Glycoanalysis Protocols

Characterization of Protein Glycosylation Elizabeth F. Hounsell 1. Introduction The majority of protems are posttransl...

Author: Elizabeth F. Hounsell

40 downloads 688 Views 15MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

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 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


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


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


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.


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).


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).


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.


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


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 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 (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,


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


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


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


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


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


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.

Glycoanalysis Protocols

Glycoanalysis Protocols

Neuropeptide Protocols

Retinoid Protocols

Polyamine Protocols

Kinesin Protocols

Chaperonin Protocols

Glycovirology Protocols

Arabidopsis Protocols

Ribozyme Protocols

Superantigen Protocols

p53 Protocols

Superantigen Protocols

HIV Protocols

Genomics Protocols

Polyamine Protocols

YAC Protocols

Immunochemical Protocols

Mycotoxin Protocols

Pesticide Protocols

Interleukin Protocols

GTPase Protocols

Kinesin Protocols

Bioconjugation Protocols

Genomics Protocols

Integrin Protocols

Integrin protocols

Flavoprotein Protocols

Genomics Protocols

Genomics Protocols

Glycoanalysis Protocols