Detailed Structural Analysis of N-Glycans Released From Glycoproteins in SDS-PAGE Gel Bands Using HPLC Combined With Exoglycosidase Array Digestions

  • Louise Royle
  • Catherine M. Radcliffe
  • Raymond A. Dwek
  • Pauline M. Rudd
Part of the Methods in Molecular Biology book series (MIMB, volume 347)

Abstract

In contrast to the linear sequences of protein and DNA, oligosaccharides are branched structures. In addition, almost all glycoproteins consist of a heterogeneous collection of differently glycosylated variants. Glycan analysis therefore requires high-resolution separation techniques that can provide detailed structural analysis, including both monosaccharide sequence and linkage information. This chapter describes how a combination of high-performance liquid chromatography (HPLC) and exoglycosidase enzyme array digestions can deliver quantitative glycan analysis of sugars released from glycoproteins in sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel bands by matching HPLC elution positions with a database of standard glycans.

1 Introduction

As the roles for posttranslational modification of proteins become more apparent, it is vital to fully characterize their glycosylation. Alterations in protein glycosylation accompany many diseases (1, 2, 3), so analysis of these glycan changes should lead to a better understanding of disease pathology and provide valuable diagnostic biomarkers (see also Chapters 6 and 8). It is also necessary to monitor glycosylation during the production of recombinant glycoproteins (GPs) to ensure a consistent therapeutic product. Glycans can be linked to protein either via the amide group of an asparagine residue (N-glycans) or to the hydroxyl group of a serine or threonine residue (O-glycans). The high-performance liquid chromatography (HPLC)-based analysis of enzymatically released, fluorescently labeled N-glycans is covered in the protocols in this chapter.

Peptide-N-glycanase F (PNGaseF) is used to remove N-glycans (4), with or without α1-6 linked core fucose. The released glycans are fluorescently labeled for sensitive detection (10 fmol) following HPLC analysis. PNGaseF will not remove any N-glycans which have an α1-3 linked core fucose such as that found on many plant derived glycans; this requires the GP to be digested with trypsin before using PNGaseA to remove the glycans (5,6). There is as yet no generic O-glycanase, making removal of intact, non-reduced O-glycans possible only through chemical methods.

As a consequence of the oligosaccharide processing pathway there is extensive diversity of glycan structures. There are a number of different ways in which the monosaccharides can be linked together; e.g., galactose can be linked to GlcNAc at the 2, 3, 4, or 6 position by either an α or β linkage, giving eight possible isomeric structures. To add to this complexity, in contrast to proteins and nucleic acids, glycans are usually branched structures.

Most GPs exist as a heterogeneous population of glycoforms in which a range of different glycans is present at each glycosylation site. For example, there are 32 different structures found on human IgG, which has only a single glycosylation site at Asn 297 on each heavy chain (7,8). The largest of these is a disialylated, digalactosylated, bisected, core-fucosylated biantennary glycan; the remaining chains are smaller, less-processed glycans lacking some of these monosaccharides. The structures of the glycans depend on the levels of specific glycosyl transferases, the availability of appropriate monosaccharides, and the sugar nucleotide donors (cell-specific glycosylation) as well as the local three-dimensional structure of the protein at the glycosylation site. Thy-1 is an example of a GP in which there is site-specific glycosylation. Each of the three glycosylation sites contain a different range of sugars (9). Local environmental factors can also play a role, and a GP produced under different fermentation conditions or grown in different cell lines can have quite different glycan structures in each case (10).

2 Materials

To eliminate as much background contamination as possible from the HPLC analysis of released glycans, high-purity chemicals (including water) are used throughout (seeNotes 1 and 2). The quality of plasticware can also affect the results, as lubricants used in their manufacture can produce peaks in the HPLC profiles (seeNotes 3 and 4). Non-powdered gloves should be used at all stages of GP purification, glycan release, and analysis; otherwise a polysaccharide ladder may be detected by HPLC that will obscure any glycans present in the sample (seeFig. 1).
Fig. 1.

NP-HPLC profiles of (A) dextran ladder; (B) contamination from the use of powdered gloves. The numbers on the profiles are glucose unit values.

2.1 Sample Preparation

  1. 1.

    Laemli sample buffer (5X): 0.04 g of Bromophenol blue, 0.625 mL of stacking buffer, pH 6.6 (seeSubheading 2.2., item 2), 1 mL of 10% SDS, and 0.5 mL of glycerol in 2.875 mL of water.

     
  2. 2.

    0.5 M of dithiothreitol (DTT): dissolve 7.71 mg of DTT in 100 µL of water (seeNote 2) and immediately freeze in single-use (20-µL) aliquots at −20°C.

     
  3. 3.

    100 mM of iodoacetamide: prepare 18.5 mg in 1 mL of water and immediately freeze in single-use (20-µL) aliquots at −20°C.

     

2.2 Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis

  1. 1.

    30% (w/w) acrylamide: 0.8% (w/v) bis-acrylamide stock solution (37.5-1; Protogel ultrapure protein and sequencing electrophoresis grade, gas stabilized; National Diagnostics, Hessle, Hull, UK). Caution: this is a neurotoxin when unpolymerized, so care should be taken to avoid exposure.

     
  2. 2.

    Stacking buffer: 0.5 M of Tris-HCl (6 g for 100 mL), adjusted to pH 6.6 with HCl (seeNote 5).

     
  3. 3.

    Gel buffer: 1.5 M of Tris-HCl (18.2 g for 100 mL) adjusted to pH 8.8 with HCl.

     
  4. 4.

    10% Sodium dodecyl sulfate (SDS): 1 g of SDS in 10 mL of water.

     
  5. 5.

    5X running buffer: 144 g of glycine, 30 g of Tris-HCl, and 10 g of SDS in 2 L of water.

     
  6. 6.

    Ammonium peroxodisulphate (APS): prepare a 10% solution (1 g in 10 mL of water) and immediately freeze in single-use (200-µL) aliquots at −20°C.

     
  7. 7.

    N,N,N,N′-tetramethyl-ethylenediamine (TEMED; seeNote 6).

     
  8. 8.

    Water-saturated butanol: Shake equal volumes of water and butanol together in a glass bottle. Use the top layer.

     
  9. 9.

    Molecular-weight markers: Sigmamarker wide range (molecular weight 6,500–205,000; Sigma-Aldrich Company Ltd., Poole, Dorset, UK). Dissolve according to manufacturer’s instructions and store frozen in 10-µL aliquots at −20°C.

     
  10. 10.

    Coomassie stain: 1.25 g of Coomassie R-250 (brilliant blue), 250 mL of methanol, 50 mL of concentrated acetic acid, and 200 mL of water.

     
  11. 11.

    Destain 1: 50% (v/v) methanol, 7% (v/v) concentrated acetic acid, 43% water.

     
  12. 12.

    Destain 2: 5% (v/v) methanol, 7% (v/v) concentrated acetic acid, 88% water.

     
  13. 13.

    Mini-gel system: XCell SureLock™ Mini-Cell (Invitrogen, Paisley, UK).

     
  14. 14.

    Gel cassettes (Invitrogen).

     

2.3 N-Glycan Release and Extraction

  1. 1.

    PNGaseF buffer: 20 mM of NaHCO3, pH 7.0 (0.168 g in 100 mL of H2O adjusted to pH 7.0 with HCl). Store frozen at −20°C in 10-mL aliquots.

     
  2. 2.

    PNGaseF stock solution: PNGaseF (Roche Diagnostics GmbH, Mannheim, Germany) made up in H2O to 1000 U/mL.

     
  3. 3.

    PTFE membrane filter (Millex-LCR, Millipore, Watford, Herts, UK).

     
  4. 4.

    1-mL Plastic syringe (PP/PE syringe, Sigma; cat. no. Z230723).

     
  5. 5.

    1% Formic acid in water.

     

2.4 Fluorescent 2-AB Labeling and Cleanup

  1. 1.

    LudgerTag 2-AB labeling kit (Ludger Ltd., Abingdon, Oxon, UK) or Glyko Signal 2-AB Labeling Kit (Prozyme, San Leandro, CA).

     
  2. 2.

    Whatman 3MM chromatography paper cut into 10- × 3-cm pieces.

     
  3. 3.

    Acetonitrile.

     
  4. 4.

    2.5-mL Plastic syringe (PP/PE syringe, Sigma; cat. no. Z116858).

     
  5. 5.

    PTFE membrane filter (Millex-LCR, Millipore).

     

2.5 NP-HPLC Profiling of 2-AB-Labeled N-Glycans

  1. 1.

    Acetonitrile: E Chromasolv® for HPLC far UV (Riedel-de Haën, Sigma; cat. no. 34888; seeNote 7).

     
  2. 2.

    2 M Ammonium formate, pH 4.4, normal phase (NP) stock solution: Weigh 184.12 g of formic acid into a 2-L glass beaker. Place the beaker in an ice bath to which salt has been added to take the temperature down to −10°C, add 1 L of water, and stir with a glass rod. Adjust the pH by adding 25% ammonia solution. This causes a rapid rise in temperature, so the ammonia must be added in small amounts. Add 4X 50 mL of ammonia, making sure that the temperature drops between each addition; then add in 5-mL aliquots until pH 4.4 is reached at room temperature. Transfer the solution to a 2-L volumetric flask and make up to 2 L with water. Store this stock solution in a brown Winchester bottle at room temperature.

     
  3. 3.

    Waters 2695 (Waters Ltd., Elstree, Herts, UK) separations module with a Waters 474 or 2475 fluorescence detector (or another HPLC system which delivers a reproducible shallow gradient and has a fluorescence detector).

     
  4. 4.

    NP column: TSK Amide-80 250- × 4.6-mm column (Anachem, Luton, UK; cat. no. TSK 13071).

     
  5. 5.

    2-Aminobenzamide (2-AB)-dextran ladder (2-AB-glucose homopolymer, Ludger).

     

2.6 Exoglycosidase Digestions

  1. 1.

    A selection of exoglycosidase enzymes such as those listed in Table 1 is needed. These can be purchased from a variety of companies such as Prozyme, Ludger, Merck Biosciences (Nottingham, UK), New England Biolabs (Hitchin, Herts, UK), or Sigma. However, the standard and specificity of these enzymes can vary, so it is prudent to check their specificity occasionally against standard N-glycans (obtainable from Ludger or Prozyme).

     
  2. 2.

    Incubation buffers: 10X 50mM of sodium acetate, pH 5.5, for mixed enzyme incubations. 5X Incubation buffers at the optimal pH are usually supplied with individual enzymes.

     
  3. 3.

    Protein-binding filters for removal of enzymes before HPLC (Micropure-EZ, Millipore; cat. no. 42530).

     
Table 1

Exoglycosidase Enzymes for Glycan Analysis

Enzyme

Abbreviation

Specificity

Arthrobacter ureafaciens sialidase

Abs

Sialic acid α2-6>3,8

Streptococcus pneumoniae sialidase recombinant in Escherichia coli

Nan1

Sialic acid α2-3,8

Bovine testis galactosidase

Btg

Gal β1-3,4>6

S. pneumoniae galactosidase

Spg

Gal β1-4

Coffee bean α galactosidase

Cbag

Gal α1-3,4,6

Bovine kidney fucosidase

Bkf

Fuc β1-6>2≫3,4

Almond meal fucosidase

Amf

Fuc β1-3,4

S. pneumoniae-N-acetylhexosaminidaseaThe exact linkage specificities of the Nary according to their concentration as well as between different recombinant species. It is therefore important to check these against standard bisected, tri-, or tetraantennary glycans if any structures other then biantennary N-glycans are present.

Sph

GlcNAc β1-2>3,4,6

Jack bean-N-acetylhexosaminidaseaThe exact linkage specificities of the Nary according to their concentration as well as between different recombinant species. It is therefore important to check these against standard bisected, tri-, or tetraantennary glycans if any structures other then biantennary N-glycans are present.

Jbh

GlcNAc, GalNAc β1-2,3,4,6

Jack bean mannosidase

Jbm

Man α1-2,3,6

aThe exact linkage specificities of the Nary according to their concentration as well as between different recombinant species. It is therefore important to check these against standard bisected, tri-, or tetraantennary glycans if any structures other then biantennary N-glycans are present.

2.7 Weak Anion Exchange-HPLC Profiling of 2-AB-Labeled N-Glycans

  1. 1.

    2 M of ammonium formate, pH 9.0. Weak anion exchange (WAX) stock solution: make up in the same way as NP stock solution (seeSubheading 2.5., item 2), but adjust pH to 9.0.

     
  2. 2.

    Methanol.

     
  3. 3.

    WAX column: Vydac protein WAX, 7.5- × 50-mm (cat. no. 301 VHP 575).

     
  4. 4.

    2-AB Bovine serum fetuin N-glycan standard. This is a mixture of mono-, di-, tri-, and tetrasialylated glycans which can be obtained by releasing and labeling the N-glycans from bovine serum fetuin (Sigma) as detailed in this chapter. Alternatively, individual N-glycans may be obtained from Ludger or Prozyme.

     

3 Methods

The GP under investigation is first run on SDS-PAGE and visualized by staining with Coomassie blue. The GP can be run as either the reduced or nonreduced GP, although reduction ensures that the protein is fully unfolded and accessible to digestion by PNGaseF, and is therefore the preferred option. Reduction of the GP also has the advantage of separating subunits which can then be analyzed separately (Fig. 2). The method described here is based on that published by Küster et al. (11),with modifications by Radcliffe et al. (12) that involve reduction and alkylation of the GP before SDS-PAGE, freezing the gel pieces before PNGaseF treatment, and extensive washing of the treated gel pieces to remove the released glycans. Further modifications include more extensive washing of the gel pieces before incubation with PNGaseF,which helps to remove background contamination; removal of the step with AG-50 (H + form) resin, which can lead to the loss of some larger glycans (it has been established that desalting is not required); and the addition of an incubation with 1% formic acid to ensure that all released glycans are converted to aldoses before fluorescence labeling. These modifications ensure maximum recovery and minimum background contamination, enabling publishable profiles to be obtained from as little as 2 µg of GP.
Fig. 2.

Human secretory IgA (SIgA) was reduced, alkylated, and then run on a 10% bis-tris gel. The N-glycans were released by in-gel PNGaseF digestion, 2-AB-labeled, and run on NP-HPLC. The HPLC traces are all on the same scale, so the relative size (glucose unit) and abundance of the N-glycans can be compared. (Reproduced from ref.10 with permission from The American Society for Biochemistry and Molecular Biology).

Following release, the glycans are labeled with the fluorophore 2-AB and separated by NP-HPLC. The advantage of using 2-AB over charged labels such as 2-aminoanthranilic acid (2-AA) is that the order of elution on the column is related to the number of sugar residues in the glycan, so that the larger glycans elute later even when sialic acid residues are present (which does not apply when using 2-AA). The elution times of glycans are expressed in glucose units (GU) by reference to a dextran ladder (13). Each individual glycan structure has a GU value that is directly related to the number and linkage of its constituent monosaccharides. Thus GU values can be used to predict structures, since each monosaccharide in a specific linkage adds a given amount to the GU value of a given glycan. The use of arrays of exoglycosidases in combination with NP-HPLC profiling enables the individual monosaccharides and linkages to be determined (13,14). In addition, WAX-HPLC can be used to separate glycans on the basis of charge (15), which can be very useful when sialic acids or sulphate groups are present.

Fingerprinting the whole undigested glycan pool from NP-HPLC profiles may be sufficient for comparing the glycosylation between batches of cultured GPs where the range of possible glycans is well-established. However, confirmation of any initial assignments by exoglycosidase sequencing is essential in most cases as a number of glycan structures can co-elute, e.g., the hybrid structure Fuc1 GlcNAc2Man4GalNAc1Gal1Neu5Ac1 (FcMan4A1G1S1) co-elutes with the complex biantennary GlcNAc2Man3GalNAc2Gal2Neu5Ac1 (A2G2S1) at GU 7.9 (14).

3.1 Preparation of Sample

  1. 1.

    Recommended amounts of sample to run are 5–10 µg of GP per well to ensure that clear separated bands are seen. Samples are usually reduced and alkylated before SDS-PAGE to ensure maximum release of glycans by PNGaseF.

     
  2. 2.

    Reduce the sample by adding 4 µL of 5X sample buffer, 2 µL of 0.5 M DTT, and water to make up to a total of 20 µL. Incubate for 10 min at 70°C. (If samples are to be run nonreduced, then incubate without DTT and do not alkylate).

     
  3. 3.

    Alkylation of reduced samples: add 2 µL of 100 mM iodoacetamide to the reduced samples and incubate for 30 min in the dark at room temperature.

     

3.2 SDS-PAGE and Preparation of Gel Bands for Glycan Removal

  1. 1.

    These instructions assume the use of an Invitrogen vertical mini gel system with freshly prepared gels (80 × 80 × 1 mm).

     
  2. 2.

    Choose a percentage gel appropriate to the size of the protein. Ten percent gel is the most commonly used; however, use 6% gel for proteins greater than 120 kDa, 12.5% gel for 70–200 kDa, or 15% gel for small (<70 kDa) proteins. Mix together the solutions for both the running gels (seeTable 2) and for the stacking gels (seeTable 3) in separate plastic tubes but do not add the TEMED until just before pouring the gel. The APS should be taken out of the freezer just before use and added penultimately.

     
  3. 3.

    Have two empty gel cassettes ready. Add the APS and the TEMED to the running gel mixture and mix well by inverting the tube (do not vortex as this can introduce air bubbles). Fill the gel cassettes with running gel up to the line approx 2 cm from the top edge. Cover with a layer of water-saturated butanol, tap out the bubbles, and leave to set for 15–20 min.

     
  4. 4.

    Pour off the butanol and rinse the gel top with water, and dry off any remaining droplets with a piece of filter paper.

     
  5. 5.

    Add APS and TEMED to the stacking gel mixture and mix well by inversion. Fill the top of the cassettes with stacking gel. Insert the comb, ensuring no bubbles remain. Leave to set for 15–20 min.

     
  6. 6.

    Prepare the running buffer by diluting 200 mL of 5X running buffer with 800 mL of water.

     
  7. 7.

    Carefully remove the combs and rinse the wells three times with running buffer, shaking out the buffer with each rinse. Peel off the tape from the bottom of the gel plate.

     
  8. 8.

    Load the gels with the sample using gel-loading pipet tips, starting with the tip at the bottom of the well and lifting it up slowly as the well fills. Load 5 µL of molecular weight markers in one or two wells.

     
  9. 9.

    Assemble the gel unit according to the manufacturer’s instructions with a magnetic stirrer in the bottom. Add a small amount of running buffer to the inner compartment and make sure there are no leaks. Fill the inner compartment, taking care not to disturb the samples in the wells. Fill the outer chamber to about three quarters. Put the lid on and connect the power supply. With the power off, set the voltage to 500 V and the current to 25 mA per gel (50 mA for two gels). Switch on and run until the blue line reaches the bottom of the cassette (∼45 min for a 10% gel).

     
  10. 10.

    Pry the gel cassette apart with a palette knife and discard the top plate. Gently drop each gel into separate plastic boxes containing enough Coomassie stain to cover and leave to stain for about 2 h (or it can be left overnight) on a platform shaker (note that the gel will shrink).

     
  11. 11.

    Tip out the stain and cover with destain 1. Place on a platform shaker for 5 min.

     
  12. 12.

    Tip out destain 1 and replace with destain 2. Place on a platform shaker for several hours or overnight until sufficiently destained (seeNote 8).

     
  13. 13.

    Photograph the gel.

     
  14. 14.

    On a clean glass plate over a light box, cut out the Coomassie-stained bands from the gel with a clean scalpel. Cut into approx 1-mm3 pieces and transfer the pieces from each band to 1.5-mL Eppendorf tubes, and freeze for at least 2 h or overnight (seeNote 9).

     
  15. 15.

    Wash the gel pieces thoroughly by adding 1 mL of acetonitrile, vortexing, and then mixing on a roller mixer for 30 min at room temperature. Pipet off and discard the liquid. Repeat this washing procedure with 1 mL of PNGaseF buffer, 1 mL of acetonitrile, 1 mL of PNGaseF buffer, and 1 mL of acetonitrile.

     
  16. 16.

    Dry the gel pieces in a vacuum centrifuge.

     
Table 2

Reagent Quantities for Two Running Gels

 

Percentage Gel

 

6%

10%

12.5%

14%

15%

17.5%

Protogel

2.4 mL

4.0 mL

5.0 mL

5.6 mL

6.0 mL

7.0 mL

Gel buffer

3.0 mL

3.0 mL

3.0 mL

3.0 mL

3.0 mL

3.0 mL

H2O

6.6 mL

5.0 mL

4.0 mL

3.4 mL

3.0 mL

2.0 mL

SDS (10%)

120 µL

120 µL

120 µL

120 µL

120 µL

120 µL

APS (10%)

120 µL

120 µL

120 µL

120 µL

120 µL

120 µL

TEMED

12 µL

12 µL

12 µL

12 µL

12 µL

12 µL

Table 3

Reagent Quantities for Two Stacking Gels

Protogel

0.665 mL

Stacking buffer

1.25 mL

H2O

3.05 mL

SDS (10%)

50 µL

APS (10%)

50 µL

TEMED

5 µL

3.3 N-Glycan Release and Extraction

  1. 1.

    The N-glycans are released by incubation with PNGaseF. The quantities given below are sufficient for one gel band of 2–3 mm in length, 10–15 mm3.

     
  2. 2.

    Prepare the PNGaseF by mixing 3 µL of the stock solution with 27 µL of PNGaseF buffer.

     
  3. 3.

    Add the PNGaseF solution to the gel pieces and leave for 10–15 min until the gel has reswollen; add more PNGaseF solution if the gel has not fully reswollen. Cover the gel with 1–2 mm of PNGaseF buffer (50–100 µL) and close the lids. Incubate overnight at 37°C.

     
  4. 4.

    Wash the gel pieces thoroughly to extract the glycans. Vortex the gel pieces and spin down in a benchtop centrifuge. Remove any supernatant and retain in a 1.5-mL Eppendorf tube. Add 200 µL of H2O to the gel and sonicate for 30 min, remove the supernatant, and add to the retentate. Repeat this procedure with a further 2X 200 µL of H2O, 200 µL of acetonitrile, 200 µL of H2O, and 200 µL of acetonitrile (the acetonitrile washes shrink the gel and help to squeeze out the water). Dry the sample down to about 500 µL.

     
  5. 5.

    Filter the glycan solution into a 1.5-mL Eppendorf tube through a 13-mm 0.45-µm low-protein-binding PTFE membrane filter using a 1-mL plastic syringe (seeNote 4), followed by 200 µL of H2O to wash out the syringe and filter.

     
  6. 6.

    Dry down the filtrate and resuspend in 20 µL 1% formic acid and leave at room temperature for 40 min to ensure all the released glycans are converted to aldoses.

     
  7. 7.

    Dry down in a vacuum centrifuge to remove the formic acid. Resuspend in a known volume of H2O. This glycan solution can be stored frozen at −20°C.

     

3.4 Fluorescent 2-AB-Labeling and Cleanup

  1. 1.

    Dry down the glycan sample in a 200-µL PCR tube. Make sure the sample is completely dry.

     
  2. 2.

    Prepare some freshly-washed 3MM chromatography strips by rinsing in three changes of water and drying in an oven at 65°C (seeNotes 10 and 11). The paper should be washed and used within 24 h, as this reduces contaminants that leach from the paper.

     
  3. 3.

    Make up the 2-AB labeling solution per the manufacturer’s instructions. This mixture is stable for several weeks if stored in the dark at −20°C in a clean glass vial.

     
  4. 4.

    Add 5 µL of the 2-AB labeling solution to the tube containing the dry glycans and cap the tube. Vortex and spin down. Incubate at 65°C in a dry oven for 30 min. Re-vortex and spin the sample, then return it to the oven for another 2.5 h.

     
  5. 5.

    Cool the sample in a freezer for 5 min and write the name of the sample in pencil at the top of the paper strip. Spot the sample onto the center line of the filter paper 1 cm from the bottom. Dry with a hair dryer.

     
  6. 6.

    Place the paper strip in a 100-mL glass beaker containing about 10 mL of acetonitrile; this should wet the paper but not touch the sample spot. Leave in a fume hood for 1–1.5 h. Any free 2-AB should run with the solvent front leaving 2-AB-labeled glycans at the origin. Remove the paper from the beaker and dry. Check that the 2-AB streak is well away from the sample spot by viewing under a UV light. Use a clean pair of scissors to cut out the fluorescent sample spot (do not draw around the spot as this will contaminate it).

     
  7. 7.

    Put the paper-sample spot into a 2.5-mL plastic syringe (seeNote 4) fitted with a 13-mm 0.45-µm low-protein-binding PTFE membrane filter. Add 0.5 mL of H2O to the syringe and leave for 10 min for the glycans to dissolve. Push the solution gently through the filter and collect. Repeat with a further 4X 0.5 mL of H2O. Dry down the solution and redissolve in 100 µL of H2O (seeNote 12). This solution of 2-ABlabeled glycans should be stored frozen at −20°C and is stable for at least a year.

     

3.5 NP-HPLC Profiling of 2-AB-Labeled N-Glycans

  1. 1.

    Waters 2695 (Waters Ltd., Elstree, Herts, UK) separations module with a Waters 474 or 2475 fluorescence detector (or another HPLC system which delivers a reproducible shallow gradient and has a fluorescence detector can be used. Set the fluorescence detector excitation and emission wavelengths to 330 and 420 nm with a bandwidth of 16 nm, set at maximum sensitivity.

     
  2. 2.

    Samples are injected in 80% acetonitrile. Take an aliquot of the sample (or standard dextran ladder) and make up to 20 µL with water, then add 80 µL of acetonitrile. It is a good idea to use only a small percentage of your sample in the first run in order to get some idea of how much must be loaded to produce a good trace. Usually about 1–5% of each gel band is sufficient.

     
  3. 3.

    Make up NP-HPLC running buffer by diluting 50 mL of NP stock solution to 2 L with water (this is solvent A). Solvent B is acetonitrile.

     
  4. 4.

    Set the HPLC to run the 30-min startup method (seeTable 4), followed by a 180-min run (seeTable 5) with no sample injection. This ensures that the column is well conditioned and helps with run-to-run reproducibility. Run a dextran ladder standard followed by the samples, with the injection volume set to 95 µL. Make sure that a dextran standard is run with each batch of samples.

     
  5. 5.

    Calibration and allocation of GU: The dextran ladder is used to calibrate the HPLC runs against any day-to-day or system-to-system changes (seeFig. 1). The GU value is calculated by fitting a fifth order polynomial distribution curve to the dextran ladder (usually glucose 1–15), then using this curve to allocate GU values from retention times (Empower GPC software from Waters can be used to calculate GU values). The GU values for neutral N-glycans are very reproducible with standard deviations of less than 0.03 between columns (13). This allows direct comparison with database values collected from a range of instruments over a long period of time. For sialylated glycans, more variation is found (±0.3 for disialylated biantennary) between columns and systems. It is therefore advisable to run sialylated samples together for direct comparison.

     
  6. 6.

    Lifetime of column: In order to maintain resolution of glycan peaks, it is important to monitor the column and change it when the peak widths at 50% height get above 0.65 min for dextran peaks 3 and 8. This is usually after about 800 runs. You may, however, need to be more stringent if you are measuring poorly-resolved peaks.

     
Table 4

NP-HPLC Startup Method

Time (min)

Flow (mL/min)

%A

%B

0

0

20

80

4

1

20

80

8

1

95

5

12

1

95

5

16

1

20

80

26

1

20

80

27

0.4

20

80

40

0.4

20

80

41

0.0

20

80

Run time 30 min. A, 50 mM ammonium formate, pH 4.4, normal phase (NP) stock solution; B, aceteonitrile.

Table 5

NP-HPLC Running Method

Time (min)

Flow (mL/min)

%A

%B

0

0.4

20

80

152

0.4

58

42

155

0.4

100

0

157

1

100

0

162

1

100

0

163

1

20

80

177

1

20

80

178.5

0.4

20

80

200

0.4

20

80

201

0.0

20

80

Run time 180 min. A, 50 mM ammonium formate, pH 4.4, normal phase (NP) stock solution; B, aceteonitrile.

3.6 Exoglycosidase Digestions

  1. 1.

    Pipet aliquots of the 2-AB-labeled glycan pool into 200-µL microcentrifuge tubes and dry down. Add 1 µL of 10X incubation buffer, pH 5.5 (or 2 µL of 5X manufacturer’s buffer), the required enzyme or array of enzymes, and H2O to make up to 10 µL. Incubate overnight (16–18 h) at 37°C. A typical set of enzyme digestion arrays is shown in Fig. 3 for human IgG glycans.

     
  2. 2.

    Prewash the Micropure-EZ enzyme removers with 200 µL of H2O in a microcentrifuge at half speed (∼7000g) for 10 min and discard washings.

     
  3. 3.

    Apply the digested glycan sample to the middle of the filter and centrifuge at full speed (∼14,000g) for 2 min. Wash out the digestion tube with 20 µL of H2O and apply to the filter, then centrifuge for another 2 min. Apply a further 100 µL of H2O to the filter and centrifuge. Dry down the sample, then redissolve in 20 µL of H2O ready for injection onto the HPLC.

     
Fig. 3.

NP-HPLC profiles of 2-AB-labeled human IgG N-glycans. The top profile shows undigested whole pool glycans, followed by a series of exoglycosidase digestions (seeTable 1 for enzyme abbreviations and specificities).

3.7 WAX-HPLC Profiling of 2-AB-Labeled N-Glycans

  1. 1.

    Set up equipment as in Subheading 3.5., step 1.

     
  2. 2.

    Inject samples in 100% aqueous solution. Dilute the required amount of sample or fetuin N-glycan standard to 100 µL, ready for 95-µL injection.

     
  3. 3.

    Make up WAX-HPLC running buffer by diluting 250 mL of WAX stock solution to 1 L with water (this is solvent A). Solvent B is 10% methanol in water.

     
  4. 4.

    Set the HPLC to run the 30-min startup method (seeTable 6), followed by an 80-min run (seeTable 7) with no sample injection. This ensures that the column is well-conditioned and helps with run-to-run reproducibility. Run a fetuin N-glycan standard followed by the samples.

     
  5. 5.

    Compare the elution positions of peaks to those of the fetuin standard (seeFig. 4). Note that the larger triantennary glycans elute before the biantennary glycans with the same charge, and that if excess 2-AB label has not been sufficiently removed, a large “hump” in the baseline which can obscure the glycan peaks is seen when the sample is run on WAX.

     
Table 6

WAX-HPLC Startup Method

Time (min)

Flow (mL/min)

%A

%B

0

0

0

100

5

1

0

100

10

1

0

100

40

1

0

100

41

0

0

100

Run time 30 min. A, 500 mM ammonium formate, pH 9.0, weak anion exchange (WAX) stock solution; B, 10% methanol in water.

Table 7

WAX-HPLC Running Method

 

Flow (mL/min)

%A

%B

0

1

0

100

12

1

5

95

25

1

21

79

50

1

80

20

55

1

100

0

65

1

100

0

66

2

0

100

70

2

0

100

77

2

0

100

78

1

0

100

90

1

0

100

91

0

0

100

Run time 80 min. A, 500 mM ammonium formate, pH 9.0, weak anion exchange (WAX) stock solution; B, 10% methanol in water.

Fig. 4.

WAX-HPLC profiles of (A) bovine serum fetuin N-glycans showing separation of the mono-, di-, tri-, and tetrasialylated glycans. The double peaks of the mono- and disialylated glycans are from the triantennary eluting before the biantennary glycans. (B) When excess 2-AB has not been sufficiently removed, the glycan peaks are obscured.

3.8 Structural Allocation

  1. 1.

    The GU value for a glycan is directly related to the number and linkage of its constituent monosaccharides; the larger the glycan, the higher its GU value. Thus GU values can be used to predict structures since each monosaccharide in a specific linkage adds a given amount to the GU value of a given glycan (seeTable 8).

     
  2. 2.

    The preliminary assignment of structures to peaks is made by matching the GU values from the undigested pool with those in a database of known structures ( refs. 13 and 14 ; see also Oxford Glycobiology Institute HPLC database). This will often result in several possible structures for each peak; therefore, exoglycosidase digestion is required to confirm structures. To make a “final” allocation, the NP-HPLC traces and GU values from the exoglycosidase digestions must be interpreted. For example, in Fig. 2, the Abs + Bkf + Sph + Btg trace shows that the glycans have digested down to GlcNAc2Man3 (GU 4.4). The trace above, which omits the N-acetylhexosaminidase (Sph), shows two peaks (GU 5.5 and GU 5.8), a biantennary glycan with and without a bisecting GlcNAc. In the trace without the fucosidase (Bkf) the two major peaks are 0.5 GU higher, as they contain core fucose plus minor peaks for these two structures without core fucose. In the trace without galactosidase (Abs + Bkf) the structures (±bisect) have 0, 1, or 2 galactose residues added (note that isomers separate: if galactose is on the upper 6-linked arm it elutes later than if the galactose is on the lower 3-linked arm). By continuing to compare successive enzyme digestions and being guided by the incremental values in Table 8, all of the structures can be allocated.

     
  3. 3.

    As there is one 2-AB fluorescent label per glycan, the fluorescence intensity is directly related to the number of moles of labeled glycans present. Thus, the areas of the HPLC peaks can be used to quantify the amounts of glycans present. Comparison of peak areas within a trace gives the relative proportions of different glycans. For actual quantification of the amounts of glycans, a 2-AB-labeled standard of known amount can be used to generate a calibration curve.

     
  4. 4.

    Additional information about the number of sialic acids present can be gained from the WAX-HPLC profiles.

     
  5. 5.

    Digestion with specific sialidases to distinguish between linkages can be carried out and then run on the WAX for comparison.

     
  6. 6.

    It is also feasible to collect these charge-separated fractions following WAXHPLC, then run them separately on NP-HPLC with or without further exoglycosidase digestions (seeFig. 5). The buffer salts used to elute the samples from WAX must be removed before NP-HPLC or exoglycosidase digestion. To do this, dry the collected WAX fractions and redissolve them in 2 mL of H2O and then redry; repeat two or three times before finally lyophilizing them overnight (14).

     
Table 8

Incremental GU Values for 2-AB-Labeled N-Glycans

Monosaccharide

Linkage

To

GU increment

Mannose

α1-2,3,6

Mannose

0.7–0.9

GlcNAc

β1-2,4,6

α-Mannose

0.5

GlcNAc (bisect)

β1-4

β-Mannose

0.2–0.4

Galactose

α or β1-3,4

GlcNAc or Gal

0.8–0.9

Fucose (core)

α1-6

Core GlcNAc

0.5

Fucose (outer arm)

α1-3,4

GlcNAc

0.8

Fucose (outer arm)

α1-2

Gal

0.5

Neu5Ac

α2-3,6

Gal

0.7–1.2

Fig. 5.

N-glycans from secretory component of human secretory IgA were separated into (A) neutral and mono- and disialylated fractions by WAX-HPLC, and then (B) fractions were collected and run on NP-HPLC. The majority of the structures were monosialylated. Note that the two disialylated structures co-eluted with monosialylated structures. (Reproduced from ref.10 with permission from The American Society for Biochemistry and Molecular Biology.)

4 Notes

  1. 1.

    High-purity reagents are used throughout, e.g., AnalaR, Ultrapure, or greater than 99%.

     
  2. 2.

    Unless otherwise stated, all solutions are prepared in water that has a resistivity of 18 MΩ-cm, is particle-free (>0.22 µm), and has a total organic content of less than ten parts per billion.

     
  3. 3.

    Sarstedt (Sarstedt Aktiengesellschaft & Co., Nümbrecht, Germany) microtubes are used throughout, as they give very little contamination on the HPLC.

     
  4. 4.

    It is important to use syringes that do not have any silicone rubber, as this leads to contaminating peaks in the sample.

     
  5. 5.

    All solutions are stored at room temperature unless otherwise stated.

     
  6. 6.

    Buy small bottles of TEMED as this may decline in quality after opening, lengthening the time gels take to polymerize.

     
  7. 7.

    It is important to use acetonitrile with low fluorescence; otherwise, a sloping baseline is seen from the gradient elution.

     
  8. 8.

    Add a small piece of clean foam rubber packaging material to destain 2, as this greatly speeds up the destain process.

     
  9. 9.

    Freezing the gel pieces helps break down the matrix a little so that more of the GP is accessible to the PNGaseF.

     
  10. 10.

    An alternative cleanup method is to use LudgerClean S/GlykoClean S glycan purification cartridges (Ludger or Prozyme) instead of 3MM chromatography paper.

     
  11. 11.

    When clean surfaces are required, e.g., for drying the paper, use clean aluminium foil as this will not contaminate the sample.

     
  12. 12.

    When drying volumes of more than 0.5 mL, recovery of the sample can be improved by redissolving the sample in reducing volumes of water, vortexing, and spinning before redrying. For example, starting with 2.5 mL, the sample is redissolved with the following volumes: 1 mL, 0.5 mL, 250 µL, 100 µL, or until the sample reaches the volume required.

     

References

  1. 1.
    Parekh, R. B., Dwek, R. A., Sutton, B. J., et al. (1985) Association of rheumatoid arthritis and primary osteoarthritis with changes in the glycosylation pattern of total serum IgG. Nature 316, 452–457.CrossRefPubMedGoogle Scholar
  2. 2.
    Block, T. M., Comunale, M. A., Lowman, M., et al. (2005) Use of targeted glycoproteomics to identify serum glycoproteins that correlate with liver cancer in woodchucks and humans. Proc. Natl. Acad. Sci. USA 102, 779–784.CrossRefPubMedGoogle Scholar
  3. 3.
    Butler, M., Quelhas, D., Critchley, A. J., et al. (2003) Detailed glycan analysis of serum glycoproteins of patients with congenital disorders of glycosylation indicates the specific defective glycan processing step and provides an insight into pathogenesis. Glycobiology 13, 601–622.CrossRefPubMedGoogle Scholar
  4. 4.
    Kuhn, P., Tarentino, A. L., Plummer, T. H., Jr., and Van Roey, P. (1994) Crystal structure of peptide-N4-(N-acetyl-beta-D-glucosaminyl)asparagine amidase F at 2.2-A resolution. Biochemistry 33, 11,699–11,706.CrossRefPubMedGoogle Scholar
  5. 5.
    Navazio, L., Miuzzo, M., Royle, L., et al. (2002) Monitoring endoplasmic reticulumto-Golgi traffic of a plant calreticulin by protein glycosylation analysis. Biochemistry 41, 14,141–14,149.CrossRefPubMedGoogle Scholar
  6. 6.
    Tekoah, Y., Ko, K., Koprowski, H., et al. (2004) Controlled glycosylation of therapeutic antibodies in plants. Arch. Biochem. Biophys. 426, 266–278.CrossRefPubMedGoogle Scholar
  7. 7.
    Wormald, M. R., Rudd, P. M., Harvey, D. J., Chang, S. C., Scragg, I. G., and Dwek, R. A. (1997) Variations in oligosaccharide-protein interactions in immunoglobulin G determine the site-specific glycosylation profiles and modulate the dynamic motion of the Fc oligosaccharides. Biochemistry 36, 1370–1380.CrossRefPubMedGoogle Scholar
  8. 8.
    Takahashi, N., Ishii, I., Ishihara, H., et al. (1987) Comparative structural study of the N-linked oligosaccharides of human normal and pathological immunoglobulin G. Biochemistry 26, 1137–1144.CrossRefPubMedGoogle Scholar
  9. 9.
    Homans, S. W., Ferguson, M. A., Dwek, R. A., Rademacher, T. W., Anand, R., and Williams, A. F. (1988) Complete structure of the glycosyl phosphatidylinositol membrane anchor of rat brain Thy-1 glycoprotein. Nature 333, 269–272.CrossRefPubMedGoogle Scholar
  10. 10.
    Parekh, R. B., Dwek, R. A., Rudd, P. M., et al. (1989) N-glycosylation and in vitro enzymatic activity of human recombinant tissue plasminogen activator expressed in Chinese hamster ovary cells and a murine cell line. Biochemistry 28, 7670–7679.CrossRefPubMedGoogle Scholar
  11. 11.
    Küster, B., Wheeler, S. F., Hunter, A. P., Dwek, R. A., and Harvey, D. J. (1997) Sequencing of N-linked oligosaccharides directly from protein gels: in-gel deglycosylation followed by matrix-assisted laser desorption/ionization mass spectrometry and normal-phase high-performance liquid chromatography. Anal. Biochem. 250, 82–101.CrossRefPubMedGoogle Scholar
  12. 12.
    Radcliffe, C. M., Diedrich, G., Harvey, D. J., Dwek, R. A., Cresswell, P., and Rudd, P. M. (2002) Identification of specific glycoforms of major histocompatibility complex class I heavy chains suggests that class I peptide loading is an adaptation of the quality control pathway involving calreticulin and ERp57. J. Biol. Chem. 277, 46,415–46,423.CrossRefPubMedGoogle Scholar
  13. 13.
    Guile, G. R., Rudd, P. M., Wing, D. R., Prime, S. B., and Dwek, R. A. (1996) A rapid high-resolution high-performance liquid chromatographic method for separating glycan mixtures and analyzing oligosaccharide profiles. Anal. Biochem. 240, 210–226.CrossRefPubMedGoogle Scholar
  14. 14.
    Royle, L., Roos, A., Harvey, D. J., et al. (2003) Secretory IgA N-and O-glycans provide a link between the innate and adaptive immune systems. J. Biol. Chem. 278, 20,140–20,153.CrossRefPubMedGoogle Scholar
  15. 15.
    Zamze, S., Harvey, D. J., Chen, Y. J., Guile, G. R., Dwek, R. A., and Wing, D. R. (1998) Sialylated N-glycans in adult rat brain tissue—a widespread distribution of disialylated antennae in complex and hybrid structures. Eur. J. Biochem. 258, 243–270.CrossRefPubMedGoogle Scholar

Copyright information

© Humana Press Inc. 2006

Authors and Affiliations

  • Louise Royle
    • 1
  • Catherine M. Radcliffe
    • 1
  • Raymond A. Dwek
    • 1
  • Pauline M. Rudd
    • 1
  1. 1.Department of Biochemistry, Glycobiology InstituteUniversity of OxfordOxfordEngland

Personalised recommendations