Advances in LC-MS/MS-based glycoproteomics: getting closer to system-wide site-specific mapping of...

Biochimica et Biophysica Acta 1844 (2014) 1437–1452

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bbapap

Review

Advances in LC–MS/MS-based glycoproteomics: Getting closer tosystem-wide site-specific mapping of the N- and O-glycoproteome

Morten Thaysen-Andersen ⁎, Nicolle H. PackerBiomolecular Frontiers Research Centre, Department of Chemistry and Biomolecular Sciences, Macquarie University, Sydney, Australia

Abbreviations:CID, collision induceddissociation; ConAdissociation; EIC (or XIC), extracted ion chromatogram; ESInance; Fuc, fucose; Gal, galactose; GalNAc, N-acetylgalactosraphy; HPLC, high performance liquid chromatography; IRMMALDI,matrix assisted laser desorption ionization;Man,maNeuAc,N-acetyl-5-neuraminic acid; Q-TOF, quadrupole-tim⁎ Corresponding author at: Department of Chemistry and

E-mail address: [email protected] (M. Tha

http://dx.doi.org/10.1016/j.bbapap.2014.05.0021570-9639/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t
a r t i c l e i n f o
Article history:Received 2 March 2014Received in revised form 23 April 2014Accepted 5 May 2014Available online 12 May 2014

Keywords:GlycoproteomicsGlycopeptideGlycoproteinMass spectrometryProtein glycosylationGlycoproteome

Site-specific structural characterization of glycoproteins is important for understanding the exact functional rele-vance of protein glycosylation. Resulting partly from themultiple layers of structural complexity of the attached gly-cans, the system-wide site-specific characterization of protein glycosylation, defined as glycoproteomics, is still farfrom trivial leaving the N- and O-linked glycoproteomes significantly under-defined. However, recent years haveseen significant advances in glycoproteomics driven, in part, by the developments of dedicated workflows and effi-cient sample preparation, including glycopeptide enrichment and prefractionation. In addition, glycoproteomics hasbenefitted from the continuous performance enhancement and more intelligent use of liquid chromatography andtandemmass spectrometry (LC–MS/MS) instrumentation and awider selection of specialized software tackling theunique challenges of glycoproteomics data. Together these advances promise more streamlined N- and O-linkedglycoproteomeanalysis. Tangible examples include system-wideglycoproteomics studies detecting thousands of in-tact glycopeptides from hundreds of glycoproteins from diverse biological samples. With a strict focus onthe system-wide site-specific analysis of protein N- and O-linked glycosylation, we review the recent advances inLC–MS/MS based glycoproteomics. The review opens with a more general discussion of experimental designs inglycoproteomics and sample preparation prior to LC–MS/MS based data acquisition. Although many challengesstill remain, it becomes clear that glycoproteomics, one of the last frontiers in proteomics, is gradually maturing en-abling a wider spectrum of researchers to access this new emerging research discipline. The next milestone in ana-lytical glycobiology is being reached allowing the glycoscientist to address the functional importance of proteinglycosylation in a system-wide yet protein-specific manner.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction to system-wide site-specific analysis ofprotein glycosylation

1.1. Structure and function of protein N- and O-glycosylation

Protein glycosylation is the covalent attachment of complex carbohy-drates or oligosaccharides (here collectively called glycans) to specificamino acid residues of the polypeptide backbone of proteins. The biosyn-thetic machinery of mammals primarily allows the attachment of glycansto asparagine (N-glycosylation) and serine/threonine (O-glycosylation)residues thereby forming two major classes of protein glycosylation.These types i.e. N-GlcNAc and O-GalNAc (mucin-type) glycosylation,which are both the focus of this review, are synthesized in a non-

, concanavalinA; CSF, cerebrospinalflu, electrospray ionization; ETD, electronamine; GlcNAc, N-acetylglucosamine;PD, infraredmultiphoton dissociation;nnose;MRM,multiple reactionmonitoe-of-flight; RP, reversed phase; SPE, solBiomolecular Sciences, Macquarie Un

ysen-Andersen).

template drivenmanner by a spectrumof glycosylation enzymes throughdifferent routes in the secretory pathway. The N-linked glycans,which in mammals are usually restricted to occupy asparagine res-idues in NXS/T, X ≠ P consensus sequences (sequons), consist of acommon chitobiose core (Man3GlcNAc2) from which a variety ofmonosaccharides and other glycan modifications may be added tothe non-reducing termini in specific linkage configurations [1].The status of the cellular glycosylation machinery and the natureof the proteins undergoing glycosylation together determine therepertoire of glycans being presented on the protein carriers thuscreating the important features of cell- and protein-specific glyco-sylation [2]. N-linked glycans are usually larger by mass and volumethan their O-linked counterparts, which, in contrast, are more

id;DDA, data dependent acquisition;DIA, data independent acquisition; ECD, electron capturetransfer dissociation; FDR, false discovery rate; FT-ICR, Fourier transform ion cyclotron reso-HCD, higher-energy collisional dissociation; HILIC, hydrophilic interaction liquid chromatog-LC–MS/MS, liquid chromatography tandemmass spectrometry; LTQ, linear trap quadrupole;ring;MSn,mass spectrometry to thenth power; nETD, negative electron transfer dissociation;id phase extraction; TMT, tandemmass tag; UPLC, ultra-high pressure liquid chromatographyiversity, NSW 2109, Australia. Tel.: +61 2 9850 7487; fax: +61 2 9850 6192.

http://crossmark.crossref.org/dialog/?doi=10.1016/j.bbapap.2014.05.002&domain=pdf

http://dx.doi.org/10.1016/j.bbapap.2014.05.002

mailto:[email protected]

http://dx.doi.org/10.1016/j.bbapap.2014.05.002

http://www.sciencedirect.com/science/journal/15709639

1438 M. Thaysen-Andersen, N.H. Packer / Biochimica et Biophysica Acta 1844 (2014) 1437–1452

frequently attached in the serine- and threonine-rich regions of mucinand mucin-like proteins. Several types of non-mucin O-glycosylationare also known to exist in mammals including O-GlcNAcylation [3], O-mannosylation [4] and O-fucosylation [5], but are not addressed in thisreview. The general structures, biosynthetic pathways and associateddisorders ofN- andO-linked glycosylation have been reviewed elsewherein detail and readers are encouraged to use these resources for a moregeneral introduction to protein glycosylation [1,6].

As expected from the conservation of mammalian protein glyco-sylation sites and glycan structures and the significant amount ofcellular energy utilized to ensure proper regulation of the biosyn-thetic machinery, N- and O-linked glycans have repeatedly beenshown to be important for generating andmaintaining the structuraland functional integrity of their carrier proteins [1]. Protein N- andO-glycosylation is also involved in development processes [7,8], inbacterial binding to the host [9] and to sustain the normal functionof individual cells and indeed also of the whole tissue, organ and or-ganism [10]. Finally, glycosylation is a highly regulated protein mod-ification in cells and is known to be affected by various pathologiesincluding, but not restricted to, cancer [11,12], inflammation [13],Alzheimer disease [14], multiple sclerosis [15], cystic fibrosis [9] rel-ative to cells exposed to ‘normal’ physiology and homeostasis.

1.2. Multiple structural layers of protein glycosylation

Interpretation of the glycosylation code has been intriguing to pastgenerations of glycobiologists. The ability to accurately characterizethe structure of glycoproteins is of importance to unravel the diversefunctions of glycans. Better understanding of protein glycosylation willfacilitate the development of next generation of glycoprotein-basedtherapeutics [16]. However, relative to simple protein modifications,protein glycosylation is more challenging to characterize due to thesub-stoichiometry and heterogeneity formed by multiple layers ofstructural diversity at the primary structural level of the glycan. Varia-tions in the monosaccharide compositions, overall topology/branchingpatterns and linkage types are common features creating a spectrumof closely related glycan species. The amount of information needed tounambiguously characterize a glycoprotein is consequentlymuch largerthan for unmodified proteins and proteins modified by other chemicalgroups e.g. phosphorylation and methylation. Even the most detailedcharacterization studies usually only capture part of the glycoproteinstructure leaving some structural aspects unknown. Specific structuralfeatures may be assumed/predicted from the established ‘synthetic re-striction rules’ derived from the well-studied biosynthetic machineryof mammalian protein glycosylation [17]. It is important to stress thatglycobiological function often can be learnt from less-than-completestructural information of glycoproteins. In other cases, the functionalunderstanding of glycobiology is hidden in the fine details of the glyco-protein structure supporting more thorough glycoprotein characteriza-tion requirements.

1.3. Bottom-up site-specific glycoprofiling of isolated glycoproteins

Ever since glycosylated proteins were discovered there has been adesire to accurately determine the glycoprotein structure. Historically,characterization of isolated glycoproteins of relatively high amounts(Npico-nanomolar) has been achieved using an array of analytical tech-niques to obtain in-depth information of the glycoprotein structure [18].This has indeed advanced our understanding of the structure/functionrelationship of many glycoproteins [19–22]. The majority of thesetechniques are based on so-called bottom-up principles where gly-coproteins are hydrolyzed and the fragments including monosaccha-rides, released glycans and proteolytically generated glycopeptides,are analyzed separately [23]. The structural information from the indi-vidual fragments is then pieced together to learn about the structureof the entire glycoprotein. Modern techniques are centered around

high performance liquid chromatography (HPLC) [24,25] and liquidchromatography conjugated on-line with mass spectrometry (LC–MS)[26–33] of fluorescently labeled and native glycans or intact glycopep-tides. Only few analytical techniques are presently capable of handlingand generating information-rich data from intact glycoproteins, inparticular when analyzed in complex mixtures e.g. from 2D gels [34],leaving the conventional bottom-upmethods still the preferred analyt-ical approach.

Analysis of glycans released from their carrier proteins represents auseful and relatively easy experimental approach with which to gener-ate in-depth glycan structural information such as knowledge of themonosaccharide composition, branching topology and stereoisomericlinkage of the monosaccharide sequence. However, although profilingof released glycans can generate a holistic view of the global cellglycome or cellular sub-glycomes, glycomics alone intrinsically lacksthe ability to provide information about the carrier protein and thespecific attachment site of each glycan. In contrast, glycopeptides (andglycoproteins) provide an avenue for establishing direct evidence ofthe connectivity between the glycan and the polypeptide backbone. Asdiscussed later, less information about the glycan structure can usuallybe obtained from glycopeptide-based approaches. Knowledge aboutthe specific glycan attachment site becomes particularly needed forglycoproteins containing multiple glycosylation sites or when mixturesof glycoproteins are analyzed. Established protocols for site-specificglycosylation analysis of isolated proteins are available [33,35];however, due to the extensive variety in glycoprotein architecturemostly arising from variations in the polypeptide sequences and posi-tions of sequons relative to available proteolytical sites such methodsmust be customized to be suitable for the glycoprotein of interest.Often glycan- and glycopeptide-based analyses are carried out in paral-lel to obtain complementary and in-depth site-specific information ofisolated glycoproteins [19,33,36–38]. Site-specific characterization andrelative quantitation of approximately 170 mammalian glycoproteinshas been recently curated by the authors based on published literature[2]. From this we estimate that glycan structures have been partly orcompletely characterized in a site-specific manner for approximatelyonly 400–500 of the ~10,000 predicted human glycoproteins [39].Together with the knowledge that protein glycosylation varies dramati-cally in a spatial and temporal manner, it is thus clear that the vastmajority of the human (let alone the whole mammalian) glycoproteomeremains uncharacterized. Hence, glycoscientists have long argued theneed for the development of analytical techniques capable of performingsite-specific analysis of protein glycosylation at the larger-scale toincrease the coverage of the glycoproteome.

1.4. Defining glycomics and glycoproteomics

Glycomics is commonly defined as the conjugate-unspecific analysisof the glycome at a system-wide level. This includes, but is not limitedto, the global analysis of glycans released from proteins derived fromisolated cells, tissues or body fluids. The term glycomics often implicitlyrefers to the analysis of the N- and/or O-glycomes, but should morecorrectly be specified in case-by-case studies to avoid confusion withthe global analysis of other glycoconjugates including glycolipids, GPIanchors or free glycans. Several semi-automated sensitive glycomicsworkflows are now available for low/medium through-put analysis ofthe N- and O-glycomes [23,35,40,41]. An excellent overview of recentdevelopments in glycan analysis by mass spectrometry was recentlypublished [42].

Glycoproteomics is the protein-specific counterpart to glycomics de-fined as the site-specific analysis of the glycoproteome at the system-wide level. Thus, glycoproteomics yields information about the proteincarriers, the glycan attachment sites and the structure and occupancyof the glycan. Glycopeptides, and much less frequently intact glycopro-teins, are the target analytes for glycoproteomics. As discussed in thiswork, techniques andworkflows for glycoproteomics remain immature

1439M. Thaysen-Andersen, N.H. Packer / Biochimica et Biophysica Acta 1844 (2014) 1437–1452

relative to the glycomics counterparts. This is, in part, due to the highermolecular complexity of the glycopeptides (two conjugated hetero-geneous biomolecules i.e. glycans and peptides) and, in part, due tothe fact that glycopeptides are only weakly detected in MS. TheweakerMSdetection of glycopeptides is a result of their sitemicrohetero-geneity, reducing the copy number of each individual glycopeptidespecies, and the lower MS signal intensity of glycopeptides compared tounmodified peptides [43]. Nonetheless, acknowledging the potentialwealth of biological information hidden in the relatively untouched N-andO-linked glycoproteomes, recent years have seen a tremendous effortto facilitate more efficient, large-scale site-specific analysis of proteinglycosylation [44–49].

Here we will review the recent advances in LC–MS/MS-basedglycoproteomics that has enabled tangible examples of in-depth analysisof theN- andO-glycoproteomes. The focuswill be strictly on the technicaladvances relating to the system-wide site-specific analysis of theglycoproteome. We will, thus, aim to deviate from other excellentreviews covering glycoprotein/glycopeptide analysis of less complexmixtures [50–54]. The main focus will be on advances in LC–MS/MSbased acquisition of data on the glycoproteome. The review will openwith amore general discussion of the importance of experimental designand sample preparation in glycoproteomics. The aim is partly to highlightthe analytical challenges of the field and to emphasize the importance ofmaintaining a holistic view of glycoproteomics experiments bycareful consideration of all aspects of the approach, and not just theLC–MS/MS-based data collection. Techniques for site-specific analysisof single/few glycoproteins and glycomics type approaches will not becovered in this review. In addition, proteomics-type approaches dealingwith the qualitative and quantitative analyses of formerly glycosylatedproteins following deglycosylation, designed to learn about theproteome-wide attachment sites of glycans [55–62], but ignoring theglycan structure itself, will not be covered in this review as it does notfall under our definition of glycoproteomics. We do recognize thatalthough site identification studies per se do not qualify under thedefinition of glycoproteomics, they still provide useful proteome-wideinformation of the attachment site provided correct experimentaldesign is applied [63]. This is particularly true when only partial degly-cosylation is used: the part of the glycan structure (e.g.N-linked GlcNAcor O-linked GalNAc) that remains linked to the identified peptides mayprovide some valuable site-specific information including knowledge ofwhich proteins carry core N-glycan fucosylation [56] and which pro-teins carry mucin core-1 type O-glycosylation [64]. In addition, severalimpressive studies have used “SimpleCells” to genetically engineercells to produce homogeneous O-glycosylation with the aim to identifya large number of human O-glycosylation sites using LC–MS/MS tech-nologies [65–68]. Although such studies are not described here in detaildue to the removal of the intrinsic biological heterogeneity of proteinglycosylation, the technology is clearly applicable to study glycosylationin these systems.

2. Experimental design in glycoproteomics

2.1. Experimental design and introduction of wanted/unwanted bias

Appropriate experimental designs are the key to test glycoproteome-centric (null) hypotheses. Similar to experimental designs in otherrelated areas such as proteomics and glycomics, the sampling and num-ber of biological and technical replicates are of immense importance toachieve a confident rejection/acceptance of the proposed hypothesis.Glycoproteomics deviates from proteomics and glycomics by the factthat N- and O-linked glycopeptides span a much larger spectrum ofstructural heterogeneity compared to unmodified peptides or releasedglycans. The heterogeneity is an inherent challenge for the samplepreparation e.g. glycopeptide enrichment/prefractionation. If a globalglycoproteome analysis is desired, the sample preparation should ideallyleave the entire set of glycosylated peptides in their original molar ratios

without introducing bias towards some species over others. The sameprinciple is true for the MS detection: an ideal setup would provide MSsignal responses reflecting the accuratemolar quantities of the glycopep-tides in the mixture. However, considering the extensive glycopeptideheterogeneity encountered in biological samples covering large di-versity in structure, mass, volume and charge, completely unbiasedglycoproteomics is difficult to achieve. As discussed below, some re-searchers may decide to target specific sub-glycoproteomes (e.g. se-creted/soluble or membrane glycoproteomes) or only subsets ofglycoproteins showing common glycan determinants. In thesecases a bias is deliberately introduced in the sample work-up toobtain the sub-glycoproteome of interest. Focusing on specific sub-glycoproteomes as opposed to the global glycoproteome may berelevant in some cases, but one should ensure that the selectedsub-glycoproteome accurately reflects the research questions beingaddressed without potential loss/gain of information. Taken together,the risk of introducing unwanted bias in glycoproteomics experimentsseems higher than for related research disciplines emphasizing theneed for careful consideration of all aspects of the experimental design.

Based on published glycoproteomics experiments, an outline of a‘generic’ workflow is presented in Fig. 1. Using this representation,glycoproteomics experiments can be crudely divided into the follow-ing components: 1) crude extraction of proteins from biologicalsamples including tissues, cells, conditioned media or bodily fluids,2) enrichment or prefractionation of glycoproteins, 3) proteolysisusing a suitable protease to generate mixtures of appropriatelysized glycopeptides and unmodified peptides, 4) enrichment ofglycopeptides to reduce their under-representation in the peptidemixture or prefractionation into multiple fractions to lower thepeptide complexity, 5) separation and detection of glycopeptidesusing LC–MS/MS, 6) quantitative and qualitative interpretation ofglycopeptide-based LC–MS/MS data and 7) statistical analysis toallow acceptance/rejection of a proposed hypothesis. Several com-ponents are heavily sample- or approach-dependent (i.e. point 2and 4) and should be considered in accordance with the proposedresearch question. This section will deal with aspects relating tothe experimental design including the hypothesis generation andsome general considerations for glycoproteomics. The next sec-tions focus on the sample handling and preparation and the subse-quent LC–MS/MS data acquisition and interpretation.

2.2. Structural information obtained from glycoproteomics experiments

Before initiating an ambitious set of glycoproteomics experiments, itis important to understand which glyco-related research question(s)may be addressed using glycoproteomics strategies. In other words,what of kind of structural information can be uniquely obtained usingglycoproteomics. This will ensure an experimental design which ismore likely to provide a confident acceptance/rejection of the proposedhypothesis.

Specifically, glycopeptide-based glycoproteomics can yieldinformation of several features of the glycoproteins in a given mix-ture: 1) qualitative description of the entire set of glycan structureslinked to defined sites on specified peptides/proteins (microhetero-geneity), 2) quantitative description of the relative molar distribu-tion of the observed glycans on the individual glycosylation sites(glycoprofiling), and 3) quantitative description of the total glycanoccupancy of the individual sites (macroheterogeneity). As de-scribed above (Section 1.2), glycans can be characterized at differentstructural resolutions. At present, glycoproteomics yields ‘low-reso-lution’ structural data on the glycan i.e. monosaccharide compositionand, at best, some topology/branching information. In contrast, thepeptide and, hence the corresponding protein, is usually confidentlyidentified using high-mass accuracy detection of the peptide precur-sor and fragment ions. However, glycoproteomics often relies on sin-gle peptide-based protein identification thus increasing the risk of

Crude protein mixture

Crude glycoprotein mixture

Mixture of glycopeptides / non-glycosylated peptides

Glycopeptide mixture

(Tissue, cell, media, body fluid)

Prefractionation/enrichment [3.2] High pH RP (C18) HPLCLectin HPLC

SCX HPLCHILIC SPE

‘Rev. glycoblot.’ (hydrazide)TiO2SPE

Proteolytic digestion [3.1] Specific proteases(e.gTrypsin/Lys-C)

Non-specific proteases(e.g. Pronase/Proteinase K)

Prefractination [3.2]

(Optional/sample dependent)

(Optional/sample dependent)Lectin HPLC

SCX HPLC

RP (C4) HPLC

SEC HPLC

Protein extraction/denaturation [3.1]

LC-MS/MS acquisition Data interpretation

Fragmentation: CID, HCD, ETD [4.4]

Res: High/High, High/low [4.2]

MS/MS trigger: DDA (int), DDA (ox.) [4.5]

LC Flow: Cap,Nano [4.1]

LC:RP (C18)/HILIC/PGC[4.1]

MS: Orbitrap, Q-TOF [4.2] New software for glycoproteomics: [4.6]Byonic, SweetHeart,GlycoFragWork, SweetSEQer,Glycopeptide Finder, GlycoPeptideSearch

Biological sample 2(Tissue, cell, media, body fluid)

Data comparison

Qualitative comparisonRel. quant. comparison

HomogenizationCell lysis

PrecipitationMWCO filter

(Sample dependent)

* Potential introduction of strong wanted/ unwanted bias

*

*

*

Settling on an experimental design [2.1]

Glycoproteome-centric research questions [2.3] Choosing the appropriate sub-glycoproteome [2.4]

Structural information obtained using glycoproteomics [2.2]

Site-specific glycoprofiles

Site-specific glycoprofiles

Sam

ple

han

dlin

g/ s

amp

le p

rep

arat

ion

Experimental design

Ion/polarity: ESI, MALDI, +/- [4.3]

Pre

-an

alys

is

LC

-MS

/MS

dat

aco

llect

ion

an

din

terp

reta

tio

n

Biological sample 1

Fig. 1.Outline of a generic workflow of glycopeptide-based LC–MS/MS-centric glycoproteomics experiments. Typically two or more samples of biological origin (here represented by redand green text/lines) are processed and analyzed individually (in parallel) using glycoproteomics. The resulting peptide glycoprofiles are then qualitatively or (relative) quantitativelycompared. The workflow is broken down into the individual components (gray boxes) forming the experimental design. The corresponding numbered sections discussing the individualcomponents are indicated. Within each component there is a variety of techniques (light gold boxes) creating multiple permutations of a glycoproteomics workflow. The preferredmethods according to the published glycoproteomics literature, if any, are underscored.


ambiguous identification. Amino acid sequence resolution is oftenneeded to confidently identify O-glycan sites whereas N-glycansites usually need less evidence due to their predictable sequonlocation (see Fig. 2 for a general representation of N- and O-glycopeptides and examples of their structural features). The siteoccupancy can only be evaluated if no prior glycopeptide enrichmentis performed since the matching non-glycosylated peptides aredepleted in this process.

2.3. Generating glycoproteome-centric research questions

The site-specific nature of glycoproteomics allows the researcher toinvestigate how cellular, molecular or systemic (whole organism) alter-ations affect protein glycosylation at the global (system-wide), yetprotein-specific level. Proteome-wide changes in protein glycosylationindicate regulation of the glycosylation machinery e.g. altered expres-sion/activity of specific glycosylation enzymes. Such regulation mayyield important cues into the direct or indirect relationship withthe perturbed physiology being investigated. In contrast, single proteinchanges in the glycoproteome indicate regulation at the protein-specific level, possibly involving changes in the following featuresof the individual protein: gene expression, ER/Golgi trafficking route/

rate, protein splice variants/polymorphisms/degradation/truncation,cross-talk with other post-translation modifications and/or changes inthe tertiary and quaternary structure. Needless to say, if the investigat-ed mixture of glycoproteins originates from multiple cell types the in-terpretation of any cellular regulation of the glycosylation machinerybecomes significantlymore difficult since each cell typemay be unique-ly regulated [69]. Hence, knowledge about the cellular heterogeneityand the cellular origin of the identified glycoproteins are invaluablewhen designing glycoproteomics experiments.

Simple mapping initiatives of the N- and O-linked glycoproteomes ofunperturbed systems are valuable considering the limited coverage anddepth of the mammalian glycoproteomes. However, most researchersare interested in performing comparative glycoproteomics studies toidentify altered protein glycosylation in the context of biomarker discov-ery or to gain mechanistic insight into specific pathogeneses [59,70–73].Comparative glycoproteome studies can be of different natures coveringi) qualitative assessment of the glycoprotein micro-heterogeneity bythe presence/absence of glycopeptides fromall proteinswithin each sam-ple, ii) relative quantitative assessment defining the relativemolar distri-bution of related glycopeptides within each sample, and iii) absolutequantitative assessment describing the absolute molar amounts of aspecific glycopeptide by comparison to a labeled internal glycopeptide

H3N+- -COOHH+

N-glycanprovides localhydrophilicity for efficientHILIC SPE enrichment

Peptide provides local hydrophobicityfor efficient RP-(C18) LC retention

Multiple basic groups ensures efficient positive ion mode ionization. Multiple charges brings ions into favourable m/z range and is advantageous for ETD

H+HH+ H+

N++

Steric hindrance and electrostatic repulsion from O-glycan may leave proximal proteolytical sites uncleaved

c/z-ions from ETD/ECD MS/MS important for peptide identification

c1 c2 c3

z13 z12

Y1

Abundant Y1-ion following CID MS/MS can be used for CID-basedMS3-peptide identification

Crucial c/z-ions from ETD/ECD MS/MS for site determination of O-glycan-

c10c9b5 b6

y8 y7

b7

b/y-ions from HCD (or Q-TOF) MS/MS important for peptide identification but N-and O-glycans are detached

Sialic acid containing glycopeptides can be enriched using TiO2 SPE or hydrazide (‘reverse glycoblotting’)

z5

B2 B3’

Y6’

Abundant B- and Y-ions from CID-MS/MS important to identify monosaccharide composition and substructures/topology

Y5

Oxonium ions (e.g. m/z 204.087) from HCD MS/MS valuable to identify glycopeptide spectrum and trigger precursor ETD

Fig. 2. Typical tryptic N- and O-linked glycopeptide encountered in glycoproteomics. Important physicochemical and structural features in the context of sample prepara-tion and LC–MS/MS-based glycoproteomics are highlighted.


analog of known amount. The former approaches (i–ii) are mostlyused due to their biologically informative and label-free nature.Substantial insight into how the glycosylationmachinery affects the indi-vidual glycoproteins can be obtained by determining the alterations inpeptide-specific glycoprofiles when two or more samples are compared.Absolute quantitative measures of glycopeptides (iii) are particularlyuseful for in-depth characterization of specific glycoproteins, for examplewhen used for therapeutics analysis.

Glycan occupancy (macroheterogeneity) is also known to affect theprotein solubility, stability, antigenicity, clearance rate, half-life andin vivo activity [74]. Thus, defining and understanding changes occur-ring in the degree of occupancy of each individual glycosylation site ofspecific proteins are of biological importance. Finally it should be men-tioned that alterations of the abundance of specific glycoproteins (butnot necessarily its glycosylation pattern) can similarly be important inthe context of the studied physiology/pathogenesis. Such alterations inprotein abundance can be routinely determined on the proteomewide level using deglycosylated proteins or peptides located outside ofthe glycosylated region by established quantitative proteomicsmethods[75,76]. Hence, such techniques will consequently not be discussedhere.

Usually the desire to perform comparative glycoproteomics is basedon prior knowledge indicating the involvement of protein glycosylationin cause/effect mechanisms in the studied system. Examples includedifferential transcription of glyco-genes and epigenetics [72,77–79],regulation of the glycomes or proteomes detected by LC–MS/MS [56,79,80], differential staining of tissues using lectins [79,81,82] or meta-bolic labeling [83], or prior measurements indicating that the activities

of the glycosylation related enzymes [84] and/or the nucleotide–sugarsubstrates [85] are being regulated. As discussed later, prior knowledgeof the proteome(s) and glycome(s) being studied is of great value forthe interpretation of glycopeptide-based glycoproteomics data.

2.4. Choosing the appropriate sub-glycoproteome

Depending on the nature of the biological samples and the researchquestion being pursued, different sub-glycoproteomes may be of inter-est. Secreted glycoproteomes, defined as the entire set of glycoproteinssecreted from cells or tissues, are a natural choice when investigatingbodily fluids including tears, saliva and plasma/serum. As such, the se-creted glycoproteomes are ideal for biomarker studies where easy sam-pling and early disease detection are key features. In contrast, tissuesamples from biopsies/surgery are often rich in membrane glycopro-teins, which, because of their cellular anchoring, often can establish amore reliable connectivity between the molecular and cellular changesbeing investigated. Thus, the membrane glycoproteome may be thechoicewhen studyingdiseasemechanisms. Ideally healthy anddiseasedtissue from the same patient should be investigated to account forindividual-specific variations in glycosylation. However, obtaining suffi-cient starting material for such glycoproteomics approaches may be anobstacle. Cultured cells are a source of both secreted and membraneglycoproteins which can be extracted from the culture medium andfrom the cellmembrane fractions, respectively. Although showing unnat-ural protein glycosylation relative to cells in their native environment[11], the glycosylation machinery of cultured cells (immortalized or pri-mary cells) can be manipulated to study the fundamentals of protein


glycosylation and identify disease mechanisms at the protein-specificlevel using glycoproteomics.

Dictated by limitations of the analytical approaches or by the questionbeing addressed, the researcher may also want to restrict the focus to ei-ther the O- or the N-glycoproteome. Most developed glycoproteomicsworkflows, including the downstream LC–MS/MS acquisition and datainterpretation, inherently favor the analysis of N-glycoproteomes overO-glycoproteomes. In addition to the fact that researchers historicallyhave been somewhat biased towards analyzing N-glycosylation (due tothe availability of enzymes that release all N-glycans from their proteincarriers and that the biosynthetic machinery of protein N-glycosylationcan be manipulated in different ways), the reasons are multi-factorial:1) the limited number of glycosylation sites inmostN-glycoproteins pro-vides little restriction to protease digestions. In contrast, the heavily occu-pied O-glycan clusters often found in mucin type O-glycoproteins candramatically shield the protein from proteolytic digestion. Hence,obtaining O-glycopeptides suitable for LC–MS/MS detection can be amajor challenge, 2) N-glycosylation is predominantly found in predict-able consensus sequences (sequons), which reduces the need for singleamino acid residue sequence resolutionwhen determining the glycan at-tachment site. The multiple O-glycan moieties occupying Ser/Thr resi-dues in a given O-glycopeptide, in contrast, require extensive peptidecoverage to unequivocally determine the glycan-peptide connectivity;this is still difficult to technically achieve, 3) the common chitobiosecore shared by allN-glycans and their well-understood biosynthetic con-straints generally allows easier topology assignment of the glycanmoietyof N-glycopeptides compared to O-glycopeptides that have beendescribed to have at least eight cores, and finally 4) the larger size N-glycans (generally 7–15 monosaccharide residues) relative to O-glycans(generally 1–6 monosaccharide residues) creates a more dense clusterof local hydrophilicity of the N-glycopeptides (Fig. 2). This physicochem-ical property of N-glycopeptides is useful to discriminate these speciesfrom their non-glycosylated counterpart using hydrophilicity-based en-richment. Together this means that N-glycoproteomics workflows todate aremoremature. However, as discussed in later sections, workflowsfor system-wide site-specific analysis of the O-glycoproteome have alsobeen presented. As an alternative approach, although not offering thesite-specific knowledge, mucin O-glycosylation can instead be studiedby mature glycomics approaches [35]. Considering the cluster-based de-sign of mucin O-glycosylation, mapping the mucin-type O-glycans in amore site-unspecific (but still protein- and cell-specific) manner maystill be biologically informative. The unique challenges of O-linked glyco-protein analysis have recently been reviewed [86].

3. Sample preparation and general considerationsin glycoproteomics

3.1. Extraction of glycoproteins and initial sample preparation

When designing a glycoproteomics experiment it is most importantto ensure that enough glycoprotein can be derived from the biologicalsamples of interest to enable sufficient glycopeptide signal responsefor all technical replicates in the downstream LC–MS/MS analysis. Sub-stantial losses in analyte yield may be expected due to the multi-stepglycoprotein sample processing. Additionally, the sub-stoichiometry ofglycopeptides relative to unmodified peptides resulting from glycopro-tein macro- and micro-heterogeneity and their lower MS signal re-sponse in LC–MS has to be taken into account [43]. Together thismeans that a glycoproteomics experiment usually needs more proteinstarting material (N100 μg) than conventional proteomics (b100 μg).

Needless to say, the initial sample preparation is heavily dependenton the nature of the biological starting material (e.g. tissue, cells, andbody fluids) and the sub-glycoproteome of interest (global, secreted,membrane glycoproteome). Techniques such as homogenization, celllysis, precipitation, liquid–liquid extraction, density centrifugation, mo-lecular weight cut-off filtering, lectin enrichment/separation and size

exclusion HPLC, methodswhich are all used in conventional proteomics[87,88], are usefulmethods to crudely extract, isolate and prefractionatemixtures of intact glycoproteins in glycoproteomics. However, the indi-vidual steps must be carefully considered to avoid significant glycopro-tein losses and introduction of unwanted bias (see Section 2.1).

Prior to the glycopeptide generating proteolysis, glycoprotein dena-turation is needed tomaximize the digestion efficiency [89]. Incompleteproteolytical digestion will further reduce the stoichiometry of the gly-copeptides and significantly complicate the relative quantitation of theobserved glycopeptide variants. As for conventional proteomics, trypsinis a favored proteolytical enzyme because of its specificity for basicamino acid residues and since it can access most proteolytic sites evenwhen relatively large N-glycans are occupying sites proximal to argi-nine/lysine residues [33]. However, some steric hindrance in particularfrom ‘bulky’ N-glycans e.g. core-fucosylated N-glycans may still be ob-served [90]. In addition, O-glycosylation rich Ser/Thr regions in the vi-cinity of arginine/lysine residues carrying sialo-O-glycans (Fig. 2) maymask the tryptic cleavage sites (personal unpublished observation).The analytical chemist should also remember that other enzymes suchas Lys-C, Lys-N, Arg-C and Glu-C may produce more suitable glycopep-tides for LC–MS/MS in particular for acquisitions involving electrontransfer dissociation (ETD) which benefit from a higher number of pos-itive charges carried by larger peptides [91]. Some researchers favor theuse of non-specific proteases including Pronase, Proteinase K and otherenzymes for the site-specific analysis of protein glycosylation of smalleryet more heterogeneous glycopeptides [32,92–95].

3.2. Glycopeptide enrichment and prefractionation prior to LC–MS/MS

Even when glycoproteins are separated from unmodified proteinsprior to digestion, the proteolytic glycopeptides will often be maskedin the LC–MS/MS analysis by the unmodified peptides originatingfrom the non-glycosylated regions of the same glycoproteins. Hence,glycopeptide enrichment is usually undertaken to lower the dynamicrange prior to LC–MS/MS in glycoproteomics. As such, glycopeptide en-richment can be viewed as one of the absolutely critical steps for en-abling in-depth glycoproteome analysis of complex biological samples.The glycopeptide enrichment can be performed using a range of HPLCor solid phase extraction (SPE) based techniques. Glycopeptide enrich-ment strategies have beenwell-covered by other recent reviews [33,44,50,58,59,96–100]. Thus, we will only briefly touch on this importanttopic.

For qualitative glycoproteomics in general and quantitativeglycoproteomics in particular it is ideally desirable to achieve unbi-ased enrichment of the entire set of glycopeptides relative to theunmodified peptides in a sample without altering their originalmolar ratios (see Section 2.1). Common physiochemical propertiesof the structurally heterogeneous glycopeptides can be used forisolation purposes. Since all N-glycopeptides share the feature ofhigh hydrophilicity localized around the N-glycan moiety (Fig. 2),HILIC is in our opinion the preferred tool to globally enrich intactglycopeptides from complex peptide mixtures. Although HILICchromatography forms a quite large family of related separationtechniques primarily as a result of the numerous stationary phasesavailable e.g. neutral polar, anionic and zwitterionic [101], the retentionmechanism of all HILIC approaches, in the context of glycoproteomics,can generally be described as hydrophilic partitioning of the glycopep-tides into the watery layer around the hydrophilic solid phase [102].Importantly, this partitioning is a fully reversible retention and glyco-peptides can be competitively eluted with hydrophilic solvents i.e.water. However, aqueous salt solutions may be needed to weaken anyadditional electrostatic interactions [103]. We have recently optimizedthe enrichment efficiency of the original HILIC SPE protocol [104,105]further by using a mobile phase containing an ion-pairing reagentwhich creates an enhanced hydrophilicity difference between glycosyl-ated and non-glycosylated peptides [106].


Although lectins selectively and reproducibly retain some glycopro-teins, other glycoproteins carrying the same repertoire of glycans, maynot be retained [107]. This seems particularly true when separationsare performed on complex protein mixtures. Together, the low affinity,broad specificity, and the shared protein/glycan epitope recognitionlimit the capacity of lectins to perform global glycopeptide enrichment.Instead lectins are more valuable for prefractionation steps [108–110]and for visualizing glyco-determinant expression using histochemicalstaining techniques, micro-arrays and lectin blotting [72,73,81,82,111].Lectins can also be used in glycoproteomics studies where wantedbias is purposely introduced to analyze specific sub-glycoproteomes.

In addition, glycopeptides from specific sub-glycoproteomesmay beenriched using other methods. For example, sialylated glycopeptides ofthe sialoglycoproteome can be reversibly captured by TiO2 [112] or hy-drazide (reverse glycoblotting) [113,114]. Following release from suchstationary phases, the native/desialylated sialoglycopeptides can thenbe analyzed using LC–MS/MS-based glycoproteomics.

Even efficiently enriched glycopeptide fractions isolated from verycomplex peptide mixtures of biological starting material may stillcontain thousands of analytes. Depending on the peak capacity of theLC separation and the MS detection speed, resolution, sensitivity anddynamic range, this complexity may still cause analytical problemsleaving a significant proportion of the analytes in themixture undetect-ed. Introducing an orthogonal off-line separation step consisting ofstrong-cation exchange [115], HILIC [37,38] or neutral/high-pH-RP[116] before or after the glycopeptide enrichment is a rational approachto lowering the analyte complexity. Consequently, the glycoproteomecoverage is increased, but at the severe expense of increased samplehandling/instrument/data analysis time. If an off-line separation stepis introduced prior to the glycopeptide enrichment step, added benefitsinclude the elimination of denaturing agents, surfactants, salts andlipids which may interfere and lower the enrichment efficiency.

4. LC–MS/MS data acquisition and interpretation in glycoproteomics

This section will discuss the recent improvements in the LC–MS/MS-based acquisition of glycoproteomics data. The individual componentsof the LC–MS/MS experiment will be covered separately includingonline-LC separation strategies, modern mass spectrometers used inglycoproteomics, ionization and quantitation aspects, fragmentationtechniques and specialized MS/MS triggering strategies. Finally, we willbriefly touch on the very recent developments in the dedicated bioinfor-matics tools useful for glycoproteomics data interpretation.

4.1. Online-LC separation strategies

As for conventional proteomics, the vast majority of glycoproteomicsstudies base their online-LC separation on C18-reversed phase(RP) HPLC/UPLC separation principles (see Table 1 for publishedglycoproteomics studies). In fact, it was not possible to find large-scaleglycoproteomics studies in the literature where any other LC-separation technique was utilized. However, it should be mentionedthat HILIC [117,118] and PGC [93–95] are indeed capable of separat-ing N- and O-linked glycosylated peptides in MS compatible solventsand column dimensions. Hence, HILIC and PGC may prove valuablefor some glycoproteomics applications considering the hydrophilicnature of glycopeptides. These stationary phases may be particularlyvaluable for smaller glycopeptides derived from non-specific prote-ase digestion which often do not have sufficient retention on RP-LCto enable separation and MS detection.

So why are RP-C18-based separations heavily favored inglycoproteomics? The already introduced and readily availabletechnology in most proteomics laboratories and its compatibilitywith other lines of proteomics research provide one obvious an-swer. In addition, RP has superior peak capacity for glycopeptidescompared to other LC-separation techniques including HILIC and

PGC (personal unpublished observation). The high peak capacityis in part facilitated by the availability of sub-2-micron RP-C18

particles made from traditional silica, monolithic silica rods orcore–shell technology. The separation is further enhanced by thenear-zero-dead volume column packing technology in fused silicatips of integrated ESI emitters. As such, glycoproteomics hasbenefitted immensely from years of proteomics-type LC optimizationwhich has enabled ultra-high separation power of (glyco)peptides. Nar-row5–15 second elution bands of glycopeptidesmay be obtained in or-dinary acetonitrile gradients. Preferred flow-rates for glycoproteomicsare on the nano-scale (200–300 nl/min) [62, 113, 119–124] matchingwell with the 50–100 μm wide inner-diameter fused silica columnsthus yielding high detection sensitivity. Nano-flow ESI-MS technolo-gies, once limited to expert proteomics laboratories, are now morewidely accessible to research communities partly by the introductionof commercial products including EasySpray (Thermo), nanoHPLCChip systems (Agilent) [125] and Triversa NanoMate (Advion) [126].However, glycoproteomicsmay still be successfully performed at higherflow rates albeit at lower sensitivity [114,127]. LC additives such asDMSO [128] and ammonium formate [129], to enhance the LC peakcapacity without compromising MS sensitivity, have further im-proved the separation efficiency of RP, but are yet to be introducedinto LC–MS/MS-based glycoproteomics.

Another advantage of RP-LC for glycoproteomics type studies is thatrelated glycopeptides sharing the same peptide moiety but conjugatedwith a variety of neutral glycans elute closely together just before thenon-glycosylated peptide variant [130,131]. Sialylated glycopeptidestend to elute at a slightly lower or higher acetonitrile concentration de-pending on the ion pairing reagent in the mobile phase [132]. In addi-tion to promoting a more confident structural identification byenabling tandemMS of the individual eluting species, the LC separationof sialoglycopeptides from their asialo-counterparts is useful to detectany in-source fragmentation, which may skew the quantitation. How-ever, the main advantage of the RP retention behavior is that the co-eluting glycopeptide isoforms can be easily identified based on theirm/z differences on the MS1-level (e.g. ΔHexNAc corresponds to Δm/z203.079 (+1), 101.540 (2+) and 67.693 (3+)). In addition, the co-elution is advantageous for quantitation purposes, particularly formore complex peptide mixtures, by limiting the interference by thestronger signals from unmodified peptides on the signal from the glyco-peptide cluster.Mostmodernmass spectrometers are sufficiently fast toisolate, fragment and detect the resulting fragment ions of multiple gly-copeptide precursor ions eluting simultaneously. However, the limitedcapacity for separating related glycopeptides by RP-LC naturally leavesthe relative retention time of the individual glycopeptides comparablyless informative for structural characterization than on other stationaryphases including PGC where efficient separation of related glycopep-tides and glycans can be obtained [133,134].

4.2. Modern mass spectrometers

Two types of mass spectrometers are heavily preferred inglycoproteomics (Table 1): Variants of Orbitrap hybrid mass spectrom-eters (Thermo) i.e. Orbitrap Elite (Hybrid Ion Trap-Orbitrap), LTQOrbitrap Velos Pro and LTQ Orbitrap XL or Q-TOF typemass spectrome-ters i.e. Synapt (Waters) or QStar (AB Sciex) [135]. The commonalitiesbetween these two generic types of mass spectrometers are high massaccuracy (low-sub ppm) and high resolution (N20,000) of precursorand fragment ions (provided ions are detected in the Orbitrap and notthe LTQ) and the capability of performing higher-energy collisional dis-sociation (HCD). HCD has certain advantages for glycopeptide analysisas discussed below. In particular the Orbitrap technology platform pro-vides additionally superior speed and sensitivity compared to othertypes of mass spectrometers which are crucial characteristics to tacklethe complexity and dynamic range of the glycoproteome. The latestmodel of the Orbitrap Fusion Tribrid mass spectrometer, which has

Table 1Chronological order of published glycopeptide-based glycoproteomics-style studies outlining the experimental design i.e. biological starting material, proteolytic enzyme, prefractionation/glycopeptide enrichment methods, LC–MS/MS instrumen-tation, fragmentation and the obtained results i.e. number of non-redundant glycopeptides, sites and proteins identified. The methods used for data processing are listed. Studies analyzing glycopeptides from isolated glycoproteins or where theglycopeptide heterogeneity has been significantly reduced/eliminated prior to LC–MS/MS analysis are not included in this table.

Biological sample analyzed/targetsub-glycoproteome (protease)

Glycoprotein or glycopeptideprefractionation/glycopeptideenrichment

Online-LCseparation(flow rate)

MS detection(resolution)a

MS/MSfragmentation

Non-redundantglycopeptides (sites)/glycoproteinsidentifiedb

Qualitative/rel. quantitativestudy (method forLC–MS/MS data processing)

Other information/comments Ref(year)

Murine (epi)dermis/membraneN-glycoproteome (trypsin)

Con A lectin/RP-HPLC(100 fractions)

No onlineLC-separation

MALDI TOF (low/low) Higher energyCID (TOF/TOF)

84 (19)/15 Rel. quantitative (Mascot, manual) N-glycome mappedsimultaneously

[131](2005)

Human serum/secretedN-glycoproteome (trypsin)

None/none Cap-RP (C18)(2 μl/min)

QSTAR Q-TOF (low/low) CID 109 (23)/15 Rel. quantitative (Mascot, manual) Medium/low confidence of CID-based assignment of peptide/site

[127](2008)

Human CSF/secreted N- andO-sialoglycoproteome (trypsin)

None/periodate oxidation,hydrazide capt., desialylation

Nano-RP (C18)(200–300nl/min)

LTQ-FTICR (high/low,high/high)

CID (DDAtop-5 MS3),ECD

N-linked: 88 (36)/23O-linked: 38 (44)/22

N-linked: rel. quant. O-linked:qualitative (Manual, Mascotv2.2 of CID MS3)

Desialylated N- andO-glycoproteomes studied

[119](2009)

Bovine serum/secretedN-glycoproteome (trypsin)

Jacalin lectin HPLC/ConA lectin HPLC

Nano-RP (C18)(250 nl/min)

Q-TOF premier (high/high) LTQ-Orbitrap (high/low)

CID, CID/ETD 48 (17)/13 Qualitative (manual, ProteinProspector v.5.2.1, BioWorksv.3.3.1)

O-glycopeptides furthervalidated using partialdeglycosylation

[124](2009)

Mice sera/secretedN-sialoglyco-proteome (trypsin)

None/NaO4 oxidation,hydrazide capt., PA-labeling

Cap-RP (C18)(30 μl/min)

4000 Q-Trap (low/low) CID (MRM) 67 (36)/26 Quantitative (Mascot v2.1, manual,ProteinPilot v3.0, MultiQuant v1.0)

Optimized “reverseglycoblotting” method

[114](2010)

C. jejuni/membraneN-glycoproteome (trypsin)

None/ZIC-HILIC SPE,soybean lectin affinity SPE,graphite

Nano-RP (C18)(200 nl/min)Cap-RP(5 μl/min)

LTQ-Orbitrap XL (high/high), HCT-IT (low-low)

CID/HCD,CID/ETD

130 (75)/Not stated Qualitative (Mascot v2.2, manual) C. jejuni has noN-glycan micro-heterogeneity

[120](2011)

Human urine/secreted N- andO-sialoglycoproteome (trypsin)



LTQ-FT-ICR (high/low,high/high)

CID (MS3),ECD

N-linked: 58 (25)/17O-linked: 63 (49)/40

N-linked: rel. quant. O-linked:qualitative (Manual, Mascot v2.3of CID MS3)

Desialylated N- andO-glycoproteomes studied

[113](2012)

N. gonorrhoeae/membraneO-glycoproteome (trypsin/AspN)

Immunoaffinity/none Nano-RP (C18)(200 nl/min)

LTQ-Orbitrap XL(high/high)

CID/HCD,HCD

18 (10)/9 Qualitative (ProteomeDiscoverer v2.0 (Sequest))

Search conducted only towardsdi-N-acetylbacillosamine(mass: 228.110 Da)

[172](2012)

Human endothelium/secretedN-glycoproteome (trypsin)

Con A lectin/ZIC-HILIC SPE Nano-RP (C18)(300 nl/min)

Orbitrap Elite (high/high,high/low)

HCD/ETD,HCD-pd-ETD

131 (118)/59 Qualitative (Manual, Byonic(ProteinMetrics))

Proteome identified,glycopeptide TMT-labeled

[121](2013)

Tomato/cell wall N-glycoproteome(trypsin)

Con A lectin/ZIC-HILIC HPLC Nano-RP (C18)(300 nl/min)

Synapt Q-TOF (high/high) CID, PIDproduct ionmode

36 (26)/24 Qualitative (Analyst v1.4.2,BioAnalysis v1.4, GlycoMod,manual)

Site occupancy also measured [123](2013)

Human CSF/secretedO-glycoproteome (trypsin)



LTQ-FTICR, Orbitrap Velos,Orbitrap XL (high/low,high/high)

CID, HCD, ETD(DDA,CID-MS2/MS3)

Not stated (106)/51 Qualitative (Manual, MascotDaemon v2.3.0)

Desialylated O-glycoproteome(core 1 structures identified),some sites still unknown

[148](2013)

Murine synaptosome/membraneN- and O-glycoproteomes (tryp-sin)

WGA lectin/high pH RP HPLC RP (details notstated)

LTQ-Orbitrap Velos(high/low)

ETD N-linked: 2080(678)/375 O-linked:463 (190)/122

Qualitative (ProteinProspector v5.9.0, manual)

Several types of O-linkedglycosylation identified andnon-consensus N-sites

[171](2013)

Rat brain/membraneN-glycoproteome (trypsin)

High-pH RP HPLC/ZIC-HILIC SPE Nano-RP (C18)(250 nl/min)

LTQ Orbitrap Elite(high/high)

CID/HCD, ETD 863 (276)/161 Qualitative (manual, byonic(ProteinMetrics))

Glycome + proteome identifiedseparately

[116](2013)

Human serum/secretedN-glycoproteome (trypsin)

Top-7 depletion/none Nano-RP (C18)(350 nl/min)

LTQ-Orbitrap Velos(high/high, high/low)

CID/HCD,CID/HCD/ETD

103 (53)/33 Qualitative (GlycoFragWork) Description of bioinformaticstool for glycopeptide ID + FDRassignment

[122](2014)

a Where no indication is given, ESI interfaces for introduction of ions into the mass spectrometer were used.b The number of non-redundant glycopeptides, sites and proteins covered was either derived directly from the text of the specific publication or extracted as accurately as possible from published data.

1444M.Thaysen-Andersen,N

.H.Packer

/Biochimica

etBiophysicaActa

1844(2014)

1437–1452


seen another round of performance-enhancing hardware iteration[128], and the Q Exactive Hybrid Ion Trap-Orbitrap mass spectrometerfrom the previous generation remain to be tested in the context ofglycoproteomics. However, their performances in conventional proteo-mics type benchmarking exercises indicate that the glycoproteomics re-searcher could be an immediate beneficiary of their implementation.

4.3. Ionization and quantitation aspects

Several otherMS features are important to considerwhen performingglycoproteomics studies. With the exception of an early glycoproteomicsreport utilizing MALDI-based ionization [131], glycoproteomics re-searchers have exclusively used ESI to ionize and introduce the glycopep-tide ions into the mass spectrometer in an online-LC-compatible manner(see examples in Table 1). Prior to MS/MS fragmentation events, it is cru-cial to obtain solid MS1-level information about the glycopeptide precur-sor ions. Such ‘survey scans’ are typically carried out at ample time andion accumulation to enhance the resulting signal-to-noise ratio. Detectionshould be performed at high mass accuracy/resolution to establish theexact monoisotopic mass (i.e.m/z and charge state) of the glycopeptides.Establishing the exact masses of the precursors is crucial for glycopeptideidentification purposes in glycoproteomics analyses of complex mixtures(discussed below). Since theMS1-scan is commonly used to establish thequantitative relationship between the related co-eluting glycopeptides itis critical to minimize the in-source fragmentation [94]. In addition, itcan be advantageous to focus on a slightly higher m/z region e.g. m/z700–2000 [116]. The lack of charge-accumulating basic functional groupsdirectly on the N- andO-glycanmoieties means thatN- and O-linked gly-copeptides generally will take up fewer positive charges (i.e. protons) permass than unmodified peptides, leaving their correspondingm/z higher.

The consensus for glycoproteomics experiments is to perform theMS event, and subsequent MS/MS events, in positive polarity mode.Positive polarity ensures good ionization efficiency and, hence, sensitiv-ity of tryptic glycopeptides containing at least one basic amino acid res-idue and an amino-group terminal (see Fig. 2 for example). In addition,several studies have shown that the relative molar ratios of relatedglycopeptides sharing the same peptide moiety can be estimated withreasonable accuracy based on their signal strength (signal height orEIC/XIC) in positive polarity mode in both MALDI and ESI-MS [33,136].By the same token, it should be stressed that glycopeptides are severelyunder-represented compared to the non-glycosylated and deglycosyl-ated counterparts when equimolar amounts are analyzed in positiveESI andMALDI-MS [43]. This bias should be adjusted for when glycosyl-ation site occupancy is evaluated. Few fundamental studies have report-ed that detection of glycopeptides is enhanced in negative polaritymode [137,138]. However, the use of negative mode ionization forlarge-scale quantitative glycoproteomics needs further investigation toreveal the true potential.

Whereas high mass accuracy in the MS1-events is a requirement forglycoproteomics, Orbitrap users can decide to detect the resulting glyco-peptide fragment ions in the Orbitrap at high resolution/mass accuracy(high/high) or in the LTQ at lower resolution/mass accuracy (high/low).The latter is faster and more sensitive, but comes at the cost of increasedFDRs in the glycopeptide identification due to the lower resolution/massaccuracy. High/low or high/high acquisitionmodes are seemingly equallypreferred approaches in glycoproteomics (Table 1).

4.4. Fragmentation techniques

Glycoscientists can now choose from a range of fragmentationmethods to dissociate positively charged glycopeptide ions usingmodernmass spectrometers. Fragmentation schemes include dissociationmethods based on ion/neutral interactions i.e. collisional induced dissoci-ation (CID) and higher-energy collisional dissociation (HCD), ion/electronand ion/ion interactions i.e. ECD and ETD (collectively called activatedelectron dissociation (ExD)) and ion/photon interactions i.e. IRMPD

and UV photodissociation [139]. Some fragmentation techniques areinstrument-specific e.g. ECD (FT-ICR) and HCD (C-trap). Negativelycharged glycopeptide ions have the potential to be fragmented usingelectron detachment dissociation (EDD) and negative ETD (nETD), butapplications to glycopeptides and indeed glycoproteomics are just begin-ning to surface [139]. The fundamental principles, behavior and utility ofthe ensemble of fragmentation techniques have been well describedwith respect to glycopeptides [44,47,50,52,59,99,100,139]. Hence,we will only briefly cover fragmentation approaches in the contextof glycoproteomics.

The compilation of glycoproteomics studies (Table 1) shows a strongbias towards three dissociation techniques: i) lower-energy CID gener-ated from linear (LTQ) and 3D (Paul) ion traps, ii) higher energy-styleCID (alternatively called HCD when performed in C-traps) obtained onQ-TOF, Orbitrap (C-trap), and TOF/TOF platforms, and iii) ETD featuredby some ion traps and hybrid-type instruments, in particular Orbitraps,and less available for Q-TOFs [140]. Information-rich and complementa-ry structural data of glycopeptides can be obtained using these threefragmentation types, thus, supporting their combined use [139]. Directcomparisons of MS/MS spectra acquired with CID, HCD and ETD onthe same glycopeptide precursor have been published [19,47,50,99,120,139]. Simplified, their key features for glycopeptide characteriza-tion/identification can be described as follows: i) lower energy CID frag-mentation predominantly causes cleavages of the glycosidic linkages ofthe glycopeptides generating B/C- and Y/Z-ions yielding information ofthe monosaccharide composition and the general topology of the N-glycan (see Fig. 2 for examples), but little information on the peptidebackbone, ii) higher-energy CID fragmentation (and HCD) generateshigh abundant low-mass oxonium- and B/C-ions and some Y/Z-ionsfrom glycosidic cleavages. In addition, series of b/y-ions from peptidebackbone fragmentation, albeit without the glycan attached, are pro-duced in this fragmentation scheme. Thus, HCD is useful to searchglycoproteome-wide MS/MS data-sets for glycopeptide containingspectra and identify the peptide carrying moieties, but without localiz-ing the specific site, iii) ETD fragmentation causes predominantly pep-tide backbone fragmentation in the form of c-/z•-ion series whilemaintaining integrity of the glycan–peptide-linkage. ETD is consequent-ly useful for both identification of the peptide moiety of glycopeptidesand site-determination. However, ETD needsmostly triply, and prefera-bly quadruply and higher charged positive ions, to yield information-rich fragment spectra. Multiplexed isobaric tandem mass tag (TMT)can be used to achieve a slightly higher average charge state of the gly-copeptide precursor which is beneficial for ETD fragmentation and si-multaneously useful for quantitative comparison of glycopeptidesbetween samples [126].

4.5. Specialized MS/MS triggering strategies

The available multiple fragmentation options leave the analyticalchemist with some flexibility in the LC–MS/MS-based acquisition ofglycoproteomics data. The fragmentation techniques have been used in-dividually in separate LC–MS/MS runs. Increased requirements of instru-mentation time and sample amounts are naturally the consequences ifmultiple individual fragmentation techniques are used to obtain thedata. However, the approach maintains a relative faster MS duty cyclesince no switching between fragmentation methods is needed. Fastduty cycles are crucial to facilitate collection of multiple MS data pointsover the short time-span of the sharply eluting glycopeptides. Multipledata points, in turn, ensure more accurate EIC-based quantitation. Plentyof examples of glycoproteomics type studies using single fragmentationevents for glycopeptide characterization are available including CID-based fragmentation to identify N-glycopeptides of human serum [127],higher energy CID (TOF/TOF)-basedN-glycoprofiling ofmurine epidermis[131] and ETD fragmentation to characterize mucin-type O-glycosylatedpeptides [133,141]. The multiple injections and sample re-analysis canbe avoided by alternating between the dissociation methods within the


same LC–MS/MS run, but at the expense of longerMS duty cycles. Severalexamples of glycoproteomics studies using alternating fragmentationmethods can be identified in literature including the use of alternatingCID and HCD (and separately ETD) to identify N-glycopeptides from therat brain [120], alternating CID and ETD to identify N-glycopeptides[142,143] or O-glycopeptides [144].

The precursor selections used in these examples are all based on theprecursor MS signal intensities from an initial MS1 event of each dutycycle. This data-dependent acquisition (DDA) and precursor selectionis then followed by multiple MS/MS events using alternating fragmen-tation techniques. An advantage is that exactly the same precursor isfragmented under different dissociation schemes, which is not guaran-teed if the same sample was re-analyzed by multiple injections usingLC–MS/MS with single (non-alternating) dissociation methods. Otherglycoproteomics specific DDAMS/MSmethods have been presented in-cluding HCD-product ion-triggered-ETD activation of N-glycopeptides[145,146]. Here, the abundant HexNAc oxonium ion at m/z 204.087 inintensity-dependent HCD MS/MS was used to identify N-glycopeptideprecursor ions ‘on-the-fly’ and only trigger ETD fragmentation forthese precursors. This approach combines the features of abundantoxonium ions resulting from the C-trap HCD fragmentation and the ca-pability of low m/z range detection at high mass accuracy in theOrbitrap. The benefits include improved dynamic range and duty cyclerelative to a traditional alternating HCD/ETD approach. Another dedi-cated DDA MS/MS method for glycoproteomics includes an MSn-typeapproach taking advantage of the abundant Y1 ions originating fromCIDMS/MS fragmentation ofN-glycopeptides (see Fig. 2) [147]. This re-sulted in the identification of 101 N-glycopeptides carrying 32 N-glycans distributed on 12 N-glycosylation sites from 12 mouse serumproteins. Similarly, an automated CID-MS2/MS3 approach was used tocharacterize 106 HexHexNAc (presumably core 1 type) O-linked glyco-sylation sites of cerebrospinal fluid (CSF) glycoproteins [148].

As opposed to DDA, it should be mentioned that data-independentacquisitions (DIAs) strategies in LC–MS/MS based experiments are be-coming more popular. DIA approaches include MS(E), all-ion-fragmentation, SWATH, multiplexedMS/MS and parallel reaction mon-itoring, which have all been applied to proteomics type experiments[149]. However, DIA approaches still need proper evaluation in aglycoproteomics context to demonstrate its potential. Other innovativestrategies for precursor selections in glycoproteomics have been used.Froehlich and colleagues presented a logical LC–MS/MS-based identifi-cation (classification) of glycopeptide precursor ions from crude pep-tide mixtures based on a discriminatory mass defect of glycopeptides.This approach takes advantage of the higher proportion of oxygenatoms in glycopeptides giving them a slightly lowermass defect relativeto unmodified peptides [150]. Such classified glycopeptide ions whichwere left unfragmented in the first LC–MS/MS analysis were targetedspecifically by an inclusion list in the subsequent analysis. High specific-ity and sensitivity of the classifier were shown by detecting N-glycopeptides from the human urinary proteome. Finally, it is relevantto mention that detection and indeed quantitation of specific glyco-peptides of known glycoproteins also can be made directly fromcomplex biological peptide mixtures without the need for detectingthe entire complement of glycopeptides in a sample. This targeted‘non-glycoproteomics’ approach, which is most useful for addressinghighly protein-specific questions about protein glycosylation, can be per-formed without any prior glycopeptide enrichment or prefractionation.The potential of this approachwas recently shown using anMSnmultiplereaction monitoring (MRM) workflow targeting the detection of specificO-glycopeptides from hemopexin and sex hormone binding globulin ina crude serum sample [151].

4.6. New software for glycoproteomics data interpretation

Once acquired, the LC–MS/MS glycoproteomics data needs to beinterpreted to capture as much information as possible. Informatics in

the broad field of protein glycosylation has been reviewed [42,152–154]. In addition, recent reviews have expertly covered the bioinfor-matics tools currently available forMS-based glycoproteomics [155–158].Readers are kindly referred to these resources for background introduc-tion and overview of bioinformatics in glycoproteomics. Here we willonly briefly draw attention to a few very recent tools and novel conceptsin thefield of glycoproteomics informatics,whichhavenot been reviewedpreviously. The following free or commercial programs have recentlyshown their ability to handle glycoproteomics data: SweetHeart [147],Byonic (ProteinMetrics) [159], GlycoFragWork [122], SweetSEQer [160],Glycopeptider Finder [94,95] and GlycoPeptideSearch (GPS) [37,38].These tools span different styles of data-input (e.g. vendor specificity,fragmentation/acquisition types) and address a wide range of data pro-cessing options (e.g. identification, quantitation) and output styles. Theirindividual advantages and limitations will not be compared here as theyare not within the scope of this review.

Taking the advances in bioinformatics into consideration, it is inter-esting to note that all glycoproteomics studies presented to date haveextensively used manual data pre-/post-processing and spectral inter-pretation and validation (Table 1). Thus, the present bioinformaticstools appear to aid and facilitate some aspects of the glycoproteomicsexperiment, but fail to provide a packaged solution from raw data gen-eration to peptide glycoprofile comparison between samples. It is ex-pected that the assisting bioinformatics tools will continue to get morewidespread in the future andwill lower the burden of manual data pro-cessing. However, until we are capable of fully understanding and opti-mizing the multiple fragmentation options for glycopeptide analysis,and implementing such knowledge into sophisticated informaticstools, manual annotation will remain as a necessary quality control tovalidate computer program-based identifications. As for other biomole-cules, implementation of false discovery rate (FDR) based glycopeptideidentification to accurately evaluate the quality of automated assign-ments is crucial to reduce the manual workload. Following confidentidentification, the next requirements are integration of EIC-based quan-titation and designing of GUIs to provide an easy overview of the site-specific glycoprofiles at the proteome-wide level from multiple sam-ples. These developments should be facilitated by drawing parallels tothe more developed informatics field in proteomics. In conclusion,based on the volume of very recent field-specific software develop-ments it is natural to envision that further bioinformatics developmentsin academic and commercial arenas will continue to complement theadvances in sample preparation and LC–MS/MS acquisition to furthermature glycoproteomics.

5. Examples of state-of-the-art glycoproteomics studies

Thorough screening of the literature reveals that a limited number ofglycoproteomics experiments have been published. The experimentaldesign and the generated data/results of these publications are summa-rized in Table 1. Several other studies of a glycoproteomics character/potential have been published, but not included here due to the low/medium-complexity of the starting glycoprotein mixture [94,95,146,156,161–164]. In addition, glycoproteomics experiments solely focusingon O-GlcNAcylation [165–167] and O-mannosylation [168] of proteinsare excluded from this review since the focus is restricted to N- and(O-GalNAc mucin-type) O-linked glycosylation. It should be stressedthat the LC–MS/MS technology and system-wide glycoproteomicsreviewed in this work are capable of studying (and discovering) newtypes of protein glycosylation e.g. O-GlcNAc phosphorylation [169],tyrosine O-linked glycosylation [65–67,170,171].

With the exception of two studies, the glycoproteomics experi-ments generally reported in the order of 20–150 non-redundant N-and O-linked glycopeptides derived from 10 to 60 glycoproteins[113,114,119–124,127,172]. Below we will briefly describe tworecent glycoproteomics studies which confidently identifiedan order of magnitude more glycopeptides/glycoproteins from


complex biological samples. As such, these state-of-the-art studies dem-onstrate the present-day potential of glycoproteomics and are indicativeof the future direction of the field. By the same token, it is emphasizedthat other analytical workflows for glycoproteomics may have similarpotential and capabilities, but their application to achieve this level ofglycoproteome coverage remain unpublished.

5.1. Glycoproteomics case study: example 1

In 2013, Medzihradszky and colleagues were the first to present atrue large-scale site-specific characterization of both N- and O-linkedglycosylation from hundreds of proteins [171]. In this impressivepiece of work, a qualitative description of the membrane N- and O-glycoproteome of murine synaptosomeswas targeted, including the in-tracellular O-GlcNAcylated glycoproteins. Murine synaptosomes werepurified using density gradient centrifugation and pooled from severalmice corresponding to a total of 30 mg of starting material. Followingthorough (glyco)protein denaturation, tryptic digestionwas performed.The resulting peptidemixturewas desalted, lyophilized and enriched inan epitope-biased fashion using wheat germ agglutinin (WGA) lectinchromatography. WGA has reportedly broad affinity towards GlcNAc,NeuAc, Gal and GalNAc moieties. Glycopeptides were eluted usingGlcNAc-competition and prefractionated using high pH RP chromatog-raphy. Twenty fractionswere collected and analyzed by LC–MS/MS. On-line separation was performed over 90 min on a RP column and theglycopeptides were detected on an LTQ Orbitrap Velos mass spectrom-eter. DDA MS/MS ETD activation of the 10 most abundant precursorions with charge state plus three or higher was performed and theresulting c-/z•-ions were detected in the linear trap (high/low). Initialdata processing included the generation of peak listswith in-house soft-ware. Peak lists were then searched against aMus musculus database ofsemi-tryptic peptides using Protein Prospector v5.9.0 allowing a total ofthree variable modifications per peptide. The variable modifications in-cluded a pre-determined set of 31 glycan modification masses of Ser,Thr, Tyr and Asn. Mass accuracy thresholds were 5–10 ppm for precur-sor ions and 0.6 Da for fragment ions. The FDR of the site assignmentswas b1% based on a so-called SLIP scoring as well as the presence of aconsensus sequence in the identified N-glycopeptides, and manual in-spection/validation of fragment spectra for the identified O-glycopeptides. The set of glycanmodificationswas determined by an in-novative two-step iterative approach searching for unknown peptidemodifications of 100–3000 Da of previously identified proteins in thesample. Most of the 31 modification masses corresponded to knownN- or O-glycan molecular masses that accurately indicated the Asn- orSer/Thr/Tyr linkage; however, some linkage ambiguity was left forsome modification masses. In total, almost 2100 N-glycopeptides wereconfidently identified covering 678 glycosylation sites and 375 N-glycoproteins. Some non-consensus glycosylation sites e.g. NVV, NNV,NAG and NFQ were found to be N-glycosylated. Covering several O-glycan core types, a total of 463 O-linked glycopeptideswere identified from 190 unambiguously identified sites on 122 O-glycoproteins. An additional 152 sites were ambiguously identified.Although the work represents an outstanding analytical achievementseveral aspects have to be considered to enable a fair comparison toother work: i) the glycan assignments were based only on mass valuesof the peptide modification and no direct evidence on the composition,sequence or linkages of the monosaccharides in the glycan was obtain-ed as the approach relied solely on ETD MS/MS, ii) different cleavageforms of the sameglycopeptidewere included in the number of total re-ported identifications reducing the number of actual non-redundantidentified glycopeptides, iii) very high amounts of starting materialwere used thus increasing the number of identifications above whatmay be achievable from other biological systems, iv) extensiveprefractionation into 20 fractions increased the LC–MS/MS time dra-matically and v) the study was mostly qualitative and only included arelative quantitative analysis of a few glycopeptides from selected

glycoproteins of interest. In conclusion, this study represents an impres-sive system-wide ETD-centric glycoproteomics approach generatingthe largest amount of site-specific glycan information to date. Althoughthe site and protein identifications are of high confidence, the attachedglycans are blindly considered as a mass modification, which leaveslittle hypothesis-generating information for the glycobiologist who be-lieves that the structural aspects of the glycan are the key to under-standing their function, to build further work on.

5.2. Glycoproteomics case study: example 2

Simultaneously, Parker and co-workers published a differentglycoproteomics approach targeting the global rat brain membraneN-glycoproteome [116]. Starting from five milligram membrane-associated proteins extracted from homogenized brain tissue, compre-hensive protein denaturation and tryptic digestion were performed.Efficient global enrichment of N-glycopeptides was enabled usingzwitter-ionic (ZIC) HILIC-SPE. The enriched glycopeptides were off-lineprefractionated into 13 fractions using neutral pH RP-HPLC. The individu-al fractions were analyzed over 150min by nano-RP-LC–MS/MS using anLTQ-Orbitrap Velos for detection. DDA-based fragmentation wasperformed using CID (high/low), HCD (high/high) or ETD (high/high) ofthe three most intense precursors in the MS1-scans. HCD spectra werefiltered for oxonium ion containing signals i.e.m/z 204.087 at high massaccuracy. The glycopeptide enriched HCD spectra were searched againsta highly targeted database consisting of concatenated rat trypticpeptides known to be N-glycosylated based on previous identifica-tion of de-N-glycosylated peptides from the same samples and thepresence of consensus sequons. Searches were performed usingByonic (ProteinMetrics), a commercial search-engine enablingspecialized searches against a targeted set of defined N-glycan struc-tures as variable Asn modifications. The N-glycan structures presentin the mixture of proteins were pre-determined by state-of-the-artglobal glycomics analysis of the same sample [173]. Manual spectralinspection and the presence of specific oxonium ions were used toenhance the confidence of the glycopeptide assignment. This manualcuration was particularly required for glycopeptides containing twodHex (fucose) vs one NeuAc (Δm = 1 Da) residue where false mon-oisotopic peak picking (Δm= 1 or 2 Da) flagged by Byonic as a “pre-cursor off-set” may result in incorrect identification. Thecorresponding ETD and CID spectra were annotated by Byonic ormanually, respectively, and used to further enhance the confidenceof the peptide identification and to differentiate N-glycopeptidescontaining isobaric N-glycans e.g. outer-arm fucosylation versuscore fucosylation. In total, 863 non-redundant N-glycopeptideswere identified from 276 sites on 161 N-glycoproteins. The identifi-cations which were based on high mass accuracy assignment of aminimum of four HCD fragment ions yielded highly confidentmatches as evaluated by a FDR of 1.8%. The FDRwas assessed by reversingthe amino acid sequences (excluding the C-terminal Lys/Arg) of the pep-tide spectral matches. Although the number of identified glycopeptideswas lower relative to the identifications presented by Medzihradszkyand colleagues [171], more structural information on the attached N-glycans was obtained from the use of multiple dissociation methods.This included the sequential CID-based monosaccharide losses and theabundant glycan sub-structure-specific oxonium ions in the HCD spectra.ETD fragmentation provided little additional peptide information in thisstudy. In conclusion, the largest global coverage of the rat brain N-glycoproteome was achieved using HILIC (as opposed to lectin enrich-ment), prefractionation and state-of-the-art glycopeptide detection andidentification. However the study had a requirement for high startingamounts, long analysis/instrument time and expertise in the sam-ple handling, data acquisition and processing. Furthermore, the ac-curacy of FDR estimations using a reversed peptide database (asopposed to a reversed protein database) requires furtherevaluation.


Taken together, the two publications present experimentally andanalytically different, yet successful, system-wide glycoproteomicsstudies that showcase the present-day potential of the field. Countlesspermutations of theworkflows and applicationsmay potentially be pur-sued, considering the many variables in the experimental design(Fig. 1). Future more powerful glycoproteomics workflows will likelyevolve rapidly from those presented here that will further increase thecoverage of the glycoproteome in various systems. Until then, thediscussed studies serve the purpose of accurately reflecting the currentstatus, capacity and feasibility of present-day glycoproteomics.

6. Conclusion and outlook

Biologists and chemists alike acknowledge that studying protein gly-cosylation in a carrier-specific context is a crucial component of under-standing the exact structure/function relationships of glycoproteins. Forseveral decades reductionist protein chemists, depending on the avail-ability of a sufficient amount of the isolated protein, have been able toaccurately describe the site-specific distribution of glycans on specificglycoproteins using a myriad of classical protein chemical techniques.The system-wide site-specific analysis of the glycoproteome, definedas glycoproteomics, has until recently been out of reach.

The site-specific analysis of N- and O-linked glycoproteins, based onanalysis of proteolytically generated glycopeptides, has benefitted im-mensely from advancements of LC–MS/MS technologies in recent de-cades. Many of these advancements have been driven by technologyimprovements within the larger proteomics research field. Due to theunique analytical challenges encountered in glycopeptide-based ap-proaches,many of these technological improvements have only becomeuseful to the glyco-scientist after tailored alterations of the workflowsand tedious method optimization. As we become better at handling,separating and detecting glycopeptides, we are about to reach a newmilestone in analytical glycobiology. This review has clearly showedthat glycoproteomics workflows are maturing to gradually allowsystem-wide site-specific analysis of protein glycosylation.

Tangible examples among the recent literature are now available,describing glycoproteomics experiments where thousands of unique,intact glycopeptides from hundreds of protein carriers are confidentlyidentified at the site-specific level from complex biological samples. Im-portantly, the technology may yield clues to the functional relevance ofthe glycan heterogeneity observed on these sites. However, it is clearthat only the top of the glycoproteome ‘iceberg’ is covered to date. In ad-dition, several aspects of glycoproteomics are still highly manual andtime-consuming, demanding expert supervision.Without question, fur-ther technology improvements on all levels from sample preparationover LC–MS/MS acquisition to automated data interpretation and FDR-based evaluation, are needed to increase our depth and coverage ofthe glycoproteome. However, analogous to other ‘omics’ disciplines,such improvements can be expected to be developed and implementedrapidly.

Wewould like to emphasize that in addition to focusing on increasingthe glycoproteome coverage by increasing thenumbers of detected glyco-proteins, researchers should be encouraged to aim to maximize theamount of glycan structural information captured in glycoproteomicstype experiments. Hence, achieving system-wide yet protein- and site-specific knowledge of the glycan branching/topologies and glycosidiclinkages in addition to the current capacity to determine only mass-based monosaccharide compositions on their specific protein carriers,will introduce another dimension of structural resolution and insight.Furthermore, the accurate description of the relative molar distri-bution of glycoprotein micro-heterogeneity and site occupancy(macro-heterogeneity) are important quantitative features in functionalglycoproteomics since pathogenesis often is associated with a quantita-tive rather than qualitative alteration of protein glycosylation.

Initiatives to study protein glycoforms at the intact glycoprotein(top-down) level have clear advantages in terms of reduced sample

preparation and analysis time and improved quantitative accuracy[33]. However, considering the challenges still associated with LC-separation, MS ionization, fragmentation and data interpretation ofthe heterogeneous intact glycoproteins, particularly the larger multi-glycosylated proteins, such holistic approaches to glycoproteomicsmay still be years, if not decades, in the making. Irrespective ofthe vast analytical challenges, the benefits provided by top-downglycoproteomics are overwhelming and it is difficult not being excitedby its (theoretical) potential.

Taking the unique analytical challenges, structural information andresearch questions into account, glycoproteomics is emerging as aseparate research discipline within the larger field of structuralglycobiology. The volume of glycoproteomics literature in high-qualityjournals and the increasing attention it draws at international researchsymposia andworkshops show that glycoproteomics has reached a crit-icalmass of dedicated researchers.Meanwhile it is important to remem-ber that glycoproteomics has evolved from the established, hithertoseparate, disciplines of glycomics and proteomics by a desire to extractthe substantial amount of biological information which sits untouchedat their interface. In addition, glycoproteomics has close analytical tiesto the more traditional fields of protein and carbohydrate chemistry.Hence, researchers from all neighboring fields are expected to have aninterest in the new emerging research discipline of glycoproteomics.

As this new arm of analytical science emerges, intuitive and simpleglyco-centric nomenclature becomes important to describe and per-ceive specific and highly technical aspects accurately, in order to avoidconfusion and trepidation from non-experts with limited knowledgeof glycoproteomics. This is particularly relevant when considering theincreased number of both technology and biology oriented publicationsexpected in the maturation process of glycoproteomics. Another chal-lenge is that the glycoproteomics data description, submission andpresentation in most field-relevant journals are still based on the indi-vidual editor-, reviewer- and author-preferences. A wide diversity ofpresentation-styles and depth of detail may leave data, results and con-clusions of glycoproteomics papers harder to interpret. Realizing theneed for data standardization, some community consensus on thismatter has recently been reached [174] based on the proteomicsmodel. As for the proteomics counterpart [175,176], glycoproteomicswould benefit from rapid field-wide implementation of these guidelinesby the relevant journals.

We will finish this review reiterating our definition of the two keyterms in the field of glycoproteomics: We suggest that glycoproteomeis used solely to describe the total complement of glycosylated proteinsin a time- and space-defined environment, whereas glycoproteomics de-scribes the system-wide site-specific study of a defined glycoproteome.The latter includes studies where intact glycopeptides or glycoproteinsare used to obtain direct evidence for the attachment of specific glycanstructures to specified sites on defined protein carriers. It does not, inour opinion, encompass studies solely detecting glycosylation sites ofpreviously occupied (deglycosylated) peptides or the site-specific anal-ysis of glycoproteins isolated to partial/complete purity. Although suchstudies remain of high importance they aremore accurately categorizedunder other descriptors.

As glycoproteomics matures as an established discipline wheresystem-wide site-specific analysis of protein glycosylation becomes lesstime- and expertise-demanding, a range of exciting avenues for targetingpreviously unaddressed biological questions in a glycoprotein-specificcontext arises. Examples include obtaining mechanistic insight into howprotein-specific glycosylation on a whole cell/tissue/organ/organismlevel is regulated in pathogenic processes, studying the interplayand cross-talk between glycosylation and other post-translationalmodifications at the proteome-wide level, learning about system-widecause/effect mechanisms of protein glycosylation in relation to diseasesof interest, discovering candidate biomarker glycoproteins displaying en-hanced sensitivity and specificity characteristics and learning how thisheterogeneous post-translational modification can fine-tune protein


function. The next analyticalmilestone in analytical glycoscience iswithinreach, andwepredict that glycoproteomicswill bring unique insights intobiology; insightswhich can only be reachedwhen considering the proteinand the attached glycan as one complete biomolecule.

Acknowledgements

This work was supported by an Australian Research Council SuperScience Fellowship for M.T.-A. Dr. Benjamin L. Parker, Garvan Instituteof Medical Research, Sydney, Australia and Prof. Giuseppe Palmisano,Institute of Biomedical Sciences, Sao Paulo, Brazil are thanked for pro-viding valuable feedback ensuring the accuracy of the review.

References

[1] K.W. Moremen, M. Tiemeyer, A.V. Nairn, Vertebrate protein glycosylation:diversity, synthesis and function, Nat. Rev. Mol. Cell Biol. 13 (2012) 448–462.

[2] M. Thaysen-Andersen, N.H. Packer, Site-specific glycoproteomics confirms thatprotein structure dictates formation of N-glycan type, core fucosylation andbranching, Glycobiology 22 (2012) 1440–1452.

[3] J. Ma, G.W. Hart, Protein O-GlcNAcylation in diabetes and diabetic complications,Expert Rev. Proteomics 10 (2013) 365–380.

[4] L. Wells, The o-mannosylation pathway: glycosyltransferases and proteins im-plicated in congenital muscular dystrophy, J. Biol. Chem. 288 (2013)6930–6935.

[5] P. Stanley, C.J. Guidos, Regulation of Notch signaling during T- and B-cell develop-ment by O-fucose glycans, Immunol. Rev. 230 (2009) 201–215.

[6] H.H. Freeze, Understanding human glycosylation disorders: biochemistry leads thecharge, J. Biol. Chem. 288 (2013) 6936–6945.

[7] R.S. Haltiwanger, J.B. Lowe, Role of glycosylation in development, Annu. Rev.Biochem. 73 (2004) 491–537.

[8] D.T. Tran, K.G. Ten Hagen, Mucin-type O-glycosylation during development, J. Biol.Chem. 288 (2013) 6921–6929.

[9] V. Venkatakrishnan, N.H. Packer, M. Thaysen-Andersen, Host mucin glycosylationplays a role in bacterial adhesion in lungs of individuals with cystic fibrosis, ExpertRev. Respir. Med. 7 (2013) 553–576.

[10] A. Varki, Biological roles of oligosaccharides: all of the theories are correct,Glycobiology 3 (1993) 97–130.

[11] M.N. Christiansen, J. Chik, L. Lee, M. Anugraham, J.L. Abrahams, N.H. Packer, Cellsurface protein glycosylation in cancer, Proteomics 14 (4-5) (2013) 525–546.

[12] M. Stuchlova Horynova, M. Raska, H. Clausen, J. Novak, Aberrant O-glycosylationand anti-glycan antibodies in an autoimmune disease IgA nephropathy and breastadenocarcinoma, Cell. Mol. Life Sci. 70 (2013) 829–839.

[13] D.W. Scott, R.P. Patel, Endothelial heterogeneity and adhesion molecules N-glycosylation: implications in leukocyte trafficking in inflammation, Glycobiology23 (2013) 622–633.

[14] S. Schedin-Weiss, B. Winblad, L.O. Tjernberg, The role of protein glycosylation inAlzheimer disease, FEBS J. 281 (2014) 46–62.

[15] A. Grigorian, H. Mkhikian, C.F. Li, B.L. Newton, R.W. Zhou, M. Demetriou, Pathogen-esis of multiple sclerosis via environmental and genetic dysregulation of N-glycosylation, Semin. Immunopathol. 34 (2012) 415–424.

[16] M. Dalziel, M. Crispin, C.N. Scanlan, N. Zitzmann, R.A. Dwek, Emerging princi-ples for the therapeutic exploitation of glycosylation, Science 343 (2014)1235681.

[17] M. Aebi, N-linked protein glycosylation in the ER, Biochim. Biophys. Acta 1833(2013) 2430–2437.

[18] H. Geyer, R. Geyer, Strategies for analysis of glycoprotein glycosylation, Biochim.Biophys. Acta 1764 (2006) 1853–1869.

[19] Z. Sumer-Bayraktar, D. Kolarich, M.P. Campbell, S. Ali, N.H. Packer, M. Thaysen-Andersen, N-glycans modulate the function of human corticosteroid-binding glob-ulin, Mol. Cell. Proteomics 10 (2011) (M111 009100).

[20] L. Dissing-Olesen, M. Thaysen-Andersen, M. Meldgaard, P. Hojrup, B. Finsen, Thefunction of the human interferon-beta 1a glycan determined in vivo, J. Pharmacol.Exp. Ther. 326 (2008) 338–347.

[21] J.S. McLellan, W.C. Ray, M.E. Peeples, Structure and function of respiratory syn-cytial virus surface glycoproteins, Curr. Top. Microbiol. Immunol. 372 (2013)83–104.

[22] B.M. Kevany, Y. Tsybovsky, I.D. Campuzano, P.D. Schnier, A. Engel, K.Palczewski, Structural and functional analysis of the native peripherin–ROM1 complex isolated from photoreceptor cells, J. Biol. Chem. 288 (2013)36272–36284.

[23] K. Marino, J. Bones, J.J. Kattla, P.M. Rudd, A systematic approach to protein gly-cosylation analysis: a path through the maze, Nat. Chem. Biol. 6 (2010)713–723.

[24] H. Stockmann, B. Adamczyk, J. Hayes, P.M. Rudd, Automated, high-throughput IgG-antibody glycoprofiling platform, Anal. Chem. 85 (2013) 8841–8849.

[25] A. Knezevic, J. Bones, S.K. Kracun, O. Gornik, P.M. Rudd, G. Lauc, High throughputplasma N-glycome profiling using multiplexed labelling and UPLC with fluores-cence detection, Analyst 136 (2011) 4670–4673.

[26] Y. Mechref, Y. Hu, J.L. Desantos-Garcia, A. Hussein, H. Tang, Quantitative glycomicsstrategies, Mol. Cell. Proteomics 12 (2013) 874–884.

[27] N. Leymarie, J. Zaia, Effective use of mass spectrometry for glycan and glycopeptidestructural analysis, Anal. Chem. 84 (2012) 3040–3048.

[28] D.J. Harvey, Analysis of carbohydrates and glycoconjugates bymatrix-assisted laserdesorption/ionization mass spectrometry: an update for 2007–2008, MassSpectrom. Rev. 31 (2012) 183–311.

[29] M. Pabst, F. Altmann, Glycan analysis by modern instrumental methods, Proteo-mics 11 (2011) 631–643.

[30] S. Hua, C. Lebrilla, H.J. An, Application of nano-LC-based glycomics towards bio-marker discovery, Bioanalysis 3 (2011) 2573–2585.

[31] S.M. Patrie, M.J. Roth, J.J. Kohler, Introduction to glycosylation and mass spectrom-etry, Methods Mol. Biol. 951 (2013) 1–17.

[32] B.H. Clowers, E.D. Dodds, R.R. Seipert, C.B. Lebrilla, Site determination of proteinglycosylation based on digestion with immobilized nonspecific proteases and Fouriertransform ion cyclotron resonance mass spectrometry, J. Proteome Res. 6 (2007)4032–4040.

[33] N. Leymarie, P.J. Griffin, K. Jonscher, D. Kolarich, R. Orlando, M. McComb, J. Zaia,J. Aguilan, W.R. Alley, F. Altmann, L.E. Ball, L. Basumallick, C.R. Bazemore-Walker, H. Behnken, M.A. Blank, K.J. Brown, S.C. Bunz, C.W. Cairo, J.F. Cipollo,R. Daneshfar, H. Desaire, R.R. Drake, E.P. Go, R. Goldman, C. Gruber, A. Halim,Y. Hathout, P.J. Hensbergen, D.M. Horn, D. Hurum, W. Jabs, G. Larson, M. Ly, B.F. Mann, K. Marx, Y. Mechref, B. Meyer, U. Moginger, C. Neusubeta, J. Nilsson,M.V. Novotny, J.O. Nyalwidhe, N.H. Packer, P. Pompach, B. Reiz, A. Resemann,J.S. Rohrer, A. Ruthenbeck, M. Sanda, J.M. Schulz, U. Schweiger-Hufnagel, C.Sihlbom, E. Song, G.O. Staples, D. Suckau, H. Tang, M. Thaysen-Andersen, R.I.Viner, Y. An, L. Valmu, Y. Wada, M. Watson, M. Windwarder, R. Whittal, M.Wuhrer, Y. Zhu, C. Zou, Interlaboratory study on differential analysis of proteinglycosylation by mass spectrometry: the ABRF glycoprotein research multi-institutional study 2012, Mol. Cell. Proteomics 12 (2013) 2935–2951.

[34] A. Rogowska-Wrzesinska, M.C. Le Bihan, M. Thaysen-Andersen, P. Roepstorff, 2Dgels still have a niche in proteomics, J. Proteome 88 (2013) 4–13.

[35] D. Kolarich, P.H. Jensen, F. Altmann, N.H. Packer, Determination of site-specific gly-can heterogeneity on glycoproteins, Nat. Protoc. 7 (2012) 1285–1298.

[36] Z. Sumer-Bayraktar, T. Nguyen-Khuong, R. Jayo, D.D. Chen, S. Ali, N.H. Packer, M.Thaysen-Andersen, Micro- and macroheterogeneity of N-glycosylation yields sizeand charge isoforms of human sex hormone binding globulin circulating inserum, Proteomics 12 (2012) 3315–3327.

[37] P. Pompach, K.B. Chandler, R. Lan, N. Edwards, R. Goldman, Semi-automated iden-tification of N-Glycopeptides by hydrophilic interaction chromatography, nano-reverse-phase LC–MS/MS, and glycan database search, J. Proteome Res. 11(2012) 1728–1740.

[38] K.B. Chandler, P. Pompach, R. Goldman, N. Edwards, Exploring site-specific N-glycosylation microheterogeneity of haptoglobin using glycopeptide CID tan-dem mass spectra and glycan database search, J. Proteome Res. 12 (2013)3652–3666.

[39] R. Apweiler, H.Hermjakob,N. Sharon,On the frequencyof protein glycosylation, as de-duced fromanalysis of the SWISS-PROTdatabase, Biochim. Biophys. Acta 1473 (1999)4–8.

[40] J. Bones, S. Mittermayr, N. O'Donoghue, A. Guttman, P.M. Rudd, Ultra performanceliquid chromatographic profiling of serum N-glycans for fast and efficient identifi-cation of cancer associated alterations in glycosylation, Anal. Chem. 82 (2010)10208–10215.

[41] S.J. North, P.G. Hitchen, S.M. Haslam, A. Dell, Mass spectrometry in the anal-ysis of N-linked and O-linked glycans, Curr. Opin. Struct. Biol. 19 (2009)498–506.

[42] M.J. Kailemia, L.R. Ruhaak, C.B. Lebrilla, I.J. Amster, Oligosaccharide analysis bymass spectrometry: a review of recent developments, Anal. Chem. 86 (2014)196–212.

[43] K. Stavenhagen, H. Hinneburg, M. Thaysen-Andersen, L. Hartmann, D. Varon Silva,J. Fuchser, S. Kaspar, E. Rapp, P.H. Seeberger, D. Kolarich, Quantitative mapping ofglycoprotein micro-heterogeneity and macro-heterogeneity: an evaluation ofmass spectrometry signal strengths using synthetic peptides and glycopeptides, J.Mass Spectrom. 48 (2013) 627–639.

[44] M. Wuhrer, M.I. Catalina, A.M. Deelder, C.H. Hokke, Glycoproteomics based on tan-dem mass spectrometry of glycopeptides, J. Chromatogr. B Analyt. Technol.Biomed. Life Sci. 849 (2007) 115–128.

[45] B. Tissot, S.J. North, A. Ceroni, P.C. Pang, M. Panico, F. Rosati, A. Capone, S.M. Haslam, A.Dell, H.R. Morris, Glycoproteomics: past, present and future, FEBS Lett. 583 (2009)1728–1735.

[46] I.M. Lazar, W. Lee, A.C. Lazar, Glycoproteomics on the rise: established methods,advanced techniques, sophisticated biological applications, Electrophoresis 34(2013) 113–125.

[47] H. Desaire, Glycopeptide analysis, recent developments and applications, Mol. Cell.Proteomics 12 (2013) 893–901.

[48] E. Ruiz-May, T.W. Thannhauser, S. Zhang, J.K. Rose, Analytical technologies for identi-fication and characterization of the plant N-glycoproteome, Front. Plant Sci. 3 (2012)150.

[49] P.G. Hitchen, K. Twigger, E. Valiente, R.H. Langdon, B.W. Wren, A. Dell,Glycoproteomics: a powerful tool for characterizing the diverse glycoformsof bacterial pilins and flagellins, Biochem. Soc. Trans. 38 (2010) 1307–1313.

[50] W.R. Alley Jr., B.F. Mann, M.V. Novotny, High-sensitivity analytical approachesfor the structural characterization of glycoproteins, Chem. Rev. 113 (2013)2668–2732.

[51] W. Song, M.G. Henquet, R.A. Mentink, A.J. van Dijk, J.H. Cordewener, D. Bosch, A.H.America, A.R. van der Krol, N-glycoproteomics in plants: perspectives and chal-lenges, J. Proteomics 74 (2011) 1463–1474.

[52] J. Zaia, Mass spectrometry and glycomics, OMICS 14 (2010) 401–418.

http://refhub.elsevier.com/S1570-9639(14)00117-4/rf0005

























































































































































[53] D. Kolarich, B. Lepenies, P.H. Seeberger, Glycomics, glycoproteomics and the im-mune system, Curr. Opin. Chem. Biol. 16 (2012) 214–220.

[54] H.J. An, J.W. Froehlich, C.B. Lebrilla, Determination of glycosylation sites andsite-specific heterogeneity in glycoproteins, Curr. Opin. Chem. Biol. 13 (2009)421–426.

[55] D.F. Zielinska, F. Gnad, J.R. Wisniewski, M. Mann, Precision mapping of an in vivo N-glycoproteome reveals rigid topological and sequence constraints, Cell 141 (2010)897–907.

[56] R. Chen, X. Jiang, D. Sun, G. Han, F. Wang, M. Ye, L. Wang, H. Zou, Glycoproteomicsanalysis of human liver tissue by combination of multiple enzyme digestion andhydrazide chemistry, J. Proteome Res. 8 (2009) 651–661.

[57] H. Kaji, T. Shikanai, A. Sasaki-Sawa, H. Wen, M. Fujita, Y. Suzuki, D. Sugahara, H.Sawaki, Y. Yamauchi, T. Shinkawa, M. Taoka, N. Takahashi, T. Isobe, H. Narimatsu,Large-scale identification of N-glycosylated proteins of mouse tissues and construc-tion of a glycoprotein database, GlycoProtDB, J. Proteome Res. 11 (2012) 4553–4566.

[58] Y. Pasing, A. Sickmann, U. Lewandrowski, N-glycoproteomics: mass spectrometry-based glycosylation site annotation, Biol. Chem. 393 (2012) 249–258.

[59] S. Pan, R. Chen, R. Aebersold, T.A. Brentnall, Mass spectrometry basedglycoproteomics—from a proteomics perspective, Mol. Cell. Proteomics 10(2011) (R110 003251).

[60] Q. Shakey, B. Bates, J. Wu, An approach to quantifying N-linked glycoproteins byenzyme-catalyzed 18O3-labeling of solid-phase enriched glycopeptides, Anal.Chem. 82 (2010) 7722–7728.

[61] H. Kaji, T. Isobe, Stable isotope labeling of N-glycosylated peptides by enzymaticdeglycosylation for mass spectrometry-based glycoproteomics, Methods Mol.Biol. 951 (2013) 217–227.

[62] B.L. Parker, G. Palmisano, A.V. Edwards, M.Y. White, K. Engholm-Keller, A. Lee, N.E.Scott, D. Kolarich, B.D. Hambly, N.H. Packer, M.R. Larsen, S.J. Cordwell, QuantitativeN-linked glycoproteomics of myocardial ischemia and reperfusion injury revealsearly remodeling in the extracellular environment, Mol. Cell. Proteomics 10(2011) (M110 006833).

[63] G. Palmisano, M.N. Melo-Braga, K. Engholm-Keller, B.L. Parker, M.R. Larsen, Chem-ical deamidation: a common pitfall in large-scale N-linked glycoproteomic massspectrometry-based analyses, J. Proteome Res. 11 (2012) 1949–1957.

[64] Z. Darula, J. Sherman, K.F. Medzihradszky, How to dig deeper? Improved enrich-ment methods for mucin core-1 type glycopeptides, Mol. Cell. Proteomics 11(2012) (O111 016774).

[65] C. Steentoft, S.Y. Vakhrushev, M.B. Vester-Christensen, K.T. Schjoldager, Y. Kong, E.P. Bennett, U. Mandel, H. Wandall, S.B. Levery, H. Clausen, Mining the O-glycoproteome using zinc-finger nuclease-glycoengineered SimpleCell lines, Nat.Methods 8 (2011) 977–982.

[66] C. Steentoft, S.Y. Vakhrushev, H.J. Joshi, Y. Kong, M.B. Vester-Christensen, K.T.Schjoldager, K. Lavrsen, S. Dabelsteen, N.B. Pedersen, L. Marcos-Silva, R. Gupta, E.P. Bennett, U. Mandel, S. Brunak, H.H. Wandall, S.B. Levery, H. Clausen, Precisionmapping of the human O-GalNAc glycoproteome through SimpleCell technology,EMBO J. 32 (2013) 1478–1488.

[67] S.Y. Vakhrushev, C. Steentoft, M.B. Vester-Christensen, E.P. Bennett, H.Clausen, S.B. Levery, Enhanced mass spectrometric mapping of the humanGalNAc-type O-glycoproteome with SimpleCells, Mol. Cell. Proteomics 12(2013) 932–944.

[68] K.T. Schjoldager, S.Y. Vakhrushev, Y. Kong, C. Steentoft, A.S. Nudelman, N.B.Pedersen, H.H. Wandall, U. Mandel, E.P. Bennett, S.B. Levery, H. Clausen, Probingisoform-specific functions of polypeptide GalNAc-transferases using zinc fingernuclease glycoengineered SimpleCells, Proc. Natl. Acad. Sci. U. S. A. 109 (2012)9893–9898.

[69] T.W. Rademacher, R.B. Parekh, R.A. Dwek, Glycobiology, Annu. Rev. Biochem. 57(1988) 785–838.

[70] X. Wei, L. Li, Comparative glycoproteomics: approaches and applications, Brief.Funct. Genomic Proteomic 8 (2009) 104–113.

[71] Y. Tian, H. Zhang, Glycoproteomics and clinical applications, Proteomics Clin. Appl.4 (2010) 124–132.

[72] H. Narimatsu, H. Sawaki, A. Kuno, H. Kaji, H. Ito, Y. Ikehara, A strategy for discoveryof cancer glyco-biomarkers in serum using newly developed technologies forglycoproteomics, FEBS J. 277 (2010) 95–105.

[73] K. Ueda, Glycoproteomic strategies: from discovery to clinical application of cancercarbohydrate biomarkers, Proteomics Clin. Appl. (2013), http://dx.doi.org/10.1002/prca.201200123 [Epub ahead of print].

[74] J. Jones, S.S. Krag, M.J. Betenbaugh, Controlling N-linked glycan site occupancy,Biochim. Biophys. Acta 1726 (2005) 121–137.

[75] D.C. Liebler, L.J. Zimmerman, Targeted quantitation of proteins by mass spectrom-etry, Biochemistry 52 (2013) 3797–3806.

[76] Y. Zhang, H. Yin, H. Lu, Recent progress in quantitative glycoproteomics, Glycoconj.J. 29 (2012) 249–258.

[77] M.K. Sethi, M. Thaysen-Andersen, J.T. Smith, M.S. Baker, N.H. Packer, W.S. Hancock,S. Fanayan, Comparative N-glycan profiling of colorectal cancer cell lines revealsunique bisecting GlcNAc and alpha-2,3-linked sialic acid determinants are associ-ated with membrane proteins of the moremetastatic/aggressive cell lines, J. Prote-ome Res. 13 (2014) 277–288.

[78] G. Lauc, A. Vojta, V. Zoldos, Epigenetic regulation of glycosylation is the quantummechanics of biology, Biochim. Biophys. Acta 1840 (2014) 65–70.

[79] H. Ito, A. Kuno, H. Sawaki, M. Sogabe, H. Ozaki, Y. Tanaka, M. Mizokami, J. Shoda, T.Angata, T. Sato, J. Hirabayashi, Y. Ikehara, H. Narimatsu, Strategy for glycoproteomics:identification of glyco-alteration usingmultiple glycan profiling tools, J. ProteomeRes.8 (2009) 1358–1367.

[80] L.R. Ruhaak, S. Miyamoto, C.B. Lebrilla, Developments in the identification of glycanbiomarkers for the detection of cancer, Mol. Cell. Proteomics 12 (2013) 846–855.

[81] N.X. Arndt, J. Tiralongo, P.D. Madge, M. von Itzstein, C.J. Day, Differential carbohy-drate binding and cell surface glycosylation of human cancer cell lines, J. Cell.Biochem. 112 (2011) 2230–2240.

[82] J. Roth, Lectins for histochemical demonstration of glycans, Histochem. Cell Biol.136 (2011) 117–130.

[83] S.T. Laughlin, C.R. Bertozzi, Imaging the glycome, Proc. Natl. Acad. Sci. U. S. A. 106(2009) 12–17.

[84] T. Miyagi, K. Takahashi, K. Hata, K. Shiozaki, K. Yamaguchi, Sialidase significance forcancer progression, Glycoconj. J. 29 (2012) 567–577.

[85] K.E. Wellen, C.B. Thompson, A two-way street: reciprocal regulation of metabolismand signalling, Nat. Rev. Mol. Cell Biol. 13 (2012) 270–276.

[86] G. Zauner, R.P. Kozak, R.A. Gardner, D.L. Fernandes, A.M. Deelder, M. Wuhrer, Pro-tein O-glycosylation analysis, Biol. Chem. 393 (2012) 687–708.

[87] P. Mallick, B. Kuster, Proteomics: a pragmatic perspective, Nat. Biotechnol. 28(2010) 695–709.

[88] J. Cox, M. Mann, Quantitative, high-resolution proteomics for data-driven systemsbiology, Annu. Rev. Biochem. 80 (2011) 273–299.

[89] D.S. Dalpathado, H. Desaire, Glycopeptide analysis by mass spectrometry, Analyst133 (2008) 731–738.

[90] N. Deshpande, P.H. Jensen, N.H. Packer, D. Kolarich, GlycoSpectrumScan: fishingglycopeptides fromMS spectra of protease digests of human colostrum sIgA, J. Pro-teome Res. 9 (2010) 1063–1075.

[91] A. Scholten, S. Mohammed, T.Y. Low, S. Zanivan, T.A. van Veen, B. Delanghe, A.J.Heck, In-depth quantitative cardiac proteomics combining electron transfer disso-ciation and the metalloendopeptidase Lys-N with the SILAC mouse, Mol. Cell. Pro-teomics 10 (2011) (O111 008474).

[92] E.D. Dodds, R.R. Seipert, B.H. Clowers, J.B. German, C.B. Lebrilla, Analytical perfor-mance of immobilized pronase for glycopeptide footprinting and implications forsurpassing reductionist glycoproteomics, J. Proteome Res. 8 (2009) 502–512.

[93] G. Zauner, C.A. Koeleman, A.M. Deelder, M. Wuhrer, Nano-HPLC-MS of glycopep-tides obtained after nonspecific proteolysis, Methods Mol. Biol. 951 (2013)113–127.

[94] C.C. Nwosu, R.R. Seipert, J.S. Strum, S.S. Hua, H.J. An, A.M. Zivkovic, B.J. German, C.B.Lebrilla, Simultaneous and extensive site-specific N- and O-glycosylation analysisin protein mixtures, J. Proteome Res. 10 (2011) 2612–2624.

[95] C.C. Nwosu, J. Huang, D.L. Aldredge, J.S. Strum, S. Hua, R.R. Seipert, C.B. Lebrilla, In-gel nonspecific proteolysis for elucidating glycoproteins: a method for targetedprotein-specific glycosylation analysis in complex protein mixtures, Anal. Chem.85 (2013) 956–963.

[96] S. Ongay, A. Boichenko, N. Govorukhina, R. Bischoff, Glycopeptide enrichment andseparation for protein glycosylation analysis, J. Sep. Sci. 35 (2012) 2341–2372.

[97] S. Ito, K. Hayama, J. Hirabayashi, Enrichment strategies for glycopeptides, MethodsMol. Biol. 534 (2009) 195–203.

[98] C.C. Chen, W.C. Su, B.Y. Huang, Y.J. Chen, H.C. Tai, R.P. Obena, Interactionmodes andapproaches to glycopeptide and glycoprotein enrichment, Analyst 139 (2014)688–704.

[99] M.V. Novotny, W.R. Alley Jr., B.F. Mann, Analytical glycobiology at high sensitivity:current approaches and directions, Glycoconj. J. 30 (2013) 89–117.

[100] J. Nilsson, A. Halim, A. Grahn, G. Larson, Targeting the glycoproteome, Glycoconj. J.30 (2013) 119–136.

[101] P.H. Jensen, S. Mysling, P. Hojrup, O.N. Jensen, Glycopeptide enrichment for MALDI-TOF mass spectrometry analysis by hydrophilic interaction liquid chromatographysolid phase extraction (HILIC SPE), Methods Mol. Biol. 951 (2013) 131–144.

[102] Y. Guo, S. Gaiki, Retention and selectivity of stationary phases for hydrophilic inter-action chromatography, J. Chromatogr. A 1218 (2011) 5920–5938.

[103] M. Thaysen-Andersen, M.R. Larsen, N.H. Packer, G. Palmisano, Structural analysis ofglycoprotein sialylation — part I: pre-LC–MS analytical strategies, Rsc Adv. 3(2013) 22683–22705.

[104] M. Thaysen-Andersen, I.B. Thogersen, H.J. Nielsen, U. Lademann, N. Brunner, J.J.Enghild, P. Hojrup, Rapid and individual-specific glycoprofiling of the low abun-dance N-glycosylated protein tissue inhibitor of metalloproteinases-1, Mol. Cell.Proteomics 6 (2007) 638–647.

[105] P. Hagglund, J. Bunkenborg, F. Elortza, O.N. Jensen, P. Roepstorff, A new strategy foridentification of N-glycosylated proteins and unambiguous assignment of theirglycosylation sites using HILIC enrichment and partial deglycosylation, J. ProteomeRes. 3 (2004) 556–566.

[106] S. Mysling, G. Palmisano, P. Hojrup, M. Thaysen-Andersen, Utilizing ion-pairing hy-drophilic interaction chromatography solid phase extraction for efficient glycopep-tide enrichment in glycoproteomics, Anal. Chem. 82 (2010) 5598–5609.

[107] A. Lee, M. Nakano, M. Hincapie, D. Kolarich, M.S. Baker, W.S. Hancock, N.H. Packer,The lectin riddle: glycoproteins fractionated from complex mixtures have similarglycomic profiles, OMICS 14 (2010) 487–499.

[108] S. Fanayan, M. Hincapie, W.S. Hancock, Using lectins to harvest the plasma/serumglycoproteome, Electrophoresis 33 (2012) 1746–1754.

[109] D.A. Butterfield, J.B. Owen, Lectin-affinity chromatography brain glycoproteomicsand Alzheimer disease: insights into protein alterations consistent with the pathol-ogy and progression of this dementing disorder, Proteomics Clin. Appl. 5 (2011)50–56.

[110] Y. Mechref, M. Madera, M.V. Novotny, Glycoprotein enrichment through lectin af-finity techniques, Methods Mol. Biol. 424 (2008) 373–396.

[111] A.S. Vercoutter-Edouart, M.C. Slomianny, O. Dekeyzer-Beseme, J.F. Haeuw, J.C.Michalski, Glycoproteomics and glycomics investigation of membrane N-glycosylproteins from human colon carcinoma cells, Proteomics 8 (2008)3236–3256.

[112] G. Palmisano, S.E. Lendal, K. Engholm-Keller, R. Leth-Larsen, B.L. Parker, M.R.Larsen, Selective enrichment of sialic acid-containing glycopeptides using titanium

































































http://dx.doi.org/10.1002/prca.201200123

http://dx.doi.org/10.1002/prca.201200123











































































































dioxide chromatography with analysis by HILIC and mass spectrometry, Nat.Protoc. 5 (2010) 1974–1982.

[113] A. Halim, J. Nilsson, U. Ruetschi, C. Hesse, G. Larson, Human urinary glycoproteomics;attachment site specific analysis of N- and O-linked glycosylations by CID and ECD,Mol. Cell. Proteomics 11 (2012) (M111 013649).

[114] M. Kurogochi, T. Matsushista, M. Amano, J. Furukawa, Y. Shinohara, M. Aoshima, S.Nishimura, Sialic acid-focused quantitative mouse serum glycoproteomics by mul-tiple reaction monitoring assay, Mol. Cell. Proteomics 9 (2010) 2354–2368.

[115] K. Lohrig, A. Sickmann, U. Lewandrowski, Strong cation exchange chromatographyfor analysis of sialylated glycopeptides, Methods Mol. Biol. 753 (2011) 299–308.

[116] B.L. Parker, M. Thaysen-Andersen, N. Solis, N.E. Scott, M.R. Larsen, M.E. Graham, N.H. Packer, S.J. Cordwell, Site-specific glycan-peptide analysis for determination ofN-glycoproteome heterogeneity, J. Proteome Res. 12 (2013) 5791–5800.

[117] G. Zauner, C.A. Koeleman, A.M. Deelder, M. Wuhrer, Protein glycosylation analysis byHILIC-LC–MS of proteinase K-generated N- and O-glycopeptides, J. Sep. Sci. 33 (2010)903–910.

[118] G. Zauner, A.M. Deelder, M. Wuhrer, Recent advances in hydrophilic interactionliquid chromatography (HILIC) for structural glycomics, Electrophoresis 32(2011) 3456–3466.

[119] J. Nilsson, U. Ruetschi, A. Halim, C. Hesse, E. Carlsohn, G. Brinkmalm, G. Larson, En-richment of glycopeptides for glycan structure and attachment site identification,Nat. Methods 6 (2009) 809–811.

[120] N.E. Scott, B.L. Parker, A.M. Connolly, J. Paulech, A.V. Edwards, B. Crossett, L.Falconer, D. Kolarich, S.P. Djordjevic, P. Hojrup, N.H. Packer, M.R. Larsen, S.J.Cordwell, Simultaneous glycan-peptide characterization using hydrophilic interac-tion chromatography and parallel fragmentation by CID, higher energy collisionaldissociation, and electron transfer dissociation MS applied to the N-linkedglycoproteome of Campylobacter jejuni, Mol. Cell. Proteomics 10 (2011)M000031–MCP000201.

[121] X. Yin, M. Bern, Q. Xing, J. Ho, R. Viner, M. Mayr, Glycoproteomic analysis of thesecretome of human endothelial cells, Mol. Cell. Proteomics 12 (2013) 956–978.

[122] A. Mayampurath, C.Y. Yu, E. Song, J. Balan, Y. Mechref, H. Tang, Computationalframework for identification of intact glycopeptides in complex samples, Anal.Chem. 86 (2014) 453–463.

[123] T.W. Thannhauser, M. Shen, R. Sherwood, K. Howe, T. Fish, Y. Yang, W. Chen, S.Zhang, A workflow for large-scale empirical identification of cell wall N-linked gly-coproteins of tomato (Solanum lycopersicum) fruit by tandem mass spectrometry,Electrophoresis 34 (2013) 2417–2431.

[124] Z. Darula, K.F.Medzihradszky, Affinity enrichment and characterization ofmucin core-1 type glycopeptides from bovine serum, Mol. Cell. Proteomics 8 (2009) 2515–2526.

[125] Y. Huang, X. Shi, X. Yu, N. Leymarie, G.O. Staples, H. Yin, K. Killeen, J. Zaia, Improvedliquid chromatography–MS/MS of heparan sulfate oligosaccharides via chip-basedpulsed makeup flow, Anal. Chem. 83 (2011) 8222–8229.

[126] H. Ye, M.T. Boyne 2nd, L.F. Buhse, J. Hill, Direct approach for qualitative and quan-titative characterization of glycoproteins using tandem mass tags and an LTQOrbitrap XL electron transfer dissociation hybrid mass spectrometer, Anal. Chem.85 (2013) 1531–1539.

[127] A. Harazono, N. Kawasaki, S. Itoh, N. Hashii, Y. Matsuishi-Nakajima, T. Kawanishi, T.Yamaguchi, Simultaneous glycosylation analysis of human serum glycoproteins byhigh-performance liquid chromatography/tandem mass spectrometry, J.Chromatogr. B Analyt. Technol. Biomed. Life Sci. 869 (2008) 20–30.

[128] A.S. Hebert, A.L. Richards, D.J. Bailey, A. Ulbrich, E.E. Coughlin, M.S. Westphall, J.J.Coon, The one hour yeast proteome, Mol. Cell. Proteomics 13 (2014) 339–347.

[129] D. Johnson, B. Boyes, R. Orlando, The use of ammonium formate as a mobile-phasemodifier for LC–MS/MS analysis of tryptic digests, J. Biomol. Tech. 24 (2013)187–197.

[130] M.Wuhrer, A.M. Deelder, C.H. Hokke, Protein glycosylation analysis by liquid chro-matography–mass spectrometry, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci.825 (2005) 124–133.

[131] R. Uematsu, J. Furukawa, H. Nakagawa, Y. Shinohara, K. Deguchi, K. Monde, S.Nishimura, High throughput quantitative glycomics and glycoform-focused prote-omics of murine dermis and epidermis, Mol. Cell. Proteomics 4 (2005) 1977–1989.

[132] G. Palmisano, M.R. Larsen, N.H. Packer, M. Thaysen-Andersen, Structural analysis ofglycoprotein sialylation —part II: LC–MS based detection, Rsc Adv. 3 (2013)22706–22726.

[133] M. Thaysen-Andersen, B.L. Wilkinson, R.J. Payne, N.H. Packer, Site-specific charac-terisation of densely O-glycosylated mucin-type peptides using electron transferdissociation ESI-MS/MS, Electrophoresis 32 (2011) 3536–3545.

[134] S. Dam,M. Thaysen-Andersen, E. Stenkjaer, A. Lorentzen, P. Roepstorff, N.H. Packer,J. Stougaard, Combined N-glycome and N-glycoproteome analysis of the Lotusjaponicus seed globulin fraction shows conservation of protein structure and glyco-sylation in legumes, J. Proteome Res. 12 (2013) 3383–3392.

[135] D. Goldberg, M. Bern, S. Parry, M. Sutton-Smith, M. Panico, H.R. Morris, A. Dell, Auto-mated N-glycopeptide identification using a combination of single- and tandem-MS,J. Proteome Res. 6 (2007) 3995–4005.

[136] M. Thaysen-Andersen, S. Mysling, P. Hojrup, Site-specific glycoprofiling of N-linkedglycopeptides using MALDI-TOF MS: strong correlation between signal strengthand glycoform quantities, Anal. Chem. 81 (2009) 3933–3943.

[137] C.C. Nwosu, J.S. Strum, H.J. An, C.B. Lebrilla, Enhanced detection and identification ofglycopeptides in negative ion mode mass spectrometry, Anal. Chem. 82 (2010)9654–9662.

[138] K. Deguchi, H. Ito, T. Baba, A. Hirabayashi, H. Nakagawa, M. Fumoto, H. Hinou, S.Nishimura, Structural analysis of O-glycopeptides employing negative- andpositive-ion multi-stage mass spectra obtained by collision-induced andelectron-capture dissociations in linear ion trap time-of-flight mass spectrometry,Rapid Commun. Mass Spectrom. 21 (2007) 691–698.

[139] E.D. Dodds, Gas-phase dissociation of glycosylated peptide ions, Mass Spectrom.Rev. 31 (2012) 666–682.

[140] M.S. Kim, A. Pandey, Electron transfer dissociation mass spectrometry in proteo-mics, Proteomics 12 (2012) 530–542.

[141] Y. Wada, M. Tajiri, S. Ohshima, Quantitation of saccharide compositions of O-glycans by mass spectrometry of glycopeptides and its application to rheumatoidarthritis, J. Proteome Res. 9 (2010) 1367–1373.

[142] W.R. Alley Jr., Y. Mechref, M.V. Novotny, Characterization of glycopeptides by com-bining collision-induced dissociation and electron-transfer dissociation mass spec-trometry data, Rapid Commun. Mass Spectrom. 23 (2009) 161–170.

[143] Y. Mechref, Use of CID/ETD mass spectrometry to analyze glycopeptides, Curr.Protoc. Protein Sci. (2012), http://dx.doi.org/10.1002/0471140864.ps1211s68(Chapter 12, Unit 12 11 11–11).

[144] F.G. Hanisch, O-glycoproteomics: site-specific O-glycoprotein analysis by CID/ETDelectrospray ionization tandemmass spectrometry and top-down glycoprotein se-quencing by in-source decay MALDI mass spectrometry, Methods Mol. Biol. 842(2012) 179–189.

[145] C. Singh, C.G. Zampronio, A.J. Creese, H.J. Cooper, Higher energy collision dissociation(HCD) product ion-triggered electron transfer dissociation (ETD) mass spectrometryfor the analysis of N-linked glycoproteins, J. Proteome Res. 11 (2012) 4517–4525.

[146] J. Saba, S. Dutta, E. Hemenway, R. Viner, Increasing the productivity of glycopep-tides analysis by using higher-energy collision dissociation-accurate mass-product-dependent electron transfer dissociation, Int. J. Proteomics 2012 (2012)560391.

[147] S.W. Wu, S.Y. Liang, T.H. Pu, F.Y. Chang, K.H. Khoo, Sweet-Heart — an integratedsuite of enabling computational tools for automated MS2/MS3 sequencing andidentification of glycopeptides, J. Proteome 84 (2013) 1–16.

[148] A. Halim, U. Ruetschi, G. Larson, J. Nilsson, LC–MS/MS characterization of O-glycosylation sites and glycan structures of human cerebrospinal fluid glycopro-teins, J. Proteome Res. 12 (2013) 573–584.

[149] K.P. Law, Y.P. Lim, Recent advances in mass spectrometry: data independent anal-ysis and hyper reaction monitoring, Expert Rev. Proteomics 10 (2013) 551–566.

[150] J.W. Froehlich, E.D. Dodds, M.Wilhelm, O. Serang, J.A. Steen, R.S. Lee, A classifier basedon accurate mass measurements to aid large scale, unbiased glycoproteomics, Mol.Cell. Proteomics 12 (2013) 1017–1025.

[151] M. Sanda, P. Pompach, J. Benicky, R. Goldman, LC-MS3 quantification of O-glycopeptides in human serum, Electrophoresis 34 (2013) 2342–2349.

[152] K.F. Aoki-Kinoshita, Introduction to informatics in glycoprotein analysis, MethodsMol. Biol. 951 (2013) 257–267.

[153] C.W. von der Lieth, T. Lutteke, M. Frank, The role of informatics in glycobiology re-search with special emphasis on automatic interpretation of MS spectra, Biochim.Biophys. Acta 1760 (2006) 568–577.

[154] S. Perez, B. Mulloy, Prospects for glycoinformatics, Curr. Opin. Struct. Biol. 15(2005) 517–524.

[155] F. Li, O.V. Glinskii, V.V. Glinsky, Glycobioinformatics: current strategies and tools fordata mining in MS-based glycoproteomics, Proteomics 13 (2013) 341–354.

[156] G. Hart-Smith, M.J. Raftery, Detection and characterization of low abundance gly-copeptides via higher-energy C-trap dissociation and Orbitrap mass analysis, J.Am. Soc. Mass Spectrom. 23 (2012) 124–140.

[157] C.L. Woodin, M. Maxon, H. Desaire, Software for automated interpretation ofmass spectrometry data from glycans and glycopeptides, Analyst 138 (2013)2793–2803.

[158] D.C. Dallas, W.F. Martin, S. Hua, J.B. German, Automated glycopeptide analysis—re-view of current state and future directions, Brief. Bioinform. 14 (2013) 361–374.

[159] M. Bern, Y.J. Kil, C. Becker, Byonic: advanced peptide and protein identificationsoftware, Curr. Protoc. Bioinformatics (2012), http://dx.doi.org/10.1002/0471250953.bi1320s40 (Chapter 13, Unit13 20).

[160] O. Serang, J.W. Froehlich, J. Muntel, G. McDowell, H. Steen, R.S. Lee, J.A. Steen,SweetSEQer, simple de novo filtering and annotation of glycoconjugate mass spec-tra, Mol. Cell. Proteomics 12 (2013) 1735–1740.

[161] O. Ozohanics, J. Krenyacz, K. Ludanyi, F. Pollreisz, K. Vekey, L. Drahos, GlycoMiner: anew software tool to elucidate glycopeptide composition, Rapid Commun. MassSpectrom. 22 (2008) 3245–3254.

[162] G. Zauner, M. Hoffmann, E. Rapp, C.A. Koeleman, I. Dragan, A.M. Deelder, M.Wuhrer, P.J. Hensbergen, Glycoproteomic analysis of human fibrinogen revealsnovel regions of O-glycosylation, J. Proteome Res. 11 (2012) 5804–5814.

[163] S.I. Snovida, E.D. Bodnar, R. Viner, J. Saba, H. Perreault, A simple cellulose columnprocedure for selective enrichment of glycopeptides and characterization bynano LC coupled with electron-transfer and high-energy collisional-dissociationtandem mass spectrometry, Carbohydr. Res. 345 (2010) 792–801.

[164] S. Joenvaara, I. Ritamo, H. Peltoniemi, R. Renkonen, N-glycoproteomics — an auto-mated workflow approach, Glycobiology 18 (2008) 339–349.

[165] H. Hahne, N. Sobotzki, T. Nyberg, D. Helm, V.S. Borodkin, D.M. van Aalten, B.Agnew, B. Kuster, Proteome wide purification and identification of O-GlcNAc-modified proteins using click chemistry and mass spectrometry, J. Proteome Res.12 (2013) 927–936.

[166] A.K. Nagel, M. Schilling, S. Comte-Walters, M.N. Berkaw, L.E. Ball, Identification ofO-linked N-acetylglucosamine (O-GlcNAc)-modified osteoblast proteins by elec-tron transfer dissociation tandem mass spectrometry reveals proteins critical forbone formation, Mol. Cell. Proteomics 12 (2013) 945–955.

[167] S.A. Myers, S. Daou, B. Affar el, A. Burlingame, Electron transfer dissociation (ETD):the mass spectrometric breakthrough essential for O-GlcNAc protein site assign-ments—a study of the O-GlcNAcylated protein host cell factor C1, Proteomics 13(2013) 982–991.

[168] M.B. Vester-Christensen, A. Halim, H.J. Joshi, C. Steentoft, E.P. Bennett, S.B. Levery, S.Y. Vakhrushev, H. Clausen, Mining the O-mannose glycoproteome reveals

































































































http://dx.doi.org/10.1002/0471140864.ps1211s68










































http://dx.doi.org/10.1002/0471250953.bi1320s40

http://dx.doi.org/10.1002/0471250953.bi1320s40































cadherins as major O-mannosylated glycoproteins, Proc. Natl. Acad. Sci. U. S. A. 110(2013) 21018–21023.

[169] M.E. Graham, M. Thaysen-Andersen, N. Bache, G.E. Craft, M.R. Larsen, N.H. Packer, P.J.Robinson, A novel post-translational modification in nerve terminals: O-linked N-acetylglucosamine phosphorylation, J. Proteome Res. 10 (2011) 2725–2733.

[170] A. Halim, G. Brinkmalm, U. Ruetschi, A. Westman-Brinkmalm, E. Portelius, H.Zetterberg, K. Blennow, G. Larson, J. Nilsson, Site-specific characterization ofthreonine, serine, and tyrosine glycosylations of amyloid precursor protein/amyloidbeta-peptides in human cerebrospinal fluid, Proc. Natl. Acad. Sci. U. S. A. 108 (2011)11848–11853.

[171] J.C. Trinidad, R. Schoepfer, A.L. Burlingame, K.F. Medzihradszky, N- and O-glycosylation in the murine synaptosome, Mol. Cell. Proteomics 12 (2013)3474–3488.

[172] J.H. Anonsen, A. Vik, W. Egge-Jacobsen, M. Koomey, An extended spectrum oftarget proteins andmodification sites in the general O-linked protein glycosylationsystem in Neisseria gonorrhoeae, J. Proteome Res. 11 (2012) 5781–5793.

[173] P.H. Jensen, N.G. Karlsson, D. Kolarich, N.H. Packer, Structural analysis of N- and O-glycans released from glycoproteins, Nat. Protoc. 7 (2012) 1299–1310.

[174] D. Kolarich, E. Rapp, W.B. Struwe, S.M. Haslam, J. Zaia, R. McBride, S. Agravat, M.P.Campbell, M. Kato, R. Ranzinger, C. Kettner, W.S. York, The minimum informationrequired for a glycomics experiment (MIRAGE) project: improving the standards forreporting mass-spectrometry-based glycoanalytic data, Mol. Cell. Proteomics 12(2013) 991–995.

[175] C.F. Taylor, N.W. Paton, K.S. Lilley, P.A. Binz, R.K. Julian Jr., A.R. Jones, W. Zhu,R. Apweiler, R. Aebersold, E.W. Deutsch, M.J. Dunn, A.J. Heck, A. Leitner, M.Macht, M. Mann, L. Martens, T.A. Neubert, S.D. Patterson, P. Ping, S.L.Seymour, P. Souda, A. Tsugita, J. Vandekerckhove, T.M. Vondriska, J.P.Whitelegge, M.R. Wilkins, I. Xenarios, J.R. Yates 3rd, H. Hermjakob, The min-imum information about a proteomics experiment (MIAPE), Nat. Biotechnol.25 (2007) 887–893.

[176] A.R. Jones, K. Carroll, D. Knight, K. Maclellan, P.J. Domann, C. Legido-Quigley, L.Huang, L. Smallshaw, H. Mirzaei, J. Shofstahl, N.W. Paton, Guidelines forreporting the use of column chromatography in proteomics, Nat. Biotechnol.28 (2010) 654.



































Advances in LC-MS/MS-based glycoproteomics: getting closer to system-wide site-specific mapping of...

Documents

Transcript of Advances in LC-MS/MS-based glycoproteomics: getting closer to system-wide site-specific mapping of...