Mass spectrometry-based proteomics in biomedical research:...

Mass spectrometry-based proteomics

in biomedical research: emerging

technologies and future strategies

Geraldine M. Walsh1,2, Jason C. Rogalski1,2, Cordula Klockenbusch1

and Juergen Kast1,2,3,*

In recent years, the technology and methods widely available for massspectrometry (MS)-based proteomics have increased in power and potential,allowing the study of protein-level processes occurring in biological systems.Although these methods remain an active area of research, establishedtechniques are already helping answer biological questions. Here, this recentevolution of MS-based proteomics and its applications are reviewed, includingstandard methods for protein and peptide separation, biochemicalfractionation, quantitation, targeted MS approaches such as selected reactionmonitoring, data analysis and bioinformatics. Recent research in many of theseareas reveals that proteomics has moved beyond simply cataloguing proteinsin biological systems and is finally living up to its initial potential – as anessential tool to aid related disciplines, notably health research. From here,there is great potential for MS-based proteomics to move beyond basicresearch, into clinical research and diagnostics.

When, in 2000, the draft of the sequenced humangenome was introduced, many new avenues ofresearch for exploring human health becameavailable. One field that experienced anexplosion of interest was proteomics, the studyof the protein complement of a cell undercertain conditions. Although these newlyuncovered genome sequences revealed whichprotein sequences could be expressed, splicing,post-translational modifications (PTMs), tertiary

structure, enzymatic activity, formation ofcomplexes and ligand interactions combine toproduce a much richer protein environmentthan what is simply coded for, and it is theseintricate and complex processes that dictate howbiological functions occur. Proteomic research isthe attempt to understand all that is occurring inthis complex environment, with the aim ofelucidating protein-level processes involved inbiological activity.

1The Biomedical Research Centre, University of British Columbia, Vancouver, BC, Canada.2The Centre for Blood Research, University of British Columbia, Vancouver, BC, Canada.3Department of Chemistry, University of British Columbia, Vancouver, BC, Canada.

*Corresponding author: Juergen Kast, The Biomedical Research Centre, 2222 Health Sciences Mall,Vancouver, BC, Canada V6T 1Z3. E-mail: [email protected]

expert reviewshttp://www.expertreviews.org/ in molecular medicine

1Accession information: doi:10.1017/S1462399410001614; Vol. 12; e30; September 2010

©Cambridge University Press 2010

Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

The first tentative steps towards massspectrometry (MS)-based proteomics started inthe late 1980s, well before the human genomewas sequenced, when the development of soft-ionisation techniques such as electrosprayionisation (ESI) (Ref. 1) and matrix-assisted laserdesorption/ionisation (MALDI) (Ref. 2) allowedMS analysis of intact biological macromoleculesfor the first time. These technologies, togetherwith the fact that peptides produced throughthe digestion of proteins with highly specificproteases are characteristic of their parentprotein, permitted protein identification bycomparison of MS data with known sequences,in silico. Progress in this field generated a greatdeal of fervour, and researchers began todevelop new techniques, as well as incorporateestablished techniques, to aid proteome analysisby MS. The years that followed saw gargantuanleaps in the capabilities of MS and relatedtechnologies.So why all the excitement? It is mainly due to

the ability of MS to obtain specific and sensitiveinformation about a complex sample quickly,over a wide dynamic range. Given that thegenome of a given species codes for manythousands of protein products [∼20 500 forhumans (National Human Genome ResearchInstitute, http://www.genome.gov)], whichcover many orders of magnitude in abundance(ten in the case of plasma) (Ref. 3), two-dimensional (2D) gel electrophoresis wasinitially the only technology capable of sensitiveand reproducible visualisation of the proteome.MS, combined with a host of affiliatedtechnologies, provided the first opportunity togo beyond gel-based visualisation, enablingdiscovery and identification of the componentsof a proteome on a large scale, to a depth thatimmunoprecipitations and 2D gels could notprovide. These proteome-wide discoveryexperiments were the basis of the initial thrustin MS-based proteomics, inspiring a rapid rateof creation and improvement of new techniquesand instrumentation in an attempt to digdeeper into the proteome, with more certainty,less sample and less time. This focus oninstrumentation brought together differentcombinations of mass analyser and ion source,and fostered the utilisation of the strengths ofdifferent mass analysers in hybrid instruments(Ref. 4). Research and development continueto produce and improve mass spectrometers

to this day (Ref. 5). Although these techniqueand technology improvements have resultedin the greatly increased utility and robustnessof MS-based proteomics, what does thismean for tangible benefits to human healthresearch? Essentially, it means that proteomicshas moved beyond simply asking the ‘what’ ofa biological question, and now can routinelyand robustly study the ‘when, where, how andhow much’. Current popular techniques andexperiment types employed in MS-basedproteomics that are now being utilised inbiomedical research are discussed in this review(Fig. 1).

Protein and peptide separation techniquesThe field of MS-based proteomics can becategorised into two broad approaches. Theincreasingly popular ‘top-down’ proteomicapproach focuses on the analysis of intactproteins, whereas the more widely used‘bottom-up’ proteomic approach focuses on theanalysis of peptides following proteolyticdigestion of proteins, and is the main topic ofthis review (Ref. 6). Because ‘bottom-up’proteomic approaches require digestion ofproteins into peptides prior to their analysis byMS, preanalytical sample processing plays animportant role and should be carefullyconsidered when designing and conductingthese types of experiments. By far the mostpopular method to prepare a proteomic sampleis enzymatic digestion using trypsin, which isvery well suited to downstream analysis by themost common MS and tandem MS (MS/MS)techniques. However, information regardingPTMs or protein isoforms could be missed, andit is often worth considering other proteolyticenzymes or applying a panel of enzymes(Ref. 7). The digestion of proteins into peptidesprior to MS analysis greatly increases thecomplexity of samples, and the separation ofthese complex samples into manageable,reproducible fractions is an issue thatproteomics has battled with since its inception.

Owing to several factors, including competitiveionisation of coeluting species, dynamic rangelimitations (the ability to analyse a weak signalin the presence of a strong signal), duty cycleconstraints (how many things can be analysedper unit of time) and resolving power, it isgenerally known that the greater the separationbefore MS sequencing, the better the results




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

(Ref. 8). Powerful separations are therefore anecessity for in-depth proteome analysis andthere has been much research into separationtechnologies that are compatible with

proteomic workflows. The art is now such thatwhat was cutting-edge experimental work fiveyears ago is now routinely performed inlaboratories all over the world. For example,

SILACN15

ICATiTRAQICPL

iTRAQICPLAQUA

Proteolyticlabelling

ICATiTRAQICPL

Globalanalysis

Subcellularfractionation

Proteinenrichment

Enrichmentof PTMs

Peptide/protein identification

(e.g. X!tandem, SEQUEST)

Bioinformatic analysis

Quantitative analysis

Quantitativeanalysis

Unbiased MS(e.g. LC-MS/MS)

Biased MSSRM

Protein elutionLysis

Optional: protein separation

Optional: peptide separation

Protein digestion (e.g. trypsin)

Cell

Labelling methods

Lysis

Workflow of typical MS-based proteomic experimentsExpert Reviews in Molecular Medicine © Cambridge University Press 2010

Figure 1. Workflow of typical MS-based proteomic experiments. (See next page for legend.)




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

separation of samples in at least twodimensions inthe liquid phase, a strategy known asmultidimensional liquid chromatography(MDLC), has many different permutations andis a large field of research on its own, leading tomany of its own reviews (Refs 9, 10).

Protein separationAlthough the final dimension of separation forMS-based proteomics is generally a reversed-phase separation at the peptide level, upstreamseparations can use a number of differentproperties for fractionation of complex samplesat the peptide or protein level. Many optionsexist for protein-level separation, each of whichcan utilise a different physical property ofproteins to obtain varying but complementaryfirst-dimension separations. Cation exchange(Ref. 11) and anion exchange (Ref. 12), forexample, offer good separation orthogonality infractionating proteomic samples before liquidchromatography (LC)-MS/MS. Two differenttechnologies, however, have emerged recentlyand joined the mainstream: chromatofocusing(CF) (Refs 13, 14, 15), a variant of ion-exchangechromatography that separates proteins basedon pH; and isoelectric focusing (IEF) (Ref. 16),which separates proteins based on theirisoelectric point (pI), as is done in the firstdimension of 2D gel electrophoresis. Thesemethods are easily automatable and providepowerful protein-level separation and usefulinformation of the physical properties of theproteins, while keeping the analytes soluble and

compatible with proteomic experiments, and aretherefore widely used (Ref. 17).

Another commonly used method for protein-level sample fractionation prior to MS isthe ‘GeLC’ approach. It harnesses the well-established ability and available equipment forrunning gels, by separating a complex sample bymolecular weight at the protein level in a single1D SDS-PAGE gel lane, and using that lane asthe first dimension in a multidimensionalseparation. After staining, the entire lane isexcised, cut into bands and each band is treatedas a fraction of the same sample. After theproteins in these bands are enzymaticallydigested, each band’s peptide mixture can beanalysed on an LC-MS/MS instrument and theresults combined. The benefits of gel-basedprotein-level first-dimension separation arethreefold: gels are often a good way of makingbiological samples compatible with MS analysis(e.g. by removal of detergents), methoddevelopment is not needed as SDS-PAGE is awell-established technique, and the number ofidentifications obtained per experiment iscurrently second to none. In fact, of theaforementioned protein-level separationtechniques, the GeLC approach has been foundto provide the highest number of confidentprotein or peptide identifications, although thealternative approach of immobilised pH gradient(IPG)-based IEF has the benefit of slightly highersample recovery over the GeLC separation(Ref. 17). A recently developed fractionationmethod termed GELFrEE (gel-eluted liquid

Figure 1. Workflow of typical MS-based proteomic experiments. (See previous page for figure.) Whole-celllysates can be used for a global proteome analysis, or more in-depth analysis and additional spatial informationcan be obtained using subcellular fractionation. Alternatively, cells can be lysed and proteins or post-translation modifications (PTMs) of interest can be isolated by affinity enrichment methods. All methodsproduce protein mixtures, which can be separated further by exploiting various protein properties such asmolecular weight or isoelectric points, and are digested in the next step. Separating the generated peptidesis recommended and leads to deeper resolution. Peptides are then analysed by MS (e.g. LC-MS/MS) andin unbiased discovery experiments peptides and the corresponding proteins are identified using database-matching search algorithms, followed by quantitative and bioinformatic evaluation of the data. Alternatively,targeted MS for specific peptides and proteins can be performed using SRM. Quantitative information canbe obtained either by label-free methods or by applying a differential isotopic labelling method at one of thestages indicated on the right: metabolic labels such as SILAC and N15 can be introduced at cell level,whereas chemical labelling methods such as iTRAQ, ICPL or ICAT are utilised either at protein or at peptidelevel. Isotopic labelling can also be introduced during proteolysis, and synthetic standard isotopic peptidescan be added to the peptide mixture (AQUA). Abbreviations: AQUA, absolute quantitation; ICAT, isotope-coded affinity tags; ICPL, isotope-coded protein labelling; iTRAQ, isobaric tags for relative and absolutequantitation; LC-MS/MS, liquid chromatography tandem mass spectrometry; MS, mass spectrometry; N15,15N isotope; PTM, post-translational modification; SILAC, stable isotope labelling of amino acids in cellculture; SRM, selected reaction monitoring.




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

fraction entrapment electrophoresis) also separatesproteins based on their size using a gel column;however, in contrast to the GeLC method, theproteins are eluted and collected in the liquidphase (Ref. 18). The researcher needs to be awareof these performance differences in first-dimension separations when deciding on thepriorities for a given experiment.

Peptide separationThe development of powerful techniques andchemistries for separation at the peptide levelhas led many MS-based workflows to forgo theaforementioned protein-level separationsentirely. Many options exist for the first-dimension separation, which have varyingdegrees of orthogonality: for example, reversed-phase chromatography, which resolves issueswith solvent compatibility; or strong ionexchange, different forms of which can providegood sample complementarity. One of the mostcommon first-dimension separations for large-scale proteomic experiments is peptide-levelstrong cation exchange (SCX) chromatography.There is no ‘best’ answer for separation; despiteeach of the techniques being optimal forparticular sample types, they will all providecomplementary results.First-dimension separations can be performed

off-line with a fraction collector, although whenthey are performed in-line with a reversed-phasecolumn as the second dimension prior to MS, ina workflow called multidimensional proteininformation technology (MudPIT) (Refs 19, 20),the experiment is capable of significantproteome coverage (approximately 60%) in one– albeit very long – experiment (Ref. 21). Thistype of workflow is described thoroughly in apublished protocol (Ref. 22). Reproducing anexperiment of this type also provides 60%of the proteome, with a large number of thepeptides sequenced being species that were notsequenced in the first experiment. It is estimatedthat it would take five MudPIT experimentsperformed in this way to achieve near-completesequence coverage, a phenomenon attributed toMS/MS peptide-sampling rates. As with anyMS-based analysis of complex samples, thelimitations of this method are time andinstrument duty cycle – issues that should beconsidered when designing experiments andchoosing which proteomic approach to use.These limitations can be attenuated by

conducting biological and technical repeats andmaximising separation and fractionation prior toMS analysis. Also, the development of dynamicexclusion lists to avoid run-to-run resequencingof peptides has recently increased the number ofextracellular proteins identified in repeatanalyses of the human embryonic stem cellsecretome by an order of magnitude (Ref. 21).Expanding the MudPIT workflow to include athird dimension of separation has also beenshown to work well (Ref. 23), and a study of theproteome of the serum of patients with sepsisutilised immunodepletion of abundant serumproteins followed by a 3D peptide-levelseparation, allowing the identification of low-abundance serum proteins while identifying tenpotential serum biomarkers for sepsis (Ref. 24).Although this type of technique shifts thelimitation of the method towards separation timeand away from the duty cycle of the instrument,the deployment of fast, ultrahigh-pressure liquidchromatography (UPLC) (Ref. 25; http://www.waters.com/waters/nav.htm?locale=en_US&cid=10136122) inmany laboratories is nowproving thatthese methods are more powerful than ever.

Biochemical fractionation methodsThe protein- or peptide-separation techniquesdescribed above allow in-depth analysis of acomplex sample. However, biochemicalfractionation procedures, which add anadditional dimension of separation, can lead toeven deeper resolution as the separationmethods described previously can be performedon a less complex sample. This can be especiallyimportant in highly complex samples, such ashuman plasma or serum, which have a highdynamic range spanning at least ten orders ofmagnitude. These samples contain a smallnumber of highly abundant proteins, whosesignals can dominate MS-based analysis.Depletion of these proteins can be highlyadvantageous in allowing access to lowerabundance species, including potential diseasebiomarkers, and there are many tried and testeddepletion strategies available (Refs 26, 27). Forcellular studies, spatial information (e.g. whichproteins are found in which organelles or whichproteins interact with each other) can beextremely important for understanding acomplex system, and can be obtained byapplying either subcellular fractionation oraffinity enrichment techniques.




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

Subcellular fractionationFor subcellular analysis, every classicalbiochemical fractionation procedure, whether itbe membrane enrichment, nucleus precipitationor mitochondria preparation, can be used as thefirst enrichment step, followed by protein/peptide separation and MS analysis. Forexample, plasma membrane lipid rafts wereenriched to follow the effects of DMSO-induceddifferentiation of HL-60 cells into neutrophils byLC-MS/MS, and out of 147 identified proteins,25 were found to be upregulated and 49 weredownregulated (Ref. 28). In a different study,membrane fractionation and the hydrazidemethod were used to isolate 25 glycoproteinsfrom breast cancer cell lines, which areconsidered putative cancer biomarkers (Ref. 29).

ImmunoprecipitationAffinity enrichment of a protein and its interactionpartners decreases the complexity of a sampledramatically and provides information about thecomposition of the interaction network. Classicalcoimmunoprecipitation, a long-establishedmethod to isolate proteins, is the first stepperformed for this approach, applying eitherantibodies against endogenous proteins orimmunoaffinity tags. Precipitated proteins areisolated afterwards and analysed as describedearlier. In contrast to an immunoblot analysis,which requires a hypothesis about interactionpartners and focuses on the identification of oneprotein, MS is an unbiased detection methodand allows the discovery of several bindingpartners at once, including unexpected ones.However, MS is a very sensitive method andtherefore stringent wash conditions, severalcontrols and careful interpretation of the resultsare required to obtain correct information fromthis type of experiment (Ref. 30). For example,the interaction network of MYC was studiedusing the tandem affinity purification (TAP)approach, which allows stringent wash steps andthereby reduces false-positive identifications; 221putative interaction partners were identified, ofwhich only 17 were known before (Ref. 31).Another approach was applied for the study ofintegrin-linked kinase (ILK), where aquantitative MS approach (see below) was usedto distinguish between proteins binding to thebait protein or to the tag itself and allowed theidentification of several novel ILK-interactingproteins (e.g. α-tubulin) (Ref. 32). Two

complementary affinity purification methodswere used to identify over 40 kinases binding todasatinib, an inhibitor with putative antitumourproperties. In a second step, phosphorylatedproteins were purified from cancer cells; 23candidates identified in both pull-downs wereanalysed in more detail regarding theirsusceptibility to the inhibitor and several ofthese kinases were found to be inhibited bydasatinib (Ref. 33).

Phosphorylation-enrichment strategiesEnriching for PTMs also simplifies a complexsample, and studying the correspondingproteins can provide detailed information aboutsignalling processes. Furthermore, even thoughPTMs can be identified by MS, the lowstoichiometry of these modifications can lead tothem being missed during analysis, a problemthat can be overcome by specific affinityenrichment. One of the major modificationstaking place during signal transduction isphosphorylation, the study of which – termedphosphoproteomics – has also pioneeredtechnology development. Prior to MS analysis,phosphoproteins can be isolated byimmunoprecipitations (e.g. by applyingantibodies against phosphotyrosines) orphosphopeptides (containing modified serines,tyrosines and threonines) can be enriched bymetal-supported chromatographies such as IMAC(immobilised metal affinity chromatography)(Ref. 34) or MOC (metal oxide chromatography)(Ref. 35) mostly utilising titanium dioxide. Thesemethods have been established and optimisedin recent years, and phosphoproteomics incombination with quantitative approaches suchas stable isotope labelling of amino acids in cellculture (SILAC) or isobaric tags for relative andabsolute quantitation (iTRAQ) (see the nextsection) now has the power to study time-dependent activation cascades (Ref. 36).

The epidermal growth factor (EGF) signallingpathway has been studied in detail by severalgroups using slightly different MS approachesand can be seen as a model system for theoptimisation of phosphoproteomics (Ref. 37).Mann and colleagues have applied manymethods, including the application ofantiphosphotyrosine antibodies (Ref. 38), the useof titanium dioxide to enrich phosphopeptides(Ref. 39) and the combination of bothenrichment approaches, on the way to




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

developing a method termed qPACE, whichallows the study of very early signalling events(Ref. 40). For quantitation, they utilised SILAC.By contrast, White and colleagues used iTRAQ tostudy EGF receptor signalling using acombinational enrichment approach applyingantiphosphotyrosine antibodies and IMAC(Ref. 41), and extended their methodology withselected reaction monitoring (SRM) experiments(which are explained in detail later in thisreview), allowing a much higher reproducibility(Ref. 42). Phosphoproteomics is now used tostudy other and unknown signalling pathways,as for example the SYK signalling cascade,which was originally described only inhaematopoietic cells but has been investigatednow in human cancer cells to shed more lighton the role of this kinase in cancer formation(Ref. 43).With the help of the enrichment techniques

described here, MS-based proteomics canachieve high spatial and functional resolution.However, as mentioned throughout this section,a quantitative dimension is also frequentlynecessary to answer many of the questionscurrently asked by researchers.

Quantitative approachesThe topic of MS-based quantitation exploded inthe mid-2000s (Ref. 44), and the current state ofthe art is reviewed thoroughly and engaginglyelsewhere (Ref. 45). Essentially, despitequantitative proteomics still being an active areaof research on its own, it is now also available tohuman health researchers who are interested instudying drug effects, biomarkers of disease andthe pathways involved in disease processes.

Isotopic labelling techniquesMass spectrometers are not inherentlyquantitative. Differences in ionisation,transmission and detection efficiency dictate thatthe intensity of a signal from a particularmolecule is a relative measure of its abundance,but not an absolute measure. For this reason, allquantitative proteomics, even ‘absolute’quantitation is relative – relative to an internalstandard’ (Ref. 45). MS-based proteomicquantitation was therefore not thrust into themainstream until 1999, when isotope-codedaffinity tags (ICATs) were introduced (Ref. 46).These tags were the first widely availablemethod to quantify the relative concentrations of

peptides or proteins in a sample, by way of anisotope-coded chemical modifier. Briefly, each oftwo samples is treated with either one of a‘light’ or ‘heavy’ chemical reagent that bindsspecifically to cysteine residues. The light andheavy tags are chemically identical, except forisotopic differences. The two samples are thenmixed and digested, and the tagged peptidesare enriched using avidin or streptavidinchromatography against the biotin moietyembedded in the tag. On performing MSanalysis on these enriched samples, thechemically identical species from the twosamples will coelute from a column and ionisewith identical efficiency; however, the peptidethat is modified with the ‘light’ form of thereagent will appear at a known lower mass inthe spectrum than the ‘heavy’ tagged equivalentpeptide from the other sample. One can thendirectly compare the peak areas of the twochemically identical coeluting peptides andthereby obtain a relative measure of theirabundance. Relative quantitation, performedthrough isotope-coding methods similar to this,is the best way to obtain information aboutquantitative differences in protein expression,especially from the complex samples usual inproteomics (Fig. 2).

One issue with the ICAT method describedabove, however, is its dependence on themodification of cysteine residues, which accountfor only 1.42% of the amino acids in a sample(Ref. 47). Many peptides, and even wholeproteins, do not contain a cysteine, and aretherefore unquantifiable by means of ICAT. Thisproblem was resolved in 2004 with theintroduction of the isobaric tags for relative andabsolute quantitation (iTRAQ) label (Ref. 48)(Applied Biosystems; http://www.appliedbiosystems.com). With this tag, initially four, andnow up to eight, samples can be comparedtogether, using labels with identical mass shifts.This is achieved through the differentialplacement of the stable isotopes onto ‘balance’and ‘reporter’ pieces of the tag, which areseparated by a labile bond. Each of the differenttag ‘flavours’ adds the same overall mass to apeptide, bound through the balance group tothe primary amines on lysine side chains andthe N-termini of peptides. On mixing of thesamples, unlike ICAT-labelled samples, taggedpeptides will appear as one signal in a normalMS scan. Only upon fragmentation does the




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

Starting culture

Culture in mediacontaining

nonlabelled (light)amino acids

Culture in mediacontaining

isotopically labelled(heavy) amino acids

Leave as control Expose to stimuli

Mix samples 1:1Analyse using protein/peptide separation

and mass spectrometry

678.3675.3

675.8678.8

676.3 679.3Inte

nsity

m/z

SILAC

MS scan

a

Cultured cells or any biological material

Expose to different stimuliLeave ascontrol

Mix samples evenlyAnalyse using protein/peptide separation

and mass spectrometry

m/z

Inte

nsity

461.7

462.2

462.7

Fragment and collect low-mass marker ions

MS scan of all combined labelledforms of peptide

116115

114 117

Inte

nsity

m/z

iTRAQ or TMT

Lyse cells, digest and label each samplewith a specific iTRAQ tag

114 115 116 117

MS scan

MS/MS scan

b

MS scan of light and heavy SILAC-labelled peptides: extracted ion chromatograms

give quantitative information

Intensities of marker ions givequantitative information

Workflows for implementation of stable isotope labelling via metabolic (SILAC) and chemical (iTRAQ or TMT) methodsExpert Reviews in Molecular Medicine © Cambridge University Press 2010

Figure 2. Workflows for implementation of stable isotope labelling via metabolic (SILAC) and chemical(iTRAQ or TMT) methods. (See next page for legend.)




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

labile bond holding the iTRAQ modificationtogether dissociate, forming an intense markerion from the reporter group, each of which willbe specific for one of the samples to becompared. This behaviour allows for highspecificity (marker ion intensity can only befrom the peptide currently being fragmented)and high sensitivity because the isobaric natureof the tag dictates that the intensities of thesignals from all samples are additive for initialdetection in the MS scan, and subsequentsequencing. This technique has also beensuccessfully used in the tandem mass tag (TMT)strategy from Thermo Scientific (http://www.piercenet.com).These types of chemical labelling strategies are a

good choicewhenprobing the proteomeof humancells or tissues that cannot or should not becultured. Unfortunately, as the combination ofthe samples to be compared occurs late in thisworkflow, there is a chance of systematic errorsduring sample handling (Fig. 2). Metaboliclabelling techniques, in which cells are grown inisotopically labelled media and compared withthose grown in normal media, have been shownto be the most accurate proteomic quantitationmethod mainly due to the ability to combinesamples very early in the procedure (e.g.immediately after lysis), thus minimising errorsinvolved with differential sample handling inthe subsequent isolation and purification steps(Ref. 38). Unfortunately, SILAC (Ref. 49) canonly be used on cells cultured in vitro, ascontrolling the isotopes available to a biologicalsample can be problematic. Although SILACquantitation on a whole mouse has beensuccessfully performed (Ref. 50), this type ofexperiment is prohibitively expensive and timeconsuming for most research projects andspecies types. Recently, however, a techniquethat uses a combination of five SILAC-labelled

cell lines, pooled together, has been introducedas a ‘physical proteome database’, which wasthen compared with a nonlabelled carcinomatissue sample, allowing the SILAC quantitationof uncultured human tissue cells (Ref. 51). Thisnew technique allowed quantitative comparisonof lobular and ductal tumours, revealingsignificant differences, with very low coefficientsof variance, in the expression of focal adhesionand glycolytic proteins, in a clinically relevanthuman tissue sample.

Label-free quantitationGiven that the intensity of the signal in a massspectrometer is innately a proxy of theabundance of the species in the sample, label-free quantitation approaches have recentlyentered the mainstream because of theirapparent ease, simplicity and cost savings. Oneof the numerous label-free quantitationapproaches is ‘spectral counting’, in which theMS/MS spectra collected for a given species arecounted and compared with those collected forthe same species in a different sample. Thistechnique uses the assumption that unbiased,intensity-based precursor ion selection leads tointense ions being selected for sequencing morefrequently. The number of MS/MS spectracollected for a given analyte would therefore bea proxy of its intensity, and therefore itsabundance. Like all label-free quantitationmethods, systematic errors in the analysis, suchas signal suppression, detector saturation anddifferential sample loading, occur. These errorsneed to be minimised and accounted for byperforming many replicates, normalising thedata, and statistical validation (Refs 45, 52).Although this approach has been shown to beadequate for quantitation of high-abundancecomponents of a mixture (Ref. 53), isotopiclabelling techniques, which correct for these

Figure 2. Workflows for implementation of stable isotope labelling via metabolic (SILAC) and chemical(iTRAQ or TMT) methods. (See previous page for figure.) (a) SILAC incorporates isotopes early in the samplepreparation procedure, maximising accuracy and reproducibility, and is generally used only for quantitation ofsamples from cells that can be cultured. Isotope incorporation into the amino acids themselves meanspeptides to be compared have different masses; therefore quantification occurs from the MS scan. (b)Chemical isotope-coded tags are applied later in the workflow, but can be applied to any biologicallyderived sample. Isobaric chemical tags (iTRAQ, TMT) add equivalent masses to the peptides in the sample,but produce specific marker ions upon fragmentation, allowing quantification from the MS/MS scan.Abbreviations: iTRAQ, isobaric tags for relative and absolute quantitation; MS, mass spectrometry; MS/MS,tandem mass spectrometry; SILAC, stable isotope labelling of amino acids in cell culture; TMT, tandemmass tag.




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

systematic errors and also allow for samplemultiplexing, are still the method of choice forquantitative proteomics. The commonly usedquantitative proteomic methods, along withtheir strengths and weaknesses, are shown inTable 1.

Targeted analysis: selected reactionmonitoring

Unlike global proteomic techniques, which operatebased on intensity-dependent fragmentation ofprecursor ions and are biased towards moreabundant proteins, SRM (also known as multiplereaction monitoring, MRM) targets predeterminedprecursor ions for fragmentation (Fig. 3). Thisallows peptides from a particular protein ofinterest to be monitored, giving access to lowerabundance species even in complex mixtures suchas plasma and serum (Ref. 54). For years, SRMhas been used effectively to detect and quantifydrugs and drug metabolites in pharmaceuticalresearch (Ref. 55). Today, it is rapidly becomingthe method of choice in many fields owing to itsconsistency, accuracy and sensitivity. This includesbasic proteome research, where advances indevelopment and validation of these assays, aswell as novel software and data repositories, areincreasing the potential of the SRM approach inwhole-proteome analysis. In clinical research, itspotential as a biomarker verification tool isthought by some to rival the standard ELISAmethod and there is huge potential for theapplication of this approach in clinical diagnostics.The amount of sample required for SRM analysisis small, and its sensitivity is high (attomole level)(Ref. 56); therefore, it is suitable for the analysis ofsamples containing small amounts of material,such as neonatal screening and therapeutic drugmonitoring, meeting the throughput requirementsof clinicians (Refs 55, 57).

Experimental designTargeting the most appropriate peptides andfragment ions for the protein of interest is key toa successful SRM experiment; therefore someprior experimental knowledge is required. Thistranslates into knowing the mass to charge ratio(m/z) of an abundant, consistently produced (or‘proteotypic’) peptide (Ref. 58) as well as the m/z of one of its fragment ions that is generatedwith high intensity. These ‘transitions’ (specificprecursor–fragment ion pairs) allow targetedanalysis of a particular peptide in a complex

mixture. There are many guidelines that can aidin the selection of appropriate transitions, basedon prior experimentation, physicochemicalparameters and in silico predictions (Ref. 59).These take into account factors that are toonumerous to describe in detail here, but areoutlined in several recent reviews (Refs 54, 60).There are multiple software options available toaid the design and optimisation of thesetransitions (Ref. 61). Many are reviewedelsewhere (Ref. 62), with a selection listed inTable 2. Although this design stage can takeconsiderable time, once transitions areestablished, they can be used indefinitely forexperiments studying the protein of interest.

SRM and quantitationThe SRM approach can be used to quantitateproteins. Relative quantitation can be conductedsimply by comparing the absolute peak area ofthe individual samples (label-free quantitation),although it is difficult to obtain precisemeasurements because of differences inionisation efficiency, analyte composition andchromatography. SRM experiments can also becombined with many of the standard isotopelabels used in quantitative proteomicexperiments, including ICAT, SILAC, ICPL andiTRAQ. Additionally, several methods that aidgreatly in speeding up the assay developmentaspect of SRM have emerged, includingdatabases such as MRMAtlas (Ref. 63) and amethod of crude synthetic peptide libraryproduction, which allow the rapid generation ofvalidated SRM assays for whole proteomes(Ref. 56). These approaches have been pioneeredusing the yeast proteome, but the developmentof databases and resources such as this forclinically relevant tissues could help thrust SRM-based quantitation firmly into the clinical arena.

Applications of SRMThe advantages of SRMexperiments have led to analmost exponential increase in the number ofstudies using this approach in recent years(Ref. 62), and SRM has now been applied tomany diverse biological questions, from thequantitation of the biomarker C-reactive protein(CRP) in the serum of patients with rheumatoidarthritis (Ref. 64) to the absolute quantitation ofthe human liver alcohol dehydrogenaseADH1C1 isoenzyme (Ref. 65) and pyruvatekinase M2 (PKM2), a potential endometrial




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

Table

1.MSqua

ntitationtech

nique

s

Tech

nique

Applic

ation

leve

lLa

bellin

gmetho

dPros

Cons

Refs

SILAC

Protein

Metab

olic

Highqua

ntita

tiveac

curacy

Man

yco

mparison

sper

experim

ent

Works

wellfor

celllines

Onlyforin

vitrocu

lture

Proline–

arginine

conv

ersion

canca

usequa

ntita

tionerrors

49,1

18,1

19,1

20,

121

N15

Protein

Metab

olic

Goo

dforau

totrop

hicsp

ecies

Exa

ctmas

ssh

iftof

pep

tides

isun

predictable

Onlybinaryco

mparison

122,

123,

124,

125,

126

ICAT

Protein

Che

mical

derivatisation

Purifica

tionpos

sible

bec

ause

ofrobus

tbiotin

tag;

reduc

tionof

sample

complexity

Targetson

lycy

steine

-co

ntaining

pep

tides

46,1

27,1

28,1

29

iTRAQ

orTM

TProtein

orpep

tide

Che

mical

derivatisation

Man

yco

mparison

sper

experim

ent

Additive

intens

ityin

MSqua

ntita

tion

multip

lexe

din

MS/M

Ssc

anEve

rytryp

ticpep

tideca

nintheo

rybe

qua

ntified

Com

pressionof

expression

ratio

sVa

riability

inlabellingeffic

ienc

yCos

tan

ddifficulty

48,1

30http://

www.

piercen

et.com

ICPL

Protein

orpep

tide

Che

mical

derivatisation

Man

yco

mparison

sper

experim

ent

Qua

ntifies

anylysine

-con

taining

pep

tide

Varia

bility

inlabellingeffic

ienc

yCos

tan

ddifficulty

131,

132

Proteolytic

labelling

Pep

tide

Duringdiges

tion

Eas

eCos

tDifferen

tialrates

oflabel

inco

rporation

Bac

kex

chan

geof

label

133,

134

Label-free

Pep

tide

Non

eEas

eCos

tNoex

trasa

mple

hand

ling

Onlyac

curate

forab

undan

tsp

ecies

Nosa

mple

multip

lexing

Differen

tials

igna

lsup

pression

effects

135,

136,

137,

138,

139,

140,

141,

142

AQUA

Pep

tide

Spikeof

synthe

sise

dpep

tide

Acc

urate,

abso

lute

qua

ntita

tion

Correctsfordifferen

cesin

analysis

Sen

sitiv

e(with

SRM)

Exp

ensive

Nee

dforsynthe

sise

dpep

tide

forea

chqua

ntified

pep

tide

Can

notbeus

edfordisco

very

143,

144

Abbreviations

:AQUA,a

bso

lute

qua

ntita

tion;

ICAT,iso

tope-co

ded

affin

itytags

;ICPL,

isotop

e-co

ded

protein

labelling;

iTRAQ,iso

baric

tags

forrelativ

ean

dab

solute

qua

ntita

tion;

MS/M

S,tan

dem

mas

ssp

ectrom

etry;N

15,1

5N

isotop

e;SILAC,s

table

isotop

elabellingof

aminoac

idsin

cellcu

lture;S

RM,s

elec

ted

reac

tionmon

itorin

g;TM

T,tand

emmas

stag.




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

cancer marker (Ref. 66). The approach has beenshown to be extremely powerful, both inconfirmation of potential biomarkers and indiscovery of novel biomarkers, as seen in arecent study that integrated high-throughputand SRM-based approaches to explore breastcancer in a mouse model (Ref. 67). SRM is also apromising tool to study splice variants that areputative biomarkers for cancer cells. Forexample, using the breast cancer mouse modelmentioned above, 216 of 608 splice variantswere found only in tumour cells and SRM isideally suited to study these (Ref. 68). UsingSRM coupled with stable isotope dilution MS(SIDMRM-MS), quantitative, multiplexed assayswere developed for the analysis of six proteinsclinically relevant to cardiac injury (Ref. 69).These widely applicable assays were conductedusing plasma samples, with proteins of interestspanning four orders of magnitude. These werequantitated using a few signature peptides fromeach target protein, with limits of quantitationranging between 2 and 15 ng/ml. Similarly, lowng/ml sensitivity quantitation of the prostate-specific antigen biomarker was achieved using LC-MS/MS SRM, from 100 μl of serum, demonstratinggood correlation with ELISA measures (Ref. 70).Even greater sensitivity was achieved in a recentstudy that captured and enriched peptides withantipeptide antibodies and then used SRM-basedanalysis to quantitate aberrant GlcNAcylated tissueinhibitor of metalloproteinase 1 (TIMP1), a proteinimplicated in colorectal cancer (Ref. 71). Followingenrichment and digestion of glycoproteins frompatients’ serum, SISCAPA (stable isotopestandards and capture by antipeptide antibodies)(Ref. 72) and SRM-MS permitted highlysensitive quantitation of TIMP1 at attomolarconcentrations. Automation and multiplexing ofthis approach shows great potential for analysinglarge numbers of biomarkers with sufficientsensitivity, reproducibility and precision for clinicalapplications (Ref. 73).Still, proven reproducibility is essential for SRM

assays to move beyond basic research and becomea force in clinical or diagnostic assay developmentand application. A recent, multisite reviewdemonstrated high reproducibility acrossdifferent laboratories using different instrumentplatforms (Ref. 74). Another recent reviewquestioned whether SRM-MS will replaceantibody-based testing in the validation ofbiomarkers (Ref. 75). SRM-MS has several

advantages over antibody assays for biomarkervalidation: SRM has exquisite sensitivity, withno crossreactivity and less specificity issues thanare often associated with antibody assays; SRMis ‘reagent independent’; SRM can be used forany MS-observable ion, making it generallycheaper than antibody assays (Ref. 76); andthese assays are quantitative and easilymultiplexed. This is a key point, as realitydictates that having a single biomarker for adisease is unlikely; panels of biomarkers are themore likely future of disease diagnostics, andSRM technologies are very well placed to studythese. Issues still remain however, particularlyregarding assay throughput and precision,which have not been thoroughly tested andcurrently do not meet the US Food and DrugAdministration (FDA) requirements for routineclinical tests (Ref. 75). Another issue is operatorfamiliarity, as these assays are only beginning toenter the mainstream, and it will take time forusers to become comfortable with applyingthese new techniques. However, the hurdlesfacing the use of SRM in biomarker validationare slowly being overcome, and althoughantibody assays will still be used for biomarkervalidation, increasingly we can expect to see theapplication of SRM assays.

Bioinformatic analysisWith the rapid development of MS-basedproteomic technologies, automated analysis ofthe qualitative and quantitative data resultingfrom large-scale proteomic studies has becomeincreasingly important and challenging (Ref. 77).The large number of MS/MS spectra generatedin a typical proteomic experiment requiresseveral stages of analysis, including statisticalvalidation of peptide and protein identifications,analysis of any quantitative information andinterpretation of the resultant protein information.

Protein identificationIdentification of peptides and their correspondingproteins is generally conducted using searchalgorithms that correlate experimental MS/MSspectra to theoretically derived spectra createdfrom known peptide sequences. There areseveral different search engines available, whichdiffer in their approaches to identifying peptidesequences. The most common search algorithmsinclude Sequest (Ref. 78), Mascot (Ref. 79) andX!Tandem (Ref. 80). It is worth noting that these




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

Ionsource

Detector

Q1Set as a mass filter

Q3Set as a mass filter

Q2Set as an ion guide

Precursorion selection

Collision cell(peptide fragmentation)

Fragmention selection

a

b

Time (min)

30.48

Inte

nsity

(cp

s)

6.0e4488.7/630.3

488.7/701.3

488.7/276.1

488.7/424.7

3.5e4

3.0e4

2.0e4

488.7/343.2

20 25 30 35 40 45 50 55 60 65 70 75 80

1.0e4

Peptide: AGFAGDDAPR

Mode of operation of selected reaction monitoring (SRM)Expert Reviews in Molecular Medicine © Cambridge University Press 2010 (part a only)

Figure 3. Mode of operation of selected reaction monitoring (SRM). (a) After protein/peptide separation,peptides elute from a reversed-phase column, ionise, and enter a triple quadrupole mass spectrometer. Thefirst quadrupole (Q1) is set as a mass filter for a specific peptide; Q2 is set as an ion guide/collision cell,where peptides selected in Q1 are fragmented; and Q3 is set as a mass filter that specifically transmits aparticular fragment ion. When specific peptide–fragment transitions occur, a signal is recorded by thedetector, which can be plotted as a chromatogram. (b) Example chromatograms from an SRM experimentare shown. The SRM tool at the Global Proteome Machine was used to design five transitions (Q1/Q3 ionpairs) for the peptide AGFAGDDAPR from the protein β-actin (ACTB). A complex mixture, digested humanplatelet lysate, was used as the test sample. The resulting transitions display sufficient intensity andspecificity to allow for positive identification of this peptide in the sample. The Q1/Q3 m/z ratios for eachtransition are displayed on the graphs in bold and the elution time (30.48) is indicated. Quantitation of thetransition can be conducted using the peak area. As SRM is a fast and sensitive method, many transitionscan be acquired serially in a short time, allowing quantitation of multiple transitions in a single experiment.SRM transitions (b) are reprinted from Ref. 59 (©2009), with permission from Elsevier. Abbreviations:Q, quadrupole; m/z, mass to charge ratio; SRM, selected reaction monitoring.




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

Table

2.Exa

mplesofse

lected

reac

tionmonitoring

(SRM)reso

urce

s

Res

ource

Brief

des

cription

Source

Ref./web

site

MRMPilo

tSug

gests/op

timises

tran

sitio

nsBuildsSRM

andMIDAS

workflows

Toolsforrev

iewingan

darch

iving

data

Com

mercial

–AppliedBiosystem

shttp://w

ww.appliedbiosystem

s.co

m

SRM

workflow

software

BuildsSRM

metho

ds

Proce

ssingan

dreview

ofresu

ltsCom

mercial

–Th

ermoScien

tific

http://w

ww.the

rmo.co

m

VerifyE

High-throug

hput

optim

isationof

SRM

tran

sitio

nsCom

mercial

–Waters

http://

www.w

aters.co

m

Mas

sHun

terOptim

izer

Autom

atically

optim

ises

data

acquisitio

nparam

etersforSRM

Com

mercial

–Agilent

tech

nologies

http://

www.che

m.agilent.com

TIQAM

(targeted

iden

tification

forqua

ntita

tivean

alysis

byMRM)

Optim

ises

SRM

tran

sitio

nsfor

iden

tificationan

dqua

ntita

tion

Free

lyav

ailable

from

Sea

ttle

Proteom

ics

Cen

treus

ingPep

tideA

tlasdatab

ase

Ref.1

45

MRMer

Man

ages

MRM-bas

edex

perim

ents

Extractsprecu

rsor

andproduc

tmas

ses

Calcu

latesrelativ

earea

under

thecu

rveforqua

ntita

tion

Free

lyav

ailable

from

Fred

Hutch

inso

nCom

putationa

lProteom

icsLa

boratory(CPL)

Proteom

icsRep

osito

ry

Ref.1

46http://

proteom

ics.fhcrc.org/

CPL/MRMer.htm

l

Pep

tideA

tlas(in

corporating

MRMAtla

s)Pub

licrepos

itory

usefulforS

RM

des

ign

Free

lyav

ailable

from

Sea

ttle

Proteom

ics

Cen

ter

Ref.1

47http://

www.pep

tidea

tlas.org

http://

www.m

rmatlas.org

Tran

che

Pub

licrepos

itory

usefulforS

RM

des

ign

Free

lyav

ailable

from

proteom

ecom

mon

s.org

https://proteom

ecom

mon

s.org/

tran

che

PRIDE(protein

iden

tification

datab

ase)

Pub

licrepos

itory

usefulforS

RM

des

ign

Free

lyav

ailable,h

ostedbyEurop

ean

Bioinform

aticsInstitu

teRef.1

48http://

www.ebi.a

c.uk

/prid

e

GPMDB(Global

Proteom

eMac

hine

datab

ase)

Pub

licrepos

itory

usefulforS

RM

des

ign

Free

lyav

ailable

Ref.1

49http://w

ww.the

gpm.org

Abbreviations

:MIDAS,M

RM-initia

teddetec

tionan

dse

que

ncingworkflow;M

RM,m

ultip

lereac

tionmon

itorin

g.




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

search engines are complementary to some extent,so it is often useful to use at least two differentalgorithms to analyse MS/MS data to increaseconfidence and sensitivity, and there are toolsavailable to aid this (Ref. 81). Validation of thepeptide and protein identifications is necessaryand is often conducted by determining falsediscovery rates using decoy databases and otherstatistical methods (Ref. 82). For de novosequencing of proteins, an approach that isuseful in the analysis of PTMs or of organismswhose genome has not been sequenced, thereare several software options available, includingPEAKS (Ref. 83).

Analysis of quantitative proteomic dataTools for the analysis of quantitative data fromproteomic experiments are continuouslyemerging and being refined. The Trans-Proteomic Pipeline (Ref. 84) is a collection ofintegrated MS/MS analysis tools, includingXRPESS and ASAPRatio that are used for therelative quantitation of isotopically labelledpeptides and proteins. MaxQuant is a recentlydeveloped software suite for the analysis andquantitation of SILAC experiments (Ref. 85).Similarly, Mascot Distiller, from Matrix Science,determines quantitation based on the relativeintensities of extracted ion chromatogramsfor precursors (http://www.matrixscience.com).This approach can be used for ‘label-free’approaches, or with any chemistry that creates aprecursor mass shift, for example 18O, AQUA,ICAT, ICPL, metabolic labelling and SILAC.ProteinPilot, from Applied Biosystems (http://www.appliedbiosystems.com), provides proteinidentification and quantitation of SILAC- andiTRAQ-based labels. For label-free approaches,there are many open-source and commercialsoftware packages available, which arediscussed in a recent review (Ref. 86). It is worthnoting that for the analysis of quantitativeproteomic data, no standard procedure has beendeveloped that is broadly applicable to allexperiment types. As is evident, many softwaretools exist, and the user still needs tounderstand what the software is doing in orderto be able to critically analyse the results.

Data-mining approachesWith the rapid growth in large-scale proteomicexperiments comes the generation of longer andlonger lists of proteins. However, the sound

biological interpretation of these data lags behind(Ref. 77). There are now several analyticalstrategies and tools available to extractbiologically relevant information (e.g. regardingprotein–protein interactions, signalling pathwaysand biological networks) from these largeproteomic datasets. These ‘data-mining’approaches have the potential to contribute to adeeper understanding of biological systems, butneed to be applied and interpreted correctly. Oneof the most powerful tools available, and oftenthe first tool used to conduct analysis on a largedataset, is Gene Ontology (GO) (Ref. 87). This is acontrolled vocabulary that is used tostandardise the way in which proteins aredescribed across different species and databases.The consistency in terminology that this ontologyprovides makes it an invaluable resource for bothexperimentalists and bioinformaticians. GOannotation of a large MS dataset can be used todetermine whether there is any enrichment ordepletion for a particular GO category, or can beused to compare two different datasets.

Pathway and network analysisAnother useful approach is pathway analysis,which explores proteomic data in terms ofbiological pathways, based on known physicaland functional interactions between proteinsthat are present. It is estimated that there arearound 300 biochemical pathway analysis toolscurrently available (Ref. 77), with the KyotoEncyclopedia of Genes and Genomes (KEGG)and Reactome representing the largestdatabases. Many of the pathway analysis toolsare freely available, but there are also somecommercially available tools – for example,Ingenuity Pathways Analysis from IngenuitySystems, and GeneGo from GeneGo Inc. With somany pathway analysis options to choose from,Pathguide (http://www.pathguide.org), whichcontains information on about 317 biologicalpathway tools, is an invaluable resource to helpguide users in selecting the most appropriateresource to use (Ref. 88). Pathguide also coverstools that model network and functionalinteraction information, which takes the databeyond pathway analysis and groups proteinsbased on participation in larger, multiproteinassemblies. For visualisation of molecularnetworks, Cytoscape is a useful open-sourceplatform, which also allows integration ofgenetic and other information (Refs 89, 90).




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

There are now several meta-databases forinteraction information, including STRING(Refs 91, 92), which generates interactionnetworks by incorporating data from manycurated databases, as well as predictedinteractions and pathway information. Data canbe input to STRING as protein lists, and it has auser-friendly interface. MiMI, from the NationalInstitute for Integrative Biomedical Informatics,merges data from numerous interactiondatabases as well as other sources and also hasa Cytoscape plug-in to allow easy visualisationof networks (Ref. 93). For all interactiondatabases, which can have high error rates, careneeds to be taken when interpreting informationand the source of the interaction informationshould be checked manually if possible.

Meta-data analysis and data integrationOne of the key challenges currently facingresearchers is the integration of all these availabledata. There are several tools available for meta-data analysis of proteomic data, including thedatabase for annotation, visualisation andintegrated discovery (DAVID), from the NationalInstitute of Allergy and Infectious Diseases(NIAID), which provides a comprehensive set offunctional annotation tools for investigators tounderstand the biological meaning behind largelists of genes. Other meta-tools include PANTHER(protein analysis through evolutionaryrelationships, http://www.pantherdb.org), whichwas designed to classify proteins (and their genes)in order to facilitate high-throughput analysis,and Babelomics, which is a suite of interconnectedtools used to functionally annotate genome-scaleexperiments. Conceptgen, from the NationalInstitute for Integrative Biomedical Informatics,is a web-based tool designed to explore networksof relationships between biological concepts(Ref. 94). The Global Proteome Machine (http://www.thegpm.org), a search engine and databasefor MS/MS data, links the protein identificationsdirectly to annotation resources such as GO andKEGG within the same platform, allowingefficient examination of the GOs and pathwaysunder- or over-represented in a particular dataset.

Discussion: biological and clinicalapplications

One of the initial goals of proteomics was thecollection of inventories of whole proteomes. Byapplying subcellular and protein/peptide

fractionation approaches, MS-based proteomicshas had marked success in this endeavour,especially with the more abundant components.Many of the proteomes relevant to humanhealth research are now well characterised,including those of human blood cells (Refs 95,96, 97, 98), plasma (Ref. 99), cerebrospinal fluid(Ref. 100), bronchial epithelia (Ref. 101),heart muscle (Refs. 102) and a variety of cancertissues and cells (e.g. Refs 103, 104, 105). TheNormal Clinical Tissue Alliance (NCTA, http://wiki.thegpm.org/wiki/Normal_Clinical_Tissue_Alliance) provides high-quality proteomiccatalogues of clinically relevant normal humantissues, such as brain and bone, and also bodilyfluids, including bronchoalveolar lavage, salivaand urine. Global proteomic analysis can also bea helpful tool to investigate cell subtypes, asshown recently by applying a proteomicapproach to demonstrate that the so-calledendothelial progenitor cells may actually bemonocytes that had taken up plateletmicroparticles (Ref. 106).

However, the real strength behind proteomicapproaches lies in the ability to compare andquantitate samples. Formerly, 2D gel analysiswas one of the only ways to gain quantitativeinformation on a set of proteins, and althoughthere are still many current publicationssuccessfully using this approach, alternativetechniques such as isotopic labelling arecurrently supplanting 2D gels as a preferredquantitation method. Many studies applyingthese approaches have successfully identifiedbiomarkers with clinical potential. For example,a recent study used a combination of murinecancer models and iTRAQ quantitation todiscover a novel, putative biomarker for gastriccancer (Ref. 107) (Fig. 4). The biomarker wasthen validated in serum from cancer patients.Quantitation is especially important in the studyof time-dependent processes, such as thechanges that take place during storage of bloodbefore transfusion. The platelet storage lesionhas been studied by applying severalcomplementary quantitative proteomicapproaches to platelets at days 1 and 7 ofstorage (Ref. 108). 2D gel electrophoresis/differential gel electrophoresis (DIGE), iTRAQand ICAT were used, resulting in 503 proteinchanges identified over the course of storage,the majority of which were identified using theiTRAQ method. Despite this, the benefit of




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

usingmultiple quantitative proteomic approacheswas evident, as less than 16% of the 503 proteinswere identified by two or more proteomicapproaches and only five proteins were identifiedby all approaches.Combining the technologies discussed in

this review is a good way of utilising the

power of MS-based proteomics for targetedclinical studies. For example, tandem affinitypurifications, GeLC-MS/MS and iTRAQquantitation have been used to map theinteractome of the drug target BCR–ABL, atyrosine kinase causing chronic myeloidleukaemia (Ref. 109). A tightly bound cohort of

a

b c

In vivo cell culture

Control(No tumor)

Low tumour burden

Blood collection via cardiac puncture

Abundant proteins depletion of plasma samples

iTRAQ replicate 1

LC-MS/MS analysis LC-MS/MS analysis LC-MS/MS analysis

iTRAQ replicate 2 iTRAQ replicate 3

Mid tumour burden High tumour burden

Sensitivity: 96%Specificity: 66%Cut-off: 111305

2-sample t-testP < 0.001n = 167

500 000

400 000

300 000

200 000

100 000

00.00

0.00

0.25

0.50

0.75

1.00

0.25 0.50 0.75 1.00Cancer Normal

Status 1 – Specificity

ITIH

3

Sen

sitiv

ity

Discovery of the potential biomarker ITIH3 for early detection of gastric cancerExpert Reviews in Molecular Medicine 2010 Published by Cambridge University Press

MKN45

Figure 4. Discovery of the potential biomarker ITIH3 for early detection of gastric cancer. (See next pagefor legend.)




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

interconnected proteins around BCR–ABL, whichremodels on inhibitor treatment, was found,suggesting that the effect of the drugs is causedby a remodelling of the BCR–ABL complexinstead of a simple inhibition of the protein itself.In a complementary approach, novel kinase andnonkinase targets of three BCR–ABL inhibitorswere discovered utilising GeLC-MS/MS, iTRAQquantitation and IMAC phosphopeptideenrichment, showing that these MS-basedexperiments are a valuable tool to discover andstudy additional drug targets (Ref. 110).However, the clinical application of MS-based

proteomics still faces several challenges(Ref. 111). Certain experiments and manyputative clinical applications require the analysisof very small amounts of cells (100–1000). MS isa very sensitive method and in principle allowsthe analysis of single cells (Ref. 112). However,one major problem in the analysis of few cellslies in the standard sample preparation anddigestion protocols used for MS-basedproteomics, during which a high percentage ofthe sample can be lost (Ref. 111). An optimisedlysis and digestion method was developed toaddress this problem, which is performed in onetube. Furthermore, by optimising the parametersof the LC-MS/MS system for the analysis ofsmall amounts of cells it was possible to analyseas few as 500 cells, from which 167 proteinswere identified (Ref. 113). Another problem stillfaced by the proteomic community isaccessibility to low-abundance proteins,particularly in the presence of high-abundanceproteins, such as in the analysis of serum.Depletion methods can be used to remove these

proteins in order to investigate lower abundanceproteins; however, some peptides and proteinsbind to these carrier proteins and are discardedthrough this procedure. As an alternativeapproach, a differential solubilisation methodwas developed to enrich for low-abundanceproteins in plasma. By analysing these enrichedfractions with high-quality MALDI-TOF (time offlight), more than 1500 peptides from a 1 μlserum sample were identified and four newpotential colon cancer biomarkers werediscovered (Ref. 114). This approach hasthe potential to greatly contribute to thediscovery of novel low-abundance biomarkers.One aspect that is especially important forthe analysis of biomarkers in serum isreproducibility: it has been shown that serumproteins are degraded by endogenous proteasesshortly after a blood draw, leading to varyingresults. However, the addition of proteaseinhibitors to the blood drawing tubes cancounteract this effect and stabilise serumproteins (Ref. 115).

As outlined in this article, MS-based proteomicapproaches are now applied to many diverseaspects of clinical research, some of which arehighlighted in Table 3, and the ultimate hope isfor the development of diagnostic andprognostic tools that will benefit human health.As more potential biomarkers move from thediscovery phase towards clinical trials, thereis the need for accurate statistical andmathematical analysis of the data, in order tobetter determine key outcomes, for exampleprecision and accuracy, using standardised testssuch as positive predictive value (Ref. 116).

Figure 4. Discovery of the potential biomarker ITIH3 for early detection of gastric cancer. (See previouspage for figure.) (a) A mouse xenograft model was used to identify putative biomarkers for gastric cancer.Tumours were induced in mice with the human gastric cancer cell line MKN45 and mice were categorisedaccording to tumour burden [low (length, L= 1–2 mm, volume, V= 2–3 mm3); mid (L= 7.5 mm, V=127–210 mm3); high (L= 15 mm, V= 726–1078 mm3)]. Plasma from these mice and a control group waslabelled with four different iTRAQ labels and studied by LC-MS/MS. Triplicates were performed to obtainhigh-quality data. (b) Thirty-one proteins were identified as putative biomarkers, and the presence of one ofthese proteins, ITIH3, was analysed in serum derived from healthy humans (normal) and gastric cancerpatients (cancer) by immunoblotting. ITIH3 levels were found to be elevated significantly (P-value <0.001) incancer patients. (c) An ROC curve was generated using the data from (b) to estimate the accuracy of ITIH3detection in gastric cancer detection. Sensitivity was determined to be very high, at 96%, whereasspecificity was slightly lower, at 66%. The area under the ROC curve was found to be 0.86 (with 0.5 being auseless and 1.0 a perfect test), which implies that ITIH3 could be a valuable biomarker in early gastriccancer detection. Figure adapted with permission from Ref. 107 (©2010 American Chemical Society).Abbreviations: iTRAQ, isobaric tags for relative and absolute quantitation; LC-MS/MS. liquidchromatography tandem mass spectrometry; ROC, receiver operating characteristics.




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

Table 3. Recent examples of clinical applications of proteomics

Application Examples

Risk determination (discovery ofbiomarkers for disease risk)

A serological proteome (SERPA) approach was used to identifyautoantibodies in melanoma patients, by identifying positivereactions between patients’ sera and proteins isolated from a G361melanoma sample (Ref. 150)N-glycoproteins from enrichedmembranes of breast cancer cell lineswere analysed to generate a set of potential biomarkers (Ref. 29)

Early detection (discovery ofbiomarkers to aid in early diagnosis ofdisease)

Plasma samples from gastric cancer patients were screened for aprotein found to be highly expressed in a mouse model of gastriccancer; it was identified as a potential biomarker expressed in early-stage gastric cancer (Ref. 107)2D gel separation and MS was applied to identify a protein that wassignificantly upregulated in hepatocellular carcinoma and elevated inplasma of patients, which could be used to detect early stages of thedisease (Ref. 151)

Verification/quantitation ofbiomarkers (validating previouslydiscovered biomarkers, quantifyingbiomarkers, developing more broadlyapplicable assays to detectbiomarkers)

SISCAPAandSRM-MSwereused toquantifyTIMP1,acolorectal cancerbiomarker, from patients’ sera at attomolar concentrations (Ref. 71)A broadly applicable multiplexed, MS-based assay was used to verifyand quantify changes of biomarker proteins associated with cardiacinjury in the low ng/ml range (Ref. 69)Isotope-labelledsyntheticpeptidesandSRMwasapplied toscreenCRP,acandidatebiomarker for rheumatoidarthritis, insmall volumesofhumanserum depleted of major plasma proteins (Ref. 64)

Proteomic classification of disease(establishing biomarker panels,determining the proteomic profile/proteomic signature of disease)

SELDI-TOF-MS ProteinChip technology identified and tested aserum profile for distinguishing hepatocellular carcinoma and livercirrhosis, and showed it could be a better diagnostic tool than apreviously established marker (Ref. 152)An MS fingerprint based on three MALDI-TOF MS peaks wasidentified that specifically separated patients with rheumatoidarthritis from healthy controls (Ref. 153)Genomic and proteomic markers of mild and moderate/severechronic allograft nephropathy in peripheral blood that could be usedto predict graft loss were identified (Ref. 154)

Characterisation of disease[identifying altered proteins orsignalling pathways in disease,characterising disease progression,(sub)classification of disease]

Phosphotyrosine affinity columns and SILAC were used to identifyand quantitate proteins dependent on SYK signalling in humancancer cells (MCF7) in order to elucidate the role of this protein intumour formation and progression (Ref. 43)A novel application of SILAC was used to identify significantdifferences in expression of focal adhesion and glycolytic proteinsbetween lobular and ductal tumors (Ref. 51)The protein–protein interaction network of the tyrosine kinase BCR-ABL, implicated in myeloid leukaemia, was charted using affinitypurification and MS (Ref. 109)

Cataloguing proteomes of diseasedtissues/cells (creating catalogues ofproteins identified in either normal ordiseased tissues or cells)

Differential expression analysis of human colorectal cancer cells in anin vitro model system was used to examine progression fromadenoma to carcinoma (Ref. 104)2D separation, DIGE and SERPA were applied to construct a proteinexpression database for human non-small-cell lung cancer (Ref. 103)Proteomic and genomic profiles of airway epithelial cells from neverand current smokers were generated and correlated (Ref. 101)

(continued on next page)




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

ConclusionThe developments discussed above point to thefuture potential of the proteomic approach forthe exploration of key questions in basic andclinical research as well as in establishing toolsfor clinical diagnostics, while emphasising thecontinued importance of technology andmethod development in pushing the boundariesof MS-based proteomics. The development ofhigh-throughput technologies such as therecently established SRM assays for mappingthe kinases and phosphatases of Saccharomycescerevisiae makes the quantitative analysis ofwhole proteomes more realistic (Ref. 56). Amajor movement already under way is theHuman Proteome Project, which will beformally launched at the HUPO 2010 congressand plans to map the human proteome in amanner analogous to the mapping of the humangenome. Three different experimentalapproaches are proposed to achieve this: MS-

based proteomics to identify and quantifyproteins in tissues and cells, generation ofantibodies against each protein to show cellularlocation, and systematic identification ofinteractors for every protein. This ambitiousproject requires established standards forproteomics-based profiling, antibody-basedprofiling and network-based profiling, as well asa massive bioinformatics effort to analyse,archive and make available the ensuing data(Ref. 117). With a projected ten-year time line,the Human Proteome Project includes a clearclinical focus, with its stated aim being thecreation of a resource immediately available tothe clinical and basic science communities in aformat that assures fundamental discoveriesand insight into diagnostic and treatmentregimens for the patient (Ref. 117). Theaspirations of the Human Proteome Projectemphasise the hoped-for impact of proteomicson human health research and highlight once

Table 3. Recent examples of clinical applications of proteomics (continued)

Application Examples

Uncovering the effects of drugs orpotential drug targets (comparingtreated versus untreated samples todetermine themechanismof action ofa therapeutic agent, uncoveringpotential drug targets)

Chemical proteomics together with immunoaffinity purification oftyrosine-phosphorylated peptides identified nearly 40 differentkinase targets of the SRC-family kinase inhibitor dasatinib (Ref. 33)Plasma proteome changes were investigated in ALS patients beforeand during immunisation with glatiramer acetate in a clinical trial(Ref. 15)A chemical proteomics affinity purification approach was used forquantitative profiling of the targets of the drugs imatinib, dasatiniband bosutinib (Ref. 110)

Monitoring disease progressiontreatment response/prognosticmarkers (assessing/monitoringdisease progression, monitoringresponse of patients to treatment,determining patient prognosis)

SELDI-TOF MS was used to profile and compare the serum ofresponding and nonresponding patients with metastatic colorectalcancer to identify biomarkers that could predict treatment responseand be used for monitoring (Ref. 155)An independentprognostic factorwas identified fordisease-freesurvivaland overall survival in patients with serous ovarian cancer (Ref. 156)Analysis of blood samples frompatientswith biopsy-confirmedacuterenal allograft rejection, chronic rejection and stable graft functionwas used to establish serum peptidome fingerprints and aid in theearly diagnosis of renal allograft rejection (Ref. 157)A novel prognostic marker for distant metastasis in non-small-celllung cancer was identified by cancer cell secretome pleural effusionproteome analysis (Ref. 158)

Abbreviations: ALS, amyotrophic lateral sclerosis; CRP, C-reactive protein; 2D, two dimensional; DIGE,difference gel electrophoresis; MALDI, matrix-assisted laser desorption/ionisation; MS, mass spectrometry;SELDI, surface-enhanced laser desorption/ionisation; SERPA, serological proteome; SILAC, isotope labellingof amino acids in cell culture; SISCAPA, stable isotope standards and capture by antipeptide antibodies;SRM, selected reaction monitoring; TIMP1, tissue inhibitor of metalloproteinase 1; TOF, time of flight.




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

again the steady march of proteomics into theclinical realm.

Acknowledgements and fundingThe authors acknowledge the Canadian Institutesfor Health Research, the Natural Sciences andEngineering Research Council of Canada, andthe Michael Smith Foundation for HealthResearch for funding. G.M.W. is supported by apostdoctoral fellowship from the CanadianInstitutes of Health Research/Heart and StrokeFoundation of Canada Strategic TrainingProgram in Transfusion Science through theCentre for Blood Research at the University ofBritish Columbia. The authors thank the refereesfor their instructive comments.

References1 Fenn, J.B. et al. (1989) Electrospray ionization for

mass spectrometry of large biomolecules. Science246, 64-71

2 Tanaka, K. et al. (1988) Protein and polymeranalyses up to m/z 100 000 by laser ionizationtime-of-flight mass spectrometry. RapidCommunications in Mass Spectrometry 2,151-153

3 Anderson, N.L. and Anderson, N.G. (2002) Thehuman plasma proteome: history, character, anddiagnostic prospects. Molecular and CellularProteomics 1, 845-867

4 Aebersold, R. and Mann, M. (2003) Massspectrometry-based proteomics. Nature 422,198-207

5 Hu, Q. et al. (2005) The orbitrap: a new massspectrometer. Journal of Mass Spectrometry 40,430-443

6 Han, X., Aslanian, A. and Yates, J.R., 3rd (2008)Mass spectrometry for proteomics. CurrentOpinion in Chemical Biology 12, 483-490

7 Swaney, D.L., Wenger, C.D. and Coon, J.J.(2010) Value of using multiple proteases for large-scale mass spectrometry-based proteomics. Journalof Proteome Research 9, 1323-1329

8 Yates, J.R., Ruse, C.I. and Nakorchevsky, A. (2009)Proteomics by mass spectrometry: approaches,advances, and applications. Annual Review ofBiomedical Engineering 11, 49-79

9 Dugo, P. et al. (2008) Comprehensivemultidimensional liquid chromatography: theoryand applications. Journal of Chromatography A1184, 353-368

10 Tang, J. et al. (2008) Recent development of multi-dimensional chromatography strategies in

proteome research. Journal of Chromatography B,Analytical Technologies in the Biomedical and LifeSciences 866, 123-132

11 Shen, Y. et al. (2004) Ultra-high-efficiency strongcation exchange LC/RPLC/MS/MS for highdynamic range characterization of the humanplasma proteome. Analytical Chemistry 76,1134-1144

12 Zhou, H. et al. (2010) New ammunition for theproteomic reactor: strong anion exchange beadsand multiple enzymes enhance proteinidentification and sequence coverage. Analyticaland Bioanalytical Chemistry 397, 3421-3430

13 Sluyterman, L.A.A. and Elgersma, O. (1978)Chromatofocusing – isoelectric-focusing on ion-exchange columns. 1. General principles. Journalof Chromatography 150, 17-30

14 Sluyterman, L.A.A. and Wijdenes, J. (1978)Chromatofocusing – isoelectric-focusing on ion-exchange columns. 2. Experimental-verification.Journal of Chromatography 150, 31-44

15 Schlautman, J.D. et al. (2008) Multidimensionalprotein fractionation using ProteomeLab PF 2D forprofiling amyotrophic lateral sclerosis immunity: apreliminary report. Proteome Science 6, 26

16 Michel, P.E. et al. (2003) Protein fractionation in amulticompartment device using off-gel isoelectricfocusing. Electrophoresis 24, 3-11

17 Fang, Y., Robinson, D.P. and Foster, L.J. (2010)Quantitative analysis of proteome coverage andrecovery rates for upstream fractionation methodsin proteomics. Journal of Proteome Research 9,1902-1912

18 Tran, J.C. and Doucette, A.A. (2008) Gel-elutedliquid fraction entrapment electrophoresis: anelectrophoretic method for broadmolecular weightrange proteome separation. Analytical Chemistry80, 1568-1573

19 Washburn, M.P., Wolters, D. and Yates, J.R., 3rd(2001) Large-scale analysis of the yeast proteome bymultidimensional protein identificationtechnology. Nature Biotechnology 19, 242-247

20 Lin, D., Alpert, A.J. and Yates, J.R., 3rd (2001)Multidimensional protein identificationtechnology as an effective tool for proteomics.AmericanGenomic/Proteomic Technology 1, 38-46

21 Bendall, S.C. et al. (2009) An enhanced massspectrometry approach reveals human embryonicstem cell growth factors in culture. Molecular andCellular Proteomics 8, 421-432

22 Delahunty, C. and Yates, J.R., 3rd (2005) Proteinidentification using 2D-LC-MS/MS. Methods 35,248-255




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

23 Wei, J. et al. (2005)Global proteomediscovery usingan online three-dimensional LC-MS/MS. Journal ofProteome Research 4, 801-808

24 Shen, Z. et al. (2006) Sepsis plasma protein profilingwith immunodepletion, three-dimensional liquidchromatography tandem mass spectrometry, andspectrumcounting. Journal of ProteomeResearch 5,3154-3160

25 Shen, Y. et al. (2002) High-efficiency nanoscaleliquid chromatography coupled on-line withmass spectrometry using nanoelectrosprayionization for proteomics. Analytical Chemistry74, 4235-4249

26 Pernemalm,M., Lewensohn, R. and Lehtio, J. (2009)Affinity prefractionation for MS-based plasmaproteomics. Proteomics 9, 1420-1427

27 Dwivedi, R.C. et al. (2010) Assessment of thereproducibility of random hexapeptide peptidelibrary-based protein normalization. Journal ofProteome Research 9, 1144-1149

28 Yanagida, M. et al. (2007) Proteomic analysis ofplasma membrane lipid rafts of HL-60 cells.Proteomics 7, 2398-2409

29 Whelan, S.A. et al. (2009) Mass spectrometry(LC-MS/MS) site-mapping of N-glycosylatedmembrane proteins for breast cancer biomarkers.Journal of Proteome Research 8, 4151-4160

30 Markham, K., Bai, Y. and Schmitt-Ulms, G. (2007)Co-immunoprecipitations revisited: an update onexperimental concepts and their implementationfor sensitive interactome investigations ofendogenous proteins. Analytical and BioanalyticalChemistry 389, 461-473

31 Koch, H.B. et al. (2007) Large-scale identificationof c-MYC-associated proteins using a combinedTAP/MudPIT approach. Cell Cycle 6, 205-217

32 Dobreva, I. et al. (2008)Mapping the integrin-linkedkinase interactome using SILAC. Journal ofProteome Research 7, 1740-1749

33 Li, J. et al. (2010)A chemical andphosphoproteomiccharacterization of dasatinib action in lung cancer.Nature Chemical Biology 6, 291-299

34 Thingholm, T.E. and Jensen, O.N. (2009)Enrichment and characterization ofphosphopeptides by immobilized metalaffinity chromatography (IMAC) and massspectrometry. Methods in Molecular Biology527, 47-56, xi

35 Sugiyama, N. et al. (2007) Phosphopeptideenrichment by aliphatic hydroxy acid-modifiedmetal oxide chromatography for nano-LC-MS/MSin proteomics applications. Molecular and CellularProteomics 6, 1103-1109

36 Rogers, L.D. and Foster, L.J. (2009)Phosphoproteomics – finally fulfilling the promise?Molecular Biosystems 5, 1122-1129

37 Dengjel, J., Kratchmarova, I. and Blagoev, B. (2009)Receptor tyrosine kinase signaling: a view fromquantitative proteomics. Molecular Biosystems 5,1112-1121

38 Blagoev, B. et al. (2004) Temporal analysis ofphosphotyrosine-dependent signalingnetworks byquantitative proteomics. Nature Biotechnology 22,1139-1145

39 Olsen, J.V. et al. (2006) Global, in vivo, and site-specific phosphorylation dynamics in signalingnetworks. Cell 127, 635-648

40 Dengjel, J. et al. (2007) Quantitative proteomicassessment of very early cellular signaling events.Nature Biotechnology 25, 566-568

41 Zhang, Y. et al. (2005) Time-resolved massspectrometry of tyrosine phosphorylation sites inthe epidermal growth factor receptor signalingnetwork reveals dynamic modules. Molecular andCellular Proteomics 4, 1240-1250

42 Wolf-Yadlin, A. et al. (2007) Multiple reactionmonitoring for robust quantitative proteomicanalysis of cellular signalingnetworks. Proceedingsof the National Academy of Sciences of the UnitedStates of America 104, 5860-5865

43 Larive, R.M. et al. (2009) Phosphoproteomicanalysis of Syk kinase signaling in human cancercells reveals its role in cell-cell adhesion. Oncogene28, 2337-2347

44 Ong, S.E. andMann, M. (2005) Mass spectrometry-based proteomics turns quantitative. NatureChemical Biology 1, 252-262

45 Elliott, M.H. et al. (2009) Current trends inquantitative proteomics. Journal of MassSpectrometry 44, 1637-1660

46 Gygi, S.P. et al. (1999) Quantitative analysis ofcomplex protein mixtures using isotope-codedaffinity tags. Nature Biotechnology 17, 994-999

47 Gevaert, K. et al. (2008) Stable isotopic labeling inproteomics. Proteomics 8, 4873-4885

48 Ross, P.L. et al. (2004) Multiplexed proteinquantitation in Saccharomyces cerevisiae usingamine-reactive isobaric tagging reagents.Molecular and Cellular Proteomics 3, 1154-1169

49 Ong, S.E. et al. (2002) Stable isotope labelingby amino acids in cell culture, SILAC, as asimple and accurate approach to expressionproteomics. Molecular and Cellular Proteomics 1,376-386

50 Kruger, M. et al. (2008) SILAC mouse forquantitative proteomics uncovers kindlin-3 as an




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

essential factor for red blood cell function. Cell 134,353-364

51 Geiger, T. et al. (2010) Super-SILAC mix forquantitative proteomics of human tumor tissue.Nature Methods 7, 383-385

52 Li, Q. and Roxas, B.A. (2009) An assessment of falsediscovery rates and statistical significance in label-free quantitative proteomics with combined filters.BMC Bioinformatics 10, 43

53 Old, W.M. et al. (2005) Comparison of label-freemethods for quantifying human proteins byshotgun proteomics. Molecular and CellularProteomics 4, 1487-1502

54 Lange, V. et al. (2008) Selected reaction monitoringfor quantitative proteomics: a tutorial. MolecularSystem Biology 4, 222

55 Kim, K. and Kim, Y. (2009) Preparing multiple-reaction monitoring for quantitative clinicalproteomics. Expert Reviewof Proteomics 6, 225-229

56 Picotti, P. et al. (2010) High-throughput generationof selected reaction-monitoring assays for proteinsand proteomes. Nature Methods 7, 43-46

57 Nagy, K. et al. (2003) Direct tandem massspectrometric analysis of amino acids in driedblood spots without chemical derivatization forneonatal screening. Rapid Communications inMass Spectrometry 17, 983-990

58 Mallick, P. et al. (2007) Computational prediction ofproteotypic peptides for quantitative proteomics.Nature Biotechnology 25, 125-131

59 Walsh, G.M. et al. (2009) Implementation of a datarepository-driven approach for targetedproteomics experiments by multiple reactionmonitoring. Journal of Proteomics 72, 838-852

60 Han, B. and Higgs, R.E. (2008) Proteomics: fromhypothesis to quantitative assay on a singleplatform. Guidelines for developing MRM assaysusing ion trap mass spectrometers. Briefings inFunctional Genomics and Proteomics 7, 340-354

61 ChamMead, J.A., Bianco, L. and Bessant, C. (2010)Free computational resources for designingselected reaction monitoring transitions.Proteomics 10, 1106-1126

62 Yocum, A.K. and Chinnaiyan, A.M. (2009) Currentaffairs in quantitative targetedproteomics:multiplereaction monitoring-mass spectrometry. Briefingsin Functional Genomics and Proteomics 8, 145-157

63 Picotti, P. et al. (2008) A database of massspectrometric assays for the yeast proteome.NatureMethods 5, 913-914

64 Kuhn, E. et al. (2004) Quantification of C-reactiveprotein in the serum of patients with rheumatoidarthritis using multiple reaction monitoring mass

spectrometry and 13C-labeled peptide standards.Proteomics 4, 1175-1186

65 Janecki, D.J. et al. (2007) A multiple reactionmonitoring method for absolute quantification ofthe human liver alcohol dehydrogenase ADH1C1isoenzyme. Analytical Biochemistry 369, 18-26

66 DeSouza, L.V. et al. (2008) Multiple reactionmonitoring of mTRAQ-labeled peptides enablesabsolute quantification of endogenous levels of apotential cancer marker in cancerous and normalendometrial tissues. Journal of Proteome Research7, 3525-3534

67 Whiteaker, J.R. et al. (2007) Integrated pipelinefor mass spectrometry-based discovery andconfirmation of biomarkers demonstrated in amouse model of breast cancer. Journal of ProteomeResearch 6, 3962-3975

68 Menon, R. and Omenn, G.S. (2010) Proteomiccharacterization of novel alternative splice variantproteins in human epidermal growth factorreceptor 2/neu-induced breast cancers. CancerResearch 70, 3440-3449

69 Keshishian, H. et al. (2009) Quantification ofcardiovascular biomarkers in patient plasma bytargeted mass spectrometry and stable isotopedilution. Molecular and Cellular Proteomics 8,2339-2349

70 Fortin, T. et al. (2009) Clinical quantitation ofprostate-specific antigen biomarker in the lownanogram/milliliter range by conventional boreliquid chromatography-tandemmass spectrometry(multiple reaction monitoring) coupling andcorrelation with ELISA tests. Molecular andCellular Proteomics 8, 1006-1015

71 Ahn, Y.H. et al. (2009) Quantitative analysis of anaberrant glycoform of TIMP1 from colon cancerserum by L-PHA-enrichment and SISCAPAwithMRM mass spectrometry. Journal of ProteomeResearch 8, 4216-4224

72 Anderson, N.L. et al. (2004) Mass spectrometricquantitation of peptides and proteins using stableisotope standards and capture by anti-peptideantibodies (SISCAPA). Journal of ProteomeResearch 3, 235-244

73 Whiteaker, J.R. et al. (2010) An automated andmultiplexed method for high throughputpeptide immunoaffinity enrichment andmultiple reaction monitoring massspectrometry-based quantification of proteinbiomarkers. Molecular and Cellular Proteomics9, 184-196

74 Addona, T.A. et al. (2009) Multi-site assessment ofthe precision and reproducibility of multiple




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

reaction monitoring-based measurements ofproteins in plasma. Nature Biotechnology 27,633-641

75 Issaq, H.J. and Veenstra, T.D. (2008) Would youprefer multiple reaction monitoring or antibodieswith your biomarker validation? Expert Review ofProteomics 5, 761-763

76 Latterich, M., Abramovitz, M. and Leyland-Jones,B. (2008) Proteomics: new technologies and clinicalapplications. European Journal of Cancer 44,2737-2741

77 Malik, R. et al. (2010) From proteome lists tobiological impact – tools and strategies for theanalysis of large MS data sets. Proteomics 10,1270-1283

78 Eng, J.K.,Mccormack,A.L. andYates, J.R. (1994)Anapproach to correlate tandemmass-spectral data ofpeptides with amino-acid-sequences in a proteindatabase. Journal of the American Society for MassSpectrometry 5, 976-989

79 Perkins, D.N. et al. (1999) Probability-based proteinidentification by searching sequence databasesusing mass spectrometry data. Electrophoresis 20,3551-3567

80 Craig, R. and Beavis, R.C. (2004) TANDEM:matching proteins with tandem mass spectra.Bioinformatics 20, 1466-1467

81 Price, T.S. et al. (2007) EBP, a program for proteinidentification using multiple tandem massspectrometry datasets. Molecular and CellularProteomics 6, 527-536

82 Li, X., Pizarro, A. and Grosser, T. (2009) Electiveaffinities – bioinformatic analysis of proteomicmass spectrometry data. Archives of Physiologyand Biochemistry 115, 311-319

83 Ma, B. et al. (2003) PEAKS: powerful software forpeptide de novo sequencing by tandem massspectrometry. Rapid Communications in MassSpectrometry 17, 2337-2342

84 Keller, A. et al. (2005) A uniform proteomics MS/MS analysis platform utilizing open XML fileformats. Molecular System Biology 1, 2005.0017(doi: 10.1038/msb4100024)

85 Cox, J. and Mann, M. (2008) MaxQuant enableshigh peptide identification rates, individualizedp.p.b.-range mass accuracies and proteome-wideprotein quantification. Nature Biotechnology 26,1367-1372

86 Wang, M. et al. (2008) Label-free massspectrometry-based protein quantificationtechnologies in proteomic analysis. Briefings inFunctional Genomics and Proteomics 7,329-339

87 Ashburner, M. et al. (2000) Gene ontology: tool forthe unification of biology. The gene ontologyconsortium. Nature Genetics 25, 25-29

88 Bader, G.D., Cary, M.P. and Sander, C. (2006)Pathguide: a pathway resource list. Nucleic AcidsResearch 34, D504-D506

89 Shannon, P. et al. (2003) Cytoscape: a softwareenvironment for integrated models of biomolecularinteractionnetworks.GenomeResearch13, 2498-2504

90 Cline, M.S. et al. (2007) Integration ofbiological networks and gene expressiondata using Cytoscape. Nature Protocols 2,2366-2382

91 Snel, B. et al. (2000) STRING: a web-server toretrieve and display the repeatedly occurringneighbourhood of a gene. Nucleic Acids Research28, 3442-3444

92 Jensen, L.J. et al. (2009) STRING8 – a global viewonproteins and their functional interactions in 630organisms. Nucleic Acids Research 37,D412-D416

93 Tarcea, V.G. et al. (2009) Michigan molecularinteractions r2: from interacting proteins topathways. Nucleic Acids Research 37, D642-D646

94 Sartor, M.A. et al. (2010) ConceptGen: a gene setenrichment and gene set relation mapping tool.Bioinformatics 26, 456-463

95 Qureshi, A.H. et al. (2009) Proteomic and phospho-proteomic profile of human platelets in basal,resting state: insights into integrin signaling. PLoSOne 4, e7627

96 D’Alessandro, A., Righetti, P.G. and Zolla, L.(2010) The red blood cell proteome andinteractome: an update. Journal of ProteomeResearch 9, 144-163

97 Pasini, E.M. et al. (2006) In-depth analysis of themembrane and cytosolic proteome of red bloodcells. Blood 108, 791-801

98 Haudek, V.J. et al. (2009) Proteome maps of themain human peripheral blood constituents. Journalof Proteome Research 8, 3834-3843

99 Omenn, G.S. et al. (2005) Overview of the HUPOPlasma Proteome Project: results from the pilotphase with 35 collaborating laboratories andmultiple analytical groups, generating a coredataset of 3020 proteins and a publicly-availabledatabase. Proteomics 5, 3226-3245

100 Pan, S. et al. (2007) A combined dataset of humancerebrospinal fluid proteins identified by multi-dimensional chromatography and tandem massspectrometry. Proteomics 7, 469-473

101 Steiling, K. et al. (2009) Comparison of proteomicand transcriptomic profiles in the bronchial airway




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

epitheliumof current andnever smokers. PLoSOne4, e5043

102 Kline, K.G. et al. (2008) High quality catalog ofproteotypic peptides from human heart. Journal ofProteome Research 7, 5055-5061

103 Buhrens, R.I. et al. (2009) Protein expression inhuman non-small cell lung cancer: a systematicdatabase. Pathobiology 76, 277-285

104 Roth, U. et al. (2010) Differential expressionproteomics of human colorectal cancer based on asyngeneic cellular model for the progression ofadenoma to carcinoma. Proteomics 10, 194-202

105 Roesli, C. et al. (2009) Comparative analysis of themembrane proteome of closely related metastaticand nonmetastatic tumor cells. Cancer Research 69,5406-5414

106 Prokopi, M. et al. (2009) Proteomic analysis revealspresence of platelet microparticles in endothelialprogenitor cell cultures. Blood 114, 723-732

107 Chong, P.K. et al. (2010) ITIH3 is a potentialbiomarker for early detection of gastric cancer.Journal of Proteome Research

108 Thon, J.N. et al. (2008) Comprehensive proteomicanalysis of protein changes during platelet storagerequires complementary proteomic approaches.Transfusion 48, 425-435

109 Brehme, M. et al. (2009) Charting the molecularnetwork of the drug target Bcr-Abl. Proceedings ofthe National Academy of Sciences of the UnitedStates of America 106, 7414-7419

110 Bantscheff, M. et al. (2007) Quantitative chemicalproteomics reveals mechanisms of action of clinicalABL kinase inhibitors. Nature Biotechnology 25,1035-1044

111 Apweiler, R. et al. (2009) Approaching clinicalproteomics: current state and future fields ofapplication in cellular proteomics. Cytometry A 75,816-832

112 Whittal, R.M., Keller, B.O. and Li, L. (1998)Nanoliter chemistry combined with massspectrometry for peptidemapping of proteins fromsingle mammalian cell lysates. AnalyticalChemistry 70, 5344-5347

113 Wang, N. et al. (2010) Development of massspectrometry-based shotgun method for proteomeanalysis of 500 to 5000 cancer cells. AnalyticalChemistry 82, 2262-2271

114 Kawashima, Y. et al. (2010) High-yield peptide-extraction method for the discovery ofsubnanomolar biomarkers from small serumsamples. Journal of ProteomeResearch 9, 1694-1705

115 Yi, J., Kim, C. andGelfand, C.A. (2007) Inhibition ofintrinsic proteolytic activities moderates

preanalytical variability and instability ofhuman plasma. Journal of Proteome Research 6,1768-1781

116 Ransohoff, D.F. (2010) Proteomics research todiscover markers: what can we learn from Netflix?Clinical Chemistry 56, 172-176

117 [No authors listed] (2010) A gene-centric HumanProteome project: HUPO – the Human ProteomeOrganization.Molecular andCellular Proteomics 9,427-429

118 Chen, X., Smith, L.M. and Bradbury, E.M. (2000)Site-specific mass tagging with stable isotopes inproteins for accurate and efficient proteinidentification. Analytical Chemistry 72, 1134-1143

119 Harsha, H.C., Molina, H. and Pandey, A. (2008)Quantitative proteomics using stable isotopelabeling with amino acids in cell culture. NatureProtocols 3, 505-516

120 Blagoev, B. and Mann, M. (2006) Quantitativeproteomics to study mitogen-activated proteinkinases. Methods 40, 243-250

121 Bendall, S.C. et al. (2008) Prevention of amino acidconversion in SILAC experiments with embryonicstem cells. Molecular and Cellular Proteomics 7,1587-1597

122 Oda, Y. et al. (1999)Accurate quantitation of proteinexpression and site-specific phosphorylation.Proceedings of theNationalAcademyof Sciences ofthe United States of America 96, 6591-6596

123 Gao, H.Y. et al. (2000) Two-dimensionalelectrophoretic/chromatographic separationscombinedwith electrospray ionization FTICRmassspectrometry for high throughput proteomeanalysis. Journal of Microcolumn Separations 12,383-390

124 Conrads, T.P. et al. (2000) Utility of accurate masstags for proteome-wide protein identification.Analytical Chemistry 72, 3349-3354

125 Wu, C.C. et al. (2004) Metabolic labeling ofmammalian organisms with stable isotopes forquantitative proteomic analysis. AnalyticalChemistry 76, 4951-4959

126 Wang, Y.K. et al. (2002) Inverse 15N-metaboliclabeling/mass spectrometry for comparativeproteomics and rapid identification of proteinmarkers/targets. Rapid Communications in MassSpectrometry 16, 1389-1397

127 Li, J., Steen, H. and Gygi, S.P. (2003) Proteinprofiling with cleavable isotope-coded affinity tag(cICAT) reagents: the yeast salinity stress response.Molecular and Cellular Proteomics 2, 1198-1204

128 Hansen, K.C. et al. (2003) Mass spectrometricanalysis of protein mixtures at low levels using




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

cleavable 13C-isotope-coded affinity tag andmultidimensional chromatography. Molecular andCellular Proteomics 2, 299-314

129 Vaughn, C.P. et al. (2006) Analytical characteristicsof cleavable isotope-coded affinity tag-LC-tandemmass spectrometry for quantitative proteomicstudies. Journal of Molecular Diagnostics 8,513-520

130 Simpson, K.L., Whetton, A.D. and Dive, C. (2009)Quantitative mass spectrometry-based techniquesfor clinical use: biomarker identification andquantification. Journal of Chromatography. B,Analytical Technologies in Biomedical and LifeSciences 877, 1240-1249

131 Munchbach, M. et al. (2000) Quantitation andfacilitated de novo sequencing of proteins byisotopic N-terminal labeling of peptides with afragmentation-directing moiety. AnalyticalChemistry 72, 4047-4057

132 Schmidt,A., Kellermann, J. andLottspeich, F. (2005)A novel strategy for quantitative proteomics usingisotope-coded protein labels. Proteomics 5, 4-15

133 Yao, X. et al. (2001) Proteolytic 18O labeling forcomparative proteomics: model studies with twoserotypes of adenovirus. Analytical Chemistry 73,2836-2842

134 Sakai, J. et al. (2005) 18O-labeling quantitativeproteomics using an ion trap mass spectrometer.Proteomics 5, 16-23

135 Fischer, B. et al. (2006) Semi-supervised LC/MSalignment for differential proteomics.Bioinformatics 22, e132-140

136 Andreev, V.P. et al. (2007) A new algorithm usingcross-assignment for label-free quantitation withLC-LTQ-FT MS. Journal of Proteome Research 6,2186-2194

137 Finney, G.L. et al. (2008) Label-free comparativeanalysis of proteomics mixtures usingchromatographic alignment of high-resolutionmuLC-MS data. Analytical Chemistry 80, 961-971

138 America, A.H. and Cordewener, J.H. (2008)Comparative LC-MS: a landscape of peaks andvalleys. Proteomics 8, 731-749

139 Liu, H., Sadygov, R.G. and Yates, J.R., 3rd (2004) Amodel for random sampling and estimation ofrelative protein abundance in shotgun proteomics.Analytical Chemistry 76, 4193-4201

140 Silva, J.C. et al. (2006) Absolute quantification ofproteins by LCMSE: a virtue of parallel MSacquisition. Molecular and Cellular Proteomics 5,144-156

141 Choi, H., Fermin, D. and Nesvizhskii, A.I. (2008)Significance analysis of spectral count data in label-

free shotgun proteomics. Molecular and CellularProteomics 7, 2373-2385

142 Asara, J.M. et al. (2008) A label-free quantificationmethod by MS/MS TIC compared to SILAC andspectral counting in a proteomics screen.Proteomics 8, 994-999

143 Kirkpatrick, D.S., Gerber, S.A. and Gygi, S.P. (2005)The absolute quantification strategy: a generalprocedure for the quantification of proteins andpost-translational modifications. Methods 35,265-273

144 Gerber, S.A. et al. (2003) Absolute quantification ofproteins and phosphoproteins from cell lysates bytandemMS. Proceedings of the National Academyof Sciences of the United States of America 100,6940-6945

145 Lange, V. et al. (2008) Targeted quantitative analysisof Streptococcus pyogenes virulence factors bymultiple reaction monitoring. Molecular andCellular Proteomics 7, 1489-1500

146 Martin, D.B. et al. (2008) MRMer, an interactiveopen source and cross-platform system for dataextraction and visualization of multiple reactionmonitoring experiments. Molecular and CellularProteomics 7, 2270-2278

147 Deutsch, E.W., Lam, H. and Aebersold, R. (2008)PeptideAtlas: a resource for target selection foremerging targeted proteomics workflows. EMBOReports 9, 429-434

148 Vizcaino, J.A. et al. (2009)A guide to the proteomicsidentifications database proteomics datarepository. Proteomics 9, 4276-4283

149 Craig, R., Cortens, J.P. and Beavis, R.C. (2004) Opensource system for analyzing, validating, and storingprotein identification data. Journal of ProteomeResearch 3, 1234-1242

150 Suzuki, A. et al. (2010) Identification ofmelanoma antigens using a serological proteomeapproach (SERPA). Cancer Genomics Proteomics 7,17-23

151 Sun, S. et al. (2010) Circulating lamin B1 (LMNB1)biomarker detects early stages of liver cancer inpatients. Journal of Proteome Research 9, 70-78

152 Chen, L. et al. (2010) Enhanced detection of earlyhepatocellular carcinoma by serum SELDI-TOFproteomic signature combined with alpha-fetoprotein marker. Annals of Surgical OncologyMar 31; [Epub ahead of print]

153 Long, L. et al. (2010) Pattern-based diagnosis andscreeningofdifferentiallyexpressed serumproteinsfor rheumatoid arthritis by proteomicfingerprinting. Rheumatology International Mar25; [Epub ahead of print]




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

154 Kurian, S.M. et al. (2009) Biomarkers for earlyand latestagechronicallograftnephropathybyproteogenomicprofiling of peripheral blood. PLoS One 4, e6212

155 Helgason, H.H. et al. (2010) Identification of serumproteins as prognostic and predictive markers ofcolorectal cancer using surface enhanced laserdesorption ionization-time of flight massspectrometry. Oncology Reports 24, 57-64

156 Li, Y.L. et al. (2010) Identification of glia maturationfactor beta as an independent prognostic predictorfor serous ovarian cancer. European Journal ofCancer 46, 2104-2118

157 Sui, W. et al. (2010) Proteomic profiling of renalallograft rejection in serum using magnetic bead-based sample fractionation and MALDI-TOF MS.Clinical and Experimental Medicine Apr 8;[Epub ahead of print]

158 Wang, C.L. et al. (2009) Discovery ofretinoblastoma-associated binding protein 46 as anovel prognostic marker for distant metastasis innonsmall cell lung cancer by combined analysis ofcancer cell secretome and pleural effusionproteome. Journal of Proteome Research 8,4428-4440

Further reading, resources and contacts

Protein and peptide separation techniquesDelahunty, C. and Yates, J.R., 3rd (2005) Protein identification using 2D-LC-MS/MS. Methods 35, 248-255

Describes in detail the strong cation exchange/reversed-phase method that is widely used for 2D LC-MS/MSanalysis.

Fang, Y., Robinson, D.P. and Foster, L.J. (2010) Quantitative analysis of proteome coverage and recovery ratesfor upstream fractionation methods in proteomics. Journal of Proteome Research 9, 1902-1912Rigorously compares the three most commonly used protein-level first-dimension separation techniques CF,IPG and GeLC.

Biochemical fractionation methodsMarkham, K., Bai, Y. and Schmitt-Ulms, G. (2007) Co-immunoprecipitations revisited: an update on

experimental concepts and their implementation for sensitive interactome investigations of endogenousproteins. Analytical and Bioanalytical Chemistry 389, 461-473Deals with immunoprecipitations and MS to identify interaction partners and mentions several validguidelines to avoid possible pitfalls.

Rogers, L.D. and Foster, L.J. (2009) Phosphoproteomics – finally fulfilling the promise?Molecular Biosystems 5,1122-1129An in-depth review about the development of phosphoproteomics in recent years.

Dengjel, J., Kratchmarova, I. and Blagoev, B. (2009) Receptor tyrosine kinase signaling: a view from quantitativeproteomics. Molecular Biosystems 5, 1112-1121An overview about recent approaches to study receptor tyrosine kinase signalling using MS-basedproteomics.

Quantitative approachesElliott, M.H. et al. (2009) Current trends in quantitative proteomics. Journal of Mass Spectrometry 44, 1637-1660

A thorough and well-written synopsis of the current state of quantitative proteomics, which discusses thestrengths and weaknesses of the methods in more depth than they have been covered here.

Targeted analysis: SRMLange, V. et al. (2008) Selected reaction monitoring for quantitative proteomics: a tutorial. Molecular System

Biology 4, 222An informative tutorial explaining many aspects of SRM, including transition design and optimisation as wellas the application of SRM for quantitative proteomics.

Yocum, A.K. and Chinnaiyan, A.M. (2009) Current affairs in quantitative targeted proteomics: multiple reactionmonitoring-mass spectrometry. Briefings in Functional Genomics and Proteomics 8, 145-157A useful review on SRM that includes advice on method development as well as a section on other MS-based targeted approaches.

(continued on next page)




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

Further reading, resources and contacts (continued)

Issaq, H.J. and Veenstra, T.D. (2008) Would you prefer multiple reaction monitoring or antibodies with yourbiomarker validation? Expert Reviews of Proteomics 5, 761-763An engaging discussion of the possible applications of SRM in clinical biomarker validation assays.

Bioinformatic analysisMalik, R. et al. (2010) From proteome lists to biological impact – tools and strategies for the analysis of large MS

data sets. Proteomics 10, 1270-1283An excellent review of available tools to extract biologically relevant information from large proteomicdatasets.

Wang, M. et al. (2008) Label-free mass spectrometry-based protein quantification technologies in proteomicanalysis. Briefings in Functional Genomics and Proteomics 7, 329-339An overview of available label-free MS-based approaches to protein quantification.

Features associated with this article

FiguresFigure 1. Workflow of typical MS-based proteomic experiments.Figure 2. Workflows for implementation of stable isotope labelling via metabolic (SILAC) and chemical (iTRAQ or

TMT) methods.Figure 3. Mode of operation of selected reaction monitoring (SRM).Figure 4. Discovery of the potential biomarker ITIH3 for early detection of gastric cancer.

TablesTable 1. MS quantitation techniques.Table 2. Examples of selected reaction monitoring (SRM) resources.Table 3. Recent examples of clinical applications of proteomics.

Citation details for this article

Geraldine M. Walsh, Jason C. Rogalski, Cordula Klockenbusch and Juergen Kast (2010) Mass spectrometry-based proteomics in biomedical research: emerging technologies and future strategies. Expert Rev. Mol.Med. Vol. 12, e30, September 2010, doi:10.1017/S1462399410001614




Mas

ssp

ectrometry-b

ased

proteomicsin

biomed

ical

rese

arch

:em

ergingtech

nologiesan

dfuture

strategies

Mass spectrometry-based proteomics in biomedical research:...

Documents

Transcript of Mass spectrometry-based proteomics in biomedical research:...