Members and sponsors are encouraged to share ideas, applications research and experiences, or technical problems by contributing to this column. In addition, short reviews covering aspects of biotechnology that are of interest to members are also encouraged. All contributions will be subject to peer review and the deadline for submitting articles for the June issue is May 1, 1994. Periodically, articles will also be solicited from non-members by the editors. Contributions may be mailed or faxed to the Editor, Ken Williams, W.M. Keck Foundation Biotechnology Resource Laboratory, P.O. Box 9812, 295 Congress Avenue, New Haven, CT 06536-0812, (203) 737-2206, Fax (203) 737- 2638. Contributions will not necessarily reflect the views of the Editorial Board or the Association.
Glycosylation is one of the more common modes of posttranslational modification of eukaryotic proteins. A variety of biological roles have been proposed for the carbohydrate of glycoproteins. These include presentation (or masking) of molecular determinants for recognition by cells, microorganisms, or other biological molecules, modulation of the half-life of glycoproteins, and physical/structural roles. There is growing evidence that the oligosaccharides of various glycoproteins, and other glycoconjugates, indeed fulfill all of the roles which have been ascribed to them. This has been recently reviewed in detail by Varki (1). Interest in the biological function of the carbohydrate part of all glycoconjugates has led to the term "glycobiology" as an umbrella for the study of the many different roles which these molecules play. This review will focus on the analysis of glycoprotein glycosylation, because glycoproteins are in many respects the best-studied class of glycoconjugates, and the ones most likely to be encountered in core facility laboratories.
Glycobiology has become important to a large segment of the protein analytical community because many of the proteins cloned and expressed as the first generation of biotechnology products are glycosylated. While early attempts were made to express mammalian glycoproteins (such as human- -1-protease inhibitor) in prokaryotic systems, the biological or pharmacokinetic properties of the recombinant proteins were often unsatisfactory. This seems to be one of the reasons that the biotechnology industry has invested heavily in the use of eukaryotic expression systems for the production of therapeutic (glyco)proteins. With the recent discovery of the selectins (2), there has also been a growing interest in the carbohydrates of glycoconjugates in their own right. A number of biotechnology companies are now attempting to develop drugs based on carbohydrates or mimetics which interfere with carbohydrate-mediated biological interactions (3).
As a result of the growing interest in glycobiology, there has also been growing interest in characterization of carbohydrates. The technology of carbohydrate analysis has lagged behind that for protein and nucleic acid analyses. Thus, while commercial sequencers for proteins and polynucleotides have been around for a while, only recently has a single, specialized automated "carbohydrate sequencer" for glycoproteins become available. This is largely due to the marked heterogeneity of oligosaccharide structures often found on a single polypeptide species. Thus, a single protein gene product exists in vivo as a complex collection of glycoprotein "glycoforms", differing only in the amount or structure of attached carbohydrate moieties. Carbohydrates are also the only biological polymers which often occur as branched structures, due to the multiple hydroxyl groups present on monosaccharides and available for glycosylation. The importance of the structural diversity of oligosaccharides to the science of glycobiology has recently been succinctly reviewed by Kobata (4). The extreme structural diversity possible for oligosaccharides is key to their postulated biological roles, but confounds structural characterization.
Glycoprotein glycosylation can be divided into two classes, referred to as N- and O-glycosylation. The best studied mode of glycosylation is the formation of an N-glycosidic linkage to Asn in the polypeptide chain. The necessary (but not sufficient) criterion for protein N-glycosylation is the presence of the sequence ("sequon") Asn-x-Ser/Thr in the polypeptide sequence (x may be any amino acid except Pro). O-glycosylation occurs at Ser and/or Thr residues in glycoproteins, although there is no known sequon for O-glycosylation. The O-linked carbohydrates tend to be shorter and simpler structures than their N-linked cousins, but at this point, they are not as well studied. O- glycosylation is nevertheless probably no less important than N-glycosylation (5).
Which proteins are glycoproteins? It has long been apparent that most mammalian plasma proteins are glycosylated, although there are significant exceptions, such as serum albumin. Many eukaryotic cell-surface receptors and other membrane-bound proteins are also glycosylated. Recent work by Hart and colleagues (reviewed in 6) has made it clear that many eukaryotic cytosolic proteins are also glycosylated, bearing residues of O-linked N-acetylglucosamine (GlcNAc). This latter type of glycosylation is "ubiquitous and dynamic", and may play a regulatory role, in apposition to phosphorylation (6). Another important class of glycoprotein glycosylation is the C-terminal, glycosylphosphatidylinositol (GPI)-glycolipid anchor found on many cell surface glycoproteins (30). Thus, one may find glycosylation in eukaryotic proteins at almost every turn. Prokaryotes, on the other hand, lack the necessary subcellular machinery to glycosylate proteins as described above. Except for the structural peptidoglycans, most prokaryotic proteins are not glycosylated.
How does one identify whether a given protein is glycosylated? Historically, one could assay a sample of a given protein for carbohydrate content by a simple, colorimetric assay such as the phenol-sulfuric acid method (7). The sensitivity of this classical method is insufficient for modern biochemistry. Methods such as periodate oxidation/Schiff's staining of SDS-PAGE gels may identify a protein as glycosylated, but are not terribly specific or sensitive. More recently, methods to identify protein glycosylation after Western blotting by immunostaining methods have been developed and commercialized by several suppliers (e.g., Oxford GlycoSystems, Boehringer-Mannheim). Typically, the blotted protein is subjected to periodate oxidation, which yields free aldehyde groups when there is carbohydrate present in the sample. The aldehyde is then reacted with a hapten (e.g., digoxygenin) modified to contain a hydrazide moiety. Anti-hapten antibody is then used to label the hapten-modified glycoprotein and detected by an ELISA strategy. Periodate oxidation under mild conditions can be used to specifically detect sialic acid-containing glycoproteins using a similar strategy. A modification of this approach above using hapten-modified lectins to detect specific carbohydrate motifs in Western blotted proteins has been reported (8). These simple, straightforward methods for initial characterization of glycoproteins are well-suited to the average biochemist or molecular biologist, but are generally inadequate for quantitative or fine structural analyses.
Carbohydrate Compositional Analysis
It is often valuable to determine the overall carbohydrate composition of a glycoconjugate. Compositional analysis gives information on the number and type of monosaccharide residues present in a glycoconjugate. In the case of glycoproteins, compositional analysis can sometimes give insight into the type of oligosaccharides present. For example, the presence of N-acetylgalactosamine (GalNAc) in a glycoprotein is typically (although not absolutely) indicative of O-glycosylation.
Monosaccharide composition analysis is analogous to amino acid composition analysis. Acid hydrolysis or methanolysis is usually used to release the monosaccharides as reducing sugars or methyl glycosides, respectively. Both GC and LC methods have been used to separate and quantify released monosaccharides. GC methods require derivatization, while LC methods have been reported with and without chemical derivatization (9, 33).
In recent years, high-pH anion-exchange chromatography with pulsed amperometric detection (HPAEC/PAD) has become a popular method for routine compositional analysis (10). While the technique is fairly simple and sensitive, a fair amount of user skill is required, and the absolute quantification and reproducibility of the method varies.
Oligosaccharide Analysis
Modern biochemistry generally requires more than compositional analysis of a glycoprotein. Complete structural analysis has been a daunting task, but developments in separation and detection technology, as well as a growing knowledge base of reported structures, is making the routine characterization of at least glycoprotein N-oligosaccharides a possibility. The remainder of this review will focus on the methodology available for the release and structural characterization of glycoprotein oligosaccharides.
Release of Oligosaccharides
The characterization of glycoprotein oligosaccharides generally requires their release from the underlying peptide. Either chemical or enzymatic methods can be used to release intact oligosaccharides for separation and structural analysis. Biochemists have tended to favor enzymatic methods, due to the mild conditions required, but recent advances in automation have increased interest in chemical methods to release oligosaccharides.
Enzymatic Methods
Virtually all known N-linked oligosaccharide structures can be released intact by use of the enzyme peptide N-glycosidase F (PNGaseF, N-Glycanase, EC 3.2.2.18). PNGase F is actually an amidase, which releases the N-linked oligosaccharide from the (poly)peptide chain as an N-glycoside, which spontaneously hydrolyzes to yield the free, intact, reducing oligosaccharide. PNGase F leaves the peptide chain intact, although the formerly N-glycosylated Asn residue is converted to Asp by the action of the enzyme. A thorough review of the use of endoglycosidases and glycoamidases is available (31).
Another useful class of enzymes for carbohydrate release are the endo-beta-N-acetylglucosidases (EC 3.2.1.96). These enzymes are true glycosidases; they cleave between the two GlcNAc residues in the N-linked oligosaccharide core. The released oligosaccharide is thus shortened by one GlcNAc residue, and the other GlcNAc residue remains linked to the peptide Asn glycosylation site. The best known enzymes of this type are Endo H and the Endo F family. These enzymes show considerable specificity for the types of N-linked structures which they will cleave. The "Endo" enzymes are especially useful for the study of oligomannosyl ("high mannose") and "hybrid"-type oligosaccharides (32).
The main drawback to using enzymes to release oligosaccharides is that peptide structure can often markedly affect the release of the oligosaccharide. It is often necessary to do considerable optimization of release protocols for each peptide, and some glycopeptides are refractory to one or more of these enzymes. For example, PNGase F can not cleave oligosaccharides from amino- or carboxy-terminal Asn residues. Further, there is no good "O-glycanase" activity which can be universally applied to de-O- glycosylate peptides.
Chemical Methods
The above-mentioned drawbacks have led to renewed interest in chemical methods to release intact oligosaccharides. One of the most useful techniques is the use of anhydrous hydrazine (hydrazinolysis) to release N-linked oligosaccharides. The basic mechanism is cleavage of all acyl bonds, including the N- glycosidic linkage of oligosaccharide to Asn and the peptide bonds in the protein itself. Conditions to release intact, reducing O-linked oligosaccharides by hydrazinolysis have recently been reported (11), although the mechanism is still unclear. A significant drawback to hydrazinolysis is the cleavage of N-acyl groups in sugar residues of the oligosaccharide (e.g., GlcNAc glucosamine). Thus it is necessary to re-N- acylate the released oligosaccharides, usually with acetic anhydride. If the sugar residues of the released oligosaccharides contained acyl groups other than acetyl (e.g., N-glycolyl-neuraminic acid, a type of sialic acid), this information could be lost after hydrazinolysis. Also, the polypeptide chain is generally degraded during hydrazinolysis, rendering it useless for peptide analyses after deglycosylation. The successful application of hydrazinolysis to release intact oligosaccharides has been something of an art, requiring considerable skill and very pure hydrazine. Recently, Oxford GlycoSystems (OGS) has developed an instrument to effect the automated release, purification, and recovery of N- and/or O-linked oligosaccharides. While very expensive, the OGS "GlycoPrep 1000" seems to be a reliable and useful tool for routine oligosaccharide release requiring relatively little user skill. The stability of phosphorylated, sulfated, and O-acylated sugars (e.g., O-acetyl-sialic acids) has not yet been well-established. Analysis of glycans suspected to contain these groups should probably be attempted using both chemical and enzymatic means.
Another method for releasing O-linked oligosaccharides from peptides involves the use of strong alkali plus reducing agent ( -elimination). While the method is very useful for O-glycosides, it yields the sugars in the form of reduced sugar alcohols (alditols), a chemical "dead end" for introduction of, e.g., a fluorescent label at the reducing terminus. On the other hand, the released sugars can be radiolabeled by using NaB3H4 as the reducing agent. Beta- elimination is useful for studies of the deglycosylated peptides, because it results in the formation of unique amino acid residues from the formerly O-glycosylated Ser and Thr residues.
Oligosaccharide Structural Analysis
Once the oligosaccharides are released, they must be separated and characterized. The separation and characterization process itself constitutes a whole branch of glycobiology, but a brief review of the mass spec, chromatographic and electrophoretic technologies in current use is within the scope of this review.
Historically, carbohydrate structures were determined using data from sequential digestions with exoglycosidases, regiospecific chemical degradation, methylation analysis (GC-MS), FAB-MS, and/or high-field proton and multidimensional NMR methods. While these methods still represent the gold standard for structure analysis, most are beyond the reach of protein analytical facilities due to cost or the lack of suitably trained personnel. Recent developments in technology have improved the ease of at least inferential oligosaccharide structural analysis and "fingerprinting" techniques (e.g., comparison with standards). Enough instrumentation and standards are now available to put some degree of oligosaccharide analysis capacity into any peptide core facility.
The Conundrum of Detection
To characterize oligosaccharides, one must separate them. To separate oligosaccharides, one must be able to find them. Most oligosaccharides lack a unique chromophore or fluorophore. Thus, considerable work has been done to develop detection methods for mono- and oligosaccharides. A review of detection methods (Table 1) highlights the advantages and disadvantages of these approaches.
The reducing terminus of an oligosaccharide is essentially an aldehyde in disguise. Thus, the chemical modification of the reducing terminus is relatively easy. As a result, derivatization of reducing sugars with a chromophore or fluorophore has been a popular route to sensitive and specific detection. Unfortunately, the reducing terminus of a sugar is less reactive than a typical aldehyde, so the conditions required for the derivatization may be harsh, and degradation may occur. Not all reducing termini will label at the same rate or to the same extent. Of the chemical methods, radiolabeling with NaB3H4 is probably the gentlest and most reliable for quantification. The continually harsher regulatory climate, though, will probably spur improvements in non-isotopic labeling methods, as has also been seen in the nucleic acid field.
The presence of multiple hydroxyl groups in oligosaccharides makes them potentially attractive electrochemical analytes. The technique of pulsed amperometry (PAD) has revolutionized the detection of underivatized carbohydrates. While PAD can be very sensitive, it is not exceptionally selective, and many interfering components (amino acids, buffer salts, etc.) can complicate analyses. The electrochemical response factor (PAD signal / mol analyte) is also rather variable from sugar to sugar. PAD can be used for quantitative analysis, but it is essential to calibrate the response factor of each analyte with an appropriate external standard.
HPLC Methods for Oligosaccharide Separation
In recent years, HPLC has been the method of choice to separate complex mixtures of glycoprotein oligosaccharides. Excellent reviews are available in the literature (e.g., 9, 33-35). Some of the currently popular approaches are summarized in Table 2.
HPAEC/PAD has been particularly valuable in separations of complex oligosaccharide mixtures. Many labs use HPAEC/PAD as a routine, fingerprint assay for glycoprotein oligosaccharide content. The use of high pH does impart some risk for base catalyzed alteration (epimerization, "peeling" reactions) of the oligosaccharides. Despite this, it is possible to preparatively isolate isomerically pure oligosaccharides from HPAEC/PAD for thorough structural characterization. HPAEC/PAD appears to be one of the best approaches to separate anionic oligosaccharides (e.g., oligosaccharides containing sialic acid, phosphate, or sulfate). Ion-exchange LC at lower pH has also been useful for separation of anionic oligosaccharides.
The use of size exclusion chromatography (SEC) has long proven useful for oligosaccharide separations. The most popular method (12) uses BioGel P4 to separate an admixture of the radiolabeled oligosaccharide alditols of interest and an internal size calibration standard consisting of a series of glucose oligosaccharides (isomaltooligosaccharides, from partially hydrolyzed dextran). Pure water is used as eluant. The separated oligosaccharides can be used for further structural analysis (e.g., exoglycosidase digestion). Note however that the isolated alditols are not chemically pure, due to the presence of unlabeled isomaltooligosaccharides. SEC is also useful as a fingerprinting technique. The major drawbacks to SEC are the inability to easily work with anionic sugars, and the insensitivity of size-based separation to many subtle, isomeric differences in oligosaccharide structures.
Method Sensitivity Specificity Comments Ref.
-------------------------------------------------------
Refractive Moderate Very Low Incompatible 9
Index (RI) ca.1 nmol with LC
gradient
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
Intrinsic Moderate Very Low Difficult to 15
Absorbance -high use for
(Low UV) nmol quantification
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
Reductive Amination
(aryl) amine +
NaCNBH3
(Chromophore, High High Reactions may 16
e.g., DAAB) nmol (aldehyde) be incomplete
Degradation of
some species
Easy to perform
(Fluorophore, High High As above 17
e.g., 2AP) pmol (aldehyde) be incomplete
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
Hydrazide
(e.g., Dansyl High High Reactions may 18
NHNH2) pmol (aldehyde) be incomplete
derivatives not
as stable as
above
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
NaB3H4 High High License required
Reduction 10 pmol (aldehyde) yields alditol 19
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
Amperometry High Moderate Response factor 20
(PAD) 10 pmol (hydroxyls) is species-
dependent
Table 2. LC and HPLC Methods in Common Use for Oligosaccharide Analysis
Method Column Comments Ref. ------------------------------------------------------- Normal phase Amine-bonded Good for underivatized silica sugars and radiolabeled alditols 21 - - - - - - - - - - - - - - - - - - - - - - - - - - - - Reverse-phase ODS-silica Generally used with derivatized sugars, can give excellent resolution (e.g. with 2AP) 22 - - - - - - - - - - - - - - - - - - - - - - - - - - - - Anion-exchange DEAE or Q Useful for sialylated, materials phosphorylated, and sulfated sugars 23 - - - - - - - - - - - - - - - - - - - - - - - - - - - - HPAEC Strong anion High resolving power; exchanger at some danger of alkaline high pH degradation 24 (Dionex CarboPac) - - - - - - - - - - - - - - - - - - - - - - - - - - - - SEC BioGel P4 Good method for purification and exo- glycosidase analysis; rather slow 12 - - - - - - - - - - - - - - - - - - - - - - - - - - - - Serial Lectin Various High resolving power Affinity immobilized with natural selectivity Chromatography lectins of lectins; laborious and time-consuming 25
Traditionally, SEC analysis has been powerful, but slow and cumbersome. Oxford GlycoSystem's GlycoMap 1000 hasrecently introduced rapid (< 24 hour), sensitive SEC analysis with on-line radioactivity and RI detectors. Dedicated computer software for instrument control, data acquisition and analysis is also available.
Many methods have been developed for the separation of oligosaccharides by reverse-phase HPLC (RP-HPLC). Since oligosaccharides are very hydrophilic, and lack a specific chromophore, virtually all of the RP-HPLC methods have entailed use of chemical derivatization to introduce a hydrophobic chromophore or fluorophore. Reductive amination using the fluorogenic compound 2- aminopyridine (2AP) has been particularly valuable for this application. Useful mapping techniques have been developed for 2AP-labeled oligosaccharides (13). TaKaRa Shuzo has introduced equipment to automate the 2AP derivatization, as well as columns for 2AP-sugar separations and a series of 2AP standards. 2AP oligosaccharides are good candidates for further structural characterization by classical means. The stability of some substituents (e.g., sialic acids) to the 2AP derivatization is still questionable.
Electrophoretic Methods for Oligosaccharide Separation
Paper electrophoresis is a time-honored approach for separation of neutral as well as charged oligosaccharides. Long run times, high voltages, and unpleasant buffer systems were the norm. Recent advances in glycobiology have led to new interest in applying electrophoresis methods for carbohydrate analysis. Gel electrophoresis methods are fast, require little equipment, and are very user friendly. Capillary electrophoresis offers the potential of highly efficient separations, high mass sensitivity, and rapid analysis times. Each of the latter methods is currently of much interest to analytical glycobiologists.
Gel Electrophoresis
The basic approach uses high-percentage polyacrylamide slab gels to separate fluorescently tagged oligosaccharides by electrophoresis. The resulting gel banding patterns are visualized by fluorography and/or sophisticated CCD imagers. A commercialized system, Fluorophore Assisted Carbohydrate Electrophoresis (FACE), has been developed by Glyko and Millipore. Interestingly, both companies are marketing hardware and reagents for FACE. FACE offers a rapid and sensitive method to separate and quantify both mono- and oligosaccharides. Demonstration of the resolving power of FACE for complex mixtures of oligosaccharides and oligosaccharide isomers has been limited.
Capillary Electrophoresis
CE separation and analysis of oligosaccharides remains an area of significant basic research. Table 3 summarizes some of the recent reports. Detection may be either by intrinsic absorbance, absorbance or fluorescence of end-labeled oligosaccharides, or by DC or pulsed amperometry. The latter modes are especially attractive due to the relative insensitivity of electrochemical methods to the "path length" of the capillary. Thus, DC amperometry or PAD might enable high concentration sensitivity as well as the traditional high mass sensitivity of CE. At this time, there seem to be no commercial products designed for CE mono- or oligosaccharide analysis, although several major suppliers are probably developing them.
Mass Spectrometry
As mentioned above, classical approaches to determination of monosaccharide composition and oligosaccharide structure have relied heavily on chemical derivatization (or fragmentation) and mass spectrometry (MS) and/or GC-MS. Mass spectrometry "is probably the most broadly applicable analytical tool in the chemical sciences...[and] in biological research as capabilities to address large molecules advance" (36). The relatively recent development of "soft" ionization techniques (FAB-MS, LD-MS, thermospray and electrospray MS) have even expanded the utility of MS for the analysis of large biopolymers. The sensitivity is often in the pmol range or better. Several modern MS methods are applicable to glycoconjugate analysis, either for purified oligosaccharides or for glycoprotein glycoforms (see, e.g., 37-40 for recent reviews).
The following section will highlight a few of these approaches (see also Table 4). While MS methods can provide very accurate masses of molecular or fragment ions, it must be remembered that many monosaccharide residues have the same mass (e.g., all neutral hexoses = ca. 162 Da), and NMR, chemical, or exoglycosidase degradation must be employed to determine the identity, linkage positions, and anomericity of monosaccharide residues by MS.
FAB-MS (also known as LSIMS) has been used for some time to determine the molecular weight of oligosaccharides. Purified glycopeptides or oligosaccharides may be analyzed directly by FAB-MS (see, e.g., 41), but derivatization of the oligosaccharides to enhance their surface activity markedly improves the sensitivity of the method (42). For example, a FAB-MS strategy to analyze oligosaccharide structures after periodate oxidation, NaBD4 reduction of the resultant aldehydes, and permethylation has been developed by Nilsson and colleagues (43).
Electrospray MS (ES-MS) has become a popular method for analysis of recombinant glycoproteins (37). Since the electrospray interface requires a continuous infusion of solvent, it is compatible with on-line LC-MS or CE-MS, thus enhancing the data obtained from a chromatographic "fingerprint" of a tryptic digest or a pool of released oligosaccharides. Since electrospray ionization produces families of multiply-charged ions, mass spectrometers with relatively modest mass ranges can be employed to analyze even large glycoproteins such as immunoglobulins with excellent intrinsic mass resolution. Thus, electrospray methods can, and have, been used to look at the glycosylation heterogeneity of intact glycoproteins such as ribonuclease (44). On the other hand, the very complex envelope of multiply- charged ions requires sophisticated deconvolution analysis to extract information on the original molecular weight(s) of the species present. Indeed, very large and heterogeneous glycoproteins can yield "a bell shaped hump" (45) of completely unresolved ions in the critical 1000-2000 Da mass region. In some cases, the best approach may be tying a high-resolution separation method (capillary HPLC or CE) to a relatively low-resolution mass spectrometer via an electrospray interface.
Analytes Capillary/Buffer Detection Ref.
-------------------------------------------------------
Mono/oligo- Fused silica/phosphate Reductive 26
saccharides pH 9.4 amination
laser-induced
fluorescence
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
Oligo- Fused silica/phosphate Low UV 27
saccharides pH 6.6 or Tricine (200 nm)
pH 8.2 + putrescine
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
Mono/oligo- Fused silica/100 mM DC amperometry 28
saccharides NaOH, LiOH, or KOH (Cu electrode)28
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
Mono/oligo- Fused silica/ PAD 29
saccharides dilute NaOH (Au electrode)
Table 4. Mass Spectrometry Methods for Glycoconjugate and Oligosaccharide Analysis
Method Application/Comments (indented) Ref. ------------------------------------------------------- CI-MS Oligosaccharide derivatization with 49 perfluorobenzylaminobenzoate - - - - - - - - - - - - - - - - - - - - - - - - - - - - FAB-MS Nilsson's strategy for structural analysis 43 Reducing-terminal modification with alkyl- aminobenzoates to enhance sensitivity 42 Analyses of mixtures of underivatized, sialylated glycopeptides 41 Sensitivity enhanced by derivatization; gives molecular ions - - - - - - - - - - - - - - - - - - - - - - - - - - - - ES-MS Selective detection of glycopeptides in 47 protein digests Analysis of glycoprotein glycoforms 44 Very sensitive; amenable to on-line detection for LC and CE; intrinsically good mass resolution; produces families of multiply-charged ions - - - - - - - - - - - - - - - - - - - - - - - - - - - - LD-MS Glycoprotein (IgG) heterogeneity 45,50 Released oligosaccharides 48 Simple, fast, relatively inexpensive; generally affords only molecular ions; mass resolution limited by length of drift tube; choice of matrix dye can influence sensitivity
Tandem MS (MS-MS) methodology, while very expensive, has proven useful for glycoconjugate and oligosaccharide analyses. For example, permethylation followed by MS-MS was used to analyze the glycoinositol phospholipid anchor of the scrapie prion protein (46), with a requirement of only ca. 100 fmol of the original protein sample. Carr and colleagues have reported the use of LC/ES-MS and LC/ES-MS/MS to selectively detect glycopeptides in proteolytic digests of glycoproteins (47).
One of the most intriguing of the recent MS technologies is matrix-assisted laser desorption MS (known as LD-MS or MALD[I]-MS). LD-MS instruments which measure the mass of the resultant ions by time-of-flight (TOF) in an evacuated drift tube are compact, inexpensive, and very easy to use. A number of LD-MS instruments are already commercially available and are ensconced in many bio-analytical laboratories. While only information on the intact molecular weight of the analyte is generated, the TOF approach permits the analysis of samples over an extremely broad mass range. Thus, it is possible to use LD-MS to look at glycosylation heterogeneity of a glycoprotein preparation (e.g., Ribonuclease B) or to determine the molecular weights of the released oligosaccharides themselves using a single instrument (48). A drawback to TOF mass spectrometers is relatively poor mass resolution, which renders the glycosylation heterogeneity of larger glycoproteins as a single, broad molecular ion peak. For example, LD-MS of human -1- acid glycoprotein produced a 36,800 molecular ion (M+H+) with a peak width of 2200 Da (40).
Oligosaccharide Sequencing
Once an oligosaccharide is isolated, the process of determining its sequence is not a trivial effort. In peptide or nucleic acid sequencing, one must determine which residue is linked to which. In oligosaccharide sequencing, one must also determine how each is linked. The holy grail for the analytical glycobiologist is the development of a generic oligosaccharide sequencer. The extreme diversity in oligosaccharides may prevent this completely, but the first steps towards automated glycoprotein oligosaccharide sequencing have been taken.
The specificity of the exoglycosidases has made them useful tools for oligosaccharide sequencing. For example, the beta-galactosidase from jack bean can cleave a nonreducing terminal galactose (Gal) residue when it is beta-linked to the 6 or 4 position of the penultimate sugar residue, but is almost inactive towards a Gal (beta,1-3) linkage. Conversely, the beta- galactosidase from bovine testes cleaves all three types of Gal linkage. Judicious use of the two enzymes can give insight into the galactosylation of a purified oligosaccharide, as long as the reaction products can be separated and quantified. Sequential exoglycosidase digestion, using a variety of purified exoglycosidases of known specificity, can be combined with the chromatographic or electrophoretic techniques mentioned herein to "sequence" an oligosaccharide. The approach is tedious, and requires considerable knowledge of the possible candidate structures of the unknown sample (or some lucky guesses).
Researchers at the Oxford Glycobiology unit and Oxford GlycoSystems have employed a parallel approach with exoglycosidases to develop an oligosaccharide sequencing strategy for tritium-labeled N-linked oligosaccharides (14). Their Reagent Array Analysis method (RAAM) uses mixtures of exoglycosidases in eight parallel digestions (and an enzyme-negative control). Each of the eight digests contains various subsets of a working panel of (five) exoglycosidases. Depending on the structure of the unknown oligosaccharide, a certain limit oligosaccharide will be produced by a given combination of enzymes. The digests are incubated, then all combined together. The combined samples are then fractionated by automated BioGel P4 SEC as described above. The set of digests produces a histogram (signature) of varying proportions of limit oligosaccharides of different sizes. The RAAM software compares the unknown's signature with its knowledge base of RAAM signatures from authentic standard oligosaccharides. The software will select the best matches, and assign an index of probability to each. The entire product, called "RAAM 1000 GlycoSequencer", has just been introduced.
The main limitation of the approach is that, in its present incarnation, it is limited to analysis of N- linked oligosaccharides. In principle, though, other exoglycosidases could be added to the scheme to extend the working space of oligosaccharide structures. The use of BioGel P4 SEC as a separation system requires that the candidate oligosaccharides for sequencing be neutral species; thus, RAAM can give no information on, e.g., the sialic acid linkages of an oligosaccharide. A two-dimensional approach would have to be used to fully characterize sialylated oligosaccharides.
Summary
A variety of recently-commercialized techniques have extended the feasibility of glycoprotein oligosaccharide analysis to the realm of the non-specialist. The sensitivity of detection and selectivity of separation strategies is adequate (barely) for modern research applications. The increased ease of oligosaccharide analysis will certainly be reflected in an increased interest in the biological effects of glycosylation, or changes in glycosylation.
Acknowledgments
I thank Dr. Jeff Rohrer of Dionex Corp. for his suggestions. I also thank Drs. Roger O'Neill and Joseph Olechno, Applied BioSystems, Inc., for comments, suggestions, and for supplying a preliminary form of Table 1.
References
Return to the The ABRF Home Page