needed for man-ual sample handling. A big advantage with the ISET method is the lowsample requirement. Only 6 µL bacterial lysate was used, which is a neg-ligible fraction of the total lysate volume after cell cultivation (0.1%). TheISET workflow could therefore offer a possibility of protein verification beforelarge-scale protein purification. Since the IMAC matrix is rather expensive,an initial screening for protein identity in a small scale could be very advan-tageous.

74


Absolute MS-based protein quantification to study the corre-lation between protein and mRNA levels (paper III)

As discussed in chapter 1, proteins are produced from mRNA in a processcalled translation. Methods for large-scale mRNA analysis, called RNA se-quencing have been developed and generation of large amounts of RNA datais today feasible [314, 315]. However, large-scale protein analysis is not asstraightforward and getting a complete view of a proteome is today verydifficult, if not impossible. The analysis of mRNA in order to gain knowl-edge about protein function is therefore a tempting approach. By analyzingwhich mRNA transcripts are present in high and low levels, some conclusionscan be drawn regarding which proteins are expressed in what cell types andat what levels. The studies that have been performed comparing mRNAand protein levels are numerous and the results have been varying. Somehave claimed that no correlation exists between mRNA and protein levelsand other studies have shown correlations as high as 0.9 (Spearman’s rankcoefficient) [40, 47, 48, 58, 316]. A trend seems to be that more recent stud-ies usually report higher correlations than do older studies, indicating thatinstrumentation and technology has had an impact in this area. In recentstudies MS is usually the method of choice for protein quantification andstrategies such as SILAC labeling [48] and label-free approaches [40,58] havebeen used for this purpose. In paper III, another approach, PrEST-SILAC,was applied in which heavy isotope-labeled protein fragments are used as in-ternal standards for absolute protein quantification. The protein fragmentsare the antigens (PrESTs) developed within the HPA project, which canbe used directly as they are (without isotopic labels) as internal standardsin mass spectrometry for quantification of proteins in heavy isotope-labeledcell lines. However, PrESTs can also be produced with heavy isotope-labeledarginine and lysine residues in order to generate "heavy" PrESTs that can beused for quantification of proteins in unlabeled samples [127], as describedin chapter 3.

In total, 57 PrESTs corresponding to 32 genes were here chosen for analysisin six cell lines (A-549, HEK 293, HeLa, Hep-G2, RT-4 and U-2 OS). Theprotein fragments were produced with heavy isotope-labeled arginine andlysine residues and the proteins were quantified in the different cell linesusing the PrEST-SILAC setup (Figure 6.6). PrESTs were spiked into the

75


light cell line heavy PrEST

mix

digestion

protein extraction

m/z

intensity

MS analysis

MS spectrum

fractionation

Figure 6.6: Workflow for the PrEST-SILAC method. HeavyPrESTs are spiked into a cell lysate. After digestion andfractionation the sample is analyzed in a mass spectrometer.Heavy to light ratios for peptides corresponding to the addedPrEST sequences are used for absolute quantification of theendogenous proteins.

cell lysates directly after cell lysis, minimizing the errors introduced due tosample handling. Samples were digested and fractionated into six portionsusing strong anion exchange chromatography in a pipette tip format prior toMS analysis to determine the absolute protein concentration. The resultingprotein copy numbers were further compared to mRNA data obtained fromRNA sequencing.

76


1e+04

1e+05

1e+06

1e+07

1e+08

1e−

02

1e+

00

1e+

02

1e+

04

FPKM (RNA sequencing)

pro

tein

copy n

um

ber

per

cell

Spearman correlation

R = 0.46

Figure 6.7: Comparison of protein and mRNA levels for 32genes in six different cell lines. A Spearman’s rank coefficientof 0.46 (p=7e−10) was determined.

Protein quantification was performed using triplicate samples and a mediancoefficient of variation of 17.2% was obtained across all cell lines. Quantifica-tion was performed using between one and 18 peptides, with an average valueof 3.8 quantified peptides per protein. To further validate the accuracy ofthe method, unlabeled (light) PrESTs were also used as internal standards.For this experiment, three of the cell lines (A-549, HEK 293 and HeLa)were cultured in medium containing heavy isotope-labeled amino acids. Theresulting copy numbers were compared and the Pearson correlation was de-termined to 0.76.

Protein copy numbers determined in the six cell lines were subsequentlycompared to mRNA data obtained by RNA sequencing. First, copy numbersfrom all proteins in all six cell lines were compared to the corresponding RNAdata. The correlation was determined to 0.46 (Spearman’s rank coefficient,p=7−10) (Figure 6.7), similar to previously published results [47, 49]. Thisrather low observed agreement was expected as several tightly controlledprocesses are involved in protein expression. Different genes are regulatedin different ways and to different extent, and regulation can occur both at

77


the transcriptional and the translational stage. For this reason investigatingcorrelations on the individual gene level would be more accurate.

●

●

●●

●

1e+03

1e+04

1e+05

1e+06

1e+07

1e+08

1e−02

1e−01

1e+00

1e+01

1e+02

1e+03


protein copy number per cell

ANXA1

A−549HEK 293HeLaHep−G2RT−4U−2OS

●

●

●

●

●

1e+03

1e+04

1e+05

1e+06

1e+07

1e+08

1e−02

1e−01

1e+00

1e+01

1e+02

1e+03



CD81


●

●

●

●

1e+03

1e+04

1e+05

1e+06

1e+07

1e+08

1e−02

1e−01

1e+00

1e+01

1e+02

1e+03



DECR1


●

●

●

●●

●

1e+03

1e+04

1e+05

1e+06

1e+07

1e+08

1e−02

1e−01

1e+00

1e+01

1e+02

1e+03



MSN


●●

● ●

●●

1e+03

1e+04

1e+05

1e+06

1e+07

1e+08

1e−02

1e−01

1e+00

1e+01

1e+02

1e+03



PTPN1


●

●●

●

●

●

1e+03

1e+04

1e+05

1e+06

1e+07

1e+08

1e−02

1e−01

1e+00

1e+01

1e+02

1e+03



RRBP1


Spearman’s rank coe�cient = 0.90

p = 0.083


p = 0.083


p = 0.083


p = 0.033


p = 0.058


p = 0.017

Figure 6.8: Comparison of protein and mRNA levels forindividual genes.

When comparing protein and mRNA abundance for every individual geneacross the six cell lines an increased correlation was observed compared tothe correlation for the whole gene set. For many genes, similar patterns inmRNA and protein levels could be observed across the cell lines. This wastrue both for genes with variation in the mRNA level across the cell linesand for genes where the levels were constant. For six genes a significantSpearman coefficient of correlation (p<0.1) could be determined. In orderto be able to analyze the whole gene set in this way, data from more cell linesis needed. All six analyzed genes showed correlations above 0.8, the averagecorrelation being 0.9 (Figure 6.8).

It is clear that there are differences in gene regulation between the studied

78


genes. In several cases, mRNA levels changed drastically with very smallchanges in protein abundance. For these genes, factors such as protein half-life or inhibitory RNA could for example play a role in the regulation ofprotein abundance. This data shows that protein and mRNA correlationcan actually be very high, although due to differences in gene regulation,comparisons of large gene sets usually show lower overall correlations. Tofurther validate this theory, more data is needed. Quantification of a largernumber of proteins would be interesting to get a larger picture of mRNAand protein correlation, but expansion of the cell line panel would also benecessary in order to get enough data for each gene for statistical comparisonsto be significant.

Antibody validation using immunoenrichment coupled to massspectrometry (paper IV)

As described in chapter 4, one strategy to analyze proteins in complex sam-ples is by coupling an immunoenrichment step to MS readout. This enablesspecific identification but also, in many cases, analysis of low abundant pro-teins, since the immunoenrichment both lowers the complexity of the sam-ple significantly and also leads to increased protein or peptide concentra-tion. However, since the availability of good, validated antibodies is limitedand different antibodies generally do not behave in the same manner, high-throughput methods are rare [244, 245]. If analyzing protein interactions,it is of primary importance that the specificity of the antibody is known,in order to be able to distinguish true interaction partners from proteinsthat either bind nonspecifically to the antibody or proteins for which theantibody shows specific cross-reactivity. Finding good controls can thereforebe problematic when dealing with target-specific antibodies, as compared totag-specific antibodies.

In paper IV, it was investigated whether antibodies generated within theHPA project could be used as capture agents in an immunoenrichment setup.The HPA antibodies are thoroughly validated, however the methods used aremostly based on detection of denatured proteins. Antibody cross reactiv-ity is determined using PrEST arrays, where antibody binding towards 384PrESTs is analyzed. Western blot (WB) analysis is further used to investi-

79


Bead coupling Immunoenrichment

MS sample prep MS analysis and data evaluation

m/z

intensity

Figure 6.9: Workflow for immunoenrichment using HPAantibodies coupled to MS readout. Antibodies are covalentlyattached to protein A-coupled beads and the beads are subse-quently incubated with a cell lysate. After wash, elution andtryptic digestion, peptides are analyzed in a mass spectrome-ter and proteins interacting with the antibodies are identified.

gate binding to endogenous proteins in cell lines, tissues and plasma. In thisstudy, a total of 30 antibodies were chosen for analysis. Twenty of these werepolyclonal antibodies generated within the HPA project and the remainingten antibodies were monoclonal antibodies generated by Atlas Antibodies.The glioblastoma cell line U251 was used for the analysis and antibodies werechosen that generated a band corresponding to the correct molecular weightin WB. To ensure that the antibodies were positioned correctly for antigenbinding, antibodies were covalently attached to protein A-coated beads. Theimmunoenrichment was performed in a 96-well filter plate format, which en-abled parallel handling of up to 96 samples. By attaching the filter plateto a vacuum device, liquid was easily drawn through the wells. The samesystem was used as in paper I. After incubation of beads and lysate, severalwashing steps were performed in order to minimize the amount of unspe-cific binding to the beads, antibodies and walls of the wells. Proteins wereeluted from the beads, digested with trypsin and analyzed using LC-MS/MS

80


(Figure 6.9). The procedure was performed in triplicate reusing the sameantibody-coupled beads for each replicate. When analyzing the peptide in-tensities of the detected proteins no drop in intensity over the replicates wasobserved, showing that the beads are reusable for at least three times.

Out of the 30 analyzed antibodies, 16 could capture their correspondingtarget in at least two out of three replicates. Seven of these were monoclonaland nine were polyclonal antibodies, resulting in a success rate of 70% and45% for the monoclonal and polyclonal antibodies respectively. There areseveral reasons that could explain why a portion of the antibodies could notcapture their target protein in this setup. The epitope could be located atthe interior of the protein, which makes it accessible to the antibody onlyif the protein is in its denatured state. WB analysis showed that thereis a clear difference between antibody recognition of a denatured proteinsample compared to a sample with proteins in their native state. It is alsopossible that the protein is inaccessible to the antibody for other reasons.A protein involved in a large protein complex might not be able to interactwith the antibody in its native state due to sterical hindrance, however underdenaturing conditions where protein interactions are lost, the protein will bemore accessible to the antibody. Another potential reason for antibodiesnot capturing their target protein in detectable amounts is that the proteinis expressed at a very low level in this cell type or not expressed at all.Proteins not detected by their corresponding antibody generally had lowermRNA levels in the U251 cells based on RNA sequencing. There is also apossibility that the antibody antigen interaction is weak and that the proteintherefore is released from the antibody during the washing steps.

Label free quantification (LFQ) was applied in order to determine antibodyspecificity. Specifically enriched proteins were identified by performing ttests where peptide LFQ intensities from samples were compared to corre-sponding intensities from a control. Proteins with a more than sixteen-foldincrease in intensity between sample and control and a p-value below 0.01were considered significant. The monoclonal antibodies only enriched theirtarget protein, whereas different contaminants and interaction partners wereidentified in the majority of the samples from the polyclonal antibodies. Ex-amples are shown in Figure 6.10. Here, intensity ratio between sample andcontrol is shown on the x axis and −log10(p-value) on the y axis. Proteins

81


●●

●

●

●

● ●

●

●●

●

● ●

●

●●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●●

●●●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

● ●

●

●

●●

●

●

●

●●

●

●

●●●●

●

●

●

●●

●

●

●

● ●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●●

●

●

●

● ●

●●

●

●

●

●

●

●

● ●●

●

●

●

●

●

●●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●●

●

●

●

●● ●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●● ●

●●

●

●●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●●

●

●●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●●

●

●●

●

●

●●

●

●

●

●

●

●

●

●

●

●●

●

●

● ●

●●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●● ●

●

●

●

● ●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●●

●

●

● ●●

●

●

●

●

●

●●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●●

●

●

●

●

●

●

●

● ●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●●

●

●●

mAb1 ( target: RBM3 )

log2(ratio mAb1 /control)

−log

10(p

−valu

e)

−6 −4 −2 0 2 4 6

02

46

8

RBM3

●●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●●

●

●●

●

●●

●

●

●

●

●●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

● ●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●●●

●

●●

●

●●

●

●●●●

●

●

●

●

●●

●

●

●●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●●

●

●●

●

●

●●

●

●

●

●

●

●●

●●

●

●

●

●

●

●

●●

●

●●

●●

●

●●●

●

●●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●●

●●

●●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●●●

●

●

●●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●●

●

●●

●

●●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●●

●●

●

●●

●

●

●

●

●

●

●

●●

●●

●

●

●

●

●

●

●

●

●

●

●●

●●

●

●

●

●

●●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

mAb6 ( target: STX7 )

log2(ratio mAb6 /control)

−log

10(p

−valu

e)

−10 −5 0 5 10

01

23

45

6

STX7

●

●

●●

●●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●●

●

●

●

●●

●●●

●

●

●

●

●●

●

●

●

●●

●

●

●

● ●●

●●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●●

●

●

●

●●

●

●

●

●

●

●

●

●

● ●

●

● ●

●

●

●

●

●

●

●

●

●

●

●● ●

●

●

●

●

●

●

●

●●

●

●

●

● ●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●●

●●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●● ●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●●

●

●

●

●●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●●

●

●

●

●

● ●●

●●●

●

●

● ●

●

●

●

●●

●

●

●

●●

●●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●●●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●●

●

●

●

●

●

●

●

●

●

pAb5 ( target: CHCHD3 )

log2(ratio pAb5 /control)

−log

10(p

−valu

e)

−6 −4 −2 0 2 4 6

01

23

45

67

CHCHD6

CHCHD3●

●

●●

●

●

●

●

●

●●

●

●

●

●

●●

●●

●

●

●

●

●

●

●●

●

●

●●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●●

●

●

●

●

●

●

●●

● ●

● ●

●

●●●

●

●

●

●●●

●

●

●

●

●

●

●

●

●

●

●

●●●●

●

●

●

●

●

●

●●

●

●● ●●

●

●

●●

●●

●●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●●

●

●

●

●

●

●

●

●

●

●●

●

● ●

●

●

●●

●

● ●

●

●

●

●

●

●

●●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●●●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●●

●

●

●

●

●

●

●

●

●

●

●●●

●

●

●

●

●

●

● ●

●

●

●●

●●

●

●

●●

●

●

●

●

●

●●

●

●

●●

●

●

●

●

●

●●●●

●

●●

●

●

●

●

●

●

●

●

●

● ●

●

●●

●

●

●

●

●

●●

●

●

●●

●

●●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●● ●

●

●●

●●

●

●

●

●

●

●●

●

●

●

●●● ●

●●

●

●

●

●

●

●

●●

●●

● ●

●

●

●

●

●

●

pAb15 ( target: NSRP1 )

log2(ratio pAb15 /control)

−log

10(p

−valu

e)

−10 −5 0 5 10

02

46

8

CSRP1

NSRP1

ATPIF1

CSRP2

Figure 6.10: Identification of target proteins after immu-noenrichment coupled to MS. Ratios from logarithmized LFQintensities are shown on the x-axis and the negative logarithmsof the corresponding p-values on the y-axis. Grey lines indi-cate threshold values of log2(ratio)=4 and p=0.01. Signifi-cantly enriched proteins are located in the upper right cor-ner. Color-coding: green = target protein, black = commoncontaminant, blue = probable interaction partner or antibodycross reactivity target.

specifically enriched by the antibody will therefore be present in the upperright section, whereas contaminants with no significant difference in inten-sity between sample and control will appear around the center. Monoclonalantibody 1 (mAb1) targets RBM3, which was clearly enriched in the cor-responding samples. The same is true for mAb6, which targets the proteinSTX7. Polyclonal antibody 5 (pAb5) targets the protein CHCHD3. Thisprotein was identified in the corresponding samples, however ended up just

82


below the significance level. The antibody significantly enriched the pro-tein CHCHD6, which is known to interact with CHCHD3 and might explainwhy it was detected [317,318]. When comparing the sequences of these twoproteins, a stretch of amino acids within the antigen region showed rela-tively high sequence similarity, wherefore antibody cross reactivity could beanother possible explanation. The target of pAb15 is NSRP1, which was sig-nificantly enriched in the corresponding samples. Three other proteins werealso enriched, however these proteins were present in a large portion of thesamples, indicating that they are contaminants interacting nonspecificallywith the antibody, beads or plastics. In total, LFQ data was obtained for14 of the target proteins and twelve of these were significantly enriched inthe corresponding sample. In the remaining two samples, the target proteinswere identified just below the significance threshold.

In summary, roughly 50% of the analyzed antibodies were functional in thedeveloped immunoenrichment setup. This indicates that a very large numberof the over 21,000 antibodies generated within the HPA project, should bepossible to use in immunoenrichment of native proteins coupled to MS read-out. These antibodies could for example be used to study protein abundancein different cells and tissues or to investigate protein interactions.

Protein quantification using immunoenrichment and mass spec-trometry (paper V)

Immunoenrichment can be performed both at the protein and the peptidelevel, as described in chapter 4. For quantification purposes, enrichmentat the peptide level is preferred, due to the straightforward incorporationof AQUA peptides for absolute quantification, as in the SISCAPA tech-nology [76]. However, incomplete digestion can hamper the accuracy ofquantification based on AQUA peptides, as they are spiked into the sam-ple after the proteolytic digestion [122]. In paper V, PrESTs were used asinternal standards, similarly as in paper III. However this strategy, namedimmuno-SILAC, also includes an immunoenrichment step where HPA anti-bodies are used to capture tryptic peptides from a digested cell lysate (Figure6.11).

First it was investigated whether the HPA antibodies were able to bind

83


light cell line heavy PrEST

mix

digestion

protein extraction

m/z

intensity

MS analysis

MS spectrum

immunoenrichment

Figure 6.11: Workflow for protein quantification usingimmuno-SILAC. Heavy PrESTs are spiked into a cell lysateand the sample is digested. Antibodies are used to enrich bothheavy and light versions of tryptic peptides and the eluate isanalyzed in a mass spectrometer. Heavy to light ratios areused for absolute quantification of the endogenous proteins.

to tryptic peptides even though they were raised towards larger proteinfragments. Epitope mapping of 941 HPA antibodies was performed usinghigh-density arrays with 12-mer synthetic peptides covering the 941 antigensequences with one amino acid shift. The number of linear epitopes for eachantibody ranged between zero and ten with on average 2.9 linear epitopes perantibody. This should give an indication of whether the antibody is suitable

84


for peptide immunoenrichment, however several other factors also affect theoutcome. For example, epitopes containing a lysine or arginine residue willnot be functional due to destruction of the epitope during tryptic cleavage.Also, some epitopes might be longer than twelve amino acids, making themundetectable on the peptide array. A subset of 150 antibodies were chosenin order to investigate to what extent the antibodies can be used in peptideimmunoenrichment and the corresponding PrEST was available for 127 ofthese. Antibodies were incubated with protein A coated magnetic beads inpools of around 50 antibodies. As little as 50 ng per antibody was required.A PrEST mix was digested with trypsin before incubation with the beadpools. After wash and elution, the samples were subjected to MS analysis.Due to the lowered complexity of the sample after immunoenrichment, a 15min LC gradient was sufficient for peptide separation. This can be comparedto the setup in paper III, where six fractions were each separated using a 3h gradient. In total, 57 out of the analyzed 127 antibodies successfully cap-tured one or several tryptic peptides corresponding to the target PrEST. Theconclusion from this experiment was that roughly half of the antibodies rec-ognize at least one tryptic peptide, making the large pool of HPA antibodiesa valuable resource for immunoenrichment also on the peptide level.

It was thereafter investigated whether the antibodies could be used to gen-erate quantitative data by immunoenrichment of a mixture of endogenouspeptides and heavy isotope-labeled internal standard peptides using theimmuno-SILAC setup. A new set of 41 PrESTs were produced with heavyisotope-labeled arginine and lysine and were spiked into a HeLa cell lysatebefore tryptic digestion. Corresponding antibodies were thereafter used tocapture the peptides from the cell lysate. After MS analysis, peptides corre-sponding to 22 of the PrESTs were identified. These PrESTs in turn corre-sponded to 20 different proteins, resulting in a success rate of roughly 50%,also in this experiment. The quantification of the 20 proteins was in parallelperformed using the PrEST-SILAC method, described in Zeiler et al [127]and in paper III. The number of peptides used for quantification was higherfor PrEST-SILAC, which was expected since no specific peptides were en-riched (Figure 6.12). In immuno-SILAC, 59% of the proteins were quantifiedwith only one peptide, whereas in PrEST-SILAC the corresponding numberwas 6%.

85


0 20 40 60

BLVRB

HPRR3210149 [73aa]

log 2

(R

atio

H:L

)

−4

−2

02

4

0 20 40 60 80

CAPG

HPRR2370009 [83aa]

log 2

(R

atio

H:L

)

−4

−2

02

4

120 140

0 20 40 60 80 100 120 140

ANXA3

HPRR2050268 [141aa]

log 2

(R

atio

H:L

)

−4

−2

02

4

0 50 100 150

SERPINB6

HPRR1440008 [145aa]

log 2

(R

atio

H:L

)

−4

−2

02

4

0 20 40 60 80 100 120

CLPP

HPRR1440011 [122aa]

log 2

(R

atio

H:L

)

−4

−2

02

4

0 20 40 60 80 100 120

DAP3

HPRR1420037 [135aa]

log 2

(R

atio

H:L

)

−4

−2

02

4

0 20 40 60 80

ERLIN1

HPRR1950257 [91aa]

log 2

(R

atio

H:L

)

−4

−2

02

4

0 20 40 60 80 100

PTPN1

HPRR1952306 [139aa]

log 2

(R

atio

H:L

)

−4

−2

02

4

0 10 20 30 40 50 60 70

DECR1

HPRR2540115 [71aa]

log 2

(R

atio

H:L

)

−4

−2

02

4

Figure 6.12: Quantitative data from the immuno-SILACand PrEST-SILAC setups. Heavy to light peptide ratios areshown for nine example proteins. Quantitative data generatedwith immuno-SILAC (yellow) and PrEST-SILAC (blue) arecompared. The median peptide ratios from the PrEST-SILACquantification are shown by the dotted lines.

It has been shown previously that the digestion efficiency between endoge-nous proteins and PrESTs is similar [127], which makes it possible to includemiscleaved peptides that are otherwise usually removed from the dataset.This trend was also seen in these results. For example, several miscle-aved peptides corresponding to the protein DAP3 were identified using theimmuno-SILAC setup. These all show similar heavy to light ratios where-fore they can actually provide additional accurate quantitative information(Figure 6.12). The chosen targets were all moderate to high abundant in the

86


●

●

●

●●●

● ●

●●●● ●

●

● ●● ●

●

●

1e+04

1e+05

1e+06

1e+07

1e+08

1e

+0

4

1e

+0

5

1e

+0

6

1e

+0

7

1e

+0

8

protein copy number per cell (PrEST−SILAC)

pro

tein

co

py n

um

be

r p

er

ce

ll (im

mu

no

−S

ILA

C)

Pearson correlation

R2=0.91

Figure 6.13: Comparison of protein copy numbers gen-erated with immuno-SILAC and PrEST-SILAC. Data fromimmuno-SILAC is shown on the y-axis and PrEST-SILACdata on the x-axis. The Pearson correlation was determinedto 0.91.

HeLa cell line, wherefore in this case proteins were in most cases identifiedwithout immunoenrichment. However, for low abundant targets enrichmentstrategies such as immuno-SILAC could be necessary. Regarding variationin quantitative data between replicate samples, immuno-SILAC generatedslightly increased coefficients of variation compared to PrEST-SILAC rang-ing between 10-40% for the majority of the proteins. This can be due to thelower number of peptides included in the quantification, where a large num-ber of quantified peptides will generally lead to a more accurate result. Whencomparing the copy numbers per cell determined using the two methods, agood correlation of 0.91 could however be observed (Figure 6.13).

Paper V shows that HPA antibodies can successfully be used for peptide im-munoenrichment, however a screening step is required in order to determinewhether the antibody recognizes tryptic peptides and also to determine thenumber of peptides that can be captured by the antibody. This setup couldbe used for rapid quantification of proteins in cell lines but also potentiallyfor quantification of lower abundant plasma proteins.

87


Concluding remarks

Proteomics is a growing area of research. Not so surprising one could think;it is hard not to be intrigued by the challenge of investigating our most im-portant building blocks. The constant improvement of MS instrumentationis most likely a great contributor to the rapid expansion of the field andwith the current development speed, an increase in sensitivity of around 70-fold has been anticipated within the coming decade [152]. Along with thetechnological development within the field, numerous innovative methodsfor protein analysis have emerged. MS, affinity reagents and the combina-tion of these are the dominant players within proteomics and are used inlaboratories worldwide.

The papers presented in this thesis are all focused toward high-throughputprotein analysis. In papers I and II, screening methods for high-throughputverification of protein products were developed in different formats. Thesescreening methods could decrease both the time and effort spent on pro-tein production and purification. In paper II, the ISET platform was usedin a new setting, for enrichment of His6-tagged proteins, hence increasingthe possible fields of application for this platform. The PrESTs generatedwithin the HPA project were in papers III and V used as internal standardsfor MS-based absolute protein quantification. A high-throughput system forproduction of heavy isotope-labeled protein fragments has been set up, en-abling production of 100 protein fragments in a single batch. Recently, thepossibility to downscale the production from a shake flask format to a deepwell plate format has been investigated, which would enable feasible, paral-lel production of several hundred heavy isotope-labeled PrEST proteins. Inpaper V, a peptide enrichment step was also included where HPA antibodieswere used to capture tryptic peptides resulting in a significant decrease insample complexity. In addition, the enrichment step could possibly enableanalysis of low abundant proteins in for example plasma, although this hasyet to be investigated. For purposes such as the analysis of protein com-plexes, enrichment at the protein level is necessary. The HPA antibodieswere in paper IV used in a high-throughput protein enrichment setup andit was shown that a large portion of the antibodies could capture the corre-sponding native full-length protein from a complex sample. This could beof great value in for example investigating protein interactions but also for

88


determining antibody specificity.

One future goal is to be able to combine the setups presented in papers IVand V, using antibodies for protein and peptide enrichment, with the ISETplatform used in paper II. This would result in a miniaturized setup for pro-tein and peptide analysis that could potentially be of value in the clinic.By combining antibodies targeting a panel of selected biomarkers, indicativefor a specific disease in vials of the ISET plate, information regarding pa-tient disease state could be generated. This could be enabled using minimalamounts of both reagents and patient sample and also in a minimal amountof time. Moreover, the protein immunoenrichment protocol presented in pa-per IV could be modified to enable investigation of protein complexes. Thisshould be done by optimizing the washing steps in order to preserve alsoweaker interactions.

Another aim is to further develop the application of PrESTs as internal stan-dards in MS-based absolute quantification setups. Additional optimizationof the PrEST production and validation process and development of stream-lined protocols for protein quantification could enable a robust system foraccurate quantification of proteins in a high-throughput fashion. In paperIII, 32 proteins were quantified using this setup, however the limit of multi-plexing probably lies far beyond this and hopefully PrEST protein standardscan in the near future be used for parallel quantification of several hundredsor thousands of proteins.

In summary, the work presented in this thesis includes the development andapplication of high-throughput methods for protein analysis and quantifi-cation. These methods can in the future be applied to determine absoluteprotein copy numbers in cells and hopefully also tissue and plasma. By usingantibodies, it is possible that also lower abundant proteins can be detectedand quantified and protein enrichment strategies enables the investigation ofprotein complexes. It is my hope that the presented investigations can be ofvalue for the proteomics research field and that other researchers can benefitfrom the results.

89


90

Chapter 7

Populärvetenskapligsammanfattning

Tillägnad min fantastiska familj

Den här avhandlingen handlar om proteiner och hur man på olika sättkan studera dem. Proteiner finns i alla våra celler och har där livsviktigauppgifter. Proteiner ser till att cellerna håller ihop, de ser till att syre trans-porteras genom blodet och de ser till att vi kan ta kål på bakterier och virussom kommer in i kroppen. Bara för att nämna några exempel. Proteinernaproduceras i en process som kallas translation och det är gener i vårt DNAsom bestämmer hur ett protein ska se ut. Människans DNA blev kartlagt2001 och idag vet vi alltså vilka gener vi har och därigenom vilka proteinersom kan produceras. Detta betyder dock inte på något sätt att vi vet vadalla proteiner har för funktion. Denna frågeställning studeras i det forsk-ningsområde som kallas proteomik.

Proteomiken kan delas upp i två olika delar: masspektrometri (MS)-baseradproteomik och affinitetsproteomik. I MS-baserad proteomik analyserar manproteiner genom att bestämma deras massa och man tittar oftast på pep-tider, det vill säga mindre delar av proteiner som man kan få fram genomatt klippa sönder ett protein med ett enzym. Detta gör man helt enkelt

91

Populärvetenskaplig sammanfattning

för att de blir lättare att detektera då. I affinitetsproteomik använder manaffinitetsmolekyler, oftast antikroppar, som kan binda till det protein man ärintresserad av för att på så vis kunna detektera det. Jag har i min forskninganvänt mig av MS-baserad proteomik men även en kombination av MS-baserad och affinitetsproteomik. Målet med forskningen som presenterasi denna avhandling var att utveckla metoder för att kunna analysera ochkvantifiera många proteiner parallellt. Jag har till exempel jobbat med MS-baserade metoder för att kunna kvantifiera proteiner i olika celler. Ett prob-lem när man vill analysera proteiner med MS är att det finns för mycketandra proteiner i lösningen som stör. Ett sätt att lösa detta problem är attanvända antikroppar som kan fiska ut de proteiner man är intresserad av.Antikroppar har en viktig roll i vårt immunförsvar där de binder till proteinerpå ytan av bakterier och virus. Detta signalerar till resten av immunförsvaretatt bakterien eller viruset är en inkräktare som då kan tas om hand av olikaimmunceller. Antikropparnas roll är alltså (bland annat) att binda starkttill andra proteiner. Man kan idag producera antikroppar som binder till detprotein man vill (målproteinet) genom att utnyttja immunförsvaret. Dettagörs i stor skala i proteinatlas-projektet (HPA) genom att immunisera ka-niner med mänskliga proteinfragment. Jag har använt antikroppar för attfånga målprotein från ett cellysat (proteiner som kommer ut i lösning närman slår sönder ett cellprov). MS användes sedan för att kontrollera vadsom hade bundit till antikropparna. I ett annat projekt användes en lik-nande strategi men här var det peptider (delar från sönderklippta proteiner)istället för proteiner som fiskades ut av antikroppar. Jag har även jobbatmed att utveckla metoder som kan användas inom HPA-projektet för attförenkla flödet för proteinproduktion och proteinrening.

Sammanfattningsvis bygger denna avhandling på utvecklingen av olika metod-er för parallell analys av proteiner. Metoderna har bland annat använts föratt kvantifiera proteiner i olika celler och för att undersöka till vilka proteinerolika antikroppar binder. Det slutliga målet är att öka förståelsen för hurolika proteiner fungerar och interagerar med varandra och att kunna identi-fiera vilka proteiner som är kopplade till olika sjukdomar. Förhoppningsviskommer denna avhandling att kunna bidra till detta.

92

Chapter 8

Acknowledgements

So now you’ve reached the last part of my thesis. I had such a great timeworking on the different projects and I have so many people to be gratefulto for this. Both people involved in the projects but also people who havesupported me in other, less research-related, ways and people who have justmade me smile when I have needed it the most.

Sophia, min underbara huvudhandledare och förebild. Jag är så glad att jagfått spendera de senaste åren med dig. Du har lärt mig så otroligt mycket,inte bara om forskning, utan även om hur viktigt det är att tro på sig självoch stå på sig. Tack för att du alltid ser lösningar på alla problem och för attdu alltid peppar och uppmuntrar. Tack för att du tar dig tid och alltid finnsdär även fast du egentligen har tusen andra viktiga saker att göra. Tackockså för att du gett mig så mycket frihet och låtit mig utvecklas.

Tack till min bihandledare Thomas för spännande samarbeten kring ISEToch roliga möten med gott godis. Tack vare dig har jag fått uppleva någotannat än bara KTH, vilket har varit otroligt lärorikt och framförallt riktigtkul.

Jenny OT, även fast det inte står på papper, känner jag att du verkligenhar varit en handledare för mig. Du har alltid tid för frågor och funderingaroch kommer alltid med kloka råd. Det har verkligen varit lärorikt att jobbamed dig. Du gör inget halvhjärtat, vilket jag verkligen beundrar. Tack föralla diskussioner och pekskrivande.

93

Acknowledgements

Tack Mathias för att du är en sådan inspirationskälla och startar projektsom HPA. Tack också för att du vågade satsa på MS-kvantifiering med tungaPrESTar på KTH fast det var något helt nytt för oss.

Till alla övriga PIs på plan3, både nya och gamla, vill jag också rikta ettjättestort tack. Tack för alla kloka ord och för att ni värnar om arbetsmiljönoch inser att det blir bättre forskning när alla hjälps åt (och ses för att drickaöl ihop).

Tack till Knut och Alice Wallenbergs stiftelse samt PIEp/Vinnovaför finansiering.

Tack till alla som jag samarbetat med i projekten under åren. Hanna Toch Louise, ni välkomnade mig in i EnBase-projektet när jag var helt nydoktorand. Tack för alla härliga stunder med filterplattor och reningar!Tack också Louise för att du introducerade mig till MS-världen. Thank youKaisa and the rest of the EnBase

R© team for nice collaborations.

Belinda och Simon, efter många finjusteringar lyckades vi äntligen få rob-otarna att göra som vi ville och vi lyckades automatisera ISET-protokollet.Tack för många häftiga och lite nördiga ISET-filmer. Ett speciellt tack till digBelinda för allt knasigt vi haft för oss nere hos er på LTH. Hydrofobpennoroch häftmassa skulle jag tidigare inte associerat med forskning, men tänk såfel jag hade. Tack också för roliga samarbeten, kurser och konferenser.

Tack Henrik J och Janne för MS-stöd i många av projekten. Tack för allhjälp med MS-prover och svar på frågor och för att MS-kön kunnat modifieraslite vid behov. Tack Erik M för alla timmar du suttit vid min dator ochprogrammerat. Utan din hjälp hade det inte blivit så mycket MS-data. TackFrida för cellodling och Anna Bä och Hanna T för roligt labbande ihopvid produktion av tunga PrESTar. Fredrik och Björn, vad kul det var attäntligen få några att prata MS med. Tack för alla diskussioner om tungaPrESTar och MS-instrument. Tack till projektledarna Sophia, Thomas,Jenny OT och Mathias som styrt projekten i mål. Tack även till Sophia,Thomas, Jenny OT, Janne och Margareta för värdefulla kommentarertill min avhandling.

Thank you Marlis and Matthias for nice collaborations and thank youMarlis for taking such good care of me during my visit in Munich. I had a

94

Acknowledgements

wonderful time and learnt so much.

För nästan 6,5 år sedan började jag som sommarjobbare i MolBio-gruppeni HPA. Tack Anja för att du anställde mig och tack Holger, Marica,Bahram, Anna S, Jenny F, Mehri, Malin och Lili för en rolig tid medmycket racklingar och sekvensning.

Tack Johan N för en fantastisk tid under mitt exjobb. Det finns ingen merpedagogisk och positiv handledare än du. Tack även Elin och Anna JP förall hjälp på Affibody under exjobbet.

Jag är otroligt tacksam att ha fått vara en del av en fantastisk forskargruppmed många kloka och fina människor. Tack Sophia för alla resor till Målaoch tack alla nuvarande och gamla medlemmar i gruppen, Hanna T, JennyOT, Micke, Sarah L, Sara K, Emma F, Johan N, Mattias, Anna K,Cajsa, Tove A, Johanna S och alla exjobbare för alla roliga stunder påfrukostar och middagar.

Alla härliga tjejer i proteinfabriken som haft hand om MS-instrumentenpå ett eller annat sätt, Kattis, Anne-Sophie, Anneli H, Anna Be ochJenny B. Tack för alla roliga, mindre roliga och frustrerande timmar vidinstrumenten. Det finns ingen större glädje än när man meckat ihop ESInhelt på egen hand.

De fina exjobbare jag fått äran att handleda under åren, Sara K, Erik Foch Jessica. Tack för att ni varit så motiverade och roliga att jobba med.Jag tror att jag lärde mig minst lika mycket som ni.

Tack till alla nuvarande och gamla rumskamrater, Sara K, Micke, Erik F,Greg, Linnea, Josefine, Wasil, Prem och Staffan för mysig gemenskapmed skratt och diskussioner. Tack Micke och Erik F för att ni alltid finnsdär för att hjälpa till med alla tänkbara datorproblem.

Tack alla fina plan3-tjejer som styrt upp stickjuntor, tjejmiddagar och an-nat skoj. Tack gamla och nuvarande medlemmar i beer foundation för godöl och roliga tillställningar i lunchrummet. Tack till alla Bio05-or på avdel-ningen som hängt ihop i snart 10 (!) år. Hanna L, Lisa, Filippa, Magnus,Johan S och Anna H, det känns som om det var igår vi stod tillsammansframför gula K:t och det kommer kännas konstigt när vi skiljs åt i den riktigavärlden. Tack Filippa och Andreas J för svettiga timmar på KTH-hallen.

95

Acknowledgements

Tack också Andreas J för att du dragit iväg mig till gymmet under höstennär jag trott att jag haft viktigare saker för mig. Tack även Emma och LanLan för beställningar och Inger och Kristina för administrativ hjälp. TackRoxana & Co för fantastiskt jobb i diskrummet.

Sara K, vad hade jag gjort utan dig? Jag kan inte tänka mig hur detkommer vara att inte sitta bredvid dig varje dag. Tack för att du alltid finnsdär oavsett om man vill skratta eller gråta. Du är nog bäst i världen på attmuntra upp när inget går som det ska. Jag hoppas du förstår att du är heltfantastisk.

Tack alla härliga sångfåglar som jag fått sjunga med i luciatåget genomåren. Hanna T, Sara K, Anneli H, Magnus, Erik F, Andreas H,Tobbe, Micke, Tarek, Erica, Josefine, Anna-Luisa och LanLan. Luciaär verkligen bästa dagen på året!

Tack alla på plan3 för att ni tillsammans gör det till världens bästa arbet-splats. Tack också till alla HPA-grupper som tillsammans gjort HPA tilldet fantastiska projekt det är. Särskilt tack till proteinfabriken för BCA-analyser m.m, tack till immunotech för hjälp med western blot och tack tillmolbio för kontrollsekvensning. Tack också för trevliga resor till Mallorcaoch Antalya.

Alla vänner utanför KTH, tack för allt kul vi haft genom åren. Tack gym-nasiegänget Linda, Sophia, Jennie, Aya, Susanna, Linn, Rickard ochMyggan för roliga middagar och söta barn. Tack Isabelle för att du ärmin närmsta vän och alltid stöttar! Tack Aida och Kalle för att ni är fan-tastiska och alltid finns där och tack till mina musketörer Johnny, Marcusoch Linda för alla fantastiska och galna stunder. Ni är bäst!

Tack även till Frinsö-klanen som får mig att glömma forskningen ibland föratt fokusera på de viktiga sakerna i livet, som fönsterfoder och elverk.

Min fina familj, Mamma, Pappa och lillebror Henning. Ni är de bästaoch jag älskar er över allt annat. Tack för att ni låtit mig gå min egen vägoch att ni stöttat mig med mina "flygande proteiner".

Slutligen, min älskade Tord, tack för att du gör mig lycklig och får mig attskratta varje dag. Tack för att du alltid är stolt över mig, speciellt när jagbehöver det som mest. Du och jag för alltid.

96

Chapter 9

Bibliography

[1] F. H. Crick. On protein synthesis. Symp Soc Exp Biol, 12:138–63, 1958.

[2] N. Gregersen, P. Bross, S. Vang, and J. H. Christensen. Protein misfolding andhuman disease. Annu Rev Genomics Hum Genet, 7:103–24, 2006.

[3] S. Khan and M. Vihinen. Spectrum of disease-causing mutations in protein sec-ondary structures. BMC Struct Biol, 7:56, 2007.

[4] F. Crick. Central dogma of molecular biology. Nature, 227(5258):561–3, 1970.

[5] B. R. Glick and J. J. Pasternak. Molecular biotechnology : principles and applica-tions of recombinant DNA. ASM Press, Washington, D.C., 3rd edition, 2003.

[6] C. Branden and J. Tooze. Introduction to protein structure. Garland : TaylorFrancis, New York, 2nd edition, 1999.

[7] M. Mann, R. C. Hendrickson, and A. Pandey. Analysis of proteins and proteomesby mass spectrometry. Annu Rev Biochem, 70:437–73, 2001.

[8] M. Uhlen, E. Bjorling, C. Agaton, C. A. Szigyarto, B. Amini, E. Andersen, A. C.Andersson, P. Angelidou, A. Asplund, C. Asplund, L. Berglund, K. Bergstrom,H. Brumer, D. Cerjan, M. Ekstrom, A. Elobeid, C. Eriksson, L. Fagerberg, R. Falk,J. Fall, M. Forsberg, M. G. Bjorklund, K. Gumbel, A. Halimi, I. Hallin, C. Ham-sten, M. Hansson, M. Hedhammar, G. Hercules, C. Kampf, K. Larsson, M. Lind-skog, W. Lodewyckx, J. Lund, J. Lundeberg, K. Magnusson, E. Malm, P. Nilsson,J. Odling, P. Oksvold, I. Olsson, E. Oster, J. Ottosson, L. Paavilainen, A. Pers-son, R. Rimini, J. Rockberg, M. Runeson, A. Sivertsson, A. Skollermo, J. Steen,M. Stenvall, F. Sterky, S. Stromberg, M. Sundberg, H. Tegel, S. Tourle, E. Wahlund,A. Walden, J. Wan, H. Wernerus, J. Westberg, K. Wester, U. Wrethagen, L. L. Xu,S. Hober, and F. Ponten. A human protein atlas for normal and cancer tissues basedon antibody proteomics. Mol Cell Proteomics, 4(12):1920–32, 2005.

[9] E. Lundberg and M. Uhlen. Creation of an antibody-based subcellular protein atlas.Proteomics, 10(22):3984–96, 2010.

97

Bibliography

[10] T. Geiger, A. Wehner, C. Schaab, J. Cox, and M. Mann. Comparative proteomicanalysis of eleven common cell lines reveals ubiquitous but varying expression ofmost proteins. Mol Cell Proteomics, 11(3):M111 014050, 2012.

[11] U. B. Nielsen and B. H. Geierstanger. Multiplexed sandwich assays in microarrayformat. J Immunol Methods, 290(1-2):107–20, 2004.

[12] M. Bantscheff, M. Schirle, G. Sweetman, J. Rick, and B. Kuster. Quantitative massspectrometry in proteomics: a critical review. Anal Bioanal Chem, 389(4):1017–31,2007.

[13] M. Bantscheff, S. Lemeer, M. M. Savitski, and B. Kuster. Quantitative mass spec-trometry in proteomics: critical review update from 2007 to the present. AnalBioanal Chem, 404(4):939–65, 2012.

[14] O. T. Avery, C. M. Macleod, and M. McCarty. Studies on the chemical natureof the substance inducing transformation of pneumococcal types : Induction oftransformation by a desoxyribonucleic acid fraction isolated from pneumococcustype iii. J Exp Med, 79(2):137–58, 1944.

[15] J. D. Watson and F. H. Crick. Molecular structure of nucleic acids; a structure fordeoxyribose nucleic acid. Nature, 171(4356):737–8, 1953.

[16] F. H. Crick, L. Barnett, S. Brenner, and R. J. Watts-Tobin. General nature of thegenetic code for proteins. Nature, 192:1227–32, 1961.

[17] T. B. Osborne. The vegetable proteins. Monographs on biochemistry. 1909.

[18] D. Perrett. From ’protein’ to the beginnings of clinical proteomics. Proteomics ClinAppl, 1(8):720–38, 2007.

[19] H. B. Vickery. The origin of the word protein. Yale J Biol Med, 22(5):387–93, 1950.

[20] J. A. Reynolds and C. Tanford. Nature’s robots : a history of proteins. OxfordUniversity Press, Oxford, 2001.

[21] F. Sanger and H. Tuppy. The amino-acid sequence in the phenylalanyl chain ofinsulin. i. the identification of lower peptides from partial hydrolysates. Biochem J,49(4):463–81, 1951.

[22] F. Sanger and H. Tuppy. The amino-acid sequence in the phenylalanyl chain ofinsulin. 2. the investigation of peptides from enzymic hydrolysates. Biochem J,49(4):481–90, 1951.

[23] L. Pauling, R. B. Corey, and H. R. Branson. The structure of proteins; two hydrogen-bonded helical configurations of the polypeptide chain. Proc Natl Acad Sci U S A,37(4):205–11, 1951.

[24] L. Pauling and R. B. Corey. Atomic coordinates and structure factors for two helicalconfigurations of polypeptide chains. Proc Natl Acad Sci U S A, 37(5):235–40, 1951.

[25] C. Tanford. How protein chemists learned about the hydrophobic factor. ProteinSci, 6(6):1358–66, 1997.

[26] W. Kauzmann. Some factors in the interpretation of protein denaturation. AdvProtein Chem, 14:1–63, 1959.

98

Bibliography

[27] R. K. Saiki, S. Scharf, F. Faloona, K. B. Mullis, G. T. Horn, H. A. Erlich, andN. Arnheim. Enzymatic amplification of beta-globin genomic sequences and re-striction site analysis for diagnosis of sickle cell anemia. Science, 230(4732):1350–4,1985.

[28] K. Mullis, F. Faloona, S. Scharf, R. Saiki, G. Horn, and H. Erlich. Specific enzymaticamplification of dna in vitro: the polymerase chain reaction. Cold Spring Harb SympQuant Biol, 51 Pt 1:263–73, 1986.

[29] E. C. Hayden. Technology: The $1,000 genome. Nature, 507(7492):294–5, 2014.

[30] E. S. Lander, L. M. Linton, B. Birren, C. Nusbaum, M. C. Zody, J. Baldwin,K. Devon, K. Dewar, M. Doyle, W. FitzHugh, R. Funke, D. Gage, K. Harris,A. Heaford, J. Howland, L. Kann, J. Lehoczky, R. LeVine, P. McEwan, K. McK-ernan, J. Meldrim, J. P. Mesirov, C. Miranda, W. Morris, J. Naylor, C. Raymond,M. Rosetti, R. Santos, A. Sheridan, C. Sougnez, N. Stange-Thomann, N. Stojanovic,A. Subramanian, D. Wyman, J. Rogers, J. Sulston, R. Ainscough, S. Beck, D. Bent-ley, J. Burton, C. Clee, N. Carter, A. Coulson, R. Deadman, P. Deloukas, A. Dun-ham, I. Dunham, R. Durbin, L. French, D. Grafham, S. Gregory, T. Hubbard,S. Humphray, A. Hunt, M. Jones, C. Lloyd, A. McMurray, L. Matthews, S. Mercer,S. Milne, J. C. Mullikin, A. Mungall, R. Plumb, M. Ross, R. Shownkeen, S. Sims,R. H. Waterston, R. K. Wilson, L. W. Hillier, J. D. McPherson, M. A. Marra, E. R.Mardis, L. A. Fulton, A. T. Chinwalla, K. H. Pepin, W. R. Gish, S. L. Chissoe, M. C.Wendl, K. D. Delehaunty, T. L. Miner, A. Delehaunty, J. B. Kramer, L. L. Cook,R. S. Fulton, D. L. Johnson, P. J. Minx, S. W. Clifton, T. Hawkins, E. Branscomb,P. Predki, P. Richardson, S. Wenning, T. Slezak, N. Doggett, J. F. Cheng, A. Olsen,S. Lucas, C. Elkin, E. Uberbacher, M. Frazier, et al. Initial sequencing and analysisof the human genome. Nature, 409(6822):860–921, 2001.

[31] J. C. Venter, M. D. Adams, E. W. Myers, P. W. Li, R. J. Mural, G. G. Sutton,H. O. Smith, M. Yandell, C. A. Evans, R. A. Holt, J. D. Gocayne, P. Amanatides,R. M. Ballew, D. H. Huson, J. R. Wortman, Q. Zhang, C. D. Kodira, X. H. Zheng,L. Chen, M. Skupski, G. Subramanian, P. D. Thomas, J. Zhang, G. L. Gabor Mik-los, C. Nelson, S. Broder, A. G. Clark, J. Nadeau, V. A. McKusick, N. Zinder,A. J. Levine, R. J. Roberts, M. Simon, C. Slayman, M. Hunkapiller, R. Bolanos,A. Delcher, I. Dew, D. Fasulo, M. Flanigan, L. Florea, A. Halpern, S. Hannen-halli, S. Kravitz, S. Levy, C. Mobarry, K. Reinert, K. Remington, J. Abu-Threideh,E. Beasley, K. Biddick, V. Bonazzi, R. Brandon, M. Cargill, I. Chandramouliswaran,R. Charlab, K. Chaturvedi, Z. Deng, V. Di Francesco, P. Dunn, K. Eilbeck, C. Evan-gelista, A. E. Gabrielian, W. Gan, W. Ge, F. Gong, Z. Gu, P. Guan, T. J. Heiman,M. E. Higgins, R. R. Ji, Z. Ke, K. A. Ketchum, Z. Lai, Y. Lei, Z. Li, J. Li, Y. Liang,X. Lin, F. Lu, G. V. Merkulov, N. Milshina, H. M. Moore, A. K. Naik, V. A.Narayan, B. Neelam, D. Nusskern, D. B. Rusch, S. Salzberg, W. Shao, B. Shue,J. Sun, Z. Wang, A. Wang, X. Wang, J. Wang, M. Wei, R. Wides, C. Xiao, C. Yan,et al. The sequence of the human genome. Science, 291(5507):1304–51, 2001.

[32] M. R. Wilkins, C. Pasquali, R. D. Appel, K. Ou, O. Golaz, J. C. Sanchez, J. X.Yan, A. A. Gooley, G. Hughes, I. Humphery-Smith, K. L. Williams, and D. F.

99

Bibliography

Hochstrasser. From proteins to proteomes: large scale protein identification by two-dimensional electrophoresis and amino acid analysis. Biotechnology (N Y), 14(1):61–5, 1996.

[33] R. Aebersold and M. Mann. Mass spectrometry-based proteomics. Nature,422(6928):198–207, 2003.

[34] B. Domon and R. Aebersold. Mass spectrometry and protein analysis. Science,312(5771):212–7, 2006.

[35] P. H. O’Farrell. High resolution two-dimensional electrophoresis of proteins. J BiolChem, 250(10):4007–21, 1975.

[36] O. Stoevesandt and M. J. Taussig. Affinity reagent resources for human proteomedetection: initiatives and perspectives. Proteomics, 7(16):2738–50, 2007.

[37] M. Clamp, B. Fry, M. Kamal, X. Xie, J. Cuff, M. F. Lin, M. Kellis, K. Lindblad-Toh,and E. S. Lander. Distinguishing protein-coding and noncoding genes in the humangenome. Proc Natl Acad Sci U S A, 104(49):19428–33, 2007.

[38] M. Mann and O. N. Jensen. Proteomic analysis of post-translational modifications.Nat Biotechnol, 21(3):255–61, 2003.

[39] O. N. Jensen. Modification-specific proteomics: characterization of post-translational modifications by mass spectrometry. Curr Opin Chem Biol, 8(1):33–41,2004.

[40] M. Wilhelm, J. Schlegl, H. Hahne, A. Moghaddas Gholami, M. Lieberenz, M. M.Savitski, E. Ziegler, L. Butzmann, S. Gessulat, H. Marx, T. Mathieson, S. Lemeer,K. Schnatbaum, U. Reimer, H. Wenschuh, M. Mollenhauer, J. Slotta-Huspenina,J. H. Boese, M. Bantscheff, A. Gerstmair, F. Faerber, and B. Kuster. Mass-spectrometry-based draft of the human proteome. Nature, 509(7502):582–7, 2014.

[41] J. Cox and M. Mann. Quantitative, high-resolution proteomics for data-drivensystems biology. Annu Rev Biochem, 80:273–99, 2011.

[42] N. L. Anderson and N. G. Anderson. The human plasma proteome: history, char-acter, and diagnostic prospects. Mol Cell Proteomics, 1(11):845–67, 2002.

[43] J. L. Luque-Garcia and T. A. Neubert. Sample preparation for serum/plasma pro-filing and biomarker identification by mass spectrometry. J Chromatogr A, 1153(1-2):259–76, 2007.

[44] P. A. McGettigan. Transcriptomics in the rna-seq era. Curr Opin Chem Biol,17(1):4–11, 2013.

[45] A. Mortazavi, B. A. Williams, K. McCue, L. Schaeffer, and B. Wold. Mapping andquantifying mammalian transcriptomes by rna-seq. Nat Methods, 5(7):621–8, 2008.

[46] S. K. Archer, D. Inchaustegui, R. Queiroz, and C. Clayton. The cell cycle regulatedtranscriptome of trypanosoma brucei. PLoS One, 6(3):e18425, 2011.

[47] R. de Sousa Abreu, L. O. Penalva, E. M. Marcotte, and C. Vogel. Global signaturesof protein and mrna expression levels. Mol Biosyst, 5(12):1512–26, 2009.

100

Bibliography

[48] E. Lundberg, L. Fagerberg, D. Klevebring, I. Matic, T. Geiger, J. Cox, C. Algenas,J. Lundeberg, M. Mann, and M. Uhlen. Defining the transcriptome and proteomein three functionally different human cell lines. Mol Syst Biol, 6:450, 2010.

[49] E. Eisenberg and E. Y. Levanon. Human housekeeping genes, revisited. TrendsGenet, 29(10):569–74, 2013.

[50] J. K. Aronson. Biomarkers and surrogate endpoints. Br J Clin Pharmacol,59(5):491–4, 2005.

[51] L. Trinkle-Mulcahy. Resolving protein interactions and complexes by affinity pu-rification followed by label-based quantitative mass spectrometry. Proteomics,12(10):1623–38, 2012.

[52] A. C. Gingras, M. Gstaiger, B. Raught, and R. Aebersold. Analysis of proteincomplexes using mass spectrometry. Nat Rev Mol Cell Biol, 8(8):645–54, 2007.

[53] E. de Hoffman and V. Stroobant. Mass spectrometry : principles and applications.John Wiley, Chichester, 3. edition, 2007.

[54] M. Karas and F. Hillenkamp. Laser desorption ionization of proteins with molecularmasses exceeding 10,000 daltons. Anal Chem, 60(20):2299–301, 1988.

[55] J. B. Fenn, M. Mann, C. K. Meng, S. F. Wong, and C. M. Whitehouse. Electrosprayionization for mass spectrometry of large biomolecules. Science, 246(4926):64–71,1989.

[56] X. Han, A. Aslanian, and 3rd Yates, J. R. Mass spectrometry for proteomics. CurrOpin Chem Biol, 12(5):483–90, 2008.

[57] H. Steen and M. Mann. The abc’s (and xyz’s) of peptide sequencing. Nat Rev MolCell Biol, 5(9):699–711, 2004.

[58] N. Nagaraj, J. R. Wisniewski, T. Geiger, J. Cox, M. Kircher, J. Kelso, S. Paabo,and M. Mann. Deep proteome and transcriptome mapping of a human cancer cellline. Mol Syst Biol, 7:548, 2011.

[59] A. H. Payne and G. L. Glish. Tandem mass spectrometry in quadrupole ion trapand ion cyclotron resonance mass spectrometers. Methods Enzymol, 402:109–48,2005.

[60] R. Craig and R. C. Beavis. Tandem: matching proteins with tandem mass spectra.Bioinformatics, 20(9):1466–7, 2004.

[61] D. L. Tabb, C. G. Fernando, and M. C. Chambers. Myrimatch: highly accurate tan-dem mass spectral peptide identification by multivariate hypergeometric analysis. JProteome Res, 6(2):654–61, 2007.

[62] J. Cox, N. Neuhauser, A. Michalski, R. A. Scheltema, J. V. Olsen, and M. Mann.Andromeda: a peptide search engine integrated into the maxquant environment. JProteome Res, 10(4):1794–805, 2011.

[63] D. N. Perkins, D. J. Pappin, D. M. Creasy, and J. S. Cottrell. Probability-based pro-tein identification by searching sequence databases using mass spectrometry data.Electrophoresis, 20(18):3551–67, 1999.

101

Bibliography

[64] J. K. Eng, A. L. McCormack, and J. R. Yates. An approach to correlate tandemmass spectral data of peptides with amino acid sequences in a protein database. JAm Soc Mass Spectrom, 5(11):976–89, 1994.

[65] J. D. Tipton, J. C. Tran, A. D. Catherman, D. R. Ahlf, K. R. Durbin, and N. L.Kelleher. Analysis of intact protein isoforms by mass spectrometry. J Biol Chem,286(29):25451–8, 2011.

[66] F. Lanucara and C. E. Eyers. Top-down mass spectrometry for the analysis ofcombinatorial post-translational modifications. Mass Spectrom Rev, 32(1):27–42,2013.

[67] X. Han, M. Jin, K. Breuker, and F. W. McLafferty. Extending top-down massspectrometry to proteins with masses greater than 200 kilodaltons. Science,314(5796):109–12, 2006.

[68] B. A. Garcia. What does the future hold for top down mass spectrometry? J AmSoc Mass Spectrom, 21(2):193–202, 2010.

[69] M. Hedhammar, T. Graslund, and S. Hober. Protein engineering strategies forselective protein purification. Chemical Engineering Technology, 28(11):1315–1325,2005.

[70] M. Neiman, C. Fredolini, H. Johansson, J. Lehtio, P. A. Nygren, M. Uhlen, P. Nils-son, and J. M. Schwenk. Selectivity analysis of single binder assays used in plasmaprotein profiling. Proteomics, 13(23-24):3406–10, 2013.

[71] J. R. Wisniewski, A. Zougman, N. Nagaraj, and M. Mann. Universal sample prepa-ration method for proteome analysis. Nat Methods, 6(5):359–62, 2009.

[72] A. Shevchenko, H. Tomas, J. Havlis, J. V. Olsen, and M. Mann. In-gel digestionfor mass spectrometric characterization of proteins and proteomes. Nat Protoc,1(6):2856–60, 2006.

[73] R. M. Branca, L. M. Orre, H. J. Johansson, V. Granholm, M. Huss, A. Perez-Bercoff,J. Forshed, L. Kall, and J. Lehtio. Hirief lc-ms enables deep proteome coverage andunbiased proteogenomics. Nat Methods, 11(1):59–62, 2014.

[74] J. R. Wisniewski, A. Zougman, and M. Mann. Combination of fasp and stagetip-based fractionation allows in-depth analysis of the hippocampal membrane pro-teome. J Proteome Res, 8(12):5674–8, 2009.

[75] M. P. Washburn, D. Wolters, and 3rd Yates, J. R. Large-scale analysis of the yeastproteome by multidimensional protein identification technology. Nat Biotechnol,19(3):242–7, 2001.

[76] N. L. Anderson, N. G. Anderson, L. R. Haines, D. B. Hardie, R. W. Olafson, andT. W. Pearson. Mass spectrometric quantitation of peptides and proteins using sta-ble isotope standards and capture by anti-peptide antibodies (siscapa). J ProteomeRes, 3(2):235–44, 2004.

[77] Y. Zhang, A. Wolf-Yadlin, P. L. Ross, D. J. Pappin, J. Rush, D. A. Lauffenburger,and F. M. White. Time-resolved mass spectrometry of tyrosine phosphorylation sitesin the epidermal growth factor receptor signaling network reveals dynamic modules.Mol Cell Proteomics, 4(9):1240–50, 2005.

102

Bibliography

[78] M. R. Thingholm, T. E.and Larsen, C. R. Ingrell, M. Kassem, and O. N. Jensen.Tio(2)-based phosphoproteomic analysis of the plasma membrane and the effects ofphosphatase inhibitor treatment. J Proteome Res, 7(8):3304–13, 2008.

[79] J. Rappsilber, M. Mann, and Y. Ishihama. Protocol for micro-purification, enrich-ment, pre-fractionation and storage of peptides for proteomics using stagetips. NatProtoc, 2(8):1896–906, 2007.

[80] N. Jehmlich, C. Golatowski, A. Murr, G. Salazar, V. M. Dhople, E. Hammer, andU. Volker. Comparative evaluation of peptide desalting methods for salivary pro-teome analysis. Clin Chim Acta, 434:16–20, 2014.

[81] M. A. Kuzyk, D. Smith, J. Yang, T. J. Cross, A. M. Jackson, D. B. Hardie, N. L.Anderson, and C. H. Borchers. Multiple reaction monitoring-based, multiplexed, ab-solute quantitation of 45 proteins in human plasma. Mol Cell Proteomics, 8(8):1860–77, 2009.

[82] D. Domanski, A. J. Percy, J. Yang, A. G. Chambers, J. S. Hill, G. V. Freue, andC. H. Borchers. Mrm-based multiplexed quantitation of 67 putative cardiovasculardisease biomarkers in human plasma. Proteomics, 12(8):1222–43, 2012.

[83] L. Anderson and C. L. Hunter. Quantitative mass spectrometric multiple reactionmonitoring assays for major plasma proteins. Mol Cell Proteomics, 5(4):573–88,2006.

[84] P. Picotti and R. Aebersold. Selected reaction monitoring-based proteomics: work-flows, potential, pitfalls and future directions. Nat Methods, 9(6):555–66, 2012.

[85] T. Shi, T. L. Fillmore, X. Sun, R. Zhao, A. A. Schepmoes, M. Hossain, F. Xie,S. Wu, J. S. Kim, N. Jones, R. J. Moore, L. Pasa-Tolic, J. Kagan, K. D. Rodland,T. Liu, K. Tang, 2nd Camp, D. G., R. D. Smith, and W. J. Qian. Antibody-free, tar-geted mass-spectrometric approach for quantification of proteins at low picogram permilliliter levels in human plasma/serum. Proc Natl Acad Sci U S A, 109(38):15395–400, 2012.

[86] P. Picotti, O. Rinner, R. Stallmach, F. Dautel, T. Farrah, B. Domon, H. Wenschuh,and R. Aebersold. High-throughput generation of selected reaction-monitoring as-says for proteins and proteomes. Nat Methods, 7(1):43–6, 2010.

[87] J. D. Venable, M. Q. Dong, J. Wohlschlegel, A. Dillin, and J. R. Yates. Automatedapproach for quantitative analysis of complex peptide mixtures from tandem massspectra. Nat Methods, 1(1):39–45, 2004.

[88] A. Panchaud, S. Jung, S. A. Shaffer, J. D. Aitchison, and D. R. Goodlett. Faster,quantitative, and accurate precursor acquisition independent from ion count. AnalChem, 83(6):2250–7, 2011.

[89] L. C. Gillet, P. Navarro, S. Tate, H. Rost, N. Selevsek, L. Reiter, R. Bonner, andR. Aebersold. Targeted data extraction of the ms/ms spectra generated by data-independent acquisition: a new concept for consistent and accurate proteome anal-ysis. Mol Cell Proteomics, 11(6):O111 016717, 2012.

103

Bibliography

[90] H. Molina, Y. Yang, T. Ruch, J. W. Kim, P. Mortensen, T. Otto, A. Nalli, Q. Q.Tang, M. D. Lane, R. Chaerkady, and A. Pandey. Temporal profiling of theadipocyte proteome during differentiation using a five-plex silac based strategy. JProteome Res, 8(1):48–58, 2009.

[91] S. E. Ong, B. Blagoev, I. Kratchmarova, D. B. Kristensen, H. Steen, A. Pandey, andM. Mann. Stable isotope labeling by amino acids in cell culture, silac, as a simpleand accurate approach to expression proteomics. Mol Cell Proteomics, 1(5):376–86,2002.

[92] N. Ibarrola, H. Molina, A. Iwahori, and A. Pandey. A novel proteomic approach forspecific identification of tyrosine kinase substrates using [13c]tyrosine. J Biol Chem,279(16):15805–13, 2004.

[93] G. T. Cantin, J. D. Venable, D. Cociorva, and 3rd Yates, J. R. Quantitative phospho-proteomic analysis of the tumor necrosis factor pathway. J Proteome Res, 5(1):127–34, 2006.

[94] S. E. Ong, I. Kratchmarova, and M. Mann. Properties of 13c-substituted argininein stable isotope labeling by amino acids in cell culture (silac). J Proteome Res,2(2):173–81, 2003.

[95] B. Blagoev, S. E. Ong, I. Kratchmarova, and M. Mann. Temporal analysis ofphosphotyrosine-dependent signaling networks by quantitative proteomics. NatBiotechnol, 22(9):1139–45, 2004.

[96] P. A. Everley, J. Krijgsveld, B. R. Zetter, and S. P. Gygi. Quantitative cancerproteomics: stable isotope labeling with amino acids in cell culture (silac) as a toolfor prostate cancer research. Mol Cell Proteomics, 3(7):729–35, 2004.

[97] T. Geiger, S. F. Madden, W. M. Gallagher, J. Cox, and M. Mann. Proteomic portraitof human breast cancer progression identifies novel prognostic markers. Cancer Res,72(9):2428–39, 2012.

[98] Q. Wu, W. Xu, L. Cao, X. Li, T. He, Z. Wu, and W. Li. Saha treatment reveals thelink between histone lysine acetylation and proteome in nonsmall cell lung cancera549 cells. J Proteome Res, 12(9):4064–73, 2013.

[99] R. Zhang, C. S. Sioma, S. Wang, and F. E. Regnier. Fractionation of isotopicallylabeled peptides in quantitative proteomics. Anal Chem, 73(21):5142–9, 2001.

[100] T. Geiger, J. Cox, P. Ostasiewicz, J. R. Wisniewski, and M. Mann. Super-silac mixfor quantitative proteomics of human tumor tissue. Nat Methods, 7(5):383–5, 2010.

[101] T. Geiger, J. R. Wisniewski, J. Cox, S. Zanivan, M. Kruger, Y. Ishihama, andM. Mann. Use of stable isotope labeling by amino acids in cell culture as a spike-instandard in quantitative proteomics. Nat Protoc, 6(2):147–57, 2011.

[102] R. R. Lund, M. G. Terp, A. V. Laenkholm, O. N. Jensen, R. Leth-Larsen, and H. J.Ditzel. Quantitative proteomics of primary tumors with varying metastatic capabil-ities using stable isotope-labeled proteins of multiple histogenic origins. Proteomics,12(13):2139–48, 2012.

104

Bibliography

[103] D. K. Schweppe, J. R. Rigas, and S. A. Gerber. Quantitative phosphoproteomicprofiling of human non-small cell lung cancer tumors. J Proteomics, 91:286–96,2013.

[104] P. L. Ross, Y. N. Huang, J. N. Marchese, B. Williamson, K. Parker, S. Hattan,N. Khainovski, S. Pillai, S. Dey, S. Daniels, S. Purkayastha, P. Juhasz, S. Martin,M. Bartlet-Jones, F. He, A. Jacobson, and D. J. Pappin. Multiplexed protein quan-titation in saccharomyces cerevisiae using amine-reactive isobaric tagging reagents.Mol Cell Proteomics, 3(12):1154–69, 2004.

[105] A. Thompson, J. Schafer, K. Kuhn, S. Kienle, J. Schwarz, G. Schmidt, T. Neumann,R. Johnstone, A. K. Mohammed, and C. Hamon. Tandem mass tags: a novelquantification strategy for comparative analysis of complex protein mixtures byms/ms. Anal Chem, 75(8):1895–904, 2003.

[106] G. Pottiez, J. Wiederin, H. S. Fox, and P. Ciborowski. Comparison of 4-plex to8-plex itraq quantitative measurements of proteins in human plasma samples. JProteome Res, 11(7):3774–81, 2012.

[107] T. Werner, G. Sweetman, M. F. Savitski, T. Mathieson, M. Bantscheff, and M. M.Savitski. Ion coalescence of neutron encoded tmt 10-plex reporter ions. Anal Chem,86(7):3594–601, 2014.

[108] P. J. Boersema, R. Raijmakers, S. Lemeer, S. Mohammed, and A. J. Heck. Multiplexpeptide stable isotope dimethyl labeling for quantitative proteomics. Nat Protoc,4(4):484–94, 2009.

[109] J. L. Hsu, S. Y. Huang, N. H. Chow, and S. H. Chen. Stable-isotope dimethyllabeling for quantitative proteomics. Anal Chem, 75(24):6843–52, 2003.

[110] M. Miyagi and K. C. Rao. Proteolytic 18o-labeling strategies for quantitative pro-teomics. Mass Spectrom Rev, 26(1):121–36, 2007.

[111] M. Schnolzer, P. Jedrzejewski, and W. D. Lehmann. Protease-catalyzed incorpora-tion of 18o into peptide fragments and its application for protein sequencing by elec-trospray and matrix-assisted laser desorption/ionization mass spectrometry. Elec-trophoresis, 17(5):945–53, 1996.

[112] S. P. Gygi, B. Rist, S. A. Gerber, F. Turecek, M. H. Gelb, and R. Aebersold.Quantitative analysis of complex protein mixtures using isotope-coded affinity tags.Nat Biotechnol, 17(10):994–9, 1999.

[113] E. C. Yi, X. J. Li, K. Cooke, H. Lee, B. Raught, A. Page, V. Aneliunas, P. Hieter,D. R. Goodlett, and R. Aebersold. Increased quantitative proteome coverage with(13)c/(12)c-based, acid-cleavable isotope-coded affinity tag reagent and modifieddata acquisition scheme. Proteomics, 5(2):380–7, 2005.

[114] A. H. America and J. H. Cordewener. Comparative lc-ms: a landscape of peaks andvalleys. Proteomics, 8(4):731–49, 2008.

[115] D. H. Lundgren, S. I. Hwang, L. Wu, and D. K. Han. Role of spectral counting inquantitative proteomics. Expert Rev Proteomics, 7(1):39–53, 2010.

105

Bibliography

[116] H. Liu, R. G. Sadygov, and 3rd Yates, J. R. A model for random samplingand estimation of relative protein abundance in shotgun proteomics. Anal Chem,76(14):4193–201, 2004.

[117] D. A. Megger, T. Bracht, H. E. Meyer, and B. Sitek. Label-free quantification inclinical proteomics. Biochim Biophys Acta, 1834(8):1581–90, 2013.

[118] R. D. Voyksner and H. Lee. Investigating the use of an octupole ion guide for ionstorage and high-pass mass filtering to improve the quantitative performance of elec-trospray ion trap mass spectrometry. Rapid Commun Mass Spectrom, 13(14):1427–37, 1999.

[119] W. Wang, H. Zhou, H. Lin, S. Roy, T. A. Shaler, L. R. Hill, S. Norton, P. Ku-mar, M. Anderle, and C. H. Becker. Quantification of proteins and metabolitesby mass spectrometry without isotopic labeling or spiked standards. Anal Chem,75(18):4818–26, 2003.

[120] W. M. Old, K. Meyer-Arendt, L. Aveline-Wolf, K. G. Pierce, A. Mendoza, J. R.Sevinsky, K. A. Resing, and N. G. Ahn. Comparison of label-free methods for quan-tifying human proteins by shotgun proteomics. Mol Cell Proteomics, 4(10):1487–502,2005.

[121] V. Brun, C. Masselon, J. Garin, and A. Dupuis. Isotope dilution strategies forabsolute quantitative proteomics. J Proteomics, 72(5):740–9, 2009.

[122] S. A. Gerber, J. Rush, O. Stemman, M. W. Kirschner, and S. P. Gygi. Absolutequantification of proteins and phosphoproteins from cell lysates by tandem ms. ProcNatl Acad Sci U S A, 100(12):6940–5, 2003.

[123] R. J. Beynon, M. K. Doherty, J. M. Pratt, and S. J. Gaskell. Multiplexed absolutequantification in proteomics using artificial qcat proteins of concatenated signaturepeptides. Nat Methods, 2(8):587–9, 2005.

[124] S. Hanke, H. Besir, D. Oesterhelt, and M. Mann. Absolute silac for accurate quan-titation of proteins in complex mixtures down to the attomole level. J ProteomeRes, 7(3):1118–30, 2008.

[125] V. Brun, A. Dupuis, A. Adrait, M. Marcellin, D. Thomas, M. Court, F. Vandenesch,and J. Garin. Isotope-labeled protein standards: toward absolute quantitative pro-teomics. Mol Cell Proteomics, 6(12):2139–49, 2007.

[126] S. Singh, M. Springer, J. Steen, M. W. Kirschner, and H. Steen. Flexiquant: a noveltool for the absolute quantification of proteins, and the simultaneous identificationand quantification of potentially modified peptides. J Proteome Res, 8(5):2201–10,2009.

[127] M. Zeiler, W. L. Straube, E. Lundberg, M. Uhlen, and M. Mann. A protein epitopesignature tag (prest) library allows silac-based absolute quantification and multi-plexed determination of protein copy numbers in cell lines. Mol Cell Proteomics,11(3):O111 009613, 2012.

[128] M. Zeiler, M. Moser, and M. Mann. Copy number analysis of the murine plateletproteome spanning the complete abundance range. Mol Cell Proteomics, in press,2014.

106

Bibliography

[129] B. Schwanhausser, D. Busse, N. Li, G. Dittmar, J. Schuchhardt, J. Wolf, W. Chen,and M. Selbach. Global quantification of mammalian gene expression control. Na-ture, 473(7347):337–42, 2011.

[130] J. C. Silva, M. V. Gorenstein, G. Z. Li, J. P. Vissers, and S. J. Geromanos. Absolutequantification of proteins by lcmse: a virtue of parallel ms acquisition. Mol CellProteomics, 5(1):144–56, 2006.

[131] Y. Ishihama, Y. Oda, T. Tabata, T. Sato, T. Nagasu, J. Rappsilber, and M. Mann.Exponentially modified protein abundance index (empai) for estimation of absoluteprotein amount in proteomics by the number of sequenced peptides per protein. MolCell Proteomics, 4(9):1265–72, 2005.

[132] P. Lu, C. Vogel, R. Wang, X. Yao, and E. M. Marcotte. Absolute protein expressionprofiling estimates the relative contributions of transcriptional and translationalregulation. Nat Biotechnol, 25(1):117–24, 2007.

[133] J. Grossmann, B. Roschitzki, C. Panse, C. Fortes, S. Barkow-Oesterreicher,D. Rutishauser, and R. Schlapbach. Implementation and evaluation of relativeand absolute quantification in shotgun proteomics with label-free methods. J Pro-teomics, 73(9):1740–6, 2010.

[134] L. Arike, K. Valgepea, L. Peil, R. Nahku, K. Adamberg, and R. Vilu. Compar-ison and applications of label-free absolute proteome quantification methods onescherichia coli. J Proteomics, 75(17):5437–48, 2012.

[135] Z. Li, R. M. Adams, K. Chourey, G. B. Hurst, R. L. Hettich, and C. Pan. System-atic comparison of label-free, metabolic labeling, and isobaric chemical labeling forquantitative proteomics on ltq orbitrap velos. J Proteome Res, 11(3):1582–90, 2012.

[136] P. Pichler, T. Kocher, J. Holzmann, M. Mazanek, T. Taus, G. Ammerer, andK. Mechtler. Peptide labeling with isobaric tags yields higher identification ratesusing itraq 4-plex compared to tmt 6-plex and itraq 8-plex on ltq orbitrap. AnalChem, 82(15):6549–58, 2010.

[137] A. Sandberg, R. M. Branca, J. Lehtio, and J. Forshed. Quantitative accuracy inmass spectrometry based proteomics of complex samples: the impact of labelingand precursor interference. J Proteomics, 96:133–44, 2013.

[138] A. F. Altelaar, C. K. Frese, C. Preisinger, M. L. Hennrich, A. W. Schram, H. T.Timmers, A. J. Heck, and S. Mohammed. Benchmarking stable isotope labelingbased quantitative proteomics. J Proteomics, 88:14–26, 2013.

[139] M. Mann. Functional and quantitative proteomics using silac. Nat Rev Mol CellBiol, 7(12):952–8, 2006.

[140] L. M. de Godoy, J. V. Olsen, G. A. de Souza, G. Li, P. Mortensen, and M. Mann.Status of complete proteome analysis by mass spectrometry: Silac labeled yeast asa model system. Genome Biol, 7(6):R50, 2006.

[141] L. M. de Godoy, J. V. Olsen, J. Cox, M. L. Nielsen, N. C. Hubner, F. Frohlich,T. C. Walther, and M. Mann. Comprehensive mass-spectrometry-based proteomequantification of haploid versus diploid yeast. Nature, 455(7217):1251–4, 2008.

107

Bibliography

[142] N. Nagaraj, N. A. Kulak, J. Cox, N. Neuhauser, K. Mayr, O. Hoerning, O. Vorm,and M. Mann. System-wide perturbation analysis with nearly complete coverage ofthe yeast proteome by single-shot ultra hplc runs on a bench top orbitrap. Mol CellProteomics, 11(3):M111 013722, 2012.

[143] A. S. Hebert, A. L. Richards, D. J. Bailey, A. Ulbrich, E. E. Coughlin, M. S.Westphall, and J. J. Coon. The one hour yeast proteome. Mol Cell Proteomics,13(1):339–47, 2014.

[144] M. Beck, A. Schmidt, J. Malmstroem, M. Claassen, A. Ori, A. Szymborska, F. Her-zog, O. Rinner, J. Ellenberg, and R. Aebersold. The quantitative proteome of ahuman cell line. Mol Syst Biol, 7:549, 2011.

[145] F. Desiere, E. W. Deutsch, N. L. King, A. I. Nesvizhskii, P. Mallick, J. Eng, S. Chen,J. Eddes, S. N. Loevenich, and R. Aebersold. The peptideatlas project. Nucleic AcidsRes, 34(Database issue):D655–8, 2006.

[146] T. Farrah, E. W. Deutsch, G. S. Omenn, Z. Sun, J. D. Watts, T. Yamamoto,D. Shteynberg, M. M. Harris, and R. L. Moritz. State of the human proteomein 2013 as viewed through peptideatlas: comparing the kidney, urine, and plasmaproteomes for the biology- and disease-driven human proteome project. J ProteomeRes, 13(1):60–75, 2014.

[147] H. A. Ebhardt, E. Sabido, R. Huttenhain, B. Collins, and R. Aebersold. Range ofprotein detection by selected/multiple reaction monitoring mass spectrometry in anunfractionated human cell culture lysate. Proteomics, 12(8):1185–93, 2012.

[148] T. A. Addona, S. E. Abbatiello, B. Schilling, S. J. Skates, D. R. Mani, D. M.Bunk, C. H. Spiegelman, L. J. Zimmerman, A. J. Ham, H. Keshishian, S. C. Hall,S. Allen, R. K. Blackman, C. H. Borchers, C. Buck, H. L. Cardasis, M. P. Cusack,N. G. Dodder, B. W. Gibson, J. M. Held, T. Hiltke, A. Jackson, E. B. Johansen,C. R. Kinsinger, J. Li, M. Mesri, T. A. Neubert, R. K. Niles, T. C. Pulsipher,D. Ransohoff, H. Rodriguez, P. A. Rudnick, D. Smith, D. L. Tabb, T. J. Tegeler,A. M. Variyath, L. J. Vega-Montoto, A. Wahlander, S. Waldemarson, M. Wang, J. R.Whiteaker, L. Zhao, N. L. Anderson, S. J. Fisher, D. C. Liebler, A. G. Paulovich,F. E. Regnier, P. Tempst, and S. A. Carr. Multi-site assessment of the precisionand reproducibility of multiple reaction monitoring-based measurements of proteinsin plasma. Nat Biotechnol, 27(7):633–41, 2009.

[149] H. Keshishian, T. Addona, M. Burgess, E. Kuhn, and S. A. Carr. Quantitative, mul-tiplexed assays for low abundance proteins in plasma by targeted mass spectrometryand stable isotope dilution. Mol Cell Proteomics, 6(12):2212–29, 2007.

[150] T. Shi, X. Sun, Y. Gao, T. L. Fillmore, A. A. Schepmoes, R. Zhao, J. He, R. J.Moore, J. Kagan, K. D. Rodland, T. Liu, A. Y. Liu, R. D. Smith, K. Tang, 2ndCamp, D. G., and W. J. Qian. Targeted quantification of low ng/ml level proteinsin human serum without immunoaffinity depletion. J Proteome Res, 12(7):3353–61,2013.

[151] M. Pernemalm and J. Lehtio. Mass spectrometry-based plasma proteomics: stateof the art and future outlook. Expert Rev Proteomics, 11(4):431–48, 2014.

108

Bibliography

[152] L. Anderson. Six decades searching for meaning in the proteome. J Proteomics,107:24–30, 2014.

[153] R. S. Yalow and S. A. Berson. Immunoassay of endogenous plasma insulin in man.J Clin Invest, 39:1157–75, 1960.

[154] L. E. Miles and C. N. Hales. The preparation and properties of purified 125-i-labelledantibodies to insulin. Biochem J, 108(4):611–8, 1968.

[155] L. E. Miles and C. N. Hales. Immunoradiometric assay of human growth hormone.Lancet, 2(7566):492–3, 1968.

[156] E. Engvall and P. Perlmann. Enzyme-linked immunosorbent assay (elisa). quanti-tative assay of immunoglobulin g. Immunochemistry, 8(9):871–4, 1971.

[157] B. K. Van Weemen and A. H. Schuurs. Immunoassay using antigen-enzyme conju-gates. FEBS Lett, 15(3):232–236, 1971.

[158] R. M. Lequin. Enzyme immunoassay (eia)/enzyme-linked immunosorbent assay(elisa). Clin Chem, 51(12):2415–8, 2005.

[159] P. Chames, M. Van Regenmortel, E. Weiss, and D. Baty. Therapeutic antibodies:successes, limitations and hopes for the future. Br J Pharmacol, 157(2):220–33,2009.

[160] P. Parham. The immune system. Garland, New York, NY, 2. edition, 2005.

[161] B. Hjelm, B. Forsstrom, J. Lofblom, J. Rockberg, and M. Uhlen. Parallel immu-nizations of rabbits using the same antigen yield antibodies with similar, but notidentical, epitopes. PLoS One, 7(12):e45817, 2012.

[162] G. Kohler and C. Milstein. Continuous cultures of fused cells secreting antibody ofpredefined specificity. Nature, 256(5517):495–7, 1975.

[163] M. Leenaars and C. F. Hendriksen. Critical steps in the production of polyclonaland monoclonal antibodies: evaluation and recommendations. ILAR J, 46(3):269–79, 2005.

[164] C. M. Hammers and J. R. Stanley. Antibody phage display: technique and appli-cations. J Invest Dermatol, 134(2):e17, 2014.

[165] H. J. Dyson and P. E. Wright. Antigenic peptides. FASEB J, 9(1):37–42, 1995.

[166] T. Olafsen and A. M. Wu. Antibody vectors for imaging. Semin Nucl Med,40(3):167–81, 2010.

[167] Consortium Structural Genomics, Consortium China Structural Genomics, Consor-tium Northeast Structural Genomics, S. Graslund, P. Nordlund, J. Weigelt, B. M.Hallberg, J. Bray, O. Gileadi, S. Knapp, U. Oppermann, C. Arrowsmith, R. Hui,J. Ming, S. dhe Paganon, H. W. Park, A. Savchenko, A. Yee, A. Edwards, R. Vin-centelli, C. Cambillau, R. Kim, S. H. Kim, Z. Rao, Y. Shi, T. C. Terwilliger, C. Y.Kim, L. W. Hung, G. S. Waldo, Y. Peleg, S. Albeck, T. Unger, O. Dym, J. Prilusky,J. L. Sussman, R. C. Stevens, S. A. Lesley, I. A. Wilson, A. Joachimiak, F. Collart,I. Dementieva, M. I. Donnelly, W. H. Eschenfeldt, Y. Kim, L. Stols, R. Wu, M. Zhou,S. K. Burley, J. S. Emtage, J. M. Sauder, D. Thompson, K. Bain, J. Luz, T. Gheyi,

109

Bibliography

F. Zhang, S. Atwell, S. C. Almo, J. B. Bonanno, A. Fiser, S. Swaminathan, F. W.Studier, M. R. Chance, A. Sali, T. B. Acton, R. Xiao, L. Zhao, L. C. Ma, J. F.Hunt, L. Tong, K. Cunningham, M. Inouye, S. Anderson, H. Janjua, R. Shastry,C. K. Ho, D. Wang, H. Wang, M. Jiang, G. T. Montelione, D. I. Stuart, R. J.Owens, S. Daenke, A. Schutz, U. Heinemann, S. Yokoyama, K. Bussow, and K. C.Gunsalus. Protein production and purification. Nat Methods, 5(2):135–46, 2008.

[168] A. L. Nelson. Antibody fragments: hope and hype. MAbs, 2(1):77–83, 2010.

[169] J. Lofblom, F. Y. Frejd, and S. Stahl. Non-immunoglobulin based protein scaffolds.Curr Opin Biotechnol, 22(6):843–8, 2011.

[170] J. Lofblom, J. Feldwisch, V. Tolmachev, J. Carlsson, S. Stahl, and F. Y. Frejd. Affi-body molecules: engineered proteins for therapeutic, diagnostic and biotechnologicalapplications. FEBS Lett, 584(12):2670–80, 2010.

[171] J. N. Ahmad, Biedermannova L. Li, J., M. Kuchar, H. Sipova, A. Semeradtova,J. Cerny, H. Petrokova, P. Mikulecky, J. Polinek, O. Stanek, J. Vondrasek, J. Ho-mola, J. Maly, R. Osicka, P. Sebo, and P. Maly. Novel high-affinity binders ofhuman interferon gamma derived from albumin-binding domain of protein g. Pro-teins, 80(3):774–89, 2012.

[172] T. Alm, L. Yderland, J. Nilvebrant, A. Halldin, and S. Hober. A small bispecificprotein selected for orthogonal affinity purification. Biotechnol J, 5(6):605–17, 2010.

[173] H. K. Binz, M. T. Stumpp, P. Forrer, P. Amstutz, and A. Pluckthun. Design-ing repeat proteins: well-expressed, soluble and stable proteins from combinatoriallibraries of consensus ankyrin repeat proteins. J Mol Biol, 332(2):489–503, 2003.

[174] D. Lipovsek. Adnectins: engineered target-binding protein therapeutics. ProteinEng Des Sel, 24(1-2):3–9, 2011.

[175] A. Skerra. Alternative binding proteins: anticalins - harnessing the structural plas-ticity of the lipocalin ligand pocket to engineer novel binding activities. FEBS J,275(11):2677–83, 2008.

[176] K. M. Song, S. Lee, and C. Ban. Aptamers and their biological applications. Sensors(Basel), 12(1):612–31, 2012.

[177] M. Uhlen, P. Oksvold, L. Fagerberg, E. Lundberg, K. Jonasson, M. Forsberg,M. Zwahlen, C. Kampf, K. Wester, S. Hober, H. Wernerus, L. Bjorling, and F. Pon-ten. Towards a knowledge-based human protein atlas. Nat Biotechnol, 28(12):1248–50, 2010.

[178] M. Lindskog, J. Rockberg, M. Uhlen, and F. Sterky. Selection of protein epitopesfor antibody production. Biotechniques, 38(5):723–7, 2005.

[179] H. Tegel, J. Steen, A. Konrad, H. Nikdin, K. Pettersson, M. Stenvall, S. Tourle,U. Wrethagen, L. Xu, L. Yderland, M. Uhlen, S. Hober, and J. Ottosson. High-throughput protein production–lessons from scaling up from 10 to 288 recombinantproteins per week. Biotechnol J, 4(1):51–7, 2009.

[180] C. Agaton, R. Falk, I. Hoiden Guthenberg, L. Gostring, M. Uhlen, and S. Hober.Selective enrichment of monospecific polyclonal antibodies for antibody-based pro-teomics efforts. J Chromatogr A, 1043(1):33–40, 2004.

110

Bibliography

[181] P. Nilsson, L. Paavilainen, K. Larsson, J. Odling, M. Sundberg, A. C. Andersson,C. Kampf, A. Persson, C. Al-Khalili Szigyarto, J. Ottosson, E. Bjorling, S. Hober,H. Wernerus, K. Wester, F. Ponten, and M. Uhlen. Towards a human proteomeatlas: high-throughput generation of mono-specific antibodies for tissue profiling.Proteomics, 5(17):4327–37, 2005.

[182] C. Algenas, C. Agaton, L. Fagerberg, A. Asplund, L. Bjorling, E. Bjorling,C. Kampf, E. Lundberg, P. Nilsson, A. Persson, K. Wester, F. Ponten, H. Wernerus,M. Uhlen, J. Ottosson Takanen, and S. Hober. Antibody performance in westernblot applications is context-dependent. Biotechnol J, 2014.

[183] F. Ponten, J. M. Schwenk, A. Asplund, and P. H. Edqvist. The human protein atlasas a proteomic resource for biomarker discovery. J Intern Med, 270(5):428–46, 2011.

[184] A. Asplund, P. H. Edqvist, J. M. Schwenk, and F. Ponten. Antibodies for profilingthe human proteome-the human protein atlas as a resource for cancer research.Proteomics, 12(13):2067–77, 2012.

[185] F. Danielsson, M. Wiking, D. Mahdessian, M. Skogs, H. Ait Blal, M. Hjelmare,C. Stadler, M. Uhlen, and E. Lundberg. Rna deep sequencing as a tool for selectionof cell lines for systematic subcellular localization of all human proteins. J ProteomeRes, 12(1):299–307, 2013.

[186] I. Matic, E. G. Jaffray, S. K. Oxenham, M. J. Groves, C. L. Barratt, S. Tauro,N. R. Stanley-Wall, and R. T. Hay. Absolute silac-compatible expression strainallows sumo-2 copy number determination in clinical samples. J Proteome Res,10(10):4869–75, 2011.

[187] J. Holzmann, P. Pichler, M. Madalinski, R. Kurzbauer, and K. Mechtler. Stoichiom-etry determination of the mp1-p14 complex using a novel and cost-efficient methodto produce an equimolar mixture of standard peptides. Anal Chem, 81(24):10254–61, 2009.

[188] T. Sano, C. L. Smith, and C. R. Cantor. Immuno-pcr: very sensitive antigendetection by means of specific antibody-dna conjugates. Science, 258(5079):120–22,1992.

[189] S. F. Kingsmore and D. D. Patel. Multiplexed protein profiling on antibody-basedmicroarrays by rolling circle amplification. Curr Opin Biotechnol, 14(1):74–81, 2003.

[190] D. M. Rissin, C. W. Kan, T. G. Campbell, S. C. Howes, D. R. Fournier, L. Song,T. Piech, P. P. Patel, L. Chang, A. J. Rivnak, E. P. Ferrell, J. D. Randall, G. K.Provuncher, D. R. Walt, and D. C. Duffy. Single-molecule enzyme-linked im-munosorbent assay detects serum proteins at subfemtomolar concentrations. NatBiotechnol, 28(6):595–9, 2010.

[191] D. M. Rissin, C. W. Kan, L. Song, A. J. Rivnak, M. W. Fishburn, Q. Shao, T. Piech,E. P. Ferrell, R. E. Meyer, T. G. Campbell, D. R. Fournier, and D. C. Duffy.Multiplexed single molecule immunoassays. Lab Chip, 13(15):2902–11, 2013.

[192] H. C. Tekin, M. Cornaglia, and M. A. Gijs. Attomolar protein detection using amagnetic bead surface coverage assay. Lab Chip, 13(6):1053–9, 2013.

111

Bibliography

[193] S. F. Kingsmore. Multiplexed protein measurement: technologies and applicationsof protein and antibody arrays. Nat Rev Drug Discov, 5(4):310–20, 2006.

[194] R. P. Ekins. Multi-analyte immunoassay. J Pharm Biomed Anal, 7(2):155–68, 1989.

[195] R. J. Fulton, R. L. McDade, P. L. Smith, L. J. Kienker, and Jr. Kettman, J. R.Advanced multiplexed analysis with the flowmetrix system. Clin Chem, 43(9):1749–56, 1997.

[196] D. A. Hall, J. Ptacek, and M. Snyder. Protein microarray technology. Mech AgeingDev, 128(1):161–7, 2007.

[197] K. Drobin, P. Nilsson, and J. M. Schwenk. Highly multiplexed antibody suspensionbead arrays for plasma protein profiling. Methods Mol Biol, 1023:137–45, 2013.

[198] D. Juncker, S. Bergeron, V. Laforte, and H. Li. Cross-reactivity in antibody mi-croarrays and multiplexed sandwich assays: shedding light on the dark side of mul-tiplexing. Curr Opin Chem Biol, 18:29–37, 2014.

[199] S. Hu, Z. Xie, J. Qian, S. Blackshaw, and H. Zhu. Functional protein microarraytechnology. Wiley Interdiscip Rev Syst Biol Med, 3(3):255–68, 2011.

[200] L. Perlee, J. Christiansen, R. Dondero, B. Grimwade, S. Lejnine, M. Mullenix,W. Shao, M. Sorette, V. Tchernev, D. Patel, and S. Kingsmore. Developmentand standardization of multiplexed antibody microarrays for use in quantitativeproteomics. Proteome Sci, 2(1):9, 2004.

[201] D. J. Phillips, S. C. League, P. Weinstein, and W. C. Hooper. Interference in micro-sphere flow cytometric multiplexed immunoassays for human cytokine estimation.Cytokine, 36(3-4):180–8, 2006.

[202] N. C. dupont, K. Wang, P. D. Wadhwa, J. F. Culhane, and E. L. Nelson. Validationand comparison of luminex multiplex cytokine analysis kits with elisa: determina-tions of a panel of nine cytokines in clinical sample culture supernatants. J ReprodImmunol, 66(2):175–91, 2005.

[203] L. Dossus, S. Becker, D. Achaintre, R. Kaaks, and S. Rinaldi. Validity of multiplex-based assays for cytokine measurements in serum and plasma from ”non-diseased”subjects: comparison with elisa. J Immunol Methods, 350(1-2):125–32, 2009.

[204] J. A. Bastarache, T. Koyama, N. E. Wickersham, and L. B. Ware. Validation ofa multiplex electrochemiluminescent immunoassay platform in human and mousesamples. J Immunol Methods, 2014.

[205] M. F. Elshal and J. P. McCoy. Multiplex bead array assays: performance evaluationand comparison of sensitivity to elisa. Methods, 38(4):317–23, 2006.

[206] M. Pla-Roca, R. F. Leulmi, S. Tourekhanova, S. Bergeron, V. Laforte, E. Moreau,S. J. Gosline, N. Bertos, M. Hallett, M. Park, and D. Juncker. Antibody colocal-ization microarray: a scalable technology for multiplex protein analysis in complexsamples. Mol Cell Proteomics, 11(4):M111 011460, 2012.

[207] J. M. Schwenk, U. Igel, M. Neiman, H. Langen, C. Becker, A. Bjartell, F. Pon-ten, F. Wiklund, H. Gronberg, P. Nilsson, and M. Uhlen. Toward next generationplasma profiling via heat-induced epitope retrieval and array-based assays. Mol CellProteomics, 9(11):2497–507, 2010.

112

Bibliography

[208] O. Poetz, T. Henzler, M. Hartmann, C. Kazmaier, M. F. Templin, T. Herget, andT. O. Joos. Sequential multiplex analyte capturing for phosphoprotein profiling.Mol Cell Proteomics, 9(11):2474–81, 2010.

[209] S. Fredriksson, M. Gullberg, J. Jarvius, C. Olsson, K. Pietras, S. M. Gustafsdottir,A. Ostman, and U. Landegren. Protein detection using proximity-dependent dnaligation assays. Nat Biotechnol, 20(5):473–7, 2002.

[210] M. Lundberg, S. B. Thorsen, E. Assarsson, A. Villablanca, B. Tran, N. Gee,M. Knowles, B. S. Nielsen, E. Gonzalez Couto, R. Martin, O. Nilsson, C. Fer-mer, J. Schlingemann, I. J. Christensen, H. J. Nielsen, B. Ekstrom, C. Andersson,M. Gustafsson, N. Brunner, J. Stenvang, and S. Fredriksson. Multiplexed homoge-neous proximity ligation assays for high-throughput protein biomarker research inserological material. Mol Cell Proteomics, 10(4):M110 004978, 2011.

[211] M. Lundberg, A. Eriksson, B. Tran, E. Assarsson, and S. Fredriksson. Homogeneousantibody-based proximity extension assays provide sensitive and specific detectionof low-abundant proteins in human blood. Nucleic Acids Res, 39(15):e102, 2011.

[212] H. Zhang and D. W. Chan. Cancer biomarker discovery in plasma using a tissue-targeted proteomic approach. Cancer Epidemiol Biomarkers Prev, 16(10):1915–7,2007.

[213] J. R. Whiteaker, L. Zhao, H. Y. Zhang, L. C. Feng, B. D. Piening, L. Anderson,and A. G. Paulovich. Antibody-based enrichment of peptides on magnetic beadsfor mass-spectrometry-based quantification of serum biomarkers. Anal Biochem,362(1):44–54, 2007.

[214] E. Kuhn, T. Addona, H. Keshishian, M. Burgess, D. R. Mani, R. T. Lee, M. S.Sabatine, R. E. Gerszten, and S. A. Carr. Developing multiplexed assays for troponini and interleukin-33 in plasma by peptide immunoaffinity enrichment and targetedmass spectrometry. Clin Chem, 55(6):1108–17, 2009.

[215] A. N. Hoofnagle, J. O. Becker, M. H. Wener, and J. W. Heinecke. Quantifica-tion of thyroglobulin, a low-abundance serum protein, by immunoaffinity peptideenrichment and tandem mass spectrometry. Clin Chem, 54(11):1796–804, 2008.

[216] J. R. Whiteaker, C. Lin, J. Kennedy, L. Hou, M. Trute, I. Sokal, P. Yan, R. M.Schoenherr, L. Zhao, U. J. Voytovich, K. S. Kelly-Spratt, A. Krasnoselsky, P. R.Gafken, J. M. Hogan, L. A. Jones, P. Wang, L. Amon, L. A. Chodosh, P. S. Nelson,M. W. McIntosh, C. J. Kemp, and A. G. Paulovich. A targeted proteomics-basedpipeline for verification of biomarkers in plasma. Nat Biotechnol, 29(7):625–34, 2011.

[217] H. Neubert, J. Gale, and D. Muirhead. Online high-flow peptide immunoaffin-ity enrichment and nanoflow lc-ms/ms: assay development for total salivarypepsin/pepsinogen. Clin Chem, 56(9):1413–23, 2010.

[218] J. R. Whiteaker, L. Zhao, L. Anderson, and A. G. Paulovich. An automatedand multiplexed method for high throughput peptide immunoaffinity enrichmentand multiple reaction monitoring mass spectrometry-based quantification of proteinbiomarkers. Mol Cell Proteomics, 9(1):184–96, 2010.

113

Bibliography

[219] J. R. Whiteaker, L. Zhao, C. Lin, P. Yan, P. Wang, and A. G. Paulovich. Sequentialmultiplexed analyte quantification using peptide immunoaffinity enrichment coupledto mass spectrometry. Mol Cell Proteomics, 11(6):M111 015347, 2012.

[220] N. L. Anderson, M. Razavi, T. W. Pearson, G. Kruppa, R. Paape, and D. Suckau.Precision of heavy-light peptide ratios measured by maldi-tof mass spectrometry. JProteome Res, 11(3):1868–78, 2012.

[221] E. Szajli, T. Feher, and K. F. Medzihradszky. Investigating the quantitative natureof maldi-tof ms. Mol Cell Proteomics, 7(12):2410–8, 2008.

[222] M. Razavi, L. E. Frick, W. A. LaMarr, M. E. Pope, C. A. Miller, N. L. Ander-son, and T. W. Pearson. High-throughput siscapa quantitation of peptides fromhuman plasma digests by ultrafast, liquid chromatography-free mass spectrometry.J Proteome Res, 11(12):5642–9, 2012.

[223] E. N. Warren, J. Jiang, C. E. Parker, and C. H. Borchers. Absolute quantitation ofcancer-related proteins using an ms-based peptide chip. Biotechniques, Suppl:7–11,2005.

[224] J. Jiang, C. E. Parker, K. A. Hoadley, C. M. Perou, G. Boysen, and C. H. Borchers.Development of an immuno tandem mass spectrometry (imaldi) assay for egfr diag-nosis. Proteomics Clin Appl, 1(12):1651–9, 2007.

[225] J. Jiang, C. E. Parker, J. R. Fuller, T. H. Kawula, and C. H. Borchers. An im-munoaffinity tandem mass spectrometry (imaldi) assay for detection of francisellatularensis. Anal Chim Acta, 605(1):70–9, 2007.

[226] J. D. Reid, D. T. Holmes, D. R. Mason, B. Shah, and C. H. Borchers. Towards thedevelopment of an immuno maldi (imaldi) mass spectrometry assay for the diagnosisof hypertension. J Am Soc Mass Spectrom, 21(10):1680–6, 2010.

[227] D. R. Mason, J. D. Reid, A. G. Camenzind, D. T. Holmes, and C. H. Borchers.Duplexed imaldi for the detection of angiotensin i and angiotensin ii. Methods,56(2):213–22, 2012.

[228] A. G. Camenzind, J. G. van der Gugten, R. Popp, D. T. Holmes, and C. H. Borchers.Development and evaluation of an immuno-maldi (imaldi) assay for angiotensin iand the diagnosis of secondary hypertension. Clin Proteomics, 10(1):20, 2013.

[229] F. Weiss, B. H. van den Berg, H. Planatscher, C. J. Pynn, T. O. Joos, and O. Poetz.Catch and measure-mass spectrometry-based immunoassays in biomarker research.Biochim Biophys Acta, 1844(5):927–32, 2014.

[230] M. Uhlen and F. Ponten. Antibody-based proteomics for human tissue profiling.Mol Cell Proteomics, 4(4):384–93, 2005.

[231] R. W. Nelson, J. R. Krone, A. L. Bieber, and P. Williams. Mass spectrometricimmunoassay. Anal Chem, 67(7):1153–8, 1995.

[232] U. A. Kiernan, D. Nedelkov, K. A. Tubbs, E. E. Niederkofler, and R. W. Nelson.Proteomic characterization of novel serum amyloid p component variants from hu-man plasma and urine. Proteomics, 4(6):1825–9, 2004.

114

Bibliography

[233] U. A. Kiernan, R. Addobbati, D. Nedelkov, and R. W. Nelson. Quantitativemultiplexed c-reactive protein mass spectrometric immunoassay. J Proteome Res,5(7):1682–7, 2006.

[234] K. A. Tubbs, U. A. Kiernan, E. E. Niederkofler, D. Nedelkov, A. L. Bieber, andR. W. Nelson. Development of recombinant-based mass spectrometric immunoassaywith application to resistin expression profiling. Anal Chem, 78(10):3271–6, 2006.

[235] U. A. Kiernan, K. A. Tubbs, K. Gruber, D. Nedelkov, E. E. Niederkofler,P. Williams, and R. W. Nelson. High-throughput protein characterization usingmass spectrometric immunoassay. Anal Biochem, 301(1):49–56, 2002.

[236] M. F. Lopez, T. Rezai, D. A. Sarracino, A. Prakash, B. Krastins, M. Athanas,R. J. Singh, D. R. Barnidge, P. Oran, C. Borges, and R. W. Nelson. Selected re-action monitoring-mass spectrometric immunoassay responsive to parathyroid hor-mone and related variants. Clin Chem, 56(2):281–90, 2010.

[237] A. Prakash, T. Rezai, B. Krastins, D. Sarracino, M. Athanas, P. Russo, H. Zhang,Y. Tian, Y. Li, V. Kulasingam, A. Drabovich, C. R. Smith, I. Batruch, P. E. Oran,C. Fredolini, A. Luchini, L. Liotta, E. Petricoin, E. P. Diamandis, D. W. Chan,R. Nelson, and M. F. Lopez. Interlaboratory reproducibility of selective reactionmonitoring assays using multiple upfront analyte enrichment strategies. J ProteomeRes, 11(8):3986–95, 2012.

[238] R. W. Nelson, D. Nedelkov, K. A. Tubbs, and U. A. Kiernan. Quantitativemass spectrometric immunoassay of insulin like growth factor 1. J Proteome Res,3(4):851–5, 2004.

[239] P. E. Oran, J. W. Jarvis, C. R. Borges, N. D. Sherma, and R. W. Nelson. Massspectrometric immunoassay of intact insulin and related variants for populationproteomics studies. Proteomics Clin Appl, 5(7-8):454–9, 2011.

[240] O. Trenchevska and D. Nedelkov. Targeted quantitative mass spectrometric im-munoassay for human protein variants. Proteome Sci, 9(1):19, 2011.

[241] O. Trenchevska, E. Kamcheva, and D. Nedelkov. Mass spectrometric immunoas-say for quantitative determination of transthyretin and its variants. Proteomics,11(18):3633–41, 2011.

[242] U. Qundos, H. Johannesson, C. Fredolini, G. O’Hurley, R. Branca, M. Uhlén,F. Wiklund, A. Bjartell, P. Nilsson, and J. M. Schwenk. Analysis of plasma fromprostate cancer patients links decreased carnosine dipeptidase 1 levels to lymph nodemetastasis. Translational Proteomics, 2(0):14–24, 2014.

[243] M. Selbach and M. Mann. Protein interaction screening by quantitative immuno-precipitation combined with knockdown (quick). Nat Methods, 3(12):981–3, 2006.

[244] A. Malovannaya, Y. Li, Y. Bulynko, S. Y. Jung, Y. Wang, R. B. Lanz, B. W.O’Malley, and J. Qin. Streamlined analysis schema for high-throughput identifica-tion of endogenous protein complexes. Proc Natl Acad Sci U S A, 107(6):2431–6,2010.

115

Bibliography

[245] A. Malovannaya, R. B. Lanz, S. Y. Jung, Y. Bulynko, N. T. Le, D. W. Chan, C. Ding,Y. Shi, N. Yucer, G. Krenciute, B. J. Kim, C. Li, R. Chen, W. Li, Y. Wang, B. W.O’Malley, and J. Qin. Analysis of the human endogenous coregulator complexome.Cell, 145(5):787–99, 2011.

[246] N. C. Hubner, A. W. Bird, J. Cox, B. Splettstoesser, P. Bandilla, I. Poser, A. Hyman,and M. Mann. Quantitative proteomics combined with bac transgeneomics revealsin vivo protein interactions. J Cell Biol, 189(4):739–54, 2010.

[247] Y. Ho, A. Gruhler, A. Heilbut, G. D. Bader, L. Moore, S. L. Adams, A. Millar,P. Taylor, K. Bennett, K. Boutilier, L. Yang, C. Wolting, I. Donaldson, S. Schan-dorff, J. Shewnarane, M. Vo, J. Taggart, M. Goudreault, B. Muskat, C. Alfarano,D. Dewar, Z. Lin, K. Michalickova, A. R. Willems, H. Sassi, P. A. Nielsen, K. J.Rasmussen, J. R. Andersen, L. E. Johansen, L. H. Hansen, H. Jespersen, A. Podtele-jnikov, E. Nielsen, J. Crawford, V. Poulsen, B. D. Sorensen, J. Matthiesen, R. C.Hendrickson, F. Gleeson, T. Pawson, M. F. Moran, D. Durocher, M. Mann, C. W.Hogue, D. Figeys, and M. Tyers. Systematic identification of protein complexes insaccharomyces cerevisiae by mass spectrometry. Nature, 415(6868):180–3, 2002.

[248] M. Varjosalo, R. Sacco, A. Stukalov, A. van Drogen, M. Planyavsky, S. Hauri, R. Ae-bersold, K. L. Bennett, J. Colinge, M. Gstaiger, and G. Superti-Furga. Interlabora-tory reproducibility of large-scale human protein-complex analysis by standardizedap-ms. Nat Methods, 10(4):307–14, 2013.

[249] G. Rigaut, A. Shevchenko, B. Rutz, M. Wilm, M. Mann, and B. Seraphin. A genericprotein purification method for protein complex characterization and proteome ex-ploration. Nat Biotechnol, 17(10):1030–2, 1999.

[250] A. C. Gavin, M. Bosche, R. Krause, P. Grandi, M. Marzioch, A. Bauer, J. Schultz,J. M. Rick, A. M. Michon, C. M. Cruciat, M. Remor, C. Hofert, M. Schelder,M. Brajenovic, H. Ruffner, A. Merino, K. Klein, M. Hudak, D. Dickson, T. Rudi,V. Gnau, A. Bauch, S. Bastuck, B. Huhse, C. Leutwein, M. A. Heurtier, R. R. Cop-ley, A. Edelmann, E. Querfurth, V. Rybin, G. Drewes, M. Raida, T. Bouwmeester,P. Bork, B. Seraphin, B. Kuster, G. Neubauer, and G. Superti-Furga. Functionalorganization of the yeast proteome by systematic analysis of protein complexes.Nature, 415(6868):141–7, 2002.

[251] R. Kittler, L. Pelletier, C. Ma, I. Poser, S. Fischer, A. A. Hyman, and F. Buch-holz. Rna interference rescue by bacterial artificial chromosome transgenesis inmammalian tissue culture cells. Proc Natl Acad Sci U S A, 102(7):2396–401, 2005.

[252] T. Berggard, S. Linse, and P. James. Methods for the detection and analysis ofprotein-protein interactions. Proteomics, 7(16):2833–42, 2007.

[253] T. P. Hopp, K. S. Prickett, V. L. Price, R. T. Libby, C. J. March, D. P. Cerretti,D. L. Urdal, and P. J. Conlon. A short polypeptide marker sequence useful forrecombinant protein identification and purification. Bio-Technology, 6(10):1204–1210, 1988.

[254] I. Poser, M. Sarov, J. R. Hutchins, J. K. Heriche, Y. Toyoda, A. Pozniakovsky,D. Weigl, A. Nitzsche, B. Hegemann, A. W. Bird, L. Pelletier, R. Kittler, S. Hua,

116

Bibliography

R. Naumann, M. Augsburg, M. M. Sykora, H. Hofemeister, Y. Zhang, K. Nasmyth,K. P. White, S. Dietzel, K. Mechtler, R. Durbin, A. F. Stewart, J. M. Peters,F. Buchholz, and A. A. Hyman. Bac transgeneomics: a high-throughput methodfor exploration of protein function in mammals. Nat Methods, 5(5):409–15, 2008.

[255] L. Trinkle-Mulcahy, S. Boulon, Y. W. Lam, R. Urcia, F. M. Boisvert, F. Vandermo-ere, N. A. Morrice, S. Swift, U. Rothbauer, H. Leonhardt, and A. Lamond. Identi-fying specific protein interaction partners using quantitative mass spectrometry andbead proteomes. J Cell Biol, 183(2):223–39, 2008.

[256] S. Boulon, Y. Ahmad, L. Trinkle-Mulcahy, C. Verheggen, A. Cobley, P. Gregor,E. Bertrand, M. Whitehorn, and A. I. Lamond. Establishment of a protein frequencylibrary and its application in the reliable identification of specific protein interactionpartners. Mol Cell Proteomics, 9(5):861–79, 2010.

[257] D. Mellacheruvu, Z. Wright, A. L. Couzens, J. P. Lambert, N. A. St-Denis, T. Li,Y. V. Miteva, S. Hauri, M. E. Sardiu, T. Y. Low, V. A. Halim, R. D. Bagshaw,N. C. Hubner, A. Al-Hakim, A. Bouchard, D. Faubert, D. Fermin, W. H. Dunham,M. Goudreault, Z. Y. Lin, B. G. Badillo, T. Pawson, D. Durocher, B. Coulombe,R. Aebersold, G. Superti-Furga, J. Colinge, A. J. Heck, H. Choi, M. Gstaiger,S. Mohammed, I. M. Cristea, K. L. Bennett, M. P. Washburn, B. Raught, R. M.Ewing, A. C. Gingras, and A. I. Nesvizhskii. The crapome: a contaminant repositoryfor affinity purification-mass spectrometry data. Nat Methods, 10(8):730–6, 2013.

[258] A. J. Tackett, J. A. DeGrasse, M. D. Sekedat, M. Oeffinger, M. P. Rout, and B. T.Chait. I-dirt, a general method for distinguishing between specific and nonspecificprotein interactions. J Proteome Res, 4(5):1752–6, 2005.

[259] M. Vermeulen, N. C. Hubner, and M. Mann. High confidence determination ofspecific protein-protein interactions using quantitative mass spectrometry. CurrOpin Biotechnol, 19(4):331–7, 2008.

[260] R. M. Kaake, X. Wang, and L. Huang. Profiling of protein interaction networksof protein complexes using affinity purification and quantitative mass spectrometry.Mol Cell Proteomics, 9(8):1650–65, 2010.

[261] M. Hilger and M. Mann. Triple silac to determine stimulus specific interactions inthe wnt pathway. J Proteome Res, 11(2):982–94, 2012.

[262] J. V. Olsen and M. Mann. Status of large-scale analysis of post-translational mod-ifications by mass spectrometry. Mol Cell Proteomics, 12(12):3444–52, 2013.

[263] J. Rush, A. Moritz, K. A. Lee, A. Guo, V. L. Goss, E. J. Spek, H. Zhang, X. M.Zha, R. D. Polakiewicz, and M. J. Comb. Immunoaffinity profiling of tyrosinephosphorylation in cancer cells. Nat Biotechnol, 23(1):94–101, 2005.

[264] G. Zhang and T. A. Neubert. Use of detergents to increase selectivity of immuno-precipitation of tyrosine phosphorylated peptides prior to identification by maldiquadrupole-tof ms. Proteomics, 6(2):571–8, 2006.

[265] J. Villen, S. A. Beausoleil, S. A. Gerber, and S. P. Gygi. Large-scale phosphorylationanalysis of mouse liver. Proc Natl Acad Sci U S A, 104(5):1488–93, 2007.

117

Bibliography

[266] A. R. Salomon, S. B. Ficarro, L. M. Brill, A. Brinker, Q. T. Phung, C. Ericson,K. Sauer, A. Brock, D. M. Horn, P. G. Schultz, and E. C. Peters. Profiling oftyrosine phosphorylation pathways in human cells using mass spectrometry. ProcNatl Acad Sci U S A, 100(2):443–8, 2003.

[267] M. O. Collins, L. Yu, and J. S. Choudhary. Analysis of protein phosphorylation ona proteome-scale. Proteomics, 7(16):2751–68, 2007.

[268] M. Fujimuro, H. Sawada, and H. Yokosawa. Production and characterization ofmonoclonal antibodies specific to multi-ubiquitin chains of polyubiquitinated pro-teins. FEBS Lett, 349(2):173–80, 1994.

[269] M. Fujimuro and H. Yokosawa. Production of antipolyubiquitin monoclonal anti-bodies and their use for characterization and isolation of polyubiquitinated proteins.Methods Enzymol, 399:75–86, 2005.

[270] M. Matsumoto, S. Hatakeyama, K. Oyamada, Y. Oda, T. Nishimura, and K. I.Nakayama. Large-scale analysis of the human ubiquitin-related proteome. Pro-teomics, 5(16):4145–51, 2005.

[271] P. Schwertman, K. Bezstarosti, C. Laffeber, W. Vermeulen, J. A. Demmers, andJ. A. Marteijn. An immunoaffinity purification method for the proteomic analysisof ubiquitinated protein complexes. Anal Biochem, 440(2):227–36, 2013.

[272] S. A. Wagner, P. Beli, B. T. Weinert, M. L. Nielsen, J. Cox, M. Mann, and C. Choud-hary. A proteome-wide, quantitative survey of in vivo ubiquitylation sites revealswidespread regulatory roles. Mol Cell Proteomics, 10(10):M111 013284, 2011.

[273] D. L. Swaney, P. Beltrao, L. Starita, A. Guo, J. Rush, S. Fields, N. J. Krogan,and J. Villen. Global analysis of phosphorylation and ubiquitylation cross-talk inprotein degradation. Nat Methods, 10(7):676–82, 2013.

[274] K. Newton, M. L. Matsumoto, I. E. Wertz, D. S. Kirkpatrick, J. R. Lill, J. Tan,D. Dugger, N. Gordon, S. S. Sidhu, F. A. Fellouse, L. Komuves, D. M. French, R. E.Ferrando, C. Lam, D. Compaan, C. Yu, I. Bosanac, S. G. Hymowitz, R. F. Kelley,and V. M. Dixit. Ubiquitin chain editing revealed by polyubiquitin linkage-specificantibodies. Cell, 134(4):668–78, 2008.

[275] M. L. Matsumoto, K. E. Wickliffe, K. C. Dong, C. Yu, I. Bosanac, D. Bustos,L. Phu, D. S. Kirkpatrick, S. G. Hymowitz, M. Rape, R. F. Kelley, and V. M. Dixit.K11-linked polyubiquitination in cell cycle control revealed by a k11 linkage-specificantibody. Mol Cell, 39(3):477–84, 2010.

[276] K. L. Guan, W. Yu, Y. Lin, Y. Xiong, and S. Zhao. Generation of acetyllysineantibodies and affinity enrichment of acetylated peptides. Nat Protoc, 5(9):1583–95,2010.

[277] Y. Y. Wang, J. Yao, P. Y. Yang, C. H. Deng, and H. Z. Fan. Immobilization ofantibodies on magnetic carbonaceous microspheres for selective enrichment of lysine-acetylated proteins and peptides. Chinese Journal of Chemistry, 30(10):2549–2555,2012.

[278] F. M. Boisvert, J. Cote, M. C. Boulanger, and S. Richard. A proteomic analysis ofarginine-methylated protein complexes. Mol Cell Proteomics, 2(12):1319–30, 2003.

118

Bibliography

[279] A. Guo, H. Gu, J. Zhou, D. Mulhern, Y. Wang, K. A. Lee, V. Yang, M. Aguiar,J. Kornhauser, X. Jia, J. Ren, S. A. Beausoleil, J. C. Silva, V. Vemulapalli, M. T.Bedford, and M. J. Comb. Immunoaffinity enrichment and mass spectrometry anal-ysis of protein methylation. Mol Cell Proteomics, 13(1):372–87, 2014.

[280] G. Nikov, V. Bhat, J. S. Wishnok, and S. R. Tannenbaum. Analysis of nitrated pro-teins by nitrotyrosine-specific affinity probes and mass spectrometry. Anal Biochem,320(2):214–22, 2003.

[281] X. Zhan and D. M. Desiderio. Nitroproteins from a human pituitary adenoma tissuediscovered with a nitrotyrosine affinity column and tandem mass spectrometry. AnalBiochem, 354(2):279–89, 2006.

[282] C. Wingren, P. James, and C. A. Borrebaeck. Strategy for surveying the proteomeusing affinity proteomics and mass spectrometry. Proteomics, 9(6):1511–7, 2009.

[283] O. Poetz, S. Hoeppe, M. F. Templin, D. Stoll, and T. O. Joos. Proteome widescreening using peptide affinity capture. Proteomics, 9(6):1518–23, 2009.

[284] S. Hoeppe, T. D. Schreiber, H. Planatscher, A. Zell, M. F. Templin, D. Stoll, T. O.Joos, and O. Poetz. Targeting peptide termini, a novel immunoaffinity approach toreduce complexity in mass spectrometric protein identification. Mol Cell Proteomics,10(2):M110 002857, 2011.

[285] N. Lion, F. Reymond, H. H. Girault, and J. S. Rossier. Why the move to microflu-idics for protein analysis? Curr Opin Biotechnol, 15(1):31–7, 2004.

[286] A. Rios, A. Escarpa, and B. Simonet. Miniaturization in analytical chemistry, 2009.

[287] S. C. Terry, J. H. Jerman, and J. B. Angell. A gas chromatographic air analyzerfabricated on a silicon wafer. Electron Devices, IEEE Transactions on, 26(12):1880–1886, 1979.

[288] M. J. Madou. Fundamentals of microfabrication : the science of miniaturization.CRC, Boca Raton, Fla., 2. edition, 2002.

[289] M. J. Madou and R. Cubicciotti. Scaling issues in chemical and biological sensors.Proceedings of the IEEE, 91(6):830–838, 2003.

[290] E. K. Sackmann, A. L. Fulton, and D. J. Beebe. The present and future role ofmicrofluidics in biomedical research. Nature, 507(7491):181–9, 2014.

[291] N. T. Nguyen and Z. Wu. Micromixers—a review. Journal of Micromechanics andMicroengineering, 15(2):R1, 2005.

[292] G. H. W. Sanders and A. Manz. Chip-based microsystems for genomic and proteomicanalysis. Trac-Trends in Analytical Chemistry, 19(6):364–378, 2000.

[293] X. Feng, W. Du, Q. Luo, and B. F. Liu. Microfluidic chip: next-generation platformfor systems biology. Anal Chim Acta, 650(1):83–97, 2009.

[294] V. Gubala, L. F. Harris, A. J. Ricco, M. X. Tan, and D. E. Williams. Point of carediagnostics: status and future. Anal Chem, 84(2):487–515, 2012.

[295] G. M. Whitesides. The origins and the future of microfluidics. Nature,442(7101):368–73, 2006.

119

Bibliography

[296] J. Gao, J. Xu, L. E. Locascio, and C. S. Lee. Integrated microfluidic system en-abling protein digestion, peptide separation, and protein identification. Anal Chem,73(11):2648–55, 2001.

[297] L. J. Jin, J. Ferrance, J. C. Sanders, and J. P. Landers. A microchip-based pro-teolytic digestion system driven by electroosmotic pumping. Lab Chip, 3(1):11–8,2003.

[298] Y. Li, R. Wojcik, and N. J. Dovichi. A replaceable microreactor for on-line proteindigestion in a two-dimensional capillary electrophoresis system with tandem massspectrometry detection. J Chromatogr A, 1218(15):2007–11, 2011.

[299] J. Ji, L. Nie, L. Qiao, Y. Li, L. Guo, B. Liu, P. Yang, and H. H. Girault. Proteolysisin microfluidic droplets: an approach to interface protein separation and peptidemass spectrometry. Lab Chip, 12(15):2625–9, 2012.

[300] J. Li, T. LeRiche, T. L. Tremblay, C. Wang, E. Bonneil, D. J. Harrison, andP. Thibault. Application of microfluidic devices to proteomics research: identifi-cation of trace-level protein digests and affinity capture of target peptides. Mol CellProteomics, 1(2):157–68, 2002.

[301] B. E. Slentz, N. A. Penner, and F. E. Regnier. Protein proteolysis and the multi-dimensional electrochromatographic separation of histidine-containing peptide frag-ments on a chip. J Chromatogr A, 984(1):97–107, 2003.

[302] B. Adler, T. Laurell, and S. Ekstrom. Optimizing nanovial outlet designs for im-proved solid-phase extraction in the integrated selective enrichment target–iset.Electrophoresis, 33(21):3143–50, 2012.

[303] S. Ekstrom, L. Wallman, J. Malm, C. Becker, H. Lilja, T. Laurell, and G. Marko-Varga. Integrated selective enrichment target–a microtechnology platform formatrix-assisted laser desorption/ionization-mass spectrometry applied on proteinbiomarkers in prostate diseases. Electrophoresis, 25(21-22):3769–77, 2004.

[304] S. Ekstrom, L. Wallman, D. Hok, G. Marko-Varga, and T. Laurell. Miniaturizedsolid-phase extraction and sample preparation for maldi ms using a microfabricatedintegrated selective enrichment target. J Proteome Res, 5(5):1071–81, 2006.

[305] H. Yan, A. Ahmad-Tajudin, M. Bengtsson, S. Xiao, T. Laurell, and S. Ekstrom.Noncovalent antibody immobilization on porous silicon combined with miniatur-ized solid-phase extraction (spe) for array based immunomaldi assays. Anal Chem,83(12):4942–8, 2011.

[306] S. Ekstrom, L. Wallman, G. Helldin, J. Nilsson, G. Marko-Varga, and T. Laurell.Polymeric integrated selective enrichment target (iset) for solid-phase-based samplepreparation in maldi-tof ms. J Mass Spectrom, 42(11):1445–52, 2007.

[307] A. Ahmad-Tajudin, B. Adler, S. Ekstrom, G. Marko-Varga, J. Malm, H. Lilja, andT. Laurell. Maldi-target integrated platform for affinity-captured protein digestion.Anal Chim Acta, 807:1–8, 2014.

[308] S. J. Lee, B. Adler, S. Ekstrom, M. Rezeli, A. Vegvari, J. W. Park, J. Malm, andT. Laurell. Aptamer/iset-ms: a new affinity-based maldi technique for improveddetection of biomarkers. Anal Chem, 86(15):7627–34, 2014.

120

Bibliography

[309] N. S. Berrow, K. Bussow, B. Coutard, J. Diprose, M. Ekberg, G. E. Folkers, N. Levy,V. Lieu, R. J. Owens, Y. Peleg, C. Pinaglia, S. Quevillon-Cheruel, L. Salim, C. Sche-ich, R. Vincentelli, and D. Busso. Recombinant protein expression and solubilityscreening in escherichia coli: a comparative study. Acta Crystallogr D Biol Crystal-logr, 62(Pt 10):1218–26, 2006.

[310] O. Brodsky and C. N. Cronin. Economical parallel protein expression screening andscale-up in escherichia coli. J Struct Funct Genomics, 7(2):101–8, 2006.

[311] R. Vincentelli, A. Cimino, A. Geerlof, A. Kubo, Y. Satou, and C. Cambillau. High-throughput protein expression screening and purification in escherichia coli. Meth-ods, 55(1):65–72, 2011.

[312] J. Panula-Perala, J. Siurkus, A. Vasala, R. Wilmanowski, M. G. Casteleijn, andP. Neubauer. Enzyme controlled glucose auto-delivery for high cell density cultiva-tions in microplates and shake flasks. Microb Cell Fact, 7:31, 2008.

[313] K. F. Wlaschin and W. S. Hu. Fedbatch culture and dynamic nutrient feeding. AdvBiochem Eng Biotechnol, 101:43–74, 2006.

[314] J. C. Marioni, C. E. Mason, S. M. Mane, M. Stephens, and Y. Gilad. Rna-seq: anassessment of technical reproducibility and comparison with gene expression arrays.Genome Res, 18(9):1509–17, 2008.

[315] Z. Wang, M. Gerstein, and M. Snyder. Rna-seq: a revolutionary tool for transcrip-tomics. Nat Rev Genet, 10(1):57–63, 2009.

[316] R. Lichtinghagen, P. B. Musholt, M. Lein, A. Romer, B. Rudolph, G. Kristiansen,S. Hauptmann, D. Schnorr, S. A. Loening, and K. Jung. Different mrna and proteinexpression of matrix metalloproteinases 2 and 9 and tissue inhibitor of metallopro-teinases 1 in benign and malignant prostate tissue. Eur Urol, 42(4):398–406, 2002.

[317] J. An, J. Shi, Q. He, K. Lui, Y. Liu, Y. Huang, and M. S. Sheikh. Chcm1/chchd6,novel mitochondrial protein linked to regulation of mitofilin and mitochondrialcristae morphology. J Biol Chem, 287(10):7411–26, 2012.

[318] J. Xie, M. F. Marusich, P. Souda, J. Whitelegge, and R. A. Capaldi. The mitochon-drial inner membrane protein mitofilin exists as a complex with sam50, metaxins1 and 2, coiled-coil-helix coiled-coil-helix domain-containing protein 3 and 6 anddnajc11. FEBS Lett, 581(18):3545–9, 2007.

121

Date post:	15-Jun-2020
Category:	Documents
Upload:	others
View:	5 times
Download:	0 times

High-throughput protein analysis using mass spectrometry ...757226/FULLTEXT01.pdf ·...

Documents