The Challenge and Promise of Glycomics

Abstract

Introduction

Glycoconjugates exert their biological functions through complex molecular mechanisms involving both direct glycan recognition, and indirect glycan contributions to conformation and expression of the glycoconjugate. Glycans are directly recognized by glycan-binding proteins (GBPs) (Figure 1). Such interactions of glycans with GBPs can promote cell adhesion, cell-matrix interactions, cellular signaling, glycoprotein folding, and intracellular/extracellular targeting to organelles. In addition, glycans attached to macromolecules exert control through indirect mechanisms on glycoprotein conformation, stability, oligomerization, cell surface resident time, and turnover. Most secreted and membrane proteins are enzymatically glycosylated on one or more amino acids (Apweiler et al., 1999; Steentoft et al., 2013; Van den Steen et al., 1998; Zielinska et al., 2012; Zielinska et al., 2010), and virtually all nuclear and DNA binding proteins, cytoplasmic enzymes involved in metabolic regulation, and some mitochondrial proteins have the O-β-N-acetylglucosamine (O-GlcNAc) modification on Ser/Thr (Bond and Hanover, 2013; Copeland et al., 2013; Hart et al., 2011; Palaniappan et al., 2013; Yi et al., 2012). An emerging paradigm of modern glycomics recognizes that higher animals express a vast repertoire of glycan structures, which comprise their overall glycomes, and recent studies are unexpectedly revealing that glycans contribute in both general and specific ways to almost all biological regulatory pathways. Thus, as might be predicted, abnormalities in glycoconjugate synthesis or turnover are associated with hundreds of different human diseases and disorders, including the Congenital Disorders of Glycosylation (CDGs) and Dystroglycanopathies such as Congenital Muscular Dystrophy (Dennis et al., 2009; Filocamo and Morrone, 2011; Freeze, 2013; Hennet, 2012; Ju et al., 2013; Ohtsubo and Marth, 2006). Developing insights into glycan functions and the complexities of glycan structures and conformations represent both the challenge and promise of glycomics, the field of science now recognized as focused on glycans, just as genomics and proteomics are focused on nucleic acids and proteins, respectively.

Open in a separate window

Figure 1

The general roles of glycans in glycoproteins involve both their direct recognition by glycan-binding proteins (GBPs), and indirect effects of glycans on glycoprotein interactions, which may be dependent on protein-protein, protein-lipid, or even carbohydrate-to-carbohydrate interactions. This includes both glycoprotein on the cell surface, cellular organelles, and in secretions, as well as intracellular glycoproteins, e.g. O-GlcNAc glycoprotein. The bottom depicts many of the major classes of glycan linkages to proteins and lipids, along with the symbol key for representing glycan structures with abbreviations of monosaccharides and other substituents.

This article will focus on some of the major challenges and promises of the emerging field of glycomics, both structural and functional. While the term glycomics usually denotes the chemical aspects of glycobiology, we use this term here as a shorthand to denote a broad set of research and knowledge in chemistry and biology of glycans in terms of structure, function, biosynthesis, role in biology and disease, etc. Our major emphasis will be on human and animal systems, but we make a point to acknowledge that glycomics and glycoscience, represent broad areas of knowledge and research encompassing human and animal biology, as well as plant, fungal, and microbial systems.

Glycan and Glycan Binding Protein (GBP) Diversity

Glycans occur as both simple and complex structures in thousands of glycoconjugates, which include glycoproteins, proteoglycans, glycolipids, and as free or un-conjugated glycans. The factors regulating glycan expression and their molecular and functional roles have long been a supremely challenging puzzle. Moreover, our knowledge of the types of glycans and the number of glycan-amino acid linkages is growing, and the nature of the “core structures” of glycans in both glycoproteins, glycolipids, and glycosaminoglycans (GAGs), and glycosylphosphatidylinositol (GPI)-anchored glycoproteins is expanding at an astonishing rate, fueled by development of genomics, proteomics, and mass spectrometry-based tools (Figure 1). Mammalian glycomes are built from 9 common sugars (Glc, GlcNAc, Gal, GalNAc, Man, Fuc, GlcA, Xyl, and NeuAc (sialic acid) with a 10^th sugar IdoA being created within pre-synthesized glycosaminoglycans. At least 9 amino acids are known to be glycosylated in nature (Asn, Arg, Ser, Thr, Tyr, Trp, Cys, Hydroxylysine and Hydroxyproline) (Spiro, 2002; Stepper et al., 2011). The surface membrane of cells may contain over 10 million glycans linked to Asn and Ser/Thr residues alone, with the concentration of terminal sugars such as sialic acid approaching 100 mM (Wang et al., 2013). The spatial and temporal organization and functions of all these mammalian glycans is largely unclear. While proteomics is defining protein-protein and protein-nucleic interactions that are estimated to be in the many hundreds of thousands (Garma et al., 2012; Venkatesan et al., 2009), with >70% of proteins encoded in the human genome containing at least one identifiable protein interaction domain (Liu et al., 2012a; Pawson and Warner, 2007), glycomics has not yet advanced enough to estimate the number of protein-glycan interactions, i.e., what can be termed the protein-glycan interactome. While there is a growing list of mammalian GBPs with defined carbohydrate-binding domains and carbohydrate-binding modules (see (Gupta et al., 2012; Taylor and Drickamer, 2011; Varki et al., 2009; Vasta and Ahmed, 2009) and www.cazy.org), many new interactions (Tateno, 2010), especially those involving GAGs, are being discovered and will require further mechanistic insights to generally define protein motifs governing glycan recognition. Because the historical term lectin refers to mainly soluble multivalent proteins capable of agglutinating cells, and lacking enzyme activity, the commonly understood definition of lectin would preclude many membrane bound monovalent proteins, antibodies that bind glycans, toxins, microbial adhesins, enzymes that bind glycans, engineered glycosidases, and GAG-binding proteins. Thus, the more general term GBPs encompasses all of these types of proteins, although the word lectin appears in many databases. We denote in Figure 1 “Direct Glycan Recognition,” where a glycan determinant is directly bound by a GBP, and “Indirect Glycan Effects” that range from glycan-glycan interactions (Handa and Hakomori, 2012), protein solubility, protection from proteolysis and immune surveillance to affecting protein conformation and associations of glycoconjugates in the plasma membrane, for example. Given the growing evidence that the human and individual animal glycomes contain thousands of glycan species (Cummings, 2009), as well as evidence for hundreds of different GBPs encoded in animal genomes (Hileman et al., 1998; Schulenburg et al., 2008; Zhang, 2010) Figure 2, it is possible that the number of protein-glycan interactions may approach that for protein-protein and protein-nucleic interactions, especially considering that many of the GBPs in animals interact with glycans derived from the microbiome, as well as those expressed by the myriad of pathogens that infect these animals (van den Berg et al., 2012).

Open in a separate window

Figure 2

Glycan-Related Genes. Compilation of data from several databases identifies glycosylation-related human and mouse genes (glycogenes). Modified and updated from Nairn et al., (Nairn et al., 2008).

Studies on glycan structures and functions are complicated by the facts that most glycans in human and animal tissues remain poorly defined structurally, expression is typically cell type-specific and developmentally- and differentiation-dependent, little is known about the factors governing glycosylation on specific proteins and at specific sites, and virtually nothing is known about the overall architecture and topology of glycan expression on and within cells. However, recent developments in this field have been spurred, not only by the important biological functions of glycans being revealed through genetic studies, but also through breakthroughs in technologies involving mass spectrometry, as well as glycan microarray analyses that demonstrate the wide distribution and binding specificities of GBPs. Together, the genetic information and GBP information are being combined to promote a functional understanding of glycans, which has been termed functional glycomics. While the current picture presented here underscores the tremendous challenges in the field, the landscape of Glycoscience is rapidly changing and the field is at the threshold of significant breakthroughs that will reverberate throughout biology. Such discoveries are highlighting the position of glycans as one of the four major classes of life’s macromolecules, on center stage of modern biomedical and chemical research (Marth, 2008).

Glycan Microarrays and Glycan Determinants

Protein-glycan interactions were historically studied using the hapten-inhibition approach in laborious precipitation or hemagglutination-type assays. Such approaches were far from high-throughput, lacked sensitivity, and required large quantities of valuable glycans, thus limiting binding studies to monosaccharides and short oligosaccharides or poorly defined, complex glycoconjugates isolated from natural sources. This approach led to the prior paradigm that GBPs have relatively low affinity and lack strong specificity. However, within the last decade and beginning with the introduction of ELISA-type assays in the late 1990’s, studies have used microarray technologies in which large libraries of complex glycans are immobilized on surfaces, akin to nucleic acid arrays. These glycan microarrays have helped to revolutionize the field of glycobiology and provided the ability to interrogate putative binding specificities of GBPs by researchers in a wide spectrum of biomedical and basic sciences, in a fashion unimaginable a decade ago. Glycan microarrays can be interrogated by indirect or direct fluorescence-based techniques to identify specific binding of glycans by lectins, antibodies, toxins, viruses, etc. (Figure 3).

Open in a separate window

Figure 3

To explore protein-glycan interactions one major technology in the field is glycan microarrays. In such an approach, individual glycans from natural sources or chemo/enzymatic syntheses, are modified to allow their automated printing and attachment, either covalently or non-covalently, to a slide-type surface in defined positions, akin to a gene array. The glycan microarray can be interrogated with a GBPs or other reagents or even cells and viruses, to identify those glycan “spots” that are recognized, which can be visualized by either direct or indirect fluorescent tagging. The spot pattern on the image, which should also incorporate replicates of each glycans, can be averaged to generate a histogram. In the example shown, the results indicate that that the GBPs in question bound strongly to one glycan, less strongly to another glycan, and did not bind appreciably to any other glycan. Such microarrays can also be prepared from glycolipids, glycopeptides, whole glycoproteins, or polysaccharides of animal, plant, or microbial origins.

The Consortium for Functional Glycomics (CFG), developed through a Large Scale Collaborative Project from NIH/NIGMS, made available a printed glycan microarray (Blixt et al., 2004) that continues to be available to the research community (Rillahan and Paulson, 2011; Smith and Cummings, 2013), and data generated is posted and publicly available (functionalglycomics.org/). Over 15,000 glycan microarrays or slides from the CFG have been used over the past decade by researchers in ~3,000 array experiments with >1,000 types of samples containing GBPs, including human and animal sera, as well as intact eukaryotic and prokaryotic cells. The database of the CFG is the world’s largest catalog of protein-glycan interactions. The work of the CFG is complementary to those of many other laboratories world-wide involved in developing glycan microarrays and probing them with a variety of GBPs and other reagents (Chevolot, 2012; de Paz and Seeberger, 2012; Liu et al., 2009; Park et al., 2013). Smaller libraries of glycans have also been usefully explored for binding to GBPs using solution-based approaches (Arata et al., 2001; Chang et al., 2011; Duverger et al., 2010; Gutierrez Gallego et al., 2004; Takeda et al., 2013). However, a need in the future is for all users of glycan microarrays to deposit their data in publicly-available databases for curation and cataloging in relation to the larger effort to define the protein-glycan interactome.

The results of glycan microarray studies using different platforms and different types of glycans are revealing that many GBPs including antibodies display relatively high affinity interactions with complex glycans, often involving 3–7 monosaccharide residues, that includes branched structures or saccharide modifications, such as sulfation or phosphorylation. Researchers are moving away from a focus on the minimal determinants that might inhibit a lectin binding to a single glycan, such as millimolar concentrations of monosaccharides or simple glycans, to the concept of the biological and physiological interactions between relevant naturally occurring GBPs and relatively complex glycan structures (Wang et al., 2013). The concept of glycan determinants (Cummings, 2009), which are the minimal glycan structures that confer the maximum glycan binding affinity, may be thought of as akin to antibody epitopes or glycotopes (Cao et al., 1996). This concept helps us to understand the contributions of individual glycan features and biosynthetic pathways to glycan recognition. However, such array studies naturally raise questions as to their physiological relevance, and there are concerns of glycan “presentation” on microarrays and the degree to which this presentation is biologically relevant to their presentation on cell surfaces or native glycoconjugates (Park et al., 2013).

It should be self-evident that binding or lack of binding of GBPs to glycans on glycan microarrays or other surfaces does not directly provide evidence of functionally significant interactions, but such results can provide hypotheses to be tested regarding particular glycan functions. In many cases, ligands predicted by glycan microarray analyses have been shown to have physiological relevance, whereas in some cases the connection is less clear. Examples of the former are results of virus binding, such as influenza viruses and noroviruses, which bind in microarrays and in vivo to sialylated (Blixt et al., 2004; Stevens et al., 2006; Walther et al., 2013) and blood group H-type glycans (Lindesmith et al., 2003; Shang et al., 2013), respectively. Another example of concordance of glycan microarray data and physiological functions are studies on CD22 (Siglec-2), which was shown to bind glycans with NeuAcα2-6-linked sialic acid on cells in vitro (Macauley et al., 2013; Powell et al., 1993; Sgroi et al., 1993), in solution studies (Powell and Varki, 1994), in glycan microarrays (Campanero-Rhodes et al., 2006; Tateno et al., 2008), and in mice (Hennet et al., 1998). A possible example of discordance between glycan microarray data and physiological ligands involves recent studies on Siglec-F in mice. Both human Siglec-8 and its murine homolog Siglec-F preferentially recognize the same sialylated glycan 6′-sulfo-sialyl Lewis X (6′su-SLe^x - NeuAcα2-3[6-SO₃]Galβ1-4[Fucα1-3]GlcNAc-R), in part based on binding to immobilized arrayed glycans (Bochner et al., 2005; Kiwamoto et al., 2013b; Tateno et al., 2005). This Siglec is expressed in eosinophils and its binding to 6′su-SLe^x has been proposed to be important in mitigating allergic eosinophilic airway inflammation. Studies in vivo indicate, however, that while St3gal3 mutants lacking the sialyltransferase had diminished Siglec-F binding and more intense allergic eosinophilic airway inflammation (Kiwamoto et al., 2013a), deletion of the sulfotransferases capable of generating 6-sulfated galactose, as determined by glycomic analysis, does not appreciably affect Siglec-F binding in a mouse model (Patnode et al., 2013). Thus, in all cases where it is experimentally feasible, it is critical to link the results of glycan microarray analyses for GBPs glycan determinants with physiological evidence for such glycan determinants being functionally important. It is also important to note that current glycan microarrays lack the full presentation of structures found in the human and animal glycomes, thus many potentially important glycans and glycan determinants are lacking, and thus lack of binding of a protein to a glycan microarray may simply indicate that the relevant glycan ligand is missing.

While the number of structurally different glycans in the human glycome is presently unknown, it is likely to be many tens of thousands. The repertoire of glycan determinants in the human glycome, however, is estimable. An approach to this problem is to conceptually assemble partial determinants or segments of glycan structures with each other in biochemically allowed and defined ways based on known biosynthetic pathways and glycan structures (Cummings, 2009; Rademacher and Paulson, 2012; Werz et al., 2007). Such a combinatorial approach permits calculations as to how many such determinants might exist in N- and O-glycans and glycolipids, as well as glycosaminoglycans (GAGs). An illustration using partial determinants shows that that there are possibly many thousands of glycan determinants (Figure 4). Few of these glycan determinants are currently available on available microarrays (the current CFG glycan microarray has 611 different glycans); thus, there is a tremendous need for chemical and enzymatic syntheses of thousands of compounds. Fortunately, the NIH has funded the development of small libraries of synthetic glycans as a first step in expanding the availability of glycans to the research community (sbir.cancer.gov/funding/contracts/fy2013_09.asp). Coupled with advances in asymmetrical synthesis of glycans (Wang et al., 2013), these efforts should be encouraged since a large number of glycans are obviously needed for structural and functional studies and as standards for nuclear magnetic resonance and mass spectrometry (MS) analyses. In any case, the need for increased number of synthetic glycans in sufficient quantities to impact glycan analysis and functional studies far exceeds their availability.

Open in a separate window

Figure 4

GBPs often recognize 2–6 linear or branched monosaccharides with additional modifications, such as phosphorylation, sulfation, O-acetylation, etc. Such recognition can be termed a glycan determinant, which is the minimal glycan structure that confers maximum binding. These glycans determinants can be thought of as being assembled from partial determinants, that in the example shown are modified galactose, modified Gal-GlcNAc, and modified GlcNAc-Man. With the partial determinants shown, which are respectively, 6, 9, and 3 in number, it is possible to assemble these into 162 different glycan determinants. Shown also are two of these as Determinant 1 and 2, which are differently recognized by a GBPs or GRM. If one considers all such known partial determinants in human glycans defined to date, it is possible to predict that there are over 5,000 glycan determinants; if the GAG sequences are also included up to pentasaccharides, then there are an additional 10,000 or so. It is likely that this is an underestimate for the total theoretical glycan determinants, since it is likely that other partial determinants will be identified in the future. In addition, it is possible that in a single branched or linear glycan, the one set of glycan determinants may attenuate the recognition of the same or a different glycan determinant on the same molecule.

Milk and Glycolipid “Metaglycomes” as a Paradigm for Glycomics

While the human glycome certainly represents a challenge in terms of defining all its structures, the meta-glycomes of cells and tissues may be amenable to total analysis. We use the term, “metaglycome,” as denoting a constituent glycome of a specific cell or tissue, as well as a type or family of glyconjugates; summing all the human metaglycomes would then define the human glycome. In fact, targeting specific metaglycomes and providing their repertoire of glycans would likely constitute a milestone toward the direction of eventually defining the human glycome. Two such metaglycomes that would be amenable to such analyses are the soluble human milk oligosaccharides or glycans and human glycosphingolipids. The sizes of both of these metaglycomes have already been estimated, with the HMO comprising several hundred glycans (Bode, 2012; Urashima et al., 2011), and the human glycosphingolipid metaglycome comprising at least 500 different glycosylated neutral and acid species, based on the Lipid Maps database (www.lipidmaps.org/), without consideration of the aglycone lipid constituents. Importantly, technologies appear in place to allow quantitative and qualitative descriptions of both the free glycans in human milk (Bao et al., 2013; Ninonuevo et al., 2006) and releasable glycans from glycosphingolipid glycomes (Fujitani et al., 2011). Interestingly, both of these metaglycomes have terminal Glc as the non-reducing sugar (lactose-type glycans in milk and lactosylceramide-type glycans in glycosphingolipids), so technologies for defining the glycans may be overlapping and synergistic. Recent technological advances indicate that glycosphingolipid metaglycomes may be approached using intact lipids, maintaining information regarding aglycone dynamics as well ((Boccuto et al., 2013)

A common conundrum for the glycomics field is: how will researchers know that they have fully defined a metaglycome? The evidence that a metaglycome is largely, if not fully, defined would be based on the pace of discovery of new glycan species within such metaglycomes in the future; thus, at some point in time the pace of discovery would lessen to the point that one could estimate that >95% of the glycans below a certain reasonable size limit, e.g. 5,000 daltons, have been described for that metaglycome. The definitive description of a metaglycome to that level, with calculated and statistical reasoning, would go far in assisting the development of technologies and bioinformatic approaches that will provide researchers with their first breakthrough paradigm in defining at least two specific component metaglycomes of the human glycome. A clear challenge to the field of glycomics is that if we cannot fully define the HMO and glycosphingolipid metaglycomes, how can we possibly imagine defining the glycoprotein-derived metaglycomes, which are clearly much more complex?

The complexity of glycomes parallels the technical difficulties in analyses, as well as the number of glycan determinants that might exist (Figure 5). Thus, one could imagine that GAGs, with their multitude of potential disaccharide repeating units, would represent the most complex set of glycans in the human glycome, while GPI-anchors and human milk oligosaccharides, may represent the relatively least complex set. Thus, future studies in glycomics should focus to some degree on the metaglycomes that are most definable, and functional glycomics can then define the recognition of those glycans by GBPs, or the specific roles of those glycans in human physiology and disease. Such a strategy could lead to the identification of the protein-glycan interactome for that particular metaglycome.

Open in a separate window

Figure 5

The different classes of glycans, from the GPI glycans within GPI-anchored glycoproteins, to the GAGs, represent different levels of complexity and numbers and diversity of glycan determinants. Of course with complexity and increased numbers of determinants, the analytical difficulties expand tremendously. Thus, a great challenge of glycomic technologies is to difficult better methods of preparing and analyzing glycans to overcome the challenges of their complexities.

In the absence of either knowing all the structures in the human glycome, or having all glycan determinants available as synthetic compounds, it is possible to use a shotgun glycomics approach, prior to definitive structural characterization, to obtain all the cellular glycans and use fractionated glycan species for functional studies on glycan microarrays and other surfaces. Such a shotgun approach has been successful in many cases for preparing total glycosphingolipid-derived glycan libraries and other types of natural glycans (Liu et al., 2012b; Song et al., 2011; van Diepen et al., 2012). The combined approach of general glycomics, the release of all the glycans from endogenous glycoconjugates and subsequent structural analysis, with functional glycomics, where individual purified glycans species are probed for their functional recognition by GBPs, is highly likely to make significant strides to unravel both the mysteries about the glycan structures that make up the glycome and their recognition by endogenous and pathogen-derived GBPs.

Glycan Recognition Molecules (GRM)

Over a hundred different plant lectins and a few from invertebrates have been identified, and many, such as Concanavalin A (Jack bean) and Helix pomatia agglutinin (snail), have been utilized successfully to study glycan expression and function. The use of glycan microarrays has allowed the binding determinants of a large number of lectins to be explored, thereby greatly facilitating their utility in testing hypotheses regarding glycan function. The repertoire of human and mouse glycan structures recognized by the plant lectins commonly available is quite limited, however, and often a single lectin binds to multiple glycan structures carrying minimal glycan determinants. In addition, analysis of the binding specificities of commercial plant lectins by the glycan microarray screening of the CFG has demonstrated remarkable inconsistencies among commercial preparations of the same lectin, including variable binding affinities and even inactive preparations. The field of Glycomics is at a point where more reliable and better-defined reagents that recognize a breadth of glycan structures are required. Moreover, if investigators encompassing the breadth of biomedical research had available these types of reliable reagents, a significant increase in our knowledge of glycan function would, no doubt, occur. The time is right to consider developing an extensive library of glycan recognition molecules (GRM) with specificities and determinants defined by binding to expansive glycan microarrays. These reagents, if made available to researchers with diverse health-related interests, would greatly accelerate the pace of glycan function discovery in physiology and disease. These reagents could also provide a direct means to target specific glycans for potential therapeutic and diagnostic applications. There are novel platform technologies in the early stages of development that can generate and select for molecules that recognize and bind specific glycan structures. These include antibodies or antibody-like proteins, as well as non-protein molecules such as nucleic acid aptamers (Li et al., 2008) and synthetic lectins (Ke et al., 2012). Since some of these glycan recognition molecules are not proteins, here we refer to this class as GRM, instead of GBP. Potential technologies for generation and selection of antibodies with specificity for specific glycan epitopes include engineering of glycans, glycopeptides, and novel adjuvants to elicit high-affinity IgG (Lakshminarayanan et al., 2012), yeast display of scFv antibodies (Zhao et al., 2011), and lamprey antibodies (Yu et al., 2012). In addition, the elucidation of glycan binding specificities of viruses, bacteria, both plant and animal lectins using glycan microarrays, as well as structural knowledge of these GBP plus many glycan-recognizing enzymes, provide a structural basis of glycan binding motifs. Knowledge of these motifs, coupled with detailed structural studies of glycan binding domains by crystallography and NMR spectroscopy, as well as molecular simulations of glycan-protein interactions, have led to an expanding understanding of how to engineer glycan recognition determinants, e.g., (Feinberg et al., 2013; Ford et al., 2003; Mercer et al., 2013). This knowledge could be exploited to assist the design and selection of GRM with glycan binding determinants and specificities of interest.

The prospect for the availability of libraries of GRM is somewhat analogous to the availability of antibodies directed against specific phospho-peptides for studies of cell signaling pathways, and the development of restriction enzymes for molecular biology. Availability of these sequence-specific reagents has allowed the investigation of signaling pathways and nucleic acids by non-expert investigators with diverse interests. Likewise, GRM would assist in testing for changes of glycan epitope expression and would also provide new tools for assessing functions by potentially blocking or crosslinking target glycans. Production of GRM will involve much more complex challenges than for peptide or phospho-peptide specific antibodies or restriction enzyme identification because of the diversity of glycan structures, and their precise binding determinants must be determined using glycan microarrays and other approaches, akin to what was done for the Cluster of Differentiation (CD) in immunology (www.sciencegateway.org/resources/prow/). Despite the obvious challenges, the availability of defined GRMs to pursue specific biological questions will likely be transformative.

Glycomic Analyses of Various Types of Glycoconjugates and Their Recognition by GBP

A clear challenge of glycomics is the complexity of analytical methods for individual classes of glycans, as discussed here. Each class requires different methods of extraction or glycan release, often different methods of analysis, or combinations of methods, such as HPLC, mass spectrometry, NMR, etc (Figure 6). Nevertheless, while complex, each of these methods has been adapted to provide deep insights into glycan structure for that particular glycan class. An all encompassing method of glycoprotein analysis at the top-down level would be ideal, and coupled with the bottom-up approaches below (Hanisch, 2012; Nicolardi et al., 2013), would help to define the relationships of glycans to their protein and lipid carriers and their relatively abundance to each other. Moreover, analytical approaches capable of probing the glycomes of living cells, as through biorthogonal methods (Belardi et al., 2012) or non-invasive magnetic resonance approaches, would provide breakthroughs into the dynamical aspects of glycomes in living tissues.

Open in a separate window

Figure 6

The general strategies for glycomic analyses typically involves isolating or generating free glycans from glycoproteins/proteoglycans and glycolipids. The obtained mixture of glycans can then be derivatized and directly analyzed by MS, or derivatized, separated by HPLC and other approaches, and further analyzed by MS or NMR. In shotgun glycomics for functional studies, the released glycans after separation can be printed to generate glycan microarrays. For site-specific glycosylation and identification of protein carriers, glycopeptides can be generated by proteolysis and then analyzed directly before or after glycan removal.

To Glycoscience and Beyond

Glycoscience has made spectacular progress in the past few years with the advent of new technologies for exploring glycan structure and function, along with insights into the genetic and molecular aspects of glycan expression and regulation. The field is challenged, however, by the complexity and dynamic nature of the glycome and lack of understanding of how glycoconjugates are expressed topologically and temporally. Understanding the functions of glycans and their higher order contributions typically require physiological studies of organisms and identification of altered pathways of anabolism and catabolism in patients (Figure 7). Thus, combinations of studies in single cell systems may fail to identify or even predict these higher order functions. As for any type of biological system, the degree of understanding can be gauged by its predictive ability, and on this score glycoscience is truly challenged. One Holy Grail for glycomics researchers is to be able to predict the glycan structure(s) to be found on any particular amino acid site on any glycoprotein of interest; another is for functional glycomics researchers to predict the molecular interactions of a glycan with GBPs and GRMs. At present, our state of knowledge is far from predictive on any level, and we are limited to identifying potential sites of glycosylation on proteins given their amino acid sequences. Defining glycosylation sites and structures of glycans at specific sites is still technically difficult and limited to laboratories with advanced tools, especially considering all the types of amino acid modifications that can occur (Fig. 1). A further challenge for glycomics is to understand the links between expression patterns and levels of glycosyltransferases/glycosidases and their localization, spatially and temporally, relative to particular glycan structures, as well as the consequences of such expressions, and that of GBPs, on biological activities. In spite of these challenges, the future is growing brighter, as our knowledge grows about glycan structures and both their direct and indirect functions through studies performed across many biological disciplines. Technological advances in mass spectrometry and other sequencing methods, glycan synthesis, bioinformatics, and growing knowledge of biological roles of glycans in development, health, and disease, notably the CDGs, provide great hope for the future of glycomics. The possibility exists to finally tie the knot linking glycans to nucleic acids, proteins, and lipids in the grand theory of everything biological, assigning glycans their place as a pillar among those macromolecules that are essential for life.

Open in a separate window

Figure 7

The higher order functions of glycans are typically revealed in physiological studies of organisms, and to a lesser extent perhaps in studies based on isolated tissues and cells. Thus, mutations of genes involved in specific glycosylation pathways of anabolism or catabolism might have little effects on cultured cells, but are deleterious to organismal development; increasing biological complexity is denoted by the green arrow. Of course, the analytical ease of defining the glycome itself is simplified by considering free or released glycans, compared to individual glycoconjugates, such as a glycoprotein, and becomes increasingly difficult with larger biological complexity to the organism itself, denoted by the blue arrow.

Acknowledgments

Abbreviations

Footnotes

References