Following steady advances in analytical technologies, our knowledge in glycomics is now increasing rapidly. Over the last decade, specific glycans have been described that are associated with a range of diseases, such as cancer and inflammation, with host–pathogen interactions and with various stages during stem cell development and differentiation. Simultaneously, deeper structural insight has been gained on glycosylated biopharmaceutical protein therapeutics manufactured in CHO (Chinese-hamster ovary) and other cell systems. This glycomic information is highly relevant for clinicians and biomanufacturing industries as a new class of glycobiomarkers emerges. However, current methods of glycoanalysis are primarily research tools and are not suitable for point-of-care on-site detection and analysis, or sensor devices. Lectin-based glycan detection provides the most promising approach to fill these gaps. However, the limited availability of lectins with high specificity and sensitivity for specific glycan motifs presents one of the main challenges in building reliable glycobiosensors. Recent reports have demonstrated the use of recombinant protein engineering, phage display and aptamer technologies in the production of lectin mimics, as well as the construction of biosensors that are capable of rapidly detecting glycan motifs at low levels in both a labelled and label-free manner. These are primarily proof-of-principle reports at this stage, but some of the approaches, either alone or in combination, will lead to functional glycobiosensors in the coming years which will be valuable tools for the clinical, biopharmaceutical and life science research communities.
- phage display
Glycans covalently linked to lipids and proteins are displayed on the surface of every cell, and the majority of secreted proteins are glycosylated. The glycosylation of proteins is one of the most common post-translational modification. Unlike the synthesis of nucleic acids and proteins, the biosynthesis of glycans is a non-template-driven process in which a large number of different branched oligosaccharide structures are generated from a handful of monosaccharides in the eukaryotic cell. As a result, the information-coding potential of glycans is enormous . Glycan-encoded information is ‘read’ by corresponding lectins (glycan-binding proteins) in numerous physiological processes such as cell–cell interaction, immune system response, protein folding, protein stability and protein trafficking [2–5].
Glycans are also involved in many chronic and infectious diseases, including cancer and bacterial and viral infections. Aberrant glycosylation, altered glycan–lectin interactions, pathogen exploitation of host glycome for invasion and targeted glycosylation for immune evasion are among the well-documented features of various diseases [6–8]. Early detection of disease is undisputedly a critical factor for successful treatment and improved quality of life, achievable through identification of earlier and more sensitive biomarkers. The disease-related glycome has become a valuable target for diagnosis of many diseases and also for therapeutic intervention.
In addition to their role in diseases, glycans are also critically important for performance of a large percentage of recombinant therapeutic glycoproteins [9,10], most of which are produced in CHO (Chinese-hamster ovary) suspension cultures. The glycosylation profile of these drugs determines their serum half-life, and pharmacokinetic, pharmacodynamic and immunogenic properties in patients. However, obtaining consistent glycosylation is a formidable challenge for biopharmaceutical industries because glycosylation is influenced by several factors in the eukaryotic cell, including culture conditions, growth phase of the cells and concurrent biochemical events . As a result, the biotechnology and biopharmaceutical industries have placed a significant emphasis on the characterization of the major carbohydrate structures on recombinant protein pharmaceuticals [10,11], a trend that has been fuelled by the requirements of regulatory agencies.
Despite its importance, characterization of glycosylation remains a considerable undertaking requiring specialized analytical expertise, expensive equipment and a serious time commitment. In the analytical laboratory setting, the glycan structures are hydrolysed chemically and/or enzymatically and then analysed using established instrumentation such as LC (liquid chromatography), MS and NMR. However, these methods are not suited to clinical laboratory or point-of-care settings. The design of convenient and HTP (high-throughput) technologies for the characterization of glycans and glycan–protein interactions is therefore a priority for researchers and clinicians alike .
Perhaps the most promising approach for rapid glycan analysis are array-based assays, comprising lectins or glycans, which have become very popular. However, there are some valid concerns about the value of the data generated from these arrays. Questions have been raised about the optimal protocols for conjugation and presentation of the lectins and glycans on solid surfaces. The source and quality of lectins themselves is also an issue for reliable and reproducible data generation and detection technologies which require extensive washing of labelled targets also cause concern [12,13].
To address these challenges, researchers have been making efforts to further optimize lectin structures, to discover lectin and glycan mimics which can be both flexible in presentation and consistent in performance, as well as to integrate lectin and glycan-recognition molecules on to label-free sensor platforms. Combinations of some of these novel technologies and sensor strategies with existing materials and methods, e.g. enzyme digestion or labelling techniques, will lead to more efficient and adaptable technologies for structural and functional characterization of the glycome. In particular, these approaches will facilitate more effective monitoring of glycan and lectin information in the context of recombinant protein production, disease diagnostic and prognostic applications.
Critical for the success of any glycoanalytical HTP platform or any glycosensors or glycobiosensors based on lectin–glycan binding is the consistent availability in large quantities of high-purity lectins or their mimics with optimal specificity and sensitivity towards the target molecule. The same is true of the availability of glycans for use as standards or tracers in these assays. As an alternative to natural lectins and purified glycans, re-engineered lectins and molecular and synthetic mimics of lectins and glycans are being developed and evaluated as affinity agents for the study of carbohydrate–protein interactions and to construct glycomic profiles for diagnostic, prognostic and engineering applications. An added advantage of these mimetic molecules is that they lack their own glycosylation, which has been reported to complicate interpretation of binding data with natural lectins .
The primary strategies being used for the identification of lectin and glycan mimics are screening of phage display libraries of lectin variants, carbohydrate-binding domains of glycoenzymes, antibody fragments and random peptide sequences, and the screening of random oligonucleotide libraries for aptamer discovery. Phage display technology is already widely used to identify proteins, peptides or scFvs (single-chain antibody variable region fragments) that bind to specific protein or hapten targets and is now being adapted for use with glycan targets. It provides an easy method to screen large combinatorial libraries of peptides/proteins. The protein/peptide variants are expressed on the surface of bacteriophage particles and the gene encoding the variant is contained within the phage. Library diversities of up to 1011 can be readily generated by a range of combinatorial and mutagenesis approaches, thus providing a large repertoire of potential binders or affinity agents. In vitro selection of specific proteins/peptides occurs by sequential enrichment of bound phage during cycles of biopanning and propagation. There are an increasing number of reports describing the use of phage display for the discovery of biomimetic peptides against carbohydrate and lectin targets .
Aptamer technology presents an alternative approach for lectin/glycan biomimetic molecule discovery. Aptamers are oligonucleotides which bind specifically to target molecules, including proteins and small molecules, with high affinity and specificity. Aptamers are usually selected by a method called SELEX (systemic evolution of ligands by exponential enrichment) technique [16,17]. In this technique, RNA or DNA aptamers are selected by multiple rounds of selection from a library of 1010–1013 random single-stranded oligonucleotide sequences. Since aptamers are nucleic acids, they can be easily synthesized and chemical modification for optimal stability is possible. Because of their ease of synthesis, controllable modification and the tremendous diversity possible, aptamers have been proposed as powerful tools for biomolecule detection, analysis and as modulators of biological interactions, comparable with antibodies. As with phage display technology, aptamer technology has been primarily used for protein-targeted affinity molecule discovery, but, in the last decade, both RNA and DNA aptamers have been described that bind to glycan moieties and to lectins (or carbohydrate-binding domains of proteins) [18–20].
Although the field of biosensors is significantly advanced for the detection of nucleic acids and proteins, it is at an early stage for detecting glycans and lectins. The most common approach used in glycobiosensors described to date is the detection of glycoconjugates based on a series of lectin/carbohydrate-recognition events on lectin arrays. Lectin arrays have effectively been used to profile glycoproteins [14,21–23], as well as the cell-surface glycans on bacteria  and mammalian cells . The profiling approach for glycan analysis enables the entire glycan structure to contribute to the overall signal output rather than a specific component of the complex glycan. The generally broad specificity and relatively low affinity of lectins and most carbohydrate binders described are also more suited to profiling rather than specific target detection . These arrays have traditionally been constructed with natural (non-recombinant) lectins from various sources, mainly from plants. Recently, Hsu et al.  have reported the use of recombinant lectins to overcome some of the reported difficulties encountered with respect to variations of purity and availability of the natural lectins. Additionally, random peptide arrays have been used for bacterial glycoprofiling . Glycan array technology has also been described for detection and characterization of lectin-like activities . Arrays have been fabricated using purified natural glycans, glycoproteins, synthetic glycans, neoglycoproteins and glycolipids using a variety of approaches  and have given information on the binding specificity of lectins, antibodies, bacterial adhesins, etc.
Fluorescence is the most common detection technology used to date in array-based glycobiosensors. Target analytes are labelled with a variety of fluorophores or quantum dots and binding is detected by scanning the array following incubation and washing steps. The cumbersome requirement to label the analyte directly can be avoided by using a sandwich format and secondary label in some, but not all, applications. Of more concern is the requirement of including washing steps to reduce non-specific binding of the fluorescent label, which can also disrupt low-affinity glycan–lectin interactions. An evanescent-field fluorescent detection approach for lectin arrays has been described, which eliminates the need for washing steps .
However, the most desirable glycobiosensors would be able to detect binding without the need to label analytes. Label-free detection techniques have been demonstrated in nucleic acid and protein biosensors and are now being applied to and optimized for the detection of glycans and lectins in proof-of-concept studies. The technologies examined include SPR (surface plasmon resonance) [30–33], electrochemical sensor methods such as EIS (electrochemical impedance spectroscopy) measurements [34–39], microcantilever deflection used in AFM (atomic force microscopy) , QCM (quartz crystal microbalance) measurements [41,42] and nanoparticle displacement methods [43,44]. These reports describe the detection of carbohydrates, ranging from monosaccharides to polysaccharides, glycoproteins and lectins. Most of these approaches have been demonstrated with purified material in standard buffers and need to be validated for sensitivity and reproducibility in natural complex or partially purified samples. The analysis of glycan-binding agents in more complex samples by using a commercially available SPR imaging device, the Biacore™ Flexchip instrument, has been described previously [27,45].
Glycans and glycoconjugates are universally distributed molecules and are involved in most physiological and pathological processes. Current glycomic methods are excellent for discovery and analysis in the sophisticated research environment, but there is a significant shortage of methods for convenient analysis of glycan–lectin interactions in the non-specialist laboratory and in point-of-care or on-site settings. There is therefore a definite need for new analytical tools that are amenable for glycobiological research and biosensor applications that are sensitive, rapid, simple, reliable and cost-effective. Advancing technologies for the discovery of natural and synthetic biomimics of lectins and glycans and modifying existing detection platforms to improve compatibility with the lectin–glycan mode of interaction, along with automation and miniaturization, will undoubtedly lead to significant developments in this area. The availability of more convenient tools for glycan analysis will increase our understanding of functional glycomics and present new opportunities for design of specific glycobiosensors tailored to their end applications.
Our work is supported by Science Foundation Ireland [Alimentary Glycoscience Research Cluster, grant number 08/SRC/B1393 and Stokes Professor for Glycosciences, grant number 07/SK/B1250], National University of Ireland Galway, Bristol-Myers Squibb, European Commission Marie-Curie Fellowship [grant number MTKD-CT-2004-013701], Irish Environmental Protection Agency [grant number 2008-EH-MS-2-S3] and Enterprise Ireland [grant number CR20070102].
Structural Glycobiology and Human Health: A Biochemical Society Focused Meeting held at Royal Holloway, University of London, 30–31 March 2010. Organized and Edited by Tony Corfield (Bristol, U.K.) and Barbara Mulloy (National Institute for Biological Standards and Control, U.K.).
Abbreviations: HTP, high-throughput; SPR, surface plasmon resonance
- © The Authors Journal compilation © 2010 Biochemical Society