Table Of Content

Characterization of Protein Glycosylation Elizabeth F. Hounsell 1. Introduction The majority of protems are posttranslationally modified, the most sigmfi- cant change being glycosylation, i.e., the attachment of one or more oligosac- charade 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 1su sed 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-glycolylneura- mmic 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 glycosylatron- specific 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). Ohgosaccha- rides or monosacchartdes released by enzymatic or chemical methods are sepa- Pro teln Glycosyla tion 3 rated by high-performance llqutd chromatography (HPLC), htgh pH amon- exchange 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). Dis- cussed m the present chapter are mass spectrometry (MS) and nuclear mag- netic 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. Proteo- glycans (see Chapter 9) also have a high carbohydrate/protein ratio. Their gly- cosammoglycan 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 1s becoming increasingly blurred; they can now be seen as a spectrum of the varying glycosylation patterns occurrmg on high and low molecular wetght, secreted,a nd cell-surface glycoprotems. As examples of this, classical mucm and proteoglycan sequences can occur on cell-membrane- attached 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 con- junction 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 Chap- ters 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 oh- gosaccharlde 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 per- formed 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, con- firmatory 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 reac- tions 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, mycobacte- ria, 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-Wacety- lated 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 1s dependent 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 ques- tion 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 sequencea nalysis 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. Protein Glycosyla tion 7 6 NMR analysis of >50 pg chromatographically pure ohgosaccharide or glycopep- tide 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 HZ0 (HPLC-grade) 2. 4% Aqueous phenol. 3 Concentrated H2S04 4 1 mg/mL Gal. 5 1 mg/mL Man 6. Orcmol (Sigma, Poole, UK) 2% (v/v) m ethanol containing 5% of H2S04 7 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 Analysis 1, 0.5M Methanohc HCl (Supelco, Bellefonte, PA) 2. Screw-top PTFE septum vials. 3 Phosphorous pentoxide. 4. Silver carbonate (Pierce and Warrmer, Chester, UK) 5. Acetic anhydride. 6 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 1MNaBH4 m 0 05M NaOH made up fresh 2 Glacial acetic acid. 3 Methanol 4 Cation-exchange column 5. PBA Bond Elut columns (Jones Chromatography, Hengoed, UK) activated with MeOH 6. 0.2M NH,, OH. 7. 0.01, 0.1, and 0.5MHCl 8. 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 elec- trochemlcal 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 .O mg 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. Protein Glycosylatlon 9 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 Chap- ters 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 Sub- heading 1.1. and Chapters 6 and 7