Cover Story  
  August 8, 2005
Volume 83, Number 32
pp. 41-50

Recent progress in arrays, functional analysis, and synthetic techniques is helping to make carbohydrates better understood and more useful


From a chemical standpoint, carbohydrates are ornery critters. They're complicated and hard to control, but recent advances show that researchers continue to make considerable progress in addressing the challenges that carbohydrates pose.

ARRAYER This robotic pin deposition unit is used at Scripps Research Institute to transfer carbohydrates onto glass slides to create carbohydrate microarrays.
Areas of investigation that have been especially active in recent years include the design and wider use of increasingly sophisticated carbohydrate arrays, collaborative studies of carbohydrate function across entire genomes of animals and microorganisms, and continued refinement of techniques for synthesizing carbohydrates and carbohydrate conjugates of growing complexity and utility.

"Over the past few years, progress in the field has been amazing," says chemistry professor Geert-Jan Boons of the University of Georgia's Complex Carbohydrate Research Center. "For example, chemists have made great strides in the chemical and enzymatic synthesis of complex oligosaccharides, and these compounds are now being used to make microarrays, which are important tools to study carbohydrate involvement in a large number of biological processes. Researchers are beginning to provide an understanding of the molecular details of carbohydrate-protein interactions. And we are starting to learn how this combined knowledge can be exploited in drug and vaccine design."

Carbohydrate microarrays, in which tens or hundreds of different sugars are bound noncovalently or bonded covalently in small spots on solid surfaces, have been under active development for a few years and are beginning to have broad use in research. The arrays are used to screen libraries of biological compounds, cell extracts, and other samples to assess the carbohydrate-binding properties of their constituents. Application areas include basic research, drug discovery, and diagnosis--as well as glycomics, the study of full complements of carbohydrates in cells, tissues, or organisms.

The arrays got an early start in the late 1980s, when glycosciences professor Ten Feizi of Imperial College London's Faculty of Medicine, Harrow, England, and coworkers extracted carbohydrates from natural sources, converted them to glycolipids, attached the glycolipids by hydrophobic adsorption to surfaces, and screened them with proteins.

It wasn't until 2002 that other groups came onto the scene. For example, a group led by biologist Denong Wang, currently at Stanford University, spotted pathogen-based polysaccharides and glycoconjugates onto glass chips by using high-precision robots developed earlier for DNA microarrays. Wang and coworkers then used the arrays to look at the carbohydrate-binding specificity of antibodies and other proteins. Also in 2002, Feizi and coworkers made microarrays of pure oligosaccharides, glycosaminoglycans, and oligosaccharide fractions from brain.

Chemistry professor Chi-Huey Wong of Scripps Research Institute, La Jolla, Calif., and coworkers constructed noncovalent arrays of oligosaccharides and used them for high-throughput analysis of sugar-protein interactions. Chemistry professor Milan Mrksich of the University of Chicago and coworkers developed arrays of monosaccharides by covalent attachment to gold surfaces. Injae Shin and Sungjin Park of Yonsei University, Seoul, South Korea, prepared arrays of mono- and disaccharides by covalent attachment to thiolated glass surfaces.

SUBSEQUENT DEVELOPMENTS in the field have focused on variations in array density and design. For example, the groups of Shin and of chemistry professor Peter H. Seeberger of the Swiss Federal Institute of Technology (ETH), Zurich, independently developed better ways of creating high-density microarrays by using high-precision robots to print carbohydrates on surfaces.

Seeberger and coworkers also prepared carbohydrate arrays on optical fibers and, in collaboration with Mrksich's group, on gold surfaces.

In 2003, a collaborative team at the Consortium for Functional Glycomics (CFG), which is based at Scripps, introduced arrays in which biotinylated polysaccharides were adsorbed to wells of streptavidin-coated microtiter plates. Later, CFG core unit director Ola Blixt, molecular biology professor and CFG Director James C. Paulson, Wong, and coworkers used robotic printing to create arrays in which natural and synthetic carbohydrates were covalently linked to activated glass slides through simple amino-reactive groups.

The CFG arrays now contain more than 275 structurally defined sugars, ranging in size from monosaccharides to decasaccharides, whereas most other glycoarrays have had only up to 50 or so such sugars on them. "Within a month it will be 300, and we're going to continuously expand the arrays," Paulson says.

CFG's robotically printed arrays are the most complete and most useful sequence-defined glycoarrays produced so far, Wong notes. For rigorous glycomics studies, however, arrays of at least 600 and ultimately up to several thousand well-defined sugars will be needed, he says. "The limiting factor will be the preparation and constant supply of thousands of saccharides."

For the past few years, glycoarrays have been applied to biomedical studies of increasing sophistication. For example, CFG arrays have been used to analyze the specificity of nearly 200 carbohydrate-binding proteins and to profile not only carbohydrate-binding biomolecules but also intact viruses. The newer CFG arrays are also "very well suited for analysis of body fluids, such as human serum, demonstrating that this system could be used in clinical and industrial applications, such as drug discovery and diagnostics," Blixt says.

Seeberger and coworkers--and, independently, Wong, Paulson, and coworkers--have used arrays of polysaccharides (or glycans, a term widely used in the field) to identify carbohydrate sequences from human immunodeficiency virus's gp120 surface glycoprotein for potential use as anti-HIV vaccines. Several of the resulting candidate vaccines are currently being tested in animal trials.

Furthermore, Seeberger's group has used carbohydrate microarrays to detect bacteria in blood and food--work with potential implications for detecting blood poisoning and food pathogens. They also have screened bacterial-resistance-causing proteins on aminoglycoside arrays as a route to discovery of novel aminoglycoside antibiotics.

Wang and coworkers, in collaboration with Jiahai Lu of the School of Public Health, Sun Yat-Sen University, Guangzhou, China, used carbohydrate arrays to study the immunogenicity of the severe acute respiratory syndrome (SARS) coronavirus. And professor of pharmacology Ronald L. Schnaar of Johns Hopkins School of Medicine and coworkers have probed carbohydrate arrays with fluorescence-labeled mammalian cells to test the function of carbohydrate-binding proteins in their natural milieu on cell surfaces, rather than as isolated species.

In collaborative studies with other groups, Feizi and coworkers used microarrays to assign ligands for three novel receptors of the innate immune system (SIGN-R1, SIGN-R3, and langerin) and assigned a carbohydrate recognition function for a key mannose 6-phosphate receptor domain, a function that could not be assigned initially by conventional biochemical methods. Currently, Feizi's group is using microarrays to identify carbohydrate ligands of other proteins of the innate immune system and to survey ligands of adhesins of bacterial and viral pathogens.

For the future, Seeberger sees greater use of carbohydrate arrays to screen blood. "Many markers in human blood bind to carbohydrates, and this is going to be a field of increasing importance," he says. "Currently, we are using arrays to screen human sera from malaria-endemic areas and have found that specific antibodies provide protection from severe disease. We basically can determine if people are going to get severe malaria or not-so-severe malaria" on the basis of the antibodies detected in the screens.

Glycominds, a biotechnology company in Lod, Israel, has commercialized IBDX, a glycoarray-based blood test for diagnosing Crohn's disease. This is the first in a series of carbohydrate-based blood test kits that Glycominds is developing for diagnosis of autoimmune and chronic inflammatory diseases, such as multiple sclerosis, antiphospholipid syndrome (a cause of thrombosis), and rheumatoid arthritis.

To a great extent, carbohydrates exert their influence in the body by interacting selectively with proteins, often on cell surfaces. Essentially, these interactions relate the glycome, the complete set of carbohydrates in an organism, to the proteome, its complement of proteins. In the past few years, the field of functional glycomics has developed to better understand these interactions.

Many research groups are currently working in functional glycomics. The projects are most effective if the resulting data can be collected centrally, and four major consortia have been organized to coordinate glycomics efforts.

Scripps-based CFG focuses on the effects of carbohydrate-protein interactions on cell communication. It's funded to the tune of about $8 million per year by a "glue grant" from the National Institute of General Medical Sciences. CFG focuses on glycan-binding proteins in mammals and in pathogens, the structures these proteins recognize, and the functional consequences of those recognition events.

In mammals such as humans and mice, "we're focusing on pure cell populations," Paulson says. "So people can find out what kind of sugar structures are predominant on the surface of a type of cell they're interested in."

Many glycan-binding proteins are involved in immune system recognition, "and this is a very hot area right now," Paulson adds. For example, DC-SIGN, a glycan-binding protein on dendritic cells that recognizes carbohydrates on HIV and mediates a critical step in HIV infection, is currently being investigated by CFG-associated groups.

CARB ANALYSIS These N-glycans from human neutrophils, designated by CFG's nomenclature system, were identified by using matrix-assisted laser desorption-ionization/time-of-flight mass spectrometry to determine their molecular weights. NeuAc is N-acetylneuraminic acid, Gal is galactose, Glc is glucose, Man is mannose, Fuc is fucose, GalNAc is N-acetylgalactosamine, and GlcNAc is N-acetylglucosamine.

Glycomics researchers have also been probing how glycan-binding proteins on the surface of one cell can recognize glycans on the same or another cell. For example, Paulson and coworkers recently found unexpectedly that the glycoprotein CD22 recognizes nearby CD22 molecules on the same cell surface.

Another group focusing on functional glycomics and carbohydrate arrays is the UK Glycochips Consortium, which started up about a year ago with an award of £3.5 million (about $6 million) from Research Councils UK. "We see our effort as being complementary to that of CFG," says chemistry professor Sabine L. Flitsch of the University of Manchester, who serves as project coordinator. At the moment, "we are funded to develop some basic array technology in the functional glycomics area."

This consortium's aim is to generate glycoarrays both from synthetic carbohydrates and from natural carbohydrates obtained from cells, tissues, and organs. These will include carbohydrates such as the glucosaminoglycans heparin and heparan sulfate, "which are very important cell matrix polysaccharides but are difficult to study," Flitsch says. Arrays are also important tools for defining the specificity of metabolic enzymes, particularly glycosyltransferases and glycosidases, she says.

A third glycomics consortium, the Zurich Glycomics Initiative (GlycoInit), based at ETH, started up a couple of years ago with funding from the institute and the Swiss government. Some of its major areas of investigation are synthetic carbohydrate chemistry (including automated synthesis); isolation, purification, and structural analysis of glycoconjugates; studies of fundamental carbohydrate biology and the role of carbohydrates in disease; and development of carbohydrate arrays and their use to identify carbohydrate-protein interactions and find inhibitors of such interactions, particularly in the context of infectious diseases. Seeberger and ETH microbiology professor Markus Aebi coordinate this consortium.

In Japan, meanwhile, biochemistry professor Naoyuki Taniguchi of Osaka University Graduate School of Medicine has helped organize the Japan Consortium for Glycobiology & Glycotechnology (JCGG). The two-year-old organization aims to help create a systems glycobiology center, in which glycobiology will join forces with nanotechnology, bioinformatics, and computation for carbohydrate research and to investigate medical applications.

"No official support for the consortium is provided by the Japanese government," Taniguchi says. Instead, scientists in the glycobiology and glycotechnology fields "took the initiative to create and support the consortium by using their research grants to facilitate exchange of scientific information, sharing of equipment and facilities, and construction of a database."

Consistent Set Of Symbols Proposed

A key achievement of the Consortium for Functional Glycomics (CFG) was its development last year of a new set of symbols to represent sugars. "Until now, everyone has tended to use different symbols," CFG Director James C. Paulson says.

CFG 's nomenclature consists of a series of colored geometric symbols, one for each type of monosaccharide, and characters that indicate the stereochemistry and connection points of glycosidic linkages between sugars in oligosaccharides. The symbols can still be interpreted correctly in black-and-white copies. In the figure, CFG's symbols are shown above International Union of Pure & Applied Chemistry (IUPAC) nomenclature for the same oligosaccharides.

CFG's system "has been fairly rapidly adopted" by many carbohydrate researchers, and IUPAC is currently evaluating the possibility of backing it officially, Paulson says.

In CFG's system, glycosidic linkages between sugars are designated by or ß--axial or equatorial orientation, respectively, of the glycosidic carbon of the first sugar--plus a number that indicates the position of the carbon in the second sugar that is receiving the glycosidic bond. The position of the anomeric carbon in the first sugar is set for each type of sugar and therefore does not appear in the CFG system, although it is indicated explicitly in the IUPAC system. NeuAc is N-acetylneuraminic acid, Gal is galactose, Glc is glucose, Man is mannose, Fuc is fucose, GalNAc is N-acetylgalactosamine, and GlcNAc is N-acetylglucosamine.

THE MASSIVE AMOUNTS of information generated by research in functional glycomics increasingly are being managed by various databases. At CFG, for example, the Glycan Binding Protein (GBP) database contains information on glycan-protein interactions and associated biological effects. And the Glycan database, a file of glycan structures, "is analogous to SwissProt for proteins or GenBank for genes," Paulson says.

Researchers browsing the Glycan database immediately can see "that different tissues produce different glycan structures, and this is the basis for differential recognition of cells by glycan-binding proteins," Paulson says. The information gives researchers who study these recognition events ideas for additional experiments, he explains. The database "has turned out to be an extraordinarily useful resource."

Other major carbohydrate databases include the Kyoto Encyclopedia of Genes & Genomes (KEGG) Glycan database at Kyoto University; GlycoSuite from Proteome Systems, North Ryde, New South Wales, Australia; and, a collection of linked carbohydrate databases at the German Cancer Research Center (DKFZ), Heidelberg.

An attempt to integrate these independent systems has been initiated by Claus-Wilhelm von der Lieth of DKFZ's spectroscopy and molecular modeling group, who helps maintain Recently, he formed the EuroCarbDB initiative, which hopes to develop standard digital glycan descriptions, software tools, and procedures that will allow easy integration of glycan databases willing to join the network. "By performing a single search, you would be able to search most of the information available, which is currently distributed over several local databases," von der Lieth says. CFG is cooperating with EuroCarbDB, and Paulson says he believes other carbohydrate database managers will, too.

One of the classic challenges in the field of carbohydrate chemistry is that carbohydrate-based compounds and conjugates are notoriously hard to synthesize. Progress has continued on strategies to simplify and automate oligosaccharide synthesis and to facilitate the construction of larger and more complex glycopeptides and glycoproteins.

A key milestone in oligosaccharide research was the development of automated synthesis techniques a few years ago by several groups--including those of Seeberger; Wong; chemistry professor Takashi Takahashi at Tokyo Institute of Technology; and professor of biological sciences Shin-ichiro Nishimura at Hokkaido University, Sapporo, Japan. "The past few years have seen a broadening of these approaches" as well as efforts to commercialize them and apply them to drug discovery, Seeberger says.

The goal of automated carbohydrate synthesis "has to be that the nonspecialist can use it," Seeberger says. No group has demonstrated that kind of nonspecialist user-friendliness yet. "The chemistry is still a little too complex, and it has not been shown that it can work for every carbohydrate structure."

Better carbohydrate building-block units are needed, yields still need to be increased considerably, and synthesizer reaction times need to be shortened, Seeberger says. Microreactors are among the tools he and his coworkers are using to screen carbohydrate-synthesis reactions for optimal reaction conditions (C&EN, Feb. 7, page 11). "Within the next two to three years, you'll see a complete solution across the board" in automated carbohydrate synthesis, he says. "It's getting pretty exciting."

In addition to advancing the construction of oligosaccharides, chemists are also making strides in the ability to assemble glycopeptides and glycoproteins. These are important synthetic targets, not only from a basic research standpoint but also because they play essential biological roles.

"From an industrial perspective, the market for therapeutic proteins is billions of dollars a year," Seeberger says, "and a lot of therapeutic proteins are glycosylated." Glycoproteins are difficult to isolate from natural sources in pure form, he says, "so there's a huge push to synthesize them."

INDUSTRIAL CARBS Researchers William J. Christ (left) and Obadiah J. Plante at Ancora Pharmaceuticals, Medford, Mass., prepare an intermediate for the production of an antimalarial carbohydrate vaccine.

SEVERAL POSSIBLE overall strategies can be used to synthesize glycoproteins. "The easiest method, and the one most drug companies use, is to make recombinant glycoproteins in mammalian or yeast cells and then remodel the carbohydrate part enzymatically, using glycosidases and glycosyl-transfer enzymes to create the sugar you want," Wong says. Nishimura and coworkers are currently investigating site-specific introduction of chemically defined carbohydrates into bacterially expressed recombinant proteins as a route to industrial production of engineered glycoprotein drugs.

A variation--offered by associate professor of biochemical engineering Tillman U. Gerngross of Dartmouth College and coworkers there and at GlycoFi Inc., Lebanon, N.H.--is genetic engineering of a humanlike glycosylation pathway into yeast to produce homogeneously glycosylated proteins similar to those needed for therapeutic use.

Bacteria have also been induced to express glycosylated proteins directly. Last year, a group led jointly by Peter G. Schultz, director of the Genomics Institute of the Novartis Research Foundation, La Jolla, and Wong showed that protein containing a glycosylated nonnatural amino acid at a defined position could be expressed in Escherichia coli in good yield. Such a carbohydrate moiety can be recognized by a carbohydrate-binding protein or subsequently modified to build more complex carbohydrates.

The Schultz-Wong approach requires the evolution of a tRNA and a corresponding aminoacyl-tRNA synthetase that do not interact with any natural tRNAs and aminoacyl-tRNA synthetases in the bacteria. The aim was "to evolve microorganisms to express human glycoproteins containing carbohydrates you want," Wong says.

"Of all of the new work I've seen on carbohydrate modifications to biologically interesting molecules, the Schultz-Wong study is the one I wish I had done," comments Harvard University professor of chemistry and chemical biology Daniel Kahne, whose research interests include glycopeptide antibiotics. "It is really nice--simple and elegant."

Glycopeptides can also be made by glycopeptide-coupling reactions. Wong and coworkers have developed enzymatic and other techniques to do this. One is intein-based coupling, a technique devised about five years ago. In this method, removal from glycopeptides of intervening domains in proteins, called inteins, causes adjacent glycopeptide fragments to link.

Glycopeptide ligation can also be carried out chemically. Chemistry professor Samuel J. Danishefsky and coworkers at Memorial Sloan-Kettering Cancer Center and Columbia University in New York City have recently been pushing the limits of chemistry-based ligation. They hope to make it possible to construct glycoproteins at will for use as potential vaccines and therapeutics.

Last year, postdoc J. David Warren, Danishefsky, and coworkers constructed two carbohydrate-containing peptides and combined them by chemical means to provide one of the largest multiply glycosylated peptides ever made synthetically in the laboratory. "It was a beautiful exercise in standing on the shoulders of others," Danishefsky says, noting that earlier work on oligosaccharide and glycopeptide synthesis by a number of groups set the stage. "The synthesis of complex, biologically important polydomain glycoproteins," he says, "is a realistic and attainable goal."


View a reaction scheme showing Danishefsky and coworkers' glycopeptide ligation.
  Chemical & Engineering News
ISSN 0009-2347
Copyright © 2005