Structural analysis of N- and O-glycans released from glycoproteins

©20

12 N

atur

e A

mer

ica,

Inc.

All

righ

ts r

eser

ved.

PROTOCOL

NATURE PROTOCOLS | VOL.7 NO.7 | 2012 | 1299

INTRODUCTIONThe surface of all cells is covered by a thick layer of glycans, which are the first level of interaction of cells with the environment. The sugars on the cell surface are all a product of the same glycosylation machinery in the particular cell, and thus profiling those glycans, often without knowing the conjugate to which they are attached, can give valuable information as to how the cell interacts with other cells. This analysis of the global glycoprofile of a sample (glycom-ics) can be useful in many contexts, as changes in glycosylation have been associated with cellular interactions in wide-ranging biological mechanisms such as cancer1,2, pathogen infection3 and lactation4.

Sugars can be attached to both lipids and proteins of mem-branes. The composition of the glycolipids and glycoproteins varies from cell type to cell type, as well as within the various intracel-lular compartments that are defined by intracellular membranes. Glycoproteins are predominantly components of the cell mem-brane, and information gained from the glycoprofile can be used in a discovery context to compare different types of samples to identify changes in glycosylation and to find biomarkers that occur in differentiation and disease5,6. This protocol applies only to the analysis of glycans released from proteins.

The use of the translational and glycosylation machinery of dif-ferent cellular expression systems to produce therapeutic glycopro-teins has grown rapidly in recent years7. In recombinant protein production, analysis of glycans can be used for quality control and for clonal selection for desired glycosylation profiles, as culture conditions are known to affect glycosylation8. As an online moni-toring tool, it can also be used to gain information about the glyco-sylation process and cellular status in order to devise nutritional feeding schemes of the cells and harvesting of glycotherapeutics with optimal glycosylation.

Robust methods for analyzing the global glycosylation, both on native proteins associated with a disease and on recombinant or plasma-purified glycoprotein therapeutics, are required.

Most glycoproteins, native as well as recombinant, are present as several to many glycoforms, and this heterogeneity is known to affect protein activity, efficacy, stability, degradation, immuno-genicity and solubility9,10.

Glycan profiling by mass spectrometry (MS) can elucidate structural features of glycans that are difficult to determine while the sugar is still attached to the protein. When combined with glycopeptide analysis (see related protocol11), these approaches provide complementary in-depth knowledge of the micro- and macroheterogeneity of glycosylation of proteins.

Development of the protocol and applications of the methodThis protocol has been applied to study the glycosylation of indi-vidual proteins (bottom-up glycomics) and of cells and whole tissues (top-down glycomics). Wilson et al.12 developed the method for the sequential analysis of N- and O-linked glycans from 2D electrophoresis–separated proteins, and Schulz et al.13 applied the approach to study the O-linked glycans from glycoproteins and mucins after electrophoretic separation. Karlsson et al.14 showed that this workflow is also compatible with nanoLC-ESI-MS analyses work-flows, if required. Karlsson et al.15, Schulz et al.16,17, Estrella et al.18 and Thomson et al.19 have applied the protocol to study O-linked mucin oligosaccharides and glycosaminoglycans. The presented protocol was part of two multilaboratory studies comparing different meth-ods for analysis of protein-bound glycans20,21. The N- and O-glycome of the milk fat globule membranes was also studied by Wilson et al.22 and showed distinct differences in the glycan epitopes present on human and bovine milk glycoproteins. More recently, this pro-tocol was applied to the detailed glycoproteomic characterization of the four glycoproteins that constitute secretory IgA (secretory component, IgA1 and IgA2 and joining chain) by Deshpande et al.23. Besides showing that the different subunits have quite dis-tinct glycosylation profiles, this study also targeted the site-specific distribution of N- and O-glycans on the various glycoproteins.

Structural analysis of N- and O-glycans released from glycoproteinsPia H Jensen1,3, Niclas G Karlsson2, Daniel Kolarich1,3 & Nicolle H Packer1

1Faculty of Science, Biomolecular Frontiers Research Centre, Macquarie University, Sydney, New South Wales, Australia. 2Department of Medical Biochemistry, University of Gothenburg, Gothenburg, Sweden. 3Present address: Department of Cardiovascular and Renal Research, University of Southern Denmark, Odense, Denmark (P.H.J.); Department of Biomolecular Systems, Max Planck Institute of Colloids and Interfaces, Potsdam, Germany (D.K.). Correspondence should be addressed to N.H.P. ([email protected]).

Published online 7 June 2012; doi:10.1038/nprot.2012.063

This protocol shows how to obtain a detailed glycan compositional and structural profile from purified glycoproteins or protein mixtures, and it can be used to distinguish different isobaric glycan isomers. Glycoproteins are immobilized on PVDF membranes before the N-glycans are enzymatically released by PNGase F, isolated and reduced. Subsequently, O-glycans are chemically released from the same protein spot by reductive b-elimination. After desalting with cation exchange microcolumns, the glycans are separated and analyzed by porous graphitized carbon liquid chromatography–electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS). Optionally, the glycans can be treated with sialidases or other specific exoglycosidases to yield more detailed structural information. The sample preparation takes approximately 4 d, with a heavier workload on days 2 and 3, and a lighter load on days 1 and 4. The time for data interpretation depends on the complexity of the samples analyzed. This method can be used in conjunction with the analysis of enriched glycopeptides by capillary/nanoLC-ESI-MS/MS, which together provide detailed information regarding the site heterogeneity of glycosylation.

©20

12 N

atur

e A

mer

ica,

Inc.

All

righ

ts r

eser

ved.

PROTOCOL

1300 | VOL.7 NO.7 | 2012 | NATURE PROTOCOLS

Lee et al.24 characterized the global glycosylation changes occurring on liver membrane glycoproteins in mice bearing an extrahepatic tumor by this method and found specific changes particularly in the glycans carrying 2,3-linked neuraminic acid. Label-free quantita-tion proteomics confirmed that these glycan structural changes were reflected by changes in the sialyltransferases and many cell surface lectins such as sialoadhesin-1 and others.

The same approach was used to have a closer look at how multi-lectin affinity chromatography (MLAC) separates glycoproteins in complex mixtures of glycoproteins from cell membranes and plasma25. Interestingly, the overall glycoprofile of the N- and O-linked glycans was almost identical in both the bound and the eluted fractions, showing that there is more than just the affinity of the lectins for a particular protein glycosylation structure contrib-uting to glycoprotein separation by MLAC. The lectins selectively and reproducibly retained some glycoproteins, but other proteins with the same attached oligosaccharide structures did not bind, indicating that the mechanisms responsible for lectin partition-ing of complex glycoprotein mixtures might not be the same as those that have been established for purified and well-characterized standard protein mixtures. Overall, these publications show the wide applicability and efficacy of the protocol presented here.

Comparison with other methodsTwo recent international interlaboratory studies, in which the two laboratories of the authors were involved, were organized by the Human Disease Glycomics/Proteome Initiative (HGPI) to compare the analysis of N- (ref. 20) and O-glycans21 released from the same samples in different laboratories. This comparison of the methods currently in use for the analysis of N- and O-glycans showed that the various approaches for performing glycomics can broadly be divided into three major groups: (i) fluorescently derivatized glycans sepa-rated by LC, (ii) permethylated glycans analyzed by matrix-assisted laser desorption/ionization (MALDI)-MS and (iii) reduced glycans separated and analyzed by LC-MS (presented in this protocol).

Fluorescent tagging by reductive amination followed by HPLC analysis. This has been one of the first sensitive approaches for HPLC-based glycan structure determination26–28. Sensitivity is achieved via fluorescent tagging by reductive amination, which com-prises a chemical modification of the analyte. Reductive amination requires a free reducing end on the released glycan molecules, which is easily obtained by enzymatically releasing N-glycans; however, the lack of a comparable enzyme for O-glycans makes this method less useful for the O-linked oligosaccharides. Hydrazinolysis (which has reduced in popularity because of its chemical toxicity) and nonre-ductive -elimination are currently the only options for releasing O-glycans with an intact reducing end. The latter method has the problem that the glycan is steadily destroyed under the basic, non-reducing conditions by ‘peeling reactions’ occurring at the released reducing terminus. This was confirmed in the Human Proteome Organisation (HUPO) study that compared approaches for O-glycan analysis and found that nonreducing -elimination was least reli-able for semiquantitative O-glycan analysis21.

Fluorescent labeling followed by LC separation methods are cur-rently probably the best for quantitative N-glycomics. It should, however, be emphasized that isomer separation is often not achieved and complex samples may result in incomplete derivatization and/or co-elution of glycans, thereby compromising the quantitation.

Permethylation of glycans followed (mostly) by MALDI-TOF-MS analysis. In principle, this approach, in which every hydroxyl group in the oligosaccharide is methylated, allows rapid screening of glycan compositions and is usually supplemented by separate tandem MS data providing more detailed structural information. The directional fragmentation of permethylated oligosaccharides allows easier interpretation of sequence and branching informa-tion, and several informatics tools have been developed to facilitate the structural determination29–32. However, this approach limits the unambiguous identification of isobaric glycan isomers. Although multiple-stage ESI-MS/MS can, to some extent, overcome these issues33, limited sample amounts derived from biological samples often do not provide sufficient material to allow for the detailed multiple tandem MSn (MS3–9) analyses that are required for unam-biguous isomeric differentiation of every composition. In addi-tion, it is possible that methylation of the structures will not be complete. This will increase sample heterogeneity and result in sample losses, especially if additional permethylation steps are required, as is the case for sulfated glycans34. Uncommon modi-fications, such as methylation and acetylation of the glycans in nonmammalian organisms, might not be accommodated by this method at all35.

Porous graphitized carbon (PGC)-LC-ESI-MS/MS. Reproducible isomer separations of low amounts (fmol) of released, reduced gly-cans have been shown to be successfully analyzed by PGC-LC-ESI-MS/MS. Recent publications have shown that hydrophilic liquid interaction chromatography (HILIC; reviewed in ref. 36) can also be used for the separation of isomeric forms of released glycans. However, in our hands, PGC has advantages over HILIC in terms of robustness, column lifetime and wide pH compatibility. The PGC-LC-ESI-MS/MS analysis approach presented here overcomes several of the above-mentioned shortcomings of other methods and can be used for the analysis of both N- and O-glycans released sequentially from complex mixtures or from single proteins by using standard proteomics LC-MS equipment.

The high resolving power of PGC requires that the reducing end of the released oligosaccharides be reduced to overcome the increased complexity caused by the capacity of the carbon matrix to separate the anomers ( , ) of reducing sugars. This unique separation capacity, however, has the advantage that iso-mers of reduced oligosaccharides with the same composition can be well resolved.

Limitations of this approachAlthough the composition of these released reduced oligosaccha-ride isomers is easy to determine from the PGC-LC/MS data, the interpretation of the different linkages of the structures from the MS/MS data is at present highly manual, and a major limitation of this approach is the lack of good bioinformatic tools for this task. The chromatographic separation of the isomers is indicative of dif-ferent linkages, some of which can be deduced from retention time and diagnostic substructure fragment ions, but it often requires other methods such as exoglycosidase treatment and rechroma-tography to be applied.

One other drawback is that the PGC column matrix is unable to retain neutral monosaccharides and some disaccharides, and thus is not useful for the detection of released O-linked Tn (GalNAc) and T (Gal( 1-3)GalNAc) antigens.

©20

12 N

atur

e A

mer

ica,

Inc.

All

righ

ts r

eser

ved.

PROTOCOL


Experimental designThe protocol consists of six major steps: (i) blotting of the sample to PVDF membrane, (ii) enzymatic release and chemical reduction of N-linked glycans, (iii) reductive release of O-linked glycans from the same protein(s) spot, (iv) desalting of reduced glycans using ion exchange chromatography, (v) PGC-LC-ESI MS/MS analysis of N- and O-glycan alditols and (vi) data analysis.

Dot-blotting/electroblotting onto PVDF membranes. Dot- blotting or electroblotting glycoproteins (which may comprise either a mixture of glycoproteins or a purified protein12) to a PVDF mem-brane makes it possible to sequentially release the N-glycans (using the enzyme PNGase F) and the O-glycans (by chemical reductive

-elimination) from the same protein sample12. This greatly reduces sample handling, contaminants and losses of glycoprotein, simplifies desalting and minimizes the amount of protein required, as the same sample is used to obtain both released N- and O-linked glycans. It also enables long-term storage of samples and allows the use of effec-tive, easy clean-up procedures in the protocol. On the hydrophobic PVDF, the protein part of the molecule binds to the membrane and the hydrophilic sugars are displayed on the surface, making them easily accessible to PNGase F. The protocol is optimized for direct loading of small amounts and volumes of sample (5–20 g of protein in 2–10 l)25,37, but it will work equally well using samples electro-blotted after 1D or 2D electrophoresis6,17,23,38. If larger volumes are required to be dot-blotted directly onto the membrane, the use of a dot-blot apparatus (e.g., Bio-Rad) is advised.

Release of N-linked glycans. The enzyme PNGase F is the gold standard for the release of intact N-linked glycans from mammalian glycoproteins. It cleaves the glycan between the innermost GlcNAc attached to the protein and the asparagine residue. (For details on the cleavage reactions, see Essentials of Glycobiology (http://www.ncbi.nlm.nih.gov/books/NBK1908/) and the Functional Glycomics Gateway (http://www.functionalglycomics.org/).) The analysis of N-glycans from the glycoprotein blot can be accomplished in several ways, depending on the information required. Treatment with PNGase F will release the N-glycans that can be analyzed by PGC-LC-ESI-MS, with or without reduction. The latter will give information on possible O-acetylation of sialic acids (Fig. 1), as the alkaline conditions of reduction will cause loss of acetyl groups. In Chinese hamster ovary (CHO) cell production of recombinant proteins, the O-acetylation of sialic acids can be substantial. This same spot can then be used for O-linked glycan release.

An alternative approach can be to treat replicate samples of blot-ted glycoprotein(s) with PNGase F and with a mixture of PNGase F/ sialidase in parallel, after which the released N-glycans are reduced (Fig. 2). The simultaneous desialylation gives information on sialic acid heterogeneity and yields informative MS/MS spectra.

The released N-glycans can also be reduced in situ by drying down the samples on the membrane after the PNGase F release, which

increases the yield of N-linked glycan alditols. However, partial release of O-linked glycans may occur under these conditions and will be detected together with the N-glycans in this approach.

Reduction of N-linked glycans and release/reduction of O-linked glycans. Released N-glycans can be analyzed in the free or reduced form (Figs. 1 and 2). The reduction of the terminal reducing sugar converts the and anomers of the reducing terminus to sugar alditols so that single chromatographic peaks are separated for each glycan composition on the graphitized carbon column in the subsequent LC-MS/MS analysis. Before reduction of the N-glycans, the reducing-end glycosylamines resulting from the deamidation mechanism of PNGase F have to be converted to hydroxyls by the addition of weak acid. The O-glycans are already in the reduced form owing to the conditions of the reductive

-elimination release from the protein. (For details on the above reactions see Essentials of Glycobiology (http://www.ncbi.nlm.nih.gov/books/NBK1908/) and the Functional Glycomics Gateway (http://www.functionalglycomics.org/)).

Desalting of reduced N- and O-linked glycans. Desalting the samples after the sodium borohydride reduction of N-glycans and sodium hydroxide/borohydride release of O-glycans is performed using cation exchange. The preparation and elution of these col-umns, as described in the protocol, is most easily done by posi-tioning the columns in 1.5-ml microcentrifuge tubes and using the centrifugal force of a small bench centrifuge to spin the liquid through. This allows for simultaneous handling of multiple sam-ples, limited only by the number of vials that fit in the centrifuge.

PGC-LC-ESI MS/MS of glycans. In principle, the released glycans can then be analyzed by any type of MS, either with ESI or MALDI ionization. The most notable drawback to MALDI-MS, in this con-text, is that neutral and acidic glycans have to be analyzed separately because of the need for different matrices and the loss of sialic acids39; to overcome this, additional chemical modification steps are required40. In addition, the identification of different glycan isomers becomes more challenging or impossible without prior chromatographic separation. ESI-MS/MS, wherein both neutral and acidic glycans can be detected in negative ion mode14,15, in

a

1,129.63!

1,183.72!

1,226.73! 1,348.43!

1,470.43!

1,512.42!

1,694.82!

1,840.52!1,877.42!

2,023.52!919.74!

1,183.72!

800 1,000 1,200 1,400 1,600 1,800 2,000m/z

+ O-Acetateb

+ 2 O-Acetate

1,204.5 1,225.5

1,180 1,190 1,200 1,210 1,220 1,230 1,240m/z

Figure 1 | Screening of N-linked glycans from erythropoietin. (a) Nonreduced N-linked glycans from 10 g of recombinant human erythropoietin (rHuEPO, biological reference standard (EPO BRP), European Pharmacopoeia Commission) released by PNGase F (eluted from graphitized carbon column RT 35.3–47.4 min). (b) Expanded region (RT 42.9–43.8 min) of a to focus on O-acetylation of N-acetyl neuraminic acid–containing structures. Major peaks and their charge states in the spectra are labeled. The glycan structures are annotated using the Consortium for Functional Glycomics (CFG) symbol nomenclature (http://www.functionalglycomics.org/static/consortium/Nomenclature.shtml).

http://www.functionalglycomics.org/static/consortium/Nomenclature.shtml

http://www.functionalglycomics.org/static/consortium/Nomenclature.shtml

©20

12 N

atur

e A

mer

ica,

Inc.

All

righ

ts r

eser

ved.

PROTOCOL


combination with the isomeric separation afforded by graphitized carbon LC, has proven to be a very powerful tool for the analysis of released nonlabeled N- and O-glycans41. PGC separates not only different monosaccharide compositions but also isobaric isomers that differ in linkage and branching42. This provides an important

extra dimension of information that can be used to determine the detailed glycoprofile of a sample. The protocol presented here has been thoroughly tested and used by the authors for its applicability to both bottom-up and top-down glycomics.

Replicates, positive and negative controls. Ideally, each sample/control is spotted in at least two replicates on the same PVDF membrane. BSA or other proteins known to be nonglycosylated can be spotted as negative controls. This will yield an LC-MS/MS trace of the background and contamination from the procedure and a list of masses that are unlikely to be sample related. A well-characterized glycoprotein such as 1-protease inhibitor, fetuin or IgG can be spotted as a positive control to test the success of the protocol.

1,840.82–

1,227.13–

1,184.82–

1,348.83–1,695.32–

1,100 1,300 1,500 1,700 1,900m/z

1,470.52–1,512.72–

1,129.63–

a

b

c

1,076.32–

1,441.42–

1,623.92–

893.62– 1,787.62–

700 900

600

1,100 1,300 1,500 1,700 1,900m/z

m/z

966.3

675.4

1,258.82–

800 1,000 1,200 1,400

Figure 2 | Screening of reduced N-linked and O-linked glycans from erythropoietin. (a) Reduced N-linked glycans from 10 g of EPO BRP released by PNGase F (summed MS profile of elution from graphitized carbon column). (b) Reduced N-linked glycans from 10 g of EPO BRP after combined release by PNGase F/desialylation using Arthrobacter ureafaciens sialidase. (c) O-Linked glycans released by reductive -elimination from 10 g of EPO BRP after release and removal of N-linked glycans by PNGase F. Only major peaks are labeled. Satellite peaks observed for each of major ions include biologically significant entities such as hydroxylation ( + 16 Da) and acetylation ( + 42 Da), as well as chemical artifacts of ESI ionization (dehydration ! 18 Da) and adducts (sodium + 22 Da and potassium + 36 Da). With careful tuning of the instrument and proper desalting, chemical artifacts are kept below 5% of the corresponding [M ! xH]x- ion.

MATERIALSREAGENTS

Ethanol (Sigma-Aldrich)Methanol (Sigma-Aldrich)Water, (Milli-Q ultrahigh-purity water (Millipore))Glacial acetic acidDirect Blue 71 (Sigma-Aldrich, cat. no. 212407-50G)Polyvinylpyrrolidone (PVP40; Sigma-Aldrich, cat. no. PVP40-50G)PNGase F (Flavobacterium meningosepticum recombinant in E. coli; Roche, cat. no. 11365185001)Sialidase Arthrobacter ureafaciens (for 3- and 6-linked sialic acid, Glyko, cat. no. GK80040) or Streptococcus pneumoniae (for 3-linked sialic acid, Glyko, cat. no. GK80040) Ammonium hydroxide solution (puriss. p.a. plus, 25% (vol/vol) NH

3 in

H2O; Sigma-Aldrich)

Sodium borohydride (NaBH4; Sigma-Aldrich)

Potassium hydroxide (KOH; Sigma-Aldrich)Hydrochloric acid (HCl)ZipTip C18 (Millipore)Cation exchange resin (AG 50W X8; Bio-Rad, cat. no. 142-1431)TopTip (empty pipette tip with frit, Glygen) ! CAUTION Alternatively, a C18 Zip-Tip or Stage Tip (Thermo Fisher Scientific) can serve as the frit in the pipette tip.Extract-Clean carbon SPE cartridge (Grace)Trifluoroacetic acid (TFA; Sigma)Ammonium bicarbonate (NH

4HCO

3; Sigma)

Acetonitrile, LC-MS grade (Fluka)Fetuin from fetal calf serum (Sigma-Aldrich)Ammonium acetate

EQUIPMENTImmobilon-PSQ PVDF membrane (Millipore) ! CAUTION For high-throughput applications, Immobilon-P PVDF 96-well plates (Millipore) may be used, but vacuum filtration of the plate is not recommended (see Step 13).Kimtech Science Kimwipes tissues (Kimberly-Clark)Pyrex Petri dish (Corning)Shaker

•••••••

•

•

••••••

••••••

•

•••

Flat-bottom polypropylene 96-well plate (Corning)Clean scalpel or scissorsParafilm or lid to seal 96-well plateIncubator at 37 °C (Boekel Scientific)Ultrasonic bathSpeedVac concentrator (Thermo Fischer Scientific)Incubator at 50 °C (Boekel Scientific)MicrocentrifugeAgilent 1100 Series LC/MSD Trap XCT Plus or equivalent (Agilent)Graphitized carbon column (Hypercarb KAPPA Capillary Column; Thermo Fisher Scientific) 250 Å, 320 m inner diameter ! 100 mm, 5 mSonicatorMicrocentrifuge tubes

REAGENT SETUP Glycoproteins Glycoprotein(s) of interest (e.g., recombinant erythropoietin) in solution. CRITICAL Should be stored according to the manufacturer’s recommendations.Direct Blue wash solution Direct Blue wash solution is 40% (vol/vol) etha-nol, 10% (vol/vol) acetic acid. It can be stored for 3 months at 20–25 °C.Direct Blue 71 stock solution Stock solution is 0.1% (wt/vol) Direct Blue 71 in Milli-Q water.Polyvinylpyrrolidone (PVP40) PVP40 solution is 1% (wt/vol) solution in 50% (vol/vol) methanol. It can be stored for 3 months at 20–25 °C.Ammonium acetate (100 mM, pH 5) Weigh the correct amount of acetic acid, add Milli-Q water and adjust the pH to 5 with ammonia. It can be stored for 3 months at 4 °C.KOH Dilute to 50 mM in Milli-Q water. It can be stored for 1 month at 20–25 °C.HCl (1 M) Prepare this solution by careful dilution from concentrated HCl.

CRITICAL When working with concentrated HCl, fume hoods must be used and proper protective measures need to be taken according to the work-place regulations of Milli-Q.NH4HCO3 in water (100 mM) Weigh the correct amount of NH

4HCO

3 and

dilute in ultrapure water. CRITICAL The solution should be filtered before dilution into mobile phases A and B. It can be stored for 1 month at 4 °C.

••••••••••

••

©20

12 N

atur

e A

mer

ica,

Inc.

All

righ

ts r

eser

ved.

PROTOCOL


Mobile phase A (10 mM NH4HCO3) Dilute 100 ml of 100 mM NH4HCO

3 in

900 ml ultrapure water. It can be stored for 1 week at 20–25 °C.Mobile phase B (10 mM NH4HCO3 in 80% (vol/vol) acetonitrile) Add 800 ml acetonitrile to 100 ml of 100 mM NH

4HCO

3 and adjust the volume to 1,000 ml

with ultrapure water. CRITICAL Addition of acetonitrile may precipitate NH

4HCO

3 in an endothermic reaction. Bringing the solution back to room temper-

ature (RT) will redissolve the precipitate. It can be stored for 2 weeks at 20–25 °C.EQUIPMENT SETUPLC system The LC system should be optimized in order to minimize dead volumes, and the integrity of the system should be verified according to the respective laboratory standard operating procedures. Sample-loading times must be adjusted to the respective flow rate of the system and tubing length. Gradients should be optimized for the samples at hand, and the postsample column washing and re-equilibration should be adjusted to the system in use. As an example, below is a table of the specific setup used in our laboratory on an Agilent 1100 Series LC/MSD Trap XCT Plus.

Graphitized carbon column

Hypercarb KAPPA capillary column (Thermo Fisher) 250 Å, 320 m inner diameter " 100 mm, 5 m

Mobile phases A: 10 mM NH4HCO3, B: 10 mM NH4HCO3 in 80% (vol/vol) acetonitrile

Flow rate 7 l min ! 1

Separation programa

Start with 100% A for 5 min, then apply a gradient to 35% A/65% B over the next 30 min, a 35–37 min sharp increase to 100% B to clean the column and then equilibrate column back to the starting condi-tions with 100% A for 30 min

Sample injection volume

5 l

aDepending on the depth of separation required, the gradient can be adjusted to be shallower (better sep-aration of isomers but longer analysis time) or steeper (quicker but less separation of structural isomers).

Mass spectrometer The MS needs to be tuned and calibrated, and the spray optimized for negative ionization according to the manufacturer’s instructions. As an example, below is a table of the specific tuning parameters used in our laboratory on an Agilent 1100 Series LC/MSD Trap XCT Plus.

Ionization mode Negative electrospray

Drying gas flow 7 liters per min

Drying gas temperature 320.0 °C

Nebulizer gas 20.0 p.s.i.

Skimmer ! 30.0 V

Capillary exit ! 180.0 V

Trap drive 59.0 V

ICC On

Maximum accumulation time 300 ms

Target 70,000 (MS/MS)

Scan range 250–2,200

Isolation window Isolation window 4.0 m/z

MS/MS fragmentation amplitude 1.0 V

Smart fragmentation option On (start amplitude 30%—end amplitude 200%)

PROCEDUREDot-blotting TIMING ~ 30 min for ten samples1| Cut a piece of PVDF membrane that will fit the number of protein spots to be applied. Approximately 1.5 cm2 is needed for each spot, which should be < 0.5 cm in diameter. Spotting with ~1 cm of spacing between spots is optimal in most cases.

2| Wet the membrane with ethanol and position it on top of an ethanol-wetted tissue. CRITICAL STEP Ensure that the surface of the membrane is dry and not covered in ethanol before spotting of the samples,

but keep the tissue underneath thoroughly wetted while applying the samples.

3| Spot the protein solution in discrete spots of maximum 2.5 l; 5 g of protein is a suitable amount. This also applies for any standard glycoprotein included. Keep the tissue wet with ethanol until all spots have been applied and are dry.

CRITICAL STEP If more than 2.5 l of sample needs to be applied, let the first spot dry and reapply another 2.5 l on top of the previous spot. The spots usually dry within 5 min.? TROUBLESHOOTING

4| Dry the membrane at RT while avoiding any contamination of the sample. PAUSE POINT To ensure that all proteins are properly bound, it is recommended that the membrane be dried overnight.

Staining of the PVDF membrane TIMING ~45 min5| Wash the PVDF membrane in a Petri dish with methanol while shaking for 15 min to rewet it.

6| Remove salts by washing the membrane for 15 min in water. CRITICAL STEP Ensure that the membrane is submerged in the water and not just floating on top. Alternatively, the mem-

brane can be turned ‘face down’ during the water-washing step. Ensure that it is turned back before starting the staining.

7| Stain the membrane in a mixture of 20 ml of wash solution and 1.6 ml of Direct Blue 71 stock solution until the spots appear. CRITICAL STEP The stain is only used to visualize where the protein spots are, so stop the staining once all spots are visible.

©20

12 N

atur

e A

mer

ica,

Inc.

All

righ

ts r

eser

ved.

PROTOCOL


8| Destain briefly in wash solution and then in water in order to remove the acid.

9| Dry the membrane. PAUSE POINT The membrane can be kept at RT for months. Just wrap it lightly in tinfoil or store it in a closed box to keep

it away from dust.

Release of N-linked glycans from PVDF (and optional desialylation) TIMING ~60 min on day 1; ~90 min on day 2 for ten samples10| Fill a number of wells in a 96-well plate with 100 l of 1% (wt/vol) PVP40 solution.

CRITICAL STEP The PVP40 solution blocks the membrane and the well so as to avoid nonspecific binding of PNGase F enzyme.

11| Carefully cut protein spots out of the membrane with a scalpel or scissors and put each spot in a separate PVP40-filled well with the protein side up.

12| Shake the plate gently for 5 min and remove PVP40 solution. CRITICAL STEP When washing and otherwise handling the PVDF membrane in the 96-well plate, take care to prevent the

membrane from turning upside down.

13| Repeat washing three times for 5 min with water to wash the spots. CRITICAL STEP It is optional at this stage to move the spots to new wells. If this is done, the new wells must be filled

with 100 l of water first in order to keep the spots wet. Do not use a vacuum manifold for washing steps, as the proteins can enter the membrane and become less accessible to the PNGase F enzyme in the next step.

14| Remove the water and subsequently add 5 l of PNGase F (0.5 U l ! 1; optional: add 1 l (5 mU) of Arthrobacter ureafaciens (cleaves 3- and 6-linked sialic acid) or Streptococcus pneumoniae (cleaves 3-linked sialic acid) sialidase if desialylation is required) and incubate for 15 min at 37 °C.

15| Add 10 l of water and incubate overnight at 37 °C. CRITICAL STEP When carrying out overnight incubations at elevated temperatures, care must be taken to minimize evapo-

ration from the wells. This can be done by adding water to surrounding wells and covering the 96-well plate with a lid or Parafilm. Ensure that the wells are completely sealed.? TROUBLESHOOTING

16| Sonicate the plate for 5 min.

17| Collect the samples in individual tubes. CRITICAL STEP Do not use any brand of low-protein-binding tubes for collecting the released glycans, as substantial

losses of the hydrophilic oligosaccharides occurs in these tubes.

18| Wash the wells two times in 20 l of water (dispense and aspirate several times), and pool the washes with the sample.

19| Add 10 l of 100 mM ammonium acetate (pH 5) to the samples and incubate them at RT for 1 h. CRITICAL STEP The addition of acid removes the glycosylamines from the reducing terminus of PNGase F–released glycans

and allows subsequent quantitative reduction.

20| Dry the samples in a SpeedVac concentrator without heating. PAUSE POINT Dried samples can be stored for months at ! 20 °C.

21| The free N-glycans can now be analyzed by carbon LC-ESI-MS/MS (Step 38; Fig. 1). Alternatively, the glycans can be reduced before analysis (Steps 22–33).

Reduction of N-linked glycans TIMING ~3 h for ten samples22| Add 20 l of 1 M NaBH4 in 50 mM KOH to the dried N-glycans and incubate for 3 h at 50 °C.

CRITICAL STEP The reduction of the terminal reducing sugar of N-glycans converts the and anomers of the reducing terminus to sugar alditols so that single chromatographic peaks are separated for each glycan composition on the graphitized carbon column in the subsequent LC-MS/MS analysis.

©20

12 N

atur

e A

mer

ica,

Inc.

All

righ

ts r

eser

ved.

PROTOCOL


23| Briefly (~30 s) spin in a mini-centrifuge to collect the sample at the bottom of the tube.

24| Add 2 l of glacial acetic acid to neutralize the reaction and mix the samples vigorously. Some effervescence might be seen upon the addition of acid.

25| Spin the samples again and subsequently desalt them as described in Steps 26–33.

Desalting of reduced N-linked glycans TIMING ~60 min for ten samples (plus drying time)26| Prepare one cation exchange column per sample by depositing cation exchange resin (~25 l packed volume) onto a ZipTip C18 tip.

CRITICAL STEP The AG 50W X8 cation exchange resin should be washed with at least three changes of methanol before use. A larger volume of resin can be prewashed and stored in fresh methanol.

CRITICAL STEP Other types of tips may be used for packing the cation exchange columns. The ZipTip C18 tips are used for their convenient size. As the glycans are not retained by the reversed phase, the C18 packing only serves as a means of support to hold back the cation exchange resin in the tip.

27| Put the columns into microcentrifuge tubes and wash the columns by centrifugation three times with 50 l of 1 M HCl followed by three times with 50 l of methanol. Transfer the columns into new microcentrifuge tubes and wash them three times with 50 l of water.

PAUSE POINT Retain a little of the last water wash on top of the column and close the lid. In this way, the columns can be prepared ahead of time and stored for a few hours if necessary.

28| Place the columns in final sample-collection microcentrifuge tubes. Apply the glycan samples to the columns and spin in a minicentrifuge, at 6,000 r.p.m. or 2,000g at 20–25 °C for about 30 s or until only ~5 l remains on top of the column.

29| Wash the original sample tube with 20 l of water. Add this to the column and spin in a minicentrifuge as before (see Step 28).

30| Elute glycans off the column twice using 50 l of water, and then spin in a minicentrifuge as before (see Step 28) and combine the eluants.

31| Remove the columns from the tubes and dry the eluted glycans in the SpeedVac concentrator.

32| Add 100 l of methanol to the samples and dry the samples in the SpeedVac concentrator. Repeat this procedure three or five times.

CRITICAL STEP The addition of methanol is used to remove residual borate by evaporation of the volatile methyl borate. Ensure that the dried samples are completely redissolved before each redrying step. It is possible that this procedure may introduce chemical O-acetylation of oligosaccharides, as well as lactonization of sialylated oligosaccharides.

PAUSE POINT Dried samples can be stored for months at ! 20 °C.? TROUBLESHOOTING

33| The reduced N-glycans can now be taken up in 10 mM NH4HCO3 and analyzed by LC-MS/MS (Step 38; Fig. 2a,b).? TROUBLESHOOTING

Release of O-linked glycans from PVDF TIMING ~10 min (plus incubation time)34| Rewet the membrane with 2.5 l of methanol.

35| Add 20 l of 0.5 M NaBH4 in 50 mM KOH to each well and incubate at 50 °C for 16 h. CRITICAL STEP When carrying out overnight incubations at elevated temperatures, care must be taken to minimize

evaporation from the wells. This can be done by adding water to surrounding wells and covering the 96-well plate with a lid or Parafilm. Ensure that the wells are completely sealed.? TROUBLESHOOTING

36| After incubation and cooling to RT, add 2 l of glacial acetic acid to neutralize the reaction. Some effervescence might be seen upon the addition of acid.? TROUBLESHOOTING

©20

12 N

atur

e A

mer

ica,

Inc.

All

righ

ts r

eser

ved.

PROTOCOL


Desalting of reduced O-linked glycans TIMING ~60 min (plus drying time)37| This is done exactly as described in Steps 26–33 (desalting of reduced N-linked glycans). Samples are ready to be analyzed by LC-ESI-MS (Step 46; Fig. 2c).

PAUSE POINT Dried samples can be stored for months at ! 20 °C.

Optional carbon cleanup steps TIMING ~10 min per sample (including column preparation) CRITICAL At this stage, the released glycans that have been desalted by cation exchange can be subjected directly to

PGC-LC-ESI-MS/MS analysis (proceed directly to Step 46); alternatively, an optional extra cleanup to protect the nano/capillary LC column from particulates can be performed using a carbon solid-phase extraction tip (Steps 38–45). This will remove any residual particles of ion exchange resin used in Steps 26–33, which might be washed into the final sample, as well as any salts that may be remaining.

38| Spin column preparation. Remove the carbon material from a carbon SPE cartridge and transfer into a fresh microcentrifuge tube. Prepare a slurry with an approximate concentration of 50 mg ml ! 1 in methanol.

39| Load approximately 25 l ( = 2.5 mg) of the slurry into an empty TopTip to form small carbon columns in the pipette tip.

40| Place tips into microcentrifuge tubes, spin down to pack column and remove any fines (small carbon particles) that may pass through the TopTip frit.

41| Wash the column by centrifugation three times with 50 l of acetonitrile containing 0.1% (vol/vol) TFA.? TROUBLESHOOTING

42| Wash the tips three times with 50 l of water containing 0.1% (vol/vol) TFA.

43| Cleanup of glycans. Load the sample dissolved in 10 l of 0.1% (vol/vol) TFA on to the top of the column and wash three times with 50 l of water containing 0.1% (vol/vol) TFA. Discard the wash.

44| Elute bound glycans two times with 10 l of 50% (vol/vol) acetonitrile containing 0.1% (vol/vol) TFA.

45| Dry the eluted sample in the SpeedVac concentrator at RT and continue with Step 46.

Graphitized carbon LC-ESI-MS/MS of glycans TIMING ~60–120 min per sample 46| The free or reduced N- and O-linked glycans are dissolved in 10 mM NH4HCO3 and analyzed by PGC-LC-ESI MS/MS.

47| Set up the HPLC system to separate glycans as described under EQUIPMENT SETUP.

48| Set up the MS instrument to separate glycans as described in the EQUIPMENT SETUP.

Data analysis TIMING variable; minutes to days per sample 49| The data require largely manual interpretation. Extracted ion chromatograms of typical glycan-related fragments in negative ion mode, such as m/z 161.05 (Hex), 179.05 (Hex + H2O); 202.1 (HexNAc), 290.1 (NeuAc), 364.1 (HexHexNAc) and 655.2 (NeuAcHexHexNAc), help distinguish glycan parent ion masses from other molecules. The GlycoMod Tool on the ExPASy Server (http://web.expasy.org/glycomod/) is used to match MS masses with possible glycan compositions. The use of freely available software tools such as GlycoWorkbench (http://www.glycoworkbench.org/) is also highly recommended29. GlycoWorkbench assists substantially in analysis of the obtained MS/MS data. Usually, the [M ! nH]n ! ions are the dominating ions, but sodiated adducts can also be detected ( + 22 amu), especially for highly sialylated species. Sialylated species can also sometimes be distinguished by the presence of satellites of N-glycolylneuraminic acid ( + 16 amu from the dominating N-acetyl neuraminic acid–containing peak, depending on the source) and by O-acetylation ( + 42 amu, only detected if the sample is not reduced; Fig. 1b). ! CAUTION Absolute quantitation of unlabeled glycans is not possible, as the ionization efficiency of neutral and negative glycans differs markedly. It is, however, possible to perform relative quantitation to compare glycan abundances across similar samples.? TROUBLESHOOTINGThe positive glycoprotein controls should indicate whether or not the protocol was performed successfully. The inclusion of a positive glycoprotein control for which the detailed glycosylation is known (e.g., 1-protease inhibitor, fetuin or IgG) will

©20

12 N

atur

e A

mer

ica,

Inc.

All

righ

ts r

eser

ved.

PROTOCOL


check for binding of the sample to PVDF, enzyme reactivity, efficient -elimination, sample reduction and desalting, satisfac-tory chromatographic separation and MS detection. The negative nonglycosylated protein control (e.g., BSA) should indicate contamination of reagents and provide a list of background masses that are not sample related. Further troubleshooting advice can be found in Table 1.

TIMINGSteps 1–4, dot-blotting: ~30 min for ten samplesSteps 5–9, staining of PVDF membrane: ~45 minSteps 10–21, release of N-linked glycans from PVDF (and optional desialylation): ~60 min on day 1; ~90 min on day 2 for ten samples (plus incubation time and sample drying time in the SpeedVac concentrator)Steps 22–25, reduction of N-linked glycans: ~3 h for ten samplesSteps 26–33, desalting of reduced N-linked glycans: ~60 min for ten samples (plus sample drying time in the SpeedVac concentrator)Steps 34–36, release of O-linked glycans from PVDF: ~10 min on day 1 (plus 16 h incubation for ten samples)Step 37, desalting of reduced O-linked glycans: ~60 min for ten samples (plus sample drying time in the SpeedVac concentrator)Steps 38–45, optional carbon cleanup step: ~10 min per sample (including column preparation)Steps 46–48, graphitized carbon LC-ESI-MS/MS of glycans: ~60–120 min per sample (depending on the complexity and assuming that the system is running satisfactorily)Step 49, data analysis: variable; minutes to days per sample (depending on complexity)

ANTICIPATED RESULTSThis comprehensive glycan analysis protocol yields information on the overall set of glycans present in microgram quantities of a protein or a mixture of proteins

TABLE 1 | Troubleshooting table.

Step Problem Solution

3 Protein spots ‘run’ on the PVDF and mix with other spots

Ensure that there is no ethanol on top of the PVDF, only on the tissue below

15,35 After incubation no liquid is present in the well anymore

Take care to properly seal the wells before overnight incubation at 37 or 50 °C, respectively

32 White residue is seen in the tubes after drying

Ensure complete removal of residual borate by taking up the sample, an extra time in methanol and re-drying

33,36 Traces of cation exchange beads are present in the sample

Perform an additional carbon cleanup step (Steps 38–45); alternatively, centrifuge the sample and carefully remove the liquid without pipetting any of the beads

41 No signals are detected Ensure that the LC-MS instrument is performing according to specifications

Do not use protein-low-binding tubes for released glycans, as substantial losses of hydrophilic oligosaccharides can occur in these vials

Two separate chromatographic peaks are detected for each N-glycan

Ensure that both the removal of glycosylamines from the reducing end and the reduc-tion of the released N-glycans are performed properly, to eliminate and anomers

8

Inte

nsity

(!1

06)

6

4

2

030 35 40 45 50 55 60 65 70

Time (min)

Figure 3 | PGC LC-MS N-glycan profile derived from 5 g of human IgG. Seventeen different N-glycan structures of 12 different compositions are separated in a single LC-MS/MS analysis.

©20

12 N

atur

e A

mer

ica,

Inc.

All

righ

ts r

eser

ved.

PROTOCOL


at femtomole sensitivity. This includes the composition of the glycans and the information on the structural isomers of each composition. A global profile of both released N- and O-glycans from the same sample can be obtained with one analytical setup. We show the data obtained by this protocol on the profile of the glycosylation of the glycothera-peutics: erythropoietin as nonreduced and reduced oligosaccharides (Figs. 1 and 2), and immunoglobulin G (IgG) as reduced oligosaccharides (Fig. 3). Analysis of N-glycans under nonreduc-ing conditions enables identification of base-labile modifications such as O-acetylation (Fig. 1b), whereas reduction before analysis (as described in this protocol; Fig. 2a,b) removes reducing-terminus anomerization and simplifies the chromatographic separation (Figs. 3 and 4). -elimination to release O-glycans (as described in this protocol) involves simultaneous reduction of the reducing end

to avoid any peeling side reactions43, and thus results in reduced oligosaccharides (Fig. 2c).PGC-LC of nonreduced glycans generally results in two distinct peaks because of the presence of and anomers on the

reducing end, and reduction is also generally recommended for N-glycans so that the exquisite separation power of PGC can be used to determine the structures of isobaric glycans (Figs. 3 and 4). This separation results in signature MS/MS spectra for individual isobaric structures (Fig. 4). The different positions of the galactose on the respective antennae (Fig. 4) can be distinguished on the basis of the particular LC retention time together with the respective MS/MS spectrum. Earlier work assessing the basic separation principles of PGC allows the assignment of particular structural features, such as antenna linkage, on the basis of the relative LC elution times42, and the assigned structure can be further verified by particular frag-ment masses in the acquired MS/MS spectrum. For example, Nakano et al.5 showed that collision-induced dissociation (CID) fragmentation of isomers of the negatively charged, singly sialylated, diantennary N-glycans, of composition Hex5HexNAc4-NeuAc, resulted in specific fragment ions that provided an additional level of confidence in the assignment of the linkages and orientation of antennae.

Another MS/MS fragment allows clear assignment of a fucose moiety attached to the core (rather than as a Lewis-type fucose) of a reduced N-glycan structure with the presence of the singly charged ion at m/z = 350.1, corresponding to the mass of the Fuc-HexNAc-ol fragment (Fig. 4, both spectra). Despite the fact that fucose migration on glycans has been reported on protonated masses in positive ion mode MS/MS spectra44, this phenomenon has not been observed on deprotonated glycans analyzed by negative ion MS, as described in this protocol.

Structural assignment of glycan structures based on LC retention time and MS/MS spectra will be further facilitated by the development and public availability of curated glycan structural knowledgebases (e.g., UniCarbKB, http://www.unicarbkb.org/)38. The recently established UniCarb-DB (http://www.unicarb-db.com/)45,46 contains chromatography and MS fragmentation data pri-marily derived by this protocol. As ion-trap fragmentation is considered to produce very reproducible fragment spectra, spectral matching will hopefully be available to assist in glycan structural determination47 in the near future using such resources.

The described protocol can be applied to the identification of subtle structural differences in cellular glycosylation between biological samples, such as cancer versus noncancer5,24, or the many other systems in which changes in

263.

226

2.2

300.

0

350.

138

2.0

407.

2

424.1

466.1

710.72–

730.22–

910.2

1,09

6.2

517.

72–

556.

12–

607.

62–

628.

22–

782.

32–#

791.

22–#

710.

72–

658.

1

526.

1

571.

2

508.

2

571.

354

7.72–

730.22–

791.22–#

782.

22–#

629.72–1,113.3

1,260.51,258.3

1,256.3 1,422.41,445.5

1,461.5

262.1

304.

1

350.

136

4.2

407.

0

424.0

454.

246

6.1

851.2

910.3

953.2

1,03

8.2

1,05

6.4

1,113.3

1,256.3

1,258.3

1,260.51,281.3

1,421.4

1,445.4

0

1

2

3

4

!104

0

0.25

0.50

0.75

1.00

1.25

400 600 800 1,000 1,200 1,400 m/z

368.

0

50.3 min

50.9 min

0

0.25

0.50

0.75

1.00

1.25

Inte

nsity

(!1

07)

Inte

nsity

(!1

04)

50.3 minEIC m/z 812.3

50.9 min

30 35 40 45 50 55 60 65 Time (min)

Figure 4 | Separation by LC enables individual tandem-MS spectra acquisition for individual glycan structures. Top: extracted ion chromatogram (EIC) of m/z = 812.3 (Hex4HexNAc4Fuc) shows the presence of two isobaric but structurally different glycan structures. Bottom: individually acquired MS/MS spectra of LC-separated N-glycan structures (m/z = 812.32 !) provide signatures allowing assignment of structure-specific features such as core fucosylation (m/z = 350.11 ! ). The antennary position of galactose was assigned on the basis of previously published rules for the elution of diantennary N-glycans5,42.

©20

12 N

atur

e A

mer

ica,

Inc.

All

righ

ts r

eser

ved.

PROTOCOL


glycosylation may be implicated16,17. For example, Nakano et al.5 showed that the drug resistance of T cell acute lymphob-lastic leukemic cells was directly associated with a change in the structure of specific N-glycan sialylated isomers on the cell membrane proteins (Fig. 5). In addition, a detailed knowledge of the global structural heterogeneity of oligosaccharide structures present on the glycoproteins in a system provides valuable input into the interpretation of glycopeptide mass data23 and makes glycosylation site assignment easier (see associated protocol in ref. 11).

ACKNOWLEDGMENTS P.H.J. was supported by the Danish Agency for Science, Technology and Innovation (grant 272-07-0066). D.K. was supported by an Erwin Schrödinger Fellowship from the Austrian Science Fund (grant J2661) and Macquarie University. We also thank M. Nakano for the preparation of Figure 5 (based on data from ref. 5).

AUTHOR CONTRIBUTIONS All authors contributed equally to this work. N.G.K. developed and validated the initial protocol and performed the analysis of recombinant EPO. P.H.J. tested, optimized and wrote the protocol. D.K. and N.H.P. co-wrote and edited the final manuscript. All authors discussed the results and implications and commented on the manuscript at all stages.

COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests.

2,6

2,6

2,6

T cell acute lymphoblasticleukemia cell line

CEM

CEM cells resistant to30 nM dEpoB

CEM cells resistant to300 nM dEpoB

P3_0_Nglycans_GA4_01_1693.d: EIC8 P3_0_Nglycans_GA4_01_1693.d: EIC1

X = 1

Extracted ion chromatogram(Hex)2 (HexNAc)2 (Deoxyhexose)1 (NeuAc)x + (Man)3(GlcNAc)2

X = 2X = 0

P3_30_Nglycans_GA5_01_1694.d: EIC P3_30_Nglycans_GA5_01_1694.d: EIC

P3_300_Nglycans_GA6_01_1695.d: EIC

P3_30_Nglycans_GA5_01_1694.d: EIC

P3_300_Nglycans_GA6_01_1695.d: EIC P3_300_Nglycans_GA6_01_1695.d: EIC

P3_0_Nglycans_GA4_01_1693.d: EIC1

6

Inte

nsity

(10

4 )In

tens

ity (

104 )

Inte

nsity

(10

4 )

Inte

nsity

(10

5 )In

tens

ity (

105 )

Inte

nsity

(10

5 )

Inte

nsity

(10

5 )In

tens

ity (

105 )

Inte

nsity

(10

5 )

4

2

0

6

4

2

0

1.0

0.8

0.6

0.4

0.2

0

1.5

1.2

1.0

0.8

0.6

0.4

0.2

0

35 40 45 50 5545

Time (min) Time (min) Time (min)

52 54 56 58

1.0

0.5

0

2.0

3

2

1

0

1.00

0.75

0.50

0.25

0

1.00

1.25

0.75

0.50

0.25

0

1.5

1.0

0.5

0

2,3

2,62,3

2,3

2,3

Figure 5 | Extracted ion chromatogram comparisons of differently sialylated biantennary glycans released from the membrane proteins of T cell acute lymphoblastic leukemia cells (CEM) compared with those of the same cell line resistant to 30 and 300 nM of the anticancer drug desoxyepothilone B (dEpoB)5.

Published online at http://www.nature.com/doifinder/10.1038/nprot.2012.063. Reprints and permissions information is available online at http://www.nature.com/reprints/index.html.

1. Dennis, J.W., Granovsky, M. & Warren, C.E. Glycoprotein glycosylation and cancer progression. Biochim. Biophys. Acta 1473, 21–34 (1999).

2. Otake, Y. et al. Isolation and characterization of an N-linked oligosaccharide that is significantly increased in sera from patients with non-small cell lung cancer. J. Biochem. 129, 537–542 (2001).

3. Karlsson, N.G. et al. Identification of transient glycosylation alterations of sialylated mucin oligosaccharides during infection by the rat intestinal parasite Nippostrongylus brasiliensis. Biochem. J. 350, 805–814 (2000).

http://www.nature.com/doifinder/10.1038/nprot.2012.063

©20

12 N

atur

e A

mer

ica,

Inc.

All

righ

ts r

eser

ved.

PROTOCOL


4. Landberg, E. et al. Changes in glycosylation of human bile-salt-stimulated lipase during lactation. Arch. Biochem. Biophys. 377, 246–254 (2000).

5. Nakano, M., Saldanha, R., Gobel, A., Kavallaris, M. & Packer, N.H. Identification of glycan structure alterations on cell membrane proteins in desoxyepothilone B resistant leukemia cells. Mol. Cell Proteomics 10, M111 009001 (2011).

6. Wilson, N.L. et al. Glycoproteomics of milk: differences in sugar epitopes on human and bovine milk fat globule membranes. J. Proteome Res. 7, 3687–3696 (2008).

7. Aggarwal, S. What’s fueling the biotech engine? Nat. Biotechnol. 25, 1097–1104 (2007).

8. Maiorella, B.L. et al. Effect of culture conditions on IgM antibody structure, pharmacokinetics and activity. Biotechnology (N Y) 11, 387–392 (1993).

9. Moll, M., Kaufmann, A. & Maisner, A. Influence of N-glycans on processing and biological activity of the nipah virus fusion protein. J. Virol. 78, 7274–7278 (2004).

10. Morais, V.A., Costa, M.T. & Costa, J. N-Glycosylation of recombinant human fucosyltransferase III is required for its in vivo folding in mammalian and insect cells. Biochim. Biophys. Acta 1619, 133–138 (2003).

11. Kolarich, D., Jensen, P.H., Altmann, F. & Packer, N.H. Determination of site-specific glycan heterogeneity on glycoproteins. Nat. Protoc. 7, 1285–1298 (2012).

12. Wilson, N.L., Schulz, B.L., Karlsson, N.G. & Packer, N.H. Sequential analysis of N- and O-linked glycosylation of 2D-PAGE–separated glycoproteins. J. Proteome Res. 1, 521–529 (2002).

13. Schulz, B.L., Packer, N.H. & Karlsson, N.G. Small-scale analysis of O-linked oligosaccharides from glycoproteins and mucins separated by gel electrophoresis. Anal. Chem. 74, 6088–6097 (2002).

14. Karlsson, N.G. et al. Negative ion graphitised carbon nano-liquid chromatography/mass spectrometry increases sensitivity for glycoprotein oligosaccharide analysis. Rapid Commun. Mass Spectrom. 18, 2282–2292 (2004).

15. Karlsson, N.G., Schulz, B.L. & Packer, N.H. Structural determination of neutral O-linked oligosaccharide alditols by negative ion LC-electrospray-MSn. J. Am. Soc. Mass Spectrom. 15, 659–672 (2004).

16. Schulz, B.L. et al. Mucin glycosylation changes in cystic fibrosis lung disease are not manifest in submucosal gland secretions. Biochem. J. 387, 911–919 (2005).

17. Schulz, B.L. et al. Glycosylation of sputum mucins is altered in cystic fibrosis patients. Glycobiology 17, 698–712 (2007).

18. Estrella, R.P., Whitelock, J.M., Packer, N.H. & Karlsson, N.G. Graphitized carbon LC-MS characterization of the chondroitin sulfate oligosaccharides of aggrecan. Anal. Chem. 79, 3597–3606 (2007).

19. Thomson, K.A., Schulz, B.L., Packer, N.H. & Karlsson, N.G. MUC5B glycosylation in human saliva reflects blood group and secretor status. Glycobiology 15, 791–804 (2005).

20. Wada, Y. et al. Comparison of the methods for profiling glycoprotein glycans—HUPO Human Disease Glycomics/Proteome Initiative multi-institutional study. Glycobiology 17, 411–422 (2007).

21. Wada, Y. et al. Comparison of methods for profiling O-glycosylation: Human Proteome Organisation Human Disease Glycomics/Proteome Initiative multi-institutional study of IgA1. Mol. Cell Proteomics 9, 719–727 (2010).

22. Wilson, N.L. et al. Glycoproteomics of milk: differences in sugar epitopes on human and bovine milk fat globule membranes. J. Proteome Res. 7, 3687–3696 (2008).

23. Deshpande, N., Jensen, P.H., Packer, N.H. & Kolarich, D. GlycoSpectrum-Scan: fishing glycopeptides from MS spectra of protease digests of human colostrum sIgA. J. Proteome Res. 9, 1063–1075 (2010).

24. Lee, A. et al. Liver membrane proteome glycosylation changes in mice bearing an extra-hepatic tumour. Mol. Cell Proteomics doi:10.1074/mcp.M900538-MCP200 (2010).

25. Lee, A. et al. The lectin riddle: glycoproteins fractionated from complex mixtures have similar glycomic profiles. OMICS 14, 487–499 (2010).

26. Natsuka, S. & Hase, S. Analysis of N- and O-glycans by pyridylamination. Methods Mol. Biol. 76, 101–113 (1998).

27. Guile, G.R., Rudd, P.M., Wing, D.R., Prime, S.B. & Dwek, R.A. A rapid high-resolution high-performance liquid chromatographic method for separating glycan mixtures and analyzing oligosaccharide profiles. Anal. Biochem. 240, 210–226 (1996).

28. Rudd, P.M. et al. Oligosaccharide sequencing technology. Nature 388, 205–207 (1997).

29. Ceroni, A. et al. GlycoWorkbench: a tool for the computer-assisted annotation of mass spectra of glycans. J. Proteome Res. 7, 1650–1659 (2008).

30. Goldberg, D. et al. Automated N-glycopeptide identification using a combination of single- and tandem-MS. J. Proteome Res. 6, 3995–4005 (2007).

31. Jang-Lee, J. Glycomic profiling of cells and tissues by mass spectrometry: fingerprinting and sequencing methodologies. et al. in Methods in Enzymology Vol. 415 (ed. Fukuda, M.) 59–86 (Academic Press, 2006).

32. Dell, A. & Morris, H.R. Glycoprotein structure determination by mass spectrometry. Science 291, 2351–2356 (2001).

33. Stumpo, K.A. & Reinhold, V.N. The N-glycome of human plasma. J. Proteome Res. 9, 4823–4830 (2010).

34. Lei, M., Mechref, Y. & Novotny, M.V. Structural analysis of sulfated glycans by sequential double-permethylation using methyl iodide and deuteromethyl iodide. J. Am. Soc. Mass Spectrom. 20, 1660–1671 (2009).

35. Srikrishna, G. et al. A novel anionic modification of N-glycans on mammalian endothelial cells is recognized by activated neutrophils and modulates acute inflammatory responses. J. Immunol. 166, 624–632 (2001).

36. Wuhrer, M., de Boer, A.R. & Deelder, A.M. Structural glycomics using hydrophilic interaction chromatography (HILIC) with mass spectrometry. Mass Spectrom. Rev. 28, 192–206 (2009).

37. Lee, A. et al. Rat liver membrane glycoproteome: enrichment by phase partitioning and glycoprotein capture. J. Proteome Res. 8, 770–781 (2009).

38. Schulz, B.L. et al. Mucin glycosylation changes in cystic fibrosis lung disease are not manifest in submucosal gland secretions. Biochem. J. 387, 911–919 (2005).

39. Harvey, D.J. Matrix-assisted laser desorption/ionization mass spectrometry of sphingo- and glycosphingo-lipids. J. Mass Spectrom. 30, 1311–1324 (1995).

40. Zaia, J. Mass spectrometry and glycomics. OMICS 14, 401–418 (2010).41. Kawasaki, N., Ohta, M., Hyuga, S., Hyuga, M. & Hayakawa, T. Application

of liquid chromatography/mass spectrometry and liquid chromatography with tandem mass spectrometry to the analysis of the site-specific carbohydrate heterogeneity in erythropoietin. Anal. Biochem. 285, 82–91 (2000).

42. Pabst, M., Bondili, J.S., Stadlmann, J., Mach, L. & Altmann, F. Mass plus retention time = structure: a strategy for the analysis of N-glycans by carbon LC-ESI-MS and its application to fibrin N-glycans. Anal. Chem. 79, 5051–5057 (2007).

43. Wada, Y. et al. Comparison of methods for profiling O-glycosylation. Mol. Cell. Proteomics 9, 719–727 (2010).

44. Wuhrer, M., Koeleman, C.A., Hokke, C.H. & Deelder, A.M. Mass spectrometry of proton adducts of fucosylated N-glycans: fucose transfer between antennae gives rise to misleading fragments. Rapid Commun. Mass Spectrom.: RCM 20, 1747–1754 (2006).

45. Hayes, C.A. et al. UniCarb-DB: a database resource for glycomic discovery. Bioinformatics 27, 1343–1344 (2011).

46. Campbell, M.P. et al. UniCarbKB: putting the pieces together for glycomics research. Proteomics 11, 4117–4121 (2011).

47. Ito, H., Kameyama, A., Sato, T. & Narimatsu, H. Preparation of a glycan library using a variety of glycosyltrasferases. Methods Mol. Biol. 534, 283–291 (2009).

Structural analysis of N- and O-glycans released from glycoproteins

Documents