Glycoanalysis Protocols

Characterization

of Protein Glycosylation

Elizabeth F. Hounsell 1. Introduction The majority of protems are posttranslationally modified, the most sigmficant change being glycosylation, i.e., the attachment of one or more oligosaccharade chains. Because of their long history, but also relative neglect until recently, the terminology for saccharides is diverse. Also a major problem in the glycosciences is that many different methods are necessary for oligosaccharide analysts, and this does not at first seem straightforward. I hope this chapter will demystify the structures and the analysis of glycoconjugates (glycoproteins, GPI-anchored proteins, glycolipids, and proteoglycans). The terminology is in fact easy to follow. It has simple beginnings: from glucose comes the generic term glycose, which 1sused m words such as glycosidic ring, glycoprotem, and so forth; from sucrose (a disaccharide of glucose and fructose) comes the word saccharide and, hence, oligosaccharide chain. In addition to glucose (Glc), there are seven other possible orientations of hydroxyl groups m hexoses of the formula C6Hi206 (from whence comes the term carbohydrate) m the series allose (All), altrose (Alt), Glc, mannose (Man), gulose (Gul), idose (Ido), galactose (Gal), talose (Tal). However, in addition to hydroxyl groups on the ring carbons, there are also acetamido groups (Fig. l), e.g., at C-2 m N-acetylglucosamine (GlcNAc) and N-acetylgalactosamine (GalNAc), and at C-5 in N-acetylneuramimc acid (NeuAc). There may also be present sulfate and phosphate esters. Other commonly occurring monosaccharides are the 6-deoxyhexose fucose (Fuc), the pentose xylose (Xyl), and the C-6 carboxyl uranic acids, glucuronic acid (GlcA), iduromc acid (IdoA), and galacturomc acid (GalA). The monosaccharides are linked together between the hydroxyl groups numbered around the glycosidic ring as shown in Fig. 1 and with a or p (anomeric) configuration, depending on the ring geometry (4C1 or lC4 From

Methods m Molecular Bology, Vol 76 G/ycoana/ys/s Protocols Edlted by E F Hounsell 0 Humana Press Inc , Tolowa, NJ

Hounsell

B

Fig. 1. (A) There are two alternative forms for portraymg monosacchartdes as shown here for P-n-N-acetylglucosamme (GlcNAc). Different monosaccharides vary by the number and orientation of then functtonal groups, I.e., OH, NHAc, and the like Compared to GlcNAc, GalNAc has the C-4 hydroxyl group above the plane of the ring. In addition to linkage to each other via one or more (giving branching) hydroxyl group, monosaccharides and ohgosaccharides are also linked to protein and hpid The mam linkages are GalNAca to the hydroxyl group of Ser or Thr (O-linked, mucm type), Xylcl to the hydroxyl group of Ser (proteoglycan type), GlcNAcS to the acetamrdo nitrogen of Asn (N-linked) or to the hydroxyl group of Ser (see Chapter 2), and GlcP to ceramtde (glycolipids). (B) Stalic acids are a family of monosaccharides where R = CH&!O-(N-acetylneurammtc acid) or CH20H-CO+V-glycolylneurammic acid); the hydroxyl groups can be substituted with various acyl substttuents, and those at C-8 and C-9 by additional sialic acid residues

for hexopyranosrde rings) and linkage above or below the plane of the rmg (Fig. 1). The analysis of glycoconjugates follows approximately the progresston in this and subsequent chapters of the book, i.e., detection of the presence of glycosylatron IS achieved by colorrmetric analysis or the use of glycosylatronspecific enzymes, the glycosyltransferases (e.g., to add radroactrvely labeled sugars; Chapter 2) and the glycosrdases; exoglycosidases to remove monosaccharides sequentially from the end distal to the conjugate linkage (Chapters 4, 14, and 16) or endoglycostdases to cleave wrthm the ohgosaccharide chain or at the conJugate-oligosaccharide linkage (Chapters 4-6 and 8). Ohgosaccharides or monosacchartdes released by enzymatic or chemical methods are sepa-

Pro teln Glycosyla tion

3

rated by high-performance llqutd chromatography (HPLC), htgh pH amonexchange chromatography (HPAEC) or gas-hqutd chromatography (GC). These methods are complemented by lectm affirnty chromatography (Chapter 3), methylation analysis (Chapter 6), and gel electrophoresrs (Chapter 8). Discussed m the present chapter are mass spectrometry (MS) and nuclear magnetic resonance (NMR) spectroscopy for the detection and charactertzatron of oligosacchartdes, glycopeptides, glycoprotems, glycolipids, and so forth. The molecules classtcally called glycoproteins comprise mammalian serum and cell membrane glycoprotems of an approxtmate molecular weight range of 20-200 kDa, havmg oligosacchande chains linked to the hydroxyl group of Ser/Thr or the nitrogen of Asn, i.e., 0- and N-linked, respectively, making up from 10-60% by weight. Mucins are traditionally defined as high molecular weight glycoproteins of lo6 kDa upward having >60% oligosaccharide, which is mainly O-linked via GalNAc-containing ollgosacchartde cores. Proteoglycans (see Chapter 9) also have a high carbohydrate/protein ratio. Their glycosammoglycan chains are disaccharide repeating umts, which, m most cases (i.e., heparin, heparan sulfate, chondromn sulfate, and dermatan sulfate), have alternating uromc acid and amino sugar residues, and a large degree of sulfation (the excepttons are unsulfated hyaluromc actd and keratan sulfate whtch is a sulfated Gal-GlcNAc repeat). The distinction between these categories of glycoconjugates 1sbecoming increasingly blurred; they can now be seen as a spectrum of the varying glycosylation patterns occurrmg on high and low molecular wetght, secreted,and cell-surface glycoprotems. As examples of this, classical mucm and proteoglycan sequences can occur on cell-membraneattached protems of relatively low molecular weight, and glycoproteins and proteoglycans are found m forms attached to the membrane by lipid-linked glycosylphosphatidylinositol (GPI) anchors. GPI-anchored glycoproteins were first found in trypanosomes, but are now known as a common membrane anchor in mammalian cells as described in Chapter 14. The present book largely restricts its analysis to mammalian glycoconjugates, but the methods are equally applicable to the glycoconjugates of microorganisms, some of which were discussed in the first edition (I). 1.1. How Do You Know You Have a Glycoprotein? Table 1 shows the different types of methods that can be used for the identification of glycosylatron. Oxidatton wtth periodate is a classical method for ohgosaccharide detection, e.g., the periodate-Schiff reagent (PAS) and Smith degradation, more recently adopted as part of a microsequencing strategy for structural analysis (2) and as commercial kits for glycoprotein detection in conjunction with lectms or antibodies (3). The phenol-sulfuric actd assay can be carried out at mtcroscale m a multtwell titer plate and read by an ELISA plate

Hounsell Table 1 Examples

of Analysis

Techniques

that Detect Carbohydrates

Blologlcal Release of monosaccharides by exoglycosldases Release of ohgosaccharldes by endoglycosidases Metabolic labeling with 3% or 3H monosaccharides Addition of monosaccharides by glycosyl transferases Binding to lectms or antlcarbohydrate antibodies Physicochemlcal Characteristic molecular weight by MS Characteristic chromatographic profile Characteristic signals m a NMR spectrum Chemical Oxidation with sodium metapenodate, which cleaves specifically between two adjacent hydroxyl groups (as m PAS) Phenol-sulfuric acid charring of mono- or ohgosaccharldes having a hydroxyl group at C-2 Reduction of mono- or ollgosaccharldes having a free reducing end after release from protein or hydrolysis of glycosldlc bonds Addltlon of a chemical label by reductive ammatlon. Nitrous acid cleavage of oligosaccharides at non-N-acetylated hexosamine residues Detection of polysulfated oligosaccharldes by dlmethylmethylene blue stammg

reader to detect down to 500 ng of monosaccharides having a C-2 hydroxyl group (e.g., Gal, Man, Glc). Reduction methods (concomitant with oligosaccharlde release for O-lmked chains) can be used to detect oligosacchandes specifically by mtroductlon of a radloactlve label and purification on a phenylboromc acid (PBA) column (4). High-sensltlvtty analysis can also be achieved by the addition

of a fluorescent label by a related technique called reductive amination (see Chapters 6-8 and 15). This relies on the fact that a reduced chain can be oxldlzed by periodate to gwe a reactive aldehyde for linkage to an amine-containing compound, or that free reducing sugars exist for part of the time m the open-cham aldehyde form. Derivatives chosen include amino-lipids for TLC overlay assays and TLC-MS analysis (4,5), UV-absorbing groups that also give sensltwe MS detection (5,6), and sulfated aromatic ammes for electrophoretlc separation (7,s). These can be detected down to the picomole

level

7.2. What Type of Oligosaccharide Sequences Are Present? Essential in any analysis strategy IS an initial screen for the types of ohgosaccharlde chain present, e.g., 0- or N-linked chains, and also for the pres-

Protein G/ycosy/ation

5

ence of any labile chemical linkages that might be destroyed by the subsequent analysts techniques used. High-sensitivity analysis by HPLC or HPAEC (9,ZU) can be achieved (see Chapters 5-7). However, the analysts method described m the present chapter using trtmethylstlyl ethers of methyl glycosides is the most widely applicable, bemg able m one run to identify pentoses (e.g., ribose, xylose, arabmose), deoxyhexoses (e.g., fucose, rhamnose), the hexoses, hexosamines, uromc acids, and siahc acids by gas-liquid chromatography (GC). GC of choral dertvatives (12) can be additionally used to determine the D and L configurations of monosaccharides. The technique of GC-MS analysis of partially methylated alditol acetates (see Chapter 6) is also a very useful technique that can identify the hydroxyl group, through which each monosacchartde is lmked, thus establishing their presence in a chain and giving vital structural mformation. This type of analysis can now be conveniently performed on bench-top GC-MS equipment at the picomole-to-nanomole level. Obtaining a high-field MS analysis of released oligosaccharide chams in their native form, e.g., by fast-atom bombardment (FAB), liquid secondary ion (LSI), matrix-assisted laser desorption (MALDI), or electrospray (ES) MS, is very useful for discovering any labile groups that would be removed by denvatizatlon. Permethylated ohgosaccharides, available as part of the route to partially methylated alditol acetates, can also be analyzed by these techniques to give additional sequence mformation. Alternative derivatives are peracetylated ohgosaccharides, which are readtly formed and extracted to give very clean samples for MS analysis (12). High-sensitivity detection of high molecular weight molecules down to a few picomoles of material can be achieved by the largest mass spectrometers, particularly of ohgosaccharides derivatized at the reducing end as discussed above. MS methods for analysis of oligosaccharides, glycopeptides, and glycoproteins are discussed below and in Chapters 2, 13, and 14. 7.3. What Is the Best Strategy for Release of Oligosaccharide Chains? When mmal clues regarding oligosaccharide types have been gamed, confirmatory evidence can be obtained by specific chemical or enzymatic release. Both types of methods have been researched extensively over the past decade to achieve a htgh degree of perfection m minimizing any nonspecific side reactions while maximizmg oligosaccharide yield. To obtain typical N- and O-linked oligosaccharides, chemical release can be best achieved by hydrazmolysis or alkali treatment. Hydrazmolytic cleavage of N-linked chains (13) has been perfected over the last two decades (Chapters 4 and 6-8). At lower temperatures, hydrazmolysis may also be useful for the release of O-linked chains (Chapters 6 and 7), but this step is more universally achieved

6

Hounsell

by mild alkali treatment (B-elimination), e.g., O.OSMsodium hydroxide at 50°C for 16 h, which m the presence of 0. 5-1MNaBH4 yields intact ohgosaccharide alditols (Chapter 11). Alkaline borohydride reduction conditions result m some peptide breakdown, whereas hydrazmolysis for release cleaves the majority of peptide bonds. Enzymatic release leaves the peptide intact and obviates possible chemical breakdown of ohgosaccharides. However, occasionally it may be necessary first to protease-digest to achieve complete oligosaccharide release, and the enzymes may not cleave all possible structures (e.g., when working with plants, algae, fungi, insects, viruses, trypanosomes, mycobacteria, and bacteria). The extent of deglycosylation can be readily judged by the detection methods discussed m Table 1. For proteoglycans and GPI anchors, an additional chemical method of release is the use of nitrous acid (Chapters 9 and 14) to cleave at non-Wacetylated glucosamme residues. Proteoglycan ohgosaccharide sequences are also obtained enzymatically by heparmases and heparatmases (for heparm and heparan sulfate), chondromnases (for chondrotm and dermatan sulfates), or endo+ galactosidases (for keratan sulfate). 1.4. What Does My Glycoprotein

Look Like?

The ohgosacchande chains of glycoprotems are fashioned by a series of enzymes acting m specific sequence in different subcellular compartments. The end product 1sdependent on a number of factors, mcludmg the untial protein message and its processing, availabihty of enzymes, substrate levels, and so on-factors that can vary between different cell types, different species, and different times m the cell cycle. It is therefore important to address the question of glycoprotem structure to specific glycosylation sites and have profiling methods capable of detecting minor changes in structure, which may be important m function and antigemcity. The followmg route is discussed m this and subsequent chapters: 1. Inmal characterization of type and amount of each monosacchartde and lmkage (HPAEC, GC, GC-MS; (ptcomole-nanomole) 2. Release of 0-lmked chains by alkali, alkaline-borohydride for hydrazinolysis and analysis by labeling and HPLC, PBA, or HPAEC 3. Protease digestion (Chapter 6) and analysis of the complete digest by high-field MS (peptide m 20-pmol digest identified) 4 HPLC peptide mappmg (Chapter 6) and microassay for glycopepttdes (see Table 1)

followed by peptide N-terminal ammo acid sequenceanalysis of identified

glycopeptides 5 Endoglycosidase release of N-linked oligosaccharides and chromatographic profiling as dtscussed m Chapters 4-7 followed by MS analysis of the separated ollgosacchartdes and pepttdes.

7

Protein Glycosyla tion

6 NMR analysis of >50 pg chromatographically pure ohgosaccharide or glycopeptide and conformational analysis by computer graphics molecular modelmg and physicochemrcal methods (Chapter 15).

2. Materials 2.7. Periodate

Oxidation

1 0 1MAcetate buffer, pH 5 5, contammg 1 n-r&&5 mM, or 15 mMsodmm periodate (see Notes 1 and 2). 2. Ethylene glycol. 3. Sodium borohydride, tritiated sodium borohydride, or sodium borodeuterrde at 1 mg/mL m 0. 1M sodium hydroxide 4 Glacial acetic acid. 5 Methanol 6 25 mMHzS04. 7 Nrtrocellulose membranes (e g , Scheicher & Schull, Dassel, Germany) or PVDF membranes (Milhpore, Watford, UK) 8 Labeling kit, e g , digoxigenin/antidigoxigenm (DIG) from Boehrmger Mannheim (Mannhelm, Germany) using DIG-succmyl-ammo-caproic acid hydrazide.

2.2. Calorimetric

Hexose Assays

1 2. 3 4 5 6. 7

HZ0 (HPLC-grade) 4% Aqueous phenol. Concentrated H2S04 1 mg/mL Gal. 1 mg/mL Man Orcmol (Sigma, Poole, UK) 2% (v/v) m ethanol containing 5% of H2S04 Resorcmol (Sigma) 5 mL 2% (w/v) m 45 mL 5M HCl and 125 mL 0 1M Cu II SO4 made up 4 h prior to use. 8 Glass Silica 60 TLC Plates (Merck, Poole, UK)

2.3. GC Composition 1, 2. 3 4. 5. 6

Analysis

0.5M Methanohc HCl (Supelco, Bellefonte, PA) Screw-top PTFE septum vials. Phosphorous pentoxide. Silver carbonate (Pierce and Warrmer, Chester, UK) Acetic anhydride. Trimethylsilylatmg (TMS) reagent (Tri-Sil, Pierce, Rockford, IL, or Sylon HTP kit, Supelco: pyrrdine hexamethyldisdazane, trimethylchlorosilane) Caution: corrosive. 7 Toluene stored over 3A molecular sieve 8. GC apparatus fitted with flame ionization or MS detector (see Chapter 6) and column, e.g., for TMS ethers 25 m x 0 22 mm id BP10 (SGE), and for partially methylated alditol acetates, 25 m x 0 22 mm id HP-5MS silicone (Hewlett Packard, Stockport, UK).

8

Hounseli

2.4. O-Linked

Glycosylafion

1 2 3 4 5. 6. 7. 8.

1MNaBH4 m 0 05M NaOH made up fresh Glacial acetic acid. Methanol Cation-exchange column PBA Bond Elut columns (Jones Chromatography, Hengoed, UK) activated with MeOH 0.2M NH,, OH. 0.01, 0.1, and 0.5MHCl HPLC apparatus fitted with UV detector (approx 1 nmol mono- and ohgosaccharides containing N-acetyl groups can be detected at 195-2 10 nm) and pulsed electrochemlcal detector (oltgo- and monosaccharides lomzed at high pH can be detected at plcomole level) Columns, reversed-phase (RP) C1s, ammo-bonded silica, porous graphltlzed carbon (Hypersll, Runcorn Cheshire, UK), CarboPac PA 100, and CarboPac PA1 (Dionex Camberley, Surrey, UK). 9 Eluents for RP-HPLC (9,ZO)* eluent A, 0.1% aqueous TFA, eluent B, acetomtrlle containing 0 1% TFA. 10 Eluents for HPAEC (9,ZO,Z4). 12 5MNaOH (BDH, Poole, UK) diluted fresh each day; 500 mA4 sodium acetate (Aldrich, Gillmgham, UK). After chromatography and detection, salt needs to be removed by a Dlonex micromembrane suppressor or by cation-exchange chromatography before further analysis, e.g , by methylatlon

2.5. NMR Analysis 1 5-mL NMR tubes (Aldrich) 2 D20: 99 96% for repeated evaporation and 100% (Sigma) for the final solution for NMR 3 Acetone 4 Access to 40&600 MHz NMR 5 PC with CD Rom and Web connection and/or high-resolution computer graphics screen plus data processmg, e.g., S.G Indy, Indigo, or O-2 (Silicon Graphics, Theale, UK)

3. Methods 3.1. Perioda te Oxidation 1. Dissolve 0. l-l .Omg glycoprotem m solution, or blot onto mtrocellulose or PVDF membranes m 20 pL of acetate buffer contammg sodium periodate (15 mA4 for all monosaccharides, 5 mA4 for aldltols, and 1 mA4 specifically for oxldatlon of slallc acids) 2 Carry out the periodate oxldatlon m the dark at room temperature for 1 h, 0°C for 1 h or 4°C for 48 h for aldltols, or 0°C for 1 h for slahc acids (see Note 1) Either* 3. Decompose excess periodate by the addltlon of 25 pL of ethylene glycol, and leave the sample at 4°C overnight.

9

Protein Glycosylatlon

4. Add 0 1MNaOH (about 1 5 mL) until pH 7.0 is reached (see Note 2) 5. Reduce the oxldlzed compound with 25 mg of reducing agent at 4°C overnight 6 Add acetic acid to pH 4 0, and concentrate the sample to dryness on a rotary evaporator. 7 Remove boric acid by evaporations with 3 x 100 pL methanol (see Note 3). 8. For Smith degradation hydrolyze the cleaved glycosldlc rings with 25 mMH2S04 at 80°C for 1 h and repeat the periodate oxldatlon step for newly exposed vlcmal hydroxyl groups Or. 9. Follow one of the commercial procedures for labelmg oligosacchandes on gels or for reductive aminatlon

3.2. Calorimetric Assays 3.2.1. Phenol-Sulfuric Acid Hexose Assay 1 Aliquot a solution of the unknown sample containing a range around 1 clg/lO pL mto a microtiter plate (see Note 4) along with a range of concentrations of a hexose standard (Gal or Man, usually l-10 pg) 2 Add 25 pL of 4% aqueous phenol to each well, mix thoroughly, and leave for 5 mm (see Note 5) 3. Add 200 pL of H2S04 to each well and mix prior to reading on a plate reader at 492 nm (see Note 6).

3.2.2. Orcinol Assay for Detection of Hexose-Positive Molecules Spotted onto TLC Plates 1. Spot between 1 and 100 nmol of hexose onto a thm-layer chromatography (TLC) plate. Aluminum-backed high-performance TLC (HPTLC) or normal TLC plates are fine. Ensure that the spot is as dense as possible (multiple additions of small volumes is best for this) using a Hamllton syringe. 2 Spray with orcmol reagent prepared m advance. 3. Incubate at 100°C for 5 mm giving a purple coloration or orange m the presence of fucose

3.2.3. Characterization

of Sialic Acid Residues

1. Hydrolyze oligosaccharldes or glycoproteins with O.OlMHCl for 1 h at 70°C to remove N-glycolyl or N-acetylneurammic acid with mostly intact 0- and N-acyl groups. 2. Hydrolyze with 0.025M (2 h) to 0. 1M (1 h) HCl at 80°C to remove the majority of slalic acids, but with some O- and N-acyl degradation. 3 Hydrolyze with 0.5M HCl at 80°C for 1 h to remove all slahc acids and fucose. 4. Analyze the released siahc acids by HPAEC (see Chapter 6), or spot onto a TLC plate, spray with resorcinol reagent, cover with a glass plate, and heat at 100°C for 5 min

Hounsell

3.3. GC Composition

Analysis (see Note 7)

1. Concentrate glycoprotems or ohgosaccharides containing l-100 pg carbohydrate and 10 pg internal standard (e g , arabmttol or inosttol) m screw-top septum vials Dry m a desiccator containing a beaker of phosphorus pentoxide 2 Place the sample under a gentle stream of nitrogen, and add 200 pL methanohc HCl (see Note 8) 3 Cap immediately, and heat at 80°C for 18 h 4 Cool the vial, open, and add approx 50 mg silver carbonate 5 Mix the contents, and test for neutrality (see Note 9) 6 Add 50 pL acetic anhydride, and stand at room temperature for 4 h m the dark (see Note 10) 7. Spm down the solid residue, and remove the supernatant to a clean vial 8 Add 100 pL methanol and repeat step 7, adding the supernatants together 9. Repeat step 8, and evaporate the combmed supernatants under a stream of nitrogen 10 Dry over phosphorus pentoxide before adding 20 pL trimethylsilylatmg reagent 11 Heat at 60°C for 5 mm, evaporate remaining solvent under a stream of nitrogen, and add 20 p.L dry toluene 12 Inlect onto a standard or capillary GC column (A typical chromatogram is shown m Fig. 2 ) 13 Calculate the total peak area of each monosaccharide by adding mdividual peaks and dividing by the peak area ratio of the internal standard Compare to standard curves for molar calculation determmation

3.4. O-Linked

Glycosylation

(see Note 11)

1 Release O-linked chains by treatment with 0 05M NaOH m the presence of 1M NaBH4 or NaBC3H14 for 16 h at 50°C 2 Degrade excess NaBH4 or NaB[3H]4 by the careful addition with the sample on ice of glacial acetic acid (to pH 7 0) or acetone (1 mL/lOO mg NaBH4) followed by repeated evaporation with methanol 3 Desalt on a cation-exchange column, and analyze by HPLC as described m Chapters 6 and 7 or HPAEC (Chapter 11) (14). Or for microscale identification

of the presence of aldttols.

4. Dissolve the sample m 200 pL 0.2M NH40H, and add to the top of a PBA minicolumn prewashed with MeOH, water and 0 2M NH40H 5 Wash the column with 2 x 100 pL 0.2MNH40H and 2 x 100 pL water 6 Specifically elute the aldttols m 1M acetic acid 7 Evaporate the sample, and re-evaporate with 2 x 100 pL water 8 Carry out periodate oxidation as described usmg conditions suitable for aldnol oxidation, e g ,5 Uperiodate for 5 mm at 0°C or for 48 h at 4°C (see Note 11) 9 Couple the reactive aldehyde to an organic amme of choice as discussed m Subheading 1.1. and Chapters 6 and 7

Protein Glycosylation 11

4 9

L&-L

1

10

d

-L

3 L

Fig 2 GC of trimethylsllyl ethers of methyl glycoside derlvatlves of monosaccharides: peaks 14 fucose; 5,7, mannose, 6,8,9, galactose; 10, glucose; 11, mositol; 12, 14, N-acetylgalactosamme, 13, 15, 16, N-acetylglucosamme (IV-acetylneurammlc acid occurs as a single peak with longer retention than the hexosacetamldo residues, glucuromc and galacturomc acids chromatograph close to glucose and galactose, respectively, rhamnose, xylose and arabmose chromatograph between fucose and mannose.)

3.5. NMR and Conformational

Analysis

1. Evaporate the purified and desalted sample three times from D20 (see Note 12). 2. Take up the sample in 400 pL D20, and add 1 pL of 5% acetone m D20 for each 50 clg sample present. 3. Transfer the sample to an NMR tube, and store capped at 4’C (see Note 13) 4. Carry out standard 1D and 2D ‘H-NMR experiments at 22’C, and assign chemlcal shifts to specific proton signals (see Note 14; Fig. 3) 5. Carry out 2D ‘H-lH mrrelated Spectroscopy (COSY) to assign signals from protons that are directly coupled. “Walk around” the glycosidlc ring from the C- 1 proton, asslgnmg each mdlvldual proton via 3J~,~ couplmg. 6. Carry out a Double Quantum-Filtered (DQF) COSY experiment to provide data similar to those avallable from COSY spectra, but the pulse sequence incorporates a “quantum filter” that reduces the signal intensity of uncoupled nuclei (smglets) and also gives better resolution. 7. Obtain correlations between every spin in a coupled system and not Just those giving rise to 3JH,H couplings, as m COSY/DQF-COSY by TOtal Carrelation SpectroscopY (TOCSY) experiments Magnetlzatlon IS transferred around the glycosldlc rmg until it IS disrupted by adjacent C-H with small coupling constants

12

Hounsell

c rc sglon I Gal

c Glc a

..J

-L---d 50

NeuAc

I 30

H3 eq

20

10

w-n

Fig. 3. The ID proton (‘H) NMR spectrum in D20 at 22°C of the ohgosaccharlde Gall3 I-3GlcNAcb l-3Gall3 1-4Glcall3 IT6 12~3 NeuAca NeuAca (ppm referenced from acetone at 2.225 ppm).

8 Carry out Triple Quantum-Filtered (TQF-COSY) expertments for ohgosaccharides wtth hydroxymethylene systems, e g., m hexopyranosides (H-6, H-6’, H-5), pentoses (H-5, H-5, H-4), and stalic acids (H-9, H-9’, H-8 and H-3,,, H-3,,, H-4) TQF-COSY experiments use a “spin filter” conceptually similar to that employed m DQF-COSY expertments, however, this results m spectra whose chief signals are those mvolvmg three or more mutually coupled spins 9 Assign through space interactions by Nuclear Qverhauser Effect SpectroscopY (NOESY) to detect spatially close protons not physically linked Use 50650 ms

Protein Glycosylation

10.

11.

12.

13

14.

13

mixing times for lower to higher molecular weight (see Note 15) Quantltate intensities for the crosspeaks to generate proton-proton distance constraints Input these constraints mto distance geometry and molecular dynamics packages to give “structures” consistent with the NOE data For small peptldes and carbohydrates, carry out Rotating frame Overhauser Effect Spectroscopy (ROESY) to measure “through space” correlations to obtain qualltative proton-proton distance mformatlon. Use Eeteronuclear iVJultlple-Bond Gorrelatlon (HMBC) to detect long-range hetereonuclear connectlvltles. Sequence the carbohydrate parts of glycopeptides by using the 13C-‘H couplings between glycosldlc bonds. Carry out HMBC expenments with lSN-labeled peptide to correlate the amide N with the Ca proton Assign overcrowded 2D ‘H-‘H correlated experiments by using the yeteronuclear Multiple Quantum Coherence (HMQC) experiment or the Heteronuclear Single Quantum Coherence (HSQC) experiment, which gives correlations between carbon and directly attached protons (since 13C chemical shifts have better dlsperslon, this allows easier spectral assignment of both nuclei) Asslgn the chemical shifts using data m the literature (15-I 7) and from Sugarbase on URL http.//bocwww them ruu.nl which also networks to the Complex Carbohydrate Structure Database (CCSD, CarbBank) Input the NOE and structural informatlon into computer graphics molecular models built using commercial software packages with added parameters m the force fields for monosaccharides (18)

3.6. Mass Specfromefry The principle of MS is that sample molecules (M) are ionized, and a proportion of the molecular ions (M+ or M-) dissociates forming fragments, fl, f2, and so forth. The masses of the molecular ions and fragment ions are determined and plotted against abundance. MS is then the separation of gas-phase ions according to their mass-to-charge (m/z) ratio, and detection and recording of the separated ions. Ionization is carried out m a matrix for the followmg reasons: 1 2. 3. 4. 5

Isolate single analyte molecules Prevent aggregation. Create a removable platform. Create a medium for ionization. Produce gas-phase ions from the sample.

The different

ionization

methods

are as follows.

The choice of technique

and detector device is described m Table 2. 3.6.7. Fast Atom Bombardment (FAB) or Liquid Secondary /on (LSI) MS Bombardment with high-energy particles of either neutral atoms (Ar; Xe) S-10 keV or ions (AP, Xe+, CS) up to 100 keV. Identify all the predicted peptides that are not glycosylated m about 20 pmol of a protease digest

14 Table 2 Mass Spectrometric Technique PD FAB/LSI MALDI ES

Hounsell Techniques Detection method TOF” Double-focusing TOF“ Quadrupole

Instrument

Molecular weight 5-20 kDa 500-3 500 kDa Up to 200 kDa Up to 100 kDa

aTlme-of-fllght

(Fig. 4). Analyze released oligosaccharides either m native form m the negative mode or when derlvatized m the positive mode

3.6.2. Plasma Desorption (PO) Iomzatlon mvolves the use of Cahformum 252, the energetic particles of whtch are absorbed by the sample causing vibrational excitation and desorption. The particles hit a thm uniform sample on a foil or mtrocellulose matrix to give ions.

3.6.3. Matrix-Assisted Laser Desorption Ionization (MALDI) The solid matrix absorbs most of the laser energy and decomposes leaving the analyte molecules free m the expanding matrix plume. Use this technique optimally for glycoprotem analysis.

3.6.4. Electrospray (ESMS) A flow of sample solution is pumped through a narrow-bore metal capillary giving a mist of fine, charged droplets. More than one charge can be acqutred by a glycoprotem, for example, giving an envelope of m/z ions, which 1s deconvoluted by the MS software to give the molecular weight. The molecular weights obtained are therefore highly accurate, and large mass molecules can be analyzed (high m, but also high z). Use this technique for glycopeptldes, peptides, and released oligosaccharides m underlvatized form.

4. Notes 1 It 1s important that the periodate oxidation 1s carried out in the dark to avoid unspecific oxldatlon. The periodate reagent has to be prepared fresh, since it is degraded when exposed to light 2. Periodate oxldatlon is one of the most reliable chemical reactions in carbohydrate chemistry Sugar residues with hydroxyl groups m vlcmal position are cleaved quantitatively between the carbons, and aldehydes are formed. The aldehydes are subsequently reduced

(%) ~WJ~~Ul

~~!Pml

Hounsell

16

3 Addmon of methanol m an acidic environment leads to the formation of volatile methyl borate. 4 Lmbro/Titertek plates (ICN) with well volume (350 pL) give consistently low backgrounds 5. Do not overfill wells during the hexose assay, since the cont. H2S04 will severely damage the microtiter plate reader if spilled. 6 Exercise care when adding the concentrated H2S04 to the phenol/aldttol mixture, since it is likely to “spit,” parttcularly m the presence of salt. It 1s advisable to perform the phenol-sulfuric acid procedure m a fume cupboard on a sheet of foil and wearing gloves and goggles 7 The use of methanohc HCl for cleavage of glycostdtc bonds and oltgosaccharide-peptide linkages yields methyl glycosides and carboxyl group methyl esters, which give acid stability to the released monosaccharides, and thus, monosacchartdes of different chemical labtllty can be measured m one run. An equihbrium of the a and j3 glycosides of monosacchartde furanose (t) and pyranose (p) rings IS achteved after 18 h, so that a charactertstic ratto of the four possible (fa, fp, pa, and pp) molecules IS formed to atd m unambiguous monosaccharide assignment. If required as free reducing monosaccharides (e g , for HPLC), the methyl glycoside can be removed by hydrolysis. 8 The reagent can be obtained from commercial sources or made m the laboratory by bubblmg HCl gas through methanol or by adding acetyl chloride to methanol 9. Solid-silver carbonate has a pmk hue in an acidic environment, and therefore, neutrality can be assumed when green coloration is achieved 10 The acidic conditions remove N-acetyl groups that are replaced by acetic anhydnde This means that the ortgmal status of N-acylation of hexosammes and siahc acids is not determmed in the analysis procedure Direct re-l\r-acetylation by the addttton of pyrtdme-acetic anhydrrde 1: 1 m the absence of silver carbonate can be achieved, but this gtves more variable results. 11 Prediction of potential 0-glycosylatton sttes on proteins can be carried out at http llwww cbs.dtu dk/servtces/NetOGlyc. Standard O-lurked chains, I e., those having the linkage 6 GalNAcal-SerlThr 3 are cleaved by periodate at the C-4-C-5 bond, thus givmg a characteristic product for chains linked at the C-3 and/or C-6 12 Evaporation from D20 serves to exchange all OH and NH groups to OD and ND, and therefore, only the CH protons are detected The experiments can be repeated in Hz0 contammg 1% D20 as a spm lock If samples must be cooled to <4”C, acetone/H20 can be used as the solvent.

Protein Glycosylatron

17

13. NMR tubes can crack if stored containing D20 at C4”C. Freeze on their sides, or keep at 4°C. If samples must be stored for more than 3 mo, add 0.02% sodium azide evaporated from D20 14. The majority of the literature data for ‘H-NMR of ohgosaccharides are for analysis at 295 K (22’C). Additional experiments can be performed at different temperatures to reveal signals near the HOD resonance. Chemical shifts are usually given with respect to the acetone CHs signal at 2.225 ppm at 295 K. 15 Oligosaccharides in the size range tetra- to heptasaccharide may not be detected by standard NOE experiments owmg to cancellation of positive and negative signals ROESY experiments should be used instead 16. The first generation of computer graphics software for molecular modeling does not adequately cater to ohgosaccharides. Several force fields have now been customized for oligosaccharides including TRIPOS and AMBER (IS). The required monosaccharide data can now be obtained from MONOBANK at the Institut National de la Recherche Agronomique (Nantes, France) There IS also a package called SWEET available on the web which provides a molecular model over the Internet.

Acknowledgments The author thanks Gall Evans, Mike Davies, David Bailey, and David Renouf for their help in preparing the chapter and for research support. References 1 Hounsell, E. F. (ed ) (1993) Glycoprotezn Analyw zn Bzomedzczne, vol 14, Methods in Molecular Biology Humana, Totowa, NJ 2. Stoll, M S., Hounsell, E. F , Lawson, A. M., Chai, W , and Feizi, T (1990) Microscale sequencing of O-linked ohgosaccharides usmg mild periodate oxtdation of aldnols, coupling to phospholipid and TLC-MS analysis of the resulting neoglycohpids. Eur J Blochem 189,499-507.

3. Haselbeck, A. and Hosel, W. (1993) Immunological detection of glycoprotems on blots based on labelmg with digoxigenin, m Methods znMolecular Biology, vol 14, Glycoprotein Anaiysls znBzomedznne(Hounsell, E. F , ed ), Humana, Totowa, NJ, pp 161-173 4. Stoll, M. S. and Hounsell, E. F. (1988) Selective purification of reduced ohgosaccharides using a phenylboromc acid bond elut column potential application in HPLC, mass spectrometry, reductive amination procedures and antigenic/serum analysis. Biomed. Chromatogz 2,249-253. 5 Lawson, A., Chai, W , Cashmore, G. C , Stall, M S., Hounsell, E. F , and Feizi, T (1990) High-sensitivity structural analyses of oligosaccharide probes (neoglycolipids) by liquid-secondary-ion mass spectrometry. Carbohydr Res 200,47-57 6. Poulter, L., Karrer, R , and Burlingame, A L. (1991) n-Alkylp-ammo benzoates as derivatismg agents in the isolation, separation and charactensation of submicrogram quantities of oligosaccharides by LSIMS. AnaE Bzochem 195, l-l 3.

78

Hounsell

7. Lee, K. B , Lmdhart, R J , and Al-Haktm, A (1991) Gel and captllary electrophoresrs based oligosaccharrde sequencing Glycoconjugate J 8,250,25 1 8 Sullivan, M T , Klock J , and Stack R J (199 1) Charactertsatron of N-linked glycoprotem ohgosaccharrdes by fluorescence assisted carbohydrate electrophoresrs (FACE) GlycoconJugate J. 8,249 9 Townsend, R R (1995)Analysrs of glycoconJugates usmg high pH amon-exchange chromatography, m Journal of Chromatography Library, vol 58 Carbohydrate Analyszs (Elassi, 2 , ed ), Elsevler, The Netherlands, pp 18 l-209. 10 Davies, M J and Hounsell, E F (1996) Carbohydrate chromatography* towards yoctomole sensmvlty Boomed Chromatogr: 10,285-289 Gerwtg, G J., Kamerlmg, J P, and Vhegenthart, J F G (1979) Determmatron of 11 the absolute configuration of monosacchartdes m complex carbohydrates by caplllary GC. Carbohydr Res 77, l-7 12 Dell, A (1987) FAB-MS of carbohydrates. Adv Carbohydr. Chem Bzochem. 45, 1972. Mrzuochr, T (1993) Mtcroscale sequencing ofN-linked oligosacchartdes of glyco13 protems using hydrazinolysrs, Bto-gel P-4 and sequential exoglycostdase drgestron, InMethods znMolecularBlology, vol. 14, Glycoproteln Analysis znBlomedicwe (Hounsell, E F , ed.), Humana, Totowa, NJ, pp. 55-80 14 Campbell, B. J , Davies, M. J , Rhodes, J. M , and Hounsell, E F. (1993) Separatton of neutral ohgosaccharide aldrtols from human meconium using hrgh pH amon exchange chromatography J Chromatogr 622,137-146 15 Hounsell, E F. (1995) ‘H-NMR m the structural and conformatronal analysts of ohgosaccharrdes and glycoconJugates Prog NMR Spectroscopy 27,445-474 16 Hounsell, E F (1994) Physrcochemtcal analysis of olrgosaccharlde determinants of glycoprotems Adv Carbohydr Chem Blochem SO,3 1 l-350 17 Hounsell, E F and Wrrght, D. J (1990) Computer-asststed mterpretatron of ‘H-NMR spectra m the analysis of the structure of ohgosacchartdes Carbohydr Res 205, 19-29 18 French, A. D and Brady, J W , eds (1990) Computer modellmg of carbohydrate molecules ACS Symposzum Serzes430, American Chemical Society, Washington, DC

Analytical Methods for the Study of OIGlcNAc Glycoproteins and Glycopeptides Kenneth D. Greis and Gerald W. Hart I. Introduction Over the past decade, numerous nuclear and cytosolic glycoproteins have been shown to be modified by single N-acetylglucosamme (0-GlcNAc) residues attached to the side chain hydroxyl groups of serme or threonme (1-3). This unique postranslational modification (0-GlcNAcylation), described m lymphocytes by Torres and Hart (4), can be found on a wide range of proteins including cytoskeletal proteins (5-Q nuclear pore proteins (9-13), transcription factors (141, RNA polymerase II (15), oncoproteins (16,17), and viral proteins (J&20), to name a few. In addrtion to the 0-GlcNAc modrficatron, all of these proteins are known phosphoprotems and many have been reported to be regulated and/or form reversible protem complexes based on their state of phosphorylation. These observations suggest that 0-GlcNAcylation also plays a role m the assembly/disassembly of protein complexes and/or the regulation of enzymes, perhaps by acting as an antagonist to phosphorylation. In fact, several studies have shown that phosphorylation and 0-GlcNAcylation can occur at the same sites on a given protein (1.5,21). For example, a recent study documents that the c-Myc oncoprotein is 0-GlcNAcylated at threonme 58, a known site of phosphorylatron and mutational hotspot for c-Myc m human Burkitt and AIDS-related lymphomas (21). Furthermore, 0-GlcNAcylation appears to be a key regulatory modification since it is highly dynamic (22,23), responsive to external cellular stimuli (24), and regulated in a cell cycledependent manner (25). However, direct evidence for the function of O-GlcNAcylation has been limited, m large part, because of the difftculty and the amount of protein required for mapping sites of glycosylation as a prerequisite to From

Methods m Molecular B/ology, Vol 76 G/ycoanalys/s Protocols Edited by E F Hounsell 0 Humana Press Inc , Totowa, NJ

19

20

Greis and Hart

molecular mampulations such as site-directed mutagenesis. The primary method for mappmg sites of O-GlcNAcylation requires the enzymatic transfer of a [3H]galactose with bovine milk galactosyltransferase to tag the O-GlcNAcmodified protein, generation of glycopeptides by proteolysts, purification of the glycopeptides by several rounds of high-performance liquid chromatography (HPLC), followed by gas-phase and manual Edman sequencing. Each of these steps is very tedious and often results in significant losses in yield as a result of mcomplete proteolysis, poor resolution of the glycopeptrdes by HPLC, and conservative pooling of HPLC fractions to maximize purity of the glycopeptides. Finally, thts method requires a minimum of about 10 pmol of pure, labeled, glycopeptide to generate unambiguous sequence and to identify the site of glycosylation. Therefore, methods to improve recoveries of glycopeptides and increase the sensmvity of the site analysts assays would be a major breakthrough m the study of O-GlcNAcylation. To expedite the purification procedure and to increase the overall recovery of glycopeptides, we have developed a method to greatly enrtch for the O-GlcNAcylated pepttdes by mteraction with immobilized Ricinus cummtuns agglutmin (RCA I) after m vitro galactosylatton and trypsm digestion (26). The major advantages of this method include the removal of all nonglycosylated peptides produced by proteolysis of the glycoprotein and the removal of any peptides generated by autodigestton of trypsm. The latter allows for the use of larger quantities of trypsin to ensure complete digestion of the glycoprotein and, therefore, increased recovery of glycopeptrdes. In order to increase the sensitivity and directly identify the sites of U-GlcNAcylation on tryptic glycopeptides, we have mvestigated the use of electrospray mass spectrometry (ESMS) Over the past 8 yr numerous mass spectrometry methods have been reported for sequencmg peptides in the low picomole to femtomole range by postsource decay (PSD) and colhston-Induced dtssocratton-tandem mass spectrometry (CID-MS/MS) (27-29). Furthermore, selective tdenttficatton of peptides wtth posttranslation modifications such as phosphorylatron (30,31) and glycosylation (32,33) has also been demonstrated. For example, N-linked glycopeptides can be tdenttfied m crude protein digests by the generation of an oxonium ion of N-acetylhexosamme (m/z 204) m precursor ion scans with CID-MS/MS (32,33). In the case of N-lurked glycopeptides, thus informatton 1s often all that is needed to determine the site of glycosylation since there is likely to be only one asparagme m the context of the N-linked glycosylatton consensus sequence of Asn-X-Ser/Thr (34). However, the ability to acquire mformation about the site of O-lmked glycosylation has been hampered by the fact that no clear consensus sequences have been identified, there are often clusters of serine and threonine residues near sites of glycosylation, and the ionization potential and the CID energy needed to

Study of 0-Glc/VAc and Peptides

21

fragment peptides for MS-based sequencing often resulted in the removal of the O-linked glycans from the peptide backbone. The latter eliminates the production of fragment tons containing the glycan, thus precluding the abthty of identifying the site of glycosylation (35,361. As a result, only a few reports have addressed the use of mass spectrometry techniques to identify the sites of O-linked glycosylation (35-3 7). Initial analysis of O-GlcNAcylated glycopeptides by ESMS indicated that the O-GlcNAc, like other O-lmked glycans, is readily removed by the ionization and fragmentation conditions, thereby eliminating the ability to identify the exact serine or theronme that was O-GlcNAcylated. However, pretreatment of the glycopepttdes with alkali, m the absence of reducing agent, to induce P-elimination of the glycan, results in the conversion of glycosylserme to 2-aminopropenoic acid and glycosylthreonine to 2-amino-2-butenotc acid, each 18 amu less than their intact ammo acid counterpart (38,391. Although peptide degradation 1soften reported as a consequence of alkaline @eltmination (40,41), tt can be mmlmized by omittmg reducing agents commonly used to prevent degradation of the released glycan and by adjusting the pH, the temperature, and the mcubation time of the reaction. The exact site of glycosylation can be directly identified by tandem ESMS sequencing of the P-eltmmated glycopeptide compared with the untreated glycopepttde by a fragment mass shift of 18 amu at the site of glycosylation (36,&j. Whereas the method described m Subheading 3.3. is for analysis of O-GlcNAcylated peptides by capillary HPLC tandem ESMS with sensitivity in the low picomolar range, this method is likely to be applicable to other forms of mass spectrometry sequencing including matrix-assisted laser desorption/ionizatton-time-of-flight (MALDITOF) and fast-atom bombardment (FAB)MS, but at different levels of sensitivity. This high-sensittvity method of identifying and analyzing O-GlcNAcylated peptides coupled with the RCA-agarose method for enriching the glycopeptides after proteolytic digestion represents a powerful new tool in determining the sites of O-GlcNAc attachment on proteins of low abundance such as transcription factors and oncogenes. 2. Materials 2.1. Trypsin Digestion of Galactosyltransferase-Labeled Glycoprotein 1. Bovine milk galactosyltransferase-labeled glycoprotem sample (see Note 1 for details). 2 Trypsm (sequencing grade) 3. Trypsin buffer 10 mA4 HEPES, pH 8 1. 4. Phenylmethylsulfonyl fluoride (PMSF).

Greis and Hart

22 2.2. Isolation of Galactosylated Peptides by RCA-Agarose Chromatography

1 RCA I agarose (4-5 mg lectin/mL of settled resin). 2. Phosphate-buffered saline (PBS): 6.7 tipotassium phosphate, pH 7.4, containing 150 n&Z NaCl and 0.02% NaN3. 3. RCA-agarose elution buffer: PBS containing 200 mA4 lactose. 4. Fluorescamine buffers: 10 mM HEPES, pH 8.1; 14% fluorescamine in acetone. 5. Spectrofluorometer. 6. Sep-Pak Cl8 cartridges from Waters. 7. Sep-Pak Buffers: 0.1% formic acid; 0.1% formic acid in 60% CH$N.

2.3. Identification

of Glycopeptides

by Capillary

LC-ESMS

1. Instruments: a. HPLC system equipped with a capillary C,s column. b. Electrospray triple quadrapole mass spectrometer. 2. HPLC buffers: 0.1% formic acid; 0.1% formic acid in 60% CH$N.

2.4. p-Nimina

tion of Glycopep tides

1. p-elimination buffer: 0.2MNaOH. 2. Neutralization buffer: 0.3M acetic acid.

3. Methods 3.7. Trypsin Digestion of Galactosyitransferase-Labeled Glycoprotein (see Note 2) 1. Dilute galactosyltransferase-labeled glycoproteins (see Note 1) into 90 & of trypsin buffer. 2. Add 10 pL of a freshly prepared trypsin solution to give a 2: 1 (w/w) final concentration of glycoprotein to trypsin. 3. Digest at 37°C for 16 h. 4. Inactivate the trypsin with PMSF and boiling for 5 min.

3.2. Isolation of Gaiactosylated Peptides by RCA-Agarose Chromatography (see Note 3 and Fig. I) 1. Equilibrate a column of RCA-agarose (0.7 cm id x 20 cm long) in PBS at room temperature at a flow rate of 0.34.5 mL/min. 2. Wash extensively (30-column vol) with PBS to remove the lactose present in the storage buffer. 3. Apply the trypsin digested glycoprotein onto the column and collect 1 min fractions for four-column volumes (about 50 fraction). This will resolve the monogalactosylated peptides that interact weakly with the lectin from the nonglycopeptides that do not interact with the column. 4. Elute the multiply galactosylated peptides with PBS containing 200 mA4 lactose for two-column volumes (approx 25 fractions).

23

Study of 0-GlcNAc and Peptides Vi 2500

2.5

+ 4

k.$ 1500 8

500

0

41

0

10

20

30

40

Fraction

50

60

70

80

90

(500 pl)

Fig. 1. RCA-agaroseprofile of an in vitro [3H]-galactosylated O-GlcNAc-modified aprotein after trypsin digestion. ([I), total peptide profile asjudged by fluorescamine assay at an emissionof 475 nm; (A), CPM of tritium. 5. Determine the elution position of the tritium-labeled glycopeptides by liquid scintillation spectrometry of a small aliquot (5%) of each fraction. 6. Determine the total peptide profile by a fluorescamine assayof a secondaliquot (2.5%) of each fraction into 2 mL of 10 mA4 HEPES, pH 8.1. Add 0.25 mL of 14% fluorescamine in acetoneand measurethe emissionat 475 nm at an excitation of 390 nm with spectrofluorometer. 7. Pool fractions containing the tritium label. This should include the weakly interacting (monogalactosylated) glycopeptide pool and the lactose-eluted (multiply galactosylated) glycopeptide pool. 8. Desaltand concentratethe glycopeptidepoolson a Sep-PakCts cartridgeequilibrated in 0.1% formic acid.After applying the sample,washwith 10mL of 0.1% formic acid, then elute the glycopeptideswith 5 x 0.5 mL fractions of 60% acetonitrileio.1% formic acid. Assay an aliquot of the flow through, wash, and each elution fraction by liquid scintillation spectrometryto determinerecovery of the glycopeptides. 9. Dry the eluted glycopeptides under reduced pressure (Speed-Vat) and store at -80°C until samplesare ready for MS analysis.

3.3. Identification of Glycopeptides by Capillary L C-ESMS (see Note 4 and Fig. 2) 1. Resuspendthe RCA-agarose enriched glycopeptides in 10 $ of 0.1% formic acid and store on ice.

Greis and Hart

24

1

Extracted ion profile (366)

8.0 211

[M+2H]++ 616

I 200

400

600

loss of 366

800

-

1000

1200

Fig 2. HPLC-ESMS of the glycopeptide YSPTS(GalPl-4GlcNAc)PSK (A) Extracted Ion profile at 366 amu (B) The average massspectrum of scans407-414 at 15.6 mm correspondmgto the elution position of the extracted ion (366) m (A)

2 Inject 1 pL of the glycopeptlde onto a capillary C,s reverse-phaseHPLC column equilibrated m 0 1% forrmc and coupledto the lomzatlon probe of an electrospray massspectrometer 3 Elute the glycopeptldes at 4 pL/mm with an Increasing gradient of CH,CN (o-60%) m 0 1% formic acid developed over 17mm. Collect positive ion spectra at an orifice potential of 70 V by scanning the first quadrapole (Ql) over a mass/charge(m/z) range of 200-2200.

25

Study of 0-GlcNAc and Peptides

4 O-GlcNAcylated pepttdes that had been enzymattcally galactosylated can be identified by the dtagnosttc fragment ton at 366 amu correspondmg to the O-linked GalPI-4GlcNAc (Fig. 2A) This fragment is liberated from a portion of the glycopeptide because of the orifice potential energy at 70 V. A glycopeptide IS identified when mass ions corresponding to the intact glycopeptide, the deglycosylated peptide, and the liberated glycan (366 amu) are present at the same retention time m the mass spectrum (Fig. 2B) In many cases, identtfication of the mass of a trypttc glycopeptrde 1s sufftcrent to predict the sequence of the peptide provided that the primary sequence of the protein is already avatlable. The primary sequence of the predicted glycopeptide can be confirmed by tandem LC-ESMS by generatmg a nested set of fragment ions that correspond to the amino acid sequence (see Sub-

beading 3.5.). 3.4. P-Elimination

of Glycopeptides

(see Note 5)

1 Dilute 4 pL of RCA-agarose enriched glycopeptide from Subheading 3.3., step 1 mto 21 pL of HPLC-grade water 2 Add 25 pL of 0 2M NaOH to give a final concentration of 0 1M NaOH 3 Incubate at 45°C for 4 h and immediately neutraltze the reaction on me with 25 pL 0 3M acetic acid 4 Dry under reduced pressure (Speed-Vat) and store at -80°C until ready for MS analysis

3.5. identification of the P-Eliminated Peptides by Capillary LGESMS (see Note 6 and Fig. 3) 1. Resuspend the P-eliminated pepttde samples m 10 pL of 0 1% formtc acid and store on ice 2 Inject 2 5 pL onto a capillary Cts column and analyze under the identical condttions as the glycopeptide m Subheading 3.3.

3.6. Product Analysis by Tandem Capillary (see Note 7 and Fig. 4)

LC-ESMS

1. In separate analyzes, Inject erther 3 pL of the RCA-agarose glycopepttde pool or 7.5 pL of the P-eliminated peptide pools onto the capillary Cts column The mass spectrometer is set to an orifice potential of 50V m the tandem MS mode such that only the target mass ton ~111 be SubJect to CID and analysis of the fragment ions 2 Elute the peptides from the capillary C,s column as m Subheading 3.3. 3. Analyze the fragment ions to confirm the sequence of the glycopeptides (Fig. 4A) and the positron of the P-ehmmated glycosylserme (2-ammopropenotc acid) or glycosylthreonme (2-ammo-2-butenoic acid) (Fig. 4B).

Greis and Hart

26

100 -

f B

[M+2H]* 425

848 25 a 251 223 n 1 200

598 685 400

600

800

1000

1200

Fig. 3 HPLC-ESMS of the glycopeptlde YSPTS(GalPl-4GlcNAc)PSK after p-elimmation as described in the text. The average of scans 443-449 at a retentton time of 15.5 mm is shown

4. Notes 1 0GlcNAcylated proteins can be specifically labeled on the O-GlcNAc residue(s) by the enzymatic transfer of [‘HIGal from UDP-[3H]Gal m the presence of bovme milk galactosyltransferase. The details of this method are described elsewhere (43-45) Since the subsequent enr&unent of the glycopeptides on the lectm column is dependent on the complete transfer of Gal to all O-GlcNAc residues, the UDP-[3H]Gal should be isotopically diluted with cold UDP-Gal to provide at least a fivefold molar excess of the nucleottde sugar to substrate acceptor Alternatively, excess cold UDP-Gal can be added to complete the reaction after the UDP-[3H]Gal has been grven sufficient time to be transferred to the O-GlcNAc 2 Trypsm is the first choice of proteolyttc enzymes because it results m peptide fragments with an ammo group at each end-one at the ammo-terminus and the other on the side-chain of lysme or argmine at the carboxyl-terminus of the peptides. Since protonation of peptides by ESMS favors ionization at amino groups, sequencmg of tryptic peptrdes by tandem ESMS is greatly enhanced when the charged groups are at opposite ends of the peptrde (46) (see Note 7 for details) The exact volume and concentration of trypsm needed is often protein dependent. Larger reaction volumes and more enzyme can be used, however, care must be taken to avoid overloadmg the RCA-agarose column wtth the digested protein (see Note 3) In addition, excesstve amounts of trypsm or impure trypsm can produce nonspecific cleavages of the protein, often at chymotrypic-like sites (47,48) Furthermore, it is essential to inactivate the trypsm fully before proceeding to the RCA-agarose column to avoid proteolytic degradation of the lectm

A

b,,.)

164 261 348 U8 634 633 720 848

b

46

Gal!31,4Gl;NAc

66

Serme 87

204

200

600

400

800

m/z

2-ammopropenoicacld 69

136

425

SO

200

400

600

m/z

Fig. 4. Tandem mass spectra of the glycopeptide YSPTS(GalP l-4GlcNAc)PSK before and after p-ehminatton. (A) The doubly protonated form of the glycopepttde (M + 2H = 6 16) was fragmented in Q2 and the product ions were analyzed m Q3 of the triple quadruple mass spectrometer as described m the text (B) The doubly protonated form of the P-eliminated glycopepttde (M + 2H = 425) was fragmented and analyzed. The insets m each panel represent the predicted b and y tons fragments from each pepttde Note the loss of the y ton at 418 in (A) and the appearance of a y Ion at 400 m (B) correspondmg to the conversion of the glycosylserme residue (87 amu) to a 2-ammopropenolc acid (691, thus tdenttfying the sue of glycosylatton.

28

3

4

5

6.

Greis and Hart Finally, reduction and alkylation of the glycoprotem prior to proteolytic digestion will often increase the yield of glycopepttdes since the digestion IS more likely to go to completion and the adverse effects of dtsulfide lmkages between pepttdes will be elimmated. It 1simportant to note that the mteraction of the m vitro galactosylated glycopeptides with the RCA lectm column 1s not a bind and elute mteraction The monogalactosylated peptides are only retarded by weak mteractton with the lectin, therefore, care must be taken to use a sufficient bed volume m a long narrow column to enhance the separation of the noninteractmg (nonglyco-) peptides and the weakly mteractmg (monogalactosylated) pepttdes A typrcal elution profile from the RCA-agarose column is shown in Fig. 1 whereas a more detailed descriptron of this lectm mteractton with galactosylated 0-GlcNAc IS provided m Hayes et al (26) Other protein assays can be used to determine the elutton position of the peptides. We chose the fluorescamme assay (49) because of its ease and its high-senstttvtty The method provided was developed on a Sciex API-III triple quadrapole electrospray mass spectrometer eqmpped with an atmospheric-pressure ton source Since the amplitude of the mass-spectral signal 1s based entirely on the concentration of the glycopepttde entering the mass spectrometer, the use of an in-lure capillary HPLC system and a steep elutton gradient of acetomtrtle is crucial to elute the glycopepttdes m a very concentrated form. This allows for the tdenttficatton of peptides in the one to five ptcomolar range to be achieved. Secondly, selective tdentttication of the galactosylated, 0-GlcNAc peptides can be achieved by mducmg partial pepttde fragmentation via an increase m the orifice potential of the mass spectrometer. The opttmal orifice potential to selectively release a small, yet dtagnosttc, portton of the O-linked GalPl-4GlcNAc (366 amu) without stgmticant fragmentatton of the pepttde backbone has been emptrtcally determine to be about 70 V (32.42) Additional background mformatton on ESMS can be found m some recent reviews (27-29) p-elimmation and reduction has been widely used to liberate and study the ohgosaccharide structures of O-lurked glycans (38,39) Although the reduction is necessary to prohibit subsequent “peelmg” reactions on the released oltgosaccharade (SO), it also severely fragments the peptide backbone (40,41) thus prohtbttmg mass analysts of the P-eltmmated peptide However, mild-base treatment m the absence of reducing agent removes the 0-GlcNAc-Gal with only limited cleavage of the pepttde backbone The condmons reported here were emptrtcally derived from several glycopeptide standards (42) p-eltmmation should result m a mass shift ofthe glycopepttde by 384 amu, corresponding to the loss of the protonated GalP14GlcNAc (366 amu) and 1 mol of H20 (18 amu) from the glycosyl-serme or glycosyl-threonme for each site of glycosylation on the peptide The example glycopeptide (M + H = 1232) m Fig. 2 IS converted to a P-eltmmated peptrde (M + H = 848) corresponding to the loss of the Gal/3 1-4GlcNAc and 1 mol of water (Fig. 3)

Study of 0-GlcNAc and Peptldes 7. Once the mass of the glycopeptide is determmed before and after /3-ehmmation (Subheadings 3.3. and 3.5.), the peptides can then be sequenced by CID and analysis of the resultmg fragment ions. The doubly protonated ion of the pepttde is selected m Ql of the mass spectrometer. Fragmentation of the doubly protonated ion will most likely generate two fragment ions from a tryptic digest because of the protonation of the ammo groups at either end of the peptide (see Note 2) The orifice potential is set to 50 V instead of 70 V to prohibit the removal of any of the glycan durmg ionization that would decrease the amount of doubly protonated glycopepttde available for focusmg and fragmentation The selected smgle mass tons from Ql are focused mto the second quadrapole (Q2). Here the ions are fragmented by colhsion with a mixture of inert gas, typically argon or nitrogen After collision of the doubly protonated ions in Q2, the masses of the fragment ions are separated and recorded m the third quadrapole (Q3) Since the most favorable fragmentation event occurs at the amide bonds of the pepttde, two smgly charged ions are produced that correspond to ions carrying a charge at the ammo-terminus of the peptide (b tons) and those with a charge at the carboxyl-termmus of the peptide (y ions) The nested set of b and y ions produced by fragmentation at the various amide bonds differ by the masses of mdividual ammo acid thereby providing the peptide prrmary sequence The mass shaft of 18 amu at the position of a serme (87-69 amu) or threonme (101-83 amu) m the primary sequence after p-ehmmation is diagnostic for the site of O-GlcNAcylatton Figure 4 shows an example of the CID spectrum of a glycopeptide before (Fig. 4A) and after (Fig. 4B) l%ehmmation Further information regarding the theory of peptide sequencmg by tandem ESMS is provrded m some recent reviews (27-29) The concentration of collision gas needed to get optimal fragmentation for sequencing will vary based on type of collision gas, the size of the collision cell, and the ammo acid sequence of the peptide As a result, opttmization of the colhsion gas concentration usmg standard glycopeptides should be done prior to analysts of low abundance samples

References 1 Greis, K D and Hart, G. W (1995) Nuclear and cytosohc glycoprotems, in Gfycoprotezns, vol II (Montreml, J., Schachter, H., and Vhegenthart, J F G., eds ), Elsevier, Amsterdam, m press 2 Hart, G W, Kelly, W G , Blomberg, M A, Roquemore, E. P., Dong, L.-Y D , Kreppel, L , Chou, T., Snow, D , and Greis, K (1994) Nuclear and cytoplasmic glycosylation is ubiquitous and has the hallmarks of a regulatory moditicatton, m Complex Carbohydrates VI Drug Research (Bock, K and Clausen, H , eds ), Munksgaard, Copenhagen, Denmark, pp 280-290 3. Hart, G. W., Kelly, W G , Blomberg, M. A , Roquemore, E P , Dong, L.-Y. D , Kreppel, L., Chou, T-Y , Snow, D , and Greis, K (1993) Glycosylation of nuclear and cytoplasmic proteins is as abundant as phosphorylation, m DNA Rephcatzon and the Cell Cycle (Wieland, F and Reutter, W , eds ), Sprmger-Verlag, New York, pp. 91-103

Greis and /-/art

30

4 Torres, C.-R and Hart, G W (1984) Topography and polypeptide distnbutton of terminal N-acetylglucosamme residues on the surfaces of intact lymphocytes. J Bzol Chem 259,3308-33 17 5 Holt, G. W., Haltiwanger, R S , Torres, C-R , and Hart, G. W (1987) Erythrocytes contam cytoplasmtc glycoprotems O-linked GlcNAc on Band 4.1 J Blol Chem. 262, 14,847-14,850. 6 Dong, D. L.-Y., Xu, Z -S., Chev((rier, M. R., Cotter, R J., Cleveland, D. W , and Hart, G W (1993) Glycosylation of mammalian neurofilaments. Locahzatton of multiple O-linked N-acetylglucosamme moieties on neurofilament polypeptides L and M. J Blol Chem 268, 16,67916,687. 7 Ku, N.-O and Omary, M. B. (1994) Identification of the maJor phystologic phosphorylation site of human keratin 18. potential kinases and a role in filament reorgamzatron J. Cell Blol 127, 161-17 1 8. Chou, C -F., Smtth, A. J., and Omary, M. B (1991) Characterization and dynamics of O-lurked glycosylation of human cytokeratin 8 and 18. J Cell Blol 115,353a.

9 Holt, G. D , Snow, C. M., Senior, A, Haltiwanger, R S , Gerace, L , and Hart, G W (1987) Nuclear pore complex glycoprotems contam cytoplasmically drsposed O-linked N-acetylglucosamme J Cell Blol 104, 1157-l 164 10. Starr, C. M. and Hanover, J A. (1990) Glycosylatton of nuclear pore protein p62 Rettculocyte lysate catalyzes O-lurked N-acetylglucosamine addition m vitro J B~ol Chem 265,6868-6873.

11 Davis, L I and Blobel, G. (1987) Nuclear pore complex contams a family of glycoprotems that mcludes ~62: glycosylation through a previously unidentified cellular pathway Proc Nat1 Acad SCI USA 84,7552-7556 12 Hanover, J A , Cohen, C K , Willingham, M C , and Park, M K (1987) O-linked N-acetylglucosamme 1sattached to protems of the nuclear pore Evtdence for cytoplasmic and nucleoplasmic glycoprotems J Blol Chem 262, 9887-9894

13 Schmdler, M., Hogan, M., Miller, R., and DeGaetano, D (1987) A nuclear specific glycoprotem representative of a umque pattern of glycosylation. J Bzol Chem 262,1254-1260. 14. Reason, A. J., Morris, H. R., Pamco, M., Marars, R., Treisman, R. H , Haltrwanger, R. S., Hart, G W., Kelly, W. G., and Dell, A (1992) Localization of 0-GlcNAc modrficatron on the serum response transcrtptton factor. J Bzol Chem 267, 16,911-16,921 15. Kelly, W G , Dahmus, M E , and Hart, G. W. (1993) RNA polymerase II is a glycoprotein. Modificatton of the COOH-termmal domam by 0-GlcNAc J BzoZ Chem 268, lo,41 6-l 0,424 16. Chou, T-Y, Dang, C V, and Hart, G. W. (1995) Glycosylation of the c-Myc transactivation domam. Proc Nat1 Acad Scz USA 92,44 17-442 1 17. Prrvalsky, M. L. (1990) A subpopulation of the avlan erythroblastosis vu-us v-erbA protein, a member of the nuclear hormone receptor family, is glycosylated J Erol 64,463-466.

Study of 0-GlcNAc and Peptides

31

18. Bjoem, S., Fosters, D. C , Them, L., Wtberg, F. C , Chrtstensen, M., Komlyama, Y., Pederson, A H , and Klsiel, W (1991) Human plasma and recombinant factor VII J Bzol. Chem 266, 11,051-11,057. 19 Benko, D. M., Haltiwanger, R. S., Hart, G. W, and Gibson, W. (1988) Vmon basic phosphoprotem from human cytomegalovirus contams O-linked N-acetylglucosamine Proc Nat1 Acad Scz USA 85,2573-2577 20. Whitford, M and Faulkner, P (1992) A structural polypeptlde of the baculovmus Autographa californlca nuclear polyhedrosis virus contams O-linked N-acetylglucosamme J Viral 66,3324-3329 2 1. Chou, T -Y, Hart, G. W , and Dang, C V. (1995) c-Myc IS glycosylated at threonine 58, a known phosphorylation site and a mutatronal hot spot m lymphomas J Blol Chem 270, 18,961-l 8,965 22 Chou, C -F , Smith, A J , and Omary, M B (1992) Charactertzatton and dynamICS of O-linked glycosylatlon of human cytokeratm 8 and 18. J Biol Chem 267, 3901-3906 23. Chou, C.-F. and Omary, M B (1993) Mttottc arrest-associated enhancement of Olmked glycosylation and phosphotylation of human keratins 8 and 18 J Blol Chem 268,4465-4472 24 Kearse, K. P and Hart, G. W (1991) Lymphocyte activatton induces rapid changes m

nuclear and cytoplasmlc glycoprotems. Proc Nat1 Acad Scl USA 88, 170 l-l 705 25 Kelly, W G , Roquemore, E P , and Hart, G W (1994) Cell cycle dependent changes m nuclear glycosylatlon, submitted. 26. Hayes, B K , Greis, K D., and Hart, G W. (1995) Specific isolation of O-GlcNAcbearing glycopeptldes. Anal Biochem 228,115-122 27. Mann, M and Wllm, M (1995) Electrospray mass spectrometry for protein characterization. Trends Blochem Scz 20, 2 19-224. 28 Smzdak, G (1994) The emergence of mass spectrometry m biochemical research. Proc Nat1 Acad Sci USA 91,11,290-l 1,297 29 Blemann, K (1992) Mass spectrometty of peptldes and protems Ann Rev Bzochem 61,977-1010

30 Taniguchl, H., Suzuki, M , Manenti, S , andTitani, K (1994) A mass spectrometrlc study on the m vlvo posttranslational modlficatton of GAP-43 J Biol Chem 269, 22,48 l-22,484.

3 1. Watts, J. D., Affolter, M , Krebs, D. L., Wange, R L., Samelson, L. E., and Aebersold, R. (1994) Identification by electrospray ionization mass spectrometry of the sites of tyrosme phosphorylatton Induced m activated Jurkat T cells on the protem tyrosme kinase ZAP-70. J Biol Chem. 269,29,520-29,529. 32. Carr, S. A., Huddleston, M. J., and Bean, M. F (1993) Selective tdenttticatton and differentiation ofN- and O-linked ohgosaccharides m glycoprotems by liquid chromatography-mass spectrometry Protein Scz 2, 183-I 96. 33. Huddleston, M J., Bean, M. F , and Carr, S A (1993) Collisional fragmentation of glycopepttdes by electrospray iomzatlon LUMS and LC/MS/MS: methods for selective detection of glycopeptides m protem digests. Anal Chem 65, 877-884.

32

Greis and Harf

34. Hart, G. W., Brew, K , Grant, G. A., Bradshaw, R. A , and Lennarz, W J (1979) Prtmary structural requtrements for the enzymatic formation of the N-glycosidic bond m glycoprotems J Blol Chem 254,9747-9753 35. Medzihradszky, K F, Gillece-Castro, B L , Settmeri, C A , Townsend, R R., Masiarz, F. R , and Burlmgame, A. L. (1990) Structure determmation of 0-linked glycopeptides by tandem mass spectrometry Blamed Environ Mass Spectrom 19, 777-781. 36. Rademaker, G J , Haverkamp, J , and Thomas-Oates, J (1993) Determination of glycosylation sites m O-linked glycopepttdes. a sensmve mass spectrometric protocol. Organic Mass Spectrom 28, 1536-1541. 37 Settineri, C A , Medzihradszky, K F , Masiarz, F R , Burhngame, A L , Chu, C , and George-Nasctmento, C. (1990) Characterization of 0-glycosylation sttes m recombmant B-chain of platelet-derived growth factor expressed m yeast usmg liquid secondary ion mass spectrometry, tandem mass spectrometry and Edman sequence analysis Blamed Envwon Mass Spectrom 19, 665-676 38. Zmn, A B , Plantner, J. J , and Carlson, D. M (1977) The Glycoconjugates Academic, New York, pp. 69-85 39. Downs, F and Pigman, W (1976) Determmatton of 0-glycosidic lmkages to L-serme and L-threonme residues of glycoprotems, m Methods HI Carbohydrate Chemzstry (Whistler, R L and BeMiller, J N., eds.), Academic. New York, pp 200-204 40 Bertohm, M. and Pigman, W (1967) Action of alkali on bovine and ovme submaxlllary mucms. J Bzol. Chem 242,3776-3781 41, Downs, F , Herp, A, Moschera, J., and Pigman, W (1973) P-Elimination and reductton reacttons and some applications of dimethylsulfoxide on submaxillary glycoprotems Blochem Bzophys Acta 328, 182-l 92 42. Greis, K D , Hayes, B K., Comer, F I, Kirk, M., Barnes, S , Lowary, T L , and Hart, G W. (1996) Selective detection and site-analysis of 0-GlcNAc-modified glycopeptides by p-ehmmation and tandem electrospray mass spectrometry. Anal Blochem 234,38-49 43. Whiteheart, S W, Passaniti,A , Reichner, J S., Holt, G D., Haltiwanger, R S , and Hart, G W (1989) Glycosyltransferase probes. Methods Enzymol 179,82-95 44. Roquemore, E P , Chou, T.-Y, and Hart, G W (1994) Detection of O-linked N-acetylglucosamme (0-GlcNAc) on cytoplasmic and nuclear proteins Methods Enzymol 230,443-460 45. Haltiwanger, R S and Hart, G. W (1993) Glycosyltransferases as tools m cell biologrcal studies, m Methods zn Molecular Bzologv, Vol 14 Glycoprotem AnalySIS in Btomedlclne (Hounsell, E F., ed ), Humana, Totowa, NJ, pp 175-187 46. Covey, T. R , Huang, E C , and Hemon, J D (1991) Structural characterization of protein tryptic peptides via hqutd chromatography/mass spectrometry and colhsion-induced dissociation of their doubly charged molecular ions Anal Chem 63, 1193-1200. 47. Keil, B. (1971) Hydrolysis: peptlde bonds, in The Enzymes Vol III (Boyer, P D , ed ), Academic, New York, pp 249-275

Study of 0-GlcNAc and Peptdes

33

48 Ken, B (1992) Essential substrate restdues for action of endopeptrdases, m Specsficlty of Proteolysis, Spnnger-Verlag, New York, pp. 66-72. 49 Udenfrtend, S., Stein, S , Bohlen, P, and Danman, W (1972) Fluorescamine a reagent for assay of ammo acids, peptides, proteins, and primary ammes m the picomole range. Sczence 178,87 1,872 50. Mayo, J W and Carlson, D M (1970) Effect of alkali and of sodmm borohydrtde at alkaline pH on N-acetylchondrosine* reduction versus cleavage Carbohydr Res 15,300-303

3 Analysis of Asparagine-Linked Oligosaccharides by Sequential Lectin-Affinity Chromatography Kazuo Yamamoto, Tsutomu

Tsuji, and Toshiaki

Osawa

1. Introduction Sugar moieties on the cell surface play one of the most important roles in cellular recognmon. In order to elucidate the molecular mechanism of these cellular phenomena, assessmentof the structure of sugar chains is mdispensable. However, it is difficult to elucidate the structures of cell surface ohgosaccharides because of two technical problems. First is the difficulty m fractionating various oligosaccharides heterogeneous m the number, type, and substitution patterns of outer sugar branches. The second problem IS that very limited amounts of material can be available, which makes it difficult to perform detailed structural studies. Lectins are proteins with sugar-binding activity. Each lectin binds specifically to a certain sugar sequence m ohgosaccharides and glycopeptides To overcome the aforementioned problems, lectins are very useful tools. Recently, many attempts have been made to fractionate oligosaccharides and glycopeptides on immobilized lectin columns. The use of a series of immobrlized lectin columns, whose sugar-binding specificities have been precisely elucidated, enables us to fractionate a very small amount of radioactive oligosaccharides or glycopeptides (-10 ng depending on the specific activity) into structurally distinct groups. In this chapter, we summarize the serial lectin-Sepharose affinity chromatographic technique for rapid, sensitive, and specific fractronatron and analysis of asparagine-linked oligosaccharides of glycoprotems. Structures of asparagine-linked oligosaccharides fall into three main categories termed high mannose-type, complex-type, and hybrid-type (1). They share the common core structure Mana 1-3(ManalA)Mar@ l-4GlcNAcP I4GlcNAc-Asn, but differ m their outer branches (Fig. 1). High mannose-type From

Methods tn Molecular Biology, Vol 76 Glycoanalysm Protocols Edlted by E F Hounsell 0 Humana Press Inc , Totowa, NJ

35

Yamamoto, Tsuji, and Osawa

36 high

mannose-type

Mann,.ZManu,‘i. j ;Manal\ 6 3ManR

Manal-2Ma”al’:’

I -4GlcNAcRI

-+GlcNAc-fin

:

al /

Manal-ZMannl\ZMan

complex-type NeuAcuZ-6GalRI-4GlcNAcRIt2Man

FUcall

al \

b 3ManRI-4GlcNAcRI-4GIcNAc-&n

6 j

al /

Ne”Aca2-6GalRI-4GI~NAcR1~2Man

hybrid-type GlcNAcRl

Manal’: I‘6 , 3Manal

‘6

Manal,:’ GdlRI4GlcNAcR1~2Man

al

4 3ManR

/

I -4GlcNAcRI

-4GlcNAc-&n

; 3

L----------------------------------J

Fig 1. Structuresof major types of asparagme-linkedohgosaccharidesThe boxed areaenclosesthe core structurecommon to all asparagme-linkedstructures ohgosaccharides have two to six additional a-mannose residues lmked to the core structure. Typical complex-type ohgosaccharides contams two to four outer branches with a sialyllactosamme sequence. Hybrid-type structures have the features of both high mannose-type and complex-type ollgosaccharides and most of them contain bisecting N-acetylglucosamme that is linked l31-4 to the J3-lmked mannose residue of the core structure. In addition, a type of carbohydrate chain, so-called poly-N-acetyllactosamme-type, has beendescribed (2-5). Its outer branches have a characteristic structure composed of N-acetyllactosamme repeating units. It may be classified to be of complex-type, however, it is antigemcally and functionally distinct from standard complex-type sugar chains (4). Some poly-N-acetyllactosamme type oltgosaccharides have branched sequences containing GalP 1-4GlcNAcP I-3(Gall314GlcNA@ 1-6)Gal units (2,3), which is the determinant of the I-antigen Other novel complex-type sugar chains having GalNAc~l-4GlcNAc groups m their outer chain moieties has been found recently (6,7), and GalNAc residues are sometimes sulfated at C-4 or sialylated at C-6. Glycopeptides or oligosaccharides can be prepared from glycoproteins by enzymatic digestions or chemical methods as discussed m subsequent chapters m this book. The most widely used means for preparing glycopeptides 1s to completely digest material with pronase Oligosaccharide can be prepared from

Sequent/al Lectin-Affimty Chromatography

37

glycoprotems or glycopeptides by treating samples with anhydrous hydrazine (8) or endoglycosidases. Since the released oligosaccharides retam their reducing termini, they can be radiolabeled by reduction with NaB3H4 (9). The prtmary ammo group of peptide backbone of glycopeptides are labeled by acetylation wtth [3H]- or [14C]-acetic anhydride (10) Before employmg columns of immobilized lectms for analyses, olrgosaccharides, or glycopeptldes should be separated on a column of QAE- or DEAE-cellulose based on amomc charge derived from siahc acid, phosphate, or sulfate restdues. Acidic ohgosaccharides thus separated should be converted to neutral ones for simplifying the following separation. To simplify discussion, the ohgosaccharides discussed here do not contain sialtc acid, phosphate, or sulfate residues, although these actdlc residues, especially sialic acid residues, are found m many ohgosaccharides. In most cases, the influence of these residues on the interaction of ohgosaccharides with nnmobilized lectins 1sweak, but where documentation of the influence of these residues is available, it IS mentioned m the appropriate sections. In this chapter we describe the general procedure of serial lectmaffinity

chromatography

of glycopepttdes

and oligosaccharrdes

using several

well-defined munobihzed lectms. 2. Materials 1. Mono Q HR5/5, DEAE-Sephacel, Sephadex G-25 (Pharmacla, Uppsala, Sweden) 2. High-performance ltqutd chromatography (HPLC), two pumps, wrth detector capable of momtoring ultraviolet absorbance at 220 nm. 3. Neurammtdase: 1 U/mL of neurammidase from Streptococcus sp (Seikagaku Kogyo, Tokyo, Japan) m 50 mA4 acetate buffer, pH 6.5. 4. Dowex SOW-X8 (50-100 mesh, H+ form) 5. Bio-Gel P4 minus 400 mesh (Blo-Rad, Hercules, CA). 6. HPLC mobde phase for Mono Q: A, 2 mA4Tris-HCl, pH 7.4, B, 2 mMTris-HCI, pH 7 4,0 5MNaCl 7. HPLC mobile phase for Blo-Gel P4: disttlled water 8. HPLC standard for Bto-Gel P4. partial hydrolysate of chttm prepared according to Rupley (II), 10 pg mtxed with 50 & distilled water Store frozen 9. Concanavaltn A, Rxznus communes (RCA) lectm, wheat germ agglutinm (WGA), Datura stramonlum (DSA) lectin, Maackla amurensls leukoagglutmm (MAL), Wistar~aflor~bunda (WFA) lectm, Allomyrzna dzchotoma (allo A) lectm, Amaranthus caudatus lectm (EY Laboratories, San Mateo, CA), Phaseolus vulgaris erythroagglutmm (E-PHA), Phaseolus vulgarzs leukoagglutinin (L-PHA) (Setkagaku Kogyo) Immobrlized lectins were prepared at a concentration of l-5 mg lectm/mL of gel (see Notes 1 and 2) or obtamed commercially (e g., Pharmacia, EY Laboratories, Blo-Rad, Seikagaku Kogyo) Galanthus nivalzs (GNA) lectm, Lens culznarzs (LCA) lectm, Pisum satwum (PSA) lectm, I&rafava (VFA) lectin, pokeweed mltogen (PWM), Sambucus nzgra L lectin (SNA)

38

Yamamoto, Tsuj~, and Osawa

10 3H-NaBH4 100 mC1 of 3H-NaBH4 (sp. 5-l 5 Cl/mmol; NEN, Boston, MA) mixed with 2 mL 10 mMNaOH, store at -80°C. 11 Tris buffered salme (TBS). 10 mA4 Tris-HCl, pH 7 4, 0 15MNaCl 12. Lectin column buffer 10 mMTns-HCl, pH 7 4,0.15MNaCl, 1 mMCaC& 1 mk! MnCl, (see Note 3). 13. N-Acetylgalactosamme (Sigma, St Louis, MO) 100 mM m lectm column buffer, store refrigerated. 14 Methyl-a-mannoside (Sigma)* 100 mM m TBS; store refrigerated. 15. Methyl-a-glucoslde (Sigma). 10 mA4 m TBS; store refrigerated. 16 Lactose (Sigma). 50 mM in TBS, store refrigerated 17. N-Acetylglucosamine (Sigma): 200 mA4 in TBS, store refrigerated

3. Methods 3.1. Separation of Acidic Sugar Chains on Mono Q HR5/5 or DEAE-Sephacel and Removal of Sialic Acids 3.1.1. Ion Exchange Chromatography 1, Equilibrate the Mono Q HR5/5 or DEAE-Sephacel column with 2 mMTns-HCl, pH 7.4 at a flow rate of 1 mL/mm at room temperature 2. Dissolve the ohgosaccharldes or the glycopeptldes m 0 1 mL of 2 mA4Tns-HCl, pH 7.4 and apply to the column. 3. Elute with 2 mMTris-HCl, pH 7 4 for 10 min, then with linear gradient (O-20%) of 2 mMTns-HCl, pH 7.4, OSMNaCl m 60 mm at a flow rate of 1 mL/mm. 4. Neutral oligosaccharlde are recovered in the pass through fraction Acldlc monoslalo-, dlslalo-, trrslalo-, and tetraslalo-oligosaccharides are eluted out successively by the linear gradient of NaCl.

3.1.2. Removal of Sialic Acid Residues 1 To IO-100 pg ohgosaccharldes free of buffers or salts add 100 pL of neurammidase buffer and 100 pL of neuramnudase, and incubate at 37°C for 18 h 2 Heat-inactivate the neurammldase by immersion In a boiling water bath for 3 min. 3. Apply to the column of Dowex 5OW-X8 (0 6 cm id x 2 5 cm), wash the column with 1 mL of distilled water, and concentrate the eluates under vacuum

Alternatlvely, add 500 $ of 0.M HCI and heat at 80°C for 30 min, and dry the sample using evaporator 3.2. Separation of PO/y-N-Acetyllactosamine-Type Sugar Chains from Other Types of Sugar Chains Poly-N-acetyllactosamme-type sugar chains vary asto the number of N-acetyllactosamme repeating units and the branching mode and the structural characterization of poly-N-acetyllactosamine-type sugar chains has been quite difficult (12). These types of sugar chams have higher molecular weights com-

39 Em-Gel I

t poly-N-acetyilactosamme-type I PWM I

I

P-4 1 WFA I complex-type GalNAc-GlcNAc

1

I other-Types

with

I

unbranched

branched

high

“$” mannosetype

I hybrid-type

hybrid-type complex-type I WGA 4 complex-type I Con A

I tn-,tetraantennary E-PHA I tnantennary +blsectmg

GN

tn-. tetraantennary I L-PHA

“i blantennary +Fuc

blantennary -Fuc

tn 2 6 dnched tetra

Fig 2 Scheme of fracttonation of asparagine-linked afftmty chromatography on tmmobthzed lectms

sugar chains by combining

pared to other high mannose-type, complex-type, or hybrid-type chains. PolyN-acetyllactosamme-type

sugar chains wrth a molecular

mass >4000 that are

excluded from the Blo-Gel P4 column chromatography (2,13) and thus are easily separated from others. 1 Equrhbrate two coupled columns (0.8 cm td x 50 cm) of Bto-Gel P4 m water at 55% by use of water Jacket 2. Elute the ohgosaccharrdes at a flow rate of 0.3 mL/mm and collect fractions of 0 5 mL Monitor absorbance at 220 nm. 3 Collect poly-N-acetyllactosamme-type oligosacchartdes that are eluted at the void volume of the column Other types of ohgosaccharldes mcluded m the column are subjected to the next separatton (Subheadings 3.3.-3.6.) illustrated m Fig. 2 The specificity of the lectms used is summartzed m Fig. 3 and Table 1

3.3. Separation of Complex-Type Sugar Chains Containing GalNAcp l-4GlcNAc Groups from Other Sugar Chains Novel complex-type ohgosaccharldes and glycopeptldes with GalNAcP l4GlcNAc$l-2

outer chains bmd to a WFA-Sepharose

column

(7).

Yamamoto,

40

Tsuji, and Osawa

r---------1

blantennary

complex-type NeuA~a2-3Ga101-4GICNAcOl-2~an~al _----___----

,

r\l~~~\ca~~~~i~~-4GlcNAcOl-2~an

trlantennaty

; GlcNAcRl : Ev_cp I ,----I I : 1 \: 4; 6 L----a-: 3ManRl-4Gl~NA~RI-4Gl~NA~ Asn

r--

L.-----A

al ‘:

complex-type Gal&l

4GIcNAcRI

Gall31 4GlcNAcRI

,:-------I I. I, ;Ma”ul.: ” I. -___--_A

GalRI-~GIcNAcRI-2Man

GalRI-4GlcNAcRI-2Man G~~RI-~G~cNAcR~ G~~R~-~G~cNAcRI

tetraantennary

\r------7 1-4 ;, 2Manal ‘~--w--s:

Manl? I -4GIcNAcR ;: al I--------‘1 1 GlcNAcRl ; ;I A*-:1 3ManR I -4GlcNAcRI

I -4GIcNAc

Asn

-4GlcNAc-Asn

;/ I

complex-type Gall31 -4GIcNAcRI ‘6 GalRl-4GlcNAcRI

/2

GalRl-4GlcNAcRI

\

GalRl-4GIcNAcR

I /2

Manal \ ManB1-4Gl~NAcRI-4GlcNAc-&n 4Ma”,l

unbranched

poly-N-acetyllactosamIne-type

Gal&l-4GlcNAcRI

branched

-I-3GalRI.4GlcNAcRl

)“- 2Man

al

poly-N-acetyllactosamnxype

Gall31

complex-type

wth

GalNAt

Man6

I -4GlcNAcRI

Man6

I -4GIcNAcR

-4GIcNAc-&n

GIcNAc

SO3 4GalNAcRI -_______--_____4GlcNAcRl-2Man N~uAcCC~-~G$!~$~C~~! >QCNAC_R!-2Man

al \ I -4GIcNAc

Am

CXI /:

Fig. 3. Structures of several complex-type ohgosaccharides. The boxed area mdlcates the characterlstlc structures recognized by several nnmobllized lectms

41

Sequential Lectm-Affimty Chromatography Table 1 Characteristic

Structures

Recognized

by Several

Immobilized

Lectinsa

RCA GNA WGA ConA LCA PSA VFA E-PHA L-PHA DSA M-

b

M-M 3” ‘M-GN.GN M-M‘3 3

G-GN-

WFA

2

-

M GN-GN

+

G-GN- 3

m

M-GN-GN

+

R

+

G-GN-M G-GN

f M-GN-GN

+

R

+

+

+

R

>

G-GN P

G-GN-M G-GN

>

G,N M-GN-GN

G-GN P

N.D

M-GN-GN

+

+

G-GN g:g

+

G-GN

a+, bound; -, not bound; G, galactose; Gn, N-acetylgalactosamme, GN, N-acetylglucosamme, F, fucose, M, mannose, N D , not determmed, R, retarded

42

Yamamoto, Tsuji, and Osawa

1 Equilibrate the WFA-Sepharose column (0 6 cm id x 5 0 cm) m lectm column buffer 2 Dissolve ohgosaccharldes or glycopeptides m 0.5 mL of lectm column buffer, and apply to the column 3 Elute (1 .O-mL fractions) successively with three-column volumes of lectm column buffer, and then with three-column volumes of 100 mA4 N-acetylgalactosamme at flow rate of 2 5 mL/h at room temperature 4 Collect complex-type sugar chains with GalNAcj31-4GlcNAc outer chams, which are eluted after the addition of N-acetylgalactosamme 5 Collect other types of sugar chains, which pass through the column

3.4. Separation of High Mannose-Type Sugar Chains from Complex-Type and Hybrid-Type Sugar Chains 3 4. I. Affinity Chromatography on Immobil/zed RCA After the separation of high molecular weight poly-N-acetyllactosammetype ollgosacchandes, a mixture of the other three types of sugar chains can be separated on a column of RCA lectin that recognizes GalPl-4GlcNAc sequence (1415). 1 Equilibrate the RCA-Sepharose column (0 6 cm Id x 5.0 cm) m TBS 2 Dissolve oligosaccharldes or glycopeptldes m 0.5 mL of TBS, and apply to the column 3 Elute (1 .O-mL fractions) successively with three-column volumes of TBS, then with three-column volumes of 50 rnA4 lactose at flow rate 2 5 mL/h at room temperature 4 Bmd both complex-type and hybrid-type sugar chains to the RCA-Sepharose column (see Note 4). 5 Collect high mannose-type ohgosacchandes, which pass through the column 6. Purify the ohgosaccharides or glycopeptides from salts and haptemc sugar by gel filtration on Sephadex G-25 column (1.2 cm id x 50 cm) eqmhbrated with dlstilled water

3.4.2. Affinity Chromatography

on immobilized Snowdrop Lectin

High mannose-type glycopeptides cifically retarded on the immobilized

that carry Manal-3Man units are spesnowdrop GNA lectm (16).

1 Equilibrate the GNA-Sepharose column (0 6 cm id x 5 0 cm) m TBS 2. Dissolve ohgosacchandes or glycopeptides m 0.5 mL TBS, and apply to the column 3. Elute (0.5-n& fractions) successively with five-column volumes ofTBS, to collect sugar chains lacking Manal-3Man units or hybrid-type, which are not retarded 4. Elute with three-column volumes of 100 mMmethyl-a-mannoside at flow rate of 2 5 mL/h at room temperature to obtain the specifically retarded high mannosetype glycopeptldes that carry Mancll-3Man units

Sequential Lectin-Affinity Chromatography

43

3.5. Separation of Hybrid-Type Sugar Chains from Complex-Type Sugar Chains 3.5.7. Affinity Chromatography on Immobilized WGA Wheat germ lectm (WGA)-Sepharose has a high aflintty for the hybrid-type sugar chains. It has been demonstrated that the sugar sequence GlcNA@l4Manp 1-4GlcNAcP 1-4GlcNAc-Asn structure IS essential for tight bmdmg of

glycopeptides to WGA-Sepharose column (I 7). 1 2 3. 4.

Equilibrate the WGA-Sepharose column (0 6 cm id x 5.0 cm) in TBS. Dissolve glycopeptides m 0 5 mL TBS and apply to the column Elute (0 5-mL fractions) successively with five-column volumes TBS Collect hybrid-type glycopeptides with a bisecting N-acetylglucosamine residue, which are retarded on the WGA column. 5 Collect sugar chains having the typical complex-type (and also high mannosetype) sugar chains eluted at the votd volume of the column with TBS

3.6. Separation of Complex-Type Biantennary Sugar Chains 3.6.7. Affinity Chromatography on Immobilized Con A Oligosaccharrdes and glycopeptides with tri- and tetraantennary complextype sugar chains pass through Con A-Sepharose, whereas biantennary complex-type, hybrid-type, and high mannose-type sugar chains bind to the Con A and can be drfferentrally eluted from the column (I&19). 1 Equilibrate the Con A-Sepharose column (0 6 cm id x 5.0 cm) in lectm column buffer 2. Pass the oligosaccharide mixture of complex-type chains obtained from the WGA column through the Con A-Sepharose column 3 Elute (1 -mL fractions) successively with three-column volumes of lectm column buffer 4 Collect ohgosacchartdes with tri- and tetraantennary complex-type sugar chains that pass through the column Complex-type biantennary glycopeptides or ohgosaccharides having btsectmg GlcNAc also pass through the column. 5 Elute (I-mL fractrons) successively with three-column volumes of 10 mk! methyl-cx-glucoside and finally with three-column volumes of 100 n-&f methyla-mannoside. 6 Collect complex-type biantennary sugar chams, which are eluted after the addition of methyl-a-glucoside 7. Collect high mannose-type and hybrid-type oligosacchartdes or glycopeptides eluted after the addition of 100 mA4 methyl-a-mannoside.

3.6.2. Affinity Chromatography on Immobilized LCA, PSA, or VFA The brantennary complex-type sugar chains bound to the Con A-Sepharose column and eluted with 10 mM methyl-a-glucoside, contam two-types of oli-

44

Yamamoto, Tsuji, and Osawa

gosaccharides, which will be separated on a column of lentil lectin (LCA) pea lectm (PSA) or fava lectm (VFA) (2&22). 1 Equthbrate the LCA, PSA, or VFA-Sepharose column (0 6 cm Id x 5 0 cm) m lectm column buffer. 2 Pass the btantennary complex-type sugar chains from the Con A column through the LCA, PSA, or VFA-Sepharose column. 3. Elute (1 .O-mL fracttons) successively with three-column volumes of lectin column buffer, then with three-column volumes of 100 mM methyl-a-mannostde at a flow rate 2 5 mL/h at room temperature 4 Collect btantennary complex-type sugar chams wrthout fucose that pass through the column 5 Elute bound btantennary complex-type sugar chams having a fucose residue attached to the mnermost N-acetylglucosamme to the column

3.6.3. Affinity Chromatography

on Immobilized E-PHA

Complex-type btantennary sugar chams having outer galactose residues and btsectmg N-acetylglucosamine are retarded by E-PHA-Sepharose (15,23). 1 Eqmhbrate the E-PHA-Sepharose column (0.6 cm td x 5.0 cm) m lectm column buffer 2 Apply the pass-through fraction from the Con A column on E-PHA-Sepharose column 3 Elute (0.5-mL fractions) successrvely wtth five-column volumes of lectm column buffer at flow rate 2.5 mL/h at room temperature 4 Collect btantennary complex-type sugar chains having a bisecting N-acetylglucosamme residue retarded on the E-PHA column (see Note 6) When elutron of the column IS performed at 4°C btantennary complex-type oltgosaccharides wtthout brsectmg N-acetylglucosamme are also retarded by the E-PHASepharose column

3.7. Separation of Complex-Type Trian tennary and Tetraantennary Sugar Chains 3.7.1. Affinity Chromatography on Immobilized E-PHA E-PHA-Sepharose interacts with high aftimty with trrantennary (having 2,4branched mannose) oligosaccharides or glycopepttdes containmg both outer galactose residues and a btsectmg N-acetylglucosamme residue (23). 1 Equilibrate the E-PHA-Sepharose column (0.6 cm td x 5.0 cm) m lectin column buffer 2 Apply the pass-through fraction from the Con A column on the E-PHA-Sepharose column 3 Elute (O-5-mL fractions) successively with five-column volumes of lectin column buffer at flow rate 2 5 mL/h at room temperature

45

Sequential Lectin-Affindy Chromatography

4 Collect retarded trlantennary (having 2,4-branched mannose) ohgosaccharldes or glycopeptldes contammg both outer galactose and bisecting N-acetylglucosamme on the E-PHA column Other tn- and tetraantennary ollgosaccharides pass through the column (see Note 7).

3.7 2. Affinity Chromatography

on Immobilized L-PHA

L-PHA, which is an lsolectm of E-PHA, interacts with trlantennary and tetraantennary complex-type glycopeptides having an a-lmked mannose residue substituted at positions C-2 and C-6 with GalP 1-4GlcNAc (24). 1 Equlllbrate the L-PHA-Sepharose column (0.6 cm id x 5 0 cm) m lectm column buffer 2. Apply the pass-through fraction from the Con A column onto the L-PHASepharose column 3. Elute (OS-mL fractions) successively with five-column volumes of lectm column buffer at flow rate of 2.5 mL/h at room temperature 4. Collect retarded trlantennary and tetraantennary complex-type glycopeptldes having both 2,6-branched a-mannose and outer galactose on the L-PHA column (see Note 8) Other tn- and tetraantennary ohgosacchandes pass through the column

3.7.3. Affinity Chromatography

on Immobilized DSA

DSA lectm shows high affimty with tn- and tetraantennary complex-type ollgosacchandes. Trlantennary complex-type ollgosaccharldes contammg 2,4-substituted a-mannose are retarded by a DSA-Sepharose column. Trlantennary and tetraantennary complex-type ollgosaccharldes having an a-mannose residue substituted at the C-2,6 positions bmd to the column and eluted by GlcNAc oligomer (25,2/j). 1. Equilibrate the DSA-Sepharose column (0 6 cm id x 5.0 cm) m TBS 2. Apply the pass-through fraction from the Con A column on DSA-Sepharose column. 3 Elute (O.S-mL fractions) successively with three-column volumes of TBS at flow rate 2.5 mL/h at room temperature to obtain retarded trlantennary complex-type sugar chains having 2,4-branched a-mannose on the DSA column 4 Elute with three-column volumes of 5 mg/mL N-acetylglucosamine ollgomer at flow rate 2 5 mL/h at room temperature to obtain bound trlantennary and tetraantennary complex-type oligosaccharides having an a-mannose residue substituted at the C-2,6 positions

3.8. Separation of PO/y-N-Acetyllactosamine-Type

Sugar Chains

High molecular weight poly-N-acetyllactosamme-type ohgosaccharides are classified mto two groups. One 1s branched poly-IV-acetyllactosammoglycan containing a Gal~l-4GlcNAc~1-3(Gal~l-4GlcNAc~l--6)Gal umt, and the

46

Yamamoto, TSUJI,and Osawa

other is linear poly-N-acetyllactosamme structure that lacks galactose residues substituted at the C-3,6 positions 3.8.1. Affinity Chromatography

on lmmobrlized PWM

Branched poly-N-acetyllactosamine-type ohgosaccharldes can be separated by the use of a PWM-Sepharose column (27). Since the sugar sequence Gal/3l4GlcNAcP 1-6Gal firmly binds to the PWM-Sepharose column, the branched poly-N-acetyllactosamme chains can be retained by the column, whereas unbranched ones are recovered wlthout any retardation (28) (see Note 9). 1. Equilibrate the PWM-Sepharose column (0 6 cm id x 5 0 cm) m TBS 2 Apply the poly-N-acetyilactosamme-type sugar chams separated on Blo-Gel P4 (see Subheading 3.2.) on the PWM-Sepharose column 3, Elute (1 0-mL fractions) successively with three-column volumes of TBS then with three-column volumes of 0 1MNaOH at flow rate 2 5 mL/h at room temperature. 4 Collect unbranched poly-N-acetyllactosamme-type sugar chams that pass through the column 5 Collect bound branched poly-N-acetyllactosamme-type sugar chams.

3.8.2. Affinity Chromatography

on Immobilized DSA

Immoblllzed DSA lectm mteracts wtth high affinity with sugar chains havmg the linear, unbranched poly-N-acetyllactosamme sequence. For the bmdmg to DSA-Sepharose, more than two intact N-acetyllactosamme repeating units may be essential (26). 1 Eqmhbrate the DSA-Sepharose column (0.6 cm Id x 5.0 cm) m TBS 2 Apply the poly-IV-acetyllactosamine-type sugar chains separated on Blo-Gel P4 (see Subheading 3.2.) on DSA-Sepharose column 3 Elute (1 0-mL fractions) successively with three-column volumes of TBS then with three-column volumes of 5 mg/mL GlcNAc ohgomer at flow rate 2.5 mL/h at room temperature 4 Collect branched poly-N-acetyllactosamme-type sugar chains, which pass through the column, separated from unbranched poly-N-acetyllactosamme-type sugar chains, which bmd

3.9. Separation of Sialylated Sugar Chains The basic GalP1-4GlcNAc sequence present in complex-type sugar chains may contam slahc acids in a2-6 or a2-3 linkage to outer galactose residues. 3.9.1. Affinity Chromatography

on Immobilized MAL

MAL (29,30) interacts with high affinity with complex-type try- and tetraantennary glycopeptldes contammg outer sialic acid residue-lmked a2-3 to

Sequential Lectm-A ffinity Chromatography

47

penultimate galactose. Glycopeptrdes containing sialtc acid linked only a2-6 to galactose do not interact detectably with the immobilized MAL (see Note 10). 1 Eqmhbrate the MAL-Sepharose column (0 6 cm id x 5 0 cm) m lectm column buffer. 2. Apply the acidic ollgosaccharrdes or glycopeptrdes separated on Mono Q HR5/5, or DEAE-Sephacel (see Subheading 3.1.1., step 1) on the MAL-Sepharose column 3 Elute (0.5~mL fractions) successively with rive-column volumes of lectm column buffer at flow rate 2 5 mL/h at room temperature. 4. Collect glycopeptrdes or ohgosaccharrdes containing a2-6-linked srallc acid(s), which pass through the column 5. Collect retarded glycopeptides or oligosaccharrdes contammg a2-3-lurked srahc acid(s).

3.9.2. Affinity Chromatography

on Immobilized Allo A

Allo A (31,321 recognizes the other isomer of sialyllactosamme compared to MAL. Mono-, di-, and trrantennary complex-type oligosaccharides contammg termmal sraltc acid(s) In ~2-6 linkage bound to allo A-Sepharose, whereas complex-type sugar chains having isomeric a2-3-linked sialrc acid(s) do not bmd to immobrlrzed allo A. 1 Equilibrate the allo A-Sepharose column (0 6 cm id x 5 0 cm) m TBS 2 Apply the acidic ohgosaccharrdes or glycopeptrdes separated on Mono Q HR5/ 5, or DEAE-Sephacel (see Subheading 3.1.1., step 1) on the allo A-Sepharose column 3 Elute (0 5-mL fractions) successively with three-column volumes of TBS and then wrth three-column volumes of 50 mA4 lactose at flow rate 2.5 mL/h at room temperature 4 Collect glycopeptrdes or oligosaccharides containing a2-3-linked srallc acrd(s), which pass through the column 5 Elute bound glycopeptrdes or ohgosaccharides having a2-6-linked siahc acrd(s) (see Note 11).

3.9.3. Affinity Chromatography

on immobilized SNA

Elderberry SNA bark lectm (33,341 shows the same sugar binding speciticity as allo A. All types of ollgosaccharides that contain at least one NeuAca26Gal unit m the molecule bound firmly to SNA-Sepharose. 1. Equilibrate the SNA-Sepharose column (0 6 cm Id x 5.0 cm) m TBS. 2. Apply the acrdrc oligosaccharrdes or glycopeptrdes separated on Mono Q HR5/5, or DEAE-Sephacel (see Subheading 3.1.1., step 1) on the SNA-Sepharose column 3. Elute (0.5-mL fractrons) successively with three-column volumes of TBS then with three-column volumes of 50 mM lactose at flow rate 2.5 mL/h at room temperature.

Yamamoto, Tsuji, and Osawa

48

4 Collect glycopepttdes or ohgosaccharides containing a2-3-linked srahc acid(s), which pass through the column 5 Elute bound glycopepttdes or ohgosaccharides having ct2-6-linked siahc acid(s) m the 50 mM lactose eluant

3.10. Summary Various itntnobilrzed lectins can be successfully used for fractionation and for structural studies of asparagme-lmked sugar chains of glycoprotems (see Note 12). This method needs ~10 ng of a radiolabeled ohgosacchartde or glycopeptrde prepared from a glycoprotem by hydrazmolysrs or by digestion wtth endo+&acetylglucosammidases or N-glycanases. The fractionatron and the structural assessment through the use of rmmobrlrzed lectms make the subsequent structural studies much easier.

4. Notes

2. 3

4

5. 6

8.

9 10

11 12

During the couplmg reactions, sugar-binding sites of lectms must be protected by the addition of the specific haptenic sugars. Immobrhzed lectm IS stored at 4°C In most cases, immobilized lectm is stable for several years Some lectms, especially legume lectins, need Ca2+ and Mn2+ tons for carbohydrate binding, so that the buffers used for the affinity chromatography on the lectm column must contam 1 n-& CaC12 and MnC12 Complex-type or hybrid-type ohgosaccharides are retarded on a column of RCASepharose rather than tightly bound when their sugar sequences are masked by sialtc acids Intact N-acetylglucosamme and asparagme residues at the reducing end are required for tight bmdmg of complex-type ohgosaccharides to both LCA-, PSA-, or VFA-Sepharose column High-afflmty interaction with E-PHA-Sepharose IS prevented if both outer galactose residues on a bisected sugar chain are substituted at posmon C-6 by siahc acid Biantennary and triantennary complex-type sugar chains having bisecting GlcNAc can be separated on a Bio-Gel P4 column L-PHA-Sepharose does not retard the elutton of sugar chains lacking outer galactose residues. WGA can be used instead of PWM Maackm amurensts hemagglutmm (MAH), which is an isolectm of MAL strongly binds to sialylated Ser/Thr-linked GalPI-3GalNAc, but not to sialylated Asnlinked sugar moleties (35) Oligosacchartdes without slahc acid(s) of mono-, dr-, tri-, and tetraantennary complex-type are retarded by the allo A lectm column More detailed reviews on the separation of ohgosaccharides and glycopeptides by means of afftmty chromatography on nnmobd~zed lectm columns have been published (36-38)

49 References 1. Kornfeld, R and Kornfeld, S (1985) Assembly of asparagme-lmked ohgosaccharides. Ann Rev Bzochem 54,63 l-664 2. TSUJI, T., Irtmura, T and Osawa, T. (198 1) The carbohydrate moiety of Band 3 glycoprotem of human erythrocyte membrane. J Blol Chem 256, 10,497-10,502 3 Fukuda, M., Dell, A , Oates, J E , and Fukuda, M N (1984) Structure of branched lactosammoglycan, the carbohydrate moiety of Band 3 isolated from adult human erythrocytes J B1o1 Chem 259, 8260-8273 4 Merkle, R K and Cummmgs, R D (1987) Relattonshtp of the terminal sequences to the length of poly-N-acetyllactosamine chains m asparagme-linked oltgosacchartdes from the mouse lymphoma cell lure BW5 147 J Bzol Chem 282,8179-8189 5. Fukuda, M. (1985) Cell surface glycoconjugates as once-dtfferentiatton markers m hematopotettc cells Bzochem Bzophys Acta 780, 119-l 50 6 Green, E. D. and Baenztger, J U. (1988) Asparagme-linked ohgosacchartdes on lutropm, follttropm, and thyrotropm. I Structural elucidatton of the sulfated and sialylated oligosacchartdes on bovine, ovine, and human pttuttary glycoprotem hormones J Blol Chem 263,25-35 7. Nakata, N., Furukawa, K , Greenwalt, D E., Sato, T., and Kobata, A (1993) Structural study of the sugar chains of CD36 purlfled from bovine mammary eptthehal cells: occurrence of novel hybrid-type sugar chains contammg the NeuSAca26GalNAcb 1-4GlcNAc and the Mana 1-2Mancx I-3ManaldMan groups Bzochemwry 32,436%4383 8 Fukuda, M , Kondo, T, and Osawa, T. (1976) Studies on the hydrazmolysts of glycoprotems Core structures of ohgosacchartdes obtamed from porcine thyroglobulm and pineapple stem bromelam J Bzochem 80, 1223-1232 9. Takasakt, S and Kobata, A (1978) Microdetermmation of sugar composmon by radtotsotope labelmg Methods Enzymol 50, 50-54. 10. Tai, T., Yamashtta, K , Ogata, M A., Kotde, N , Muramatsu, T., Iwashtta, S , Inoue, Y, and Kobata, A (1975) Structural studies of two ovalbumin glycopeptides m relation to the endo-b-N-acetylglucosamintdase specific@ J Biol Chem 250,856%8575 11. Rupley, J A. (1964) The hydrolysis of chttin by concentrated hydrochlortc actd, and the preparation of low-molecular-weight substrates for lysozyme Bzochem Blophys Acta 83,245-255 12 Krusms, T , Fmne, J , and Rauvala, H (1978) The poly(glycosy1) chains of glycoproteins Charactertzaton of a novel type of glycoprotem saccharides from human erythrocyte membrane Eur J Blochem 92,289-300 13 Yamamoto, K , TsuJi, T , Tarutam, 0, and Osawa, T (1984) Structural changes of carbohydrate chams of human thyroglobulin accompanymg malignant transformations of thyroid grands Eur J Blochem. 143, 133-144 14. Baenziger, J U and Ftete, D. (1979) Structural determmants of Rzcznus communzs agglutmm and toxin spectfictty for ohgosacchandes J B~ol Chem 254,9795-9799 15 Irtmura, T , TSUJI, T , Yamamoto, K , Tagamt, S., and Osawa, T (198 1) Structure of a complex-type sugar cham of human glycophorm A. Blochemzstry 20,5(X&566

Yamamoto, Tsuji, and Osawa

50

16 Shibuya, N., Goldstein, I J , Van Damme, E J M , and Peumans, W. J (1988) Bmdmg properties of a mannose-specific lectm from the snowdrop (Galanthus nzvalu) bulb J Btol Chem 263,728-734. 17 Yamamoto, K , TSUJI, T , Matsumoto, I , and Osawa, T (198 1) Structural reqmrements for the binding of ohgosacchartdes and glycopepttdes to tmmobthzed wheat germ agglutmm. Btochemrstry 20, 5894-5899 18 Ogata, S., Muramatsu, T , and Kobata, A. (1975) Fractionation of glycopeptides by affinity column chromatography on Concanavalm A-Sepharose J Btochem 78, 687-696.

19 Krusms, T , Fmne, J , and Rauvala, H. (1976) The structural basis of the different affinities of two types of acidic N-glycosidic glycopepttdes from Concanavalm ASepharose FEBS Lett. 71, 117-l 20 20 Kornfeld, K , Reitman, M L , and Kornfeld, R (198 1) The carbohydrate-bmdmg spectfictty of pea and lenttl lectms J Btol Chem 256,6633-6640. 2 1, Katagirt,Y ,Yamamoto, K., Tsujt, T., and Osawa, T. (1984) Structural requirements for the bmdmg of glycopeptides to mrmobilized fictafaba (fava) lectm. Carbohydr Res 129,257-265 22 Yamamoto, K , TSUJI, T , and Osawa, T. (1982) Requirement of the core structure of

a complex-type glycopepttde for the bmdmg to tmmobthzed lentil- and pea-lectms Carbohydr

Res 110,283-289

23. Yamashtta, K., Hitot, A , and Kobata, A (1983) Structural determinants of Phaseolus vulgarts erythroagglutmatmg lectm for ohgosaccharides J Btol Chem 258, 14,753-14,755 24 Cummmgs, R D and Kornfeld, S (1982) Characterization of the structural determinants required for the high affimty mteraction of asparagme-linked ohgosaccharides with tmmobilized Phaseolus vulgarts leukoagglutmating and erythroagglutmatmg lectms J Btol Chem 257, 11,230-l 1,234. 25 Cummings, R. D and Kornfeld, S. (1984) The distributton of repeating [Galpl4GlcNAcP l-3] sequences m asparagme-lmked ohgosacchartdes of the mouse lymphoma cell line BW5 147 and PHAR2 1 J Btol Chem 259,6253-6260. 26 Yamashtta, K , Totam, K. T., Ohkura, Takasaki, S , Goldstein, I J., and Kobata, A (1987) Carbohydrate bmdmg properties of complex-type ohgosacchartdes on immobilized Datura stramomum lectm. J Btol Chem 262, 1602-1607 27 Irtmura, T. and Nicolson, G L (1983) Interaction of pokeweed mitogen with poly(Nacetyllactosamme)-type carbohydrate chains. Carbohydr Res 120, 187-l 95 28. Kawashima, H., Sueyosht, S., Li, H., Yamamoto, K., and Osawa, T. (1990) Carbohydrate bmding specificities of several poly-N-acetyllactosamine-bmding lectms Gtycoconpgate

J 7,323-334.

29 Wang, W.-C. and Cummmgs, R D (1988) The mnnobihzed leukoagglutmm from the seeds of Maackza amurensts binds with high afftmty to complex-type Asnlinked ohgosacchartdes contammg terminal siahc acid-hnked a-2-3 to penultimate galactose residues J. Biol Chem. 263,4576-4585 30 Kawaguchi, T., Matsumoto, I , and Osawa, T (1974) Studies on hemagglutmms from Maackta amurensts seeds. J Btol Chem 249, 2786-2792.

Sequential Lectin-Affinity Chromatography

51

3 1. Sueyoshr, S., Yamamoto, K., and Osawa, T (1988) Carbohydrate binding specificity of a beetle (Allomyrzna dzchotoma) lectm. J Bzochem 103, 894-899 32. Yamashita, K., Umetsu, K , Suzuki, T., Iwakl, Y, Endo, T., and Kobata, A. (1988) Carbohydrate bmdmg specificity of immobilized Allomyrzna dzchotoma lectm II J Bzol. Chem. 263, 17,482-17,489. 33. Shibuya, N , Goldstein, I J , Broekaert, W F, Lubakl, M N , Peeters, B , and Peumans. W J. (1987) Fractronation of slalylated ohgosaccharides, glycopeptldes, and glycoprotems on mnnobihzed elderberry (Sambucus nzgra L ) bark lectm. Arch Biochem Bzophys. 254, l-8 34. Shibuya, N., Goldstein, I J., Broekaert, W. F, Lubaki, M. N , Peeters, B , and Peumans, W. J (1987) The elderberry (Sambucus nzgra L ) bark lectm recognizes the NeuSAca2-GGallGalNAc sequence J Biol Chem 262, 1596-160 1 35. Konami, Y, Yamamoto, K , Osawa, T., and Irimura, T. (1994) Strong affinity of Maackza amurenszs hemagglutmm (MAH) for slalic acid-contammg SerlThr-linked carbohydrate chains of N-terminal octapeptldes from human glycophorin A. FEBS Lett 342,334-338 36. Osawa, T and TsuJi, T (1987) Fractionation and structural assessment of ohgosacchandes and glycopeptldes by use of immobilized lectins Ann Rev Bzochem 56, 2 l-42 37 Osawa, T. (1989) Recent progress m the appltcatlon of plant lectms to glycoprotem chemistry. Pure Appl Chem 61,1283-1292. 38. TSUJI, T., Yamamoto, K., and Osawa, T (1993) Affimty chromatography of ohgosaccharides and glycopeptides with immobilized lectms, m Molecular Interactzons zn Bzoseparatzons (Ngo, T T., ed ), Plenum, New York, pp 113-l 26

4 Exoglycosidase Sequencing of /Winked Glycans by the Reagent Array Analysis Method (RAAM) Sally Prime and Tony Merry 1. introduction In comparison to the sequencmg of other biological macromolecules such as nucleic acids or proteins, the sequencing of oligosaccharides (glycans) of glycoprotems generally requires specialized expertise and facilities Whereas techniques such as nuclear magnetic resonance can give full sequence mformation (l-3), they are only available m relatively few laboratories and are by no means routme (discussed further in Chapter 11)

7.7. Enzymatic Sequencing Enzymatic methods are widely used m the determination of glycoprotem glycan sequence because of then ability to determine, m many cases unambiguously, full and accurate sequence mformation (4-7). Moreover, when coupled with suitable labeling and detection techniques (8), these methods are applicable to relatively small amounts (-10 pmol/digestion) of sample, where many other techniques cannot reliably be used. Exoglycosidases are enzymes that cleave nonreducmg terminal glycoside linkages to release monosaccharides. They are highly specific to anomeric configuration and often also to monosaccharide type and linkage (4). Some are also specific to other structural features, such as local and nonlocal branching (9) Thus, a particular exoglycosidase will cleave a termmal monosaccharide only if all its specificity requirements are met. If loss of monosaccharide does take place, then the identity, and m many cases linkage of that monosaccharide, are determined. From

Methods Edlted by

m Molecular E F Hounsell

Biology, Vol 0 Humana

53

76 Glycoanalysu Protocols Press Inc , Totowa, NJ

54 7.2. Detection

Prime and Merry Methods

The detection of cleavage of monosaccharides can be monitored m a number of ways. Usually one measures some property of the unknown glycan substrate before and after mcubation with a specific exoglycosidase, the value of which can be related to the number of monosaccharrdes lost. Because of the complex branching structure of glycoprotem glycans, there may be several nonreducmg terrmm accessible to the enzyme. Therefore, it is important that the property can be measured with sufficient accuracy to determme the precise number of monosaccharides cleaved. This is most easily done by measuring the size or mass of the glycan (611). An effective technique m current use is gel permeation chromatography of fluorophore-labeled glycans (10), which measures the hydrodynamic volume of a glycan relative to an internal standard of glucose ohgomers (12,12). This simple technique has the merit of being independent of the column or instrument used, and is highly reproducible. Also, because of the large number of studies on standards (11,12), it is now possible to predict the hydrodynamic volume of any glycan to within 2% of the observed value on a universal scale of glucose units (GU) m the range from 1 to 23 GU (6) (see Table 1). The GU contribution of a monosaccharide may depend on its position and branching arrangement (11) and correction must be made for derivatized glycans. For the alditol derivative (e.g., borotritide labeled glycans), add 0.5 GU. A technique using 2-aminobenzamide (2-AB) fluorophorelabeled glycans has been recently introduced (8) and for these the correction IS, GU (2-AB) = GU (unreduced) x 1.02 - 1.97. 1.3. Sequential

Sequencing

To begin with, the glycans must be released from the glycoprotem, either enzymatically (13,M) or chemically (15,16). The glycan pool is then fluorescently labeled to improve sensitivity of detection. The reducmg terminus is labeled because it is retamed after exoglycosidase digestion, and this allows the progressive digestion of the molecule to be monitored without loss of signal. The pool of labeled glycans can then be separated chromatographitally m one or more dimension(s) and fractions of pure glycans recovered. Now a single fraction with known hydrodynamic volume can be treated with exoglycosidase, and then chromatographed to determme any induced change in hydrodynamic volume (18,19). This process of digestion and recovery can be continued until the entire structure has been degraded and all the sequence information that can be obtained using available exoglycosidase has been deduced. It is clear from the aforementioned description that sequential sequencing is an iterative, ad hoc, relatively slow and labor-mtensrve technique. A single cycle requires an enzyme incubation that may require up to 18 h for

RAAM Sequencmg Table 1 Rules for Predicting of an Oligosaccharide

55 the Hydrodynamic in GW

GU contribution

Monosaccharide Galactose Mannose Fucose N-acetylglucosamme

N-acetylgalactosamme Glucose

Volume

11 0.9 0 5 (outer arm) or 1. 0 (core) 2 when not attached to a branching point 3 3 for both GlcNAc at a double branching point when one 1s attached l-6 and the other l-2; 3.8 for both GlcNAc at a double branching point when one is attached l-4 and the other 1-2; 4.2 for all three GlcNAc at a triple branching point attached l-21416, 0 5 when bisectmg at the middle mannose, and both complex arms are extended beyond the mannose; 1.O when blsectmg, but one of the complex arms terminates at mannose; and, 1 5 when bisecting, but one of the mannose branches IS entirely missing. 2 1

“The total hydrodynamic volume of an unreduced ohgosaccharlde ISequal to the sum of the mdlvrdual monosaccharide components computed as shown m this table

complete digestion to occur, followed by an hour or so of manual clean-up procedures, followed by a chromatography run of several hours’ duratton, during which fractions must be collected

and the fraction containing

the residue

reconcentrated ready for the next cycle. Another difftculty associated with sequential sequencing is that before commencing each cycle one has to decide which enzyme to apply. This decision IS by no means obvious wrthout some knowledge of the biosynthetic pathways for protein glycosylation, from which one can model classes of expected sequence and branching patterns. The work of sequencing goes very much hand-in-hand with the study of the biological activity of glycosyltransferases (20,21). For instance, N-linked glycans all share the common core sequence (Man)3(GlcNAc)2 (20; see Chapter 3). Thus structure is derived from the common

precursor structure (Man)3(Man)g(GlcNAc)2,

whose transfer to Asn-

X-Ser/Thr sequences of a nascent polypepttde IS the mmal step m IV-glycan biosynthesis. Once attached to the polypeptrde, the precursor structure IS first

56

Prime and Merry

trimmed by a-glucosidases and a-mannosrdase, giving htgh mannose-type sequences. Further processmg by glycosyltransferases leads to the hybrid and complex series of structures An exceedingly large number of sequence possibilities exist, depending on the order of action and avatlability of the glycosyltransferases and then donor substrates, and how far this processmg continues. However, because of restricted substrate specificity of glycosyltransferases, the number of possible sequences is limited to perhaps a few thousands. It 1s only these “biosynthetically possible” sequences, therefore, that need to be considered when interpreting the results of sequencing experiments The best approach to data interpretation is to employ a computer program that can apply the rules of biosynthesis to generate all possible sequences, and then simulate the result of defined exoglycosidase digestions on each structure. It can then report the complete set of structures that are consrstent wtth the observed results 1.4. Reagent Array Analysis Method (RAAM) The RAAM is an alternative sequencing strategy which greatly simpltfies an otherwise long, iterative process requiring a considerable level of expertise (17). The RAAM usesmixtures of exoglycosidases rather than single enzymes. The sample is divided into several equal aliquots, and each aliquot is incubated with a different mixture of enzymes, the reagent array (22). The actual enzyme mixes used in the array are predefined to be suitable for sequencing all types of glycan belonging to a particular class (e.g., neutral N-linked glycans produced by animal cells). Each mix digests the sample glycan from the nonreducing terminus until a glycan fragment is left that none of the enzymes can digest further. This fragment is called a “stop point fragment.” With a well-designed array, most of the many thousands of different sequence possibilities will produce a unique set of stop point fragments that can be reconstructed, generally with the aid of software, to give the sequence of the origmal glycan based on measured values of either the hydrodynamic volumes orthemolecular weights of the stop point fragments. By pooling individual digests and analysmg them simultaneously m the RAAM strategy an entire sequence may be determined in the same time it takes to perform a single cycle of sequential analysis (1723) Reduction of the number of mampulations and associated losses also allows smaller sample amounts to be used. 1.5. Designing an Enzyme Array Careful design and formulation of the enzyme array is critical to the success of the analysis. As with the sequential sequencmg approach, the exoglycosidases used must be highly pure and free of other glycosidic contammants to avoid ambiguous results. Table 2 lists the exoglycosidases used in the arrays

RAA M Sequencing Table 2 Exoglycosidases

Commonly

57

Used for Oligosaccharide

Class

Speciticrty

Fucosidases a-Fucostdase I (Almond) EC 3 2.1.111 a-Fucosidase (Bovine kidney) EC3.2.1.51 Galactosidases /3-Galactosidase (Bovme testes) EC 3.2.1 23 P-Galactosidase (S pneumonzae) EC 3 2.1.23 Hexosamimdases P-Hexosaminidase (Jack bean [Curnavulzu]) EC 3 2 1.30 P-Hexosammtdase (S pneumonzae) EC 3.2.1.30

Mannosidases a-Mannosidase EC 3.2.1 24 P-Mannosidase EC32125 a-Mannosidase [ CunuvulzuJ)

Sequencing

Fucal-3/4 (to GlcNAc) Fucal-2131416 (rate reduces with Increasing substrate complexity) GalPl-3/4>>/31-6

GalNAc/GlcNAcPl-2/3/4/6 GlcNAcP I-2Man>>GlcNAcp l-3 Gal>GlcNAc$1--6Gal (24) (steric hindrance if a GlcNAc IS pl-6 hnked on the Man l-6 if a brsectmg GlcNAc IS present; altered spectfictty noted at higher enzyme concentrations (25)

(Aspergdlus saztoz)

Manal-2Man

(Helzx pomutiu)

ManI3 1-4GlcNAc

(Jack bean EC 3 2 1.24

Manal-* Man Concentration and steric effects can result in nonremoval of apparently free Man residues (e.g., hybrid N-glycans) (25)

Sialtdases (Neurammldases) a-Siahdase (Arthrobucter ureufaczens) EC32 118 a-Sialidase (Newcastle disease virus) EC 3 2 1.18

(25)

NeuSAc2-6Gal>NeuSAc2-3Gal> NeuSAc2-8NeuSAc NeuSAc2-3Gal>NeuSAc2-SNeuSAc

*Linkage positron not specified

drscussed in this chapter, and a few others that are mentioned in the text, The set of enzymes used to constttute the mixes are chosen to be as specific as possible to the monosacchartdes and linkages present m the class of glycan to be sequenced. The mixes themselves are then designed to produce the maxrmum differentiation of stop-point fragment patterns over the range of struc-

58

Prime Table 3 A Diagonal

and Merry

Array

Glycosldase 0 1 2 Galpl-3/4 oxxxxxx Fucal-* oxxoxxx HexPl-* oxxxoxx Manal-2/3/6 0 X X Manpl-4 oxxxxxo *Linkageposltlonnot specified Table 4 Mannosidases

3

4

5

6

X

X

0

X

2 0 x x

3 x 0 x

4 x x 0

Deleted

Glycosldase Galj3l-3/4 Fuca l-* HexPl-* *Linkageposition

0 0 0 0

1 x x x

not specified

tures to be sequenced.The diagonal reagent array 1susually a good starting point from which an optimized array can be designed. A set of enzymes ISchosen, such that, when actmg together, total hydrolysis of all glycoslde lmkages in all neutral N-lmked structures IS guaranteed. This set is then formed into an array where there 1sone control mix contaming no enzymes, one contammg them all, and then as many mixes as necessary where one of the enzymes IS absent, but all others are present (Table 3). Such an array will produce stop-pomt fragments that define most features of the monosaccharide sequence of an unknown N-glycan, but will not define the linkages. Before going further, the basic array can be simplified by removing both the mannosldases (Table 4). This IS a design declslon that is justified by the overwhelmmg evidence that linked N-glycans m animals have a common core structure (20). We therefore decide to design an array that sequencesonly to that core structure, rather than to the final reducing residue. The al -3/4 specific fucosidase (Almond meal) is introduced to distinguish between core and outer fhcose (Table 5). Finally another enzyme IS added that is particularly useful in its specificity towards GlcNAc linkages, the P-N-acetylhexosamimdase of Streptococcus pneumonme (Table 6). Generally speaking, this array is suitable for sequencing neutral linked glycans of human or Chinese hamster ovary (CHO) origin (among others), but IS not designed for material from yeasts, insects, or plants. A modified RAAM enzyme array (Table 7) ISused for sialylated N-links, which used two slalidases of different specificity.

59

RAAM Sequenang Table 5 linkage-specific

Fucosidase

Glycosldase

0

1

la

2

3

4

4a

Galfl l-314 Fucal-* Fuca l-3/4 HexP I-*

oxxoxxx 0 x 0 0 oxxxxoo

0 x

0 0

0 0

x 0

0 x

Added

*Linkage posltlon not specified

Table 6 Linkage-

and Arm-Specific

NAcetylhexosaminidase

Glycosidase

0

1’

la

la’

2

2’

3

3’

4’

GalP l-3/4 Fuca l-* Fucal-314 HexP l-* HexP l-2

oxxxxooxxx oxxoooooox oooxxooooo 0

x

0

0

0

x

0

0

Glycosldase

0

1

2

3

4

a&alldase (A ureuficzens) a&alldase (Newcastle disease vu-us) P-Galactosidase (Bovine testes) /3-Hexosammldase (D~plococcus pneumonzae)

0 0 0

x 0 x

0 x x

x 0 x

0 0 x

0

x

x

0

x

0

1

Added

x

ooxoxxxoxo

*Lmkage posltlon not specified

Table 7 A RAAM Array for Sialyated

NGlycans

2. Materials

2.1. Kits and Enzymes All kits and enzymes were obtained from Oxford GlycoSciences (OGS; Abmgdon, UK). 1 2. 3. 4. 5

Glycan release kit for hydrazinolysis (K300) Peptide-N-glycosrdase F (EC 3.2.18) (E.5006). Fluorescent labeling with 2-ammobenzamlde-SignalTM labeling kit (K400) Deacidificatlon/desialylatlon of glycans-SlgnalTM deslalylatlon kit (K04). Neutral glycan sequencing-RAAM neutral N-glycan array (K468) (for enzymes, see Table 2)

6. Slalylated glycan sequencing-RAAM see Table 2).

slalylated N-glycan array (for enzymes,

60

Prime and Merry

2.2. Chemicals 1 2 3. 4

Triethylamme, acetomtnle, n-butanol, ethanol (Aldrich, Milwaukee, WI) Bovine serum fetum (Sigma, St LOUIS, MO) Trifluoroacettc acid (TFA, Pierce, Rockford, IL) Dextran cahbratton standard of partially hydrolyzed dextran (Oxford GlycoSctences, cat no 4503) 5 Reagents for 2-AB labeling were from Oxford GlycoSciences

2.3. Equipment 1 GlycoPrep 1000 automated hydrazmolysts (Oxford GlycoSciences) 2 A high-performance liquid chromatography (HPLC) system e g , gtlson 3 15 and 3 16 pumps fitted with a Gllson 121 fluorescence detector (excitation at 330 nm, emtsston at 420 nm) 3 GlycoSep C HPLC column (Oxford GlycoScrences) for high resolution fractionation of charged glycans. 4 RAAM 2000 GlycoSequencer (Oxford GlycoSctences) for fractionatton of neutral glycans and RAAM sequencing This mstrument comprtses a pulse-free solvent delivery system, a glycan sizing column maintained at 55°C refractive index and fluorescence detectors, and the RAAM GlycoSequencer software package (Eve ver 3 1 0 Oxford GlycoSciences, 1996) 5 Freeze drying was performed on a Vntis Freezemobile lyophthzer (see Note 2) 6 Samples of >O 25 mL were drted on a Savant vacuum concentrator connected to an Edwards vacuum pump (see Note 2)

3. Methods 3.7. Glycan Release and Labeling 1. For automated hydrazmolysis (27) m a GlycoPrep 1000, prepare the samples (10 pg to 2 mg of glycoprotein) by complete desalting (for example by dialysis for 48 h at 4°C against a 0 1% solution of aqueous trtfluoroacettc acid (TFA) and lyophtltzation for at least 24 h (at a vacuum of C 10 m&bar) 2 For manual hydrazmolysis (I5,26) use the Glycan Release Ktt as descrtbed by the manufacturers. 3 Release N-linked ohgosacchartdes with pepttde-N-glycosidase F under the conditions supplied by the supplier (see also refs. 28 and 29) 4. Dissolve 25 pmol to 50 nmol of released, desalted glycan pool (see Note 1) in 5 pL of a solution of 70% dimethyl sulfoxtde, 30% glacial acetic acid contammg 0 25M 2-AB, and 0 1M sodium cyanoborohydride 5. Incubate for 2 h at 65°C. 6. Separate labeled glycans from unreacted dye by absorption onto a hydrophtlic filter in the present of acetonitrile and subsequent elutton with water (see Note 2).

61

RAAM Sequencing Table 8 Significance of RAAM Match for Best Matching Structure

Quality

Match quality range

Significance

>90 80-90 75-80 70-75 <70

Very good Good Marginal (high) Marginal (low) Poor

3.2. Deacidification/Desilylation of Glycans Prior to Gel Permeation Chromatography in Water 1 Dissolve the glycan pool m 40 pL of water and transfer to a 0.5 mL Eppendorf tube 2 Incubate with 2.0 U/mL of siahdase (Arthrobacter ureufaczens)m 100 mMammomum acetate pH 4.5 at 37°C for 14-I 8 h or 4 h prior to the sialylated N-glycan array. 3 Desalt on a tandem column of Dowex AGl (acetate form) over AG50 (triethylamme form) usmg 150-200 pL bed of each resin 4. Filter through a 0.5 ~JWpolytetrafluoroethylene (PTFE) filter (see Note 3) 5 Transfer the sample onto the top of the column bed and elute with three lots of 500 @ water 6. Evaporate the sample to dryness (see Note 2) 7 Resuspend the sample in 80 pL water evaporate to dryness 8 Run on the RAAM 2000 GlycoSequencer in high resolution mode. 9. Collect fractions of suitable size (generally 100-200 pL) correspondmg to the peak of interest, pool, and dry.

3.3. Incubation with Optimized RAAM Enzyme Arrays (see Tables 6 and 7 and Note 4) 1. Divide the pure unknown (and possibly slalylated) glycan into five equal allquots. Treat wtth the enzyme mixes. Depending on the posltlon and linkage of the slalic acids, various stop-pomt fragments are generated, some bearing sialtc acid restdues (see Note 5). 2. For the neutral glycan array wtth nine mixtures, dissolve the sample m 100 pL water tranfer 10 & accurately to small volume (500 pL) tubes 3 For the sialylated glycan sequencing array (five mixtures), dissolve the sample m 60 pL water accurately transfer IO pL to five small volume tubes 4. Dry the samples with a centrlfLlga1 evaporator (see Note 2) 5 Add to each tube 10 pL of the exoglycosldase mix from the array 6. Incubate the tubes at 37’C for 16-20 h 7 At the end of the mcubatlon, add 10 pL pyridme to each tube to stop the reactlon (see Note 7)

62

Prime and Merry

8 Pool the contents of all the tubes over a mixed bed column prepared by the sequential addition of 0 5 mL 50% aqueous suspensions of the followmg resins a Protein-bmdmg resin (Mimetic Blue AX6LSA) b Weak-anion exchange resin (Dowex AG 3&-OH- form). c Catton exchange resin (Dowex AGSO-OH+ form) d Strong amon exchange resm (Dowex AG l-OHform) 9. Elute the glycans with 3 x 1 mL water 10 Filter through 0 5 m PTFE filter (see Note 3) 11 Evaporate to dryness 12. Resuspend the sample m 200 pL of water and take to dryness.

3.3 7. Sialylated Glycan Array 1 Deactivate the enzymes and remove (see Note 5) 2 Pool the residual stop-pomt fragments from each mtx 3 Perform a second digest with the broad spectficity sialidase to remove any remaining siahc residues, leavmg behind neutral stop-point fragments that now express “memory-origmal” sialylated structures. 4. Pool the resultant desialylated (1.e , neutral) stop-point fragment and analyze usmg gel permeatton chromatography and the data analysis software used to assign the structure

3.4. Example of a RAAM Analysis of Neutral and Charged Complex Type Glycans 1. Prepare solutions of NA2F (a btantennary core fucosylated complex type glycan) (Oxford GlycoSciences, cat no C-024301) and Al a monosialylated derivative (Oxford GlycoSciences, cat no C- 124300) m water and take ahquots contammg 500 pmol of each. 2 Evaporate to complete dryness m 0.5-mL Eppendorf tubes 3. Label with 2-AB (8) as described in Subheading 3.2. 4. Take a lOO-pmol aliquot of each purified glycan fraction and analyze by RAAM. Two different enzyme arrays should be used, suitable for sequencing either neutral or stalylated N-glycans, based on the designs descrtbed m the text (Tables 6 and 7) 5 Set up the enzyme digests at 37°C for up to 18 h. 6 Analyze the exoglycostdase digestion pools (the glycan stop-pomt fragments) by gel permeatton chromatography on the RAAM 2000 GlycoSequencer with water elution at a flow rate of 70 pL/mm over 27 mL, held constant for a further 7 mL (Fig. 1) Process the chromatographic data by the Eve software which will perform a search on the database to find possible matching structures (Fig. 2). In the case of sialylated glycans use the sialylated glycan array and completely desalt the resulting stop-point fragments before the analysis (Fig. 3) Analyze the chromatographic data with the Eve software (Fig. 4)

Fig. 1. The RAAM type glucan.

2000 profile of a biantennary

core fucosylated complex-

Experimental signature

92

Predicted Database Reference: Hydrodynamic Volume: Molecular Weight:

Al- -a-

-2.1,

Structure

Key

m3n2-9374 12.2 GU 1788 Da

-3 0 A

-31\

L-Fucose D-Mannose D-Galactose N-acetyl-Dglucosamine alpha linkage beta linkage

Fig. 2. Comparison of the experimental and theoretical signatures showing elution positions and intensities and the predicted structure from the profile in Fig. 1.

63

pmoUml 6 5 4 3 2 Vol(ml)

0.00

Experimental

Theoretical 7

13

I 12

11

IO

i 9

/ / A 8

I 6

7

signature

signature: Sialylated N-glycan array 78 w 5

I 4

3

2

GU

Fig. 3. The RAAM 2000 profile and experimental and theoretical signatures of a sialylated biantennary complex-type glucan. Predicted

Structure

Database Reference: Hydrodynamic Volume: Molecular Weight:

m3n2-13+3 11.2 GU 1933 Da

&l-. 3

02-6

4q-

_ 4i--

I

0

N-acetylneuraminic

acid

Fig. 4. The predicted structure based on the experiment shown in Fig. 3. 64

RAAM Sequencing

65

3.5. Troubleshooting 1 The nine-mixture array m Table 6 represents a compromise between complexity and accuracy In particular, by choosmg the 3/4 specific galactosidase (bovine testes), it is not posstble to state whether a galactose is linked to the 3 or 4 position on GlcNAc. Although a l-4 specific P-galactosidase is available (S pneumonzae), this particular enzyme was chosen because both types of linkage can arise, and the array would need to be a lot more complex to accurately identify that particular linkage A second limitation IS that the original assumption that the enzymes chosen would guarantee hydrolysis of all N-glycans down to the common core structure may m certain cases be invalid. Unusual modifications exist, such as a-galactose, which is found as a terminating residue m some complex N-glycan and xylose, which IS typically found in the N-glycans from plant glycoproteins 2 Glycoprotem glycans often occur m complex mixtures of isomeric and nomsomeric forms It is necessary to fractionate the glycan pool m a charge separation mode of chromatography such as GlycoSep C (Oxford GlycoSciences) and subsequently to characterize and, if necessary, refractionate the isolated peaks on another chromatographic dimension, e g., the RAAM signature of an impure sugar is recognizable because it usually contains too many peaks in noninteger proportrons. The matchmg software may report many different structures of lowmatch quality. In this case it is still sometimes possible to identify the likely predominant structure If the major peaks show good agreement of GU values with the theoretical peaks m one of the structures in the match table, then that structure is more likely to be correct than others with poorer agreement With msufficient sample, the signal-to-noise ration m the RAAM signature will be too low to define the RAAM peaks clearly. Even if they are identified, they may not be m the correct relative amounts. In performmg this comparison, the composition of the RAAM array is taken into account to calculate the digestion product of each mix The software must also predict the hydrodynamic volumes of the various theoretical digestion products to compare theoretical peak posittons with the experimental ones (See comparison of theoretical and experimental signatures in Fig. 3.) 3 These methods perform automatic calibration of the chromatographic peaks with respect to the internal standard dextran hydrolysate, and measure the relative peak areas (Fig. 2). These results are automatically interpreted by comparing the experimental chromatographm signature with simulated signatures from all the possible types of N-glycan structure that can be generated from the biosynthetic rules incorporated into the software. Sometimes more than one possible structure will give the same signature of stop-point fragments, which cannot be distmguished. We refer to this feature as degeneracy It is one of the strong pomts of the computer data analysis, that where degeneracies do occur, all possibilmes are listed It should be emphasized that there are limitations in the specificity of the enzymes chosen for sequencing, but these apply to any enzymatic sequencing technique rather than to RAAM itself. The degree of similarity between the theo-

66

Prime

and Merry

retrcal and experimental signatures is measured using the Kolmogorov-Smtmov statrsttc (30) that takes mto account both peak posmon and peak mtensrty, and a figure called the “Match Quality” IS derived A value of 100 for match quality indicates perfect agreement between the experimental and theoretical signatures. The degree of confidence to be attributed to a result of gtven match quality has been determined empirically from a large body of experimental results and 1s summanzed m Table 8. 4 Although this chapter illustrates the use of gel permeation chromatography m water GlycoSequencer), alternative analytical procedures can be used to analyze the stop-point fragments from RAAM dtgests, provrded they fulfill the followmg requirements* a The measurement should be of a property of the stop-pomt fragments from the RAAM dtgests that is likely to give a different result for different fragments b For the purpose of data analysis, this property must also be accurately predtctable for any given glycan structure c If the stop fragments are to be pooled before analysts (as we have done here) the measurement should also be quantitative, i.e., it must gave the molar ratios of the different stop-point fragments. However, most of these techniques lack the predrctabrhty and precision that IS so vital to data Interpretation m any but the most simple of sequences. An exception to thts 1s the use of a column where separation is related to the size of the molecule as recently described by Guile et al. (31) Moving to higher level of complextty, improved accuracy of size measurement can be acheived by molecular weight determmatron by mass spectrometry (32), although complete sequence mformatron can be determined from enzymatic digestions on HPLC (33) 5 An enzyme array is designed to sequence glycans of a particular class One of the fundamental assumptions is that the sample will be digested fully down to a known core structure by one of the mixes m the array If a sample gives an unmtelhgible signature, and m particular seems to be at least partially resistant to all of the enzyme mixes, then the presence of a blocking group should be suspected. Examples of uncharged blocking groups are a-o-glucose, a-o-glucose, and CC-DN-acetylgalactosamme Likely charged blocking groups would be sulfate, phosphate, modified sialic acids, and uronrc acids.

4. Notes 1 For enzymatic release (13) the enzyme should be m a glycerol free formulation and the reaction carried out in the absence of detergent (28). The mcubatron mixture must be completely desalted before labelmg (29) 2 It is most important to use an efficient vacuum system (15-20 millibar) to dry the sample traces of volatile maternal which may be present A second evaporation step may help m the complete removal of poorly volatile material It is important to follow drying procedures carefully to remove volatile salts remaining m the sample after RAAM gel permeation chromatographic purificatton of the sample

RAA M Sequencing

67

During RAAM, ensure that the correct amount of each resin is used in the cleanup columns and that the column beds are thoroughly washed before use 3 Because of the difficulty of wetting the PTFE membrane with water It 1sstrongly recommended that Luer lock syrmges (Becton Dickinson) are used for all operation requiring filtration 4 All enzymes used for sequencing are tested for contammating glycosldase actlvlty by assays with appropriate mixtures of glycans. 5 All the enzymes must be completely deactivated before the different mixes are combined. Ensure that the stop solution 1sthoroughly mixed with the sample, and wash each ahquot separately onto the resin bed usmg the recommended amount of water before adding the next one If the enzyme reactlons are not stopped, then the stop-point fragments ~111 be partially degraded, leading to small extra peaks In the RAAM signature at lower hydrodynamic volume

References 1 Brockbank, R L and Vogel, H. J (1990) Structure of the ohgosaccharide of hen phosvltm as determined by two-dimensional NMR of the intact glycoprotem. Bzochemzstry 29,5574-5583

2 Homans, S W (1992) Homonuclear three-dlmenslonal NMR methods for the complete assignment of proton NMR spectra of ohgosacchandes-apphcatlon to Gal beta l-4 (Fuc alpha l-3) GlcNAc beta 1-3Gal beta 1qGlc Glycobzology 2, 153-159 3. Van Halbeek, H (1994) IH Nuclear magnetic resonance spectroscopy of carbohydrate chams of glycoprotems Methods Enzymol 230, 132-168 4 Jacob, G S and Scudder, P (1994) Glycosldases in structural analysis. Methods Enzymol 230,28&299

5. Maley, F , Tremble, R. B., Tarentmo, A. L., and Plummer, T H., Jr. (1989) Characterization of glycoprotems and then associated ohgosaccharldes through the use of endoglycosidases Anal Blochem 180, 195-204 6 Kobata, A , Yamashita, K , and Takahashl, S (1987) Blogel P-4 column chromatography of obgosaccharldes Effective size of ohgosaccharldes expressed m glucose units Methods Enzymol 138,84-93. 7 Parekh, R B , Tse, A G , Dwek, R. A., Willlams, A F, and Rademacher, T W (1987) Tissue-specific N-glycosylatlon, site-specific ohgosacchande patterns and lent11 Iectin recogmtlon of rat Thy- 1. EMBO J 6, 1233-l 244 8 Blgge, J C., Patel, T P, Bruce, J. A., Gouldmg, P. N., Charles, S. M , and Parekh, R B (1995) Non-selective and efficient fluorescent labellmg of glycans usmg 2-ammobenzamlde and anthramlic acid Blochemzstry 230,229-238 9 Yamashlta, K , Ohkura, T , Yoshlma, H., and Kobata, A (198 1) Substrate speclficlty of dlplococcal beta-hr-acetylhexosamlmldase A useful enzyme for the structural analysis of complex type asparagme linked sugar chains Bzochem Bzophys Res Commun 100,226-234

10. Dwek, R A., Edge, C J., Harvey, D. J , Wormald, M R , and Parekh, R B. (1993) Glycoform analysis of glycoprotems Ann Rev Blochem 52,65-l 0 1

68

Prime and Merry

11 Kobata, A (1994) Size fracttonatton of ohgosacchartdes Methods Enzymol 230, 200-208 12 Yamashita, K , Mizuochi, T., and Kobata, A. (1982) Analysis of ohgosacchandes by gel filtration. Methods Enzymol. 83, 105-l 26 13 Plummer, T. H. and Tarantmo, A. (1991) Purtficatton of the ohgosaccharide-cleaving enzymes of Flavobactertum mentngosepttcum Glycobiology 1,257-264 14 Tarentino, A L and Plummer, T H , Jr. (1994) PNGase enzymatic deglycosylatton of asparagme-linked glycans. purttication, properttes and spectficny of ohgosaccharade-cleaving enzymes from Flavobactertum mentngosepttcum Methods Enzymol 230,4457 15. Takasakt, S , Mtzoucht, T., and Kobata, A. (1982) Hydrazmolysts of aspargmelmked sugar chams to produce free oltgosacchartdes. Methods Enzymol 83, 263-268 16 Patel, T, Bruce, J , Merry, A, Btgge, C , Wormald, M , Jacques, A , and Parekh, R (1993) Use of hydrazme to release intact and unreduced forms of both N- and O-linked oligosacchartdes from glycoproteins. Btochemtstry 32,679-693 17 Edge, C J , Rademacher, T W., Wormald, M. R., Parekh, R B , Butters, T D , Wmg, D R., and Dwek, R. A (1992) Fast sequencing of oligosacchartdes. the reagent-array analysis method Proc Natl. Acad Scz USA 89, 6338-6342 18. Mtzuocht, T , Yonemasu, K., Yamashita, K., and Kobata, A. (1978) The asparagme linked sugar chains of subcomponent C 1 q of the first component of human complement J Btol Chem 253,7404-7409. 19 Parekh, R. B Dwek, R. A, Thomas, J. R , Opdenaker, G,, and Rademacher, T (1989) Cell-type- specific and site-specific N-glycosylatton of type I and type II human tissue plasmmogen activator Biochemtstry 28, 76447662. 20 Kornfeld, R and Korufeld, S (1985) Assembly of asparagine-lmked ohgosacchartdes Ann Rev Bzochem 54,63 l-664 21. Schachter, H and Brockhausen, I. (1992) The biosynthesis of serme (threonme)N-acetyl lactosamme-linked carbohydrate moieties, m GZycoconjugates-Composztzon,Structure and Functzon (Allen, H J. and Klsatlhs, E. C , eds.), Dekker, New York, pp. 263-290 22 Parekh, R B , Deannley, J , Ventom, A., Edge, C , and Prime, S (1996) Ohgosaccharade sequencmg based on exo- and endo-glycostdase digestton and liquid chromatography products J Chromatogr 720,263-274 23 Treumann, A, Lifely, M R., Schneider, P, Ferguson, M A J (1995) Primary structure of CD52 J Btol Chem 270,60884099. 24 Yamashtta, K., Kamerlmg, J. P , and Kobata, A (1983) Structural studies of the sugar chams of hen ovomucoid Evidence mdtcating that they are formed mainly by the alternate btosynthettc pathway of asparagme-linked sugar chains. J Bzol Chem. 258,3099-3 106 25 Ichtshima, E., Arai, M , Shigematsu,Y, Kumagai, H , and Sumida-Tanaka, R (198 1) Purification of an acidic a-D-mannosidase from Aspergrllus sartot and specttic cleavage of 1-2a-o-mannosidic linkage m yeast mannan Biochem Bzophys Acta 658,45-53

RAA M Sequencing

69

26. Lt, Y. T (1967) Studies on the glycosidases m Jack Bean Meal J, Bzol. Chem 242, 5474-5480 27. Merry, A H , Bruce, J., Btgge, C , and Ioannides, A (1992) Automated simulta-

28.

29

30

31

32

33

neous release of intact and unreduced N- and O-linked glycans from glycoprotems. Brochem. Sot Trans 20,91. Nons, G. E , Flaus, A. J., Moore, C. H , and Baker, E N (1994) Glycobiology: purflcatton and crystallization of the endoglycostdase PNGase F, a peptide N-glycostdase from Flavobacterlum menmgoseptlcum J Mel Biol 241,624-626. Davies, M J., Smith, K. D , and Hounsell, E F (1994) The release of ohgosacchandes from glycoproteins Methods Mol. Blol 32, 129-14 1 Press, W. H., Flannery, B P., Tenkolsky, S A., and Vetterlmg, W. T. (1988) Statistical descrtptton of data, m Numerical Reczpes, Cambridge Umverstty Press, Cambndge, UK, pp. 49@-494. Guile, G G , Rudd, P M., Wing, D. R., Prime, S B , and Dwek, R. A (1996) A rapid high resolution method for separating oligosaccharide mixtures and analysing sugar prints. Anal Blochem 240,2 10-226. Harvey, D J., Rudd, PM., Bateman, R H , Bordoh, R S , Howes, K , Hoyes, J. B , and Vtckers, R G. (1994) Examination of complex oligosaccharides by matrixasststed laser desorption/iomzation mass spectrometry on time-of-flight and magnettc sector instruments. Orgamc Mass Spectrom 29,753-765 Rudd, P M., Morgan, B. P., Wormald, M. R , Harvey, D J , van den Berg, C W , Davis, S J , Ferguson, M. A J , and Dwek, R. A. (1997) The glycosylation of the complement regulatory protein, human erythrocyte CD59 J Blol. Chem 272, 7229-7244

5 HPAEC-PAD Analysis of Monosaccharides Released by Exoglycosidase Digestion Using the CarboPac MA1 Column Michael Weitzhandler, Jeffrey Rohrer, James R. Thayer, and Nebojsa Avdalovic 1. Introduction As discussed in Chapter 4, exoglycosidases are useful reagents for the structural determination of glycoconjugates. Their anomeric, residue, and linkage specificity for terminal monosaccharides have been used to assess monosaccharide sequence and structure in a variety of glycoconjugates (I). Their usefulness depends on the absence of contaminating exoglycosidases and an understanding of their specificity. Digestions of oligosaccharides with exoglycosidases give two classes of products: monosaccharides and the shortened oligosaccharides. Most assays of such reactions have monitored the reaction by following oligosaccharides that are labeled at their reducing ends. In these assays, after exoglycosidase digestion the shortened oligosaccharide retains the label at its reducing end. The other digestion product, the released monosaccharide, does not carry a label and thus cannot be quantified. Additionally, identification of any other monosaccharide that could be the result of a contaminating exoglycosidase activity would not be possible. Quantitative measurement of all products (all released monosaccharide[s] as well as the shortened oligosaccharide product) would be useful because it would enable the determination of any contaminating exoglycosidase activities by determining the extent of release of other monosaccharides. High pH anion exchange chromatography with pulsed amperometric detection (HPAEC-PAD) detects the appearance of monosaccharide product(s), the shortened oligosaccharide product(s) as well as the disappearance of the oligosaccharide substrate(s) in a single chromatographic analysis without labeling. Thus, From:

Methods in Molecular Edited by: E. F. Hounsell

Biology, Vol. 76: Glycoanalysis Protocols 0 Humana Press Inc., Totowa, NJ

71

72

Weitzhandler et al.

has been used extensively to monitor the activtttes of several different exoglycosidases on glycocoqugates, usually using the CarboPac PA1 column to separate the digestion products (see refs. 1-S). A problem encountered when using HPAEC-PAD to motutor exoglycostdase digesttons is that N-acetylglucosamme (GlcNAc) is not baseline resolved from other glycocoqugate monosaccharides on the CarboPac PA1 column usmg the isocratic condittons that successively baselme resolve glycoprotem hydrolyses products: fucose, galactosamme, glucosamme, galactose, glucose, and mannose (9). We recently discovered that the CarboPac MA1 column, developed for the separation of neutral sugar alditols (10,12), gives an isocratic baseline separation of GlcNAc, GalNAc, fucose, mannose, glucose, and galactose, and stmultaneously resolves many neutral oltgosacchartdes. This separation extends the usefulness of the CarboPac MA1 column to the assay of reducing monosaccharides released by exoglycosidases In the followmg, an assay of exposed GlcNAc after p-N-acetylhexosammldase treatment of a variety of glycocoqugates is shown. To determine the suitabthty of HPAEC-PAD and the MA1 column for analyzing both released monosaccharide and ohgosacchartde products m a single analysis, we subjected an asialo agalacto biantennary oligosaccharide standard (Table 1, structure 3; Fig. lA, peak 3) and an astalo agalacto tetraantennary ollgosaccharide standard (Table 1, structure 5; Fig. lB, peak 5) to Jack bean l3-N-acetylhexosammidase dtgestton. In addttton to differences m numbers of antennae (2 vs 4) and retention times (22.3 vs 25.2 min), these two oligosaccharides differ in that the tetraantennary oligosaccharide has termmal GlcNAc linked to mannose l3(1+4) and p(l+6) m addition to the l3(1+2) linkages present m the biantennary oltgosaccharide standard. The complete dtsappearance of the asialo agalacto btantennary substrate 1s indicated by the disappeance of peak 3 (Fig. 1A; compare dashed vs solid line m bottom tracing). The complete disappearance of the asialo agalacto tetraantennary substrate is indicated by the disappearance of peak 5 (Fig. 1B; compare dashed vs solid line in bottom tracing) The expected digestion products of both structures 3 and 5 would be the released monosaccharide GlcNAc, (see Fig. lA,B; peak at 15 6 mm) and the released, shortened oligosaccharide product, Man3GlcNAcz (Table 1, structure 1; see Fig. lA,B, peak at 17 5 mm). In addition to being useful for monitoring the B-N-acetylhexosammldase release of terminal GlcNAc from isolated ohgosacchartdes, HPAEC-PAD and HPAEC-PAD

the MA1

column

can be used to directly

monitor

the presence

of termmal

GlcNAc on glycoprotems. Such an assaycould be useful for momtormg termtnal carbohydrate modifications in therapeutic glycoprotems; these modifications have been shown to affect the stability and efficacy of therapeutic

Table 1 Oligosaccharide

Substrates

for Exoglycosidase

Analyses

Structure

Retention Time

1 ManjGlcNAc2

175mm Mana( 1+6) .L Manp( 1+I)GIcNAcfi( t Mana( l-+3)

1--&)GlcNAc

2 FucosylatedManJGlcNAcz Mana( l-+6) Fuca( 1+6) d -1 Manf3( 1+l)GlcNAc~( 1+d)GlcNAc t Man(al-+3)

13 7mm

3 Asralo agalacto blantennary GlcNAcP( 1+2)Mana( 1+6) -1 Man( p 1+GlcNAc( T GlcNAcP( 1+2)Mana( 1+3)

22 3 mm

p l-+I)GlcNAc

4 Asalo agalacto brantennary, Core Fuc Fuca( l-6) GlcNAcj3( 1+2)Mana( l-+6) J J Manp( 1+4)GlcNAc~( l+l)GlcNAc ? GlcNAcp( 1-+2)Mana( l-+3)

187mm

5 Aslalo agalacto tetraantennary GlcNAcp( l-+6) 1 GlcNAcp( 1-+2)Mana( 1+6) L Manp( 1+d)GlcNAcp( t GlcNA@( 1+2)Mana( l-+3) t GlcNAcP( 1+I)

252mm

73

1+4)GlcNAc

74

Weitzhandler et al. A

Fig. 1. HPAEC-PAD of P-N-acetylhexosamimdase digests of oligosaccharlde standards. (A) DIgestIon of an aslalo agalacto blantennary ohgosacchartde standard (B) Digestion of an aslalo agalacto tetraantennary oligosaccharide standard. Peaks 1, 3, and 5 refer to structures 1,3, and 5 identified m Table 1. (Bottom tracmgs in A and B: dashed line, chromatography of substrate;sohd line, chromatography of digestion products.)

glycoproteins (12). The P-N-acetylhexosamimdase release of terminal GlcNAc from a monoclonal IgG is shown m Fig. 2A (bottom tracing; compare solid vs dashed line). Additlonally, the absence of contammatmg exoglycosidases is apparent by the absence of release of any other monosaccharides. To assesswhether the released GlcNAc was derived from a GlcNAc-terminated N-linked olrgosaccharlde, the monoclonal IgG was treated with peptldeN-glycosidase (PNGase F), an amidase that nonspecifically releases N-lmked oltgosaccharides from glycoprotems. The N-linked ohgosaccharide map of the monoclonal IgG 1s shown m Fig. 2B (bottom tracing; sohd line). The major PNGase F product peak eluted between 18.5 and 19 min with a retention time slmllar to a core fucosylated aslalo agalacto biantennary ohgosaccharide standard (18.7 min, see Table 1, structure 4; see also Fig. 2B, peak 4). The second PNGase F released peak had a retention time of 21 min and could represent a monogalactosylated, biantennary oligosaccharlde with core fucosylatlon, as has

75

CarboPac MA 7 Assay of Terminal GlcNAc

GIcNAc

B

0

5

10

15

Time

20

25

30

(min)

Fig. 2. HPAEC-PAD of P-N-acetylhexosamimdase dtgests of of a monoclonal TgG (A) Bottom tracing 1s P-N-acetylhexosammidase digests of a monoclonal IgG (dashed lure, chromatography of IgG substrate; sohd Ime, chromatography of drgestron products) Top tracmg 1schromatography of monosaccharide standards; Peak a, fucose, b, GlcNAc; c, GalNAc, d, mannose; e, glucose; f, galactose. (B) Bottom tracing 1s PNGase F drgest of a monoclonal IgG (dashed lme, chromatography of IgG substrate; solid bne, chromatography of PNGase F-released N-lurked ohgosacchartdes from the monoclonal IgG; arrows indicate ohgosaccharrde peaks). l3-N-acetylhexosamnudase treatment of PNGase F-released N-linked ohgosacchartdes 1s shown m the second to bottom tracing Note appearance of product peaks correspondmg to GlcNAc and fucosylated Man3GlcNAcz. Peaks 2 and 4 represent chromatography of ohgosaccharrde standards correspondmg to structures 2 and 4 in Table 1.

been reported m other monoclonal IgGs (7). To assesswhether the PNGase F released oligosaccharides were terminated with GlcNAc, the PNGase F released oligosaccharides were treated with P-N-acetylhexosaminidase. The results of j3-N-acetylhexosammidase treatment of the PNGase F released ohgosaccharides are depicted in Fig. 2B (second from bottom tracing). Both peaks corresponding to PNGase F released oligosaccharides disappeared and two new peaks appeared. The two new peaks had retention times corresponding to released GlcNAc (15.6 mm) and to a fucosylated Man3GlcNAcz standard (17.5 mm, see Table 1, structure 2; see also Fig. 2B, peak 2). Thus, the

76

Weitzhandler et al.

two N-linked ohgosaccharlde peaks released from the monoclonal IgG are fucosylated and have terminal GlcNAc. To obtam further structural mformatlon regarding GlcNAc linkages, one could use the Streptococcuspneumomae j3-N-acetylhexosammldase, an exoglycosldase which shows much more efficlent cleavage of the GlcNA@ 1+2)Man if the Man residue 1snot substituted with GlcNAc at the C-6 (13). This method 1s also useful for monitoring the j3-N-acetylhexosamimdasepreparation for contammatmg exoglycosldase actlvlties as other monosaccharide digestion products would be readily discernable. 2. Materials 1. 2 3 4

HPLC grade deionized water (see Note 1). 50% NaOH solution (w/w) (Fisher Sclenttfic, Pittsburgh, PA) (see Note 2) Reference-grade monosaccharldes (Pfanstlehl Laboratones, Waukegan, IL) Ollgosaccharlde standards Man3GlcNAcz (Table 1, structure 1) fucosylated MaqGlcNAcz (Table 1, structure 2), aslalo agalacto blantennary (Table 1, structure 3), aslalo agalacto blantennary, core fucose (Table 1, structure 4), and aslalo agalacto tetraantennary ollgosaccharlde (Table 1, structure 5) (Oxford GlycoSclences, Abmgdon, UK) P-N-acetylhexosammldase, Jack bean (Oxford GlycoSclences) (see Note 3) PNGase F (New England BloLabs, Beverly, MA) (see Note 4). 25-mL Plastic pipets (Fisher Sclentlfic). 1 5-mL Polypropylene mlcrocentrlfuge tubes, caps, and 0 rmgs (Sarstedt, Newton, NC). 9 Autosampler vIaIs* 12- x 32-mm disposable, limited-volume sample vials, Teflon! silicone septa, and caps (Sun Brokers, Wllmmgton, NC). 10 Nylon filters (Gelman Sciences, Ann Arbor, MI) 11 The chromatograph (Dlonex, Sunnyvale, CA) consists of a gradlent pump, a PAD II or PED, and an eluent degas module (EDM). The EDM is used to sparge and pressurize the eluents with helium The system was controlled and data were collected using Dlonex AI450 software. Sample inJection was accomplished with a Spectra Physics SP8880 autosampler (Fremont, CA) equipped with a 200~p.L sample loop The Rheodyne (Cotati, CA) injection valve is fitted wrth a Tefzel rotor seal to wlthstand the alkalmlty of the eluents. We also used a DX 500 BloLC system (Dlonex) configured for carbohydrate analysts with PeakNet software (Dlonex).

3. Methods 3.7. P-N-Acetylhexosaminidase

Digestion

1 Reconstitute approx 2 clg each of the neutral agalacto blantennary and agalacto tetraantennary ohgosacchandes (Table 1, structures 3 and 4, respectively) m 10 pL of 25 mM sodmm citrate-phosphate buffer, pH 5 0 2. Add 0.1 U of Jack bean /3-N-acetylhexosamlmdase m 2 p.L of 25 mA4 sodium citrate-phosphate buffer, pH 5.0

CarboPac MA 1 Assay of Terminal GlcNAc

77

3. Incubate the digest for 20 h at 37°C. 4 InJect 10 pL of each digest directly onto the column. 3.2. PNGase F Digestion 1. Reconstitute approx 100 pg of a monoclonal IgG in 10 pL of 2.5 mM sodium citrate-phosphate buffer, pH 5.0 2. Add 2 pL of the PNGase F preparation 3 To an identlfical PNGase F digest of the same monoclonal IgG, add 0 1 U of Jack bean P-N-acetylhexosammldase m 2 & of 25 mA4 sodium citrate-phosphate buffer, pH 5 0 4. Separately inject 10 & of each digest directly onto the column.

3.3. Chromatography

and Defection

of Carbohydrates

Separations of monosaccharides and ollgosaccharides can be achieved using a Dionex BtoLC system equipped with a CarboPac MA1 column (4 x 250 mm) and a CarboPac MA1 guard column working at an isocratic concentration of 480 mM NaOH and a flow rate of 0.4 mL/mm at ambient temperature over 35 min. Separated mono- and ollgosaccharldes are detected by PAD with a gold electrode and triple-pulse amperometry (E, = 0.05 v, t] = 420 ms; E2 = 0.80 V, t2 = 360 ms; E, = -0.15 V, t3 = 540 ms), measuring at 1000 nA full scale. 4. Notes 1. It is essential to use high quality water of high resistivity (18 MeQ) and to have as little dissolved carbon dloxlde m the water as possible Blologlcal contammatlon should be absent. The use of fresh Pyrex glass-dlstllled water is recommended The still should be fed with high-renstlvlty (18 MeR) water. The use of plastic tubing m the system should be avoided, as plastic tubing often supports mlcroblal growth Degas appropriately. 2. It is extremely important to mmimize contamination with carbonate Carbonate, a divalent anion at pH 5 12, binds strongly to the columns and Interferes with carbohydrate chromatography Thus carbonate IS known to affect column selectivity and produce a loss of resolution and efficiency. Commercially available NaOH pellets are covered with a thin layer of sodium carbonate and should NOT be used. Fifty percent (w/w) sodium hydroxide solution is much lower m carbonate and IS the preferred source for NaOH Diluting 104 mL of a 50% NaOH solution mto 2 L of water yields a 1.OMNaOH solution. Degas appropriately. 3 This enzyme has a broad specificity, cleaving nonreducing terminal /3-Nacetylglucosamme residues (GlcNAc) and P-N-acetylgalactosamine (GalNAc) with l-2,3,4, and 6 linkages. 4. PNGase F 1s an amldase from Flavobacterwm meningoseptlcum The enzyme cleaves between the innermost GlcNAc and asparagme residues of high mannose, hybrid, and complex ohgosacchandes from N-linked glycoprotems

78

Weitzhanciler et al.

Acknowledgment We thank Sylvia Morris for her work on the manuscript.

References 1. Jacob, J. S. and Scudder, P (1994) Glycosidases in structural analysis Methods Enzymol 230,28O-299 2 Scudder, P, Neville, D C A, Butters, T. D , Fleet, G. W J , Dwek, R A, Rademacher, T. W , and Jacob, G S. (1990) The isolation by ligand affimty chromatography of a novel form of o-L-fucosidase from almond J Blol Chem 265, 16,472-l 6,477. 3 Butters, T D., Scudder, P., Willenbrock, F W, Rotsaert, J. M V, Rademacher, T, Dwek, R A , and Jacob, G. S. (1989) A serial afftmty chromatographlc method for the purification of charonza lampas a-L-fucosidase, m Proceedings of the 10th Znternatzonal Symposzum on Glycoconjugates, Sept. IO-15 (Sharon, N , Lis, H , Duksm, D , and Kahane, I., eds ), Magnes, Jerusalem, Israel, pp. 3 12,3 13. 4 Davidson, D. J and Castellmo, F. J (1991) Structures of the asparagme-289-linked ohgosacchartdes assembled on recombinant human plasmmogen expressed m a Mamestra brasszcae cell lme (IZD-MB0503) Blochem&ry 30,668%6696 5 Grollman, E. F , SaJt, M , Shimura, Y., Lau, J. T , and Ashwell, G (1993) Thyrotropm regulation of sialic acid expression m rat thyroid cells J B~ol Chem 268, 3604-3609 6 Willenbrock, F W, Neville, D. C A, Jacob, G. S , and Scudder P. (1991) The use of HPLC-pulsed amperometry for the characterization and assay of glycostdases and glycosyltransferases Glycoblology 1,223-227. 7 Weitzhandler, M , Hardy, M , Co, M S., and Avdalovic, N (1994) Analysis of carbohydrates on IgG preparations, J Pharm Scz 83, 1670-l 675 8. Lm, A. I , Phthpsberg, G A., and Halttwanger, R S (1994) Core fucosylatlon of high-mannose-type ohgosaccharides m GlcNAc transferase I-deficient (Let 1) CHO cells. Glycobiology 4, 895-90 1, 9. Hardy, M. R., Townsend, R. R., and Lee, Y. C (1988) Monosaccharide analysis of glycoconjugates by amon exchange chromatography with pulsed amperometric detection Anal Biochem 170, 54-62. 10. Chou, T. Y., Dang, C. V, and Hart, G. W. (1995) Glycosylation of the c-Myc transactivation domain Proc. Nat1 Acad SCI. USA 92, 4417-442 1 11. Lm, A. I , Polk, C., Xiang, W. K , Phihpsberg, G. A., and Haltiwanger, R S (1993) Novel fucosylation pathways m parental and GlcNAc transferase I deficient (Let 1) CHO cells. Gfycobzology 3, 524 (Abstract) 12. Varki, A. (1993) Biological roles of oligosacchandes. all of the theories are correct Glycobzology 3,97-130. 13. Yamashita, K , Ohkura, T.,Yoshima, H. and Kobata, A (198 1) Substrate specificity of Dtplococcal P-N-acetylhexosammrdase, a useful enzyme for the structural studies of complex type asparagine-linked sugar chams. Blochem. Biophys Res Commun 100,226-232.

HPLC and HPAEC of Oligosaccharides and Glycopeptides Michael J. Davies and Elizabeth F. Hounsell 1. Introduction

The purification of oligosaccharides and glycopeptides presents many problems, not the least of which is the diversity of possible structures. To date no single chromatographic or electrophoretic method has been proven to separate all ohgosaccharide isomers m a buffer system that 1scompatible with immediate nuclear magnetic resonance spectroscopy (NMR) or mass spectrometry (MS) analysis. In addition, the inherent nonchromogemc nature of ohgosaccharides often makes their on-line sensitive detection very difficult. This chapter details the range of high-resolution chromatographic techniques available to glycobiologists either utilized singly or m concert to achieve the separation of pure oligosacccharides. Included are methods for the release of ohgosaccharides and monosaccharides from glycoprotems, as well as methods for then separation and sensitive detection. There are several described methods for the release of olrgosaccharides and monosaccharides, all of which have their advantages and disadvantages. For N-linked oligosaccharides, the prmcipal methods of oligosaccharide release are hydrazmolysis and enzymatic treatment with peptide-N-glycosidase F (PNGaseF). Hydrazinolysis will result in the nonselective release of all oligosaccharides, though some loss of sialylation can occur and the reaction involves the use of anhydrous hydrazine, which is highly mflammable and toxic. Digestion of glycoproteins with PNGaseF will in theory release all oligosaccharides nonselectively, though often the conditions will have to be optimized for each glycoprotem studied to ensure maximum release. The release of O-linked oligosaccharides can be achieved by either alkaline elimmation (Chapters 1,9, and 11) or hydrazinolysis (Chapters 7 and 8). However, alkaline From

Methods In Molecular B/o/ogy, Vol 76 Glycoanalym Protocols Edited by E F Hounsell 0 Humana Press Inc , Totowa, NJ

79

80

Davies and Hounsell

elnnmatton will result m the terminal N-acetylgalactosamme being reduced to an alditol, thus a reoxtdation step has to be Introduced before subsequent labelmg with sensitive UV or fluorescent chromophores can occur (I). Hydrazmolysis ~111release O-lmked sugars m an unreduced form. However, its efficacy for the release O-lurked sugars from high molecular weight mucms is yet to be proven. Enzymatic release of O-linked sugars is limited, since the enzyme O-Glycanase (Chapter 11) will only optimally release the disaccharide GalPl-3GalNAc (2,3). The release of monosaccharides for compositional analysts usually involves an acid hydrolysis mvolvmg either trifluoroacetic acid (TFA) or HCl, each acid resulting m different quantitative release efficiencies of some monosaccharides, followed by analysis by high pH amon exchange chromatography (HPAEC) with pulsed amperometric detection (PAD). Alternatively, methyl glycosides can be prepared by methanolysis (4), which can then be derivatized with trimethylsilylating agents and analyzed by gas-ltquid chromatography (GC) or GC-MS The methanolysis method has the advantage (Chapter 1) that monosaccharides are more reliably detected (HPAEC-PAD ts prone to ammo acid mterference, for example, but see this chapter, Subheadings 2.6., 3.5, and Note 13). However, the sensitivity obtained is much less than by HPAEC-PAD Glycopeptides and pepttdes are usually separated by means of reversedphase (RP) chromatography. However, the complexity of peptrde maps generated from multiply glycosylated protems often results m very complex chromatograms with several coelutmg peptides and glycopeptides. Subsequent separation of coelutmg fractions on a hydrophobic mteraction column (HIC), such as a porous graphmzed carbon (PGC) column, will often resolve peptide mixtures that coelute on RP, and if the peptide is only a few residues m size, resolution of individual glycoforms may be achieved ($6). Once a glycan library has been generated from a glycoprotem by either chemical or enzymatic methods, there are several chromatographic systems by which pure ohgosaccharide isomers can be obtained. To date, the highest resolution of sralylated ohgosaccharides has been achieved by HPAEC-PAD (7). However, the aggressive eluants used in the separation (O.lM NaOH and 1M NaOAc) make preparative separations very difficult. In addrtion, the resolution of neutral oligosacchartde alditols (8) and sulfated oligosaccharides (unless they are reduced to alditols) is poor by this method (9). Normal phase (NP) and RP chromatographies are limited in their resolvmg power of underivitized ohgosaccharides, and the sensitivity of detection by either UV or refractive index (RI) 1salso Inadequate. The use of PGC, e.g., Hypercarb S (Hypersil, Runcorn, UK) or GlycoSep H TM (Oxford GlycoSciences, Abingdon, UK) will give improved separations of underivatized oligosaccchartdes (S,lO, 11) as will the use of an inert amon-exchange column, e.g., GlycoSep CTM (Oxford GlycoSciences) (12) run either as an anion-exchange column or a hydrophillic mter-

HPLC and HPAEC

81

action column. The addition of a fluorophore to oligosacchartde will greatly Increase the sensitivity of that obtainable by PAD). This also results in power of NP and RP chromatographies, allowing

the reducing terminus of an of detection (often m excess

an increase in the resolving them to be used as part of a

two-dimensional (Chapter 7) or a three-dimensional (13) mapping strategy. Although filter-based fluorescence detectors give improved detection over UV and RI methods, the greatest Improvement in sensitivity 1s seen with modern

monochromator-based detectors. The addition of a fluorophore also increases the sensitivity of detection, if not the efficiency of separation, for PGC and GlycoSep CTMcolumns. Size-exclusion chromatography (SEC) on Bio-Gel P4 columns has been used extensively for the separation of neutral ohgosaccharides, though the resolution

of linkage isomers is not possible by this technique

(14). The complete sequencing of purified oligosaccharides can be achieved by MS (IS), GC-MS, fast-atom bombardment (FAB)-MS, or tandem MS, NMR (Chapter l), or an optimized array of exoglycosidases and either Bio-Gel P4 or NP chromatographies (Chapter 4 and ref. 16). A combination

of chromatographles

will probably

be necessary to isolate

pure oligosaccharide isomers from an oligosaccharlde library. In this chapter, we describe several complementary chromatographic techniques for the punfication of N- and O-linked

2. Materials 2.1. Instrwnen

ohgosaccharides

and glycopeptides.

ta tion

1 Blocompatible (PEEK or Tltanmm) ternary or quaternary HPLC. 2 HPLC-compatible detectors: pulsed electrochemical detector e.g., Dionex (Camberley, UK), ED-40, PED-2 or Antec (Leiden, The Netherlands) Decade, UV detector (ideally monchromator-based dual-wavelength instrument), fluorescence detector e.g., Waters (Watford, UK) 474. 3. Fraction collector and apparatus for postcolumn reagent addition (see Note 1) 4. Solvent degassing system (e g., helium sparging or on-line vacuum degassing system). 6 Bench-top GC-MS, such as Hewlett Packard 5980 GC with 5972A MSD or 6890 GC with 5973 MSD and HP-5MS capillary column, and 0 54-mm id retention gap. 7. Lyophillzer. 8. Bench-top mtcrocentrlfuge 9 Rotary evaporator e g., Savant SpeedVac (Life Sciences International, Runcom, UK).

2.2. Peptide Mapping 1 TPCK-treated chymotrypsin-free trypsm (bovine pancreas EC 3.4.2 1.4) (Sigma, Poole, UK) 2. Trypsm digestion buffer* 100 ml4 ammonium bicarbonate, pH 8.0

Davies and Hounsell

82

3 Blocompatible HPLC with UV detector 4 Reversed-Phase (C,,) HPLC column, e g., Hypersll ODS 250 x 4 6 mm, Hypersll (Runcorn, UK) 5 Eluant A 0.1 % TFA m acetomtrlle (HPLC grade) 6. Eluant B. 0 1% TFA m water (HPLC grade).

2.3. PNGaseF Treatment 1, 2 3 4

PNGase F (recombinant) and digestion buffer (Oxford GlycoSaences) (see Note 2). Blogel P2 gel-filtration resin (Bio-Rad, Hercules, CA) Microcon 10 concentrators (Amlcon, Beverly, MA). Toluene.

2.4. Hydrazinolysis 1. 0.5~mL Screw-capped V-bottomed Reacti-vlalsTM (Pierce and Warrmer, Chester, UK) Anhydrous hydrazme (Pierce and Warriner). Whatman CF- 11 cellulose chromatography medium. Reagent A: butanol.ethanol acetic acid 4.1:0.5 (v:v v). Reagent B. butanol.ethanol.water 4: 1 1 (v:v*v) Reagent C acetic anhydnde:methanol2*5 (v*v) Reagent D. 0 2M sodium acetate Reagent E* O.lM copper acetate Sep-Pak CIR cartrldge (Waters, Watford, UK) 10 Chelex 100 (H+ form) Blo-Rad, UK. 11 AG50 (H+ form) Bto-Rad, UK 2 3. 4 5 6. 7 8 9

Glycorelease also be used.

N- and O-Glycan

recovery kit (Oxford

GlycoSclences)

can

2.5. Acid Hydrolysis 1 Hydrochloric acid (constant boiling) 2 Van-clean 3 mL screw cap vials (or other acid washed vials) (Pierce and Warriner).

2.6. HPAEC-PAD 1. Blocompatible ternary HPLC system with pulsed electrochemical detector (Dlonex PED2, ED-40 or Antec Decade) and post column addition apparatus. 2. Ammo trap HPLC guard column (Dlonex). 3. Borate trap HPLC guard column (Dlonex) 4. CarboPacPA1HPLC column and guard (Dionex) 5. CarboPac PA100 HPLC column and guard (Dlonex). 6 Eluant C = 100 mMNaOH (made up from 50% NaOH solution, BDH, Poole, UK) 7 Eluant D = 300 mA4NaOH 8. Eluant E = H,O. 9. Eluant F = 1 5 mM sodium acetate (ACS grade, Aldrich, Glllingham, UK) m 50 mM NaOH.

83

HPLC and HPAEC 10 Eluant G = 1M sodmm acetate m 100 mMNaOH 11 Eluant H = OSM sodmm acetate m 100 mM NaOH

2.7. Fluorescence 1. 2. 3 4

Labeling and SEC

Signal 2-AB labeling kit (Oxford GlycoSciences) Dowex AG5OW-X8 (H+ form) (Oxford GlycoSciences). Dowex AG 1-X8 (acetate form) (Oxford GlycoScrences). RAAM 2000 Glycosequencer (Oxford GlycoSctences) or Bra-Gel P4 column (100 x 2 cm) m a waterJacket at 55°C and HPLC pump with refractrve index and fluorescence detectors

2.8. HPLC 1 Ternary brocompatrble HPLC system wrth erther UV or fluorescence detectton. 2. GlycoSepCTM HPLC column (Oxford GlycoScrences). 3. Hypercarb S HPLC column 100 x 4.6 mm, (Hypersrl) or GlycosepHTM HPLC column (Oxford GlycoSctences) 4 Eluant J = H,O. 5 Eluant K = acetomtrrle (HPLC or fluorescence grade, Fisher Chemrcals, Loughborough, UK) 6 Eluant L = 0.5M ammomum acetate, pH 4.5.

2.9. GC-MS Methylation

Analysis

1. Bench-top GC-MS system fitted wrth 0 54-mm id retention gap and 0.2-mm id HPS-MS capillary column (or alternatrve column of equivalent polarrty) 2 5-mL V-bottomed Reactr-vials (Pierce and Wanner) 3 0 5-mL V-bottomed Reactt-veals (Prerce and Warrmer). 4. Anhydrous DMSO (Sigma) 5 Anhydrous sodium hydroxide pellets (BDH, Poole, UK) 6. Methyl iodide (Sigma) 7 Trifluoroacetic acid (Pierce and Warrmer). 8 Chloroform 9 Pyndine. 10. Acetic anhydride 11. Ammomum hydroxide 12 Sodium borodeuteride (Sigma) 3. Methods

3.7. Protease Digest 1. Dissolve 1 mg of trypsm in 500 $ of trypsm dtgestton buffer immediately before It is requn-ed (see Note 3). 2. Dissolve either native or reduced and carboxy-methylated (see ref. 17) glycoprotein m enzyme/buffer solution at 20 pg enzyme/l mg glycoprotem with 5 pL of toluene to prevent bacterial growth, and incubate for up to 72 h at 37°C wrth a further additton of enzyme (10 pg enzyme/l mg glycoprotem) after 24 h

Davies and Hounsell

84 3 Wash the digest three times with 0 1 mL HZ0 4. Lyophihze the digest prior to chromatography

3.2. Peptide Mapping 1. Wash the RP column m 100% eluent A for 30 mm at a flow of 1 mL/mm 2 Next wash the column m 100% eluent B for 30 min at a flow of 1 mL/mm 3 Eqmhbrate the column in 98% B 2% A at a flow of 1 mL/min

4 Inject approx 20 nmol of glycopeptide mixture m water. 5. Elute the peptides and glycopeptides with the followmg gradient at a flow of 1 mL/mm. OmmA=2%B=98% 10 mm A = 2% B = 98% 90 mm A = 82% B = 18% lOOmmA=82%B=18% llOmmA=2%B=98% 6. Collect and pool fractions as necessary based on UV detectron, either smglewavelength at 2 10 nm or dual-wavelength of 2 10 nm for peptrdes and 206 nm for sugars (see Note 4). 7 Further fractionate coelutmg glycopeptides on a porous graphrtized carbon HPLC column under the same elution conditions as the C 18 column (5,6)

3.3. N-Linked

Oligosaccharide

Release by PNGaseF

1. Dry pure, desalted glycoprotem mto a 2-mL screw-top Eppendorf 2 Suspend glycoprotein m 200 pL of PNGaseF digestion buffer (see Note 5) 3 Add 1 U/O.5 mg of glycoprotem of PNGaseF and 5 p.L of toluene (to prevent bacterial growth) (see Note 6) 4 Incubate at 37°C for up to 72 h 5. Centrifuge the sample briefly, and transfer to Mrcrocon 10 concentrator 6. Spin the concentrator at 14,000g for 20 min to separate protein and oligosaccharide components (see Note 7) 7 Transfer the filtrate to a 2-mL Bio-Gel P2 column m a glass Pasteur pipet equihbrated wrth water 8 Elute oligosaccharides with 800 pL of HPLC grade water (see Note 8) 9. Dry desalted ohgosaccharides for further analysrs.

3.4. N-linked and O-linked Oligosaccharide by Hydrazinolysis (18)

Release

1 Dry salt-free glycoprotem mto a V-bottomed Reacti-vial, and remove from the lyophrlizer immedrately before the reaction is due to commence 2. Using a clean, dry, acid-washed glass plpet, transfer 100 pL of anhydrous hydrazme to the vial and cap rmmediately. For the release of N-linked glycans, mcubate at 95°C for 5 h. For the release of O-linked glycans, incubate at 60°C for 5 h (see Note 9).

85

HPLC and HPAEC

3. Allow the Reacti-vial to cool to room temperature, and transfer the reaction mrxture to a 1-mL cellulose (Whatman CF- 11) microcolumn washed with reagent A 4 Wash the column with 3 x 1 mL reagent B. 5 Re-N-acetylate the glycans on the column by addition of 1.4 mL reagent C for 30 mm at room temperature. 6. Wash the column with 4 x 1 mL reagent B followed by 1 mL methanol (see Note 10). 7 Elute the oltgosacchartdes with 2 x 1 mL reagent D 8. Complete the re-hr-acetylatron with 0.1 mL acetic anhydride for 30 mm at room temperature. For selective purrficatton of N-Ltnked glycans, proceed to step 11. 9 Wash a Sep-Pak Cl8 cartridge with 2mL methanol and 2 mL Hz0 10 Transfer sample contammg O-glycans to the cartridge, and collect the eluate. Elute the remaining glycans with 0.5 mL of H20 11 Add 100 pL of reagent E, and incubate at room temperature for 30 mm. 12 Partially dry the glycans before removal of cations on a microcolumn of Chelex 100 H+ form (0.4 mL resin) above Dowex AGSOW Ht form (1 mL resin) 13 Elute the ohgosaccharides from the column with 4 x 0 5 mL H20, and lyophtllze for further analysis

3.5. Acid Hydrolysis

and HPAEC-PAD of Monosaccharides

1 For the analysts of neutral and amino monosaccharides, dry 10 @ (see Note 11) of pure glycoprotem m a clean glass screw-capped vial 2 Add 100 pL 2M HCL (see Note 12) under a stream of NZ, and seal the vial 3 Incubate at 100°C for 2 h 4 Dry the sample, and wash with 3 x 100 & water. 5 Wash a CarboPac PA1 column fitted with ammo trap guard column and borate trap with 100% eluant C for 30 min at a flow of 1 mL/min, and with addmon of eluant D as postcolumn reagent at a flow not >l mL/min (see Note 13) 6. Equilibrate the column m 98% eluant E/2% eluant F at a flow of 1 mL/mm 7. Inject the sample in water with 100 pmol of deoxyglucose as internal standard 8 Elute the sample using the followmg gradient at flow 1 pL/min (see Fig. 1) OmmE=98%F=2% 35minE=98%F=2% 36 mm E = 0% F = 0% C = 100% 43 min E = 0% F = 0% C = 100% 44 min E = 98% F = 2% 9. Detect monosacchartdes (PAD and gold electrode) with the followmg pulse potentials (see Note 14) E, = + O.lV Tr = OS- 0.72s E2 = + 0 7V T2 = 0.73s - 0.85s E3=-03VT3=086s-1 2sT,,,=O52s-0.72s 10. Quantrtate the monosaccharrdes relative to hydrolyzed monosaccharide standards, and assign peak tdenttttes relative to the retention of the deoxyglucose internal standard

Davies and Hounsell

1cmo~mu -~

o------~-l0

I ‘15

---___ 20

-~

I

25

Time (mms)

Frg. 1 HPAEC-PAD of monosaccharrde standards separated on a CarboPac PA1 column with ammo trap and borate trap guard columns

11 For the analysis of siahc acids, dry 100 pg (see Note 11) of glycoprotem ma clean glass screw-top vial. 12 Add 100 $ of 0 1M HCl, seal the vial, and incubate at 70°C for 1 h (see Note 15). 13. Dry the sample, and wash three times with 100 pL water 14 Wash CarboPac PA1 column fitted with a PA1 guard and borate trap guard column with eluant G for 30 min at 1 mL/mm. 15 Wash the column with eluant C for 30 mm at 1 mL/mm. 16. Equrhbrate the column m 95% C, 5% G at 1 mL/mm 17 Inject the sample m water 18 Elute the sample using the following gradient a flow of 1 mL/mm* OminC=95%G= 5% 5mmC=95%G= 5% 30mmC=70%G=30% 35 mm C = 70% G = 30% 36 mm C = 95% G = 5% 19. Quantrtate NeuAc and NeuGc present wrth known standards run on the same day (Fig. 2) 20 Wash the column wrth 100% eluant G, and store m 100% C eluant. Wash the pumps with water prior to turning the system off

HPLC and HPAEC

87

/-----’

OL

!

I

I

5

10

15

------~

-_

1

20

Tlme(mms)

Fig 2. HPAEC-PAD of Neu5Ac and Neu5Gc on a CarboPac PA1 column with PA1 and borate trap guard columns

3.6. HPAEGPAD

of Sialylated N-Linked

Chains (Fig. 3)

1. Wash a CarboPac PA 100 column fitted a with PA 100 guard and Borate Trap with 100% eluant H at a flow of 1 mL/mm for 30 min. 2. Wash the column m 100% eluant C at a flow of 1 mL/min for 30 mm (see Note 16) 3. Equilibrate the column m 98% C, 2% H. 4 Inject the sample in water (see Note 17). 5. Elute the sample using the following gradient at a flow of I mL/mm. OmmC=98%H= 2% 5mmC=98%H= 2% 40 mm C = 60% H = 40% 45 mm C = 60% H = 40% 48 min C = 98% H = 2% 6 Wash the column in 100% H and store in 100% C; wash the pumps with water.

3.7. Fluorescence

Labeling

with 2-Aminobenzamide

(2-A B) (19)

1 Dry salt-free glycans mto a 0 5-mL Eppendorf tube (see Note 18) 2 Prepare labeling reagent of 70% DMSO, 30% glacial acetic acid containing 0 25M2-AB and 0 1MNaCNBH3 (see Note 19).

Davies and Hounsell

88

10

I 30

20

I 40

I 50

Tune(mms)

Fig 3. HPAEC-PAD of N-lmked oligosaccharldes released by PNGaseF from fetum on a CarboPAC PA100 column with PA100 and borate trap guard columns.

3. Add 5 p.L of labeling reagent, and incubate the sample at 65°C for 2 h 4. Centnfuge the sample briefly 5 Transfer to a hydrophlhc separation dtsk (supplied with labeling kit) washed with 1 mL water, 1 mL 30% acetic acid, and 1 mL acetomtrile. 6 Load the sample onto disk and leave for 15 min (see Note 20) 7. Wash the tube with 100 pL acetomtnle, and add to disk 8 Wash disk with 1 mL acetomtrile followed by 5 x 1 mL 4% water in acetomtrile. 9 Elute the sample with 3 x 0.5 mL water (see Note 21). 10 Dry the sample to about 100 p.L 11. Prepare 150 r.lr, of AGSO-X12 resm m a mlcrocolumn, and wash with 5 mL 1 5% Triethylamme m water followed by 3 x 1 mL water (see Note 22) 12 Add 150 pL of AGl-X8 (acetate form) to the mlcrocolumn takmg care not to disturb the AG50 resm 13. Wash with 0.5 mL water. 14. Load sample m 100 & water, and elute 4 x 0.4 mL water 15 Filter sample through a 0.45~pm filter, and dry for further analysis.

89

HPLC and HPAEC

90000

A3

T

A4

I ~ 0

I

_

10

-,--------+~-20

+ 30

40

----+--50

60

Tune (mm)

Fig. 4. GlycoSepC chromatography 2-AB-labeled glycans released from fetum with PNGaseF The labels Al, A2, A3, and A4 represent the numbers of sialic acids present.

3.8. Preparative 1 2 3 4. 5

HPLC on a GlycoSep CTMHPLC Column (Fig. 4) (12)

Wash the column with eluant J for 30 mm at a flow of 0 4 mL/mm. Wash the column with eluant K for 30 mm at a flow of 0 4 mL/mm Wash the column with eluant L for 30 min at a flow of 0.4 mL/mm (see Note 23) Rewash the column with eluant J for 30 mm at a flow of 0.4 mL/mm Equilibrate the column m 25% eluant J, 75% eluant K at a flow of 0.4 mL/mm.

6. Inject the 2-AB-labeled

sample m a 70:30 v/v mixture of acetomtrlle /water (see

Note 24), with fluorescence detection using an excitation h = 330 nm and an excitation h = 420 nm 7. Elute the sample with the followmg gradlent with fraction collection at a flow of 0.4 mL/mm (see Note 25). 0 mm J = 25% K = 75% 5 mm J = 25% K = 75% 30 mm J = 37 5% K = 62.5% 50 mm J = 40% K = 0% L = 60% 55mmJ=40%K=O%L=60% 60 mm J = 25% K = 75% 8. Pool and dry peaks 9 Wash the column with eluants K and L, and store m 75:25 (v/v) acetonMe/ water

Davies and Hounsell

90

3.9. Porous Graphitized Carbon Chromatography and O-linked Oligosaccharides (Fig, 5) (5,6,10,11) 1 2 3 4 5 6 7.

8 9

of N-linked

Wash the column with eluant J for 30 mm at a flow of 0 75 mL/mm (see Note 26) Wash the column with eluant K for 30 mm at a flow of 0 75 mL/mm Wash the column with eluant L for 30 mm at a flow of 0 75 mL/mm. Rewash the column with eluant J for 30 mm at a flow of 0 75 mL/mm Equihbrate the column m 78% eluant J, 20% eluant K, and 2% eluant L at a flow of 0 75 mL/mm (see Note 27) Inject the 2-AB-labeled sample m water with fluorescence detection using an excitation h = 330 nm and an excitation h = 420 nm Elute the sample with the followmg gradient with fraction collection at a flow of 0 75 mL/mm (see Note 28) OminJ=78%K=20%L=2% 2 mm J = 78% K = 20% L = 2% 35minJ=58%K=40%L=2% 70mmJ=O%K=30%L=70% 72 mm J = 0% K = 30% L = 70% 74 mm J = 78% K = 20% L = 2% Pool and dry peaks (see Note 29) Wash the column with eluants K and L, and store m 75.25 (v/v) acetomtrile/

water 3.10. Further Methods ofAnalysis The methods described above along with those described m Chapter 7 by Hase and Natsuka can be used singly or in concert to generate samples contammg smgle-oligosaccharrde isomers from either whole glycoprotems or fractionated glycopeptides. A range of techniques is available for the further charactertzatton of ohgosaccharide structures. 3. IO. 1. RAA M Sequencing The structure of purified oligosacchartdes can be determined by the use of an optimized array of enzymes with subsequent analysis by Bio-Gel P4 chromatography (Fig. 6). The use of pattern matching software to identify the orrgrnal saccharrde from the Bra-Gel P4 profile IS described by Prime and Merry m Chapter 4. More recently, a similar sequencing technique using a NP HPLC column has been described (16). 3.10.2.

NMR

High-field NMR will provide both structural and conformatronal mformatron about ohgosaccharides. However, rt 1s hmited m that relatively large amounts of material are required. The techniques mvolved are drscussed m Chapter 1.

Tim (minr)

A2

,,,,,,,,,,,,,,,,,,,,,,“7,,‘,,,,,,,,,’,,,,, 10

20

‘,““, 30

lime (mins)

40

JO

,~~rJ,-rl-rT-,--r R”

1 70

Fig. 5 (A,B) PGC chromatography of the 2-AB derivatives of slalylated N-linked glycans Al-A4 as separated by GlycoSep CTM Chromatography (see Fig. 4)

A3

D

A4

l”‘~~,~“‘l”“~““f”“~‘“‘l”“~‘“‘l”“~””l’”’~””i””~”” 0

30

0

0

10

70

1

Ttme (mins)

Fig 5. (C,D) (continued) PGC chromatography of the 2-AB derivatives of sialylated N-linked glycans Al-A4 as separated by GlycoSep C? Chromatography (see Fig. 4).

93

HPLC and HPAEC

10

30

20 Volume

40

50

(ml)

Fig. 6. Bio-Gel P4 chromatography of oligosaccharldes released from fetum by PNGaseF after slahdase dIgestIon. The peak labels are the retention In terms of glu-

coseunits (obtained by comparisonwith a hydrolyzed dextran standard) The peak at 16.59 GU corresponds to a tetra-antennary ollgosaccharlde lacking fucose, whereas the peakscorrespondmgto 14.17 GU an 11.I7 GU are trlantennary and blantennary

oligosaccharldesrespectively 3.7 0.3. Mass Spectrometry and Methylation Analysis Several mass spectrometric techniques have been applied to the analysis of ohgosacchandes and glycopeptldes. FABMS and LSIMS will provide molecular weight information on peptldes, though the iomzation of glycopeptldes 1spoor and the direct coupling to HPLC columns is unreliable. Native and permethylated oligosacccharides can also be analyzed by LSIMS, although only structural and not linkage information is obtained. MALDI-TOF MS will readily ionize oligosaccharides, peptides, and glycopeptides, although only molecular weight information will be obtained, but at greatly improved sensitivity over FABMS. ES-MS-CID-MS (tandem MS), as described by Treumann et al. m Chapter 14 for the analysis of phospatldyhnositols, is equally applicable to the sequencing of peptides, glycopeptldes, and oligosaccharldes (20-24). This technique 1s also relatively easily interfaced to HPLC systemsusing microbore and capillary columns, and thus presents a powerful tool in the analysis of most classes of glycoconjugate The complexltles of LC-MS mstrumentatlon and analysis are beyond the scope of this chapter,and the reader 1sreferred to ref. 24 for tirther details.

Davies and Hounsell GC-MS has been used extenstvely as described below for the determination of monosaccharide linkages and IS readily available to many laboratories with relatively cheap bench-top GC-MS instruments. This procedure 1sderived from that described by Cmcanu and Kerek (25). 1 Dry a mmimum of 20 nmol of desaltedoligosaccharidesin a 5-mL Reactl-vial 2 Dissolve the glycan m 100 pL of anhydrous DMSO and somcate under an inert atmosphere for 15 mm 3 Add 100 ~,ILof a suspension of approx 2 mg anhydrous NaOH m 200 pL DMSO 4 Somcate for 15 mm m an inert atmosphere. 5 Add 200 pL methyl iodide and somcate for 15 mm 6 Stop the reaction by the addltlon of 4 mL H20, and extract the permethylated ohgosaccharldes with 3 x 300 pL chloroform 7 Wash the chloroform phase with 10 x 4mL water, dry under Nz, and then lyophlllze from 200 pL HZ0 The ohgosaccharldes can now be derivltlzed to partially methylated aldltol acetates as below or be analyzed directly at this stage by FABMS or LSIMS. 8 Dry the permethylated ohgosacchande m a 0.5-mL Reactl-vial and hydrolyze with 100 $2M TFA for 1 h at 100°C 9 Cool the reaction and evaporate 3 x 100 & of methanol. 10 Reduce the permethylated monosaccharides with 50 mM NaBD4 m 50 n&f NH40H (see Note 30) for 4 h at room temperature (or at 4°C overnight) 11 Neutralize the NaBD4 with glacial acetlc acid at 0°C 12. Wash the sample with 3 x 100 p,L methanol 13. Dry the oligosaccharide under a gentle stream of N2, and re-ll’-acetylate with 150 $ pyndine:acetlc anhydride 1: 1 for 18 h at room temperature. 14. Evaporate the excess pyridine/acetic acid under NZ, and wash the sample and evaporate three times from 20 pL toluene 15. Inject the samples using an autosampler mto a bench-top GC-MS system using cool on-column Injection with a 0.54~mm id deactivated retention gap. Run a temperature gradient from 6O-265°C (increasing at 5”Umm) with a constant, vacuum-compensated flow of 1 mL/mm. The mass spectrometer can be operated in either Scan or SIM mode with electron lomzatlon (see Note 31) 4. Notes 1 The apparatus for the delivery of the postcolumn reagent should be of a pneumatic type, since pump-based methods of solvent delivery will generate excesssively noisy baselines owing to the lack of pulse dampenmg. 2. PNGaseF preparations should be free of glycerol, since this can interfere with subsequent fluorescence labeling reaction efficlences. 3 Although we have used trypsm as an example, the decision as to which protease to use for a primary digest should be based on the amino acid sequence of the protein. Ideally, the protease should be able to generate glycopeptides containing

HPLC and HPAEC

4.

5

6

7 8.

9

10. 11

12.

95

a single glycosylation site. If there is no suitable protease available, then chemical methods of cleavage, such as CNBr cleavage, can be used We do not recommend the use of Pronase, smce it is generally impure and may have contaminatmg polysaccharides or glycosidases present. Although this method describes the fractionation of glycopeptides with fraction collection and then further analysis of the released sugars from each glycopeptide, increasingly on-line mass spectrometric methods are being used (see Subheading 3.10.3. and refs. 2&24). The use of ESMS coupled to CID-MS has greatly simplified the peptide mappmg process, and complex peptide maps can be sequenced in a single chromatographic analysts Although for many glycoproteins PNGaseF will give effective deglycosylation of native protems, in some cases, it will be necessary to denature the protein to ensure maximum deglycosylation Denaturation can either be carried out by means of a proteolytic digest or by boiling the reaction mixture (without enzyme) in 0 5% SDS and 5% P-mercaptoethanol If the sample is denatured in SDS, then the PNGaseF digest must be carrted out in the presence of a nomomc detergent (e.g , 10% n-octyl glucoside or nonidet P-40) It is advisable to assay the degree of glycosylation both before and after PNGaseF digestion by HPAEC-PAD. Extra care should be taken to ensure the thorough desalting of denatured samples to ensure that subsequent fluorescence labeling reactions are not affected Deglycosylation can also be performed using PNGaseA This is especially relevant if plant glycoproteins are being studied, since PNGaseF will not cleave oligosaccharides with core fucose residues in an al-3 linkage. Protein components can also be removed by precipitation with ice-cold ethanol Bio-Gel P2 has an exclusion limit of approx 1 8 kDa. If a 2-mL column has an exclusion volume of about 600 pL, then elution of the column with 800 pL of water should elute all N-linked ohgosaccharides. If in doubt, fractions can be assayed for hexose using the phenol-sulfuric acid method (Chapter 1). The reaction should be mcubated either m an oven or a heating block but not m a water bath. Owing to the highly toxic and flammable nature of hydrazme, all manipulations that involve its use should be carried out m a fume cupboard with both additional skin and eye protection (and see Chapter 8, Subheading 3.4.2.) If the column is not washed with methanol, the final eluant will consist of an immiscible butanol/water mixture, which will prove difficult to dry. The amount of glycoprotein reqmred for hydrolysis will depend on the sensmvity of the electrochemical detector. A first-generation detector, such as the Dionex PAD-2 or PED-2, ~111 require approx 5 ug of glycoprotein (assummg 15% glycosylation) for monosaccharide analysis and 50 pg for sialic acid analysis. A second-generation electrochemical detector, such as the Dtonex ED-40, will require approx 1O-fold less material. Monosaccharide hydrolysis has frequently been performed using 2M TFA, although the use of TFA can result m the epimerization of mannose to glucose and thus give spurious results. Longer incubations with 2M HCl or 4M HCl will give slightly more accurate estimations of ammo sugars (GlcNHz and GalNH& but with decreased responses for hexoses (26)

96

Davies and Hounsell

13. A CarboPac PA1 0 column (Dionex) 1salso available for the analysts of monosaccharides. Thts IS a solvent-compattble column (the PA1 is intolerant of organic buffers) and shows better tolerance of dissolved O2 in buffers Dissolved O2 m the buffers can be reduced at gold electrodes to form H20, whtch can then be further reduced to H20, which generates a characteristic dip m the baseline of electrochemical detectors (27). The use of a borate trap guard column reduces the peak quenching effects caused by borate either m the buffers or Introduced durmg sample preparation (28) The ammo trap column will chelate any ammo acids released durmg hydrolysis, ensurmg they are eluted after all monosaccharides durmg the column regeneration step. Lysme is known to elute very close to GalNH, and also to have a quenching effect on late-running monosacchartdes, such as mannose (28). As far as posstble, plastic reagent bottles should be used, and reagents contammg sodmm acetate should be filtered before use 14 These are generalized pulse potentials set up for a PED-2 detector To obtam optimal detection efficiency, they may have to be adjusted to suit the detector being used The gold electrode should be cleaned approxtmately every 3-4 wk or when sensittvtty drops Cleanmg should be done by rubbing the electrode surface with a pencil eraser followed by washmg with large amounts of distilled water Care should be taken to avoid air bubbles m the reference cavity of the electrode, since this will result in drtftmg baselines The reference electrode should be replaced tf tt is obvtously “plugged” (very seriously dtscolored), and baseline stability cannot be obtained by polishing the working electrode 15. HPAEC-PAD analysis of stalic acids will only determine the presence of Neu5Ac or NeuSGc. To determine the presence of other O-acetylated variants, released sialtc acids can be fluorescently labeled with 1,2-diammo-4,5-methylenedtoxybenzene (DMB) and separated on an RP HPLC column (29) A commercial kit for this 1savailable from Oxford GlycoSctences 16. Separation of neutral N-linked oligosacchartdes may be achieved using 250 mM NaOH (30) or slower acetate addition Oltgosacchartde ltbrartes that contain only a2-3 linked stalic acids (e.g., those isolated from protems expressed m CHO cells) can be separated at higher resolutton at pH 5 0 (31). 17. Fluorescence detection can be used m place of PAD for the detection of charged fluorescently labeled ohgosacchartdes, but not for neutral labeled ohgosaccharides, since the loss of the reducing anomeric carbon durmg the derivatizatton process will significantly reduce retention. 18. In addttton to mtroducmg 2-AB groups, reducttve aminatton can also be used to introduce other fluorescent labels, such as 2-ammopyndme (Chapter 7) or ANTS (Chapter 8) Five microliters of the 2-AB-labelmg reagent described are sufticlent to label up to 50 nmol of ohgosaccharide. 19 If poor solubthty of the reductant (NaCNBHs) 1sobserved, this can be improved by the addition of 10 $ of water to the labeling mtxture prior to addmg tt to the samples 20 Care should be taken that the flow rate through the disk is approx 1 drop/s and that air bubbles do not form below the disk. An bubbles can be removed by gentle pressure on the disk, though it 1svery difficult to remove them completely

HPLC and HPAEC

97

21 If the samples are to be analyzed by HPLC using a UV or a filter-based fluorescence detector, then the samples are now ready for analysis However, if the samples are to be analyzed by a monochromator-based detector or by Bio-Gel P4 chromatography, then further cleanup should be performed. UV detection of 2-AB-labeled glycans can be carried out at 2.54 nm 22 Desalting may also be carried out by Bio-Gel P2 chromatography 23 To ensure high-purity eluants are always obtained, it is recommended that ammomum acetate is obtained by titrating the relevant acid (e.g., 0 5M HPLCgrade acetic acid) to the relevant pH with HPLC-grade ammomum hydroxide Ammomum formate may also be used as an eluant, at similar pH values 24 Iqectmg the sample m water will result m the sample elutmg m the void volume In the uuttal part of the gradient, the column is operating as a hydrophdltc mteraction column and will thus separate neutral ohgosaccharide isomers As the ammonmm acetate gradient Increases, the column will function as an amon exchanger, separating ohgosaccharides on the basis of charged groups 25 The precise elution gradient can be varied to suit the diversity of oligosaccharides being studied The pH of the ammonium acetate will greatly affect the resolution and retention, and can be tailored to suit the analytes being investigated (12). 26 This flow rate IS for a 100 x 4 6 mm Hypercarb S column. The smaller Glycosep HTM column should not be run at flow rates in excess of 0 5 mL/mm 27. Great care should be taken to ensure thorough equilibration of the column prior to mlectmg samples, since these columns are very sensitive to changes m orgamc phase composition 28. PGC columns can also be run with 0 1% TFA as the mobile phase modifier However, this will cause quenching of fluorescence detection, resultmg in a loss of sensitivity 29 If chromatography on GlycoSep CTM or PGC columns fails to produce pure ohgosacchartde isomers, then either RP or NP columns as described m Chapter 7 can be used to purify the oligosacchandes further 30. Sodium borodeutende solutions should be made up ~4 h before they are required 31 The NaOH/DMSO suspension will deprotonate all the free hydroxyl groups and acetamido NH groups, allowmg the methyl iodide to react with the unstable carbamons to form a permethylated (O-Me) oligosaccharide. Fragmentation of these oligosaccharides by FABMS with LSIMS will generate oligosaccharide-specific fragmentation patterns This allows oligosaccharide sequence mformation to be derived. However, the nature of the origmal glycostdic lmkage IS not determmed at this stage A tetrasaccharrde Hex-HexNAc-HexNAc-Hex will generate the fragments: Hex-HexNAc and Hex-HexNAcHexNAc The further hydrolysis of the permethylated oligosaccharides will produce monosaccharides with free hydroxyls at the position of the original glycosidic linkage. Reduction of the monosaccharides with NaBD, will reduce the monosaccharides to their alditols with the anomeric carbon bemg monodeuterated The acetylatton of the remammg free hydroxyls will generate volatile species for analysis by GC-MS The

98

Davies and Hounsell retention times of the partially methylated aldttol acetates will Identify the monosaccharide type (e g , galactose, N-acetylgalactosamine, N-acetylglucosamme), whereas the fragmentation pattern in the mass spectrometer will identify the ongmal linkage substitutions The fragment ions are generated by fragmentation between C--C bonds, depending on the substituants with methoxymethoxy fragments being more common than methoxy-acetoxy fragments, which m turn are more common than acetoxy-acetoxy fragments Thus, a 3-linked galactose will have a different characteristic set of tons from a 4-linked galactose

References 1 Stoll, M. S , Hounsell, E. F , Lawson, A. M , Chal, W , and Felzl, T (1990) Microscale sequencing of O-linked ohgosaccharldes using mild pertodate oxidation of aldnols, couplmg to phosphohpid and TLC-MS analysis of the resultmg neoglycohptds Eur J Blochem 189,499-507 2 Umemoto, J , Bhavanandan, V P, and Davidson, E. A (1977) Punficatlon and properties of an endo-a-N-acetylgalactosamimdase from Dlplococcuspneumonlae J Bzol Chem 252,860%8614. 3 Fan, J-Q , Kadowaki, S., Yamamoto, X., Kumagal, H., and Tochtkura, T (1988) Purificatton and charactertsatton of endo-a-N-acetylgalactosammtdase from Alcallgens sp. Agrzc Blol Chem 52, 1715-1723 4 Kamerlmg, J P, Gerwig, G J , Vliegenthart, J F G , and Clamp, J R (1975) Characterization by gas-hqutd chromatography-mass spectrometry and proton-magnetic-resonance spectroscopy of pertrimethylsilyl methyl glycostdes obtained m the methanolysis of glycoprotems and glycopepttdes. Bzochem J 151,491-495

5. Davies, M J , Smith, K D , Cart-tithers, R A , Chai, W , Lawson, A M., and Hounsell, E F (1992) Use of a porous graphmsed carbon column for the high performance hquld chromatography of ohgosacchartdes, aldnols and glycopeptides with subsequent mass spectrometry analysis. J Chromatgr 609, 125-13 1 6 Fan, J.-Q., Kondo, A.,Kato, I , and Lee,Y.-C (1994) High-performance liquid chromatography of glycopeptldes and ohgosacchartdes on graphtttzed carbon columns Anal. Biochem 219,224-229. 7 Townsend, R R (1995) Analysis of glycoconJugates using high-pH anion-exchange chromatography J Chromatogr Library vol 58: Carbohydrate Analysis (El Rasst, Z , ed ), Elsevier, The Netherlands, pp 18 l-209 8 Reddy, G P. and Bush, C A (1991) High-performance anion exchange-chromatography of neutral milk ohgosacchartdes and ohgosacchartde aldttols derived from mucm glycoproteins. Anal. Biochem. 198,278-284. 9 Shibata, S., Mtdura, R J , and Hascall, V. C (1992) Structural analysts of the lmkage region ohgosacchartdes and unsaturated dtsacchartdes from chondrottm sulfate using CarboPac PA1 J Blol Chem 267,65484555 10. Davies, M. J., Smith, K D , Harbm, A -M., and Hounsell, E. F. (1992) Htgh-performance liquid chromatography of oligosaccharide alditols and glycopeptides on a graphmzed carbon column J Chromatogr. A 609, 125-l 3 1.

HPLC and HPAEC

99

11 Davies, M. J. and Hounsell, E. F. (1996) Compansion of separation modes of highperformance liquid chromatography for the analysis of glycoprotem- and proteoglycan-derived oligosaccharides. J Chromatogr A 720,227-233. 12 Guile, G R , Wong, S Y C , and Dwek, R. A. (1994) Analytical and preparative separation of anionic oligosaccharides by weak anion-exchange high performance liquid chromatography on an inert polymer column Anal Blochem. 222,23 I-235.

13 Takahashi, N (1996) Three-dimensional mapping of N-linked oligosaccharides using anion-exchange, hydrophobic and hydrophilic mteraction modes of htghperformance liquid chromatography. J. Chromatogr A 720,2 17-225 14 Yamashita, K., Mizouchi, T., and Kobata, A (1982) Analysis of ohgosaccharides by gel filtration. Methods Enzymol 83,625-63 1. 15 Settinen, C. A and Burlingame, A L. (1995) Mass spectrometry of carbohydrates and glycoconJugates J Chromatogr: Library vol 58: Carbohydrate Analyw (El Rassi, Z , ed.), Elsevier, The Netherlands, pp. 447-5 14 16. Guile, G R., Rudd, I? M., Wing, D. R., Prime, S. B., and Dwek R. A. (1996) A rapid high-resolution high-performance liquid chromatographtc method for separating glycan mixtures and analyzing ohgosaccharide profiles. Anal Blochem. 240,210-226 17. Smith, K. D., Davies, M. J., Bailey, D., Renouf, D. V., and Hounsell, E. F. (1996) Analysis of the glycosylation patterns of the extracellular domain of the epidermal growth factor receptor expressed in Chinese hamster ovary fibroblasts. Growth Factors 13, 121-132 18 Patel, T., Bruce, J., Merry, A., Bigge, C., Wormald, M., Jaques, A., and Parekh, R. (1993) Use of hydrazine to release m intact and unreduced form both N- and Olinked ohgosaccharides from glycoprotems Biochemistry 32,679-693 19 Bigge, J C , Patel, T. P , Bruce, J. A., Goulding, P. N., Charles, S. M , and Parekh, R B. (1995) Nonselective and efficient fluorescent labelmg of glycans using 2-amino benzamide and anthranihc acid. Anal Bzochem 230,229-238 20. Huddleston, M. J., Bean, M. F , and Can; S. A. (1993) Collisional fragmentation of glycopeptides by electrospray iomzation LC/MS and LC/MS/MS: methods for selective detectton of glycopeptides m protein digests. Anal Chem 65, 877-884. 2 1. Guzzetta, AW., Basa, L J., Hancock, W. S., Keyt, B. A., and Bennet, V Y. T. (1993) Identification of carbohydrate structures in glycoprotein peptide maps by the use of LC/MS with selected ion extraction with special reference to tissue plasmmogen activator and a glycosylation variant produced by site directed mutagenesis. Anal. Chem 65,2953-2962.

22. Carr, S. A., Huddleston, M. J., and Bean, M. F (1993) Selective identification and differentiation ofN- and O-linked oligosaccharides m glycoprotems by liquid chromatography-mass spectrometry. Protein Scl. 2, 183-196 23. Rhemhold, V N., Rhemhold, B. B and Costello, C. E. (1995) Carbohydrate molecular weight profilmg, sequence, linkage, and branchmg data. ES-MS and CID. Anal Chem. 67,1772-1784

24. Chapman, J. R. (ed.) (1996) Protein and Peptzde Analyszs by Mass Spectrometry, Humana Press, Totowa, NJ

100

Davies and Hounsell

25 Cmcanu, I and Kerek, F. (1984) A sample and rapid method for the permethylatton of carbohydrate Carbohydr Res 131,209-217. 26 Fan, J.-Q., Namtkt, Y, Matsuoka, K , and Lee, Y. C (1994) Compartson of actd hydrolytic condmons for Am-lmked ohgosaccharides. Anal Bzochem. 219,375-378 27. Wertzhandler, M , Slmgsby, R , Jagodzmskt, J , Pohl, C , Narayanan, L , and Avdalovic, N. (1996) Ehmmatmg ammo actd and pepttde interference m hrgh performance amon exchange pulsed amperometrtc detection glycoprotem monosaccharide analysts Anal Blochem 241, 128-136 28 Rocklm, R. D., T&en, T R , and Marucco, M. G. (1994) Maxrmtzmg stgnal-tonoise ratio in direct current and pulsed amperometnc detection J Chromatogr A 671,109-l 14 29 Hara, S.,Yamaguchl, M., Takemori,Y , Furuhata, K , Ogura, H , and Nakamura, M (1989) Determmatton of mono-O-acetylated N-acetylneurammtc acrds m human and rat sera by fluorometrlc high-performance hqutd chromatography. Anal Bzochem 179, 162-l 66. 30. Cooper, G. and Rohrer, J S (1995) Separation of neutral asparagme-linked oligosacchartdes by htgh-pH anton-exchange chromatography with pulsed amperometnc detection. Anal Blochem 226, 182-l 84 3 1, Watson, E , Bhide, A , Kenney, W C , and Lm, F -K (1992) High-performance amon exchange chromatography of asparagme-lurked ollgosacchartdes. Anal Bzochem 205,90-95.

7 Analysis

of I’S and O-Glycans by Pyridylamination

Shunji Natsuka and Sumihiro

Hase

1. Introduction

The pyridylaminatlon method was originally described in 1978 as a means of analyzing glycan structures with high-sensitivity (I). Subsequently, the method has been applied to structure analyses of glycans including glycosldase digestion (Z), 2D-mapping by various kmds of high-performance llquld chromatography (HPLC) (3), partial acetolysis (4), Smith degradatton (51, methylatlon analysis (I), nuclear magnetic resonance (6), mass spectrometry (3, and lectin-affinity chromatography (8). Glycans on glycoconjugates are liberated by hydrazinolysls followed by N-acetylatlon (9), glycopeptidase digestion (I&II), or endoglycoceramldase digestion (12,13) Hydrazmolysls 1s ordinarily used to liberate N- and/or 0-glycans from glycoprotems using the condltlons described in Subheading 3.1. Glycans in the reaction mixture can be directly pyridylaminated wlthout any purification of the liberated glycans. Reducing ends of the glycans are tagged with 2-ammopyndme by reductive ammatlon (Fig. 1). Since the fluorescence intensities of pyrldylammo (PA)-derivatives from N-glycosldes are almost the same (Table 1; Hase, S., et al., unpublished data), their peak-area ratios are considered as their molar ratios. PA derivatives of glycans with fluorescence and a positive charge have the followmg advantages: 1. Detection sensitivity IShigh; 0.02 pmol of a PA-glycancanbe detectedwith commercially available HPLC apparatus. 2 Excellent separationIS achievedby reversed-phaseHPLC 3. The PA group 1schemically stableunder the condltlons for structureelucldatlon 4. A posltlve charge 1suseful for separationof PA-glycansby Ion-exchangechromatography or electrophoresls From

Methods m Molecular Wology, Vol 76 Glycoanalysa Protocols Edrted by E F Hounsell 0 Humana Press Inc , Totowa, NJ

101

Natsuka and Hase

102 H,OH 2.Aminopyridine H, OH

HO H

+.-. AcOH

NHAc

CH=N

RO

i

NHAc

NHAc

Fig. 1. Scheme of the pyridylamination

Table 1 Fluorescence Intensities Linked Sugar Chain9

reaction.

of PA-/V-

PA-GlcNAc2 PA-Xylomannose

1.05 0.99

M6B

1.oo

PA-Biantennary PA-Biantennary-NeuAc2 PA-Triantennary PA-Tetraantennary

1.07 0.95 0.95 0.95

OFigures indicate peak areas per mole when PA-sugar chains were analyzed as in Subheading 3.3.3. M6B (Fig. 8B) is taken as unity. Amounts of PA-sugar chains were determined by gas-liquid chromatography after methanolysis.

PA-glycans principles:

are purified

by three kinds of HPLC with different

separation

1. Anion-exchange HPLC. 2. Size-fractionation HPLC. 3. Reversed phase HPLC.

Furthermore, the additivity rule, which correlates elution times in reversedphase chromatography with chemical PA-glycans, as described in Note 1.

structures,

supports

the analyses of

2. Materials

2.1. Liberation

of N- and 0-Glycans

from Glycoproteins

1. Anhydrous hydrazine (Pierce, Rockford, IL) (see Note 2). 2. Screw-cap test tube (13 x 100 mm) with a Teflon seal. 3. Dowex 5OW-X2 (20@400) cation exchanger (Bio-Rad Laboratories, Hercules, CA). The resin is activated, and made up to H+ form according to the manufacturer’s protocol.

fyridylamination

103

2.2. Pyridylamination

fur Glycan Labeling

(see Note 3)

1. A glass test tube (10 x 100 mm) tapered at the bottom or a Reacti-Vial (1 mL). 2. 2-Aminopyridine: Colorless leaflet crystals are recrystallized from the commercial reagent (pale yellow) using hexane, and stored in a desiccator. Caution: This compound is toxic, and an irritant to the skin, eyes and mucosa. Avoid inhalation. 3. Coupling reagent: Dissolve 552 mg of 2-aminopyridine in 200 pL of acetic acid. (When the reagent is diluted with 9 vol of water, the pH of the solution should be 6.8.) The reagent should be stored at below -20°C in a tube sealed with Paratilm. 4. Reducing reagent: Prepare just before use by dissolving 200 mg of boranedimethylamine complex ([CH&NH.BHs: Aldrich, Milwaukee, WI) in a mixture of 50 pL of water and 80 pL of acetic acid. Caution: This compound is corrosive to the eyes, skin and mucosa. Avoid inhalation. Store desiccated at -15°C or below (flash point 43°C). 5. A TSK-gel HW-40F column (1.6 x 40 cm, Tosohaas, Philadelphia, PA) washed well with O.OlMammonium acetate, pH 6.0 (O.OlM acetic acid is titrated to pH 6.0 with 1.5M aqueous ammonia). Sephadex G-15, Sephadex G-25 or Bio-Gel P-2 can also be used equally well.

2.3. Separation

of PA-Glycans by HPLC

1. A Mono-Q HR 5/5 column (0.5 x 5.0 cm, Pharmacia, Uppsala, Sweden). A Tosoh TSK-gel DEAE-SPW column (0.75 x 7.5 cm) can also be used: solvent A: 0.7 mM aqueous ammonia (pH 9.0); solvent B: 0.5Mammonium acetate (the pH of 0.5M acetic acid is adjusted to 9.0 with 41M aqueous ammonia). 2. Shodex Asahipak NH2P-50 (0.46 x 5 cm, Keystone Scientific, Bellefonte, PA): solvent A: acetic acid-acetonitrile-water (3:930:70) titrated to pH 7.0 with aqueous ammonia; solvent B: acetic acid-acetonitrile-water (3:200:800 [v/v/v]) titrated to pH 7.0 with aqueous ammonia. The conditions are a modification of those reported (24). MicroPak AX-5 (Varian Aerograph, Walnut Creek, CA), Tosoh TSK-gel Amido-80 or YMC PA-03 (YMC, Morris Plains, NJ) are also usable provided a precolumn (silica gel, 0.75 x 7.5 cm) is placed between the injector and pump to prevent damage to the separation column. 3. A Nacalai Cosmosil5C 18-P column (0.46 x 15 cm, JM Science, Buffalo, NY) or other C,s reversed-phase column. For N-glycans (Subheading 3.3.3.): solvent A: O.lMammonium acetate, pH 4.0 (O.lM acetic acid is titrated to pH 4.0 with 4M aqueous ammonia); solvent B: 0. 1M ammonium acetate, pH 4.0, with 0.5% 1-butanol. For O-glycans (Subheading 3.3.4.): solvent A: O.lMammonium acetate, pH 6.0 (O.lM acetic acid is titrated to pH 6.0 with 4M aqueous ammonia); solvent B: O.lM ammonium acetate, pH 6.0, with 1.O% 1-butanol. For sialyloligosaccharides (Subheading 3.3.5.). Solvent A: O.lM acetic acid. Solvent B: 0. IA4 acetic acid with 0.5% 1-butanol. 4. HPLC apparatus and a fluorescence spectrophotometer equipped with a l-cm cuvet, flow cell (8-16 pL), 150-W xenon lamp, and two monochrometers.

Natsuka and Hase

104

5 A water bath (90 and SO’C), small centrifuge, and lyophilizer 6. PA-glycans are available from several suppliers (PanVera, Madison, WI; Seikagaku America, Rockvtlle, MD, Wako Chemicals USA, Richmond, VA).

3. Methods 3.7. Liberation

of N- and 0-Glycans

from Glycoproteins

(15)

1. Lyophilize glycoprotein(s) (~2 mg) m a screw-cap test tube 2. Add anhydrous hydrazme (0.2-0.3 mL) and seal the tube. Caution: Anhydrous hydrazme 1s a strong base, highly toxic, flammable, and corrosive 3 Heat for 10 h m boiling water for N-glycan liberation, or for 50 h at 60°C to release both N- and 0-glycans. 4 Remove hydrazme by repeated coevaporatton (three to five times) with toluene zn vucuo using a trap (-50°C or below) 5. Add 200 p.L of saturated NaHCOs aqueous solutton (freshly prepared) and 8 pL of acetic anhydride for re-hr-acetylation 6. Stand for 5 mm 7 Add the same soluttons as in step 5. 8 Stand for 30 mm 9 Add Dowex 5OW-X2 (H+) up to a pH of 3 0 10. Pour the suspension mto a small glass column (0 5 x 10 cm) and take the effluent. 11 Wash the resin with five column volumes of water. 12. Combine the washings and the effluent. 13 Concentrate the solution using a rotary evaporator or a concentrator without operating the heater.

3.2. Pyridylamination

for Glycan Labeling (15)

1 Lyophiltze the sample prepared as in Subheading 3.1. or 0 05-50 nmol of a glycan(s) m a glass test tube tapered at the bottom or m a Reactt-Vial. 2 Add 20 & of coupling reagent to the residue (the reagent 1swarmed just before use) and mix well. Seal the tube, and spm down the reaction mixture to the bottom. 3. Heat the mixture at 90°C for 60 mm to form a Schiff base. (Care should be taken to heat the whole tube.) Cool the tube to room temperature 4 Add 70 pL of reducing reagent and mix well 5 Reseal the tube, spin down, and heat at 80°C for 35 mm. 6. Remove most of the excess reagents by either of the followmg methods. Caution: Hydrogen 1s released during the reactton, therefore do not heat the sealed glass tubes to open them a. Method A: Add 300 pL of 25% aqueous ammonia to the reaction mixture and mix well. Then add 300 pL of water-saturated chloroform and vortex vigorously Almost all the excess 2-aminopyridrne 1s moved into the organic layer under the alkaline condition. Separate the two layers by

105

Pyrdylamina tion

Eluuon volume

(ml)

Fig. 2. Typical elution profile of a pyridylamination reactton mixture by gel filtration using a TSK-gel HW-40F column (1.3 x 20 cm). Arrows A and B indicate the elutron positions of PA-mannose and 2-ammopyrtdme, respectively. The horizontal bar shows the elution position of PA-oligosaccharides.

centrifugation and repeat chloroform-extraction twice against a water layer. Remove the ammonia from the water layer by a concentrator without operatmg the heater b. Method B* Add 40 & of a mixture of methanol and triethylamme (3.1 [v/v]) to the reaction mixture and mix well. Then add 40 pL of toluene Remove excess reagent by flushing with nitrogen (300 mL/mm) at 50°C for 12 mm under a vacuum at about 150 mmHg (see Note 3). Repeat the evaporatton three to five times with 60 pL of toluene/methanol (2 l), but use 50 pL of toluene for the final repetmon. 7 To remove the minute amounts of reagents still remaining, carry out gel filtration on a column of TSK-gel HW-40F (Fig. 2) with O.OlM ammomum acetate, pH 6 0. Collect the PA-glycan fraction, and lyophihze to remove the ammonmm acetate

3.3. Separation of PA-Glycans by HPLC 3.3.1. Anion-Exchange Chromatography (16) 1 Before injection of approx 10 samples, wash Mono-Q HR 5/5 with 6% acetic acid for 10 mm and then 0.4M aqueous ammonia for 10 mm at a flow rate of 1 mL/mm. 2 Equilibrate the column with solvent A. 3. Inject a sample made up to pH 9.0, and carry out gradient elution as shown in Fig. 3A at a flow rate of 1.O mL/mm and a column temperature of 25°C 4 Detect PA-glycans using an excitation wavelength of 310 nm and an emissron wavelength of 380 nm Under these conditions, PA-glycans are separated according to the number of negative charges, such as sialic acid residues (Fig. 3B)

106

Natsuka

I

i

212 ;7 Time (mm)

and Hase

5’0

AI I

I 0

5 Elutlon

A2

10 time

(mln)

Fig 3 (A) Gradient pattern of anion-exchange chromatography (B) Elutron profile of PA-glycans from fetum on a Mono-Q HR 515 column As-As indicate asialo, monoslalo, disialo, and trisialo-PA-glycans, respectively 3.3.2. Size-fractmation

HPLC

(2)

1 Wash Asahtpak NH2P-50 with methanol, and equihbrate the column with solvent A 2. Inject a sample (
107

Pyndylamination

0

, , 03

I 3s Time

I 0

5

(mm)

IS

IO Plutlon

I Sj

time

20

25

(min)

Ftg 4 (A) Gradient pattern of size-fracttonatton HPLC. (B) Elutton profile of PA-tsomaltoollgosacchartdes on an Asahtpak NH2P-50 column Numbered arrows indicate PA-tsomalto-ohgosacchartdes wrth correspondmg degree of polymertzation. N-acetylhexosamme residue, 0 6 glucose unit, “btsectmg” N-acetylglucosamme residue, 0 2 glucose unn fucosylal-6 residue, 0 5 glucose unit The values are almost independent of their linkage pomts and anomertc conflgurattons.

3.3.3. Reversed-Phase

HPLC for N-Glycans (2) (see Note 7)

1 Wash a Cosmos11 5C 18-P column wtth methanol, and equtllbrate tt wtth the startmg eluent (buffer A-B = 95.5 [v/v]). 2. Inject a sample, and carry out gradient elution as shown m Fig. 5 using solvent A* O.lM ammomum acetate, pH 4 0, and solvent B: 0.M ammonmm acetate, pH 4.0, with 0 5% I-butanol at a flow rate of 1 5 mL/mm and a column temperature of 25°C The elutton posmons of twelve kinds of high-mannose type glycans are shown m Fig. 6 (3) 3, Detect PA-glycans with an excitation wavelength of 3 15 nm and a fluorescence wavelength of 400 nm.

3.3.4. Reversed-Phase

HPLC for 0-Glycans

1 Wash a Cosmosil5C 18-P column wtth methanol, and equilibrate it with the starting eluent (buffer A*B = 99 1 [v/v])

Natsuka and Hase

108

Time (mm)

Fig 5 Gradtent pattern of reversed-phase HPLC for N-glycans

. M9A .

.

VaA Ma8 . M’A

l

Mk .

.

M7B M’C

M’D

. . MGA MGB 8 3 4

ML .

6

0

M5A I 10

I 05

Eluuon rime relative

to M5Aon

reversed

phase HPLC

Frg. 6. Two-dimenstonal sugar map of PA-ohgomannose-type glycans The abscissa shows the elution time relative to M5A (23 min) on a Cosmosil 5C18-P reversedphase column, and the ordinate indicates the molecular size of PA-glycans on an Asahipak NH2P column by size-fractionation HPLC. Structures of the PA-ollgomannose-type glycans show in Fig. 8A. 2 Inject a sample, and carry out gradient elution as shown in Fig. 7A using solvent A: O.lM ammonium acetate, pH 6.0, and solvent B* O.lM ammomum acetate, pH 6 0, with 1% 1-butanol at a flow rate of 1 5 mL/min and a column temperature of 25°C 3 Detect PA-glycans as described in Subheading 3.3.3., step 3. A PA-glycans separation pattern is shown in Fig. 7B

Pyridylamination

109

Time (mm)

I 0

10 llurlon

20 time

30

< 40

(mln)

Fig 7. (A) Gradient pattern of reversed phase HPLC for U-glycans. (B) Elution profile of PA-glycans on a Cosmostl 5C18-P column under the condttions for O-glycans described m Subheading 3.3.4. Peaks a-f indicate GalP l-3GalNAc-PA, GalB l6GalNAc-PA, Gall31 -4GalP l-3GalNAc-PA, GlcNAcP 1-4GlcNAc-PA, GalNAccL l-6 GalNAc-PA, and GalNAca I-~(FUCCXI-2)GalP I-3GalNAc-PA, respecttvely.

3.3.5. Reversed-Phase

HPLC for PA-Sialyloligosaccharides

1 Prepare a Cosmostl SC1 8-P column as described m Subheading 3.3.3., step 1. 2. Inject a sample, and carry out gradient elutton as shown m Fig. 5 usmg solvent A: 0 1M acetic acid, and solvent B: O.IMacettc acid with 0.5% I-butanol at a flow rate of 1.5 mL/mm and a column temperature of 25’C 3. Detect PA-glycans as descrtbed in Subheading 3.3.3., step 3. Separation of PA-stalylbtantennary glycans 1s described in ref. I7 This HPLC is suitable for purtflcation of PA-stalyloltgosaccharides for better separation from PA-asralyloligosacchartdes, though the procedure given m Subheading 3.3.3. can also be used

A r‘----‘--‘-----------------------------------------~ D3 (-0.22)Manal~2Mar1al~~

3ManaL, I, :Mar$l-4GlcNAcPl-4GlcNAc-PA D2(0.16) Manalr2Manal' -------- I Dl (-0.06)Manal-2Manal~2Manal/ L-----------------------------------------~

I1 I II : 6 I

M5A (1 00)

(-0: 6)

B M9A

Man~2Mana&, MdWdQ ManUZMana~ Mana2Mana2Mana~

*

‘MWQ

Man~4GlcNAc~4GlcNAc-PA

h47C

Mana2Mad

Man~4GlcNAc~4GlcNAc-PA M?!“Cd

ManUZManab MBA

‘Mana\ Md”R Man~4GlcNAcp4GlcNAc-PA ManUZMan122Mana I’

3/

M7D

ManRZManaG,

MBC

M6.A

~MdflUQ

MdW3.

3/

f

Manq Mancl2Mana3/ Man~4ClcNAc~4ClcNAe-PA ManaZManaZMana f

‘M.3”CZ< MB”= Manp4GlcNAcp4GlcNAc-PA ManRZMana f

3/

MB

mnac.

‘MXIUQ Ma”a3’ Man~4GlCNAC~4GlcNAc-PA blana2mna Y

‘ManaQ M=

Manp4GlcNAcp4GlcNAc-PA

ManaZMand

nana3’

Mana\

Yanaq raq

M7B

Man~4GlcNAc~4GlcNAc-PA

Ma”a3’

ManR2Mana6,

M?A

Man~4GlCNAc~4GiCNAc-PA

ManRZManCG,

‘Ma”UQ MBB ManU2Mana 3f Man~4GlcNAc~4GlcNAc-PA Mana2Mana

‘M,“U< ManaZMancd ManaZManaf

Mad ManRZManaZMana

M”““~ M5A

Manp4GlCNAcp4GlcNAc-PA f

Ma”= 3’

Manp4GlcNAcp4GlcNAc-PA Yd”CX

J

Fig. 8. (A) Structures and notations of the PA-ollgomannose-type glycans (B) C and D l-3 refer to the corresponding mannose residues m an ohgomannose-type glycan. The values m parentheses are the partial relatwe elutlon times (see Note 1) of the correspondmg mannose residues

110

4. Notes I Aditivtty rule for reversed-phase HPLC ($18) A sugar residue m a PA-glycan makes an mtrmsic contrtbution to the (relattve) elutton time of the PA-glycan This contribution is referred to as the “partial relative elution time, Et (18) To compensate for devtattons as a result of column aging, the composmon of the eluents, and so forth, relative elution times are used (M5A 1s taken as unity m Fig. 6). The addittvny rule IS expressed as follows n E=EO+.ZEl I = 1

where E is the relative elution time of the PA-glycan m question, E, is the relative elution time of a base PA-glycan, Ei is the partial relative elutton time of sugar residue I, and n is the number of sugar residues required to construct the PA-glycan in question from the base PA-glycan A partial relative elutton time for a D3-mannose (see Fig. 8A for nomenclature), for example, is calculated to be -0 22 by subtracting the relative elutton time of M5A (1 00) from that of M6A (0 78) (Fig. 6 and see Fig. 8B for structures and abbrevtattons) Four partial relative elutton times for mannose residues C, Dl, D2, and D3 (Fig. 8A) are calculated from the relative elutton times of at least five PA-glycans, for instance, M5A, M6A, M7A, M8A, and M9A. Relative elution times for seven other PA-glycans can be calculated using the four partial relative elution times, even tf PA-glycans are not available A relative elution time for M6B, for example, 1s calculated to be 0.84 by adding the relative elutton time of M5A (1 .OO) and the partial relative elubon time of C-mannose (-0 16) The same rule is applicable for the sugar residue m PA-derivatives of the oligomannose, N-acetyllactosamine (complex type) and xylomannose types (19), and of sialylobgosacchandes (17) Among 70 PA-glycans tested, the errors m culculated elution times were less than a few percent, Since the partial relative elutton times are dependent on the column, column temperature, HPLC apparatus, and method of gradient elutton, these condittons must be kept constant 2. More detatls about anhydrous hydrazme can be found in Chapter 8 and the chapter by T Mtzuochi m the previous edition (20). Commercial kits are also avatlable for hydrazinolysts (see Chapter 4) 3. An apparatus for pyrtdylammation, GlycoTAG, 1s available from Takara Shuzo, Kyoto, Japan

References 1 Hase, S , Ikenaka, T , and Matsushima,Y. (1978) Structure analyses of oligosaccharides by tagging of the reducing end sugars with a fluorescent compound Blochem Blophys Res Commun 85,257-263.

2 Hase, S , Koyama, S , Datyasu, H , Takemoto, H , Hara, S , Kobayashi,Y, Kyogoku, Y , and Ikenaka, T. (1986) Structure of a sugar chain of a protease inhibitor isolated from Barbados pride seeds J Blochem 100, l-10

112

Natsuka and Hase

3 Hase, S , Natsuka, S Oku, H , and Ikenaka, T. (1987) Identification

method for twelve oligomannose-type sugar chains thought to be processmg mtermediates of glycoproteins. Anal Blochem 167,321-326 4 Natsuka, S , Hase, S , and Ikenaka, T (1987) Fluorescence method for the structural analysis of oligomannose-type sugar chams by partial acetolysis. Anal Bzochem 167,154-159 5 Hase, S., Ktkuchi, N., Ikenaka, T., and Inoue, K (1985) Structure of sugar chains of

the third component of human complement. J Blochem 98,863-874 6 Koyama, S , Daiyasu, H , Hase, S , Kobayashi, Y, Kyogoku, Y, and Ikenaka, T

7

8.

9.

10 11.

12.

13.

14.

(1986) ‘H-NMR analysts of the sugar structures of glycoprotems as their pyridylammo derivatives. FEBS Lett. 209,265-268 Gu, J., Hiraga, T., and Wada, Y. (1994) Electrospray iomzatron mass spectrometry of pyridylaminated olrgosaccharide derivatives. sensitivity and m-source fragmentation Biol Mass Spectrum 23,2 12-2 17. Mega, T., and Hase, S (199 1) Determmation of lectm-sugar bmdmg constants by mrcroequmbrium dialysis coupled with high performance liquid chromatography. J. Blochem 109,600-603. Yosizawa, Z., Sato, T , and Schmid, K (1966) Hydrazinolysis of a,-acid glycoprotein. Blochem Blophys Acta 121,417-420. Takahasht, N (1977) Demonstration of a new amidase acting on glycopeptides Bzochem Biophys Res Commun 76,1194-1201 Plummer, T H., Jr., Elder, J. H , Alexander, S., Phelan, A. W., and Tarentmo, A. L. (1984) Demonstration of peptide: N-glycostdase F activity in Endo-P-Nacetylglucosammtdase F preparatrons. J Blol Chem 259, 10,70&10,704. Ito, M. and Yamagata, T. (1986) A novel glycosphingolipid-degrading enzyme cleaves of the linkage between the ohgosaccharide and ceramtde of neutral and acidic glycosphmgohpids J Biol Chem 261, 14,278-14,282 LI, S. C., DeGasperi, R., Muldrey, J. E , and Li, Y. T (1986) A umque glycosphmgoliptd-splitting enzyme (ceramrde-glycanase from leech) cleaves the linkage between the ohgosaccharide and the ceramide. Blochem Bzophys Res Commun. 141,346-352 Melhs, S. J and Baenztgar, J. U. (1983) Size fractionatron of anionic oligosaccharides and glycopeptides by high-performance hqutd chromatography. Anal Blochem 134,442-449.

15 Kuraya, N. and Hase, S (1992) Release of 0-lmked sugar chains from glycoprotems with anhydrous hydrazme and pyridylammatton of the sugar chains with improved reaction conditions. J Biochem 112, 122-126 16. Yamamoto, S., Hase, S., Fukuda, S., Sano, O., and Ikenaka, T. (1989) Structures of the sugar chams of interferon-y produced by human myelomonocyte cell lme HBL38. J. Blochem 105,547-555. 17. Yamamoto, S., Hase, S., Yamauchi, H., Tanimoto, T., and Ikenaka, T. (1989) Studies on the sugar chains of interferon-y from human peripheral-blood lymphocytes J Bzochem. 105,1034-1039

Pyridylamina tion

113

18. Hase, S., and Ikenaka, T (1990) Estimation of elution times on reverse-phase HPLC of pyrrdylammo derivatives of sugar chams from glycoprotems Anal Bzochem. 184,135-138

19. Hase S , Ikenaka, K., Mikoshiba, K , and Ikenaka, T. (1988) Analysis of tissue glycoprotein sugar chains by two-dimensional high-performance hqmd chromatographic mappmg. J Chromatogr (Bromedical Aphcations) 434,s 140 20. Mizuochi, T. (1993) Mrcroscale sequencmg of N-linked oligosaccharides of glycoproteins using hydrazinolysis, Bio-Gel P-4, and sequential exoglycosidase digestion, m Methods in Molecular Biology, vol 14 * Glycoproteln Analysis In Blomedwne (Hounsell, E. F, ed.), Humana, Totowa, NJ, pp 55-68.

Polyacrylamide Gel Electrophoresis of Fluorophore-Labeled Carbohydrates from Glycoproteins John C. Klock and Christopher

M. Starr

1. Introduction Prior to 1980, most methods for analysis of glycoprotem carbohydrates uttltzed column, thin-layer, and paper chromatography, gas chromatography, mass spectroscopy,and rarely nuclear magnetic resonance spectroscopy.These methods required relatively large amounts of materials (mtcromoles), specrahzed training and experience, and in some cases,significant capital equipment outlays. Because of these restrtctrons, convenient carbohydrate analysts on small samples was not available to most biologists. Recently, improvements m chromatographtc methods, labelmg methods for carbohydrates, carbohydrate-specific enzymes, and higher resolution electrophoresis methods have allowed carbohydrate analysts to be done on nanomolar amounts of maternal Because of these improvements, today’s biologist now has an improved ability to evaluate the role of carbohydrates m their research and development work. Polyacrylamtde gel electrophoresis (PAGE) of fluorophore-labeled carbohydrates has also been referred to as fluorophore-asststed carbohydrate electrophorests, or FACE @.The technique was first developed m England by Williams and Jackson (I-4) and utilizes reductive ammatton of carbohydrates by low molecular weight, negatively charged or neutral fluorophores and electrophoresis on 2040% polyacrylamtde slab gels. Thusmethod permits separation of charged or uncharged sugars or oligosaccharides with high resolution and can detect single hydroxyl anomeric differences between mono- and oltgosaccharides of sugars with otherwise rdentrcal molecular weight, charge, and sequence. The separation of sugars by electrophorests IS largely empuical, and tt IS not always possible to predict relative mobrllttes of structures. For the From

Methods Edited by

m Molecular E F Hotmsell

Bfology, Vol 0 Humana

115

76 Glycoanalysfs Protocols Press Inc , Totowa, NJ

Klock and Starr

116

most part the successof this approach has been based on experimental observations and the use of highly specific reagents, enzymes, and standards. What follows m this chapter are descriptions of the materials and methods required to perform two of the most common manipulations of oligosaccharides used in biologic research today. The method with some modification IS useful for analysis of a variety of carbohydrates mcludmg reducing and nonreducmg sugars, substituted and unsubstituted monosaccharides, and ohgosaccharides from glycoproteins, proteoglycans, glycohpids, and polysaccharides. In this chapter only profiling of N-linked and O-linked glycoprotein ohgosaccharides is discussed. The method is also useful for a variety of other manipulations including structure-function studies, preparative work, and synthesis of oligosaccharides that are not discussed in detail here. Other reviews of this technique have been published previously (5,6) and a number of research studies have been performed using this method (7-21). 2. Materials

2.1 N-/inked Oligosaccharide Analysis For profilmg of Asn-linked ohgosaccharides released by peptide N-glycosidase F. 1 N-linked Gels: 10 x 10cm low fluorescenceglassplateswith 0.5 mtn spacersand

2 3. 4. 5 6 7 8.

S-well combs tilled with 20%T acrylamide:brs m 0 448MTrrs acetate buffer pH 7 0 and containing a 5 mm stack of 10% polyacrylamrde in the same buffer (gels should be made fresh or obtained precast from a commercial source) Releasing enzyme. peptrde N-glycosrdase F (from commercral sources) Running buffer: 50 mMTns trrcme buffer (pH 8.2). Enzyme buffer: 100 n-&I sodium phosphate buffer (pH 7.5). Labeling dye: 1M 8-aminonaphthalene-1,3,6-trisulfomc actd (ANTS) m 15% acetic acid (reagent is stable for 2 wk at -70°C in the dark) Reducing agent: 1M sodmm cyanoborohydride (NaBHsCN) m drmethyl sulfoxrde (DMSO) (reagent 1s stable for 2 wk at -7O’C) Sample loading solutron 25% glycerol with thorin 1 dye (store at 4“C). Tracking dye. Mrxture of thorm 1, bromphenol blue, direct red 75, and xylene cyanole in water Electrophoresis gel box (with 2-sided cooling). Deionized or drsttlled water

9 10 11. Assortedpipetmg devices including a O-10 pL positive displacementpipet (e.g., Hamilton syringe). 12. Centrifugal vacuum evaporator.

13. Oven or water bath at 45°C and 37°C. 14 Sodium dodecyl sulfate (SDS), 5% 15. P-Mercaptoethanol @ME). 16 7.5% nomdet P-40 (NP-40)

Polyacrylamide 17 18. 19. 20. 21

Gel Electrophoresis

117

100% Cold ethanol (undenatured) Microcentrifuge tubes (1 5 mL). Mlcrocentrlfige Chicken trypsin inhibitor control (optional). Maltotetraose or partially hydrolyzed starch standard (optlonal).

2.2. O-linked Glycoprotein Oligosaccharide Analysis For profiling of Ser/Thr-linked oligosaccharldes released by hydrazme. 1 O-linked gels: 10 x 10 cm low fluorescence glass plates with 0 5-mm spacers and eight-well combs filled with 35%T acrylamldesbls in 0.448MTt1s acetate buffer and contaming a 5-mm stack of 16% polyacrylamide m the same buffer (gels should be made fresh or obtained as precast gels from a commercial source) 2. Gel running buffer, 50 mM Tris-glycine (pH 8.2) 3. O-linked cleavage reagent: anhydrous hydrazme, 1 mL amp Hydrazine is toxic and flammable; discard ampoule and residual contents after using once; dispose of safely according to your institution’s regulations. 4 Re-N-acetylatlon reagent’ acetic anhydride 5. Re-N-acetylatlon buffer: 0 2M ammomum carbonate, pH 9 4. 6 Desalting resin* Dowex AG50X8. 7. Tracking dye, Mixture of thorm 1, bromphenol blue, direct red 75, and xylene cyanole in water 8 Sample loading solution 25% glycerol with dn-ect red 75 (store at 4°C). 9 Labeling dye* 1M 8-ammonaphthalene-1,3,6-trisulfomc acid (ANTS) m 15% acetic acid (reagent 1sstable for 2 wk if stored m the dark at -70°C) 10 Reducing agent 1M sodium cyanoborohydrlde (NaBH$N) m DMSO (reagent IS stable for 2 wk at -70°C) 11. DIstilled-delomzed water 12 Microcentrifuge tubes (2 mL, glass-lined). 13. Assorted plpeting devices including a O-100 pL capillary positive displacement pipet 14. Centrifugal vacuum evaporator 15. Oven or water bath set at 37°C 16. Sand filled heat block set at 60°C. 17 Vacuum desslcator. 18. Microfuge 19 Phosphorus pentoxide (P205). 20. Bovine submaxillary mucin control (optional). 21 Maltotetraose or partially hydrolyzed starch standard (optional).

3. Methods 3.1. Principle Using fluorescent PAGE, individual oligosaccharides can be quantified to obtain molar ratios, to obtain degree of glycosylation, and to detect changes m

Klock and Starr

118 the extent or nature of glycosylation. four steps*

Oltgosaccharlde

profiling

involves

1 Release of the ohgosaccharides from the glycoprotem enzymatically or chemically, 2 Labeling of the mixture of released oligosaccharides with a fluorescent tag; 3 Separation of the fluorophore-labeled ohgosacchartdes by PAGE; and 4 Imagmg of the gel enher on a UV lightbox to obtain qualitattve band mformatton or using a commercial imaging system to determine the amount of ohgosaccharide present m each band and the relative mobility of the bands Once separated on the gel, tndivldual purified for further study.

3.2. Preparation

ollgosaccharlde

bands can also be

of Glycoprofein

1 Isolate the glycoprotem according to your usual procedures The sample should be relatively salt-free and contain no extraneous carbohydrates (e g , sephadexpurified material contains large amounts of glucose) (see Note 1) 2. If the volume of the glycoprotem solution required is >lOO pL, dry the glycoprotem m a 1 5 mL mtcrocentrtfuge tube. Generally 5tK200 pg of glycoprotem is required for N-linked analysis (see Note 2). For analysis of O-linked ohgosacchartdes 100-500 pg of glycoprotem may be required 3 For the N-linked obgosaccharide control (see Note 3) remove 100 mg of the chicken trypsm mhtbttor m a 45 mL ahquot, and place it m a 1 5 mL microfuge tube. Proceed with the enzymatic dtgestron as described below For the O-lmked ohgosacchartde control remove 50 pL (100 pg) of bovine submaxillary mucm, and place it in a reaction vial Proceed with lyophthzatton, P20, drying and process along with the sample glycoprotem (see Subheading 3.4.2.) Store remaining glycoprotems at 4°C for future use

3.3. Release of Asn-Linked Oligosaccharides with Pepfide N-Glycosidase F 1 Add an equal volume of enzyme buffer to the glycoprotem m solution or drssolve the dried glycoprotem m 22 5 pL of water and add 22 5 mL enzyme buffer, 2 SDS IS often required to completely denature the glycoprotem prior to enzymatic digestion To denature the protein add SDS to 0.1% (1 .O pL of 5% SDS to 45 pL reaction) and P-ME to 50 mM (1 5 pL of a 1.10 dilution of 14 4A4 stock P-ME to 45 pL reaction) Boil for 5 mm (see Note 4), cool to room temperature and add NP-40 to 0.75% (5 pL of 7.5% NP-40 into 45 pL) Mix with finger fltcks 3. Add 2.0 @ (5 U of pepttde-N-glycostdase F or as specified by the manufacturer) of enzyme to the glycoprotem sample MIX with finger flicks and centrifuge for 5 s. Store remaining enzyme at 4°C 4 Incubate sample for 2 h at 37’C. 5 Precipitate protein by adding 3 vol of cold 100% ethanol.

Polyacrylamide

Gel Electrophoresis

119

Keep samples on ice for 10 mm Spm samples m microcentrifuge for 5 mm to pellet protein 6. Remove the supernatant and transfer to a clean 1 5 mL mtcrocentnfuge tube. IMPORTANT! Do not discard the supernatant. It contains the released carbohydrates! 7 If a large amount of protein was digested (>250 pg) 5-10% of the released ohgosacchartdes may remain m the pellet. The recovery of these ohgosaccharides can be accomplished by drying the pellet completely m a centrifugal vacuum evaporator or lyophthzer Add 50 pL HZ0 to resuspend, then 150 pL 100% cold ethanol and precipitate on me Centrifuge and combine the supernatants. 8 Dry the supernatants m a centrifugal vacuum evaporator or lyophthze to a translucent pellet At this pomt samples may be stored at -20°C or proceed with the fluorophore labeling procedure described below

3.4. O-Linked Oligosaccharide 3.4.1. isolation of Glycoprotein

Release Using Hydrazine

1. Isolate glycoprotem according to your usual procedures. The purified glycoprotem should be prepared m a non-Trts buffer containing a mmimum amount of salt The presence of nonvolattle salts may cause the breakdown of the ohgosacchartdes during hydrazmolysts. If the glycoprotem 1sm a buffer contammg salt tt is recommended that the sample be dialyzed against distilled water to remove salts prior to chemical dtgestion

3.4.2. Hydrazinolysis 1 Dry 100-500 pg of glycoprotem m a glass-lined reactton vial, usmg a centrifugal vacuum evaporator or lyophthzer The actual amount of glycoprotem required will depend on the size of the protein and the extent of glycosylatton (see Note 2) 2 The sample must be completely dry before hydrazinolysts Following lyophihzatton, dry the sample overnight under vacuum m the presence of P205 to remove all traces of HZ0 Place samples m a desstcator flask with a beaker contammg a small amount of P205. Attach the dessicator directly to the pump without a coldtrap-any water remaining in the sample will be trapped by the P205 3. Open a fresh ampoule of anhydrous hydrazme O-linked cleavage reagent Add 50 pL of hydrazine to the dried sample using a glass transfer pipet or a positive displacement capillary ptpet (metal or plastic should not be used). Resuspend the dried sample. Overlay the sample wtth dry nitrogen and cap tightly Hydrazine is very hygroscopic. Discard unused hydrazine according to your hazardous waste regulations. Do not reuse. 4. Incubate samples for 3 h in a sand bath or dry heat block set at 60°C (do not use a water bath) to release 0-linked ohgosacchartdes (higher temperatures may result m the degradation of O-linked sugars or m the release of non-O-linked sugar chains, such as N-lurked sugars, from the sample tf they are present) 5. Dry samples in vacuum evaporator on low heat setting

720

Klock and Starr

3.4.3. Re-N -Acetylation Procedure 1. Add 30 pL of Re-N-acetylation Buffer to the dried pellet from step 5 above 2. Resuspend by vortexmg. Spin 2 s in a microfuge. 3. Add 2 $ of re-N-acetylation reagent to the solution Mix well Spin 2 s m a microfuge. 4 Incubate tubes on ice for 15 mm. 5. Following the 15 mm mcubatron, stop the reaction by adding 60 pL of the desaltmg resin. Desalting resin 1s prepared by adding 0 5 g of Dowex AG50X8 to 0 7 mL water Invert or vortex the resm immediately prior to removmg the 60 pL for each tube Incubate the resin with the sample at room temperature for 5 mm mixmg by placing the tube on a shaker or by continuously mvertmg the tube to keep the resin suspended. 6 Briefly centrifuge to pellet the resm, remove the supernatant (save supematant) and wash the resm 2X with 120 pL of water for 2 mm each (save supernatant) 7 Combme the resin supematants m a 1.5 mL microcentrifuge tube 8 Dry the supernatants m a vacuum evaporator on low heat setting

3.5. Labeling Oligosaccharides 3.5.1. Preparation of Samples and Standards If quantitatton of the olrgosaccharide bands in the samples is required, then one must compare the intensity of an internal standard (e.g., maltotetraose) band wrth the intensity of sample bands and it is, therefore, essential that the standard is present on each gel used (see Note 5 for preparation and use of this material).

3.5 2. ANTS Labeling 1 Prepare the labelmg dye as 1MANTS in 15% acetic acid (dye solution can be stored m the dark at -70°C for up to 2 wk). 2 Prepare 1M solution of NaBHsCN in DMSO and mix well by vortexing until crystals are completely dissolved (thus reducmg agent can be stored for 2 wk at -7O’C) 3. Add 5 pL of labeling dye to each dried ohgosaccharide pellet. Mix well until the ohgosacchartde pellet is dissolved. 4 Add 5 PL of reducing agent Mix well by vortexing. Centrifuge 5 s m microcentrifuge. 5. Incubate samples at 45°C for 3 h (temperatures higher than 45°C or times longer than 3 h can destroy or modify carbohydrates, e.g , sialic acids) Greater than 90% of the ohgosaccharides are labeled under these conditions. As a convenient altemative samples can be labeled at 37’C (not 45°C) overnight (or approx 16 h) These latter condttions result m labelmg of >90% of the ohgosaccharides (see Note 6). 6 After labelmg, dry the samples m centrifugal vacuum evaporator for approx 15 mm or until the sample reaches a viscous gel stage

Polyacrylamide

Gel Electrophoresis

121

3.6. Electrophoresis 3.6.1. Preparation of a Sample for Electrophoresls 1. Resuspend the dried fluorophore “labeled” oligosacchartde in 5-20 pL Hz0 The actual volume of Hz0 used to resuspend the sample ~111 depend on the amount of ohgosaccharide present m the sample (start with 10 pL, this will enable the sample to be diluted further if necessary) 2. Remove an aliquot of the sample (generally l-2 pL) and dilute it with an equal volume of sample loading solution Load the entire ahquot mto one lane of a gel. Best results are obtained by loading 4 pL/lane on a 10 x 10 cm gel with eight lanes

3.6.2. Electrophoresis 1 For N-linked analysis chill the running buffer to 46°C prior to use For N-lmked analysis perform electrophorests at a buffer temperature of 5-8°C. All O-linked gels are run at 15-20°C. 2. For N-linked analysis set up electrophoresis with a recirculatmg chiller and place the electrophoresis tank containing a stir bar on a mechanical stirrer. Connect the gel box cooling chamber to a refrigerating cuculator. Turn on the circulator and stirrer and set the coolant temperature to 5°C. 3. For N-linked analysis pour the precooled runmng buffer mto the electrophoresis tank up to the appropriate level The temperature of the buffer should be momtored during the run using a thermometer inserted through the hole m the lid or other method. The temperature will probably increase a few degrees during electrophoresis, but should not exceed 10°C. For O-linked gels the temperature should not exceed 23°C 4. Determine the number of gels required for the samples prepared. Each gel should contain eight lanes The outside lanes should be used for the tracking dye and glucose polymer standard leaving the stx inner lanes for samples and quantitation standard (maltotetraose). 5. Gently remove the comb(s) from the gel(s). To avoid distortmg the wells, gently wiggle each comb to free the teeth from the gel, then lift up slowly until the comb is released. 6 Place the gel cassette(s), one on each side of the center core unit of the gel apparatus with the short glass plate against the gasket. Be sure the cassette is centered and that the cassette is resting on the “feet” at the bottom of the apparatus. If only one gel IS being run place the buffer dam on the other side 7. It is essential that the wells of the gel are thoroughly rmsed out with the runnmg buffer from the upper buffer reservoir prior to sample loading. This is best accomplished by using a syringe with a blunt needle (a Pasteur pipet IS not recommended because of the possibihty of breakage into the wells) 8. With the core unit containmg the gels placed securely on the bench, load samples into the wells by underlaying the upper buffer. Use flat sequencing pipet tips to

122

9 10. 11 12.

13

14

15.

16.

17.

18

Klock and Starr load by delivering the sample to the bottom of each well Opttmal resolution will be achieved by using 4 pL of sample per lane Note For the most rehable quantttation of ohgosaccharide bands the use of a positive displacement pipet (e g., Hamilton syringe) 1srecommended. Load 4 pL of the standard m a lane when prepared as described m Note 5 Load 2 pL of tracking dye in a lane directly from the vial Load 4 pL of each labeled oltgosaccharide sample m a lane Samples should be diluted 1’ 1 m the sample loading solution (see Note 7). To prevent possible lane distortions as a result of different loading volumes it 1s recommended that 4 pL of Sample Loading Solution be loaded m any unused lanes. Best results are obtained when the same volume of sample 1s added to each lane Place the core unit containing the loaded gels mto the electrophorests tank and connect the power cords to the electrophorests tank then connect the power supply Place the thermometer mto the lower buffer chamber through the hole m the lid For N-linked analysts the mmal temperature of the lower buffer must be 5-8°C for O-lurked analysts 15-20°C IS optimal Turn on the power supply and select the proper current Gels should be run at a constant current of 15 mA/gel(30 mA for 2 gels, 15 mA for 1 gel and 1 buffer dam). Limits on the power supply should be set for 1OOOVand 60W These run conditions will result m voltages of 100-4OOV at the beginning of the run and may approach 800-1OOOV at the end of the run If the initial voltage is stgnificantly different check to be sure that the leads are connected properly and that the buffers are at the recommended levels (see Notes 8 and 9) Most N-linked oltgosacchartdes fall in the Glucose4-Glucose12 range (also referred to as G4-Gl2 or DP4-DP12), so the time of electrophoresis should be adjusted to optimize the separation of this region of the gel. Most O-linked ohgosaccharides run m the Gl-G6 range (DPl-DP6) Momtor the electrophoresis by followmg the migration of the fast moving thorm dye (orange band) Generally, electrophorests of N-linked oltgosaccharides is complete when the orange dye Just exits the bottom of the gel in approx 1 to 11/4 h. For the O-linked oltgosacchartdes the gel run 1s complete when the orange dye IS 1 cm above the bottom of the gel In a darkened room, the migratton of the labeled oligosacchartdes can be momtored directly during electrophoresis by turning off the power supply, removing the leads and the gel box cover and holding a hand-held UV light over the gels. The run can be continued by reposttionmg the gel m the electrophoresis box and reconnecting the power supply as described m step 15 The amount of time the current is off should be as short as possible (~5 min) to minimize diffusion of the oltgosacchartdes m the gel When the electrophoresis 1s complete, turn off the power supply, Disconnect the power cords from the power supply and the electrophoresis tank Turn off the refrigerated cooler and discard the buffer (see Note 10)

Polyacrylamide

Gel Electrophoresis

I

2

3

123

4

5

6

7

Fig. 1. Profiles of N-linked oligosaccharides from several glycoproteins. ANTSlabeled oligosaccharides released by peptide N-glycosidase (PNGase F) from six different glycoproteins are shown. Lane 1 contains an oligosaccharide ladder standard of partially hydrolyzed wheat starch with G4 representing glucotetraose; lane 2, chicken trypsin inhibitor; lane 3, bovine fetuin; lane 4, human a-acid glycoprotein; lane 5, bovine ribonuclease B; lane 6, human chorionic gonadotropin (hCG); and lane 7, chicken ovalbumin. The profiles show a wide variety of different glycosylation patterns, indicating the minimum number (some oligosaccharides will comigrate) of different types of oligosaccharide present and their relative quantities. The images presented in this review were obtained by imaging the gels following electrophoresis using a CCD (charge-coupled-device)-based imaging system (Glyko).

3.7. Gel Imaging CAUTION: UV protective eyeware or faceshield should be worn. Avoid prolonged exposure to UV light. 1. Allow UV lightbox to “warm-up” for at least 2 min in order to get maximum intensity output. The lightbox must be long-wave UV and have a peak output at approximately 360 nm; this is not the type of box typically used for ethidium bromide-stained DNA gels. 2. Remove the tape from the gel cassette, which may be fluorescent, and clean the surfaces of the cassette if it is required to image the gel within the cassette. The cassette glass must be of special low-fluorescence type to obtain an image of a gel within the cassette. If the glass is not of a low fluorescence type and if the type of gel cassette being used permits disassembly, remove the gel completely from the cassette and place it on the UV lightbox (see Notes 11 and 12). 3. An electronic image of the N-linked oligosaccharides from several glycoproteins is shown in Fig. 1; an electronic image of the O-linked oligosaccharides from several different glycoproteins is shown in Fig. 2. In each image you can also see

124

Klock and Starr

4 NeuAc .Gd n GalNAc

1

2

3

4

5

6

Fig. 2. Profiles of O-linked oligosaccharides from several glycoproteins. ANTSlabeled Ser/Thr-lined oligosaccharides released using hydrazine at 60°C from four different glycoproteins are shown. Lane 1 contains an oligosaccharide ladder standard of partially hydrolyzed wheat starch with G4 representing glucotetraose; lane 2, porcine stomach mucin, Type II; lane 3, bovine submaxillary mucin, Type I; lane 4, bovine fetuin; and lane 5, human chorionic gonadotropin (hCG). Lane 6 contains the NeuAc(a2-3)Gal@1-3)[NeuAc(a2-6)]GalNAc and Gal@l-3)GalNAc standards. The images presented in this review were obtained by imaging the gels following electrophoresis using a CCD-based imaging system (Glyko). the position of maltotetraose and a ladder of maltooligosaccharides from partially hydrolyzed wheat starch (see Notes 9,10,1316 for problems with bands).

3.8. Gel Handling 1. If the gel is no longer needed it should be properly discarded. 2. Following imaging of the oligosaccharide gels, the glass plates can be separated and the gels dried on a flat bed gel drier between sheets of Teflonm membrane at 80°C for 1 h. After the gel is dry, carefully peel the Teflon sheets away from the gel. Gels dried in this manner can be stored indefinitely in a dark dry location and can reimaged at any time with minimal bleaching. 3. Following imaging, the gel cassette can be placed back. in the electrophoresis apparatus and the run continued in order to improve the resolution of the oligosaccharide bands. In this case, the upper buffer should be saved and reused until the run is finally terminated. Note that diffision of carbohydrate bands and subsequent poor resolution will occur if the time between electrophoresis and imaging exceeds 15 min.

4. Notes 1. The glycoprotein sample should ideally first be dialyzed against distilled water and stored lyophilized in a 1.5 mL microfuge tube. If the sample needs to be in a buffered solution, one can place the sample in 50 mA4 sodium phosphate buffer,

Polyacrylamide Gel Electrophoresis

2

3.

4.

5

6

125

pH 7.5, at a final concentration of at least 100 yg/50 @., or 2 mg/mL. Best results are obtamed if the total salt concentration of the solution is < 100 mM. The use of a Trts-based buffer is not recommended If a detergent is required, the sample may be suspended in up to 0.1% SDS, 0.75% NP-40, or 1% n-octyl P-Dglucopyranoside If desired, the sample may also contain 0 05% sodium azide. We recommend that you use at least 250 ug of glycoprotem for analysis The actual amount of glycoprotein required for profilmg will depend on the size of the protein, the amount of glycosylation, and the degree of oligosaccharide heterogeneity. In general, the amount of glycoprotem required increases with the size of the protein or the degree of heterogeneity and decreases with the percent of glycosylatton As a genera1 guideline, one would start with approx 50-100 pg to profile the N-linked oligosaccharides of a 60 kDa glycoprotem that contains lO-20% carbohydrate by weight. For O-lurked ohgosacchartde analysis we suggest 100-500 pg of starting glycoprotem Thts amount would normally provide sufftcient material for several electrophoretic runs For isolatuon of individual ohgosaccharides, and carrying out sequencing, additional material will be required. The control for N-linked profiling consists of trypsm inhtbitor that is used as a control for enzyme digestion and fluorophore labeling This control should be included in the analysis for the following reasons* a. If this control is used for the first ttme tt will help the user to become familiar with the procedures; b If the profile obtamed looks appropriate then this assures the user that things are working properly, and c. In an unknown sample that may not contain N-linked ohgosaccharides, the user can be certain that the reagents are good and that the release and labeling procedures were performed properly. Similarly a control for O-linked profilmg such as bovine submaxillary mucm should be used. Some protems will precipitate when boiled, i.e., unmunoglobulms. The followmg procedure should be used if your protein precipitates. Add SDS/B-ME at the recommended concentration and incubate for 5 min at room temperature, add NP-40 according to duections, and then add PNGase F and incubate overnight The quantitation control consists of maltotetraose (Glucose J This control should be prepared so that once prepared 5 pL will contain 200 pmol of maltotetraose (standards prelabeled with ANTS are also available commercially). Accurate quantitation will be achieved when usmg an electronic imaging system. If usmg a commercial imaging system, refer to the manual for a detailed description of the quantitatton procedures. The stoichiometry of labeling 1s such that only one molecule of fluorophore is attached to each molecule of ohgosacchande When labeling 20 nmol or less of total sugar using the reagents and labeling conditions described, the fluorophore labeling efficiency is >95% (5). Labeling more than 20 nmol in each reaction will result in reduced labeling effictency. When labeling >20 nmol it is recommended that an internal labelmg control is included

726

Klock and Starr

7 Sample handlmg and storage’ a Always avoid exposing labeled samples and dyes to light or excess heat; b Labeled samples are stable when stored for 3 mo at -70°C m the dark, c Unused solutions of the dye and reducmg agent can be stored for as long as 2 wk at -70°C Thaw immediately before use 8 Band distortion m gels caused by vertical streaking or smearing may result if the sample IS overloaded Use a maximum of l/5 of the volume of the labeling reaction for each lane. The sample may have a high concentration of salt Remove salts by dialysis, desalting column, and so on, prtor to enzymatic drgestion Distorted sample wells m gel may be caused by tearing of wells when the comb was removed Remove the comb slowly usmg a gentle back and forth rocking motion and lift vertically Alternatively, gels may have been in contact with upper buffer too long prior to sample loading. Samples should be loaded within 5 mm of placmg the gel m the upper buffer tank 9 Voltage and/or current leaks can result when high voltages are used If at the begmnmg of the run the voltage is >4OOV or readings are unstable, turn the power off before checkmg the followmg possible electrical leak, check for cracks m glass plates Remove inner core assembly and check for buffer leak between gaskets and cassette plates If leaks are evident check that the plates are clean and not cracked or chipped, and that they are installed properly 10 Buffers should not be reused as they have fluorophore contamination after use Reuse of buffers may result m no bands being visible on the gel owmg to “washout” of the fluorophore-labeled oligosaccharides 11 Accurate quantification is essential for detailed carbohydrate analysis Although oligosacchande patterns on PAGE gels can be viewed and photographed on a standard laboratory UV Itghtbox, it is not reliable for accurate quantification Images of gels can be recorded usmg a Polaroid camera The proper choice of light source, filters, and film must be made. A filter must be fitted to the camera lens that completely covers the glass of the lens (stray UV contacting the lens ~111 cause it to fluoresce and subsequently lower the sensitivity of the film) A suitable filter will have no Inherent fluorescence, peak transmisston at approx 500 nm and bandwidth of 80 nm FWHM A medium speed, medium resolution, Polaroid film is recommended Use Polaroid 53 film for cameras which use single 4 x 5” sheet film, use Polaroid 553 film for cameras that use 8 sheet film cartridges. To visualize the carbohydrate bandmg patterns, the low fluorescent glass cassette contaming the gel (or the gel removed from the casette) is placed on a longwave UV lightbox with a peak excitation output at approx 360 nm Photograph the gel using the lowest practical f-stop setting on the lens with the gel fillmg as much of the frame as possible. E g., exposures at f5.6 using Polaroid 53 film have ranged from 540 s using the eqmpment specified above. Keep UV exposure of the gel to a minimum to prevent bleaching. Develop the film according to the manufacturer’s mstructtons

Polyacrylam/de Gel Electrophoresrs

127

12 For electromc archiving and quantttatton, several types of imaging systems are available To give best results these systems must have an lllummation source with an excttatton wavelength of 365 nm and a 520 nm emission filter placed m the light path between the gel and the image capturmg devtce The use of an internal standard in the gels is also required for quantttatlon Followmg electrophoresis, the gel is inserted mto the imager under long-wave UV excitation, and an electronic image of the fluorescent carbohydrate banding pattern of the gel IS acquired by the imager’s CCD as a digital image The gel image 1s displayed on a computer screen using the imaging software The tmaging system should allow for detection and quantlficatton of mdtvtdual carbohydrate bands mto the low picomole range of 1.6-300 pmol In practice, the most useful and accurate range of the tmager for band quanttficatton 1sbetween 5 and 500 pmol of carbohydrate and this range was used for the experiments described m this chapter 13. You may have “smile effect” gel dtstortions at both sides of the gel This can happen if he gel is not being cooled umformly. Check that the cooling system is on and working properly Check the buffer temperature Check that the power supply is set for the proper current level 14 You may have band dtstortions or “fuzzy bands ” This can be caused by wells that may have not been rinsed thoroughly with electrophorests upper buffer prior to loading samples or the current may have not been set properly, 1.e , the current was too high 15 Incomplete re-N-acetylatton may result m little or no labeling of the released oltgosaccharides, presumably by the hydrazide mterfermg with the reductive aminatton using ANTS To check the re-N-acetylation reagents you can use glucosamme that will migrate at a DP of approx 2 5 on the gel when N-acetylated and below DPl when unacetylated You can also use N-acetyl-glucosamme as an internal control at the begmning of the experiment and take it through hydrazmolysls and re-N-acetylatlon expecting the same migrations as stated 16 Ohgosaccharides are small molecules that can diffuse rapidly m the gel matrix. Band diffusion and resulting broad or “fuzzy” bands can occur tf a Electrophoresis is run too slowly, b. Electrophorests is run at higher than optimal temperatures, c. Electrophoresis is stopped and started repeatedly; and d The gel 1sremoved for vtsualization for longer than 10 mm and then re-electrophoresed

References 1. Williams, G. R and Jackson, P (1992) Analysts of carbohydrates US Patent 5,508 2 Jackson, P. (1990) The use of polyactylamtde-gel electrophoresls for the high resolutron separation of reducing sugars labeled with the fluorophore Il-ammonaphthalene-1,3,6-trtsulfomc acid. Bzochem J 270,705-7 13.

128

Klock and Starr

3. Jackson, P. and Wtlhams, G R. (1991) Polyacrylamtde gel electrophoresis of reducing saccharides labeled with the fluorophore 8-aminonaphthalene- 1,3,6trisulfonic acid: applicatton to the enzymatic and structural analysis of ohgosaccharides Electrophoresls 12, 94-96 4 Jackson, P. (1994) The analysts of fluorophore-labeled glycans by high-resolution polyacrylamide gel electrophorests Anal Blochem 216,243-252. 5. Starr, C , Masada, R. I., Hague, C , Skop, E. and Klock, J. (1996) Fluorophoreassisted-carbohydrate-electrophoresis, FACE@ m the separation, analysis, and sequencmg of carbohydrates. J Chromatogr A 720,295-32 1. 6. Masada, R I , Hague, C., Seid, R., Ho, S., McAhster, S , PIglet, V, and Starr, C (1995) Fluorophore-asslsted-carbohydrate-electrophoresis, FACE@, for determmmg the nature and consistency of recombinant protein glycosylation. Trends Glycoscl Glycotech. 7, 133-147. 7. Hu, G. (1995) Fluorophore assisted carbohydrate electrophorests technology and applications J. Chromatogr 705,89-103 8. Higgins, E and Friedman, Y (1995) A method for momtormg the glycosylatton of recombinant glycoprotems from conditioned medium, using fluorophore assisted carbohydrate electrophorests. Anal. Blochem 228,22 l-225 9. Basu, S. S , Dastgheib-Hossemt, S , Hoover, G., Lt, Z , and Basu, S (1994) Analysts of glycosphingohpids by fluorophore-asststed carbohydrate electrophorests usmg ceramtde glycanase from Mercenarra mercenarra Anal Blochem 222,27O-274 10 Flesher, A. R., Marzowski, J, Wang, W., and Raff, H. V (1995) Fluorophore-labeled carbohydrate analysts of mnnunoglobulin fusion proteins. correlation of oligosaccharade content with m viva clearance profile. Blotech Bzoeng 46,399-407 11 Lee, K B., Al-Haktm, A., Loganathan, D., and Lmhard, R J (199 1) A new method for sequencing carbohydrates using charged and fluorescent conJugates Carbohydr Res 214, 155-162. 12. Stack, R. J. and Sullivan, M T. (1992) Electrophoretic resolution and fluorescence detection of N-lmked glycoprotem ohgosacchartdes after reductive aminatron with 8-aminonaphthalene- 1,3,6-trisulfomc acid Glycobzology 2, 85-92 13. Roy, S. N., Kudryk, B., and Redman, C. M ( 1995) Secretion of biologtcally active recombmant fibrinogen by yeast J Bzol Chem 270,23,761-23,767. 14 Denny, P C., Denny, P A , and Hong-Le, N. H (1995) Asparagme-linked ohgosaccharides m mouse mucm. Glycobzology 5,589-597 15 Qu, Z , Sharkey, R M., Hansen, H. J , Goldenberg, D. M , and Leung, S -0 (1997) Structure determination of N-linked ohgosaccharides engineered at the CH 1 domain of humanized LL2 Glycobiology 7,803-809. 16 Prieto, P A., Mukerp, P., Kelder, B., Erney, R , Gonzalez, D , Yun, J S., Smnh, D F., Moremen, K. W., Nardelli, C., Pierce, M., LI, Y., Chen, X., Wagner, T. E., Cummings, R D., and Kopchich, J J. (1995) Remodeling of mouse milk glycoconjugates by transgemc expression of a human glycosyltransferase J Brol Chem 270,29,5 15-29,5 19. 17. Pirie-Shepherd, S , Jett, E. A., Andon, N. L., and Pizzo, S. V. (1995) Siahc acid content of plasmmogen 2 glycoforms as a regulator of fibrmolytic activity. J, Bzol Chem 270,5877-5881.

Polyacrylamide Gel Electrophoresis

129

18 Goss, P E , Baptiste, J., Fernandes, B., Baker, M., and Denms, J. W. (1994) A phase I study of swamsonme m patients with advanced malignanctes Cancer Res. 54, 1450-1457. 19. Sango, K , McDonald, M. P, Crawley, J N., Mack, M. L., Ttfft, C J , Skop, E , Starr, C. M , Hoffmann, A., Sandhoff, A., Suzukt, K., and Prota, R. L. (1996) Mice lacking both subunits of lysosomal beta-hexosammidase display ganghosidosis and mucopolysaccharidosts Nature Genet 14, 348-352. 20 Masada, R. I., Skop, E , and Starr, C. M (1996) Fluorophore-assisted-carbohydrate-electrophoresis for qualtty control of recombmant protem glycosylatton Blotechnol Appl Blochem 24, 195-205 21 Kumar, H. P. M., Hague, C , Haley, T., Starr, C. M., Besman, M. J , Lundblad, R., and Baker, D (1996) Elucidation of N-linked oligosacchartde structures of recombinant factorVI1 using fluorophore assisted carbohydrate electrophorests. Blotechnol. Appl Biochem 24,207-2 16

9 Analysis of Glycosaminoglycans and Proteoglycans Christopher

C. Rider

1. Introduction The btosynthests of glycosaminoglycans (GAGS) and proteoglycans appears to be a ubrqurtous functton m mammahan cells Some brologtcal sources, notably connective tissue, produce large quantities of GAGS that can be readily detected by colortmetrrc assays and, therefore, may be investigated by wellestablished techniques that are fully described elsewhere (see Subheading 1.3.). However, most tissues and cell cultures will yield only submtlhgram amounts of these macromolecules. This means that then mvestigatton reqmres radlolabelmg. For proteoglycans bearing sulfated GAGS, [35S]sulfate IS a relatively specific and readily incorporated radtolabel. [35S]sulfate IScomparattvely inexpensive, and is efficrently detected by both hqutd scmtrllatton countmg and fluorography. A major disadvantage IS the relatively short half-life of 35S,88 d, whtch ltmtts the time available for postmcorporatton analysts. However, an advantage of radiolabelmg IS that only those macromolecules synthesized during the labeling period will be studied Previously synthesized macromolecules that may be parttally degraded will not be detectable. After metaboltc labelmg rt IS necessary to remove unmcorporated [35S]sulfate. This can readily be achieved by gel filtratton chromatography or by exhaustive dialyses. Gel filtratton is more convement from the point of view of disposal of radiotsotope, an Important constderatton smce by far the majortty of radtolabel employed will remam unmcorporated.

From

Methods Ed&d

by

m Molecular E F Hounsell

Biology,

Vol 76 Glycoanalysls

0 Humana

737

Press

Inc ( Totowa,

Protocols NJ

132 7.2. Anion-Exchange

Rider Purification

of GAGS and Proteoglycans

The desalted preparation will still include macromolecules with nonGAG sulfate radiolabel, in particular, proteins with sulfated tyrosmes or sulfated oligosaccharides. GAGS have much higher anionic charge densities and, therefore, then separation from other macromolecules is achieved effectively by anion-exchange chromatography. Anion-exchange chromatography of proteoglycans together with their extraction from labeled cell cultures has been fully covered elsewhere (1). However, several manufacturers now produce ionexchange membrane cartridges that offer the advantages of fast flow rate, high capacity in a low bed volume, and a rigid bed to facilitate column washing Since anion-exchange separation 1sbased on charge density, for cells producmg multiple GAG chain types, the resultmg chromatographlc profiles may show resolution of the bound fraction into multiple peaks 1.3. Calorimetric Estimation of GAGS GAGS may be quantified calorimetrically by the reaction of their uromc acid residues with carbazole. A widely used protocol is one originally reported by Bitter and Muir (2). This has been fully described elsewhere (34). Although the Bitter and Muir assay 1swidely used for the analysis of GAGS, it 1s not reliable for concentrations below approx 5 pg/mL. A more sensltlve colonmetnc method is a dimethylmethylene blue dye-bmdmg assayreported by Farndale et al (5) The variant of this assaypresented m this chapter 1sour adaptation of it for microtiter plate format. This version is very sensitive, having a working range of O-4 pg GAG The Farndale method also has the advantages of convenience, and of avoiding the concentrated sulfuric acid used in the Bitter and Muir method (2) 1.4. Selective Enzymic and Chemical Degradation of GAGS It is often of interest to liberate intact GAG chains from their polypeptide cores m order to determine their size and other charactenstlcs. This may be achieved by alkaline p-ellmmation, a procedure resulting in the release of all glycans that are O-linked to serme or threomne. If required, the free glycosamlnoglycans can be separated from liberated O-linked oligosaccharldes by gel filtration. The procedure descrtbed here uses a relatively low concentration of alkali, which should minimize release of N-linked oligosaccharides. Nonetheless, some cleavage of polypeptide chains may occur. The class of GAG present may be established through the use of several well-established selective enzymatic and chemical methods of degradation. Susceptibility or resistance to these methods differentiates between the GAG types. Moreover, the GAG fragments so obtained may be subject to further

Glycosaminoglycans

and Proteoglycans

133

analysis so as to yield structural informatlon on the chains from which they have been derived. Chondroltm sulfate and dermatan sulfate may be degraded with chondroltm ABCase, which degrades both, or chondroltin ACase, which 1s specific for chondroitin sulfate and does not digest the lduronic acid-rich chains of dermatan sulfate (formerly termed chondroltm B). Use of chondroltin ABCase is described later m the chapter. The selective partial degradation of heparin and heparan sulfate may be achieved usmg one or more of the commercially available heparmase or heparltinase preparations. The descnptlon, use, and substrate specificity of these enzymes 1sfully covered elsewhere (6). Heparm and heparan sulfate differ from other GAGS by the possession of unacetylated glucosamme residues. This hexosamine, n-respective of whether it 1sN-sulfated or not, is subject to deaminatlve cleavage m the presence of nitrous acid. Such treatment of a GAG will, therefore, result m chain cleavage at the site of each nonacetylated glucosamine, and the extent of depolymerlzatlon will depend on the frequency and distribution of this particular monosaccharide. 1.5. Determination of the Size of Proteoglycans and GAGS The size and dlstrlbutlon of either intact proteoglycans or elimmated GAGS has been conventionally determined by gel filtration under dissociating conditions (4M guamdine hydrochloride) on an appropriate grade of Sepharose CL (1,2). This 1sa reliable technique, but also laborious, especially when multiple samples are to be compared. Large proteoglycans, such as the cartilage chondroitin sulfate proteoglycan (M, l-4 x 106), cannot be resolved by electrophoresls on polyacrylamide gels becauseof their size.However, a polyacrylamlde-agarose gel method has been devised for this purpose by McDevitt and Muir (7), and a full description of this procedure is available elsewhere (I). Proteoglycans synthesized by many cell types have smaller hydrodynamic sizes, approaching those of large polypeptldes. Therefore, conventional sodium dodecyl sulfate polyacrylamide slab gel electrophoreis methods may be used. In this laboratory, we routinely employ the protocol of Laemmh, a complete account of which 1sto be found elsewhere (8; and see vol. 1 of this series). However, most laboratories will have their own variant procedure, which is likely to prove satisfactory (see Note 1). Determination of molecular weight is not straightforward, because even proteoglycan molecules with the same core polypeptlde will display conslderable size heterogeneity This arises from disperse lengths of the GAG chain and even variation in the number of chains carried. Dispersity 1sparticularly evident on gel electrophoresls with its high resolution of molecular weight. Proteoglycans will separate as a smeared band extending over a considerable apparent A4, range.

Rider

134

The size of liberated GAG chains may be determined by gel filtration. Compared to proteins, free GAGS migrate anomalously rapidly through dlscontinuous SDS polyacrylamlde gels However, size separation of GAGS can be performed by the contmuous polyacrylamlde gel electrophoresls method of Hampson and Gallagher (9), which 1s described here adapted to the mmigel format. By using 12-15% polyacrylamide gels, relatively large GAGS can be separated. However, it 1s possible to resolve small oligosacchandes, such as those obtained by fragmentation of the intact GAG chains, on 20% polyacrylamlde gels.

2. Materials 2.7. Metabolic Labeling of GAGS and Proteoglycans by Lymphocytes and Lymphoma Cells

Synthesized

1 2 3 4

Sterile, tissue-culture grade 24-well plates. Sterile plastic Pasteur pipets Sterile 50-mL conical centrifuge tubes Powdered, sulfate-free RPM1 1640 Dutch Modlficatlon medmm, contammg HEPES, but without magnesium sulfate or sodium bicarbonate (Glbco, BRL, Life Technologies, Paisley, Scotland, UK). The powdered medium for 1 L 1sdissolved m 950 mL of high-punty water containing MgCl2,6H20,82 5 mg, and NaHC03, 2 g (The pH 1sroutmely 7.3, but should be adjusted with HCl or NaOH If necessary ) The volume is made up to 1 L, and the medium is sterile-filtered mto presterihzed media bottles Samples from the begmnmg and end of the filtration may be collected into small sterile Petri dishes and Incubated at 37’C for 48 h to check for sterility This reconstituted medium has a shelf-hfe of 3 mo at 5°C 5 Heat-mactlvated fetal calf serum, L-glutamme, pemclllm-G, and streptomycm sulfate 6 [35S]sulfurlc acid m water (ICN Blomedlcals, High Wycombe, UK) 7. Desalting column of BloGel P-6DG or Sephadex C-SO (medmm), of bed volume at least 10 times the sample volume Prepacked disposable columns are especially convement given that considerable levels of [35S]sulfate may be absorbed on the column. The column is equilibrated before use with elutlon buffer, which may be selected accordmg to the subsequent analyses to be applied to the sample (see Notes 2 and 3)

2.2. Anion-Exchange 1 2 3 4

Purification

of GAGS and Proteoglycans

DEAE MemSep 1000 ion-exchange cartridge (Mtlhpore Corp , Bedford, MA) Filter device (0 22 pm) HPLC grade water Sodium acetate buffer (50 n&Q, pH 5.8 containmg 180 mMNaCl,6Murea, and 0.1% (w/v) Zwlttergent 3-08 (Calbiochem-Novablochem, Nottingham, UK) (see Notes 2 and 3) 5 In-line 280-nm absorbance monitor and perlstallc pump.

Glycosaminoglycans and Proteoglycans 2.3. Calorimetric

Estimation

135

of GAGS

1 2. 3 4. 5.

1,9-Dlmethylmethylene blue (Taylor’s Blue) (Aldrich, Gillingham, UK) Glycme, free acid, at least 99% pure NaCl, analytical grade HCl(0 1M) 100 mL. 96-Well optical grade, flat-bottomed mlcrotlter plates (e.g , Dynatech Laboratories, Billingshurst, UK, Catalog No 655 170); sterile or ELISA grade plates are not necessary 6. ELISA plate reader with 525 nm optical filter

2.4. Selective

Degradation

of GAGS and Proteoglycans

1 2 3 4 5

HCl(0 1M). Sodium borohydrlde (see Note 4) Glacial acetic acid NaOH (0 1M) I-Butyl nitrite (Aldrich, stored at 5’C) freshly diluted m 4 vol ethanol on day of use (potential carcinogen, see Note 5). 6 Cetylpyrldmm chloride (10% [w/v]) m distilled water. 7 Chondroitin ABCase (chondroltin ABC lyase) from Proteus vulgarzs (ICN Biochemlcals, Seikagaku Kogyco, Tokyo, Japan, or Sigma, Poole, UK). The enzyme 1s to be reconstituted m 0 OlMTrls-HCl, pH 8 0, contammg 50% glycerol and will remam active m storage at -2OOC for at least 1 yr 8. Standard heparm and chondroltm sulfate preparation made up at 2 mg/mL in distilled water.

2.5. Size Determination

of GAG Chains 6y Gel Electrophoresis

1. Vertical slab mmlgel apparatus (Mini-Protean II, Blo-Rad, Hemel Hempstead, UK, or equivalent) and constant voltage DC power supply with maximum voltage output of at least 250 V 2 Glycme buffer (0 2M), containing 2.5 mMEDTA and 5 tiNaN adjusted to pH 8.9 by addition of solid Trls base This 1s2X electrophoresls buffer Store at 5°C. 3 Acrylamide (40 g) and 0 13 g blsacrylamide dissolved in 100 mL distilled water Store at 5OC 4 Stam; 0 08% (w/v) aqueous azure A. 5, Ammonium persulfate 6. TEMED (N,N,N’,N’,-tetramethylethylendiamme) 7. Size standard heparm and chondroitin sulfate preparations (Most commercial suppliers will provide size data on request, but these are not quoted here because of hkely batch-to-batch varlatlon ) For ohgosacchande resolutions, a low molecular weight clinical grade heparin such as Clexane (Enoxaparin, Rhone Poulenc Rohrer) 1s a convenient mixture of heparm oligosaccharides ranging from tetramers to 18-mers 8 Sample buffer comprlsmg 2.5 mL Tris-glycme buffer, 2 0 mL distilled water, 0 5 mL analar glycerol, and sufficient Bromophenol blue and Phenol red to provlde an mtense color

Rider

136

3. Methods 3.1. Metabolic Labeling of GAGS and Proteoglycans by Lymphocytes and Lymphoma Cells

Synthesized

1 In the week of use, prepare the complete labelmg medium by supplementmg the reconstituted, sulfate-free medium with 2 mM L-glutamme, 14 pL/L, 2-mercaptoethanol and 60 mg/L pemclllm-G. At this stage, two sources of sulfate are also added, streptomycm sulfate, 1 55 mg/L, and 1% fetal calf serum (see Notes 6 and 7). 2 Harvest the cells by spinning at 500g for 10 min at room temperature After resuspension in labeling medium, count the cells and assess vlablllty by dye exclusion Populations with vlabllltles ~90% are unsuitable for study. 3. Wash the cells m labeling medium, and finally resuspend at 107/mL (Verify this density with a further count ) 4 Place 150 pCl [35S]sulfate m each labeling well, add 1 mL of cell suspension/ well, and immediately transfer the plate to an Incubator at 37OC with a humldltied atmosphere with 5% C02. (At least one well 1s incubated without label, so that cell vlablhty may be reassessed during the labeling incubation ) 5 After 4 h, terminate the mcubatlon by transferring the well contents to Ice-cold centrifuge tubes after thorough resuspension of the cells with a Pasteur pipet Spm these cell suspensions immediately at 5OOg for 10 min at 4”C, wash the cell pellets twice with ice-cold phosphate-buffered salme, and combme these washings with the first supernatant. 6 Desalt labeled cell extracts and labeling supernatants by gel filtration to separate macromolecular, incorporated sulfate from unmcorporated radlolabel The resultmg elutlon profile will possess a minor void volume peak of macromolecular incorporated radiolabel followed by a major peak of free sulfate at the bed volume. Baseline separation is not usually obtained, because some mcorporatlon will be into molecules of intermediate size. Notable among these are the sulfated glycohpids Selected peak fractions should, therefore, avoid such material

3.2. Anion-Exchange

Purification

of GAGS and Proteoglycans

1. Filter and degas all solvents and buffers on the day of use 2. Flush the cartridge with 10 mL of filtered HPLC grade water at a flow rate of 10 mL/h, and then equilibrate with 10 mL of chromatography buffer 3. Filter the desalted proteoglycan sample dissolved m, or dialyzed agamst, chromatography buffer, through a 0 22-p membrane, apply to the cartridge, and elute w;th 20 mL of buffer. The condltlons of chromatography are such that only highly anionic molecules bmd to the column. 4. Once the eluted radloactlvlty and 280-nm absorbance have fallen back to low plateaus near zero, elute the bound proteoglycan with a 40 mL linear salt gradient rising to 1 8MNaCl in chromatography buffer. 5. Wash the cartridge m sequence with the followmg filtered solutions (see Note 8) a 10 mL HCl(O.5A4). b HPLC-grade water until the pH is >4.0

Glycosaminoglycans and Proteoglycans

240 2 5 e : e 2 .-I

137

-

160-

60-

6

12

Fraction

16

24

30

number

Fig. 1 Analytical scale anion-exchange chromatography on DEAE MemSep 1000 cartridge of crude proteoglycan preparattons obtained from the culture medium condittoned by a murme lymphoma cell lme The separation was performed as described m the text. A linear salt gradient from 0 15-1 8M NaCl was started at fraction 13 The fraction volume was 1.8 mL 3, Proteins concentration; 0, [35S] radioactivtty. c 10 mL NaOH (0 SM. d. HPLC-grade water until the pH is ~8.0. e Sodium azrde (0.02% [w/v]) The unit may then be stored at 5’C 6. An example of the separation of proteoglycan by ion-exchange chromatography on a DEAE Memsep cartridge IS shown in Fig. 1 The bound peak, fractions 18-2 1, is proteoglycan, whereas the minor breakthrough peak, fractions 3-8, is sulfated protein. These separations readily resolve proteoglycan from protem as shown in Fig. 1, and many-fold purification IS achieved m one step. However, the proteoglycan fraction is still likely to contain contammatmg proteins Further ton-exchange separations under similar conditions will be required to remove these. It will be beneficial to alter the nature of the anion-exchange medium, and Mono-Q (Pharmacia LKB Biotechnology, Uppsala, Sweden) is an attractive choice for subsequent chromatography

3.3. Calorimetric

Estimation

of GAGS

1 Dissolve 3.04 g glycme and 2.37 g NaCl in 95 mL O.lM HCl. 2 Place on a hot-plate stirrer and raise to slightly above laboratory temperature 3. Stir vtgorously and add 16 mg dimethylmethylene blue gradually over a pertod of approx 2 mm. 4 After a further 2 h stirrmg make up to 1 L with deiomsed water and contmue stirrmg overnight 5. Store the reagent at laboratory temperature m a fotl-covered dark glass bottle away from direct sunlight This is stable for 6 mo (see Note 9)

Rider

138 .6

1

0

1

2

Heparrn

3

I4

Fig. 2 Standard curve for the microtiter plate calorimetric assay of heparm The assay was performed as described m Subheading 3.3. Each point is the means of tripltcates. 6. Make up GAG standards O-4 pg m a final volume of 25 pL and plate out m microtiter plate wells, together with unknowns appropriately diluted and also m a final volume of 25 pL All samples should be plated out m triplicate. 7 Add 200 & of dye solution and read at 525 nm nnmedlately curve 1s shown m Fig. 2 (see Note 10).

A typical standard

3.4. Selective Degradation of GAGS and Profeoglycans 3.4.1. Alkaline p- Elrtnina t/on 1. Dissolve buffer-free, lyophilized samples m 1 mL 0 1M NaOH. 2 Add soled sodium borohydride, 0 038 g, to each sample. 3. Incubate for 18 h at 37°C Place the samples on ice and neutralize by slow, stepwlse

20 $.., additions of glacial acetic acid over a 1 h period

(Neutrality

should be attamed after 160 pL of added acid, at whtch time no further effervescence ~111 be seen on addltlon )

3 4.2. Nitrous Acid Cleavage of Heparin/Heparan

Sulfate

the salt-free sample to be treated m 1 mL dlstllled water (see Note 11). Set up controls, comprising two samples of heparm and two samples of chondroitin sulfate, contammg 2 mg GAG m 1 mL distilled water m clear glass tubes. (One of each pair will serve as treated control, with the other as untreated control.) 2. Add 0.5 mL 1M HCl to each sample and control 1. Dissolve

Glycosaminoglycans

and Proteoglycans

139

3. Add 0 5 mL of ethanohc butyl nitrite to each sample and treated control. Add 0 5 mL ethanol to untreated controls 4. Incubate for 2 h at room temperature with brief vortex mtxmg every 15-20 mm, and stop the reaction by neutrahzatton with 0.5 mL 1MNaOH (see Note 11) 5. Check that selective cleavage of heparinlheparan sulfate GAGS has occurred by addmg 100 p.L 10% (w/v) cetylpyridium chloride solutton to each of treated and untreated controls Flocculant prectpttatton of intact GAGS will occur within 15 mm

3.4.3. Chondroltin ABCase Digestion of GAGS 1 Prepare the GAG samples and controls in dtsttlled water, as for nitrous acid cleavage above (see Note 3) 2 Set up 1 mL of each sample and control m mdivtdual screw-capped plastic centrifuge tubes Add 50 $ of 0.2M Trts-HCl buffer, pH 8 0 3. Reconstitute the lyophiltzed enzyme at a concentration of 2 U/mL m 0 OlM Tris-HCl, pH 8 0, containing 50% (v/v) glycerol, where 1 U is the enzyme acttvtty capable of liberating 1 pm01 of disaccharide/mm at 37°C. Add 100 pL of this enzyme solution/digestion, and 100 pL of enzyme buffer to undigested controls 4. Add 1 drop of tolueneltube, seal the cap tight, and incubate for 24 h at room temperature 5 Check controls for digestion as described for nitrous acid degradation using cetylpyrtdmmm chloride prectpttatton (see Subheading 3.4.2., step 5) 6. The digested sample sould be boiled for 20 mm to destroy enzyme activity

3.5. Size Determination

of GAG Chains by Gel Electrophoresis

1. To resolve oligosacchartde fragments of GAG chains (see Fig. 3) cast a 20% acrylamtde gel by adding 6.25 mL Tris-glycme buffer, to 6.25 mL acrylamtde/ btsacrylamide solution 2. Add 0 55 mL freshly prepared 10% (w/v) aqueous ammomum persulfate and 11.75 pL TEMED with thorough mixing 3. Leave to polymertze (The cast gel may be stored for several days at 5°C provided tt is sealed to prevent tt from drying out; see Note 12.) 4 Fill the assembled gel apparatus with Tris-glycme buffer diluted with an equal volume of distilled water. 5 Pre-electrophorese the gel for 30 mm at 150V 6. Fill sample wells with 20 ~18GAG m 10 pL sample buffer 7. Run at 200V After approx 1 h, when the Phenol red 1sapproaching the bottom of the gel, switch off the power supply, remove the gel from the tank, and measure the migration distances of the two dyes (The migration distance of the Bromophenol blue should be 3/4 that of the Phenol red ) 8 Immediately transfer the gel to the stam solution, and leave on a rotory shaker for 15 mm Destain in several changes of disttlled water, and dry the gel as soon as the background 1s sufficiently clear (see Note 4)

140

Rider

1086’

Fig. 3. Resolution of heparin oligosaccharides on a 20% polyacrylamide gel. The vertical axis shows oligosaccharide size in terms of number of hexose residues per oligosaccharide. Track a, low molecular weight heparin (Clexaner) used as calibrant. Tracks b-i, oligosaccharide fragments from enzymically depolymerized heparin resolved by preparative gel filtration on Bio-gel P4; b and d, hexasaccharide, c, octasaccharide, e, decasaccharide,f, dodecasaccharide, g, tetradecasaccharide, h, hexadecasaccharide, i, octadecasaccharide. Some of the heterogeneity observed in tracks b-i, occurs because the oligosaccharide pools have not been completely separated from each other, but heterogeneity also arises because there are multiple structural variants within each size pool.

4. Notes I. Gel-filtration columns and electrophorersis gels are calibrated with globular protein standards, either in native conformation or denatured according to the system. The highly anionic linear chains of GAGS adopt extended conformations in solution. Therefore, proteoglycans, even in the presence of denaturants, will run anomalously when compared to standards. 2. The presence of either 4M guanidine hydrochloride or 6iU urea in the elution buffer will minimize binding of contaminating proteins to the GAGS and proteoglycans. Guanidine hydrochloride is likely to be more effective in this regard, but will interfere with any subsequent anion-exchange purification. It is more likely to denature the core polypeptide, thus destroying its biological activity. It is also a skin irritant. The buffer should also contain a detergent to minimize nonspecific binding. The zwitterionic detergent, Zwittergernt 3-09 (Calbiochem), at a concentration of 0.1% (w/v), is suitable. 3. To avoid proteolysis, routinely add on the day of use 10 mM-amino-hexanoic acid, 10 rnM N-ethylmaleimide, 1 mM benzamidine HCI, and 0.2 rnM phenylmethylsulfonyl flouride. The latter must be introduced by dissolving in 50 pL dimethyl sulfoxide, which is then added to 100 mL of eluant buffer with vigorous stirring. Where proteolysis may be a particular problem, a supplemen-

Glycosaminoglycans and Proteoglycans

4.

5 6

7

8.

9.

141

tary cocktail of antipam leupeptin, aprotimn, and chymostatm 1semployed These are made up as a mixture, each 0.5 mg/mL m buffer, stored frozen, and added 1% (v/v) on day of use. Sodmm borohydride may undergo explosive hydration on exposure to a humid atmosphere and IS also highly toxic It must be stored under dry nitrogen and opened only m a fume hood with the sash down as far as possible Safety glasses should be worn, as should laboratory overalls with the cuffs tucked inside rubber gloves, The smallest bottle supplied, 10 g (Sigma), is adequate for treatment of many samples Sodium borohydride is not essential for the elimmation reaction, but serves to reduce the liberated glycan m order to prevent its degradation by “peeling reactions.” Warning: Butyl nitrite is a potential carcmogen; thus, it and soluttons contammg it should only be handled m a fume hood. Gloves should be worn. The labeling medmm employed should be a sulfate-free variant of the most appropriate medmm for the particular cells or tissue preparation being studied Usually, magnesium chloride is substituted for magnesium sulfate in the formulation. Undialyzed fetal calf serum is hkely to have a sulfate concentration of around 1 mA4 Therefore, m the case of cells that require concentrations of 5% or above for survival during labeling, it may be worthwhile trying dialyzed serum The aforementioned protocol uses low-sulfate medium to give high mcorporations of radiolabel. A major potential drawback is that cells exposed to sulfate starvation may become deficient m the synthesis of sulfated polysaccharrdes. It is, therefore, important to check that the rate of sulfate incorporation is linear throughout, and ideally beyond, the labeling period. However, this will not rule out the possibility of qualitative changes in sulfation induced by sulfate deprivation Silbert’s laboratory has shown that medium sulfate concentrations of 0.1 mM result m undersulfation of glycosammoglycan chains, which remam of normal length. Such undersulfated chains may have further structural deticienties In particular, the undersulfated dermatan/chondroitm sulfate glycosaminoglycans synthesized by sulfate-deprived fibroblasts were found to have a markedly reduced degree of epimerization of glucuromc acid to iduromc acid (10) Therefore, if detailed structural studies of sulfated products are envisaged, it may be preferable to raise the sulfate content of the labeling mcubation and sacrifice high incorporatrons of radiolabel Refer to the manufacturer’s Instructions. It is particularly important that the cartridge is not allowed to dry out The maximum operating pressure is 1.3 bar (18 psi) that should allow a maximal flow rate of 20 mL/mm The aforementioned flow rate is therefore highly conservative, and flow rates of 2-5 mL/mm are recommended by the manufacturer However, since the nominal bed vol IS 1 4 mL, 10 mL/h allows several bed vols of eluate change per hour The sensitivity of the assay depends on obtaining a high dye concentration, as near to saturation as possible Lower dye concentrations will give a poorer dose response m the standard curve (see Fig. 2). The dye concentration of freshly made up or stored reagent stocks should give an A5z5”,,, of circa 0 4 for 100 pL

Rider plated out m a mtcrotiter well m the absence of GAGS. Since there may be some batch to batch vartatton m dye quality, where low absorbances are obtained, workers may find it worthwhtle to experiment with minor variations m the procedure for makmg up the dye reagent 10 The assay principle 1s that on bmdmg of the basic dye to acidic GAGS, there 1s a shift in the peak absorbtion readily vtsuahzed as a change from blue to a pmkishpurple. A number of consequences arise from the tonic nature of the Interaction. Firstly, GAG standards of dtffermg charge denstty will give differing standard curves. The caltbratton of the assay is, therefore, not absolute m terms of GAG mass, but simply relative to the standard employed. Secondly, buffer salts will also bmd the dye, giving the same metachromatic shift This will be observed as an Increase m the blank values obtained where buffer alone 1s added to the dye, with consequent flattening of the standard curves It is, therefore, essential that standards are dtluted m the same buffer as the unknowns, and that this should be the lowest ionic strength practtcable This is a parttcular problem when assaymg ion-exchange fractions where the ionic strengths are high, and changing m the gradient fractions where GAGS and proteoglycans are eluted (see Subheading 3.2.) In some instances where the amount of GAG is high, this problem may be avoided by the several-fold dilution of the sample for assay Where this is not possible, the assay can only be used semtquantitattvely to detect GAG-postttve fractions. Finally, since the dye neutrahzes actdtc charges on the GAGS, precipttatton of the dye complexes occurs, hence the requirement to read wells nnmediately after dye additions 11 As with the electrophoresis of proteins on polyacrylamide gels, the crosslmkmg of the gel determines the resolution range Therefore tf the mterest IS m resolvmg intact GAG chains rather than ohgosacchande fragments, the acrylamtde percentage m the gel should be reduced to 12-15% This should be achieved by addmg less acrylamtde solutton, mamtammg the volume of the mixture by addmg distilled water. 12 Excessive destaining tends to wash smaller ohgosacchartdes out of the gel. More sensitive detection can be achteved by counsterstammg with ammomacal silver (11,12) These two references also describe higher resolution separations which are obtamed through the use of gradient gels

References 1 Yanagtshtta, M , Mtdura, R J , and Hascall, V C (1987) Proteoglycan. isolation and purtflcattons from tissue culture, m Methods zn Enzymology, vol 138 (Gmsburg, V., ed.), Academic, Orlando, FL, pp. 279-289. 2. Bitter, T. and Muir, H M (1962) A modtfied uromc acid carbazole reactton Anal Bzochem 4,330-334 3 Carney, S L (1986) Proteoglycans, m Carbohydrate Analyszs-A Practzcal Approach (Chaplin, M F and Kennedy, J. F , eds ), IRL Press, Oxford, pp 97-142 4 Beeley, J. C (1987) Glycoprotezn and Proteoglycan Technzques Elsevter, Amsterdam.

Glycosaminoglycans

and Proteoglycans

143

5 Farndale, R W , Buttle, D J , and Barrett, A J (1986) Improved quantttatton and dtscrtmmatton of sulphated glycosammoglycans by the use of dtmethylmethylene blue Blochim Bzophys Acta 883, 173-177 6. Lmdhart, R. J , Turnbull, J E , Wang, H. M., Longanathan, D , and Gallagher, J T. (1990) Exammatton of the substrate spectfictty of heparm and heparan sulfate lyases Bzochemzstry 29,2611-26 17 7. McDevitt, C A and Muir, H (1971) Gel electrophorests of proteoglycans and glycosammoglycans on large-pore composite polyacrylamtde-agarose gels Anal Blochem 44,16 12-l 622 8 Hames, B D. (1990) An mtroduction to polyacrylamtde gel electrophoresis, m Gel Electrophoresls-A Practical Approach, 2nd ed. (Hames, B. D and Rtckwood, D , eds ), IRL Press, Oxford, pp 1-147 9. Hampson, I. N and Gallagher, J T (1984) Separation of radtolabelled glycosammoglycan ohgosacchartdes by polyacrylamide-gel electrophorests Bzochem J 221, 697-705

10. Silbert, J. E., Palmer, M E , Humphrtes, D E , and Silbert, C. K (1986) Formation of dermatan sulfate by cultured human skm Iibroblasts. J Bzol Chem 261, 13,397-13,400 1I Turnbull, J E and Gallagher, J. T. (1988) Ohgosacchartde mappmg of heparan sulfate by polyacrylamide-gradient-gel electrophorests and electrotransfer to nylon membrane Blochem J 251,597-608 12. Lyon, M and Gallagher, J T. (1990) A general method for the detectton and mapping of submtcrogram quanttttes of glycosammoglycan ohgosacchartde on polyacrylamtde gels by sequenttal stammg with azure A and ammoniacal stlver Anal Blochem 185,63-70.

The Use of Cell and Organ Culture for the Study of Secreted Mucins Anthony P. Corfield, Adil Aslam, Stephen Wood, Baldev Singh, and Christos Paraskeva 1. Introduction Mucms are major differentiated products of mucosal cells throughout the body and are thus important markers of normal and disease development. The use of human colomc cell lines as a model of the adenoma-carcmoma sequence is of particular interest because it allows the changes in expression of mucms to be studied during the development and progression of disease (1,2). Recently, the importance of proliferation, differentiation, and apoptosis has attracted attention to the use of culture systemsfor the study of cell behavior in normal and disease processes (3,4). In the same way, tissue obtained from patients at surgery or as biopsies can be placed m short-term primary or organ culture to study similar changes in disease ($6). Glycoconjugates, including glycoproteins, glycolipids, and proteoglycans are expressed by all mucosal cells as secretory, intracellular, and membrane components. The mucins are a family of high molecular weight glycoprotems, characteristically located in the vesicles of of Goblet cells m VIVO(7-9) and coded for by a family of genes, the MUC genes (10,11). Histochemical analysis of the mucm content of Goblet cells has indicated that these molecules are changed during many mucosal diseases (12,13) Biochemical analysis of the mucin changes may be limited because of the paucity of material available from mucosae in general and the difficulty in obtaining normal material for comparison (8,13). Improvements m the study of glycoprotems especially mucins have been achieved through: From

Methods E&ted

by

III Molecular E F Hounsell

Bology,

Vol 76 Glycoanalysrs

0 Humana

145

Press

Inc , Totowa,

Protocols NJ

Corfield et al. 1 The use of defined humanmucosalcells that can be grown in long-term culture (13,

2 Apphcatlon of metabolic labeling techmques(5,6,14), and 3 Improvements m separationand nonradloactlve detection systems(1.5-17 and see Chapter 11)

Radioactive tracer methods allow relatively small numbers of cells and tlssue fragments (biopsies) to be analyzed, and cell culture also gives access to larger amounts of the mucins produced by the mdlvldual cell lines (14,18). The mucins are very high molecular weight glycoprotems that aggregate and form gels on secretion at the mucosal surface (7-9,19). They contam typltally 70-80% of their dry weight as carbohydrate m O-glycosidic linkage to threonine and serme residues m the mucm polypeptide. Most of the polypeptide 1sresistant to proteolytlc attack because of the high level of glycosylatlon However, naked peptlde regions are susceptible to proteases and these contam the bulk of disulfide bridges present. Most mucms are made up of native macromolecules composed of subunits linked through dlsulfide bridges. Reduction and alkylation of these dlsulfide bridges allows separation of the individual subunits and together with proteolytlc dlgestlon results m the loss of vlscoelastic and gel forming properties (7-9,19). Separation methods for mucms have rehed on the properties of these molecules, typically then- buoyant density on density gradients, their high molecular weight on gel filtration, their charge on Ion-exchange chromatography, and combmatlon of molecular sizeandcharge on agarosegel electrophoresls(17,20). These methods have been apphed to mlcroscale radlolabeled mucms (6,14), and to larger amounts of mucms from cell culture or resected tissue (14,21) Although these separations will result m pure mucm fractions m most cases, contamination with proteoglycans and nucleic acids remains a posslblhty and must be analyzed. Identification and ehmmatlon of these contaminants may be necessary depending on the mucm data required. Accordingly, the type of experiment, either metabolic labelmg or direct cell or tissue extractlon, will each require the appropriate controls. The importance of the secreted mucins as major differentiated products of the mucosal cells can also be followed at the level of RNA expression. Sequence data for the MUC genes IS available and both zn sztu hybndlzatlon, and Northern analysis has been used to follow mucosal cell MUC gene expression m response to a variety of stimuli and in disease (22-24). The methods described here cover the detection, metabohc labeling and isolation of secreted mucms and analysis of MUC gene expression from cultured cells (primary cultures and cell lines), organ culture and macroscopic tissue samples. The data refer to human colorectal cells and tissue but slmllar systems and prmclples apply to a wide range of other tissueswhere mucins are produced.

Study of Secreted Mucins

147

2. Materials 1 Biopsies or trssue from surgery’ Place these on lens tissue soaked m culture medium at room temperature before mcubatlon Cut the tissue wtth a scalpel to a size of approx 2-4 mm*, taking care to remove all layers except mucosa and place on a steel grid m culture dishes 2. Fetal bovine serum (FBS)-batch testing IS essential 3. Cell culture media. a. Standard growth medium Dulbecco’s modified Eagle’s medium (DMEM) containing 2 nnI4 glutamme, 1 pg/mL hydrocortrsone sodium succmate, 0.2 U/mL insulin, 100 U/mL pemcdlm, 100 ug/mL streptomycin and 20% FBS. b Washing medmm* The same as the standard growth medium, but wtth 5% FBS, double the concentration of of pemctlhn and streptomycin and 50 ug/mL gentamycm c Digesting solution. The same as the washmg medium, but contammg 5% FBS together with 1.5 mg/mL collagenase (Worthmgton type 4) and 0 25 mg/mL hyaluromdase (Sigma, Poole, UK, type 1) d 3T3 Conditioned medmm. DMEM supplemented with 10% FBS, 2 n&’ glutamme, 100 U/mL pemctllm, 100 pg/mL streptomycm, and put onto 24 hpostconfluent 3T3 cell layers for 24 h. After condmonmg, the medium IS filtered through a 0.2 pm filter (Nalgene, Mtlton Keynes, UK) and further supplemented to give 20% FBS, 1 pg/mL hydrocortisone sodmm succmate, 0 2 U/mL insulin 4 Collagen (human placental type 4, Sigma) at 1 mg/mL prepared m 1 part glacial acetic acid to 1000 parts stenle tissue culture grade distilled water and, and stored at 4°C 5 Dtspase solution Dlspase (Boehrmger, Lewes, UK; grade 1) 1s prepared m DMEM containing 10% FBS, sterile filtered, and stored at -20°C 6. SWISS 3T3 cells from the American Ttssue Type Culture Collection (ATCC No CCL92) 7. Trypsin 0.1% by weight m 0.1% EDTA. 8. Acrldme

orange

(Sigma)

9 Organ culture medmm* Minimal Eagle’s medium, containing 10% FBS, 10 nuV sodium bicarbonate, 2 mA4 glutamme, 50 U/mL streptomycm, 50 U/mL pemclllm, 50 pg/mL gentamycm, and 20 mA4HEPES, pH 7.2. 10. PBS/rnhlbttor cocktail in PBS and 6M guanidine hydrochloride, 1 mA4 phenylmethylsulfonylfluortde, 5 nnI4 EDTA, 0 1 mg/mL soybean trypsm mhlbltor, 5 mMN-ethylmaletmide, 10 tibenzamldine, and 0 02% sodmm aztde Prepare inhibitor cocktarl fresh as required. Make up the solutton from stocks of concentrated guamdme hydrochlorrde as m step 13 11. Dlthlothreitol (Sigma) 12. Sodmm iodoacetamrde (Sigma). 13. Guamdme hydrochlortde, approx 7M stock solution m PBS, treated wtth charcoal, passed through filter paper and finally through a 0.45 pm Mlllipore (Watford, UK) filter The stock solution can be stored at room temperature and dllutlons prepared usmg PBS

148

Corfield

et al.

14. Sepharose CL 2B (Pharmacia, Mtlton Keynes, UK) 1 x 30 cm or 2.5 x 80 cm m all glass columns equilibrated m 4M guanidine hydrochloride/PBS or 10 m&J Trts/HCl, pH 8.0. 15 1% (w/v) Agarose gels for electrophoresis Agarose (SeaKem LE agarose, Flowgen, Stttmgbourne, UK) is made up at 1% m the running buffer. Gels 15 x 15 cm are run m a standard submarine horizontal electrophoresrs apparatus 16 Buffers for electrophoresis and vacuum blottmg. a Running buffer for agarose gel electrophoresis 40 mM-Tris-acetate 1 mM-EDTA pH 8 0 contaming 0 1% sodium dodecyl sulfate (SDS) (add from 10% SDS stock) b Sample buffer for agarose gel electrophorests 40 r&4-Tns-acetate 1 mA4-EDTA pH 8 0 contammg 0 1% SDS with 10% glycerol and 1% bromophenol blue c Vacuum blottmg buffer 3.3M sodium citrate pH 7.0 containing 3MNaCl. 17 Markers for electrophorests Rainbow markers, high molecular weight range (Amersham International, Amersham, UK) maximum 200 kDa (myosm) and IgM, 990 kDa (Sigma). 18 Immobllon P (PVDF) membrane (Milllpore, Watford, UK) 19. Periodic acid Schtff reagent (PAS) commercial solution (Sigma) 20 Precipitation buffer, 95% ethanol/l% sodium acetate cooled to -70°C 2 1 Proteoglycan degrading enzymes and mcubation buffers a Chondroitmase ABC from Proteus vulgarls (Boehrmger Mannheim, Lewes, UK) Incubation buffer, 250 mMTris-HCl, 176 mM sodium acetate, 250 r& sodium chloride, pH 8 0 b. Hyaluronidase from bovine testis (Sigma) Incubation buffer PBS c. Heparmase types II and III from Flavobacterzum heparrnum (Sigma), mcubation buffer 5 w-sodium phosphate, 200 mM-NaCl pH 7 0. 22 Sephadex GlOO (Pharmacta) Use 30 x 1 cm all glass columns equthbrated and run m 10 mM Trrs/HCl pH 8.0 23. Guamdmmm lsothtocyanate (Sigma) 24 RNA ladders (0.24-9.5 Kb) (Gtbco-BRL, Edinburgh, UK). 25 Prime-It (Stratagene, UK). 26 Buffers for ollgonucleotide hybridizatron: sodium chloride/phosphate/EDTA, 3.6M sodium chloride, 0.2M dibastc sodium phosphate, 0 02M EDTA standard saline/citrate, 3M sodmm chloride, 0.3M trlsodmm citrate, Denhardts solution (Sigma), Salmon sperm DNA (Sigma) 27. Hybond-N, (Amersham) 28. Hyperfilm-MP (Amersham).

3. Methods 3.1. Cell and Organ Culture (see Notes 1-3) 1. To prepare collagen-coated flasks, coat tissue culture flasks (T25 25 cm2) with a thin layer of collagen solution (Subheading 2., item 4; 0 2 mglflask), and allow to dry at room temperature in a lammar flow hood for 2-4 h (see Note 4). 2 Grow Swiss 3T3 cells (Subheading 2., item 6) on collagen on plastic tissue culture flasks in DMEM containing 10% calf serum until they are 24 h postconfluent

149

Study of Secreted Muans

3. Lethally irradiate the cells with 6OKGray (6mrads) of y radiation, or treat with 10 pg/mL mitomycin C (Sigma) for 2 h. 4 Wash the cells, and produce a single cell suspension by pipeting. The cells can either be used immediately as feeders or stored at 4°C as a single-cell suspension for up to 4 d (see Note 5)

3.1.1. Primary Culture-Enzyme

Digestion (see Notes 3, 6, and 7)

Wash the tumor specimens (adenoma and carcinoma) four times in washing medium (Subheading 2., item 3b) and cut with surgical blades to approx 1 mm m a small vol of the same medium Wash the tissue four times m washing medium, and place m digestton solution (Subheading 2., item 3c) Roughly 1 cm3 is put in 20-40 mL of solution Rotate at 37°C overnight (12-l 6 h). Mix the suspension by pipetmg to improve the separation of the epithelial elements from the stroma resulting from enzymic digestion. Filter the suspension through 50-mm mesh nylon gauze, or repeatedly allow to settle out under gravity and collect the pellets. The large clumps of cells and epitheha1 tubules (organoids that contain the maJority of the epithehal cells) are separated from the single cells (mostly from the blood and stroma) and cell debris Wash the cell pellets three times, and place m culture on collagen-coated T25 flasks m the presence of Swiss mouse 3T3 feeders (approx 1 x lo4 cells/cm4) at 37°C m a 5% CO2 m au incubator (25). In some situations, 3T3-conditioned medium can be used instead of adding mouse 3T3 cells directly to cultures (see Note 5)

3.1.2. Long-Term Culture of Adenoma Cell Lines (see Note 8) 1. Prepare culture conditions for adenoma cell lines as previously described for primary cultures 2 Pass the adenoma cells as clumps of cells using sufficient dispase Just to cover the cells, and incubate for approx 30 min at 37°C Remove the cells as a sheet, and ptpet to remove them from the flask and to break up the sheets mto smaller clumps of cells (25). 3. Wash the clumps of cells, and replate under standard culture condmons. Reattachment of cells may take several days, and during medium changing, any floatmg clumps of cells must be centrifuged and replated with the fresh medium.

3.7.3. Long-Term Culture of Carcinoma Cell Lines (see Note 8) 1 Grow the carcinoma cell lines m tissue culture plastics without collagen coating and wtthout 3T3 feeders m DMEM supplemented with 10% FBS and 1 mA4glutamme 2. Passage as single cells using 0 1% trypsin in 0 1% EDTA.

3.1.4. Apoptotic and Differentiating

Cells (see Note 9)

1. During routme culture of cell lines remove floating cells with the medium and pellet by centrifugatton Most cell lines give rise to floating cells, the majority of which undergo apoptosis.

Corfield et al

150

2. Stain the cells with 5 pg/mL acrrdme orange in PBS and examme mnnedtately by fluorescence

microscopy

to detect condensed, brtght chromatm

Count at least

300 cells 3, Extract DNA from 1O6cells and electrophorese on 2% (w/v) agarose gel contammg 0 1 pg/mL ethtdmm bromtde at 40V until the dye front has migrated 3-4 cm Run the DNA from an equivalent number of attached cells as a control, and use dexamethazone-treated mouse thymocytes as a positive control for DNA laddermg (see Note 9)

3.1.5. Organ Culture (see Note 2) Place the bropsres or tissue of 24 mm2, singly or up to SIX per dish on lens tissue placed over a stainless steel grid in culture dishes with a central well containing 2 mL of medium (Subheading 2., item 9) The orrentatron of the tissue 1s with the luminal surface uppermost

3.2. Collection of Secreted and Cellular Material 3.2. I. Collection of Radioactive Fractions After Metabolic Labeling in Cell Culture (see Note 10) 1 Add the radtoacttve precursor (e g , o-[3H]-glucosamme to the standard growth medium and incubate for times between 4 and 96 h, depending on the type of radioactive precursor and the nature of the cells (see Notes 11-13) 2 Collect the medium and wash the cells with a further 5 mL of fresh nonradioactive medium. 3 Irrigate the flasks with 5 mL of 6M guamdme hydrochlortde in PBS/mhtbitor cocktail (Subheading 2., items 10 and 13, see Note 14) contammg 10 mM dtthiothrettol and scrape the cells off wtth a cell scraper. 4 Wash the cells twrce with 6M guamdme hydrochloride PBSJmhtbitor cocktail containing 10 mM dtthtothrettol and pool the total washings. 5. Adjust the dithiothreitol washmgs to a 2 5X molar excess with sodium iodoacetamtde and incubate for 15 h at room temperature in the dark (see

Note 15) 6 Dialyze the secreted medium and dithtothrettol wash material extenstvely agarnst three changes of 6h4 guanidine hydrochlortde m PBS. 7 Homogenize the washed cell pellet in 1mL of 6M guamdme hydrochloride PBS/ mhtbttor cocktatl with an Ultraturrax for 10 s at maximum setting, on me. 8. Centrifuge the homogenate at 100,OOOg for 60 min, and decant the supernatant Resuspend the membrane fraction in 1 mL of 6Mguanidme hydrochloride PBS/ inhibitor cocktail

3.2.2. Collection of Radioactive Fractions After Metabolic Labeling in Organ Culture (see Note 10) 1. Add the radioactive precursor to the organ culture medium m the plastic culture dishes and place m an incubator at 37T in an atmosphere of 95% an or oxygen/

151

Study of Secreted Mucms

2

3

4 5

5% carbon dtoxtde Contmue the mcubatton for pertods of up to 24 h for colonic tissue (see Notes 11-13) After mcubatron, remove the medmm from the the central well and wash the tissue and dtsh with 1 mL of PBS Pool the medmm and washmgs, and dtalyze against 5 L of dtsttlled water, or three changes of 1 L of 6Mguamdme hydrochloride PBS over 48 h Homogenize the tissue m 1 mL of PBS/mhtbttor co&tall or 6M guamdme hydrochlonde PBS/inhibitor cocktatl m an all-glass Potter homogemzer ensurmg complete disruption of the mucosal cells (about 20 strokes) Remove connective tissue tf present Centrifuge the homogenate at 12,000g for 5 min, and separate the supernatant soluble fraction from the membrane pellet Resuspend the membrane pellet m 1 mL of 6A4 guamdme hydrochlortde PBS/ * mhtbltor cocktail.

3.2.3. Collection of Fractions from Nonradioactwe

Cell Cultures

1, Remove the medmm from the flasks and scrape the cells and gel layer mto PBS/ mhtbttor cocktail contammg 6M guamdine hydrochloride 2 Sediment the cells and cell debrts by centrlfugatton at 100,OOOg for 30 mm and aspnate the the soluble and gel layers from the pellet 3 Solubtltze the gel layer by contmued sturmg m 6M guanidme hydrochloride at 4°C with 10 mA4 drthtothrettol and 5 n&f EDTA for 15 h at 37°C followed by the addttton of a 2 5 molar excess of rodoacetamtde over dtthtothrettol and Incubate at room temperature for 15 h m the dark (see Note 15). 4 Centrrfuge soluble fractions for 60 mm at 100,OOOg and discard the pellets

3.2.4. Collection of Fractions from Nonradioactive Specimens

Surgical Tissue

1, Pin out the surgical specimen if possible on a dtssection board and irrigate with 6M guamdme hydrochloride PBS/mhibitor cocktail 2. Scrape the mucosal surface usmg a glass slide and wash the combined scrapings into PBS/mhrbttor cocktail contammg 6M guamdme hydrochlortde 3. Homogemze the scrapings m an all glass Potter homogenizer to achteve complete disruption of the mucosal cells, and centrifuge the suspenston at 100,OOOg for 60 mm to sedtment all membranes 4. Separate the soluble and gel (If present) fractions and ensure solubilizatton of the gel samples usmg dithtothrettol and todoacetamide as described in Subheading 3.2.3., step 3 (see Note 15)

3.3. Separation of Mucins from the Fractions After Culture (see Note 76) 3.3.1. Density Gradient Centrifugation

Obtained

1. Make up the samples from the fractrons prepared m Subheadings 3.2.1.-3.2.4. m 4M guamdme hydrochlortde/PBS to a concentration of approx l-5 mg/mL or

Corfield et a/.

152

2. 3

4

5.

containing a suitable amount of radloactlvlty (e.g., >lO,OOO cpm) for subsequent analytical techmques. Add sohd CsCl to give a density of about 1 4 g/mL and stir at room temperature for 15 h Load the samples into centrifuge tubes and centrifuge at 100,OOOg for 48 h at 10°C to obtain a CsCl density gradient. Aspirate 0.5 mL samples from the top of the tube by plpet, or drain from the bottom after piercing the tubes. Weigh the samples to obtain the density of each fraction Slot blot ahquots (5-50 pL, after dllutlon If necessary) of each fraction onto Immobllon P membrane and visualize with a carbohydrate stain (see Note 17) (IS) Quantify the results usmg a densitometer (16) For the radioactive samples each slot blot 1s cut out and and placed m scintillation cocktail for quantltatlon (see Note 18). Pool the carbohydrate contammg fractions located at densltles between 1 30 and 1 55 g/mL

3.3.2. Gel Filtration (see Notes 16 and 19) 1 Make up samples m 10 mA4Tris-HCl pH 8 0 or 4M guamdine hydrochloride/PBS buffer to give concentrations of l-5 mg/mL or by radloactlvlty (e g , >5000 cpm), and load onto columns of Sepharose CL 2B (Subheading 2., step 14) Elute the column with the same buffer and collect fractions (l-5 mL). 3 Pool the mucm containing fractions Identified in excluded or included volumes.

3.3.3. Agarose Gel Electrophoresis and Vacuum Blotting (see Notes 20 and 21) 1 Mix samples containing 10-500 pg mucin or >lO,OOO cpm with 50 @ sample buffer and load onto honzontal 1% agarose gels Rainbow markers and IgM are run as migration markers (see Note 21). 2. Run the gels at 20V for 18 h at room temperature. 3. Remove the gels after electrophoresis and stam directly using Coomassle blue stam (see Note 22) or submit to vacuum blottmg. 4. Gels are blotted onto Immobllon P membranes using a standard apparatus m vacuum blotting buffer for 2 h at 40mbar. The success of the transfer can be visualized immediately If rambow markers are mcluded on the gels. 5. Probe the blot membranes with chemical stains (e.g., PAS), lectm conjugates or antibodies using standard techmques for vlsuahzatlon (see Note 22).

3.3.4. Concentration (see Note 23)

of Mucin Samples After Purification

1 Dlalyse the samples against four to five changes of 5 L of distilled water or against two changes of 1 L of 4M guamdme hydrochloride. Salt-free samples can be freeze dried to concentrate (see Note 23). 2. Mix a solution containing mucin (l-5 mg or >lO,OOO cpm) with 4 vols of preclpltation buffer (Subheading 2., item 20) cooled to -70°C and leave for 45 mm at

Study of Secreted Mucins

153

-70°C Return to room temperature for centrtfugatton at 12,000g for 20 mm and collect the pellet

3.3.5. Identification and Removal of Proteoglycans (see Notes 24 and 25) 1. Mix samples contammg 0 5-2.0 mg mucm or >lO,OOO cpm with. a. Chondroitinase ABC (5 U/mL) and incubate for 16 h at 37°C m mcubatton buffer (Subheading 2., item 21a), b. Hyaluronidase ( 10 mg/mL) and incubate for 16 h at 37°C in mcubatton buffer (Subheading 2., item 21b); and c Heparinase II and III ( 100 mU of each enzyme) and incubate for 16 h at 37°C m mcubatton buffer (Subheading 2., item 21~). Carry out control mcubattons under the same condittons wtthout enzyme. 2. Stop the mcubations by addition of 1 mL of 10 mMTrts-HCl, pH 8 0 and load the products onto columns of Sephadex GlOO (Subheading 2., item 22), elutmg with the same buffer and collectmg 1 mL fractions. 3 Test the individual fractions for radtoacttvity or carbohydrate by slot blotting or by colorimetrrc analysis (see Note 25)

3.4. Analysis of MlJC gene Expression in Cell Cultures (see Notes 26 and 29) 1. Extract the total RNA from approx lo* cells by single step preparation using guamdmmm tsothiocyanate-phenol-chloroform (see Note 27) 2. Load 20 pg total RNA on a 0 9% denaturmg agarose gel in the presence of 3% formaldehyde. Load 4 pg RNA ladders (0 24-9.5 Kb) for size reference on the same gel. Run at 150V for 3 h or until the Bromophenol Blue loading dye has run at least 8 cm down the gel 3 Transfer the separated RNA onto nylon membrane (Hybond-N, Amersham, Subheading 2., item 27) by captllary blotting, overmght Fix the RNA by heating m an oven at 80°C for 2 h (see Note 28). 4 Prepare, label, and purify 25 ng of suitable MUC gene probe with [32P]dCTP to give approx 11GBq/mol (see Note 26) 5. Prehybrtdrze the membranes m 50% formamtde, 5X sodium chlortde/phosphate/ EDTA, 0.3% SDS, 5X Denhardts solutton and 200 pg/mL salmon sperm DNA (Subheading 2., item 26) and mcubate at 65°C overnight (16-22 h) 6 Hybridize the membranes with 1.85 MBq of the radtolabeled cDNA probe m the same buffer at 65°C overnight (16-22 h) 7 Wash the hybndized membranes under high stringency conditions (Subheading 2., item 26), two washes with 2X standard salme/cttrate, 0.1% SDS 30 mm at room temperature, one wash with 0.1% standard saline/citrate, 0 1% SDS for 1 h at room temperature, one wash with 0.1% standard saline/citrate, 0 1% SDS at 55-65Y for 30 min. 8 Expose the membrane to Hypertilm-MP (Amersham; Subheading 2., item 28) with an mtensifymg screen at -70°C for at least 24 h

154

Corfield et al

4. Notes The production and secretion of mucus glycoprotems by cultured colomc cells should be examined using cells at different stages of confluency, because this may alter the differentiation properties of the cells and, therefore, the amount and type of mucm produced. When using organ and primary cultures, it 1s Important to consider the heterogeneous nature of the cell types m the tissue (1 e , stromal elements and lymphoid cells m addition to the colomc epithelial cells) It may not be clear which cell type is producing the glycoprotems The primary culture techniques described can be used for normal adult colon However, these are not as reproducible as those with the adenomas and carcmomas, and there are more problems from contammatmg stromal elements There are at present no normal adult colomc epithelial cell lines, only adenoma and carcinoma cell lines (1,26) Collagen-coated flasks are necessary to obtain efficient attachment of primary cultures and some adenomas and carcmomas to the flasks, and to retain the optimum differentiated characteristics of the cells (25) The use of 3T3 cell feeders requires controls to determine which cell type 1s producing the glycoprotems of interest This can be achieved using 3T3-conditroned medmm where mucm production is being assessed, or removing the 3T3 feeders from the flask once the eprthelmm has grown Colorectal adenomas mvarrably need digestion with enzymes because of their organization into well-differentiated glandular structures With carcmomas, it IS possible to adopt a nonenzymic approach with surgical blades to release small clumps of tumor cells that can be cultured (27) When using colomc cell lines it 1s Important to check the true colomc epithelial nature using a battery of markers, mcludmg antikeratm antibodies, ultrastructural analysis showing the presence of desmosomes, and other colomc differentiation markers (26). Although many tumor cell lines, especially colon carcmomas, can be grown m sample media without 3T3 feeders and without collagen coatmg, the colomc cells retam better differentiated phenotypes when using the more complex culture conditions described for primary cultures Prohferatmg cells m the bottom of the colomc crypts migrate to the upper half of the crypt where they differentiate These differentiated cells migrate to the top of the crypt and there is evidence that they die by apoptosrs and that apoptosis may be the termmal stage of differentiation. The relationship between proliferation, differentiation, and apoptosis may be studied using the culture system described here We have shown that cultured

colomc

normal

adenoma

and carcmoma

cells

spontaneouslydie by apoptosism vitro and the levels of apoptosiscan be modulated by dietary short chain fatty acids (butyrate, acetate, and propionate) and bile acids(4). During routme culture of colomc epithehal cells, somecells detach from the flask and float m the medium (3) These cells contam condensedchromatm that can be detected with acridine orange stammg In addition, characteris-

Study of Secreted Mums

10.

11

12

13.

14

15.

16.

17

155

tic DNA laddermg resulting from mternucleosomal fragmentation can be seen after analysis of total cellular DNA. Dexamethazone-treated mouse thymocytes are a convenient source of cells to use as a positwe control for DNA laddermg In radlolabelmg experiments, the total amount of mucm is often small, and slgmficant losses owing to nonspecific adsorption onto plastic and glass vessels, silicone rubber, and dialysis tubing may occur Treatment of all vessels and tubmg with 1% Trlton m PBS before use improves yields The choice of radloactlve precursor is important Radloactlve glucosamme is most commonly used because It IS incorporated mto N-acetyl-o-glucosamme, N-acetylo-galactosamine and the slallc acids, major monosaccharide components of mucms In tissues other than hver such labels as mannose and fucose may be randomized to other monosaccharides before they are transferred to glycoprotems The metabolic rate of the cells should be considered for optimum labeling with the radioactive precursor. The type of isotope will govern the the amount to be added, typically [14C]- and [35S]-precursors m the range 185-l 850 kBq/expenment and [3H] precursors m the range 370-1850 kBq/expenment Short-term labeling of 2-4 h may not result m labeling of secreted material and could require higher doses of radloactlve precursor (1 85-3 7 MBq/expenment) Longer labelmg periods may reflect synthesis, catabolism, and recycling of glycoprotems. Dual labeling expenments with, e g , [35S] and [3H] need to be planned such that the relative mcorporatlon of each Isotope 1sreadily detectable m the isolated product, thus, conslderatlon of Notes 11 and 13 IS necessary Organ culture expenments should be controlled by histochemical criteria to ensure the integrity of the tissue during the mcubatlon period Diseased tissue may show signs of degradation during acceptable culture times for normal samples Increased mcorporatlon of radloactlve precursors may be achieved by reduction of the concentration of the same nonlabeled compound m the medmm (monosaccharides or ammo-acrds) for the labeling penod. However, this should be balanced against any changes m the growth of the cells or tissue under these “depleted” condltlons A protease mhlbltor cocktail is needed to avoid the degradation of mucms by bacterial enzymes and m cell homogenates Soybean trypsm inhibitor and PMSF are the most important for colomc tissue The intestinal mucms are present as adherent gels that are not always soluble m concentrated guanidme hydrochloride alone (28) In order to achieve complete solution reduction and alkylatlon 1s necessary This leads to the formation of mucm subunits that can be ldentlfied on agarose gel electrophoresls (2 7) The sequence of purification steps m mucm purlficatlon 1sslgmficant If density gradient centrlfugatlon IS followed by gel filtration, the lower molecular weight subunits or degradation products may be identified The use of a general carbohydrate stain IS useful to detect mucms on slot blots The PAS stain can be used and sensitivity 1simproved on the membranes as salt 1s eliminated (see Note 18) Lectms can also be used as general probes Wheat Germ Agglutmm-horse radish peroxldase conjugate has been found to be a satlsfactory and sensitive probe for mucms on slot blots (29)

156

Corfleld et al.

18 Owing to high salt concentrations m density gradient centrlfugatlon experiments many colortmetrlc assays and some radloactlve scmtlllation cocktails are mefficlent. Extensive dialysis of each fraction may allow colonmetrlc assays to be performed, but with small amounts of metabohcally labeled material this results m significant losses (see Note 10) The slot blottmg technique (Note 17) 1s more rehable and sensitive for calorimetric and radloactive detection 19. Gel filtration 1sthe most rapld and convenient method to obtain a mucm-enriched fraction from crude culture samples for comparative studies (5,6) It 1s also a starting point for preparation of native mucms for further purification and analySIS (see Note 16) Automated FPLC systems can also be used to analyze samples in the same way, but are not as flexible m preparation of larger mucm fractrons. 20 The separation of mucms on agarose gel electrophoresis allows the largest mucm molecules to be analyzed without the uncertainty of SDS-PAGE systems where mcomplete mlgratlon of the sample mto the gels 1s common and quantltatlve results are dlfflcult to obtain (17) 21 Markers for agarose gel electrophoresis reflect the separation of molecules on a basis of their size and charge This is m contrast to the normal condltlons used for proteins and some glycoproteins on SDS-PAGE. Accordmgly, the use of markers on agarose gel electrophoresls can give only an estimate of relative mlgratlon and not of molecular size 22 Protein stains for mucins are usually very poor Coomassle blue and sliver stains frequently give negative results. Chemical stains for carbohydrate, lectm conjugates, or specific antlbodles are the most useful. 23 Where possible mucm samples should be kept m solution Preferably m 4M guanidine hydrochloride. Desalting and/or freeze drying may result m the production of a residue that cannot be resolublhzed Concentration of mucm samples may be difficult, and where samples are treated by such methods, assessment of mucm loss IS advisable. Mild precipitation methods at reduced temperature or with specific antlbodles for smaller samples are best, but reqmre suitable controls for efficiency. 24 The proportion of proteoglycan m colomc cell and tissue samples 1s normally low, but in cases of cell transformation or selection of subclones (30), care must be taken to control high molecular weight material with a buoyant density m the range 1.35-l .60 g/mL for the presence of proteoglycans. In metabohc labeling experiments the presence of cells having a high turnover rate for proteoglycans, e.g., fibroblast hyaluronan synthesis may significantly add to the proportlon of labeled material isolated m “mu&’ fractions Both of these situations require the analysis of suspected mucm products usmg enzymlc degradation 25 The assay of carbohydrate breakdown products of proteoglycan degradation must be carried out using liquid assay systems as these products are of low molecular weight and will not be detectable using membrane blotting techniques Assay for total carbohydrate or hexosamine IS appropriate. 26. Suitable probes for the analysis of the major MUC genes so far detected in the human colon have been described (10) These are MUC 2, MUC3 (II), and MUC

Study of Secreted Mucins

157

4 (31). Labeling with [35P]dCTP can be carried out routmely usmg commercially available kits, e g., Prime-It (Stratagene), followed by routme purtficatton (23) The polydtspersity of the human MUC gene messages often yields smears rather than sharp bands on Northern analysts 27. Total RNA can be readily and rapidly prepared from cultured cells using single step extraction m guamdmmm isothiocyanate-phenol-chloroform (32) 28. The use of Hybond-N allows satisfactory transfer and fixing of RNA on membranes without the need for a vacuum oven, an otherwise expensive item of equipment. 29 The condttions for the prehybndization and hybrtdtzation may vary, the conditions quoted here are based on an earlier method (23), and our current applications with the probes detailed before (21,23).

References 1 Wtlltams, A. C., Browne, S. J., Manning, A M., Hague, A , van der Stappen, J W J , and Paraskeva, C (1993) Btologtcal consequences of the genetic changes which occur during human colorectal carcmogenests Sem Cancer Bzol 4, 153-159 2 Wtlliams, A. C , Harper, S A , and Paraskeva, C (1990) Neoplastic transformation of a human colomc epithehal cell line* experimental evidence for the adenoma to carcmoma sequence Cancer Res 50,4724-4730 3. Hague, A., Manning, A M., Hanlon, K , Huschtscha, L I , Hart, D., and Paraskeva, C (1993) Sodium butyrate induces apoptosts in human colonic tumour cell lines m a p53 independent pathway: imphcations for the possible role of dietary fibre m the prevention of large bowel cancer Znt J Cancer 55,498-505 4 Hague, A , Elder, D. J E , Hicks, D. J., and Paraskeva, C (1995) Apoptosts in colorectal tumour cells mductton by the short chain fatty acids butyrate, proptonate and acetate and the bile salt deoxycholate Int J Cancer 60,400-406 5. Corfield, A. P., Warren, B. F, Bartolo, D. C C , Wagner, S A, and Clamp, J R. (1992) Mucin changes m ileoanal pouches monitored by metabolic labelling and histochemtstry. Br J Surg 79, 1209-1212 6 Prober-t, C. S J , Warren, B F , Perry, T , Mackay, E. H., Mayberry, J F., and Cortield, A P (1995) South Asian and European cohtics show charactertsttc differences m colonic mucus glycoprotem type and turnover Potential identification of a lower risk group for severe dtsease and cancer Gut 36,696-702 7. Strous, G J. and Dekker, J (1992) Mucm-type glycoprotems. Crzt Rev Bzochem A401 Blol 27,57-92 8. Forstner, J. F. and Forstner, G. G (1994) Gastromtestmal mucus, m Physzology of the Gastrolntestlnal Tract (Johnson, L R., ed ), Raven, New York, NY, pp. 1245-1283 9. Allen, A , Hutton, D. A., Pearson, J. I?, and Sellers, L. A (1990) The colomc mucus gel barrier: structure, gel formation and degradatron, m The Cell Bzology of Znjlammation in the Gastrorntestznal Tract (Peters, T J , ed ), Corners Pubhcattons, Hull, UK, pp 113-125. 10 Gendler, S. J. and Spicer, A. P. (1995) Epithehal mucin genes. Ann. Rev. Physzol 57,607-634

158

Corfield et al

11 Gum, J R (1992) Mucm genes and the proteins they encode structure dtversity and regulatton Am J Respw Cell Mol Blol 1,551-564 12 Fihpe, M. I (1989) The hrstochemtstry of human Intestinal mucms changes m disease, m Gastrowtestmal and Oesophageal Physzology (Whitehead, R , ed ), Churchill Livmgstone, Edinburgh, pp. 65-89. 13 Jass, J. R and Roberton, A M (1994) Colorectal mucm htstochemtstry m health and disease* a crtttcal review Path01 Int. 44,487-504. 14 Vavasseur, F., Dole, K , Yang, J , Matta, K. L , Corfield, A. P, Myerscough, N , Paraskeva, C., and Brockhausen, I. (1994) 0-glycan biosynthesis m human colomc cells durmg progression to cancer Eur J Bzochem 222,4 15-424 15 Thornton, D J., Carlstedt, I , and Sheehan, J K. (1994) Identification of glycoprotems on mtrocellulose membranes and gels, m Basic Protein and Peptlde Protocols (Walker, J M , ed ), Humana Press, Totowa, NJ, pp 119-128 16. Thornton, D J , Holmes, D F , Sheehan, J K., and Carlstedt, I (1989) Quantttatton of mucus glycoproteins blotted onto mtrocellulose membranes Anal Blochem 182, 160-164 17 Thornton, D J , Howard, M , Devme, P. L , and Sheehan, J K (1995) Methods for separation and deglycosylatlon of mucin subunits Anal Bzochem 227, 162-l 67 18 Cortield, A P , Clamp, J R , Casey, A D , and Paraskeva, C (1990) Characterization of a siahc acid-rich mucus glycoprotem secreted by a premahgnant human colorectal adenoma cell line /nt J Cancer 46, 1059-l 065 19 Carlstedt, I., Sheehan, J K., Corfield, A P., and Gallagher, J T (1985) Mucous glycoprotems a gel of a problem Essays Bzochem 20,40-76 20 Carlstedt, I and Sheehan, J K (1984) Macromolecular properties and polymeric structure of mucus glycoprotems, m Mucus and Mucosa Pttman, London, pp 157-172 2 1 Thornton, D. J , Devme, P L , Hanski, C , Howard, M., and Sheehan, J. K (1994) Identtficatton of two maJor populations of mucms m respiratory secretions. Am J Resplr Crlt Care Med 150,823-832. 22 Audle, J. P., Janm, A , Porchet, N , Copm, M. C , Gosselm, B , and Aubert, J. P (1993) Expression of human mucm genes m respiratory, digestive and reproductive tracts ascertained by m sttu hybridization J Hzstochem Cytochem 41, 1479-1485. 23. Ho, S. B , Ntehans, G A, Lyftogt, C , Yan, P S , Cherwitz, D L., Gum, E T, Dahiya, R., and Kim, Y. S (1993) Heterogeneity of mucm gene expressron m normal and neoplastic tissues Cancer Rex 53, 64 l-65 1. 24. Myerscough, N , Warren, B. F , Gough, M., and Corfield, A P (1995) Expression of mucm genes m ulcerative colitis Blochem Sot Trans 23, 5368 25 Paraskeva, C , Buckle, B G., Sheer, D , and Wtgley, C B (1984) The isolatton and characteristtcs of colorectal epnhehal cell lines at different stages m malignant transformation from familial polypos~s coli pattents Int J Cancer 34,4!9-56 26. Labotsse, C L. (1989) Differentiation of colon cells m culture, m The Cell and Molecular Bzology of Colon Cancer (Augenhcht, L H , ed ), CRC Press, Boca Raton, FL, pp 2743

Study of Secreted Mucms

159

27. Leibovitz, A , Stmson, J C , McComb, W B , McCoy, C E., Mazur, K C., and Mabry, N D (1976) Classification of human colorectal adenocarcmoma cell lmes Cancer Res 36,3562-3569 28 Asker, N., Baekstrom, D , Axelsson, M A. B., Carlstedt, I , and Hansson, G C (1995) The human MUC2 mucm apoprotem appears to drmerize before O-glycosylation and shares epitopes with the “insoluble” mucm of rat small mtestme Blochem J 308,873-880 29 Ayre, D , Hutton, D A , and Pearson, J P (1994) The use of wheat germ agglutmm to improve bmdmg of heterogeneous mucm species to nitrocellulose membranes Anal Blochem 219,373-375. 30. Ogata, S , Chen, A , and Itzkowitz, S H (1994) Use of model cell lmes to study biosynthesis and biological role of cancer-associated sialosyl-Tn antigen Cancer Res 54,4036-4044 31 Porchet, N , Nguyen, V C , Dufosse, J , Audie, J. P., Guyonnet-Duperat, V, Gross, M. S , Dems, C., Gand, P, Bemherm, A , and Aubert, J P (1991) Molecular clonmg and chromosomal locahzation of a novel human trachea-bronchial mucm cDNA contammg tandemly repeated sequences of 48 base pan-s Bzochem Blophys Res Commun 175,414-422 32 Chomczynski, P. and Saccht, N (1987) Single-step method of RNA isolation by acid guanidme thiocyanate-phenol-chloroform extraction. Anal Bzochem 162, 156-159

Purification of Gastrointestinal Mucins and Analysis of Their O-Linked Oligosaccharides Barry J. Campbell

and Jonathan

M. Rhodes

1. Introduction The role of mucus m the protection of intestinal, bronchial, nasopharyngeal, and cervical mucosae has long been recognized, but poorly understood. Research has been stimulated by the expectation that an underlying mucus abnormality might be present, not only in diseasessuch as cystic fibrosis where there is an obvious physical alteration in mucus, but also in conditions such as inflammatory bowel disease, peptic ulceration, and intestinal cancer where mucus abnormalities at present seem more subtle (1). Progress has been relatively slow, however, not least because of the difficulties m obtaining a pure but undamaged mucus glycoprotem (mucm) preparation. Intact mucm molecules are very large (MW l-20 x lo6 Da), form gels and are heavily O-glycosylated (accounting for over 70% of their dry weight), and are often highly charged because of sialylation and/or sulfation. As a result they stick to chromatographic media, will only enter very low percentage polyacrylamide gels, and are often heavily contaminated with other proteoglycans, DNA, and cellular debris. The choice of mucm purification technique depends on the purpose of the study. If simple quantificatton is the aim then a single separation through a small gel filtration column (e.g., Sepharose CL-4B or CL-2B) may be quite adequate (2,3). In studies undertaken to assesstertiary structure, the requirements are much more stringent with a need not only to achieve extremely high purity but also to avoid degradatron. In that case density gradient centrtfugation m the presence of guanidmmm chloride 1sthe preferred technique (4). We have been particularly Interested in the changes in oligosaccharide side-chain structure that may occur m mflammatory and cancerous intestinal disease. For From

Methods Edlted

by

m Molecular E F Hounsell

Bology,

Vol 76 Glycoanalysrs

0 Humana

161

Press

Inc , Totowa,

Protocols NJ

162

Campbell and Rhodes

these studies a very high degree of purtty was essential, but preservation of mucm terttary structure was less important. Moreover, if a reasonable number of samples are to be studied to allow comparison between disease and controls the method needs to be reasonably straightforward. We describe here a one-step htgh performance gel filtratton process that we have found to gave consistently htgh mucm purity with a much clearer separation of mucm from nonmucin components than can be achieved by conventional gel filtration. An enzyme-linked mucm assay is described that allows reproducible mucm quantification. Two different processes are then described for release of O-linked oligosaccharide chains of mucins. The first 1s a conventional alkaline-borohydride hydrolysis that releases all O-linked chains as ohgosaccharide alditols that can then be separated by htgh-pH amon exchange chromatography (HPAEC) This techmque allows ready profilmg with separation based on oligosaccharide alditol size, charge, and glycostdtc linkage-type, but with most mucms will yield at least 20 different alditols, tdentitication of which will be demandmg. The second process described, uses a specific enzyme (endo-a-N-acetylgalactosamuudase [O-Glycanase]) to release a specific O-lurked oligosaccharide (m this case galactose pl-3 N-acetylgalactosammea-) that can be readtly identified and quantified to address a specific question. We used this technique to resolve the controversy about the increased expression of this structure as an oncofetal antigen in colon cancer (5).

2. Materials 2.1. Sample Preparation 1. Phosphate-buffered saline: 0 OlM sodium phosphate-buffered 0 14M sodium chlorrde, pH 7 4 2 Protease inhibitors Aprotnnn (store at 4”C), leupeptm (store at -2O”C), benzamidme hydrochloride (Hazard: Toxic) (store under nitrogen at 4°C) and thimerosal (Hazard: Toxic, mercuric poison) Prepare inhibitor cocktail fresh Just before use 3. Alumuuum mesh (W David & Sons, Wellmgborough, UK) 4 Culture medium* RPM1 contauung 2 mM glutamme, 10% (v/v) fetal calf serum, 100 pg/mL gentamicm (Hazard: Teratogen) (store at 4°C) and 60 U/mL nystatm (store at -20°C). 5 Falcon Organ tissue culture dishes, 60 x 15 mm style with center well (Becton Dickinson, Lincoln Park, NJ) 6 Incubator with protected cncuitary to permit use of 95%02/5%C02. 7 Ultrasonrc dismtegrator (with variable wave amplitude) (Caution: Ultrasonic device; use ear protection ) 8. A Polytron homogenizer (Kinematica, Kriens-Luzem, Switzerland) 9 PDlO Sephadex GM25 gel columns (5 x 1 6 cm, Pharmacia, Uppsala, Sweden) for rapid desalting and buffer exchange.

Analysis of Gastromtestinal Mucins

163

10 Freeze-dryer 11 Metabolic labeling studies 250 @I N-acetyl-D-[ 1-3H] glucosamme (2-10 Cl/ mmol) (Hazard: Radioactive source) Divide into 2 pCi ahquots and store at 4OC 1 mC1 [JSS]-sulfate (250-1000 mCl/mmol) (Hazard: Radioactive source); store at 4°C 12 A liquid scintillation counter, scmtlllation fluid for aqueous samples, and polypropylene scintillation vials. 13 Bicmchoninic acid (BCA) protein estimation kit (Sigma, Poole, UK) 14. Spectrophotometer (UV and vlsable absorption wavelengths)

2.2. Purification

of Crude Mucus Glycoproteins

(Much)

1 Chloroform:methanol(2: 1 v/v) (Hazards: Suspected carcinogen/Toxic and volatile; use m a solvent fume cupboard). 2. Collagenase (EC 3.4.24 8), ovme testes hyaluromdase (EC 3 2 1.35), and protease-free chondroltm ABC lyase (Proteus vulgms, EC 4 2.2 4) (Boehrmger Mannhelm, Lewes, UK) 3 Phosphate-buffered salme, pH 7 4: 0 02Msodmm phosphate, pH 5 8; 0 OlMTrlsHCI, pH 8 0 4. Gel filtration Sepharose CL-2B chromatography media (Pharmacla), empty PDlO columns, bottom and top filters (Pharmacla) (see Note 1) Store packed columns at 4°C in 0.2 pm filtered, degased elutlon buffer containing 0 02% sodium azlde (Hazard: Toxic) 5. Tween-20 (Polyoxyethylene sorbltan monolaurate). 6 Glass filter-funnel (porosity ~10 pm) 7. O.lM Tris-HCl, pH 8.0 8 Superose 6 prepacked HRl0/30 FPLC column (Pharmacla) Store in 0 2 pm filtered, degased buffer containing 0.02% sodium azlde. 9. A boroslhcate glass pump fast protein hqmd chromatography (FPLC) system (Pharmacla). 10 Antibodies Antimucm (colonic/pancreatlc) monoclonal antibody CAM 17 1 (Euro DPC, Llanbens, UK) (see Note 2) Peroxidase-conjugated rabbit antimouse Immunoglobulins (Dako, High Wycombe, UK) Store m dark at 4°C 11. Flat-bottomed microtlter plates (Type; Imm 4, cobalt irradiated) (Dynatech, ChantlIly, VA). 12 Carbonate coating buffer 35 mA4 sodium bicarbonate- 15 mA4 sodium carbonate, pH 9.0. 13. ELISA phosphate-buffered salme (PBS): 8 g/L NaCl, 0.2 g/L KH2P04, 1 15 g/L Na2HP04, 0 2 g/L KCl, pH 7.2 containing either 0.1% Tween-20 or 0.05% (v/v) Tween-20 and 0 05% (w/v) bovine serum albumin 14 10 mg Ortho-phenylenedlamme dihydrochloride (OPD) tablets (Hazard: Toxic, carcinogen); store at 4°C 15 30% w/w H202 (Hazard: Corrosive oxidizing agent) 16 2X 0 05M Sodturn citrate-o. 1M phosphate pH 5 0; Store at 4°C and dilute 1: 1 with deionized water when required

Campbell and Rhodes

164 17. 4M H2S04. 18. ELISA plate reader (49&492

nm filter).

2.3. Isolation of Mucin O-Linked Oligosaccharicie by Alkaline-Borohydride Degradation

Alditols

1 Sodmm borohydrtde (NaBH4) (Hazard: Harmful to respiratory mucous membranes) Prepare 2MNaBH4 in 0 1MNaOH fresh on day of use. Trmated NaB3H4 (10 Ci/mmol, NEN, Stevenage, UK) (Hazard: Radioactive) can be used if radiolabelmg of 0-glycans IS required durmg the P-ehmination reaction, use 2M NaB& containing 25 mC1 NaB3H4 Store tritiated borohydride m 5 mCi ahquots at -70°C m 10 mM NaOH 2 Glass 3 mL Reactt-vial with Teflon-lmed screw caps (Pierce, Rockford, IL) 3 50% v/v glacial acetic acid m HPLC water. 4 HPLC grade water (Rathburn Chemicals, Walkerburn, Scotland) 5 Dowex 50-X 12 H+-form resm (Sigma) 6 Fume cupboard (for solvent use) 7 Methanol (AnaLaR grade) (BDH, Poole, UK) (Hazard. TOXIC) 8. A vacuum-centrifugal evaporator 9 0.2 pm Anotop-IC sample filters (Whatman, Maidstone, UK)

2.4. Release of the Core 1 Disaccharide from Mucins

(Galpl-3GalNAca-)

1 Bio-Spm chromatography columns, packed with 0 8 mL Bio-Gel P30 polyacrylamide gel matrix (Bio-Rad, Hercules, CA) (see Note 3). Store at 4°C m 0 15M sodmm chloride- 17 5 mM sodium citrate pH 7 0 contammg 0 02% (w/v) sodium azide as preservative 2. 0-Glycanase Streptococcus (Dlplococcus) pneumomae endo-a-N-acetylgalactosammidase (EC 3 2 1.97) (Oxford Glyco, Abmgdon, UK) (see Note 4) Store at -20°C for up to 6 mo, but avoid repeated freeze-thawing. 3 0. 1M Sodtum citrate-phosphate pH 6 0, make up with HPLC grade water 4 Dowex 50-X12 H+-form resin. 5 Flat-bottomed microtiter plates (Dynatech) and ELISA plate reader (490-492 nm filter). 6. D-Galactose (Sigma) 7. 4% w/v Phenol m deiomsed water (Hazard: TOXIC, potential carcinogen) Handle m a solvent fume cupboard. Prepare fresh on day of use, store at 4°C 8. Concentrated H2S04 (Hazard: Corrosive; handle wtth great care). 9 Fast-delivery multichannel pipet

2.5. Removal

of Mucin Terminal Sugar Residues

1 0 05M H,SO+ 2 Glass 1 mL Reacti-vials

with Teflon-lined screw caps (Pierce, Rockford,

IL)

Analysis of Gastrointestinal

Mucins

165

3. Sapomficatton reagents, 0.05MNaOH and 0 lMH,SO+ Clostrzdlum perfrlngens sialidase (neurammldase) (EC 3.2.1 18) (Oxford GlycoSctences, Abmgdon, UK) (see Note 5). Supphed at 5 U/mL and can be stored at 4°C for several months (or at -2O’C). 5. Bovine epididymis a-L-fucosidase (EC 3 2 1 51) (Sigma) (see Note 6). Do not freeze; store at 4°C. 6. 0 OlM Sodium citrate-phosphate pH 6.0. 7. 0.05M Methanoltc-HCl. Prepare fresh by adding 0 35 vol acetyl chlortde to 100 vol anhydrous methanol (store over a molecular sieve 4A) 8 1.OMNaOH. Make fresh on day of use, using 12 5MNaOH (AnaLaR grade) and HPLC grade water.

4

2.6. High-pH Anion-Exchange Chromatography (HPAEC) Separation and Detection of Mucin O-Linked Oligosaccharides 1 HPLC system: Dtonex quartemary advanced gradient HPLC pump coupled to a pulsed amperometric detector (PAD) (see Note 7) (Dionex, Camberley, Surrey, UK). A Dionex eluent degas module to sparge and pressurize the eluents with helmm gas Rheodyne mJectton valve (Rheodyne, Cotati, CA) equtpped wtth Tefzel rotor seal to withstand alkalinity of the eluents Chromatographic data plotted by an integrator (Spectra Physics, San Jose, CA). 2 Two CarboPac PA100 (250 x 4 mm) 8 5 pm anion-exchange columns (Dionex) and one CarboPac PA100 (50 x 4 mm) guard column (see Note 8) 3 Postcolumn accessories Pneumattc controller, pressurizable eluent reservoir (preferably up to 50 psi), 3-way liquid mixing Tee, 500 pL bead-packed reactton co11and Tefzel eluent lme assemblies (Dtonex, UK) 4 12SM(50% w/v) NaOH solution (AnaLaR grade) (BDH) NaOH pellets are not acceptable! 5 Sodmm acetate (AnaLaR grade) (BDH) 6 HPLC grade water (Rathbum Chemicals). 7. HPAEC eluents: Ohgosaccharide alditol and Galpl-3GalNAc analysts (Eluent A) 0.08MNaOH; (Eluent B) 0.5Msodium acetate in 0 08MNaOH. Monosaccharide analysis; 15 mMNaOH Postcolumn reagent; 0.3MNaOH Mix eluents well, vacuum filter and degas. Make fresh on day of use and store in closed helmm gas pressurized vessels (see Note 9) 8 Glass vacuum filtration apparatus, connected to a vacuum pump. 9. 0.2 pm (pore-size) Anodtsc filters (Whatman, Matdstone, UK) 10. o-Melibiose (o-galactose CLI-6-D-glucose) (BDH) 11. n-Galactosepl-3N-acetyl-o-galactosamine (GalPl-3GalNAc) (Russell Fme Chemicals, Chester, UK) 12 Bovine testis P-galactosidase (EC 3.2.1 23) (Oxford GlycoSciences). 13. 0 1M Sodium citrate-phosphate pH 4.0. 14 o+Fucose, N-acetyl-n-galactosamine and o-galactose (Sigma) 15 0.2 pm Anotop-IC sample filters (Whatman)

166

Campbell and Rhodes

3. Methods 3.1. Sample Preparation 3 1.1. Extraction of Mucin from Surgical Resection Specimens 1 Dice the frozen tissue spectmen (l-2 mm3 size pieces) using a sterile surgical blade Alternatively for large normal tissue specimens, remove mucosa using a scrape technique with a strong glass shde 2 Homogenize the tissue (100 mg/mL), using a Polytron, m me-cold phosphatebuffered salme pH 7 4 containing a cocktail of protease mhtbttors (final concentratron; 2 @4 aprotinm, 10 pA4 leupeptm, 5 mM benzamidme hydrochloride and 0 0 1% w/v thtmerosal) 3 On me, to prevent the sample overheatmg, ultrasomcate the homogenate wtth 5 x 30 s bursts of ultrasomc waves at an amplitude of 8 microns peak to peak 4 Centrifuge at 105,OOOg for 60 mm at 4°C 5 Load 2 5 mL aliquots of supernatant onto multiple PD 10-Sephadex GM25 (5 x 1 5 cm) desaltmg columns Elute wtth phosphate-buffered saline, pH 7 4 6 Collect the void volume fractions contammg the high molecular weight protems (1 e , the crude mucms) 7. Pool and lyophdtze m a freeze drier set at 47°C and at a pressure of 54 mbar for 18 h.

3.1 2. Blosynthesls of Mucin from Endoscopic Biopsies 1 Collect endoscoptc biopsies on lens tissue soaked m culture medium. Alternatively multiple mucosal specimens, each weighing about 20 mg, can be taken from a pinned out resectton specimen usmg standard colonoscoptc biopsy forceps or a sterile scalpel 2. Place biopsies, luminal surface uppermost, on an alummmm mesh (maximum of three biopsies per dish) 3. Float the alummmm mesh on 1 mL of culture medmm m the organ culture dish central well Take care not to submerge the biopsies 4 Place 2 mL of sterile detomzed water to the outer well so as to prevent biopsies drying out during mcubation 5 Incubate for 18 h at 37”C, in 95% 02/5% CO*. 6 Collect both the media (secreted mucm) and the biopsies (stored mucm) 7 Place the btopsies m 5 mL ice-cold phosphate-buffered salme, pH 7 4 (3 biopsies/5 mL) 8 On ice, ultrasomcate with 4 x 15 s bursts or until the biopsies appear a ghostly white. 9 Centrifuge the samples at 105,OOOg for 60 mm 10. Isolate the crude mucm from the supernatant (as m Subheading 3.1.1., steps 5 and 6).

3.7.3. Metabolic incorporation of Radiolabeled

Substrates into Mucin

1 Proceed as m Subheading 3.1.X, steps 1 and 2. 2 Float the alummmm mesh on 1 mL culture medmm containing 50 pCt [35S]sulfate and 2 pCt [3H]-N-acetyl-D-glucosamme (see Note 10)

Analysis of Gastrointestinal

Muclns

167

3 Proceed as m Subheading 3.1.2, steps 4-10. 4. Purify the mucm, followmg Subheading 3.2. 5 Quantify label incorporated mto mucin by mlxmg 50-pL ahquot of Superose 6purified mucm with 5 mL of scmtlllant. 6 Count using a [3H]/[35S] dual label detection program

3.2. Purification of Crude Mucus Glycoproteins 3.2.1. Removal of Glycolipids (see Note 11) 1 2 3. 4 5.

(Mucin)

Reconstitute the lyophdlzed mucin m 1 mL delomzed water Add 2 vols of chloroform:methanol(2 1 v/v) Stand at room temperature for 15 mm, vortex mlxmg every 5 mm Separate the chloroform and aqueous layers by centrifuging for 10 mm at 1OOOg Collect and dry the aqueous layer under vacuum m a centrifugal evaporator.

3.2.2. Removal of Glycosaminoglycans

(see Note 1 I)

1 Redissolve lyophlllzed mucm m phosphate-buffered salme pH 7 4 contaming 1 mg/mL collagenase. 2 Incubate for 6 h at 37°C 3 Apply 200-pL ahquots to Sepharose CL-2B mml-columns (5 x 1 5 cm) and elute with the mcubatlon buffer 4. Monitor fractions at OD 280 nm and by ELISA using the CAM17 1 antlmucm monoclonal antlbody (see Subheading 3.2.4.) 5 Collect, pool, and lyophilize the mucin in the Sepharose CL-2B void volume fractions (see Fig. 1) 6. Repeat steps 1-5 usmg hyaluronrdase (1 mg/mL) m 0 02Mphosphate buffer pH 5.8 7. Repeat steps l-5 usmg protease-free chondroitm ABC lyase (0.02 U/mL) m 0.01 MTris-HCl pH 8 0

3.2.3. High-Performance (see Note 72)

Gel Filtration of Mucin Using FPLC

1 Redissolve the lyophlhzed mucin m application buffer (O.lM Tns-HCI, pH 8 0) at a concentration of 5-10 mg/mL 2 Inject mucm samples via a 200~pL sample loop onto a 30 x 1 cm Superose 6 FPLC column. 3. Elute at room temperature with O.lM Tris-HCl, pH 8.0, at a rate of 15 mL/h, using a FPLC system (see Note 13) 4. Monitor fractions contmuously at an optical density (OD) of 280 nm and by CAM 17.1 ELISA. 5 Collect and pool the pure mucm (void volume) fractrons (see Fig. 2) from multiple column runs. 6. Dialyse the mucin against 2-3 changes of 5 L of delonised water, at 4°C for 48 h 7 Remove a 100~p.L allquot for protein estimation of purified mucm using a bicinchoninic acid protem assay kit (see Note 14) 8. Divide mucin into 100~pg protein aliquots, lyophlhze, and store at -80°C.

168

Campbell and Rhodes

~~05~~~, 2 0

.‘--cF, 1

2

3

ELUTION

4

5

6

VOLUME

7

8

9

10

(ml)

Fig 1 Sepharose CL-2B mini-column chromatography Elution profile of ahquots of (A) [3H]GlcNAc-labeled colomc mucm, (B) blue dextran (MW 2 x lo6 Da) and bovine serum albumin (MW 6 7 x lo3 Da) (C) [3H]GlcNAc-labeled colomc mucm “spiked” by addmon of [3H]GlcNAc showing clear separatton of mucm (first peak m void volume) from GlcNAc (D) Shows CAM 17 1 (antimucm antibody) bindmg of Sepharose CL-2B separated mucm showing a single peak correspondmg to the mucm fractions Reprinted from ref. 2 with permission.

3.2.4. ELlSA for Gastrointestinal Mucins Using Antimucin Monoclonal Antibody CAM1 7.1 1. Place 50-pL samples into each well of a flat-bottomed microtiter plate (see Note 15). 2. Add 50 pL of carbonate coating buffer pH 9.0 to each well Agitate for 5 mm on an orbital shaker. Incubate overnight at 4T.

Analysis of Gastrointestinal Mucins

0

10

20

Fraction

169

30

40

50

80

number

Fig 2. Purttication of mucus glycoprotem from resected colonic mucosa usmg a 30 x 1 cm Superose 6-FPLC column, eluted at 15 mL/h. 60 x 0.5 mL fractions were monitored both at OD 280 nm and by ELISA (lower panel) usmg the CAM17 1 monoclonal antimucm antibody V,, blue dextran exclusion V,, phenol red dye (withm sample) 3. Wash the plate twice with 120 pL ELISA-PBS/O.Ol% Tween pH 7 2. 4. Add 120 p.L ELISA-PBS pH 7.2 containing 0 05% Tween and 0 05% bovine serum albumm (as a nonspecific blocking agent). Incubate for 1 h at 37°C 5. Remove, blot the plate briefly, and then add 100 pL of the CAM 17 1 antimucm monoclonal antibody (1: 10) m ELISA-PBS/O 01% Tween, pH 7 2 Incubate for 2 h at 37°C 6 Remove and wash the plate three times 7 Add 100 pL peroxidase-conjugated antimouse Ig (1:600) m ELISA PBS/O 0 1% Tween, pH 7.2. Incubate for 2 h at 37°C 8 Remove and wash the plate 5 times. 9. Add 100 pL of substrate (0 2 mg/mL OPD in 1.1 citrate-phosphate pH S.O.deionized water containing 1 6 pL/mL 30% H202). Leave for 10 min at room temperature. 10 Stop reaction with 100 & 4M H2S04 11. Read OD at 490-492 nm using an ELISA plate reader.

170

Campbell and Rhodes

3.3. Isolation of Mucin O-Linked Oligosaccharide by Alkaline-Borohydride Degradation

Al&to/s

1. Redissolve purified mucm (up to 500 pg protein) with 1 mL HPLC grade water 2 Add an equal volume of freshly prepared 2A4 NaBH4 in O.lM NaOH m a clean glass 3 mL Reacti-vial (Caution: see Note 16). 4 Incubate for 16 h at 50°C (see Note 17) 5. Following reductron, on ice, carefully degrade excess borohydride with dropwise addition of 50% v/v glacial acetic acid in HPLC water, to pH 5 0 6 Load reactions on Dowex 50X-12 H+ form resin (3 x 0 5 cm column) and elute with 3 column volumes of HPLC grade water 7 Collect, pool, and lyophihze the combined effluent and eluent 8 In a fume cupboard, remove boric acid (as methylborate) by coevaporation with four additions of 500 pL methanol (either reduced under a stream of nitrogen or under vacuum m a centrifugal evaporator). 9 Redissolve sample m 500 pL HPLC grade water and filter through 0 2-w Anotop-IC disposable filters.

3.4. Release of Core 1 Disaccharide (Galpl-SGalNAca-) from Mucin Using Endo-a-N-Acetylgalactosaminidase (o-Glycanase) 3.4.1 The Removal of Glycerol from 0-Glycanase (see Note 18) 1 Invert a Bio-Spin P30 polyacrylamide Bio-Gel column (0 8 mL column volume) several times and allow the buffer to dram by gravity. 2 Wash the column with 300 pL 0 OlMcitrate-phosphate pH 6 0, place m a collection tube and centrifuge for 2 mm at 11OOg. Repeat four times 3 Make 6 mU O-Glycanase up to 100 pL with 0.0 1M citrate-phosphate pH 6 0 and load sample carefully and directly to the center of the column, dropwise (see Note 19). 4. Centrifuge for 4 mm at 1,OOOgand collect the excluded O-Glycanase 5 Pass the excluded O-Glycanase through a second Bio-Spin column to maximize glycerol removal

3.4.2. Hydrolysis of Gal/? l-3GalNAca-

from Mucin Using 0-Glycanase

1. Reconstitute mucm (100 pg wrth 90 pL 0 1M citrate-phosphate pH 6 0 contammg 100 pg/mL bovme serum albumm and 0.02% w/v sodium azrde Mix well 2. Add 0.6 mU deglycerolated O-Glycanase (10 pL) 3. Incubate for 18h at 37°C (see Note 20). 4. Load reactions on Dowex 50-X12 H+ form resin (3 x 0 5 cm column) and elute with 3 column volumes of HPLC grade water 5. Collect the effluent and eluent, pool (2.5 mL volume). 6. Load onto a PD 10-Sephadex GM25 column and elute with HPLC grade water to isolate liberated disaccharide from intact mucm (see Note 21) 7. Monitor fractions using the CAMl7.1 ELISA and a phenol-sulfurtc acid assay for neutral sugars (see Subheading 3.4.3.)

171

Analysis of Gas tram tes tlnal Mucins Protem Neutral

0

2

4

Elutlon

6

volume

l2flOnm) sugar (490nm)

8

--c

10

12

(ml)

Fig 3 Separation of U-Glycanase-liberated disaccharide GalP I-3GalNAc from intact mucm and enzyme using PDlO-Sephadex GM25 gel filtration Samples were loaded on the column in a 2 5 mL volume. Protein was monitored at OD 280 nm and eluted m fractions 2 56 5 mL, neutral sugar (Gall3 1-3GalNAc) was monitored using a phenol-sulfuric actd assay and eluted m fractions 6.5-10.5 mL (indicated by arrows) 8. Collect the fracttons contammg neutral sugar (see Fig. 3), pool and lyophthze 9. Redtssolve sample m 500 pL HPLC grade water and filter through 0.2~pm Anotop-IC disposable filters

3.4.3. Phenol-Sulfur/c

Aad Assay for Neutral Sugars

1 2 3 4 5.

Place lo-p.L samples mto flat-bottomed microtiter plate wells (see Note 22) Add 100 pL of 4% (w/v) phenol (m deionized water) to each well Mix/aspirate five times Leave to stand for 5 mm Rapidly add 150 pL of concentrated H$04 to the wells using a fast-delivery multichannel pipet (Hazard: see Note 23). 6. Immediately mix, with 5 aspnattons as standard procedure, to generate a htgh reaction temperature (see Note 24) 7 Leave the plates to cool for 20 mm. 8. Read samples at 490 nm using an ELISA plate reader.

3.5. Removal of Terminal Sugar Residues from Much 3.5.1. Mild Acid Hydrolysis of Terminal Sugar Residues 1. Dissolve the lyophihzed mucin (100 pL) with 250 ,uL 0 05M H2S04 m a clean glass 1 mL Reacti-vial 2. Incubate for 2 h at 80°C (see Note 25) 3 Cool the vials and evaporate the reaction to dryness under a stream of nitrogen. 4. Redissolve m 0 lMcitrate-phosphate buffer pH 6.0 to a final volume of 2.5 mL

172

Campbell and Rhodes

5. Load onto a PDlO-Sephadex GM25 column and elute with 0 1M citrate-phosphate, pH 6 0 6. Collect the mucm (void volume) and lyophihze (see Note 26).

3.5.2. Selective Enzymatic Removal of Sialic Acids 1 Saponify the mucm (100 pg) m 50 pL 0 05MNaOH m a glass Reacti-vial, at 4°C for 30 mm (see Note 27) 2. On ice, neutralize the reaction with 25 pL 0 1M HzS04 3 Pass the sample through a column of Bio-Gel P30 (0 8 mL) to remove chloride ions. 4 Add 20 @ Clostrzdwn perfrwgens neurammldase (5 U/mL) and make total reaction volume up to 150 pL using 0.25M sodium acetate pH 5.0 and deiomzed water (to a final concentration of 0 05M sodium acetate pH 5 0) 5. Incubate for 18 h at 37°C. 6. Make up volume to 2.5 mL with 0 1M citrate-phosphate pH 6.0 and load onto a PD 10 Sephadex GM25 column 7 Collect the void volume and lyophilize the desialylated mucm (see Note 26)

3.5.3 Selective Enzymatic Removal of Fucose Residues 1, Dissolve the lyophihzed mucm (100 pg) with 100 pL 0 OlM citrate-phosphate pH 6 0 contammg 0 2 U/mL bovine epididymrs a-fucosidase in a clean glass 1 mL Reacti-vial 2 Incubate for 18 h at 25°C 3 Remove fucose by exclusion chromatography using PDlO Sephadex GM25 column chromatography 4 Collect the void volume and lyophihze the defucosylated mucm (see Note 26)

3.5.4. Chemical Removal of O-Sulfate Esters (see Note 28) 1 Dissolve 100 pg mucm samples with 200 pL 0 5M fresh methanohc-HCl, m a clean glass Reacti-vial with Teflon-sealed cap 2. Heat at 32°C for 4 h 3. On ice, neutralize with 1MNaOH. 4. Dialyze the desulfated mucm against three changes of 5 L of HPLC grade water, at 4°C for 48 h 5 Lyophihze and store the mucm at -80°C.

3.6. High-pH Anion-Exchange Chromatography (HPA EC): Separation and Defection of Mucin O-Linked Oligosaccharides 3.6. I. HPAEC Analysis of Alkalme-Borohydride-Released Mucm Oligosaccharide Alditols (see Note 29) 1. Inject aliquots (50 and 100 of released mucin oligosaccharide alditols, containing 320 ng of D-melibiose as Internal standard, onto the two CarboPac PA1 00 columns (connected m series) eqmhbrated in Eluent A (0.08M NaOH)

Analysis of Gastrointestmal Mucins

1

1

B

173

n

2

F m 8 m

*m .s rL

^

-

i :I

Fig. 4. HPAEC separation of Gal0 1-3GalNAc. All samples contamed 320 ng mehbtose (2) as internal standard. (A) 0 5 ug standard Galpl-+3GalNAc (1) (arrows) and GalPl-3GalNAc liberated by O-glycanase from (B) 100 pg antifreeze glycopeptrde AFGP-I and (C) colomc adenocarcmoma mucm. PAD output was at 1 pA The mmal peak at approximately 3 8 mm 1s the solvent front Reprinted from ref. 5 with permission 2 Elute with an mcreasmg gradient of Eluent B (0 5M sodium acetate m 0 08M NaOH) at a flow rate of 0.7 mL/min (see Note 30) 3. Profile the chromatogram obtained using a range of well-characterized human ohgosaccharlde alditol standards (see Note 31) 4 Re-equilibrate column before each subsequent sample apphcatton (see Note 32)

3.6.2. HPAEC Analysis of 0-Glycanase-Hydrolyzed

Galp l-3GalNAc

1. Inject a lOO+L ahquot of the O-Glycanase-released product, contammg 320 ng of D-melibiose as internal standard, onto the two CarboPac PA100 columns (connected m series) eqmhbrated m Eluent A (0 08M NaOH) 2 Elute with an mcreasmg gradient of Eluent B (0 5M sodmm acetate in 0.08M NaOH) at a flow rate of 0 7 mL/min (see Note 30 and Fig. 4) 3 Quantify Gall3 1-3GalNAc released using a cahbratron curve for standard amounts of Gall3 1-3GalNAc (see Note 33). 4 Re-equilibrate column before each subsequent sample application (see Note 32)

174

Campbell and Rhodes

3.6.3. HPAEC Analysrs of Constrtuent Monosaccharides of 0-Glycanase-Released Products. 1 Lyophtlize a 200~pL aliquot of the 0-Glycanase-released product (see Note 34). 2 Redissolve with 20 pL of 0 1M citrate-phosphate pH 4 0 containing 0.5 U/mL bovme testrs P-galactosrdase 3 Incubate for 6 h at 37°C 4 Take a l-lo-pL ahquot of the P-galactostdase reaction mixture and make up to 100 pL total volume with HPLC grade water contammg 100 ng a-t.-fucose as an internal standard 5 InJect directly onto a single CarboPac PA100 column, equilibrated with 15 mM NaOH 6. Elute samples isocratically with 15 mA4 NaOH for 20 mm, at a flow rate of 0 8 mL/mm Contmuously add postcolumn reagent (0 3M NaOH) to the column effluent at 0 6 mL/min to facrhtate detection (see Note 35 and Fig. 5) 7 Quantify monosacchartdes released using N-acetylgalactosamme and galactose standards. 8 Regenerate the column before each subsequent sample analysts (see Note 36)

4. Notes When preparing the Sepharose CL-2B “slurry,” wash the gel with high quality detomzed water on a glass filter-funnel (porosity ~10 pm) so as to remove the ethanol present as a gel preservative Add eluent buffer to make a suspension (-2-3 g gel/mL) To improve column packing, eluent should contam 0 1% (v/v) Tween-20 and be degassed before packing The gel suspension should appear homogeneous (i e , free from aggregates) when packing the empty PDIO columns Allow to dram by gravity at the temperature they will be used at Care should be used m applymg the top filters so as not to compress and distort the swollen gel CAM 17.1 1s an anttmucm mouse monoclonal antibody (6) with specificity for colomc and pancreatic mucms Specificity appears to be to the sralyl I blood group antigen (JD Milton, personal commumcatton) Homemade columns packed with Bto-Gel P30 (0 8 mL) can be used, but for convenience the columns should be made suitable for use with 1.5 mL mtcrocentrifuge or 12 x 75 mm collectton tubes. 0-Glycanase from Streptococcus (D~plococcus) pneumonzae is also avatlable from the Genzyme Corporatton, USA and from Boehrmger Mannhelm, UK (under the name of 0-glycostdase). The enzyme spectfically hydrolyzes the glycostdtc bonds formed by N-acetylgalactosamme and the hydroxyl group of either serme or threomne only when m the sequence GalPl-3GalNAcal-0-Ser/Thr (1 e , the Thomsen-Frtedenretch (TF) blood group antigen) (7,8). It will not act on GlcNAcP l-3GalNAca 1-0-Ser/Thr, GalNAca I -0-Ser/Thr, nor on GalP l3GalNAcal-0-Ser/Thr tf substttuted either on the Gal or on the GalNAc by sialtc acrd or fucose. Hence, prior removal of these sugar residues and/or O-sulfate esters can be necessary to reveal “cryptic” TF antigen (see Subheading 3.5.)

1

2 3

B

2 D p?

a

i 1 -

I b z

IB N

3 3 2

I

C

-1

-

1

-- -

--

--

1 -I

E

D

F

3

91

L

q3

\-

L

I

I

10

I

L

Ib

I

20

TIME

3 N

-

’

i

(minutes)

Fig. 5 Monosaccharide composition analysis of P-galactosidase-treated O-glycanase reaction products using single CarboPac PA 100 HPAEC. All samples contamed 100 ng a-L(-)-fucose (1) as internal standard (retention time, 5.27 mm). (A) P-galactostdase alone, (B) standard N-acetylgalactosamme (2) and galactose (3) (both 300 ng); (C) j3-galactosidase-treated standard GalPI-3GalNAc, (IkF) P-galactosidase treated 0-glycanase-liberated disaccharide from antifreeze glycopeptide AFGP-I, colonic adenocarcmoma mucm and mild acid-treated normal colomc mucm respectively PAD output was at 1 uA Reprinted from ref. 5 with permission

175

Campbell and Rhodes

176

perfringens sialldase (neuramimdase) has the broadest specificity of the available slalldases cleaving a2-3, ~~2-6, and 1~2-8 linked slahc acids It IS also highly active against mucms and is, therefore, the choice for nonselective mucm desialylatlon It will cleave 9-O-acetyl stahc acid, but not the 4-O-acetyl analog and hence prior saponification of the mucin is recommended before any enzymatic slahc acid removal. For more specific removal of (a2-3)-linkages use Newcastle Disease Vn-us slahdase and for ~2-6 >> a2-3-linked slahc acids use Arthrobacter ureafaczens slalidase It should be noted that other sources of a-L-fucosldase, e g., from bovine kidney, although suitable for the hydrolysis of linkages between fucose and a variety of substrates (mcludmg phenols, oligosacchandes, glycopeptldes, and glycollplds), m our experience does not remove fucose residues from native glycoprotems. Pulsed amperometrlc detectlon (PAD) was performed with a gold workmg electrode and a Ag/AgCl reference electrode Set the detector output at 1 cvl\ the detection potential (El) to +O.OW, the oxldatlon potential (E2) to +0.6V and the reduction potential (E3) to 4 6V. Time constants for each applied potential (on Range 2) should be set to t, = 720 ms (posltlon 9), t, = 300 ms (position 2) and t, = 240 ms (position 1) respectively. Because of the sensitivity of the PAD, the adequate Integrator settings are; attenuation = 1024, and peak threshold = 10,000 For the resolution of neutral and slalylated ohgosacchatlde or GalP 1-3GalNAc derived from mucm, use two CarboPac PA 100 columns connected m series with a CarboPac PA100 guard column For the analysis of constituent monosaccharides, use a single CarboPac PA-100 column m series with the CarboPac guard To prepare Eluent A (0.08M NaOH) take 6.4 mL of 12 5M (50% w/v) NaOH, make up to 1 L with HPLC water Eluent B (0 5M sodium acetate-O 08M NaOH) is composed by adding the same amount of 12SM NaOH and made up to 1 L with 0.5Maqueous sodium acetate (41 g anhydrous sodium acetate in 1 L HPLC water) The eluent used for HPAEC monosaccharide analysis (15 mM NaOH) IS composed by adding 1.2 mL of 12.5M NaOH, make up to 1 L with HPLC water. Postcotelumn Reagent; 24 0 mL of 12.5M NaOH, make up to 2 L with HPLC water. All eluents must be vacuum filtered through 0.2 pm and degassed by helium gas spargmg (20-30 mm) so as to mnnmize absorption of COz. Incorporation of [3H]-GlcNAc mto mucm 1sincluded to ensure adequate labeling of ammo-sugars (GlcNAc, GalNAc, and N-acetylneuramnnc acid) as an Index of new glycoprotem synthesis Inltlally there is a lag of &8 h when the rate of mucm synthesis IS low, followed by a period when the rate of mucm synthesis (total of mucm m biopsy and culture medium) IS roughly linear for up to 26 h Radlolabelmg of mucm in the biopsy specimen homogenate increases gradually to a peak at around 20 h (2) Sepharose CL-2B purified mucm contains a small amount of glycosammoglycan and glycolipid contammatlon as demonstrated by the mcubatlon of, or extraction of, [3H]GlcNAc-labeled mucm (purified by Sepharose CL-2B) with hyaluromdase or chloroform methanol respectively Loss in void volume recovery IS typically small, <5% and 2.5%, respectively.

5 Clostridwm

6

7

8

9

10

11.

Analysis of Gastrointestinal

Mucins

177

12 Minimal contammatton of mucm with DNA remains (9), but 1s not likely to be a problem for studies of mucm oligosacchartde primary and secondary structure as there is no deoxyribose m glycoprotein 13 For optimum resolution of mucin on the Superose 6 column, the sample loaded should not exceed 200 p.L,in volume nor contain more than 5-10 mg of protein In addmon, do not run the column at a contmuous back-pressure above 1.2 MPa and tf required follow the manufacturers cleaning procedures to reduce pressure. It is also important to allow an adequate equilibration period between each sample run particularly when runnmg media samples containmg high amounts of nonincorporated radiolabeled substrate and phenol red pH-indrcator 14. Protein is expressed relative to a standard of bovine serum albumin However, it should be noted that the protein content of mucin is probably underestimated because of the heavy glycosylation of the protein core Therefore, the assay should be used to compare different mucm samples from a similar source rather than to provide absolute quantlficatton of mucm content 15. Avoid using the outermost “ring” of wells on any ELISA plate as these can lead to false readmgs, therefore, coat sample wells 2-l 1 m any one row only Cahbrated standards of desalted Superose 6-purified human colomc mucm should be included on each plate for quantification and as a positive control. Blanks should also be included (1.e , buffer plus coating buffer) 16 Caution: During the p-ehmmatron reaction, Hz gas is produced and, therefore, the reaction vial should not be closed tight since the cap may blow off. If tnttated borohydnde (NaB3H4) is used, radioactive 3H, is released during the reaction and particularly durmg destructton of excess borohydride with addition of 50% acetic acid. Therefore, the use of a radtoactive designated fume cupboard is essential. 17 Colomc and respiratory mucms are parttcularly rich in acidic ohgosaccharide side-chains with a high degree of sulfatlon. The release of mucm O-linked oligosaccharide alditols under the alkaline borohydride condmons described here can result in some loss of alkali-labile O-sulfate esters. If sulfated ohgosaccharide alditols are to be investigated alkaline-borohydride treatment of mucin should be performed under the following conditions; 2MNaBH4 in 0.05MNaOH at 4YC for 16 h to ensure minimal sulphate loss (10). 18. The efficient removal of glycerol (added as an enzyme stabilizer) from commercial 0-Glycanase is necessary for the subsequent HPAEC analysis of hydrolyzed GalSl-3GalNAca-, as glycerol was found to be highly PAD actrve (see Fig. 6). Deglycerolated 0-Glycanase, however, does not store well so it is best to remove the glycerol on day of use 19. The optimum volume of sample 1sbetween 50 and 100 pL. Applying more or less results m poor recovery of desalted sample. Likewise, sample directed down the side of the gel also results m poor recovery. 20. Approprtate control samples, incubated in the absence of 0-Glycanase and incubated wtth 0-Glycanase followmg pretreatment with bovine testis S-galactostdase, should be included for each set of 0-Glycanase assays. Each assay should also include a positive control for 0-Glycanase-mediated TF-antigen release (1 e ,

Campbell and Rhodes

178 B

1

N

i

B z___ -

"

P %.-

I

rz 3

I

I

I

1

I

I

I

I

0

5

10

15

0

5

10

15

TIME (mid

TIME (min)

Fig 6 HPAEC separation of GalPl-3GalNAc-hberated from antlfreeze glycopeptldes (1) by (A) standard 0-Glycanase and (B) BloGel P30-deglycerolated 0-glycanase The glycerol present m commercial 0-Glycanase IS highly PAD active preventing Galpl-3GalNAc quantification Sample B contamed 320 ng mehblose (2) standard. PAD output was at 1 pA. from a Galpl-3GalNAca(TF antigen)-expressing substrate, such as antifreeze glycopepfides (100 pg) or aslalofetum (Type I) (200 clg)(II, 22). 0.6 mU of 0-Glycanase releases -400 ng Galpl-3GalNAc/pg antifreeze glycopeptlde I and -45 ng/pg asialofetuin (5) 21 Elution must be with HPLC grade water for problem-free detectlon of saccharides by HPAEC-PAD analysis. 22 Include a water blank and a standard curve of 0 5-20 ~18D-galactose (in trlphcate) with each assay on every plate The within-assay coefficient of variation for standards should be less than 10%. 23 Hazardous reactIon This 1sa very exothermlc reaction mvolvmg addition of concentrated sulfuric acid and an aqueous sample! Appropriate precautions should be taken; wear safety glasses, gloves and labcoat as standard 24 This micro phenol-sulfuric acid assay for neutral sugars IS a scaled down version of the method of Dubois et al. (23) The phenol-sulfuric acid reaction is based on the heat generated by mlxmg concentrated sulfuric acid with water and m this

179

Analysis of Gastrointestmal Mucins

0

’

I

Normal colontc mucu7

I

Mid

acid-treated mucln

Fig 7 Analysis of GalP I-3GalNAc, expressed as ng/pg protein, after 0-Glycanase treatment of normal and mild acid-hydrolyzed normal colomc mucm Galpl-3GalNAc expression was srgrnficantly Increased m all nme mucm samples followmg mild acid treatment to remove siahc acids and fucose (** p < 0 001)

25.

26

27. 28

method the small diameter of the ELBA plate wells (-50 mm) allows good mixmg without disstpatmg the heat too rapidly. Thus this generates the high temperature required for increased assay sensitivity Mild acid hydrolysis of mucm under the condttions described has been demonstrated to significantly reveal 0-Glycanase-hydrolyzable Gall3 1-3GalNAc (5) (see Fig. 7) In other studies, we have demonstrated that lectm blotting after mild acid hyrolysis (0 0544 H2S04, 3 h at SO’C) abolishes Limaxflavus agglutmm (LFA) and Ulex europaeus agglutmm (UEA-1) bindmg to pancreatic cancer-related serum mucus glycoprotem (14). Similarly, 1.5 h incubation of human meconium glycoprotems with 0.05MH2S04 at 100°C removes -80% of fucose restdues (15) In addmon, the included fractions from the PDlO Sephadex GM25 columns (711 mL) (see Fig. 3) can be collected, lyophihzed, and redissolved in 500 pL HPLC water for the quantificatton of released fucose and/or sialic acid residues by HPAEC (5,16). The method of saponification described here results m substantial removal of the staltc actd 0-acetyl groups, which confer resistance to the action of sialidase, without stgmficant nonspecific desialylation Efficiency of desulfation should be assessed using a duplicate sample of Superose 6-purified [35S]-sulfate/[3H]-GlcNAc-labeled mucin This IS easily

180

Campbell and Rhodes

achieved by analysis of the [35S]/[3H] content of the mucin elutmg at the void (Superose 6) before and after methanolic-HCl treatment 29. Neutral and fucosylated mucm reduced ohgosaccharides are m general not sufficiently retained for adequate separation and reproducible chromatography using the CarboPac anion-exchange phase. However, the dual CarboPac column method described here is an effective method that sigtuficantly increases retention and separation Retention times for neutral (RJ increase with mcreasmg oligosaccharide alditol chain length and branching, monosaccharide aldnols are resolved clear of the solvent front (R, = 4-5 mm), neutral dtsacchandes 5-7 mm, trisaccharides 5-10 mm and pentasaccharides 8-l 8 mm Oligosaccharide aldltol isomers differing only by termmal p( l-3)- or p( 1-4)-linked terminal galactose can be separated (16). Acidic sialylated ohgosacchartde alditols are retained longer (monosialylated, 25-33 mm and dlsialylated, 3545 mm) and this method is also useful for the isomeric separation of a(2-3)- and a(2-6)-linked sialic acids on stalylated ollgosaccharides 30 HPAEC analysis of oligosaccharide alditols and the core 1 disaccharide Galpl3GalNAc: The elutton protocol is as follows Time % Eluent A p/o Eluent B (0.08MNaOH) (0.5MNa acetate-O 08M NaOH) (ml@ 0 100 0 0 10 100 30 45 50 55 75

80

20

100 0 0 100 100 0 100 0 3 1. A range of well-characterized neutral, sialylated, and sulfated ohgosacchande alditol structures derived from gastromtestmal mucm sources, such as human mecomum (15), human respiratory mucin (IO), and human or bovine submaxillary mucin (17,18), should be used for HPAEC profilmg. 32. It should be noted that a high vanatton m retention times can be obtained between different chromatographic runs unless the CarboPac anion-exchange column IS regenerated with strong alkaline solution (i e., 0 5M sodium acetate-O 08M NaOH) and then re-equilibrated in 0.08MNaOH for a period of 20-25 mm before mitiatmg the next chromatogram (see Note 30) This phenomenon on this high pH anion-exchange phase has been previously described (16,19). 33 For the quantification of 0-Glycanase-hydrolyzed Gall3 1-3GalNAc, the dual CarboPac PA-100 anion-exchange column should be calibrated with standard amounts of Galpl-3GalNAc (0.0625-1.0 pg recovered after PDlO column separation (% analytical recovery 96 7 + 3 2 (mean LSD). The calibration curve was linear (r = 0.99) using peak height and internal standard ratio quantification analysts (area under the curve was also linear (r = 0 97) over the same concentration range). Typical within assay coefficient of variation was 9.3% and between assay coefficient of variation was 10 1%. The lower limit of detection for Galpl3GalNAc (on 1 pA PAD output) is -0.02 pg.

Analysis of Gastrointestrnal Muds

181

34 Use a positive control sample of Galj31-3GalNAc hydrolyzed by 0-Glycanase from anttfreeze glycopepttdes (100 pL) asialofetum (200 pL) and/or 10 pg standard Gall3 I-3GalNAc incubated under the same conditions (see Fig. 5) 35 PAD conditions should be as descrtbed m Note 7 However, to optimize detector sensitivity and avoid baseline drift at low NaOH concentrations the addition of 0 3MNaOH postcolumn (I e , before it enters the detector) is required to Increase pH 212. Use of a thinner electrode gasket (0 005”) gives the best signal to noise ratio. The flow rate at which the postcolumn reagent is added should be reproducible between runs and must be < to the gradient pump flow rate; a pressure of -12-14 psi (eluent vessel must be able to withstand higher pressure) should achieve this flow rate with no sigmficant increase in gradient pump pressure Beware of air bubbles m the beaded mixing cot1 preventing efficient pH Increase. Note* always turn off the gradient pump before turnmg off the postcolumn eluent stream. 36. Retention times will steadily decrease unless the column IS regularly regenerated with strong alkali. Regenerate the CarboPac PA100 anion-exchange column for 5 mm with 20% 0 5M sodium acetate-15 mA4 NaOH and then re-equilibrate for 30 mm with 15 mA4 NaOH before each subsequent sample application (see Note 32)

Acknowledgments Funding was obtained from the Medical Research Council, the National Association

for Colttis

and Crohn’s

Disease and the North

West Regional

Health Authority. The authors would like to thank the Gastroenterology Unit, the Departments of Surgery and Pathology (Umverslty of Liverpool and the Royal Liverpool University tissue specimens

Hospital

Trust) for their cooperatton

in obtatnmg

References 1. Rhodes, J M., Campbell, B. J , and Finme, I. A. (1994) Mucus and the gastromtestinal tract, m Recent Advances zn Gastroenterologv 10 (Pounder, R , ed ), ChurchillLtvmgstone, Edinburgh, pp 57-79. 2. Finme, I A., Dwarakanath, A. D., Taylor, B. A., and Rhodes, J. M. (1995) Colomc mucin synthesis is increased by sodmm butyrate Gut 36,93-99. 3. Dwarakanath, A. D., Campbell, B. J , Tsai, H H., Sunderland, D , Hart, C. A , and Rhodes, J M. (1995) Fecal mucinase activity assessedin inflammatory bowel dtsease using 14C-threonme labelled mucin substrate. Gut 37, 58-62 4 Sheehan, J. K and Carlstedt, I (1987) Size heterogeniety of cervical mucus glycoproteins-studies performed with rate-zonal centnfugatton and laser light scattering. Biochem J 245,757-762

5. Campbell, B J., Fmnie, I A., Hounsell, E. F , and Rhodes, J. M. (1995) Direct demonstratton of increased expression ofThomsen-Friedenreich (TF) antigen from colomc adenocarcmoma and ulcerative colitis mucm and its concealment in normal mucin f, Clin Invest 95. 571-576

182

Campbell and Rhodes

6 Makm, C. A (1986) Monoclonal antibodies raised to colorectal carcmoma antigens Ann R Co11 Surg Eng 68,298-301 7. Endo,Y and Kobata, A (1976) Partral purtficatton and characterlsation of an endoa-N-acetylgalactosammidase from the culture medmm ofDlplococcuspneumonrae J Blochem 80, l-8. 8 Umemoto, J , Matta, K L., Barlow, J J., and Bhavanandan, V P (1978) Actron of endo-a-N-acetylgalactosammidase on synthetic glycosldases mcludmg chromogemc substrates Anal. Bzochem 91, 186-193 9 Parker, N , Raouf, A. H , Finme, I. A, Ryder, S D , Campbell, B. J , Tsar, H H , Iddon, D., Milton, J D , and Rhodes, J M. (1993) High performance gel filtration using monodisperse highly cross-lmked agarose as a one step system for mucm puriiicatron Blamed Chromatogr 7,68-74 10 Lo-Guidice, J. M , Wleruszeski, J M , Lemome, J , Verbert, A , Roussel, P., and Lamblm, G (1994) Sialylation and sulphatron of the carbohydrate chains m respiratory mucms from a patient with cystic fibrosis. J Btol Chem 269,18,794-l 8,8 13 11 Shier, W. T , Lm, Y., and DeVries, A L (1975) Structure of the carbohydrate of antifreeze glycoprotems from an antarctic fish FEBS Lett 54, 135-138 12 Spiro, R. G and Bhoyroo, V. D. (1974) Structure of 0-glycosrdrcally linked carbohydrate units of fetum J Blol Chem 249,5704-5717. 13 Dubols, M , Gilles, K. A, Hamilton, J K., Rebers, P. A, and Smith, F (1956) Colorometric method for determination of sugars and related substances.Anal Chem 28,3X)-356. 14 Chmg, C. K and Rhodes, J M (1990) Purification and characterrsatlon of a peanut-agglutmin-bmdmg pancreatic serum-related mucus glycoprotem. Int J Cancer 45,1022-1027 15 Hounsell, E. F., Lawson, A. M , Feeney, J., Goal, H C , Pickermg, N. J , Stoll, M S , Lui, S. C , and Feizi, T (1985) Structural analysis of the 0-glycosldically linked core-region ohgosaccharides of human meconmm glycoprotems which express oncofetal antigens Eur J Blochem 148,367-377, 16. Campbell, B J , Davies, M , Rhodes, J. M , and Hounsell, E F (1993) Separation of neutral ohgosaccharide alditols from human mecomum usmg high pH amon exchange chromatography J. Chromatogr. (Biomed Appl.) 622, 137-146 17. Klein, A., Carnoy, C., Wieruszeski, J M., Strecker, G., Strang, A. M , van Halbeek, H , Roussel, P , and Lamblm, G (1992) The broad diversity of neutral and sralylated ohgosacchandes derived from human salivary mucms. Btochemlsfry 31,6 152-6165 18 Chai, W., Hounsell, E F, Cashmore, G C , Rosanktewicz, J R ,Bauer, C J , Feeney, J., Frizi, T , and Lawson, A M (1992) Neutral ohgosaccharides of bovine submaxillary mucm a combined mass spectrometry and ‘H-NMR study Eur J Bzochem 203,257-268 19 O’Lloyd, K and Savage, A. ( 199 1) High performance anion-exchange chromatography of reduced ohgosacchartdes from stalomucms GlycoconJ J 8,493-498

12 TLC in Structure and Recognition Studies of Glycosphingolipids Johannes

Miithing

Introduction Glycosphmgoliptds (GSLs) are primarily located m the plasma membrane of animal cells, but are also found in association with mtracellular organelles (1,2) GSLs consist of two structural elements: a hpophthc membrane anchor, the ceramtde portton, which is formed by a long chain aminoalcohol and a fatty acid, and a hydrophilic carbohydrate moiety, which protrudes from the cell surface. Structures and functions of GSLs have been widely reviewed (3-s). They are divided into three main classes:neutral GSLs and actdtc siahc actdand sulfate-contaming GSLs, the ganghosides and the sulfatides, respectively. Most GSLs can be grouped mto one of four mam structural families: the ganglio-, globo-, lacto-, and/or neolacto-series. Structures of neutral GSLs and gangliostdes used m this protocol are listed m Tables 1 and 2, respectively. Their nomenclature follows the IUPAC-IUB recommendations (9) and the nomenclature of Svennerholm (10). Sialic acids are destgnated according to the suggesttons of Reuter and Schauer (11). GSLs are usually extracted with chloroform/methanol-mixtures (2: 1, 1: 1, 1:2, each by vol) with a recommended tissue-to-solvent ration of 1:20 to 1:40 (v/v). Anion exchange chromatography is convenient as the first purtfication step of crude GSL extracts (3) because of the complete separation of neutral GSLs and gangliosldes. The use of various anion-exchange DEAE-coupled matrices has been reported (12-16). Further improvements in separation were the application of DEAE-Fractogel (17) and strong amon exchanger Q-Sepharose (18) and TMAE-Fractogel (19). Thereafter, the neutral and the acidic lipid containing fractions undergo a mild alkalme treatment to saponify contaminating phosphohpids usually followed by further purification of neu1.

From

Methods m Molecular Bology, Vol 76 Glycoanalysm Protocols Edited by E F Hounsell 0 Humana Press Inc , Totowa, NJ

183

184

Mthing

Table 1 Structures

of Neutral

GSLs

Glycosphmgoltpld LacCer GgOse$er GgOseJer GbOse$er GbOse&er Forssman GSL LcOse&er nLcOse&er nLcOse&er

Table 2 Structures Gangllostde” GMS GMI

GDh %b @lb

Structure GalPI-4Glcj31-1Cer GaINA@ 1-4GalP 1-4GlcP 1- 1Cer GalP l -3GalNAcj3 1-4GalP l -4GlcP 1- 1Cer Galcll-3GalP I-4GlcP I- 1Cer GalNAc~l-3GalaI-3Gal~1-4Glc~l-lCer GalNAca1-3GalNAc~1-3Galal-3Gal~l-4Glc~1-1Cer GlcNAcP 1-3GalP 1-4GlcP 1- 1Cer GalP 1-4GlcNAcP 1-3GalP l -4Glcj3 1- 1Cer Gal~I-4GlcNAc~l-3Gal~l-4GlcNAc~l-3Gal~l4GlcPl-1Cer

of Gangliosides Structure I13Neu5Ac-LacCer I13Neu5Ac-GgOse&er IV3Neu5Ac, I13Neu5Ac-GgOse&er I13(Neu5Ac)2-GgOse&er IV3Neu5Ac,I13(Neu5Ac)2-GgOse&er

aOnly ganglrosldes with NeuSAc-substltutlon are shown

tral GSLs as their peracetylated derivatives by Flortsil-chromatography (20) and ganghosides on silica gel, e.g., Iatrobeads (3,21). It should be menttoned that alkaline labile O-acetyl- or lactone-derivatives of stalic acids are nreversably disintegrated by thusprocedure. Details concermng the mentioned procedures have been recently published by Kmep and Ebel(22). Because of its resolving power and easyhandling, high-performance thm layer chromatography (HPTLC) has become the standard tool for separation of neutral GSL and ganghosrde mixtures for analytrcal and preparative appltcatrons (3,2326). The goal of this chapter is to provide some special advice for successful analyttcal TLC. The ohgosaccharrde portions of GSLs can be vtsuahzed by conventional staining with orcinol, and siahc acid with resorcmol. Recently developed TLC immunostainmg (overlay technique) enables sensitrve detection of distmct GSLs m complex mtxtures directly on the TLC plate by use of specific anti-carbohydrate antibodies. The TLC-immunostaimng procedure is described for two carbohydrate recognizing antibodies, i.e., anti-nLcOse&er and antt-

TLC of Glycosphingolipids

185

G&NeuSGc) antibodies. These practical applications will prove the overlay technique as a powerful tool in the specific detection of GSLs 1.1. Application of Samples and Development of Chroma tograms Slhca gel 60 precoated alummlum, plastic, and glass-backed plates are commonly used. HPTLC plates from E. Merck (Darmstadt, Germany) are generally recommended for diverse applications (27,28). To achieve optimal separation of distinct GSL bands, probes can be applied manually with a special microsyringe and microdispenser or with an automatic applicator. Umform distribution of the sample is a prerequisite for high resolution chromatography and correct quantification performed by TLC scanning (see Subheading 1.2.). Best chromatographic results are generally obtamed by delivery of samples to TLC plates with a commercially available automatic applicator (19). GSLs separate on slltca gel coated TLC plates because of polarity dlfferences caused by distinct length and composition of the carbohydrate moiety, fatty acid substltutlon of the ceramide portion and, m case of gangllosides, number and position of siahc acid residues. Most often used solvents for neutral GSLs and ganghosldes are mixtures of chloroform, methanol, and aqueous salts. Ammonium hydroxide in the aqueous phase changes the migration of several ganghosides, and this alkalizatlon is particularly useful to dlstmgulsh Neu5Ac from Neu5Gc substituted gangliosides (29,30). Mobtlities and compactness of ganglioside bands are markedly altered by the presence of salts. Halogen salts m the aqueous phase were found to result in excellent separations, while nonhalogen salts gave poor resolutrons (31) It 1s assumed, that strongly ionized cations effectively associate gangliosides and that such ganglloslde-ion complexes are Jointly separated on silica gel, whereas for TLC of neutral GSLs, no salts are needed. Chromatography is usually performed m a standard developing tank, saturated with a gas atmosphere of the solvent. It is advisable to seal the tank by means of a weight on the top of the glass cover. Sealants should be used sparsely. A stainless steel frame is favorable for exact vertical TLC. 1.2. Staining and Quantification After separation all GSLs can be visualized with a number of general carbohydrate stams. Orcmol-sulfuric spray is the most commonly used for detectlon of lipid bound carbohydrates (32) (see Fig. 1A). Neutral and sulfated GSLs, as well as ganghosides, will give pmkish-violet spots on a white background, whereas the color intensity is proportional to the number of monosaccharides m the carbohydrate moletles. Specific detection procedures for staining of gangliosldes on thm-layer chromatograms have been reviewed m detail by several

MOthing

186 LacCer c

abcdefg

h

Fig. 1. TLC immunostain of neutral GSLs with anti-nLcOse&er antibody. (A) Orcino1 stain (see Subheading 3.2.); (B) immunostain (see Subheading 3.3.). Neutral GSLs were chromatographed in solvent 1 (see Subheading 2.1.). Lane a: 10 pg of neutral GSLs from human granulocytes; lanes b-h 1 pg each LacCer (b), GbOse$er (c), GgOsesCer (d), GbOse&er (e), nLcOse&er (f), Forssman GSL (g), GgOse&er (h). LacCer, lactosylceramide; nLc4, nLcOse&er; nLc6, nLcOse&er. For details see ref. 42. Structures of neutral GSLs are listed in Table 1.

.3G&NeuSAc) ‘lG,gw-=)

G MI GlL G Dlh G Tlb-

: .,:

abed

I

abed

Fig. 2. TLC immunostain of G&NeuSGc) in ganglioside tractions of hybridoma cell lines. (A) Resorcinol stain (see Subheading 3.2.); (B) immunostain (see Subheading 3.3.). Gangliosides were chromatographed in alkaline solvent 3 (see Subheading 2.1.). Lane a: 30 pg of human brain gangliosides; lanes b-d 10 pg each of hybridoma M-2E6 (b), BT190/3 1 (c) and E4 (d). For details see ref. 30. Structures of gangliosides are listed in Table 2,

authors (3,23-25’. Resorcinol-hydrochloric acid spray is specific for sialic acid and, therefore, remains the most selective and sensitive tool for ganglioside detection on TLC plates (33) (see Fig. 2A).

TLC of Glycosphmgolrprds

187

For quantlficatlon of relatively small amounts of GSLs, direct densltometric scanning of stained bands 1s a simple but rapld method (3,27) and has been proved to be convenient for a reproducible quantlficatlon ofpicomolar amounts of mdivldual ganghoslde species (34). 1.3. TLC lmmunosfaining The TLC lmmunostaming technique 1san easy,rapid, and sensltlve method to investigate specific binding of ligands to separated GSLs (35). Contrary to glycoprotems, GSLs contain one hapten ollgosaccharlde per molecule and, therefore, once purified and structurally characterized, the GSL represents the functionally active carbohydrate sequence for the protem receptor under study A huge variety of related overlay methods for direct bmdmg of antibodies, toxins, lectms, and other proteins as well as related compounds have been developed as recently reviewed by Mtithing (26) After chromatography of GSLs (see Subheading 3.1.) glass-backed HPTLC plates need plastic fixation. The rationale for plastic coating 1sto prevent flakmg of silica gel from the support during the incubation and washing steps. The hydrophobicity of the plate protection is believed to induce a slmllar presentation of GSLs as m the plasma membrane. The plate 1sthen overlayed with the primary anti-GSL antibody followed by incubation with the secondary labeled antibody and dye solution for color development. A companion plate 1s chromatographed in the same tank under identical conditions and subjected to chemical detection (see Subheading 3.2.; see Figs. 1 and 2). The procedure described (see Subheading 3.3.) has been published by Bethke et al. (36) and was used with some modifications (37). Several welldefined carbohydrate-specific antibodies have been reported (38,39) and some are now commercially available. In some cases It is convenient to analyze gangllosides by combined neuramimdase treatment followed by lmmunostaming of respective asialo-gangllosides (#U-42). The detection limit of respective GSLs, which as a general rule spans a range from 1 to 100 ng, depends on several factors, e.g., dilution of the primary antibody, generally the use of a monoclonal antibody or a polyclonal antiserum or antibody-antigen avidity. 2. Materials 2.1. Application of Samples and Development 1. Silica gel 60 precoatedglass-backedHPTLC plates without fluorescent mdlcator for nano-TLC (10 x 10 cm, thxkness 0.2 mm, Cat. No 5633, E Merck, Darmstadt, Germany) 2 Prewarmed oven at 110°C 3. Silica gel with moisture mdlcator for drying (blue gel; Merck Cat No. 101925).

188

Mmng

4. Phosphorous pentoxtde (SICAPENT, Merck Cat No. 100543) drying reagent for destccators (see Note 1). 5 Desiccator for storage of heat-acttvated TLC plates 6 Template for sample apphcatton (homemade) 7. 5 $ or 10 pL High-precision mlcrosyrmge (Hamrlton Bonaduz AG, Bonaduz, Switzerland) 8 Microdrspenser (PB600- 1, Pat. No 3 16 1323, Hamrlton) 9 Automatrc TLC apphcator AS30 (Desaga, Heidelberg, Germany) 10 Nitrogen cylinder when automatic TLC apphcator IS used. 11, Chloroform (see Note 2) of analytical grade (preferably drstllled d avatlable) 12. Methanol (caution: highly flammable) of analytical grade (preferably distrlled tf available). 13 Detomzed water (preferably drstrlled rf available) 14 20 mM Aqueous calcmm chlorrde 15 2.5M Ammomum hydroxrde 16. Solvent 1 for separation of neutral GSLs* Chloroform/methanol/water (120/70/ 17, each by vol) 17 Solvent 2 for separation of ganghosides Chloroform/methanol/20 mA4 CaClz (120/85/20, each by vol) 18. Solvent 3 for separation of ganghosides: Chloroform/methanol/20 mM CaC& m 2 5MNHdOH (120/85/20, each by vol). 19 Standard developmg tank with mner drmensrons length 21 cm; wtdth 9 cm, hetght 21 cm (Desaga), lined wrth filter paper. 20 Srlicone grease (Wacker-Chemte GmbH, Munchen) 2 1 Tank cover wetght (4 kg, homemade) 22 Stainless steel frame for vertical development of HPTLC plate (homemade) 23 Chemtcal fume hood 24. Han dryer or vacuum pump.

2.2. Staining and Quantification 1 Orcmol (3,5-drhydrotoluene-monohydrate; Merck Cat No. 820933; protect from light) 2 Sulfuric acid (95-97%; Merck Art No. 10073 1; see Note 3). 3 Resorcmol (Merck Cat No 107590). 4. Hydrochloric actd (fummg 37%; Merck Cat No. 100317, cautron: causes burns) 5 0. 1M aqueous copper sulfate. 6 Protectrve frame (spray box; Desaga). 7 Fine mist sprayer SGl (Desaga) 8. 10 x 10 cm clean glass cover plate. 9 Clamps 10. Prewarmed oven at 100°C. Il. Densitometer CD60 (Desaga).

189

TLC of Glycosphingolipids 2.3. TLC Immunostaining

1 Phosphate-buffered salme: 140 mMNaC1,2 7 mA4KC1,8 mMNa2HP04, 1 5 mM KH2P04, adjusted to pH 7.3 2. Fatty acid free bovine serum albumin, receptor grade, lyophilized (BSA; Serva Fembiochemica GmbH & Co. KG, Heidelberg, Germany, Cat No. 11924) 3. Tween-21 (ICI Specialty Chemicals, Essen, Germany; Cat. No. A:3506). 4 Solutton A PBS supplemented with 1% BSA (v/w), freshly prepared 5 Solution B: PBS with 0 05% Tween 21 (v/v), freshly prepared 6 Anti-GSL primary monoclonal antibody or polyclonal antiserum (see Note 4). Polyclonal antibodies used in this protocol: (a) Anti-nLcOse&er (see Fig. 1) and (b) anti-G&NeuSGc) antisera (see Fig. 2), both from chicken Dilution’ 1.1000 m solution A 7 Alkaline phosphatase-conjugated secondary antibody directed against the Ig species of the first anti-GSL antibody. Used in this protocol. AftImPure rabbit antichicken IgY (IgG, H + L, 0 6 mg/mL, contains 15 mg/mL bovine serum albumin stabilizer and 0.05% sodium azide as preservative; Dianova GmbH, Hamburg, Germany; Cat No 303-055-033) Dilution: 1.2000 m solution A 8 0 1MGlycme buffer (pH 10 4) supplemented with 1 miVZnCl* and 1 mMMgC12. 9 5-Bromo-4-chloro-3-mdolylphosphate tolmdme salt (BCIP, Btomol Femchemirahen GmbH, Hamburg). 10. 0.05% (w/v) BCIP m glycme buffer (freshly prepared) 11 Bottle-top dispenser l&50 mL (Brand GmbH+Co , Wertheim, Germany) 12 Chromatographed HPTLC plate (see Subheading 3.1.) 13 n-hexane (caution: highly flammable) 14. Polyisobutylmethacrylate (Plexigum P28, Rohm, Darmstadt, Germany) 15 Plexigum fixatton solution* add 15 g of Plexigum beads under stirrmg to 400 mL n-hexane and contmue stirring for 45 mm Decant and use the supernatant for silica gel fixation (see Subheading 3.3.; see Note 5) 16. Small developing tank plus cover weight (2 kg, homemade) for plastic fixation of HPTLC plates Inner dimensions: length, 11.5 cm; width, 6.0 cm; height, 12.5 cm (Desaga). 17 HPTLC plate incubation chambers (plexiglass, homemade) Inner dimensions. length 10 5 cm; width 10 5 cm; height 3 0 cm. 18. Desiccator. 19 Phosphorus pentoxide (SICAPENT, Merck, see Note 1).

3. Methods 3.1. Application

of Samples and Development

of Chromatograms

1. Before use, activate the plates for 30-45 minutes at 110°C cool down and then store over heat actrvated silica gel with moisture indicator and auxiliary P,05 m a desiccator (several months, see Note 6) 2. Add 100 mL of desired solvent (see Subheading 2.1.) to the glass tank lined with filter paper and cover it tightly with a top weight (see Note 7).

190

Miithing

3 Allow the vapors m the tank to equilibrate for at least 3 h, preferably overnight 4 Place an activated HPTLC plate mto an apphcatlon template and spot the GSLsample with a glass microsyrmge as a series of mlcrodroplets with a mlcrodispenser to form a 5 mm band 10 mm parallel to the bottom of the plate, being careful not to scratch the slhca gel. 5 Alternatively, apply the probes to an HPTLC plate with an automatic TLC applicator 6. Place the plate, fixed m a stainless frame, uprlght m the tank, close it properly and allow the plate to remam undisturbed 7 Chromatography 1s fimshed when the ascending solvent front has reached a lme approx 1 cm below the top of the plate 8. Take out the plate and allow the solvent to evaporate m a fume hood (see Note 2) at ambient temperature 9 Before mltiatlon of the staining procedure (see Subheading 3.2.), traces of solvent are removed by gentle heatmg with a hair dryer or preferably by incubating the plate 5 mm under vacuum.

3.2. Staining and Quantification 1 Prepare 0.2% orcmol reagent by dtssolvmg 0.2 g orcmol in 100 mL H2S04/H20 (3 1, v/v) (see Note 8) 2 Resorcmol-HCl-Cu*+ spray 1sprepared by mixing 10 mL of 2% aqueous resorclno1 stock solution with 80 mL concentrated HCl and 0.25 mL of 0 1M CuSO, Fill up to a total volume of 100 mL with water (see Note 9) 3 Place the developed plate (see Subheading 3.1.) upright m a protective frame 4. Initiate the spray stream off the side from the plate surface, moving m a zigzag pattern beyond the chromatogram until the plate IS adequately moistened (see Note 10) 5 In case of resorcmol-staining, cover the HPTLC plate before heating. Clamp the chromatogram and the clean glass plate of equal size together and place it m an oven at 100°C Neutral GSLs will become visible after 10-15 min as pinklshviolet and ganghosldes after 20-30 mm as blue-violet spots. Orcmol and resorclno1 stained chromatograms of neutral GSLs and ganghosldes are shown in Figs. 1A and 2A, respectively (see Notes 11 and 12) 6 Transfer the stained chromatograms to a fume hood and allow the acids to evaporate. 7 Remove residual acids by incubating stamed plates under vacuum for 5 mm before transfer to the densitometer 8, Colored spots are scanned at 440 nm and 580-620 nm for orcmol- and resorclnol-stained chromatograms, respectively, and compared with known amounts of authentic GSL standards on the same plate (see Note 13).

3.3. TLC lmmunostaining 1 Put the chromatographed unstained HPTLC plate (see Subheading desiccator and dry over P205 for 30 mm under vacuum.

3.1.) into a

TLC of Glycosphingol/pids

191

2 Add 20 mL of Plexlgum fixation solution (see Subheading 2.3.) mto a small glass tank and cover It tightly with a top weight (see Note 7) 3 Place the thoroughly dried HPTLC plate immediately m the tank with the fixation solution as follows. use a pair of long forceps, grasp the top of the plate and place it mto the tank oriented with the spotted samples above the level of the fixation solution. Allow the top edge to lean against the rear of the tank. 4 Chromatograph until the solvent front has reached the top Take the plate out and control the success of fixation (plastic front) facing an electric bulb. In case of incomplete coating repeat step 3 immediately 5. Remove the plate and put It uprlght under a fume hood to completely an dry (see Note 14). 6 Transfer the silica gel fixed TLC plate mto an mcubatlon chamber (see Subheading 2.3.) and overlay with 40 mL of solution A for 15 min (see Note 15) This and all the followmg steps are performed at room temperature 7. Remove solution A by suction and incubate with 40 mL of the diluted primary anti-GSL antibody (see Subheading 2.3.) for 1 h 8 Remove first antibody by suction and wash the plate thoroughly and gently four times with solution B usmg a dispenser 9 Cover the plate with 40 mL of the diluted alkaline phosphatase-labeled secondary antibody for 1 h 10 Remove secondary antlbody by suction and wash twice with solution B, followed by twofold washing with 37°C prewarmed glycme buffer (see Note 16) 11. Take off the glycme buffer and overlay the plate with 40 mL 0 5% BCIP (w/v) in glycme buffer at 37°C. A blue indigo-like stable stain appears within l-3 h 12 Wash carefully with glycme buffer (see Note 17), remove the plate and place It upright on filter paper and allow to dry TLC immunostamed chromatograms of neutral GSLs and ganghosldes with polyclonal chicken anti-nLcOse&er and antlG&NeuSGc) are shown in Figs. 1B and 2B, respectively

4. Notes 1 P205 is corrosive and causes bums. Handle and open contamer with care. Do not breathe dust Never add water to this product. 2 Chloroform bears possible risks of irreversible effects. May cause birth defects Danger of serious damage to health by prolongued exposure Cancer suspect. 3 Causes severe burns Never add water to this product 4 Anti-GSL antibodies are available through several suppliers (mcomplete list) a American Type Culture Collection (ATCC, Rockvllle, MD); provides several hybrldomas producing anti-GSL monoclonal antlbodles b GlycoTech (Rockville, MD) c Matreya Inc. (Pleasant Gap, PA) d Pallmann GmbH (Munchen, Germany). e. Wako Chemicals (Dallas, TX) 5 This fixation solution has to be stored in a properly closed flask and 1s stable for several months at ambient temperature

Miithing 6. Exhausted sihca blue gel changes Its color from dark to light blue, and P205 becomes blue and gets lumpy. Both should be regularly replaced by freshly heatactivated gel and fresh P,05 to guarantee water-moist free atmosphere 7. Sealants should be used, but sparsely Otherwise, silicone grease smears mto the solvent resultmg in insufficient chromatographies 8 The orcmol reagent is stable for at least 2 wk and should be stored at +4”C m the dark 9 The 2% aqueous resorcmol stock solution and the resorcinol-HCl-Cu2” spray reagent are both stable for at least 2 wk and should be stored at +4”C m the dark 10. Orcmol and resorcmol spray are aggressive reagents because of its high concentration of H2S04 and HCl, respectively, and should be used m a chemical fume hood or m a spray box with ventilation Both solutions have to be stored m properly closed flasks 11 With the orcmol reagent, the plate should be sprayed thoroughly until becoming obviously moist. If stammg IS weak, spraying and heating may be repeated 12 Usmg the resorcmol spray, the plate should be moistened moderately to prevent runnmg of the dye. Covermg the TLC plate with a glass plate, the acid cannot evaporate resultmg m diffuse stammg when too intensive moistened 13 Gangliosides scanned with a densitometer using HPTLC plates (Merck, Cat 5633) give linear response m the range of 0 15-3 0 ug siahc acid per band (34) 14. Commercially available TLC plates with plastic-support do not need fixation However, because of higher resolution of GSLs the author prefers glass-backed silica gel HPTLC plates. This type of plate needs fixation to prevent flaking of the silica gel The plastic coating obtamed by chromatographmg the plate in the Plexigum solution IS highly recommended instead of dipping the plate in fixation solution (22,26) The latter procedure sometimes leads to shedding of the plastic-fixed silica gel layer. Usually, before starting with TLC immunostammg assays, check the batch of HPTLC plates since not all batches of plates are likewise appropriate, particularly m keepmg the silica gel on the glass support after fixation 1.5. BSA is used to prevent unspecific antibody bmdmg to the TLC plate Alternatively, e.g., gelatm can be used for protein blocking The author prefers BSA blocking, particularly in case of polyclonal antibodies Since our polyclonal antibodies are raised m ammals by immumzmg with GSLs coupled to permethylated BSA, these antisera also recognize BSA to a limited extent Using BSA throughout TLC immunostammg, a “negative stam” of unrecognized GSLs can be observed because of low background staining This reaction is favorable m identifying antibody specificities (see Fig. 2). 16 The washmg step with glycme buffer is recommended to remove residual phosphate from the PBS washing steps, because alkaline phosphatase activity is diminished m the presence of phosphate ions The reaction IS faster at 37°C than at room temperature 17 Flakmg of the plastic-fixed silica gel layer has been observed using water for final TLC washmg.

TLC of Glycosphingoliprds

793

References 1. Symmgton, F W., Murray, W. A , Bearman, S. I., and Hakomori, S-I. (1987) Intracellular localtzation of lactosylceramide, the major human neutrophll glycosphmgoliptd J Bzol Chem 262, 11,356-l 1,363 2 Glllard, B K , Thurmon, L T., and Marcus, D M (1993) Variable subcellular locahzatton of glycosphmgohptds. Glycobiology 3,57-67. 3. Ledeen, R W, andYu, R K. (1982) Ganghosides. structure, tsolation, and analysts Methods Enzymol 83, 139-191. 4. Karlsson, K A. (1989) Animal glycosphmgohpids as membrane attachment sites for bacteria Ann. Rev Blochem 58,309-350 5. Igarasht, Y., NoJiri, H , Hanat, N., and Hakomort, S.-I. (1989) Ganghostdes that modulate membrane protein function. Methods Enzymol 179,52 l-541 6. Stults, C. L M., Sweeley, C. C., and Macher, B. A (1989) Glycosphmgohptds: structure, biological source, and properties Methods Enzymol 179, 167-2 14. 7. Schwarzmann, G and Sandhoff, K. (1990) Metabolism and mtracellular transport of glycosphmgollplds Bzochemzstry 29, 10,865-lo,87 1 8 Zeller, C. B. and Marchase, R B (1992) Ganghostdes as modulators of cell function Am J Physzol 262, C134l-Cl355 9. IUPAC-IUB Commtsston on Biochemical Nomenclature (1977) The nomenclature of lipids Eur J Bzochem 79, 1 I-21 IO Svennerholm, L. (1963) Chromatographic separation of human bram gangliostdes J Neurochem l&613-623. 11. Reuter, G. and Schauer, R. (1988) Suggesttons on the nomenclature of siahc acids Glycoconj J 5, 133-135. 12 Momoi, T., Ando, S , and Nagal, Y (1976) Hugh resolution preparative column chromatographtc system for ganghostdes using DEAE-Sephadex and new porous slhca, latrobeads BlochIm. Blophys Acta 441,488-497 13 Iwamon, M. and Nagat,Y (1978) A new chromatographtc approach to the resolution of individual ganghosldes Gangliostde mapping Biochlm Bzophys Acta 528,257-267 14. Kundu, S. K , Chakravarty, S K., Roy, S. K , and Roy, A. K (1979) DEAE-silica gel and DEAE-controlled porous glass as ion exchangers for isolation of glycolipids. J. Chromatogr 170,65-72. 15. Itoh, T , Lt, Y.-T, Lt, S -C, and Yu, R. K (198 1) Isolation and charactertzatton of a novel monosialosylpentahexosyl ceramtde from Tay Sachs brain J 5~01 Chem 256,165-169. 16 Ando, S , Wakt, H , Kon, K., and Kishimoto, Y (1987) Up-to-date chromatography ofganghosides, m Ganglzoszdes andModulatlon ofNeurona1 Functions (Rahmann, H , ed.), Springer Verlag, Berlin, pp 167-177 17 Tanaka, T., Arai, Y , and Krshimoto,Y. (1989) Characterization and regional dlstributton of individual ganghosides m goldfish central nervous system. J Neurochem 52, 193 l-1936 18 Hnabayashi, Y , Nakao, T , and Matsumoto, M. (1988) Improved method for largescale purification of brain gangltostdes by Q-Sepharose column chromatography. J. Chromatogr 445,377-384.

194

Miithmg

19. Muthmg, J and Unland, F. (1994) Improved separation of isomertc ganghostdes by amon-exchange high-performance liquid chromatography J Chromatogr B 658,39-45 20 Saito, T and Hakomori, S.-I ( 197 1) Quantitative isolation of total glycosphmgohpids from animal cells J Lzpzd Res 12,257-259

21 Ueno, K , Ando, S., and Yu, R K (1978) Ganghosides of human, cat, and rabbit spinal cords and myelm J Lzpzd Res 19,863-87 1 22 Kmep, B and Ebel, F (1993) Lipids and glycoliprds m Methods of lmmunologzcal Analyszs (Masseyeff, R F, Albert, W H , and Staines, N A, eds ), VCH Verlagsgesellschaft, Wemheim, pp 90-105 23. Skipskt, V P (1981) Thm-layer chromatography of neutral glycosphmgohpids Methods Enzymol 35,396-425 24 Kundu, S K. (198 1) Thm-layer chromatography of neutral glycosphmgohpids and ganghosides Methods Enzymol 72, 185-204 25 Schnaar, R L. and Needham, L K (1994) Thin-layer chromatography of glycosphmgolipids Methods Enzymol 230,37 l-389 26 Muthmg, J (1996) High resolution thm-layer chromatography of gangliosides J Chromatogr A 720,3-25 27 Ando, S , Chang, N -C., and Yu, R K. (1978) Hugh-performance thin-layer chromatography and densitometric determmatron of brain ganglioside compositions of several spectes Anal Bzochem 89,437--450 28 Muthing, J. (1994) Improved thm-layer chromatography separation of ganghosides by automated multiple development, J Chromatogr B 657, 75-8 1 29. Nagai, Y and Iwamon, M (1980) A new approach m the analysis of ganglioside molecular species, m Structure and Function of Ganglzoszdes (Svennerholm, L , Mandel, P , DreyfuD, H., and Urban, P F., eds ), Plenum Press, New York, NY, pp. 13-21 30. Muthmg, J , Steuer, H , Peter-Katalinic, J , Marx, U , Bethke, U ,Neumann, U , and Lehmann, J (1994) Expression of ganghosides G&NeuAc) and G,,(NeuGc) m myelomas and hybndomas of mouse, rat, and human origin. J. Bzochem 116,64-73. 3 1 Ando, S , Wake, H , and Kon, K (1987) New solvent system for high-performance thin-layer chromatography and htgh-performance liquid chromatography of ganghostdes. J Chromatogr 405, 125-134 32. Svennerholm, L (1956) The quantttatlve estimatton of cerebrosides in nervous tissue J Neurochem 1,42-53 33 Svennerholm, L. (1957) Quantrtatrve estimation of stahc acids. Bzochzm Bzophys Acta 4,6046 11 34 Mullm, B R , Poore, C. M B., and Rupp, B. H (1984) Quantitation of ganghosrdes m the picomolar range J Chromatogr 305, 5 12,5 13 35. Magnam, J L., Brockhaus, M , Smith, D. F., and Ginsburg, V (1982) Detection of glycohptd ligands by drrect bmdmg of carbohydrate-bmdmg protems to thin-layer chromatograms. Methods Enzymol. 83,235-241 36 Bethke, U , Muthmg, J., Schauder, B , Conradt, P , and Muhlradt, P F. (1986) An improved semi-quantitative enzyme unmunostainmg procedure for glycosphmgo-

TLC of Glycosphmgohprds

37

38

39

40

41

42.

195

lipid antrgens on high performance thm layer chromatograms J Immunol Methods 89,111-116 Muthmg, J and Zrehr, H (1990) Rapid detection of extended gangho-senes ganghosrdes with termmal GalNAcP l-4Gal sequence on high performance thm layer chromatography plates Boomed Chromatogr 4, 70-72 Ferzi, T. (1985) Demonstration by monoclonal antibodies that carbohydrate structures of glycoprotems and glycoliprds are onto-developmental antigens Nature 314,53-57 Magnani, J. L , Spttalmk, S. L , and Gmsburg, V. (1987) Antibodies against cell surface carbohydrates determmation of antigen structure Methods Emymol 138, 195-207. Sarto, M., Kasar, N , and Yu, R K (1985) In srtu rmmunologrcal determmatron of basic carbohydrate structures of ganghosrdes on thin-layer plates Anal Bzochem 148,54-58 Muthing, J and Muhlradt, P F (1988) Detection of ganghosrdes of the G,rb-type on high-performance thin-layer chromatography plates by m-ununostaming after neurammrdase treatment Anal Bzochem 173, 1O-1 7 Muthmg, J and Neumann, U (1993) Selective detectton of termmally 1~2-3 and a2-6 sralylated neolacto-series gangliosides by mnnunostammg on thm layer chromatograms Bzomed Chromatogr 7,158- 16 1

13 Strategies for Preliminary Characterization of Novel Amphoteric Glycosphingolipids Roger D. Dennis, Giinter Lochnit, and Rudolf Geyer 1. Introduction The generally accepted paradigm for the classification of glycosphingohpids into neutral and acidic classeshas proved to be an oversimpltfication, at least, when considering members of the Invertebrata. The detailed occurrence and frequency of electrtcally neutral, but amphotertc moiety-contammg glycosphingohptds amongst invertebrates would justify the establishment of a third class, namely, zwittertomc glycosphingoliptds. Zwtttertonic glycosphingolipids have been characterized in members of the phyla Annehda, Arthropoda (Crustacea), Arthropoda (Insecta), Mollusca (freshwater Bivalvia), Mollusca (marine Gastropoda), and Nematoda (Ascaridida), whereby the sugar-amphoterrc moiety has been identrfied, respectively, as Gal-phosphocholme (I-#), Glc-phosphonoethanolamme (S), GlcNAc-phosphoethanolamme (6-8), Manphosphoethanolamine (9), Gal-phosphonoethanolamme (10,11), and Hex/ HexNAc-phosphocholine (12,13) (exact sateof attachment to be determined). This diversity of zwitteriomc glycosphingolipids amongst investigated members of the Invertebrata points to then brological Importance, but as yet, unknown functional significance. In the course of our studies on the zwtttertonic species of the insect, Calliphora vicina, and the parasitic nematode, Ascaris suum, we have developed schemes for their rsolatron, classrficatron, and preliminary characterization, which will be presented in this chapter. The glycolipids studied have originally been isolated, as for glycosphmgoliprds, however, until proved chemically or enzymatically to be of this third lipidclass, e.g., cleavage by the enzyme endoglycoceramidase, they will be desrgnated as glyco(sphmgo)llpids in this chapter. The shorthand designations for these categories of lipid will be N (neutral)-, Nz (zwttterionic)-, and A (acidic)From

Methods Edlted

by

m Molecular E F Hounsell

B/ology,

Vol 76 Glycoanalysts

0 Humana

197

Press

Inc , Totowa,

Protocols NJ

Dennis, Lochnit, and Geyer

198

I

Ascaris suum Exuactton Extractton Extraction

Glyco(sphmgo)hpld

(step1) wtth CMW (10 10 1, v/v/v) with CMW (30 60 8, v/v/v) wth IHW (55 20 25, v/v/v)

raw extract

(step

2)

Acetone-wash Dtalysls Glyco(sphmgo)hptd

extract

(step 3)

DEAE-Sephadex Dlalysts

Neutral and mttertomc

glyco(sphmgo)ltplds

chromatography

(step 4)

Actdic glyco(sphtngo)hptds

Gel filtration

I Neutral glyco(sphmgo)hplds

(step 5)

Zwttenomc

Peracetylatlon Flons~l chromatography

glyco(sphmgo)hplds

(step

6)

S~hca gel chromatography

Deacetylatlon Preparattve HPLC Indtvtdual

N-glyco(sphmgo)hptds

Indtvtdual

Nz-glyco(sphmgo)ltplds

Scheme 1. Generalprotocol for the lsolatron and fractronatron of neutral (N-) and zwitterionic (Nz-) glyco(sphingo)hplds, as exemplified by the parasitic nematode, Ascarls suum.

glyco(sphingo)lipids. The tsolation and fracttonatton of A. suum Nzglyco(sphmgo)hprds was achieved by the following protocol (see Scheme 1) The tissue was extracted three ttmes wtth different combmations of organic solvent and rotary evaporated to dryness (see Subheading 3.1.1., step 1). The raw extract was washed with cold acetone to remove most of the trrglycerrdes, prior to dialysis (see Subheading 3.1.2., step 2). The drred preparatton was segregated into Its N-/Nz- and A-glyco(sphmgo)lrpid fractions by DEAESephadex, anion-exchange chromatography, followed by dialysis (see Subheading 3.1.3., step 3). The dried, N-/Nz-glyco(sphingo)hprd fraction was

Novel Amphoteric Glycosphingolipids

199

separated as to its N- and Nz-glyco(sphmgo)lipid fractions by adsorption column chromatography on silica gel (see Subheading 3.1.4., step 4). The N-glyco(sphingo)lipid fraction was further purified by peracetylation, Florisil column chromatography, deacetylation and silica gel chromatography, prior to fractionation mto mdrvtdual components on a column of porous silica gel (Iatrobeads) by high-performance hquid chromatography (HPLC) (see Subheading 3.1.5., step 5). The Nz-glyco(sphingo)hpid fraction was directly fractionated mto mdividual components by silica gel column chromatography (see Subheading 3.1.6., step 6). The preliminary, structural charactertzation of the total and individual zwittertonic glyco(sphingo)lipids was undertaken according to: their high-performance thin-layer chromatography (HPTLC)-migration properties and spray reagent characteristics; the conclusions from specific, chemical and enzymatic analyses; and, for the HPLC-chromatographic differentiation of then isolated, pyridylaminated oligosaccharides (PA-ohgosaccharides), released by endoglycoceramidase-cleavage. The isolation of the glyco(sphmgo)lipid-derived ohgosaccharide chains as their PA-derivatives allowed. their manipulation as water-soluble molecules; then detection at high sensitivity, i.e., in the subpmole range, and, facile application of the methodologies developed for the structural analysis of complex carbohydrates. 2. Materials 1. Sol A* chloroform/methanol/water (30:60*8, v/v/v); Sol B: chloroform/methanol/O&V aqueous sodium acetate (30:60 8, v/v/v). 2 Rotary evaporation. performed under reduced pressure at 37°C m the presence of 1O-20% (v/v) 2-propanol 3 Dralysrs tubing wrth approx 10 kDa molecular weight cut-off from Serva (Hetdelberg, Germany) 4. DEAE-Sephadex A-25 Cl- form (Pharmacta, Freiburg, Germany) converted to CH$OO- form for use swell 100 g m 2 L distilled water 24 h at room temperature; wash four times with 1 L sol B; leave in last wash 48 h at 4°C wash four times with 1 L sol A and check for complete salt removal (evaporation of an altquot from the final column wash, on a glass plate, for the detection of salt crystals) Store in 500 mL sol A at 4°C. 5. Sthca ge&a (70-250 mesh; Merck, Darmstadt, Germany). 6 HPTLC-Silrca geleO plates purchased from Merck For HPTLC, drssolve glyco(sphmgo)hplds at 2 pg/pL m either chloroform/methanol/water (10.10: 1, v/v/v) or (60:35.8, v/v/v). For reproducibrhty, perform HPTLC according to Nores et al. (I+), suspend the plate m a tank and saturate the atmosphere for 15 mm by a motor-driven propeller m the lid before lowering into the runnmg solvent. For the separation of Nz-glyco(sphingo)lipids use, l-dimensional HPTLC: chloroform/methanol/water (10: 10.3, v/v/v); 2-dlmensronal HPTLC: first dlmensron

Dennis, Lochnit, and Geyer

7.

8 9.

10

11.

12

13

chloroform/methanol/water (10,10:3, v/v/v), second dimension as acidic runnmg solvent chloroform/methanol/water/acetic acid (25.15.4.2, v/v/v/v), as alkaline running solvent chloroform/methanol/2.5% (w/v) aqueous ammonia (10*10*3, v/v/v), or as neutral running solvent chloroform/methanol/water (10 10.3, v/v/v) Dry the plates prior to spraying or mnnunostammg Spray reagents orcmol/HzS04, 0.2% (w/v) orcmol in 2MH2S04; Dragendorff’s reagent and molybdenum blue purchased from Sigma (Deisenhofen, FRG), nmhydnn, 0.3% ninhydrm (w/v) m 1-butanol, acidify 100 mL with 3 mL acetic acid, resorcmol/HCl, add 10 mL 2% aqueous resorcinol (w/v), 250 pL 100 mM aqueous CuSO,, and 10 mL distilled water to 80 mL concentrated HCl (leave at room temperature for 4 h prior to use; stable 15-20 d in dark at 4°C) For heating use a drying oven (preferably with ventilator) Mouse myeloma protem TEPC 15 (IgA, kappa, clarified ascites M7269, Sigma) and peroxidase-labeled second antibody (Sigma) Reversed-phase cartridge desalting. Chromabond C 18ec (Macherey-Nagel, Duren, Germany), pre-equilibrate m two cycles of 10 mL methanol and chloroform/methanol (2 1, v/v), and 10 mL methanol, equilibrate m 10 mL TUP (theoretical upper phase-chloroform/methanol/l 00 n-&Z aqueous KC1 [3:98*74, v/v/v]); take up the sample in 5 mL TUP, sonicate, load, wash with 5 mL TUP and 10 mL methanol/water (1: 1, v/v); elute with 5 mL methanol and 10 mL chloroform/methanol (2.1, v/v). Dry down under N2 stream Normal-phase HPLC 2-propanol/n-hexane/water (55:44.1, v/v/v; eluant A) and (55.35.10, v/v/v, eluant B) are degassed by saturation with helium and stored m closed, pressurized vessels with helium purgmg. Use a porous silica gel column (4.6 x 250 mm; Iatrobeads 6RS-8010, 10 pm; Macherey-Nagel). Store the column m the absence of water, m 2-propanolln-hexane (55 45, v/v) Silica gel column chromatography with linear gradient. eluant A, chloroform/ methanol/water (65.25.4, v/v/v) and eluant B, chloroform/methanol/water (5*70:25, v/v/v) HPTLC-Imunostammg coat plates by dippmg for 60 s m 0.5% (w/v) polyisobutylmethacrylate (Plexigum P28; Rohm and Haas; Darmstadt, Germany) in n-hexane/chloroform (9 1 [v/v], derived from a 5% [w/v] stock solution in chloroform), air-dry for 10 mm Block the plates at 37’C for 60 mm with blocking buffer, 1% BSA (w/v) in PBS (O.lM PBS, pH 7 2) Incubate the plates at 37°C for 2 h with the blocking buffer-diluted, first antibody (Sigma), wash plates SIX times with 5 mL of PBS. Incubate the plates at 37°C for 2 h with blockmg buffer-diluted, peroxidase-conjugated rabbit antimouse Ig antibodies (Dako Diagnostika, Hamburg, Germany), wash plates SIX times with 5 mL of PBS Develop the plates no longer than 15 mm at room temperature in the dark with the substrate solution: 20 mL PBS; 320 pL diammobenzidine (40 mg/mL 3,3’-diammobenzidme tetrachloride, Sigma); 80 & NiC12 (80 mg/mL); and, 4.8 pL 30% (w/w) H,Oz. Stop the reaction by a PBS-wash and air-dry the plates Water 1-butanol (water-saturated) phase-separation. make up sample to 500 PL with distilled water, extract twice with 500 pL I-butanol and phase-separate at

Novel Arnphoteric Glycosphingolipids

14. 15. 16 17. 18

19

20 21

22

23.

24

25

201

2500 rpm for 5 min Backwash organic phase twice with 500 pL distilled water and phase-separate Dry down the aqueous phase m a Speedvac and dry the organic phase under a N2 stream or by rotary evaporation under reduced pressure at 43°C. Phospholipase buffer: 100 mM Trrs-HCl buffer, pH 7 4, containing 0.1% (w/v) sodmm deoxycholate, store at 4°C Phospholrpases (Boehrmger-Mannheim, Mannheim, Germany). Sphmgomyelinase and endoglycoceramidase (Sigma). Teflon-lined, screw-capped, conical microvtals (Zmsser, Frankfurt, Germany). Couplmg reagent 2-ammopyndme in acetic acid (560 mg 2-ammopyrtdme m 200 @glacial acetic acid, 1.e ,73% [w/w] for quantitative pyrtdylammation, PA) (15), reducing agent N,N-dimethylamine borane m acetic acid (40 mg dimethylamme borane in 200 pL glacial acettc acid) Fractogel TSK HW-40 (F) (23@-450 mesh; Merck) equilibrate gel m 10 mA4 ammomum acetate buffer, pH 6.0 Add 1.2 mL 96% (w/v) acetic acid to approx 700 mL distilled water, adjust to pH 6.0 with 25% (w/v) ammonia and make up to 1 L with drstrlled water. Fluorescence spectrophotometer. excitation and emission wavelengths (320 nm and 400 nm, respectively); for NHz-HPLC at 3 10 nm and 380 nm, respectively. Anion-exchange HPLC: distilled water (eluant A) and 500 mM KH*PO, buffer, pH 4.4 (eluant B), filter through 0.4~um filter, degas and store in pressurrzed vessels (see Subheading 2., item 10). Use column (4.6 x 250 mm) of Micropak AX-5 (5 pm; Varran, Walnut Creek, CA) Reverse-phase HPLC usmg the condmons of Ohara et al (16) as modified by Lochmt and Geyer (I 7), 50 m&I acetic acid-trtethylamine buffer, pH 5 0 (eluant A) and the same, but containing 0.5% 1-butanol (v/v, eluant B), degas and store in pressurized vessels (see Subheading 2., item 10). Use column (4 6 x 250 mm) of ODS Hyperstl(3 pm; Shandon, Frankfurt, Germany). NH,-HPLC* 200 macetic acid-trxethylamme buffer, pH 7 3/acetomtrrle (25:75, v/v) (eluant A) and the same, but at 50:50 (eluant B), degas and store m pressurized vessels (see Subheading 2., item 10). Use MN-Carbohydrate column (4.6 x 250 mm; Macherey-Nagel) HPLC-equipment a gradient pump equipped with solvent conditioner, an mJector, a column oven and a fraction collector as basrc equipment; PA-oligosaccharide analysts required a fluorescence spectrophotometer and integrator All chemicals and organic solvents purchased were of analytical grade,

3. Methods

3.1. Nz-glyco(sphingo)lipid

Isolation and Fractionation

3.1.1. Step 1 1 Pulverize undamaged, washed worms (800 g wet weight) m a Waring blender and lyophrbze.

202

Dennis, Lochnit, and Geyer

2. Perform consecutive extractions. twice m chloroform/methanol/water (10: 10.1, v/v/v), somcate for 10 mm and extract for 30 mm at 50°C; once in 1500 mL sol B, sonicate for 10 min and extract overnight at 4”C, twice m 2-propanol/n-hexanelwater (55:20 25, v/v/v), sonicate for 10 min and extract for 30 mm at 50°C. 3 Following each extraction procedure, centrifuge at 3000g for 10 mm and retain the pellet for further extraction 4 Rotary evaporate the combined extracts to dryness under reduced pressure at 37”C, desiccate overmght over P205

3.1.2. Step 2 (see Note 2) I Treat the raw extract with 600 mL acetone for 2 h at 4°C. 2. Separate and discard the supernatant on centnfugatlon at 3000g for 10 mm 3. Dissolve the retained pellet in 40 mL chloroform/methanol/water (10: 10.1, v/v/v), somcate and dialyze against distilled water for 3 d at 4°C (four changes of 5 L) 4. Extrude the contents of the dialysis tubing, sonicate, and rotary evaporate to dryness

3.1.3. Step 3 (see Note 3) 1. Dissolve and re-extract the preparation in 80 mL and 45 mL sol A with somcatlon, respectively; clear the solution each time by centrifugation at 3000g for 10 mm. 2 Apply the pooled samples to a (30 x 100 mm) column of DEAE-Sephadex at a slow drop-rate of 1-2/s 3. Wash the column with 500 mL sol A to obtam the N-/Nz-glyco(sphmgo)hpld fraction 4 Elute the column with 500 mL sol B to collect the A-glyco(sphingo)lipid fraction 5. Rotary evaporate both fractions to dryness, desalt the latter by dialysis against distilled water and rerotary evaporate to dryness.

3.1.4. Step 4 1 Dissolve the N-/Nz-glyco(sphmgo)llpld fraction m 5 mL chloroform/methanol (9: 1, v/v), sonicate and apply to a sihca gel column (30 x 150 mm) equrhbrated in chloroform/methanol (9.1, v/v). 2. Elute and prefractionate N-glyco(sphingo)hpids three times with 650 mL chloroform/methanol/water (65.25:4, v/v/v). 3. Elute the Nz-glyco(sphingo)hpid fraction with one cycle of 900 mL and 100 mL chloroform/methanol/water (10*70:20, v/v/v; use the latter step to ascertain the complete elution of components). 4 Rotary evaporate fractions to dryness, monitor for N- and Nz-glyco(sphingo) lipids by HPTLC (see Subheading 3.2., step 3) and where appropriate recombine subfractions.

3.1.5. Step 5 1 Desiccate the N-glyco(sphingo)lipld fraction overnight over PZ05. Peracetylate m 15 mL pyridine/acetic anhydnde (2: 1, v/v) for 18 h at room temperature in the

Novel Amphoteric Glycosphingoliplds

2. 3.

4.

5.

6.

203

dark and under argon, rotary evaporate together wtth 30 mL toluene to dryness and redestccate over P20, Take-up m 25 mL 1,2-dichloroethane (DCE)/n-hexane (4: 1, v/v) and apply to a Florist1 column (30 x 150 mm) equilibrated in DCE/n-hexane (4.1, v/v) Perform successive washes of 400 mL DCE/n-hexane (4: 1 and 1’ 1, v/v) and DCE; elute with 800 mL DCE/acetone (1.1, v/v), rotary evaporate to dryness and destccate overnight over P205 (see Note 4). Dtssolve residue m 20 mL methanol and deacetylate by addition of 2 mL 30% sodium methylate (w/v) for 3 h at room temperature. Neutralize with 10% (w/v), methanohc acetic acid, rotary evaporate and dialyze. Remove lipid contammants and free fatty acids on a silica gel column (30 x 50 mm) equtbbrated in chloroform/methanol (98:2, v/v) Dtssolve the sample in 5 mL chloroform/methanol (98:2, v/v), sonicate and apply to the column* subfractionate wtth 600 mL and 50 mL chloroform/methanol (98.2, v/v), three trmes 300 mL chloroform/methanol (95:5, v/v), 300 mL acetone/methanol (90.10, v/v), 300 mL chloroform/methanol (80:20, v/v), and 300 mL chloroform/methanol/water (10:70.20, v/v/v). Rotary evaporate to dryness, desalt on reverse-phase cartridge, momtor for N-glyco(sphmgo)ltpids by HPTLC and pool correspondmg fractions. To Isolate mdivtdual N-glyco(sphrngo)liprd components, drssolve the sample m 1.5 mL eluant A, sonicate and inject 500 pL onto the HPLC column. Elute at room temperature with a linear gradient from 100% eluant A to 100% B in 30 mm, continue wtth 100% eluant B for 10 mm. Collect 2 mL fractions at a flow-rate of 2 mL/min; re-equilibrate column to startmg condittons for 30 mm Dry the fracttons under N2 stream, monitor for glyco (sphmgo)ltptds by HPTLC (see Subheading 3.2., step 3), pool the corresponding fracttons and rotary evaporate to dryness (see Subheading 2., item 10 and Note 5).

3.1.6. Step 6 (see Subheading

2., /tern 11)

1 To Isolate mdtvidual Nz-glyco(sphmgo)hptd components, prepare a 10 x 500 mm silica gel column equilibrated in eluant A at room temperature. 2. Dissolve the sample m 5 mL eluant A, sonicate and apply to the column. 3. Eluate at room temperature with a linear gradient of 100% eluant A to 100% eluant B m 300 min, contmue run with 100 mL of eluant C (methanol/water [70:30, v/v]), to ensure complete elutton of glycoliptds. 4 Collect 4 5 mL fractions at a flow-rate of 1.5 mL/mm. 5 Dry the fracttons under N2 stream, monitor for glyco(sphmgo)hpids by HPTLC (see Subheading 3.2., step 3), pool the correspondmg fractions and rotary evaporate to dryness (see Note 6)

3.2. HPTLC Characteristics 1. Run one-dimensional HPTLC, observe bandmg patterns and mrgratron properties against standards of known structure, run two-dimensional HPTLC,

204

Dennis, Lochnit, and Geyer

t

b

2nddimension

Fig. 1. Two-dimensional HPTLC-separation of the N- and Nz-glyco(sphingo)lipid fractions of A. suum adult worms. HPTLC plates; running solvents, first direction: chloroform/methanol/water (10: 10:3, v/v/v), second direction: chloroform/methanol/ 2.5% (w/v) aqueous ammonia (10: 10:3, v/v/v); visualization: orcinol/H2S04 spray reagent, 10 min at 120°C. The neutral species are designated by the slope N and the zwitterionic species by the slopes Nz.

2. 3.

4.

5. 6. 7.

observe the divergent slopes of N- and Nz-glyco(sphingo)lipid fraction-components (Fig. 1). Routinely for lipids, develop in an Iz-saturated atmosphere for 30 min, positive as brown, fix staining temporarily by glass plate-cover. Routinely for sugars, i.e., glycolipids, spray heavily with orcinol/H2S04, develop at 110°C for 3-4 min (10-15 min; heating without ventilator), positive as bluepurple (see Note 7). For choline, spray uniformly with Dragendorff’s reagent, develop at room temperature for 15-20 min, positive as orange on yellow background. For the immuno-detection of phosphocholine, apply the mouse myeloma protein TEPC15 (IgA, kappa, clarified ascites; M 7269, Sigma) at a dilution of 1:2500 and the peroxidase-conjugated, second antibody at 1: 1000. For phospholipids, spray lightly with molybdenum blue, develop at room temperature for approx 10 min, positive as blue. For free-NH* groups, spray lightly with ninhydrin, develop at 110°C for l-2 min (5-l 0 min), positive as red-purple (see Note 8). For sialic acid, i.e., gangliosides, spray uniformly with resorcinol/HCl, develop at 110°C for 5-10 min (10-20 min) covered with a preheated glass plate, positive as blue-grey, remove plate to release HCI fumes (see Note 9).

Novel Amphotenc Glycosphlngolipids

205

3.3. Chemical and Enzymatic Analyses 3.3.1. Chemical Analyses 1. Dephosphorylate 170 pg of dried sample, m a cap-fitted polyethylene tube, with 150 pL 48% (w/w) HF for 24 h at 4°C (see Note 10 and Fig. 2) Remove excess reagent by Speedvac or under N2 stream and monitor glyco(sphmgo)hpids by HPTLC (evtdence for phosphate mono- and drester bonds) 2 Deammate 40 ug of dried sample, with 30 s somcatton and 10 s vortexing, m 200 pL freshly prepared 100 r&4 sodium acetate buffer, pH 4.0, contammg 0.1% (w/v) Nonidet P40 and 250 mM NaN02 for 10.5 h at room temperature Terminate the reaction with 6 pL 6M HCl, perform a water. I-butanol phaseseparation, rotary evaporate the organic phase to dryness and monitor glyco(sphingo)liptds by HPTLC (evidence for ammo sugars and/or phosphoethanolamme-contammg glycoliprds). 3. Saponify 60 pg of dried sample, with 30 s somcatron and 10 s vortexmg, in 500 pL 50% (v/v) aqueous methanol-25 pL 5M aqueous KOH for 4 h at 50°C Neutralize with 9 pL glacial acetic acid; perform a water: I-butanol phase-separation, rotary evaporate the organic phase to dryness and monitor glyco(sphmgo)hpids by HPTLC (evtdence for ester-linked fatty acids) 4 Release the sugar-zwttteriomc complex from 50 pg of the dried sample by mild actdtc hydrolysis, with 30 s somcatron and 10 s vortexmg, u-r 500 PL 1M methanolic HCl (derived from acetyl chloride) for 20 h at 80°C under N, Rotary evaporate to dryness, repeat twtce with 500 pL dtstdled water, perform a water. lbutanol phase-separation, dry the aqueous phase by Speedvac and monitor by HPTLC (running solvent, pyndme/ethyl acetate/acetic acid/water [6*3 1.3, v/v/ v/v]) for sugar-/phosphate-posltlve spots (6). 5 Release the sugar-zwttteriontc complex’s components from 50 pg of the drred sample by strong acidic hydrolysis, with 30 s somcatton and 10 s vortexmg, m 500 pL 6M aqueous HCl for 5 h at 100°C under Nz Rotary evaporate to dryness, repeat twice with 500 pL distilled water; perform a water. 1-butanol phase-separatton, dry the aqueous phase by Speedvac and monitor by HPTLC (running solvent, pyridme/ethyl acetate/acetic acid/water [6.3: 1:3, v/v/v/v]) for separate sugar- and phosphate-positive spots (6).

3.3.2. Enzymatic Analyses 1. Solubtltze dried samples (40-60 t.rg) by 30 s sonmatron and 10 s vortexmg m 100 pL enzyme buffer, incubate with enzyme for 16 h at 37°C. 2 For phosphohpase AZ, the buffer (see Subheading 2., item 14) contains 1 mMCaC1z and 50 U of bee venom enzyme 3. For phosphohpase C, the buffer (see Subheading 2., item 14) contams no CaC12 but 1 U Bacdlus cereus enzyme (phosphattdylrnositol-specific) 4. For phosphohpase D, the buffer (see Subheading 2., item 14) contains 2 5 mM CaC& and 10 p.L rabbit normal serum (as a source of glycosylphosphattdylmositol-specific enzyme).

[M+Na]

18

=I732

A 16

17.

15. 14. 13,

[M+H]-=171(

12. 11. 10. 9.

-c 150 25 p:

145.

B

140, 135, 130, 125. 120, 115. 11 0 105 100. 55 90 85. a0 75 70 65 60 55 50

800

lau

urn

12w

1300

1100

1530

1Em

1700

18m

m/z

Fig 2. Analysis of Nz-glyco(sphmgo)ltptds by matrtx-assisted laser desorptton/ tonizatton ttme-of-fhght mass spectrometry (MALDI-TOF-MS). A smgle Nz-glycosphingoliptd compound, obtamed from A suum by following the described fracttonatton protocol, was analyzed before (A) and after(B) HF-treatment by MALDI-TOF-MS m the positive-ton lmear mode using 2,5-dihydroxybenzoic acid as matrix The mass difference of m/z 165 reflects the release of one phosphocholine substnuent

Novel Amphoteric Glycosphu?goliplds

207

5. For sphmgomyelmase, add successively 40 pL chloroform, 600 pL dtethyl ether, 20 pL Triton X-100, 80 pL IA4 sodium acetate buffer, pH 7.6, 40 & 400 mA4 aqueous MgCl*, and 2 U Staphylococcus aureus enzyme Star, by magnetic flea, throughout the mcubatron period 6 For endoglycoceramtdase, 50 mM sodium acetate buffer, pH 5 0, contains 0 1% (w/v) sodium taurodeoxycholate and 0 25 mU Rhodococcus G-14-2 enzyme (see Note 11).

7 After mcubation, perform a water: 1-butanol phase-separation, rotary evaporate the organic phase to dryness and monitor glyco(sphingo)liptds by HPTLC

3.4. Pyridylamino (PA)-Oligosaccharide HPLC-Analyses 3 4.1. Pyridylamination of Zwitterionic Oligosaccharides (see Note 12) 1 Cleave 500 ug Nz-glyco(sphingo)liptd fraction with 1 mU endoglycoceramtdase (see Subheading 3.3.2., step 6, defined as mitral amount), perform a water lbutanol phase-separation and transfer the aqueous phase to a teflon-lined, screwcapped, conical microvtal and lyophrbze. 2 Add 20 pL coupling reagent, shake and react for 60 mm at 90°C Remove excess 2-ammopyndine under N2 stream at 60°C with two additions of 200 pL methanol. 3 Add 20 p.L reducing agent, vortex and react for 50 min at 80°C. Remove excess dtmethylamine borane, remaining traces of 2-ammopyrtdine and reaction by-products by TSK-HW-40 (F) gel filtratron column chromatography. 4. Prepare a 16 x 800 mm column equiltbrated m eluant 10 m&I ammonmm acetate buffer, pH 6 0, at room temperature 5 Apply reaction mixture directly to column 6 Elute, collect 6 mL fractions at a flow-rate of 12 mL/h 7 Profile the run by contmuous-flow monitormg with a fluorescence spectrophotometer detector (Fig. 3A), pool corresponding peak fractions and lyophilize.

3.4.2. PA-Nz-Oligosaccharide (see /Vote 73)

Separation by Anion-Exchange

HPLC

1 Take IOO-p.L altquots m disttlled water (1% of mittal amount) and inject onto the HPLC column. 2. Elute the column at room temperature with a lmear gradient from 100% eluant A to 60% eluant B in 60 mm; at a flow-rate of 1 mL/mm Detect PA-oligosaccharides by fluorescence spectrophotometry 3 Re-equilibrate the column with eluant A for 15 mm

3.4.3. PA-Nz-Oligosaccharide (see Note 73)

Separation by Reversed-Phase

HPLC

1 Take 100~pL analytical abquots m distilled water (1% of nutial amount) and inject onto the column 2 Elute the column at 40°C with a linear gradient from 100% eluant A to 50% eluant B in 50 mm, continue the run with 50% eluant B for 15 mm at a flow-rate of

Denms, Lochnit, and Geyer

0

60

120

180

10

240

300

360

420

20

480

540

30

600

660

40

720

780

840

900

50

60

50

60

n

r-l

ll f-f! 10

20

r-n

A n

30

Elutton time (mln) Fig 3

nn

r-l

\

40

Novel Amphoteric Glycosphingolipids

209

1 mL/min. Detect PA-oltgosaccharides by fluorescence spectrophotometry (Fig. 3B). 3. Re-equthbrate the column with eluant A for 20 mm

3.4.4 PA-Nz-Oligosaccharide

Separation by NH,-HPLC

1. Lyophihze an mmal amount and take up in the original volume with distilled water/acetomtrtle (25.75, v/v), take 100 pL analytical or 500 pL preparative aliquots and inject onto HPLC column. 2. Elute the column at room temperature with a linear gradient from 100% eluant A to 100% eluant B m 60 mm; at a flow-rate of 1 mL/mm Detect PA-oligosaccharides by fluorescence spectrophotometry (Fig. 3C). 3 Re-equilibrate the column with eluant A for 30 mm 4. Isolate the PA-oltgosacchartde peaks on elutton, lyophthze and store fractions at -20°C

4. Notes 1. Even when not stated, dried glyco(sphmgo)liptds are always subjected to a 30 s somcatton and 10 s vortexmg when solubthzed m either aqueous solutton or organic solvents 2 Since the Nz-glyco(sphmgo)hpids of A suum have been shown to be sapomfication-resistant (12), m any future work-up procedure a saponificatton step will be inserted followmg the acetone-wash and preceding dtalysis (7) 3. To obtain any strongly acidic A-glyco(sphmgo)liptd fraction, re-equilibrate the DEAE-Sephadex column (step 3 in Scheme 1) with 200 mL methanol and re-elute with 500 mL 0.8M methanohc ammonmm acetate, followed by rotary evaporatton and dralysrs. No strongly acidic A-glyco(sphmgo)hpids were apparently present m the A suum extract. 4. It was the anomalous behavior of a group of peracetylated glycosphmgoltplds of C vicina, that led to the eluctdation of phosphoethanolamme-containing Nz-glycosphingolipids (6,7). The N-acetylation of these nmhydrm-postttve glycosphingolipids yielded ninhydrm-negative products that migrated faster on HPTLC. Peracetylated, PE-containing Nz-glycosphingohptds could only be eluted from a Florisil column with methanol, and not the standard DCE/acetone. This suspected property of an anionically charged glycosphingohpid was confirmed by the retention of N-acetylated, PE-containing Nz-glycosphingohptds on DEAE-Sephadex Fig 3 (prevzous page). Isolatton and separation of zwtttertomc PA-glycans. Ohgosacchartde derivatives obtamed from Nz-glyco(sphmgo)ltptds by endoglycoceramtdase digestton and pyridylammation were separated from excess reagent and reaction by-products by gel filtration using a TSK-HW40 (F) column (A) PA-Nzohgosacchartdes (pooled as indicated by hortzontal bracket) were analyzed by reversed-phase HPLC (B) and preparatively fractionated by NH*-HPLC (C) Individual PA-Nz-glycans were pooled as indicated.

210

Denms, Lochnit, and Geyer

5. Small scale, orgamc solvent samples/fractions are always dried under a N, stream (with warmmg up to 37’C in order to accelerate the evaporation procedure). 6 A number of alternative methods are available for the separation of Nz- from N-glycosphingolipids silica gel column chromatography with methanol (7,18), sequential elution from QAE-Sephadex and DEAE-Sephadex columns (5); bandscraping from HPTLC-plates (7). 7. Glycosphingolipids can be semi-quantified on chromatograms, when sprayed with orcinol/HsSOd, by scanning with a Shimadzu (Tokyo, Japan) dual wavelength flying-spot scanner at 440 nm (reflectance mode). (The longer times of color development given correspond to the use of an oven without ventilator.) 8 If ninhydrm is used to detect free-NH* groups, pretreatment with I*-vapor is to be avoided. An alternative spray reagent to detect free-NH* groups nondestructively is fluorescamme (available commercially from Sigma). 9. Spray reagent-treated and mununo-stained HPTLC-plates are stored with a glass plate-cover at -20°C wrapped m thick alummmm foil 10. Mild HF-treatment allows the removal of substituents linked via phosphomonoor phosphodiester bonds without degrading the carbohydrate chain of the glyco(sphmgo)lipid, as demonstrated by matrix-assisted laser desorption/iomzanon time-of-flight mass spectrometry (MALDI-TOF-MS). The mass difference observed before and after HF-cleavage directly reflects the number of phosphocholme (PC) or phosphoethanolamine (PE) residues released (see Fig. 2, for the technical details of this method, see Hensel et al. /19/) 11. On the cleavage of glyco(sphingo)lipids by endoglycoceramidase, the lipid moiety is defined as a sphmgolipid. 12 Free amino groups must be blocked by either acetylation or methylation prior to pyridylamination, m order, to avoid complete loss of labeling due to sidereactions. N-Acetylation is routinely performed by incubation with 190 pL distilled water, 10 J.LLpyridme and 20 pL acetic anhydride for 30 mm at room temperature, before rotary evaporation to dryness. Selective methylation of ammo groups is performed by incubation with 200 pL 0.75M sodium carbonate and 20 pL methyl iodide for 2 h at 5O”C, followed by desalting (see Subheading 2., item 9). 13. Under the conditions used the PA-Nz-oligosaccharides were not resolved by either anion-exchange or reverse-phase HPLC (see Fig. 3B), most probably being because of the presence of the zwttterionic moiety. However, this technique is still relevant for the separation of PA-N- and PA-A-oligosaccharides into their indivtdual components

References 1 Noda, N., Tanaka, R., Miyahara, K., and Kawasaki, T. (1992) Two novel galactosylceramides from Marphysa sanguznea. Tetrahedron Lett. 33,7527-7530 2 Noda, N., Tanaka, R , Miyahara, K., and Kawasaki, T. (1993) Three glycosphmgolipids having the phosphocholme group from the crude drug “~u-yu” (the earthworm, Pheretima asiatica). Chem Pharm Bull. 41, 1733-1737

Novel Amphoteric Glycosphingolipids

211

3. Noda, N., Tanaka, R., Miyahara, K., and Kawasaki, T. (1993) Isolation and characterization of a novel type of glycosphmgohpid from Neanthes diverszcolor Bzochzm Bzophys Acta 1169,30-38 4 Sugtta, S., FuJii, H., Inagaki, F., Suzuki, M , Hayata, C., and Hart, T. (1992) Polar glycosphingolipids m Annehda. A novel series of glycosphingolipids containing choline phosphate from the earthworm, Pheretima hilgendorji J Biol Chem 267, 22,595-22,598.

5. Itonori, S., Kamemura, K., Narushima, K., Sonku, N , Itasaka, O., Horn, T , and Sugita, S. (199 1) Characterizatton of a new phosphonocerebroside, N-methyl-2ammoethylphosphonoglycosylceramide, from the antarctic krill, Euphausza superba. Bzochzm Bzophys Acta 1081,321-327. 6. Weske, B., Dennis, R.D., Helling, E, Keller, M., Nores, G.A , Peter-Katalinic, J , Egge, H , Dabrowskt, U , and Wtegandt, H (1990) Glycosphmgolipids in Insects. Chemical structures of two variants of a glucuromc acid-containing ceramide hexasaccharide from pupae of Callzphora vzczna(Insecta:Diptera), distmguished by a N-acetylglucosaminebound phosphoethanolamme stdechain Eur J Bzochem 191,379-388. 7. Helling, F., Dennis, R.D., Weske, B., Nores, G.A , Peter-Katalmic, J , Dabrowskt, U , Egge, H., and Wiegandt, H (1991) Glycosphmgohptds m insects The amphoteric moiety, N-acetylglucosamine-linked phosphoethanolamme, dtstmgmshed a group of ceramide ohgosaccharides from the pupae of Callzphora vzczna (Insecta. Diptera) Eur J Bzochem 200,409-421. 8 Itonori, S , Nishizawa, M , Suzuki, M., Inagaki, F., Hori, T., and Sugita, M (1991) Polar glycosphingolipids m insects. chemical structures of glycosphingoltpid series containing 2’-ammoethylphosphoryl-(+6)-N-acetylglucosamme as a polar group from larvae of the green-bottle fly, Luczlza Caesar. J. Bzochem. 110,479-485. 9 Itasaka, 0 and Hori, T (1979) Studies on glycosphingohpids of fresh-water bivalves V The structure of a novel ceramide oligosacchande contammg mannose-6-phosphate found m the bivalve, Corbzcula sandal J Biochem 85, 1469148 1 10. Arakt, S , Abe, S , Odam, S,. Ando, S., FUJI’, N., and Satake, M (1987) Structure of a trtphosphonopentaosylceramtde containing 4-0-methyl-N-acetylglucosamme from the skm ofthe sea hare, Aplysza kurodaz J Bzol Chem 262, 14,141-14,145. 11. Arakt, S., Abe, S., Ando, S., Fujii, N., and Satake, M. (1987) Isolation and characterization of a novel 2-ammoethylphosphonoglycosphmgohpid from the sea hare, Aplysra kurodar. J Biochem 101, 145-152. 12. Dennis, R. D , Baumeister, S., Smuda, C , Lochmt, G , Waider, T , and Geyer, E (1995) Inittatton of chemical studies on the unmunoreactive glycolipids of adult Ascarzs suum. Parasztology 110,61 l-623. 13 Lochnit, G., Dennis, R D , Muntefehr, H., and Geyer, R (1995) Structural analysis of glycolipids of the parasitic nematode Ascarzs suum. Glycoconj J, 12, 532,533. 14 Nores, G. A., Mizutamari, R K., and Kremer, D. M. (1994) Chromatographic tank designed to obtain highly reproducible high-performance thin-layer chromatograms of gangliostdes and neutral glycosphingohpids. J Chromatogr 686, 155-l 57. 15. Hase, S. (1994) High-performance liquid chromatography of pyrtdylammated saccharides. Methods Enzymol 230,225-237

212

Dennis, Lochnit, and Geyer

16 Ohara, K., Sano, M , Kondo, A., and Kato, I. (199 1) Two-dimensional mapping by hrgh-performance ltquid chromatography of pyridylammo ohgosaccharldes from various glycosphmgolipids J Chromatogr 586,35-4 1 17 Lochnit, G. and Geyer, R (1995) Carbohydrate structure analysis of batroxobm, a thrombm-like serme protease from Bothrops moojenz venom. Eul: J. Bzochem. 228, 805-8 16 18 Nores, G. A , Dennis, R D , Helling, F , and Wregandt, H (1991) Human heterophile antibodies recogmzmg epitopes present on insect glycoliprds. J Bzochem 110, l-8 19 Hensel, J , Hmtz, M , Karas, M , Lmder, D., Stahl, B , and Geyer, R. (1995) Localization of the palmrtoylatron site m the transmembrane protein p12E of Frrend murine leukaemia VKUS Eur J Brochem 232,373-380

14 Analysis of the Carbohydrate and Lipid Components of Glycosylphosphatidylinositol Structures Achim Treumann, M. Lucia S. Giither, Pascal Schneider, and Michael A. J. Ferguson 1. Introduction Glycosylphosphatidylinositols (GPIs) are a family of structures that contain the structural motif’ Mana 1-4GlcNH2a 1-Gmyo-Inositol-1 -Pod-lipid. This common substructure suggests that this family of molecules are btosynthetitally related and differentiates them from other glycosylated phosphomositides, such as the glycosylated phosphatidylinosrtols of mycobacteria and the glycosylated mositol phosphoceramides of yeasts and plants. The GPI family can be convemently divided mto two groups based on structural homology and function. The first group (1-28) are the membrane protein anchors (Fig. 1) that are found covalently linked to the C-termim of a wide variety of externally disposed plasma membrane proteins throughout the eukaryotes These GPI anchors afford a stable attachment of proteins to the membrane and can be viewed as an alternative mechanism of membrane attachment to a single-pass hydrophobic transmembrane peptide domain. For recent reviews of GPI anchor structure, brosynthesls, and function see refs. 29-31. The second group of GPI structures have only been found m protozoan organisms. These molecules exist as free glycophospholipids, such as the glycoinositol phospholipids (GIPLs) of the Leishmania, Trypanosoma cruzi, Leptomonas, Herpetomonas, and Phytomonas (29,32-34), or attached to phosphorylated repeating units as in the hpophosphoglycans (LPGs) of the Letshmama (29,35). In this chapter protocols specifically designed to analyse the protein-linked GPI anchors, although they are also applicable to the GIPLs and, to some extent, the LPGs will be described. From

Methods /n Molecular B/ology, Vol 76 Glycoanalysls Protoco/s Edlted by E F Hounsell 0 Humana Press Inc , Totowa, NJ

213

214

Treumann et al. PO.-CH2-CHz-NKCO-Protein I

Rl

Pm,sln T brucer

I72

R3

“SG

rbruce,

PARP

T cnlz,

lG7

Tcrwl

TcgS

T cn,z,

RI

RS

It6

~G~l*4 rldylated

[

Lipld

pdylactossmlna

I

sidechain

f

PSP

Toxcpfarma

gp23

Plasm&urn

Hex

AEP

1

Mana-P04.

]

I

AChE

Rat Thy-l Prp

Mouse

NCAM

Bovine

5’.NT

l-2

EtNPOa.

posltlanr

unknown,

fPGalNAc

EtNPa

WJNAc

EINPO,

[

fNANA-Hex-HexNAc

f@GaINAc

I nd

I *EtNPOI?

EtNPOI

1

(17.16)

nd

(19)

nd

fHexNAc7

EtNPa?

fEtNPO,?

nd

MDP

fGalBl-3GdNAc

fEtNP017

discylglycad

livman

UDP

fGd,N-3GdNAc

EtNPOq nd

Human

AChE

EtNPOa7

fEtNPO4

Human

PLAP

Human

CD52

f&Ml

Human

CD59

*aMan

EtNPOI7 +PGalNAc

(

EtNPOI

(1% (16)

nd

EtNPOq

(14) and

dlacyiglycad

nd

nd

nd alkyiacylglycerd 1

fEtNPOa’

palmitoyl

alkylacyl~ycemi dlacyiglycad

1

(13)

caamlde

Padns

1

* myrlrtoyl

cnamlt!0 caamlde diacylglycnd

PsA

(10.11)

(12)

diaeyiglycerd

1

(6.7) (8.9)

:!:%wd dlacylglycerd

aMan

Paramecwm

Hamster

palmitoyi

slkyiacylglycerd PGalNAc

(2.3) (4 5)

alkylscylg(ycerd

lelrhmanla

Torpedo

+

:Y%!Y~~d

[

Rsh (1)

lyso-rcylglycerd

antlge”

muclns’

0 dscddeum

Inmltd-myl

diacylglycerd

nd

fpalmltoyl

(20) (21) (22) (23) (23) (24) (25) (26) (27.26)

Fig 1. GPI anchor structures All GPI anchors attached to protem contam the conserved structure shown above with various substltuents (RI-RS) and lipids, as mdlcated. Some structures contain an additional fatty acyl chain attached to the 2-posltlon of the myoinosltol ring All metazoan organisms contam at least one, and sometimes two, extra ethanolamine phosphate (EtNPO,) substltuents in addltlon to the one used as a bridge to the protein C-terminal ammo acid. When a substltuent is known to be attached to a certain sugar residue but the linkage position 1s unknown, this 1s mdicated by a question mark. Square brackets are used to show substltuents for which the site of attachment has not been determined. The f symbol indicates that the associated residue 1s found on only a proportion of the structures AEP IS 2-ammo-ethylphosphonate.

The analyses described depend upon the presence of non-Wacetylated glucosamme (GlcN) in all GPI structures. The free amino group of the GlcN residue can be exploited by the nitrous acid deammatlon reaction that converts the GlcN to 2,5anhydromannose and simultaneously releases the phosphatidylmositol (PI) moiety (Fig. 2). The released PI moieties can be isolated by solvent extraction for molecular species analysis by negative ion electrospray ionizatron mass spectrometry (ESIMS), which gives the primary spectrum of

1

, IES-MS-CID-MS

(see

,I 9

Fig. 3)

“29HGH2

10

9

E;N

CH2-992 0

aB31i4

=O

E;N Pb 14 Menal-2Manal\

Manal-2Mana1, &al

-4AHM -

~&AC

,,4GsNAcJ~4~M EtN

(see

Fig. 5)

Fig. 2. General scheme for GPI structure determination. The PI fraction released by nitrous acid deammation (HONO) is recovered by solvent extraction and analysed by ES-MS and ESMS-CID-MS. The aqueous phase is reduced with sodium bototritiide (NaB31%) and the radiolabeled GPI neutral glycan fraction is released by aqueous I-IF dephosphorylation (aq. I-IF). The neutral glycan fraction is separated into individual glycan species by Dionex HPAEC followed by Bio-Gel P4 gel-filtration. The glycan structures suggested from the chromatographic properties on HPAEC and Bio-Gel P4 (see Table 1) are used to predict the glycan structure and a series of exoglycosidase digestion are used to confirm the structure. The exoglycosidase digestion products are conveniently analysed by HPTLC.

Treumann et al.

216 Table 1 Chromatographic

Properties

of GPI Neutral Dlonex HPAEC Du values

AHM Manal

-4AHM

Manal-GManal-4AHM Manul-ZManul-bhianal-4AHM Manul

-ZManal -6M

BbGel P4 Gu values

Glycans HPTLC Tu values

10

17

00

1 1

23

02

22

32

08

25

42

17

36

52

26

38

61

32

44

68

38

40

68

38

47

76

46

30

5.2

26

30

57

23

3s

65

29

30

67

28

nut -4AHM r

GalalMaim1 -2Manal-GM

ml-4AHM

Gala1 -6Galal-3 Manal-ZManal-GM

I

“I

Galal-6G

nul-4AHM

Ial-3 P

GalalManal-ZManul-GM

ml-4AHM 9

Gala1 -ZGalul-6Galal-3 Manal

-ZManal-6M

Galal-PGalal-6G

ml -4AHM lal-3

9

f Gala1 -2 Manal

-4AHM Manal-ZManul-GM

ml-4AHM 3

GalNAcpl-4 Manal-ZManal-2Manal-GManal-4AHM GalNAcpl -4 Manal

-2Manal-6M

Galpl-3GalNAcpl-4

ml -4AHM 9

the [M-H]- pseudomolecular ions of the PI species, and electrospray ionization-mass spectrometry-collision induced dissociation-mass spectrometry (ESIMS-CID-MS), which gives the daughter ion spectra of the [M-H]pseudomolecular ions. Examples of these spectra and then mterpretatton are given (Fig. 3). Followmg butan- l-01 extraction, the aqueous phase (containing the protein linked to the GPI-glycan) can be reduced with NaB3H4 to convert the aforementioned 2,5anhydromannose residue to [ l-3H]2,5-anhydro-

r “A

885

C ml2 865

m/z 241

m/z 283

0

Fig. 3 Examples of negative ion ESMS and ESMS-CID-MS spectra of vanous species derived from GPI anchors. (A) ESMS spectrum of the PI fraction released from porcine kidney membrane dlpeptidase (23). (B) ESMS-CID-MS daughter ion spectrum of the m/z 865 [M-H]- parent ion of the major diacyl-PI

species observed in panel A. (C) Fragmentation scheme for the daughter spectrum shown m panel B.

218

Treumann et al.

m/z 823

m/z 241

Fig. 3. (contznued) (D) ESMS spectrum of the PI fraction released from i? CYUZZ mucins (9). (E) ESMS-CID-MS

ion of an alkylacyl-PI

daughter Ion spectrum of the m/z 823 [M-H]-

parent

species observed m panel D. (F) Fragmentation scheme for the

daughter ion spectrum shown m panel E. The structure of the cychc alkyl-glycero-

phosphate ion (m/z377) 1shypothetical. (G) ESMS-CID-MS

daughter ion spectrum of

the m/z 892 [M-H]- parent Ion of a ceramide-PI species observed in panel D (H) Fragmentation scheme for the daughter spectrum shown m panel G

124

J

1

w

lode

4

283 303 f

%-

200

IK

m/z1023

400

' Dale

1000

600

8+ II m/z303

m/z283

mz255 (weak)

8 0

+

:A

t

Fig. 3. (continued) (I) ESMS spectrum of the PI fraction released from human CD524 antigen (26) (J) ESMS-CID-MS daughter ion spectrum of the m/z 1123 [M-HI-parent ion of the dlacyl (acyl)PI speciesobserved m panel I. (K) Fragmentation scheme for the daughter ion spectrum shownm panel J. (L,) ESMS spectrum ofthe PI fmctton released from T brucei PARP. (M) ESMS-CID-MS daughter ion spectrum of the m/z 837 [M-H]- parent ion of the bsoacyl(acyl)PI speciesobserved m panel I. (N) Fragmentation scheme for the daughter ion spectrum shownm panel M. The structureofthe cychcglycerophosphate ion (m/z 153) is hypothetical.

Treumann et al.

220

LlOO-

037

865 863

so0

610

620

830

640

650

660

070

690

660

900

Dale

I --

M 10

I

263

255

600

500

700

600

900

Dale

N m/z 153

m/z 283 ox

e 0 ee,

0 H

1

Fig 3. (continued) Panels A and B are reproduced from ref. 23, panels D, E, and G are reproduced from ref. 9 and panels I and J are reproduced from ref. 26 with the permission of The American Society for Blochemlstry and Molecular Biology

Glycosylphosphatidylrnos,rol DU

Structures 1

2

3

”

6 1

A

0---0

221 4

-3orlu

.J,10 *

20 Ttme tmtnsl

Gu?'J7654

3

’ B

2

I

--42Gu

0* 0640

08.20

Gu98765

10 00 TNne thrsl 4

3

2

1

2.5 -

o-l--.I'..n'. 06 40

08 20

10 00 T/me fhrsl

Fig 4. HPAEC and Bio-Gel P4 chromatography of GPI neutral glycans. The neutral glycan fraction isolated from porcine kidney membrane dipeptidase was chromatographed by Dlonex HPAEC (panel A) and the 2.5 Du and 3.0 Du peaks were rechromatographed on Bio-Gel P4 (panels B and C, respectively) The numbers above each chromatogram refer to the elution positions of the co-injected glucose ohgomer internal standards. The exoglycosidase sequencing of two of the resolved glycans, with the chromatographic properties 2.5 Du/4 2Gu and 3.0 Du/6 7 Gu, can be seen in Fig. 5

(AHM). The radiolabeled GPI glycan(s) may be isolated, followmg dephosphorylation with 50% aq. HF, by Dionex hrgh pH anion exchange chromatography (HPAEC) followed by gel-filtration on Bio-Gel P4 (Fig. 4). The chromatographic propertres of the glycans on these two systems are used to predict the glycan structure (Table 1). The glycans are then subjected to van-

mannitol

222

Treumann et al. Mllna1-2Mana\

hlanal.~aMl\ %an.l.4AnM

lam4

GalpI-3GaINAq31 hiMar.-\

/"

+eTEtG hanal-4AHM

-1 t&na1-2Mana1\

%4.mal.4AnM

he9

Gdl?.l.3GdNb&

f

If

+l

fan96

Manal-4AHM

haio

GaK31-3G~l

.

/'

t JmH

+BTBG

Ime GalNAc~1 'YZ lane7

bne"

Mmul.4AnM

lam 1.7

+JBAM IanBe

AIM

he13

mm'c /_

m

&

1 o-

0123

78

D 910ll1213

Fig. 5. Exoglycosidase and HPTLC analysis of purified GPI neutral glycans. The sequences of exoglycosidase digestions, and the structures of the products, are shown at the top of the figure and the relevant HPTLC lanes are indicated to the right of each structure. The enzymes used were Manctl-2Man specific Aspergirrus saitoi o-mannosidase (ASAM), jack bean a-mannosidase (JBAM), bovine testes S-galactosidase (BTBG), and jack bean g-hexosaminidase (JBBH). The lanes marked (D) contain NaB%I,+educed dextran oligosaccharitol standards. Note: The faint lower bands seen in lanes 10, 11, and 12 represent structures with only one aMan residue removed by the JBAM digestion due to steric hinderence. This figure is adapted from ref. 23 with the permission of The American Society for Biochemistry and Molecular Biology.

ous series of exoglycosidase digestions and the products are analyzed by highperformance thin layer chromatography (HPTLC) to confirm the predicted structure (Fig, 5). The data obtained can supply a complete primary structure for a GPI anchor, except for the location(s) of the ethanolamine phosphate substituent(s). The latter information, if required, can be deduced by a modification of the exoglycosidase sequencing protocol given here that is described in ref. 36. Examples of the approach described in this article can be found in refs. 5,9,23,26.

Glycosylphosphatidylmositol

Structures

223

2. Materials 1 HPLC-grade water 2. Screw-top Eppendorf tubes 3 HPLC-grade butan- l-01 shaken with an equal volume of HPLC-grade water m a clean bottle 4. 0.3M Sodmm acetate buffer, pH 4.0. Made by tttrating sodmm acetate solutton (0.3M final) to pH 4 0 with glactal acettc actd. Stable at room temperature for several months 5 1M Sodmm mtrite Prepare just before use. 6. 0.4M Boric acid. Stable at room temperature for several months. 7 2M Sodmm hydroxide Aristar-grade NaOH (BDH-Merck, Poole, UK). Store at room temperature m plastic for up to 1 wk 8 Sodium borotrtttide. Dissolve 100 mCt of NaB3H4 (NEN, Stevenage, UK, lO15 Ct/mmol) m freshly prepared 100 mMNaOH to a final concentration of 36 mM. Score the glass vial containing the NaB3H4 powder under vacuum wtth a glass knife and then wrap in parafilm before breaking open (to prevent loss of contents when air enters). Transfer the solution in 20-pL aliquots to Eppendorf tubes that are stored for up to 2 mo at -70°C 9. 1M Sodium borohydrrde NaBH4 is dissolved m water just before use 10 1M Acetic acid* stable at room temperature for several months. 11. Mrcrodralyser (e.g., Pierce 66320 Microdtalyser System 100, Pierce, Chester, UK) 12 50% Aqueous hydrogen fluoride (HF). Aristar-grade (BDH-Merck). Store m 1-mL aliquots m Eppendorf tubes at -2O’T Caution: htghly corrostve 13 Saturated lithium hydroxide. Prepare by shaking 20 g of LrOH wrth 40 mL of water Store at room temperature m a well-sealed plastic tube for up to 3 mo Caution: highly corrostve 14 Solid sodium hydrogen carbonate (NaHC03). 15 Acetic anhydrtde 16. Dowex AGSO-X 12,200-400 mesh (Bto-Rad, Hemel Hempstead, UK) converted to the H+ form by washing with >lO vol lMHC1 and >20 vol water. Store wrth an equal volume of water at 4°C. 17 Toluene: Aristar-grade (BDH-Merck). 18. Whatman 3MM chromatography paper (3-cm wide roll). 19. Butan-1-al/ethanol/water mixed m the ratio 4 1.0.6, by volume. The paper chromatography tank should be lined wtth filter paper with some of the solvent m the bottom. The solvent may be stored for up to 1 mo m a well sealed glass bottle. 20 2 mL Plastic syringes 2 1, 2M Acetic acid stable at room temperature for several months 22 Chelex 100 (Na+) (Bto-Rad) Store with an equal volume of water at 4°C 23. Dowex AG3-X4,200-400 mesh (Bio-Rad) converted to the OH- form by washing with >lO vol 1MNaOH and >20 vol water. Store with an equal volume of water at 4’C. 24. QAE-Sephadex-A25 (Pharmacra, Milton Keynes, UK). Swollen m water and washed with >lO vol water Store with an equal volume of water at 4°C

224

Treumann et al.

25 HPLC-grade chloroform and HPLC-grade methanol mixed m the ratro of 2.3, by volume, and contammg 1 nuI4 ammonia added from a concentrated (35%, w/v) ammonia solution 26. Soybean phosphatidyhnosrtol (Sigma) dissolved in chloroform/methanol (2.3, v/v), 1 nuI4 ammonia, at a concentration of 10 ng/pL 27. Access to a triple-quadrupole mass spectrometer with an electrospray source (e g , VG Quattro) 28 Dextran grade C (BDH-Merck) 29 0 1M Hydrochloric acid. 30 Access to a Dlonex HPAEC system 3 1 Access to a Bio-Gel P4 system 32 Alltech Micro-Spm centrifuge filters (Alltech 2494, Carnforth, UK). 33 Jack bean a-mannostdase (JBAM) (Boerhmger Mannhelm, Lewes, UK) Store enzyme suspension at 4°C. Just before use, centrifuge an ahquot of enzyme suspension, remove the ammonium sulfate supernatent and redissolve the pellet in 0. 1M sodmm acetate buffer, pH 5.0, to yield a solutron of 25 U/n& 34. Coffee bean a-galactostdase (CBAG) (Boerhmger Mannheim). Store enzyme suspension at 4°C. Just before use, centrifuge an aliquot of enzyme suspension, remove the ammonium sulphate supematent and redissolve the pellet m 0 IM sodmm acetate buffer, pH 6 0, to yield a solution of 25 U/mL 35. Jack bean P-hexosaminidase (JBBH) (Sigma, Poole, UK) Dissolve m 0 1M citrate-phosphate buffer, pH 4 2, at 4 U/mL Store frozen in 30-& aliquots at -20°C. Just before use add 3 pL of freshly prepared 100 nnI4 o-mannolc-y-lactone (Genzyme, West Malling, UK) Do not refreeze the enzyme 36 Aspergzllus saztoz a-mannosidase (Oxford GlycoSystems, Abingdon, UK) Prepare according to manufacturer’s mstructions. 37 Jack bean P-galactostdase (JBBG) (Sigma). Just before use, dialyze against 0 IM citrate-phosphate buffer pH 4 2 and adjust to 4 U/mL 38. Bovine testes P-galactosidase (BTBG) (Boehringer Mannheim) Just before use, dialyze against 0 1M curate-phosphate buffer pH 4.2 and adJUst to 0.5 U/mL 39 Aluminium-backed slhca-gel60 HPTLC plates (BDH-Merck, Art 5547) 40. HPLC-grade propan-l-01, HPLC-grade acetone and HPLC-grade water mixed m the ratio 9:6 5 and 5.4 1 (by volume). Mixed solvents can be stored for up to 1 wk m a well-sealed glass bottle 41 En3Hance aerosol spray cans (NEN). 42 Cling-film or Saran-wrap plasttc film 43. Kodak XAR-5 film and film cassette 44. DuPont Lightening-plus mtenslfymg screen (Sigma)

3. Methods 3.1. Preparation of the PI Fraction and the [3H]-Labeled Neutral Glycan Fraction 1. Remove salts from the glycoprotein solutton by dialysis agamst water (see Note 1) 2 Transfer the glycoprotein solution (containing >l nmol glycoprotein) to an Eppendorf tube and extract four times with 2 vol of butan-l-01 saturated wrth

Glycosylphosphatidylmositol

3 4 5 6

7.

8 9. 10 11 12. 13. 14.

15. 16. 17.

18.

Structures

225

water (see Note 2). For each extraction, vortex for 1 mm, centrifuge in a microfuge at maximum speed for 5 mm and remove the clear upper butan-l-01 phase Leave behind any interface material Freeze-dry the aqueous phase plus interface and redissolve the glycoprotem in 15 pL 0.3Msodium acetate buffer (pH 4.0) with the aid of a sonicatmg water bath. Add 7 5 pL of freshly prepared 1M sodium nitrite and mcubate for 1 h at room temperature. Add a further 15 pL of 0.3Msodmm acetate buffer (pH 4 0) and 7.5 pL of freshly prepared 1M sodium nitrite and incubate for a further 2 h at 37“C (see Note 3). Extract the reaction mixture once with 100 pL butan-l-01 saturated with water then twice with 50 pL butan- l-01 saturated with water (leave the interface behmd during the extractions) and pool the upper butan-l-01 phases m an Eppendorf tube (see Note 4) Store this PI fraction at -20°C prior to ESMS analysis. Transfer the aqueous phase to a good fume hood and add 10 pL 0 4M boric acid followed quickly by 20 pL of 2MNaOH (see Note 5) and 10 pL 36 mMNaB3H4 m 100 mA4 NaOH. Incubate for 2 h at room temperature then add 20 pL 1M NaBHa and incubate for a further 1 h. Destroy the excess reducmg agents by adding 20 pL ahquots of 1M acetic acid until effervescence ceases (see Note 6). Freeze-dry the products using a 250 mL round glass flask attached to a vacuum pump via a liquid nitrogen cold-trap (see Note 7). Redissolve the products in 50 pL water and transfer to a microdialyzer with a further 25 pL of washings (see Note 8) Dialyze against water for >3 h. Freeze-dry the dialyzed products m an Eppendorf tube Add 50 pL of cold 50% aq HF and mcubate at 0°C on ice/water for 48-60 h (see Note 9). Freeze 270 l.tL (see Note 10) of saturated LiOH m an Eppendorf tube and add the aq. HF digest to the frozen solution and vortex Centrifuge the mixture to remove the LrF precipitate and transfer the supernatent back to the ongmal Eppendorf tube that contained the aq. HF digest Wash the LiF pellet twice with 50 pL water and pool the supernatents. Add 40 mg of solid NaHC03 to the pooled supernatents and cool to 0°C on ice/water. Add three ahquots of 10 @., acetic anhydride at 10 min intervals (do not vortex) then allow the mixture to come to room temperature (see Note 11) Pass the reaction mixture through a column of 1 mL Dowex AG50-X12(H+) and elute with 4 mL water. Dry the eluate on a rotary evaporator and remove residual acetic acid by coevaporation with toluene (twice with 50 clr, toluene) Redissolve the radiolabeled glycans m 25 pL water and transfer to a 3 x 40 cm strip of Whatman 3MM chromatography paper. SubJect to downward paper chromatography for 48-60 h usmg butan- 1-al/ethanol/water (4: 1:0.6, v/v) as solvent (see Note 12).

19. Locate the labeled glycans using a linear analyzer (scanner) and cut out the relevant stnp of the chromatogram (see Note 13).

226

Treumann et al.

20 Roll up the paper strip and place in the barrel of a 2 mL plastic syrmge and hang the syringe barrel m a 15 mL centrifuge tube. Add 30 pL water per cm2 of paper and leave for 5 mm. Elute by centrifugatlon (5 min, 2,000g). Repeat the process of wetting and elution four more times. 2 1. Dry the eluate by rotary evaporation and redissolve m 200 pL 2Macetlc acid (see Note 14) 22 Transfer the solution to an Eppendorf tube and incubate at 100°C for 1 h 23 Dry m a Speedvac concentrator and remove the residual acetic acid by coevaporation with toluene (twice with 50 & toluene). 24. Redissolve in 200 pL water, pass through a column of 0 1 mL Chelex lOO(Na+) over 0.2 mL Dowex AGSO-X 12(H+) over 0.2 mL Dowex AG3-X4(OH-) over 0 1 mL QAE-Sephadex-A25(OH-) and elute with 1 mL water (see Note 15). 25 Dry the eluate and redissolve in 100 J.IL water Store the labeled neutral glycan fraction at -20°C prior to carbohydrate sequencing

3.2. ESMS Analysis of the PI Fraction 1. Wash the PI-contammg pooled butan-l-01 phases from step 3.1.6. by adding 200 pL water saturated with butan- l-01 and vortexmg for 1 mm. Centrifuge and transfer the upper washed butan- l-01 phase to a 2 mL glass vial and dry under a stream of N2 2 Redissolve the PI fraction m 100 pL chloroform/methanol (2:3, v/v) containmg 1 mM ammonia. 3 Pump chloroform/methanol (2.3, v/v) containing 1 mM ammoma into the electrospray source of a mass spectrometer at 5-l 0 +/mm and estabhsh a steady ion beam of solvent ions m negative-ion mode 4 Tune the ion source by optlmlzmg the response for the m/z 833 [M-H]pseudomolecular ion of soybean phosphatidylmosltol (see Note 16). 5. Inject IO-20-pL aliquots of the PI fraction and collect multiple scans over the mass range m/z 400-1400 (see Note 17) 6. Optimize the collision-induced-dissoclatlon conditions for daughter-ion scanning modes using the soybean PI standard (parent ion m/z 833) (see Note 18). 7 One at a time, select the PI [M-H]- pseudomolecular parent ions observed m Subheading 3.1., step 6 and collect the daughter ion spectra for these PI species (see Note 19). The daughter ron spectra (Fig. 3) can be readily interpreted to provide mformatlon on the class of PI species present (36) and, in some cases, the precise molecular species (see Note 20)

3.3. Fractionation

of the GPI Neutral Glycans

1 Dry the labeled GPI neutral glycan fraction from Subheading 3.1., step 25 and redissolve in 25 & water containing 75 pg of P-glucan ohgomer internal standards (see Note 21). 2 Separate glycans by high-pH anion exchange chromatography (HPAEC) on a Dionex CarboPac PA1 column using a Dlonex Blo-LC chromatograph (see Note 22) and record the elution positions of the labeled glycans in Dlonex units (Du) (see Note 23).

Glycosylphosphaticiylinositol

227

Structures

3. Pool the peak fractions and dry by rotary evaporatton Remove residual acetic acid by coevaporation twice with 50 pL toluene. 4. Redissolve peak materials in 100 pL water containmg 250 pg of P-glucan oligomer internal standards, filter through a 0.2 pm spin-filter and apply to a Bio-Gel P4 gel-filtration system (see Note 24) and record the hydrodynamic volume of the glycans m glucose units (Gu) (see Note 25). 5. Look up the Gu and Du values in Table 1 to see if the chromatographic properties of the glycan(s) correspond to known structures.

3.4. Sequencing

of the GPI Neutral Glycans

1. Using the GPI neutral glycan structure(s) suggested from the chromatographtc properties, devise an appropriate sequencing strategy (see Note 26). 2. Perform the appropriate series of exoglycosidase digestions (see Subheading 3.5.) and analyze the products by HPTLC and fluorography (see Subheading 3.6.).

3.5. Exoglycosidase

Digestions

1 Dissolve purified GPI neutral glycan samples (5,000-20,000 cpm) in enzymecontaining buffers and digest for 16 h at 37’C. 2. Inactivate the enzymes by heatmg to 100°C for 5 mm and desalt the products by passage through a column of 0 2 mL Dowex AGSO-Xl2(H+) over 0.2 mL Dowex AG3-X4(OH-) and elute with 2 mL water. 3. For JBAM use 20 pL 25 U/mL JBAM m 0 1M sodium acetate, pH 5 0 4. For CBAG use 20 pL 25 U/mL CBAG in 0 1M sodium acetate, pH 6.0 5. For JBBH use 30 pL 4 U/mL JBBH in 0 Wcitrate-phosphate, pH 4 2, containing 10 mM o-mannoic-y-lactone as a mannosidase inhibitor 6. For Aspergdlus salto Manal -2Man specific a-mannosidase (ASAM) use 10 pL 1 mU/mL ASAM in 0 1M sodium acetate, pH 5.0 7. For JBBG use 30 pL 4 U/mL JBBG m 0 1M citrate-phosphate, pH 4.2 8. For BTBG use 20 pL 0.5 U/mL BTBG m O.lMcitrate-phosphate, pH 4.5.

3.6. HPTLC Analysis of Exoglycosidase

Digests

1. Dry the desalted digests (Subheading 3.5., step 2) on a rotary evaporator, redissolve m 100 p.L water and transfer to Eppendorf tubes. 2. Dry in a Speedvac and redissolve in 4 pL 40% propan-l-01 m water. 3 Apply each sample, 1 pL at a time, to a 10 cm aluminium-backed Si-60 HPTLC plate as a band 0.5-cm wide and 1 cm from the bottom of the plate. 4. Apply standards of [3H]AHM (5000 cpm) and NaB3H4-reduced P-glucan ohgomers (25,000 cpm) to both ends of the plate (see Note 27). 5 Develop the HPTLC plate with propan-1-al/acetone/water (9:6.5, v/v), allow the plate to au-dry in a fume hood, then develop with propan- I-al/acetone/water (5:4: 1, v/v) and allow the plate to air-dry in a fume hood. 6. Spray plate in a fume hood with En3Hance spray, allow to dry in a fume hood and wrap with one layer of plastic cling-film 7. Place in a film cassette against Kodak XAR-5 film and a DuPont Lightening-plus mtensifymg screen and leave for 4-7 d at -70°C prior to developing

Treumann et al. 4. Notes 1 The nitrous acid deamination step is pH dependent and will be performed m a small volume of 0 3M sodium acetate buffer It is important to remove as much buffering capacity as possible before deamination. Ideally, the control for the analyses should be an equivalent sample of the glycoprotein that will receive sodium chloride m place of sodium nitrite m Subheading 3.1., steps 4 and 5 However, when the sample quantity is limttmg a sample of the original buffer that the glycoprotein was m should be processed in exactly the same way as the glycoprotem sample 2 The sample must be free of contaminating phospholiptds and detergents prtor to nitrous acid deammation so that the subsequent solvent extraction will contam only the released PI moieties of the GPI membrane anchor This exhaustive preextraction 1susually sufficient to remove phosphohpids and detergents It is worth saving the last of the four pre-extractions for ESMS analysis in case spurious ions are detected m the post-nitrous acid extract 3. The nitrous acid IS generated in sztu by the action of weak acid on sodium nitrite The glycoprotem solutton often turns brown during deammatton procedure 4 The PI moieties released from the glycoprotem by nitrous acid deammation of the GlcN restdue are isolated by solvent extractton with butan-l-01. 5 The exact volume of 2MNaOH to be used should be estimated emputcally Measure the volume (x mL) required to adJust the pH of 6 mL of the 0 3M sodmm acetate buffer (pH 4 0) plus 3.0 mL of the 1M sodium nitrite plus 2 0 mL of the 0 4M boric acid to between pH 10.0 and 11 5 (as measured on a pH meter) The amount of 2MNaOH to be used in Subheading 3.1., step 7 will be x x 5 pL. 6. Caution: Tritium gas is produced at this step. Use an appropriate fume hood. 7. Caution: Trmated water ts recovered in the cold-trap Dispose of this material carefully. 8. It is best to dedicate a microdialyzer for this purpose as radioactive contammatton of the plastic IS inevitable. As an alternative, gel-filtration on Sephadex GlO (using water as eluant) may be used for desalting instead of dialysis. 9. Use a positive-displacement pipet with a plasttc ttp to ahquot cold aq. HF 10 When setting up the aq. HF dephosphorylatton, ptpet 50 pL of aq. HF into several Eppendorf tubes Use these blanks to assesshow much saturated LtOH IS required to slightly under-neutralize 50 pL of aq. HF (1.e , the volume that adJusts the pH to about 3.5-4 0, as Judged by pH paper). 11. This N-acetylatton step IS designed to replace any N-acetyl-hexosamme N-acetyl groups lost during the aq HF step 12 The downward paper chromatography step removes many of the radiochemtcal contammants from the NaB3H4 reductton step 13 The labeled glycans will be found at or near the origin. If a scanner 1s not available, simply cut out the regton from -1 cm to +3 cm from the origin. 14 Subheading 3.1., steps 22 and 23 are designed to allow complete desialylation of the labeled glycans These steps can be omitted if the anchor is known not to contain sialic acid

Glycosy/~hosphati~y/~~ositol Structures 15. Further rad1ochem1cal contaminants from the NaB3H, reduction step will be removed by this mixed-bed Ion-exchange column 16 The main molecular species of soybean PI 1s sn-1-palmitoyl-2-llnoleoylphosphatldyllnos1tol (MW 834) Upon introducing a sample of soybean PI dlssolved in chloroformmethanol (2.3, v/v) containing 1 mMammon1a at 10 ng/pL, the capillary voltage, high-voltage lens and cone and skimmer voltages of the electrospray source should be adjusted in an iterative process to obtain a maximum ion current for the [M-H]- 1on of this PI species at m/z 833. The values will vary according to the instrument type and the state of the ion source Typical values for a VG Quattro I Instrument are* Capillary voltage, 2 5-3 0 kV; high-voltage lens, 0 4-O 6 kV, cone voltage, 5&65 V, skimmer voltage, 5-10 V higher than the cone voltage. 17 The mass range m/z 400-1400 (scan time 10 s) 1s sufficient to observe all conceivable lyso-PI species, PI species and (acyl)PI species. The first analysis may be performed at low-resolution (high-sensitivity) and subsequent analyses may be performed over a narrower mass range of interest at higher resolution. If the glycoprotein IS known to be sensitive to the action of bacterial PI-PLC (1 e , the PI moiety 1snot acylated) then the ions observed 1n the primary spectrum may be readily confirmed as PI species by performIng a parent ran scanning 1n MS 1 for the daughter 1on [inosltol- 1,2-cyclic phosphate]- at m/z 24 1 detected 1n MS2. The parent 1on spectrum shows only the series of ions that give rise to this diagnostic PI fragment ion This kind of primary spectrum tilter1ng can be very convenient if the primary spectrum contains non-PI contaminants 18 Daughter ions are generated by focusing one parent 1on at a time into a colllslon cell filled with argon at a pressure of 2.0-3.0 x lo9 mbar (2.0-3.0 Pa) The ions are accelerated into the collision cell (at low mass resolution 1n MS 1) through an adjustable potential difference (collision energy voltage) This voltage should be optimized while focussrng the m/z 833 parent 1on of the soybean PI standard into the collision cell 1n order to produce the max1mum 1on current for the m/z 241 daughter 1on corresponding to [1nos1tol- 1,2-cychc phosphate]-. Typical values for the VG Quattro I Instrument are 50-65 V. Other lenses will also need to be adjusted to optimize ion transmission between the quadrupoles. 19. If the parent ions are intense, daughter 1on spectra may be collected from m/z 50 to Just above the m/z value of the parent 1on However, 1f the parent ions are less Intense the daughter 1on spectra can be recorded over the mass range m/z 2OCk 520 (scan time 5 s) where most of the informative fragment ions 11e 20 The rules of daughter 1on spectra interpretation are as follows: a. The presence of an Intense [lnositol- 1,2-cyclic phosphate]- daughter 1on at m/z 24 1 m-mediately identities the parent 1on as a PI species and further 1nd1cates that the 1nos1tol ring 1s not acylated at the 2-position (26) b The presence of m/z 241 plus one fatty acid carboxylate 1on (see Fig. 3B,C) IS consistent with either a diacyl-PI species, where both fatty acids are the same (23), or a @so-acyl-PI species. The mass of the parent ion should readily dlst1ngu1sh these two posslb111tles

Treumann et al. c. The presence of m/z 241 plus two fatty acid carboxylate ions is consistent with a diacyl-PI specieswhere the fatty acids are different. The relative positions of the fatty acids can not be assignedfrom thesedata. d. The presenceof m/z 24 1 plus one fatty acid carboxylate ion plus one cyclic alkyl-glycerophosphate fragment ion (seeFig. 3E,F) indicates an alkylacylPI species(9). These are generally I-alkyl-2-acyl-PI species. e. An even integer of the molecular massof the [M-l]- parent ion, the presence of m/z 241 plus m/z 259, the [inositol-monophosphatel- ion, and an absence of fatty acid carboxylate ions andcyclic alkylglycerophosphate fragment ions, indicates a ceramide-PI structure (9), see Fig. 3G,H. The absenceof fatty acid carboxylate fragment ions and/or long chain base-containing fragment ions in these daughter ion spectra precludes complete molecular species assignmentfrom these data alone. f. The virtual absenceof m/z 241, andthe presenceof two strongandoneweak fatty acid carboxylate ions (seeFig. 3J,K) indicatesa diacyl-(acyl)PI specieswhere the weakestcarboxylate ion (usually palmitate) representsthe fatty acid originally attachedto the 2-positionof the inositol ring (26) and the strongercarboxylate ions representthe fatty acidsoriginally attachedto the glycerol backbone. g. The virtual absenceof m/z 241, and the presenceof one strong and one weak fatty acid carboxylate ion plus one cyclic alkyl-glycerophosphate fragment ion, indicates an alkylacyl-(acyl)PI species,where the weakest carboxylate ion (usually palmitate) represents the fatty acid originally attached to the 2-position of the inositol ring and the stronger carboxylate ion represents the fatty acid originally attached to the glycerol backbone. The position of the fatty acid on the glycerol backbone is generally the sn-2-position. h. The virtual absenceof m/z 24 1, the presenceof a strong ion at m/z 1.53 (cyclic glycerophosphate) and the presenceof one strong and one weak fatty acid carboxylate ion (see Fig. 3M,N), indicates @so-acyl(acyl)PI specieswhere the weakest carboxylate ion representsthe fatty acid originally attached to the 2-position of the inositol ring and the stronger carboxylate ion representsthe fatty acid originally attached to the glycerol backbone. The position of the fatty acid on the glycerol backbone is generally the sn- 1-position. 21. The P-glucan oligomers (GIc,-Glc& are prepared by partial acid hydrolysis of 100 mg of dextran (BDH-Merck, grade C) in 2 mL O.lM HCl, 4 h, 100°C. The acid is removed by passageof the hydrolysate through a column of 1.0 mL of Dowex AG3-X4(OH-) eluted with 3 mL water. The resulting set of P-glucan oligomers(at about 25 mg/mL) are stored at -20°C. 22. Dionex HPAEC should precedeBio-Gel P4 chromatography becauseit removes residual radiochemical contaminants that will otherwise contaminate the BioGel P4 system. The following programme is routinely used for the resolution of GPI glycans: Flow rate = 0.6 mlimin, buffer A = O.l5MNaOH, buffer B = 0.1 SM NaOH, 0.25M sodium acetate, starting conditions 95% A, 5% B followed by a linear gradient to 70% A, 30% B over 75 min at 0.6 mL/min. A wash cycle of 100% B for 10 min is followed by re-equilibration in 95% A, 5% B for at least

Glycosylphosphatidylinositol

Structures

237

15 min. The /3-glucan standards are detected by pulsed-amperometric detection (Dionex) and the pH of the eluate is lowered by passage through a Dionex ARRS anion suppressor prior to detection of the labeled GPI glycans by a Raytest Ramona on-line radioactivity detector equipped with a 0.2 mL solid scintillator X-cell. All data are collected and analyzed using the Raytest Ramona data system. Fractions (1 mL) are collected for pooling radioactive peaks (23). The absolute retention times of glycans can vary substantially on this HPLC system from day to day. However, the elution positionsof the GPI glycans relative to the set of /3-glucan internal standardsis almost constant. The elution position is expressed in so-called “Dionex units” (Du) by linear interpolation of the radioactive peak between adjacent P-glucan peaks.The Du value of a glycan hasno specific meaning other than asa fixed chromatographic property. An example of HPAEC chromatography of a labeled GPI neutral glycan fraction is shown in (Fig. 4A). 24. A commercial Bio-Gel P4 system (Oxford GlycoSciences GlycoMap) may be used in high-resolution mode or a column, 1 m x 1.5 cm, of Bio-Gel P4 (minus 400 mesh)should be packed after removing fines. The column isjacketed at 55°C and eluted with water using an Pharmacia P500 FPLC pump at 0.2 mL/min. Samplesare applied via a Rheodyne low-pressureinjector and the column eluate is monitored for the fl-glucan internal standardsusing an on-line refractive index monitor (e.g., Erma 75 12) and for the radiolabeled GPI neutral glycans by an online radioactivity monitor (Raytest Ramonaequipped with a 0.2 mL solid scintillator X-cell). All data are collected and analyzed using the Raytest Ramona data system. Fractions (1 mL) are collected for pooling radioactive peaks. 25. The GPI neutral glycan hydrodynamic volumes are expressedin glucose units (Gu) by linear interpolation of the elution position of the radioactive peak between adjacent P-glucan internal standards.Examples of Bio-Gel P4 fractionation of GPI neutral glycans can be seenin Fig. 4B,C. 26. The generalrules for selectingappropriateexoglycosidasedigestionsare asfollows: a. If the Gu and Du values suggesta simple linear structure such as Mancll2Manal-6Mana 1-4AHM (4.2 Gu, 2.5 Du) then one should digest aliquots of the purified glycan with ASAM and with JBAM (see Subheading 3.5.). The results (Fig. 5, lanes l-3) will show the removal of one and three aMan residues, respectively. These digests alone will formally define the glycan as Mancll-2Manal -?Mana 1-?AHM. However, combined with the chromatographic properties of the parent glycan this sequencemay be deduced, with reasonablecertainty, to be Mana l -2Mana 1-6Mancxl -4AHM. b. If the Gu and Du values suggest a branched structure such as Manal2Manal-6(Gal~l-3GalNAc~l-4)Manal-4AHM (6.7 Gu, 3.0 Du) then one should select two setsof digestions.The first set (in order) is BTBG, JBBH, ASAM and JBAM. An aliquot is savedafter each digestion for HPTLC analysis(Fig. 5, lanes4-8). This setof digestionswill indicate the sequenceGall31?HexNAcP 1-?(Mana I-2Mana 1-?Mana 1-?AHM). The second set of digestions (JBAM, BTBG, JBBH, JBAM) is designed to locate the site of the Gal-HexNAc- side-chain. In this case, the first JBAM digestion will

232

Treumann et al.

remove the aMan restdues up to the attachment sue (Fig. 5, lanes 9 and 10) The subsequent BTBG and JBBH digesttons will reveal the Manal4AHM and the second JBAM digestton ~111produce free AHM (Fig. 5, lanes 1 l-l 3) The combined digestion results alone define the glycan as Mancll-2Mancx lT(GalP 1-VGalNAcB 1-‘?)Mana 1-?AHM. However, combined wrth the chromatographtc properttes of the parent glycan thts sequence may be deduced, with reasonable certainty, to be Mana l -2Mana 1-6(GalB 1-3GalNAcP l4)Manal-4AHM Note, Sensitivity to BTBG and resistance to JBBG may be used to confirm a GalP 1-3HexNAc lmkage. 27 Standards of [3H]AHM and NaB3H4-reduced B-glucan ohgomers may be prepared as follows. a [3H]AHM standard Dtssolve 25 pg of o-glucosamme hydrochlorrde m 50 p,L 0 1M sodmm acetate buffer pH 4 0, deammate by mcubatton with 50 & freshly prepared 0.5M sodmm nitrite for 2 h at room temperature, add 50 pL 0 4M boric acid and 25 $2MNaOH, followed by 10 p.L 36 m&I NaB3H4 (l&15 Wmmol, NEN) dissolved in O.lM NaOH (2 h at room temperature). Complete the reductron by adding 20 pL freshly prepared 1M NaBH4 (2 h room temperature) Destroy excess reductant by adding 20-Ccr,ahquots of 1Macetlc actd unttl effervescence stops Caution: trmum gas produced. Pass the products through 0 3 mL AGSO-X12(H+), elute with 1.5 mL water and rotary evaporate the eluate Caution: trrttated water removed here, use a dedicated cold-trap Add 0 25 mL 5% acetrc actd m methanol and dry agam, repeat, add 0.25 mL methanol and dry. Dissolve the products in water and SubJect them to paper chromatography as described m (3.1.18.) but only for 16 h The labeled [3H]AHM (which will migrate about 6 to 8 cm from the ortgm) IS recovered from the paper by elutmg the regron from +4 cm to +l 1 cm three times with 500 pL water. Concentrate by rotary evaporation and pass through a mixed bed ton-exchange column of 0.1 mL Chelex lOO(Na+) over 0.2 mL AGSO-X12(H+) over 0 2 mL AG3-X4(OH-) over 0 1 mL QAE-Sephadex A25(OH-) with 2 mL of water Adjust the solution to 5000 cpm/pL m 40% propan-l-01 and store at 20°C b. NaB3H4 -reduced P-glucan oligomer standards. Take 20 pL of the 25 mg/mL stock solution of P-glucan ohgomer standards (descrtbed m Note 21) and reduce by adding 5 & 36 mMNaB3H4 in 0. 1MNaOH (1 h, room temperature). Add 20 pL 1M NaBH, to complete the reduction (1 h, room temperature) Destruction of excess reductant and radtochemrcal purlficatton are as described above (see Note 27a) for [3H]AHM, except that the-l to +4 cm region of the paper chromatogram should be eluted three times wtth 300 pL water

Acknowledgments This work was supported by The Wellcome Trust and by The Howard Hughes Medical Institute. Achim Treumann thanks the EU for a Human Capital and Mobility fellowship and Pascal Schneiderwas an EMBO long-term fellow. We thank Ian Brewis, Anthony Turner, and Ngel Hooper (University of

Leeds)for permission to use some figures from our Joint publication,

Glycosylphosphatidylinositol

Structures

233

References 1 Ferguson, M A J , Homans, S W, Dwek, R A., and Rademacher, T W (1988) Glycosylphosphatidylmosrtol moiety that anchors Trypano~oma brucel Variant surface glycoprotein to the membrane Sczence 239,753-759 2. Field, M. C., Menon, A. K., and Cross, G. C (199 l)A glycosylphosphatrdylmosttol protein anchor from procycltc stage Trypano~oma brucez hptd structure and brosynthesis EMBO J 10,273 l-2739. 3 Ferguson, M. A J , Murray, P,, Rutherford, H., and McConvrlle, M (1993) A sample purtficatton of procychc acidic repetmve protein and demonstration of a sralylated glycosylphosphattdylinosrtol membrane anchor Bzochem J 291,5 l-55, 4 Guther, M L S., Almerda, M L C D, Yoshrda, N, and Ferguson, M A J (1992) Structural studreson the glycosylphosphatrdyhnositol membrane anchor of TYypanosoma cmczzlG7-Antigen The structure of the glycan core J Biol Chem 267,6820-6828 5 Herse, N., Almeida, M L C. D , and Ferguson, M. A. J (1995) Charactensatron of the lipid moiety of the glycosylphosphatrdylinosrtol anchor of Trypanosoma cruzz 1G7-anttgen. Mol. Blochem Parasltol 70,7 1-84 6. Couto, A S., Lederkremer, R. M D , Colh, W., and Alves, M. J. M (1993) The glycosylphosphatrdylmosrtol anchor of the trypomastigote-specific Tc-85 glycoprotein from Trypanosoma cruzi Metabolic labeling and structural studies Eur J. Blochem 217,597-602 7 Abum, G., Couto, A. S., Lederkremer, R. M., Casal, 0 L., Galh, C., Colll, W , and Alves, M J. (1996) Trypanosoma cruzz. the Tc85 surface glycoprotem shed by trypomasttgotes bears a modrtied glycosylphosphatrdylmosrtol anchor Exp Parasltol. 82,290-297

8. Previato, J. 0 , Jones, C , Xavier, M T., Watt, R., Travassos, L R , Parodt, A J , and Mendoca-Prevtato, L (1995) Structural characterrsatron of the maJor glycosylphosphattdylmosrtol membrane anchored glycoprotem from eptmastlgote forms of Trypanosoma CYUZIstrains J Blol Chem 270,7241-7250 9 Serrano, A A., Schenkman, S , Yoshtda, N , Mehlert, A, Richardson, J, M , and Ferguson, M A. J. (1995) The lipid structure of the glycosylphosphattdyhnosrtolanchored mucm-like stahc acid acceptors of Trypanosoma cruzz changes during parasite dtfferentiatton from eprmasttgotes to infective metacyclic trypomastrgote forms. J. Blol Chem 270,27,244-27,253. 10. Schneider, P., Ferguson, M. A. J., McConvtlle, M. J., Mehlert, A , Homans, S W, and Bordier, C. (1990) Structure of the glycosylphosphattdyhnosttol membrane anchor of the Lershmanza maJor promastigote surface protease J Blol Chem 265, 16,955-l 6,964 11 McConville, M. J., Colhdge, T A. C., Ferguson, M. A J., and Schneider, P (1993) The glycomosrtol phosphohprds of Lezshmania mexzcana promastrgotes. Evrdence for the presence of three drstmct pathways of glycohptd biosynthesis J Blol Chem 268, 15,595-l 5,604. 12 Tomavo, S., Dubremetz, J -F., and Schwarz, R. T. (1993) Structural analysis of glycosylphosphatidylinositol membrane anchor of the Toxoplasma gondzz tachyzotte surface glycoprotem gp23. Bzol. Cell 78, 155-162

234

Treumann et al.

13 Azzouz, N , Strrepen, B , Gerold, P , Capdevrlle, Y , and Schwarz, R T (1995) Glycosylmosrtol-phophoceramrde m the free-living protozoan Paramecium przmaurelra modtficatton of core glycans by mannosyl phosphate EMBO J 14, 442224433 14 Gerold, P., Schofield, L , Blackman, M J., Holder, A A , and Schwarz, R T. (1996) Structural analysis of the glycosylphosphatrdylmosltol membrane anchor of the merozorte surface proteins-1 and -2 of Plasmodlum falclparum Mol Blochem Parasltol , m press 15. Fankhauser, C , Homans, S W., Thomas-Oates, J E , McConvrlle, M J , Desponds, C , Conzelmann, A , and Ferguson, M A J (1993) Structures of glycosylphosphatrdylmosttol membrane anchors from Saccharomyces cerevzslae J Bzol Chem 268,26,365-26,374 16 Haynes, P. A., Gooley, A. A , Ferguson, M. A J., Redmond, J W., and Willrams, K L (1993) Post-translational modrflcatrons of the Dzctyosteltum discozdeum glycoprotein PsA Glycosylphosphatrdylmositol membrane anchor and cornpositron of O-lurked ollgosacchandes Eur J Blochem 216,729-737 17 Butrkofer, P , Kuypers, F A , Shackleton, C , Brodbeck, U , and Streger, S (1990) Molecular species analysis of the glycosylphosphatrdylmosrtol anchor of Torpedo marmorata acetylcholmesterase J Blol Chem 265, 18,983-18,987 18. Mehlert, A , Varon, L , Srlman, I , Homans, S W , and Ferguson, M. A J (1993) Structure of the glycosylphosphatidylmosrtol membrane anchor of acetylcholmesterase from the electric organ of the electrx fish, Torpedo callfornlca Blochem J 296,473-479 19. Homans, S W , Ferguson, M A J., Dwek, R. A , Rademacher, T W., Anand, R , and Williams, A F. (1988) Complete structure of the glycosylphosphatrdylmosrtol membrane anchor of rat brain Thy-l glycoprotem Nature 333,269-272 20 Stahl, N , Baldwm, M A , Hecker, R., Pan, K -M , Burlmgame, A L , and Prusmer, S. B (1992) Glycosylmosrtol phosphohprd anchors of the scrapre and cellular prron proteins contam sralrc acrd. Blochemlstry 31,5043-5053 2 1 Mukasa, R , Umeda, M , Endo, T , Kobata, A , and Inoue, K (1995) Characterisation of glycosylphosphatrdylmosrtol (GPl)-anchored NCAM on mouse skeletal muscle cell lme C2C12 the structure of the GPl glycan and release during myogenesis Arch Blochem Blophys 318, 182-190. 22. Taguchr, R., Hamakawa, N., Harada-Nrshtda, M., Fukur, T ,NoJima, K , and Ikezawa, H. (1994) Mtcroheterogenerty m glycosylphosphatrdylmosrtol anchor structures of bovine liver 5’ nucleotrdase. Blochemrstry 33, 1017-1022. 23 Brewis, I A , Ferguson, M A J , Mehlert, A, Turner, A. J., and Hooper, N M (1995) Structures of the glycosylphosphattdylmosrtol anchors of porcine and human renal membrane dipeptidase Comprehensive structural studies on the porcine anchor and mterspecles comparison of the glycan core structures. J Blol Chem 270, 22,946-22,956. 24 Deeg, M. A., Humphrey, D. R., Yang, S. H., Ferguson, T R., Reinhold, V N , and Rosenberry, T L (1992) Glycan components m the glycomosrtol phospholrprd anchor of human erythrocyte acetylcholmesterase. J Blol. Chem 267, 18,573- 18,580

Glycosylphosphatidylinositol

Structures

235

25 Redman, C. A , Thomas-Oates, J. E., Ogata, S , Ikehara, Y., and Ferguson, M A J (1994) Structure of the glycosylphosphatidylmosrtol membrane anchor of human placental alkaline phosphatase. Bzochem. J 302,86 l-865 26 Treumann, A , Lifely, M R , Schneider, P., and Ferguson, M A J (1995) Primary structure of CD52 J Btol Chem 270,608&6099 27. Nakano,Y., Noda, K , Endo, T., Kobata, A , and Tomita, M (1994) Structural study on the glycosylphosphatidylmosrtol anchor and the asparagme-lmked sugar cham of a soluble form of CD59 m human urme. Arch Bzochem Bzophys 311,117-126 28 Sugita, Y., Nakano, Y, Oda, E., Noda, K., Tobe, T, Mmra, N -H., and Tomrta, M (1993) Determmation of carboxyl-termmal residue and disulfide bonds of MACIF (CD59), a glycosylphosphatldylmosnol-anchored membrane protem J Bzochem 114,473-477.

29 McConville, M J and Ferguson, M A J (1993) The structure, biosynthesis and function of glycosylated phosphatldylinosltols in the parasitic protozoa and higher eucaryotes Blochem J 294,305-324 30. Brown, D. A (1992) Interactions between GPI-anchored protems and membrane lipids. TrendsCell Blol 2,338-343 3 1 Stevens, V L (1995) Biosynthesis of glycosylphosphatidylmosltol membrane anchors Bzochem J 310,361-370 32 Redman, C A, Schneider, P, Mehiert, A., and Ferguson, M A. J (1995) The glycomositol-phosphohpids of Phytomonas Btochem J 311,495-503. 33 Prevrato, J O., Mendonca-Previato, L , Jones, C , and Fournet, B (1992) Structural characterisation of a novel class of glycophosphosphmgohpids from the protozoan Leptomonas-samuellJ Blol Chem 267,24,279-24,286

34 Routrer, F H , da Silvena, E X , Wait, R , Jones, C , Previato, J 0 , and MendoncaPreviato, L (1995) Chemical characterisatron of glycosylmositolphosphohpids of HerpetomonassamuelpessoalMel Blochem Parasltol 69,6 1-69

35 McConville,

M J , Schnur, L F , Jaffe, C , and Schneider, P (1995) Structure of m Old

Lezshmanzahposphosphoglycan: inter- and intra-specific polymorphism World species Bzochem J 310, 807-818.

36 Schneider, P and Ferguson, M A J (1995) Microscale analysis of glycosylphosphatrdylmosrtol structures Me6h Enzymol. 250,614-630

15 Conformational Analysis of Biantennary Glycopeptides with a Resonance Energy Transfer Technique Kyung Bok Lee, Pengguang and Yuan Chuan Lee

Wu, Ludwig Brand,

1. Introduction The carbohydrate components m glycoprotein receptors, antibodies, enzymes, toxins, and hormones are mcreasmgly gaining attention as biologtcal recognition signals (1,Z). It is, therefore, important to study the conformation of ohgosaccharides m solution to better understand the mechamsm of carbohydrat+protem interactions. X-ray diffraction (3), electron spin resonance (4), computer molecular modeling (5), and nuclear magnetic resonance (NMR) have been used to determine ohgosaccharide conformation. Proton NMR has played a major role in determmation of the solution conformation of oligosaccharides or glycopeptides by measuring the nuclear Overhauser effect (NOE) that is suitable for the distance of 2-5A (t&8] (see Chapter 1). However, in measuring the long range (lO-50A) distances m biological systems, resonance energy transfer measurements are more suitable and more sensitive tools (only nanomole-range samples are required) (9,10). In order to apply resonance energy transfer techniques to conformational studies of complex type branched oltgosaccharides, fluorescence donor and accepter molecules must be introduced at specific sites on different branches. Usmg such an approach, Rice et al. (11) were able to characterize conformations of triantennary glycopeptides m solutton by using time-resolved resonance energy transfer. This technique was extended to study the solution conformation of biantennary glycopeptides Examples of modtfication of biantennary glycopeptides with donor and accepter probes and measurement of energy transfer are described in thts chapter. From

Methods m Molecular Biology, Vol 76. Glycoanalys~s Protocols Edlted by E F Hounsell 0 Humana Press Inc , Totowa, NJ

237

Lee et al.

238

6’5’ 4’

~O~WC%hN~-Gal~(1-4)GlcNAc~(1-2)Manu(l-6)--..

A

SAr1(2-6)~l~(l-4)G~NAc~(l-2)Mana(l-3)

anp(l-4)-R

4 RI-flunrcrcent-IRheled

6’

5’

glyeopeptide

1

4’

Galp(i-4)GlcNAc$(l-2)Mana(l-6). Manp(l-4)--R ~~~~~~~~h~~-SAa(2-6)Gal~(1-4)GlcNAc~(1-2)Mana(l-3) 6 5 Iii-fluorescent-labeled

/3 4 glycopeptlde

II

II = Gl~NA~~(I-4)Gl~NAr~(I-)~ coon

Fig 1 The structure of btfluorescent-labeled

1.1. Strategy for Introduction of Fluorophores Terminal Sites of Biantennary Glycopeptides

glycopeptldes

at Specific

In order to specrfically Introduce a fluorophore at a defined posmon of each branch in the brantennary glycopepttdes, monostalylated glycopepttdes from bovme fibrmogen IS a convenient startmg material because of its ready avallability (12). Micromole range separation of monosiaylated glycopepttdes can be accompltshed by utiltzmg diethylaminoethyl (DEAE)-cellulose amonexchange chromatography. Because monosialylated glycopepttdes from bovine tibrmogen are stalylated on the Manal-3Man branch only, we are able to exploit this specific feature (Fig. 1). The terminal Ga16’ on the Manal-6Man branch IS oxtdized by galactose oxtdase and the termmal stalrc acid on Manal 3Man branch IS oxtdtzed with pertodate under controlled condtttons (13) This strategy allows us to mtroduce fluorophores at specrfic branches of btantennary glycopepttdes. N-Naphthyl-2-acetylatton of the monostalylated glycopeptrdes enhanced its hydrophobtctty to be retained on a Cs reverse phase-high performance liquid chromatography (RP-HPLC) column and allows separation of the glycopeptides with NeuAc from glycopepttdes wtth NeuGc The alternative approach is to use partial oxtdatlon (see Note 1) of the C-6 posmon of one of the terminal galactose residues on desralylated glycopeptides with galactose oxidase followed by reductive ammation wtth 2-(dansylamtdo)ethylamine and extensive HPLC separation of modified glycopepttdes.

Conformational

Analysrs by Resonance Energy

239

In this case, the characterization of positional isomers of fluorescent glycopeptides by NMR is very difficult because of the overlapping signals. The asstgn-

ment of the location of the dansyl group was possible only after removmg the underivatized branch by sequential exoglycosidase digestion 2. Materials 1 Bovine fibrinogen (Mtles Inc , Kankakee, IL) 2. Pronase (CalBiochem, La Jolla, CA) 3 Neuramnudase from Arthrobacter ureafaczens (Boehrmger Mannheim, Indianapohs, IN). 4 Glycopeptidase A (Seikagaku America Inc., Rockvrlle, MD) 5 Sodium cyanoborohydrtde (Caution: corrosive, hygroscopic). 6 Borane-pyridine complex (Aldrich, Milwakee, WI) 7. 2-ammopyridme (Caution: irritant) 8 1-hydroxybenzotnazole (Aldrich). 9. Dicyclohexylcarbodnmtde (Caution: highly toxic, corrostve) 10 2-Naphthylacetic acid (Nap) 11 Z-(Dansylamrdo)ethylamme (Molecular Probes, Eugene, OR) 12. Sodium periodate. 13. Sodium metabisulflte. 14. Galactose oxidase (Sigma, St Louis, MO) I5 Catalase (Sigma) 16 P-o-Galactosidase from bovine testes (Boehrmger Mannhelm) 17 P-N-Acetyl-n-glucosammidase from beef kidney (Boehrmger Mannhelm) 18 a-o-Mannostdase (V-Labs, Covmton, LA). 19 Stahc acid assay kit (Kyokuto, Tokyo, Japan) 20 DEAE-cellulose (DE 52, Whatman, Clifton, NJ).

2.1. HPL C Systems A Gilson HPLC System equipped with two Gilson Model 302 pumps, a Rheodyne 7125 inJector, a Fiatron TC 50 column heater, an ISCO V4 UV detector, Cis and Cs Spherisorb HPLC columns (0.46 x 25 cm, Phase Separation, Norwalk, CT), Hypersil C,s RP-HPLC column (Alltech Associates, San Jose, CA), and Perkm-Elmer LS40 scanning fluorescence detector. Eluant 1; 50 mM ammonium acetate, Eluant 2; 50% acetomtrile m 50 mA4 ammonium acetate. 2.2. High-Performance Anion Exchange Chromatography (HPAEC) Dionex BioLC system (Sunnyvale, CA) equtpped with CarboPac PA1 column (0.4 x 25 cm) and a pulsed amperometric detector (PAD-II), Helmm gas (99.995%), 0.2M NaOH solutron, 0.3M NaOH solution, 1.OM sodium acetate (HPLC grade) solution. Solutrons of 0.2M NaOH and 0.3M NaOH are prepared by diluting 50% w/w NaOH solution (Fisher Scientific).

Lee et al. 3. Methods 3.1. Preparation

of Glycopeptides

by Pronase Digestion

1 Dissolve 5 g of bovme fibrinogen m 50 mL of 0 lMTns-HCI buffer, pH 7 5 2 Add 50 mg of Pronase (see Note 2) and Incubate at 37’C for 5 d with dally addltlons of 50 mg of Pronase Freeze-dry the digestion mixture 3 Dissolve the freeze-dried sample m 20 mL of pyndme-acetic acid buffer, pH 4 7 4. Apply the sample to a Sephadex G-50 column (5 x 200 cm), elute with the same buffer as described above and collect 20-mL fractions 5 Assay the fractions (50 pL) by phenol-sulfuric acid method (see Chapter 1) and by absorbance at 280 nm 6 Pool the glycopeptides fractions, freeze-dry and keep at -20°C The final recovery yield is about 40 pm01 from 5 g of bovine fibrinogen

3.2. Separation

of Glycopeptides

on DEAE-Cellulose

1 Equilibrate a DEAE-cellulose column (1 5 x 20 cm) with 2 Mphosphate buffer, pH68 2. Dissolve the glycopeptldes (10 ~01) m 3 mL of the 2 mA4 phosphate buffer, pH68 3 Apply the sample to the DEAE-column, and wash the column with 20 mL of equlhbratmg buffer Elute with a convex gradient of 2-150 &phosphate buffer, pH 6 8, generate m a mixmg chamber of 80 mL at a flow rate of 0 3 mL/mm 4 Collect 2 2-mL fractions and assay for the glycopeptldes by the phenol-sulfuric acid method and slalic acid assay kit (14). Glycopeptides are separated mto neutral, monoslalyl, and dlslalyl fractions. 5 Pool the glycopeptldes fractions, freeze-dry, and desalt on a Sephadex G- 10 column (1 x 50 cm) equilibrated with 10 mA4 ammonium acetate, momtormg AZZOnm at 2 absorbance unit full scale (AUFS) 6 Collect the glycopeptides, freeze-dry, and keep at -20°C.

3.3. Analysis of Glycopeptides 1 Dissolve the dried monoslalyl glycopeptldes (10 nmol) m 50 pL of 10 mM ammonium acetate, pH 5 0 2 Add 1 mU of neurammldase from Arthrobacter ureafaczens m 50 pL of 10 rnM ammonium acetate, pH 5 0 and Incubate at 37°C for 16 h 3. Analyze by HPAEC (see Chapter 6) Elute the column with 100 mM sodium hydroxide at 1 n&/mm, with a linear gradient of sodium acetate from O-400 mMm 30 min (Fig. 2)

3.4. Structure

Characterization

of Monosialyl

Oligosaccharide

1. Dissolve the dried monosialylated glycopeptides (15 nmol) m 20 pL of 50 mM sodium-citrate buffer, pH 3.5 2 Add 0.4 mU of glycopeptidease A and mcubate at 37°C for 15 h (see Note 3) 3 Apply the reaction mixture to a Sephadex G-15 column (1.6 x 30 cm) and elute the column with water.

Conformational

241

Analysis by Resonance Energy A

NeuGc NeuAc

I 20

I 30

Time (min) Fig. 2. Analysis of monosralylglycopeptrdes by HPAEC (A) Chromatogram of rsomertc monostalyl glycopeptrdes; (B) chromatogram of the peaks after neurammtdase treatment 4. Collect the vord volume peak fractions and freeze-dry 5. Dissolve the dried ohgosaccharides in 40 pL of 2-aminopyrrdme solutron (1.0 g/580 PL of concentrated hydrochlorrc acid, see Note 4) and heat at 90°C for 10 mm in a heating block. 6. Add 10 $ of NaBHsCN solutron (20 mg/12 pL of water, freshly prepared) and heat at 100°C for 1 h. 7 Apply the reaction mtxture to a Sephadex G- 10 column (1 6 x 30 cm) and elute with 10 mA4 ammonmm acetate 8. Monitor the eluate by A ?s4,,,,,, collect the void volume peak fractions and freeze-dry. 9 Analyze the pyridylammo (PA)-ohgosaccharide with a RP-HPLC column (Hypersil ODS 5 micron, 0.45 x 25 cm) by comparmg with standard PA-oligosaccharides (Fig. 3, see Note 5 and Chapter 7) The column is eluted with a gradient of Eluant 2 from 5 to 20% oyer 35 mm at a flow rate of 1 mL/mm Monitor the eluate with a fluoromonitor (excrtatton 320 nm, emrssron 400 nm)

Lee et al.

242

20

10

30

Time (min) B Perk,

NeuAca(Z-6~al~(l-4)GlcNAcp(l-Z)Ma~(l-6I,

Manp(l-4)R

NeuAco(Z-6)Galp(l-4)Gl~NAcP(1-2)MaIla(l-3~ Peak2

Galp(l-4)GlcNAcp(l-Z)Malrr(l-6)-,Menp~,~4~R NeuAca(Z-6)Gal~~-4)GlcNAc~l-Z)Mam(l-3)/

Peak3

Peak.

NeuAca(Z-6)Galp(l-4)GlcNAcp(l-2)Mam(l-6)~Man~(,~4p Galp(l-4)GlcNAcp(l-2)Mam(l-3)’ Galf3(l-4)GlcNAcP(l-2)Mam(l-6~Manp(,-4~R Galp(l-4)GlcNAep(l-2)Manz(l-3)/ R= GlcNAq3(1--4)GlcNAcp-PA

Fig. 3. RP-HPLC analysis of PA-oligosaccharides. (A) The mixture of biantennary PA-ohgosaccharide was separated on a C,s column (Hypersll ODS 5 mtcron) (B) Structure of peaks 1,2, 3, and 4

3.5. Preparation of Doubly Fluorescence-Labeled Glycopeptide 3 5.1. Oxidation of Terminal Gal on Mana 1-6Man Branch and Modification with 2-(dansylamido)ethylamine

I

1 Dissolve the monoslalylated glycopeptldes (3 pool) m 200 pL of 100 mM phosphate buffer, pH 7.0, contammg 50 pg/mL catalase 2. Add 20 U of galactose oxldase in 20 pL of 100 Mphosphate buffer, pH 7.0, to the above solution (described m step 1) and incubate at 37°C for 18 h 3. Apply the reaction mixture to a Sephadex G-10 column (1 6 x 30 cm) and elute with distilled water (monitormg A220 “,,, at 2 AUFS) 4 Collect the void volume peak fractions and freeze-dry 5. Dissolve the dried sample m 1 mL of 100 mM sodmm phosphate buffer, pH 6 5 6 Add 20 mg of 2-(dansylamtdo)ethylamme borane-pyndme complex (neat)

m 200 @ of ethanol and 20 & of

Conformational

Analysis by Resonance Energy

243

A

I 0

20

40

60

Time (min) Fig. 4 RP-HPLC analysis of dansyl-labeled biantennary glycopeptides after Nnaphthyl-2-acetylation (A) The dansyl-labeled biantennary glycopeptides were purtfied on a C, scolumn. Peak 1 contams Asp and Glu and peak 2 contains Asp, Glu, and Gly (B) Peaks 3 and 4 represent the products of N-naphthyl-2-acetylation of mixture of peaks 1 and 2, respectively. (C) Chromatogram of the neurammidase treated products of mixture of peaks 3 and 4

7. Incubate the reaction mixture at 37°C for 16 h 8. Apply the reaction mixture to a Sephadex G- 10 column (2.5 x 100 cm) and elute with 10 mM ammonium acetate (monitoring A254“,,, at 2 AUFS) 9. Collect the product peak and freeze-dry 10. Dissolve the modified glycopeptide m water and mlect ( 100 nmol/SO pL) on to the C ,s RP-HPLC column (0.46 x 25 cm), momtormg A2s4“,,, at 2 AUFS. Elute the column at 1 mL/mm with a linear gradient of Eluant 2 from 30 to 70% over 55 mm (Fig. 4A) 11, Pool the product peaks from multiple chromatographtc runs, freeze-dry and keep at -20°C.

244

Lee et al.

3.5.2. Conjugation of 2-Naphthylacetic to Dansyl Labeled Glycopeptides

Acid

1 2-Naphthylacetic acid (11 umol), 1-hydroxybenzotrtazole (10 pool), dtcyclohexylcarbodttmtde (10 pmol) are mixed m 1 mL of N’N-dtmethylformamtde (DMF) and stir at room temperature overnight 2 Filter the reaction solutton through a Whatman No 1 filter paper drsc. 3 Dtssolve the dried dansyl-labeled glycopepttdes (1 pool) m 1 mL of DMF containing 10 pmol of activated 2-naphthylacettc acid and incubate at 55°C for 6 h 4. Apply the reaction mixture to a Sephadex G-10 column (2 5 x 100 cm) equrlrbrated with 50 n-u!4 ammonmm acetate (monitoring A254“,,, at 2 AUFS) 5 Collect the void volume peak fractions and freeze-dry 6 Dissolve the dried products m water and purify the products on a C,s RP-HPLC column (Fig. 4B) The elutton condtttons are same as described m step 4.

3.6. Preparation of Doubly Fluorescence-Labeled 3.6.1. Conjugation of 2-Naphthylacetic Acid to Monosialylglycopeptides

Glycopeptide

II

1. 2-Naphthylacetrc acid (1 1 mol) 1s activated by mrxmg 1-hydroxybenzotrrazole (1 mol) and dtcyclohexylcarbodumlde (1 mol) m 5 mL of DMF and stir overnight at room temperature. 2 Filter the reaction mixture through a Whitman No 1 filter paper disc 3 React the glycopeptrdes (10 ~01) in 5 mL of DMF containing 1 mol of activated naphthyl acetic acid at 55°C for 6 h. 4. Apply the reaction mixture to a Sephadex G- 10 column (2 5 x 100 cm) and elute with 20 mM ammonmm acetate, detecting peak by AZs4“,,, at 2 AUFS 5 Collect the void volume peak fractions and freeze-dry 6. Dissolve the dried products m water and apply to a Cs RP-HPLC column (0 46 x 25 cm) Elute the column at 1 mL/min with a linear gradient of Eluant 2 from 5 to 50% over 40 min, monitoring A254“,,, at 2 AUFS For preparative separation, the Cs column is eluted tsocratically with 8% of Eluant 2 in Eluant 1 7 Collect and combme glycopepttdes peaks from multiple chromatographrc runs and freeze-dry

3.6.2. Periodate Oxidabon of Terminal Siallc Acid on Mana 1-6Man Branch 1 Dissolve the naphthyl-2-acetylated

glycopeptides (1 5 pool) m 6 mL of 10 ti

sodium phosphatebuffer, pH 7.0 (chllled in an Ice bath) 2 Mix 6 mL of 2 mA4 perrodate m 10 rnA4 phosphate buffer, pH 7 0, with above solutton (This solutton should be prepared fresh Chilled m an ice bath ) 3. Keep the reaction mixture m the dark at 0°C for 20 mm 4. Terminate the reaction by adding 200 pL of 1M sodmm metabtsulfite (freshly prepared) m the phosphate buffer, pH 7 0, and leave the mrxture m the dark for 10 mm

Conformational

245

Analysis by Resonance Energy

5 Apply the reaction mtxture to a Sephadex G- 10 column (1 6 x 30 cm) and elute with water (A2*s ,,,). 6 Collect the void volume peaks fractions and freeze-dry

3.6.3. Conjugation of 2-(dansy/amido)ethy/ to Periodate-Oxidized Glycopeptides

Amine

1 Dtssolve the periodate-oxidized monosialylated N-2-naphthylacetylated glycopeptide (1 pool) m 1 mL of sodmm phosphate, pH 6.5. 2 Add 10 mg of 2-(dansylamido)ethylamme in 200 pL of ethanol to the above solution 3 Add borane-pyridine complex (10 @, neat) and keep at 37°C overmght 4 Apply the reaction mixture to a Sephadex G-10 column (1 6 x 30 cm), elute with 50 mM ammonmm acetate (momtormg A234“,,, at 0 5 AUFS) 5 Collect the void volume peak fractions and freeze-dry 6 Dtssolve the drted products m water and apply to a Cs RP-HPLC column (0.45 x 25 cm), elutmg with a linear gradient of Eluant 2 from 20 to 60% over 45 mm at 1 mL/mm (momtormg Ass4,,,,, at 2 AUFS). 7 Pool the product peak fractions from multiple chromatographm runs and freeze-dry. 8. Keep the sample at -20°C until resonance energy transfer measurement

3.7. Stepwise Digestion with Exoglycosidase

of Fluorescence-Labeled

Glycopeptides

1 Dissolve the dried fluorescence-labeled glycopepttde I (250 nmol) m 100 pL of 10 mM ammonmm acetate, pH 5 4 2 Add 25 mU of neurammidase m 7 pL of 10 mA4 ammonmm acetate and incubate at 37°C for 18 h 3. Apply the reaction mixture to a C t s RP-HPLC column (0.46 x 25 cm) elute at 1 mL/min with a linear gradient of Eluant 2 from 30 to 70% over 55 mm (momtormg A2s4 nmat 2 AUFS) 4. Pool the product peaks from multiple chromatographtc runs and freeze-dry. 5 Keep a portion of the product (50 nmol) for resonance energy transfer measurement 6. Dtssolve the remainder m 100 pL of 100 mM cttrate-phosphate buffer, pH 4 3 7 Add 100 mU of P-n-galactosidase and mcubate at 37°C overnight 8. Apply the reaction mixture to a C1s RP-HPLC column (0 46 x 25 cm) and elute as described m step 3 9 Pool the product peaks from multiple chromatographic runs and freeze-dry. 10 A portion of the product (50 nmol) is used for resonance energy transfer measurement. 11 Dissolve the remainder m 100 pL of citrate-phosphate buffer, pH 4.3, and mcubate with 2 5 U of l!I-N-acetyl-o-glucosaminidase m 50 pL of 3 2M ammonmm sulfate at 37°C overnight. 12 Purify the resulting product with a Cis RP-HPLC column (0 46 x 25 cm) eluted as descrtbed m step 3

Lee et al.

246

13 Pool the product peak fractions from multiple chromatographlc runs and freeze-dry 14 Keep a portion of the product (50 nmol) for resonance energy transfer measurement 15 Dissolve the remainder m 50 pL of sodium acetate buffer, pH 4 3 16 Add 1 U of P-D-mannosldase and incubate at 37°C for 18 h 17 Apply the reaction mixture to a Cl8 RP-HPLC column (0 46 x 25 cm) and elute as described m step 3 18 Pool the product peaks from multiple chromatographlc runs and freeze-dry 19. Keep the final product for resonance energy transfer measurement. 20 Sequentially digest glycopeptlde II with exoglycosldase (wlthout mltlal neurammldase dlgestlon) and purify as described in step 3

Part II: Resonance 4. Introduction

Energy Transfer

Measurement

Distance measurements between two chromophores by resonance energy transfer methods relies on a parameter referred to as the Forster distance (9). FGrster distance defines the intrinsic or potential interaction between a donor and an accepter and its value varies with the spectroscopic characteristics of the donor-accepter pairs. If two probes are placed wlthm a distance equal to the Forster distance of this pair, then the decay rate of the donor 1s equal to the energy transfer rate to the accepter. Typically the Forster distance of a partlcular pair of probes 1s measured or selected first. This IS covered m Subheading 6.1. For a simple average distance measurement, several methods are available, either m a conventional fluorometer, or m a time resolved fluorescence mstrument. The procedures are in Subheading 6.2. Distance distribution measurement is generally performed with a time-resolved fluorescence mstrument. The steps are m Subheading 6.3

5. Materials 1 2 3 4 5 6 7. 8

9.

A pair of quartz cuvets for absorption measurement A quartz fluorescence cuvet (Hellma, Forest Hill, NY, Uvomc, PIamvlew, NY) A spectrophotometer for absorption spectrum A fluorescence spectrophotometer (SLM, Rochester, NY, Perkm-Elmer, Norwalk, CT; Shlmadzu, Colombia, MD) A reference compound with known quantum yield (see Note 6) An optional reference compound with known fluorescence spectral shape in the wavelength range of interest A solution of accepter with optical density not exceedmg 1 at its peak absorption A solution of donor with optical density not exceeding 0 1 at Its peak absorption HPLC grade buffer should be used to mmlmize background fluorescence (see Notes 7 and 8) A solution of a reference compound with known quantum yield (see Note 9)

Conformatronal

Analys/s by Resonance Energy

247

10. An optional solutton of a reference compound wtth known spectral shape (see Note 10) 11 A personal computer for numerical calculation 12. Access to a time-resolved fluorescence mstrumentation, either m single-photon counting mode or in phase/modulation mode A ptcosecond laser ltght source is preferred 13. Access to computer software capable of analyzing experimental data by nonlmear regression methods and data deconvolutton Steps 12 and 13 are for ttmeresolved measurement

6. Methods 6.1. Quantum Yield and F&Her

Distance

1 Measure the absorptton spectrum A(h) of the accepter wtth a known concentranon C and calculate the molar absorptton coefficient E(X) as a function of wavelength h: a(h) = A&)/C, assummg that the cell length IS 1 cm (see Note 11) Measure the absorptton spectrum of the donor. Prepare a donor solutton with opttcal denstty at the peak less than 0.1. Measure the donor emtsston spectrum on a fluorescence spectrophotometer with a smgle excitatton wavelength (see Note 12) Correct the fluorescence spectrum using F,(h) = F(h)lS(h), where F(h) is the measured emtssion spectrum of donor and S(h) IS the instrument response flmctton (see Note 13) 5 Prepare a solutton of reference compound of known quantum yteld, wtth optical denstty less than 0 1 at the donor excttatton wavelength as described m step 3 6. Measure the fluorescence emtsston spectrum of the reference compound under identtcal instrumental conditions as those of the donor Correct the spectrum (see Note 14). 7 Sum over all emtsston wavelengths to get the total corrected mtenstty of the donor, I, and that of the reference compound, Ire+ 8 Calculate the quantum yield of the donor 4 = I &.fAref/(AIref), where A is the absorbance of the donor, and &rand Arer are the known quantum yield and the absorbance of the reference compound 9 With the values obtained from steps 1 and 4, calculate the overlap Integral between the donor and acceptor as step 15

where the wavelength h 1sm unit of nanometer, molar absorptton coeffictent E 1s m M-l cm-l, and the sum 1sdone over all wavelength range. 10. Calculate the Forster distance R0 (m A, no conversion m units in J) as Ro6 = 8 785 x 10” u2+Jln4

where n is the index of refraction of the solution (see Note 15) and tc2 IS the ortentatton factor, generally taken as Z/3 (see Note 15)

Lee et al.

248 6.2. Steady-State Energy Transfer and Average Distance Measurement 6.2.1 Donor Quenching Method-Steady-State Measurement at a Single Wavelength

1, Prepare one denvatlve-contammg donor only and the other derlvatlve containmg both the donor and the accepter Adjust the samples to the same donor concentratlon (or determme the donor concentrations so that a proper scalmg can be done) 2 Excite at the donor absorption wavelength and measure the emlsslon spectrum of the donor-alone derivative and that of the derivative contammg the donoraccepter pair (see Note 17). 3 Measure the intensity at the peak emlsslon wavelength for the donor, ID, and that for the donor-accepter, IDA at the same wavelength Normalize them to the same donor concentration Calculate the energy transfer efficiency E = 1 - IDAlID

4 Calculate the average distance between the donor and accepter (see Note 18) r=&(l/E-

1)“6

6.2.2. Accepter Enhancement Method: Steady-State Measurement at a Single Wavelength 1. Measure the fluorescence intensity of the derivative containing the accepter only at its peak emlsslon with excitation wavelength at the donor absorption region. 2 Measure the fluorescence intensity of the accepter m the derivative contammg the donor-accepter pair with the same excitation wavelength as m step 1 3 Compare the accepter fluorescence intensities m the two denvatlves. The increase m the accepter fluorescence m the derivative containing the donor-accepter pair 1sa result of energy transfer from the donor, after normalizing the concentration of the accepters (see Note 19)

6.2.3. Donor Quenching Method: Time-Resolved Measurement at a Single Wavelength 1 Measure the fluorescence decay of the derlvatlve contammg

donor only (see

Note 20)

2. Measure the fluorescence decay of the derivative contammg donor-accepter pair with the same excltatlon wavelength 3. Analyze the donor-only decay with a sum of exponential and obtain its average lifetime zD (see Note 21). 4 Analyze the donor decay m the derivative with the donor-accepter pair and obtain its average lifetime tom 5 Calculate the energy transfer efficiency using E = 1 - (z~A/zD).

Conformatlonal

Analysis by Resonance Energy

Table 1 Average Donor-Accepter Removeda

Distances

as a Function

Antenna 6

249 of Sugars

Residues

Antenna 6’

Sugar residues removed from antenna 6’

Avg distances (+O 1A)

Sugar residues removed from antenna 6

Avg dtstances (k0 1A)

Intact Gal 6’ GlcNAcS Man4’

17.2 17 2 17 1 17 3

Intact NeuSAc and Gal6 GIcNAc.5 Man4

17.5 17 7 17 3 174

aTemperature 20°C Table 2 Average Donor-Accepter

Distances

as a Function

Intact btantennary samples 6 6’

Temp (“C) 1 20 40

174 17 2 170

18 3 17 5 166

of Temperaturea Single chain isomers 6 6’ 17.5 173 17 0

18 1 174 16 5

aDlstance IS In A 6. Calculate the average distance between the donor and the accepter, using r = R”(lIE-

1)t16

The average distance obtamed for biantennary Tables 1 and 2

6.2.4. Curve-Fitting

glycopepttde

are shown m

Method: Steady-State Measurements

1 Prepare soluttons of the derivatives with the donor-only and that with the accepteronly. Measure the exact concentrations (see Note 22). 2. Measure the donor fluorescence and divide the spectrum by its concentration to get the normalized spectrum, F,(h) 3. With the same excitation wavelength, measure the emtsston spectrum of the accepter Normalize to its concentration to obtain, F,(h) 4 Prepare a solutton of the denvattve containing the donor-accepter pair and measure its concentration 5. Measure the emission spectrum of the derivative with the donor-accepter pan with the same excitation wavelength as in step 2 Normalize to its concentratlon,

FDA@)

Lee et al.

250

.Q .? P a a

2ooc

0.5

,

,

Ef c4

/’

: , , \

I

I

---

mtenna

6

-

mtenna

6’

\ /L

I

1.0 -

I

.--

_.-0.0

/

I

I

c I

1

.

--_

I

I

I

, I 1

I I

4o”c

0.5 -

.-0.0 - - -,- - , 5

, cc I_

/

I

I

,’

’ 1

\

I 25

10

30

Fig 5. Distance dlstrlbuttons of antennae 6 and 6’ m the intact btantennary glycopeptrdes. The distance drstrrbutrons were determined at 1, 20, 40°C 6 Use the followmg equation to obtain the energy transfer efticlency FDA(V = %I( 1 - -w-Lo> + (&A+~~D)~A@) where EA and ED are the molar absorptton coeffkrents of the accepter and donor at the same excnatton wavelength, respectively (see Note 23). 7. Calculate the average distance between the donor and accepter as in Subheading 6.2.1.

6.3. Distance Distribution

Measurement

1. Prepare a solution of the derivative with donor-only, measure the donor fluorescence decay. 2. Analyze the decay in terms of a sum of exponential lD(z) = C, a,exp[-t/r,]

Conformational

251

Analysis by Resonance Energy

25

5 I&m

I

(A)

I

I

5 Dklce

I 25

(A)

Fig 6 (A) Distance dlstrlbultlons of antenna 6’ in the partially digested blantennary glycopeptldes as a function of sugar residue removed from antenna 6 (B) Distance distributions of antenna 6 m the partially digested blantennary glycopeptldes as a flmctlon of sugar residue removed from antenna 6’ The distance distrlbutlons were determined at 20°C Sugar residues trunmed are indicated in the figure

where a, and z, are the amplitude and lifetime of the ith exponential (see Note 21). 3. Prepare a solution of the derivative with donor-accepter, and measure the donor fluorescence decay 4. Analyze the decay m terms of a distance dtstnbutlon, p(r), using the parameters obtained m step 2 as input. IDA(t) = Ip(r) C, a,exp[-tlz,( 1+ f&Jr} 6)]dr Obtain the distribution of distance between the donor and accepter (see Note 24) The distance distributions of biantennary glycopeptlde are shown in Figs. 5 and 6.

252 7.

Lee et al

Notes

1 The glycopepttdes produced by pronase dtgestton are too hydrophilic to be separated by RP-HPLC columns (Cs and C,,) To facilitate momtormg the galactose oxidase reaction on destalylated glycopepttdes by RP-HPLC. asialoglycopeptides are N-2-naphthylacetylated The oxidation of the terminal galactoses on N-2-naphthylacetylated glycopeptides can be monitored on a Cl8 column (0 46 x 25 cm), tsocratically eluted at 1 mL/mm using 12% of Eluant 2 (A,,, ““, at 0 02 AUFS) Under these condtttons, the dtaldehyde product elutes earlier than the monoaldehyde and unoxidized glycopeptides. 2 To prevent microbial growth durmg the long mcubation, the pronase solution IS sterilized by passing through a mtrocellulose filter (0 22 p pore size) Alternatively, toluene (20 &/mg pronase) can be added 3 The released ollgosacchartdes can be monitored using HPAEC equtpped with CarboPac PA-l column (0 4 x 25 cm) and a pulsed amperometnc detector (PAD-II) The column 1s eluted with a linear gradient of sodmm acetate from 0 to 300 mA4 over 30 mm m 100 mA4 sodium hydroxide at a flow rate of 1 mL/mm The pH of the solution is 6 8, when diluted with 10 vol of water These conditions were developed to prevent destalylation of sialylated oltgosaccharides Many authentic PA-oligosaccharide standards are available commercially (Nakano Vinegar Co., Handa-city, Japan or Takara Shuzo, Seta, Japan) The choice of a particular donor-accepter pair depends on the approximate dimension of the system to be studied. One can choose a pan of probes with a Forster distance as large as possible if the dimension of the system is not known Since the Forster distances of many pairs of probes have been measured (Z5), they can be used as an mtttal guide m selecting a donor-accepter pan compatible with the system of interest In general, the donor must be fluorescent and the accepter can be either fluorescent or nonfluorescent Similarly, reference compounds with known quantum ytelds and spectral shape m the range of 300 to 600 nm can be found (15) Since only the donor quantum yield IS needed to determine the Forster distance, the reference compound is selected with a fluorescence spectral shape close to that of the donor 7 Most fluorescence compounds are light sensitive Avoid long-term exposure even to room light 8 The properties of free donor may differ from those of the donor attached to another molecule This can be a change either m its quantum yield or m us spectral shape, both of which may affect the value of the Forster distance 9 The reference compound with known quantum yield can be dissolved m a different solution from that of the donor of interest. 10 The optional reference compound with known spectral shape is used only when the mstrument response function of a fluorescence spectrophotometer is not known. Otherwise, one can skip this step 11 If the sample cannot be weighed accurately, a molar absorption coefficient at a particular wavelength from the literature may be used. The concentration IS then calculated and the molar absorption coefficient as a function of wavelength IS

Conformat/ona/ Analysis by Resonance Energy

12

13

14

15

16.

17.

18

253

obtained. Make sure that the wavelength interval IS the same m the absorptton and fluorescence measurements, e g , 1 nm, so that subsequent calculattons can be easily performed. Two types of spectra can be obtained on a fluorescence spectrophotometer One is the excitation spectrum m which the emtssion wavelength is fixed and the excttation wavelength is scanned For a single donor system, the excttatton spectrum closely mimics the absorption spectrum m shape The second type IS the emission spectrum where the excttation wavelength is fixed and the emission wavelength 1s scanned For an emission spectrum, set the sltt width of both excitation and emtsston monochromators at 2 or 4 nm, and set the excitation polanzer m the vertical directton and the emission polarizer at 54 7°C from the vertical (called the magic angle). Set the excnatton wavelength at the peak absorption and start to scan the emission wavelength 5 nm above the excttatton If the donor emtsston spectrum overlaps with tts absorptton spectrum, shaft the excttatton to a lower (shorter) wavelength and obtam the entire donor emission spectral curve Most mstruments have a built-m mstrument response function to correct for the variation of light transmtsston and detection efficiency Otherwise, the instrument response at a certain wavelength range can be obtamed with reference to a fluorescent molecule wtth known spectral shape In thts case, S(h) = F&)/F,(h), where F(h) is the measured emtssion spectrum and F,(h) IS the known and corrected spectrum of the reference compound. The donor quantum yield IS measured with reference to a standard Look for a fluorophore as a standard wtth known quantum yield and absorptton and emtsston spectra similar to those of the donor (15) If the solution IS aqueous with moderate salt concentratton, take the refraction index as 1 4. When the buffers are very different for the donor and for the reference compound, then make a correction m the quantum yield and Forster distance (15) In many references the orientation factor remains a confusmg issue Thts arises primarily from the misunderstanding of the contrtbutton of ortentatton of two probes to the average dtstance determmatlon This has been clarified (16) and the value of 213 can be used If the sample fluorescence mtensity is very low, the background fluorescence may be significant Always check the buffer so that the background contribution can be corrected The use of 213 m the orientation factor does not introduce more than 10% error m distance calculation m most cases. There may be some deviation from the 2/3 average, when both probes are oriented at a partrcular dtrectton and are mrmobile. For solution studies, this can only occur when both probes are inside of a macromolecule and are stertcally constrained. In almost all energy transfer experiments, one or more probes are introduced by chemtcal modification. Two or more single bonds between the probes and the sites of modification are sufficrent to give enough sampling on the mutual ortentation between the two probes and lead to an averaged ortentatton factor not far from 213

254

Lee et al.

19 In general, accepter enhancement is more suitable for quahtative measurement of energy transfer because more parameters are required to calculate energy transfer efficiency (9) 20 The average hfetime of a probe is proportional to its steady-state Intensity and can be obtamed without the exact concentration measurement Either smgle photon-countmg method or phase method can be used In either method, amplitude (not intensity) averaged hfetime should be used Fluorescence lifetime IS an important consideration m choosmg a donor besides quantum yield If the lifetime IS too short, then it is difficult to measure energy transfer efficiency m the donor-accepter pan where the donor lifetime is further quenched by the accepter 21 Always start with a smgle exponenttal decay model m analyzing the fluorescence decay. The goodness of a fit can be Judged by the reduced-X*, which should be close to 1.O for a satisfactory fit Plotting of the residuals or the weighted residuals (difference between experimental and tit data) and the autocorrelation of the restduals is very useful. If a fit IS good, both should be randomly distributed about zero, If one exponential model does not fit, then try a two-exponential model Sometimes there may be a contribution of either Raman scattering or stray light leakage to the detector A scattered hght term may be used As the number of exponenttals increases, more fitting parameters are used Thus as a general rule, use a mnnmum number of parameters m a fit 22 This method is useful when the emission of the donor overlaps with that of the accepter and as a result, it may be dtflicult to obtain either the fluorescence mtensity of the donor or that of the accepter 23, The energy transfer efficiency can be obtained by nonlinear least regression methods using a computer Many statistical software packages contam nonlinear regression routmes Several levels of complexity can be used If the molar absorption coefficients or concentrations cannot be measured accurately, then a more elaborate global analysts procedure may be utilized (I 7). 24 The distance integration should cover as large a region as possible If one distance distribution is not sufficient, use two distance distributions (II) The use of 213 in the orientation factor may make the width of a distance distribution appear wider Thus p(r) contains a contribution from the orientation In general, the distribution of orientation factor is very difficult to model due to its peculiar shape (26) If there are flexible linkages between the probes and the sites of modiflcatton, the contrtbution of ortentation to the distance distribution should be small

References 1. Dwek, R. (1993) Basis and prmciples of glycobiology. FASEB J 7, 133&1336 2. Opdenakker, G , Rudd, P , Pontmg, C , and Dwek, R (1993) Concepts and prmctples of glycobiology FASEB J 7, 1330-1337 3 Srikrishnan, T , Chowdhary, M. S., and Matta, K L (1989) Crystal and molecular structure of methyl 0-a-o-mannopyranosyl-( 1-2)-cr-o-mannopyranoside Carbohydr Res 186, 167-175

Conformational

Analysis by Resonance Energy

255

4. Davoust, J , Mtchel, V, Spik, G , Montreml, J , and Devaux, P F (198 1) Flextbthty of bt- and trtantennary glycans of the N-acetyllactosamme type FEBS Lett 125, 27 l-276 5 Imberty, A., Gerber, S , Tran, V., and Perez, S (1990) Data bank of three-dtmenstonal structures of dtsaccharides, a tool to build 3-D-structures of ohgosaccharides GlycoconJ J 7,27-54. 6 Edge, C J., Joao, H C , Woods, R. J , and Wormald, M R (1993) The conformattonal effects of N-lmked glycosylation Bzochem Sot Truns 21,452-455 7. Homans, S. W (1990) A molecular mechanical force field for the conformattonal analysts of ohgosacchartde comparison of theorettcal and crystal structures of Man1 3Manl4GlcNAc. Bzochemzstry 29,91 l&91 18 8 Wormald, M. R and Edge, C J. (1993) The systematic use of negative nuclear Overhauser constraints m the determmatton of oltgosacchartde conformattons appltcatton to sralyl-Lewis X Carbohydr Res 246, 337-344 9 Stryer, L (1978) Fluorescence energy transfer as a spectroscoptc ruler Ann Rev Blochem 47,8 1!&846 10 Clegg, R M (1995) Fluorescence resonance energy transfer Curr Bzotechol 6, 103-l 10 11. Rice, K. G., Wu, P., Brand, L., and Lee,Y. C (1993) Modrficatton of oltgosaccharide antenna flexibility Induced by exoglycostdase trimming Blochemlstry 32, 7264-7270 12 Debetre, P, Montreal, J , Mocker, E , van Halbeek, H , and Vhegenthart, J F G (1985) Primary structure oftwo maJor glycans of bovine fibrmogen Eur JBzochem 151,607-611 13 Lee, K. B and Lee, Y C (1994) Transfer of modtfied sialic acids by Trypanosoma cruzl trans-stahdase for attachment of functional groups to ohgosacchartde Anal

Blochem 216,358-364 14. Sugahara, K., Sugimoto, K , Nomura, 0 , and Usul, T (1980) Enzymattc assay of serum stahc acid. Clm Chrm Acta 108,493498. 15 Wu, P. and Brand, L ( 1994) Resonance energy transfer* methods and appltcatton Anal Blochem 218, I-13. 16 Wu, P. and Brand, L. (1992) Orientation factor m steady-state and time-resolved resonance energy transfer measurements. Blochemlstry 31, 7939-7947 17 Flammon, P.J., Cachra, C , and Schretber, J. P (1992) Non-linear least-square methods applied to the analysis of the fluorescence energy transfer measurements. J Blochem Bzophys Methods 24, l-l 3.