by Jeffrey M. Perkel
When pathogens first enter the human body, just what exactly do they see? According to Nicolle Packer, professor of chemistry and biomolecular sciences at Macquarie University in Sydney, Australia, the answer is "a carpet of sugars."
These sugars—whether attached to proteins on epithelial cell surfaces, or in the mucus layer that coats these cells—provide docking sites for pathogens to literally gain a foothold in the body, and Packer has dedicated her lab to understanding these interactions, the better to possibly thwart them.
Researchers who study sugar chains and sugar-modified proteins and lipids are called glycobiologists, and for a long time, theirs was a relatively lonely discipline. That's because most biologists sidestepped these molecules (or at least, their modifications) if at all possible, dismissing glycan analysis as "too difficult," says Packer.
To a certain extent, that belief was well founded. Unlike proteins, which are linear chains of substantially different amino acids, glycan chains can branch, connect via a variety of linkages, and involve sugar building blocks that are all more or less identical in terms of mass and chemical properties. Further, unlike the genetic code, no known template dictates the structure of glycan chains, and few software tools exist to make sense of the mass spec data they generate in any event.
"It's very much where proteomics, I believe, was 10, 12 years ago," says Packer.
Yet, with the growing recognition of glycobiology's impact on development, immune function, cell-cell communication, and disease, the field's ranks are growing. The glycobiologist's tool of choice: mass spectrometry.
Packer uses a Bruker Daltonics HCT Ultra Ion Trap for her analysis of glycoproteins and oligosaccharides. In a typical analysis, she takes a biological sample, such as human milk, cleaves sugars off the proteins with enzymes such as PNGase F or via a chemical reaction called beta-elimination, fractionates the glycan chains on a graphitized carbon column, and then analyzes the chains via tandem mass spectrometry.
Graphitized carbon, she says, is to glycans what reverse-phase chromatography is to peptides. "It can actually separate by minutes sugars which have got the same mass but different structure."
Her choice of the HCT Ultra, meanwhile, stems both from its ability to perform multiple rounds of tandem MS analysis (that is, to fragment a glycan over and over again until its structure has been determined) and also because it supports ETD (electron transfer dissociation) fragmentation. ETD, she explains, enables her lab to determine the amino acid sequence of glycopeptides without losing the attached glycan chains, something that is not possible using standard collision-induced dissociation methods.
Carlito Lebrilla, chair of the Department of Chemistry at the University of California, Davis, favors high mass accuracy and high resolution instruments for his research into the sugar content of serum and human milk.
High mass accuracy instruments, he says, enable his team to identify oligosaccharides and glycopeptides, and to assess sample purity. "These can all be done with other instruments, but the high mass accuracy gives you more confidence that you are seeing what you think you are seeing," he says.
Lebrilla's instrument collection includes two Varian FT-ICR mass spectrometers and an Agilent HPLC-Chip/TOF system, both of which he uses to ferret out potential biomarkers of such diseases as prostate, breast, and ovarian cancer. The work, he says, has multiple levels of complexity.
The simplest level involves releasing all glycans from a solution and profiling them, to identify those that change consistently during disease, for instance. The next level involves associating given sugars to specific proteins. And the final layer of complexity: tying each oligosaccharide to a specific site on the protein. That, he says, "is very complicated," because each possible glycosylation site is heterogeneous. In other words, any given site can contain from zero to perhaps dozens of different modifications.
To attack such a problem, Lebrilla and his team purify glycopeptides from biological fluids, digest them with proteinases that break down everything except those peptides protected by a glycan chain, and then analyze the resulting pieces by mass spec, using the high mass accuracy of the instrument to identify both the peptide sequence and the glycan structure. "If you have very high mass accuracy, you can actually assign the glycans with a specific peptide," he says.
Martin Larsen, research associate professor at the University of Southern Denmark in Odense, also favors high mass accuracy instruments for his research into post-translational modifications. Among Larsen's instruments are two Thermo Orbitraps, a Waters QToF Premier, and a trio of tandem MALDI-TOFs from Waters, Bruker Daltonics, and Applied Biosystems. He also has an ETD-enabled LTQ from Thermo on loan.
One of Larsen's projects involves identifying glycosylation differences between normal and cancerous tissues, especially sialylated glycosylation. Because sialylated glycopeptides represent such a small fraction of peptides overall, Larsen enriches for these via HILIC (hydrophilic interaction chromatography), or normal-phase chromatography, and titanium dioxide chromatography.
HILIC appears to be the best general method (as opposed to more specific approaches involving antibodies or lectins) for separating glycopeptides from unmodified peptides, says Ron Orlando, director of the Mass Spectrometry Facility at the University of Georgia Complex Carbohydrate Research Center in Athens. "The presumed procedure by which HILIC works is you get a sphere of hydration around the particles, and you then get some partitioning of hydrophilic things into that hydration sphere," explains Orlando. "So the glycans should fit very nicely into the hydrated sphere around the stationary phase, and the peptides then go through, and they do."
Titanium dioxide, in contrast, captures highly negatively charged molecules, Larsen says. By combining it with HILIC, his lab can obtain highly enriched preparations of sialylated glycopeptides, excluding neutral and non-sialylated glycopeptides. "This is the power of TiO2 chromatography, which bind phosphopeptides and sialylated glycopeptides very efficiently," he says.
Orlando's lab uses a variety of high mass accuracy instruments, including a Thermo LTQ ion trap, a Thermo LTQ-FT, and two Waters QToFs, to cut down on false positives in his glycobiology research. He also uses the hardware to drive some interesting and sorely needed methods development for glycobiologists.
For instance, Orlando's team has come up with the glycan equivalent of such proteomics tools as SILAC and Applied Biosystems' iTRAQ and ICAT reagents.
SILAC is a metabolic labeling approach for relative quantification experiments that labels one protein sample in vivo with isotopically "heavy" amino acids, providing a mass shift relative to protein grown with non-isotopically labeled amino acids. In Orlando's variant, called IDAWG (for the University of Georgia Bulldogs, and to contrast with ICAT), cells are provided with N15-glutamine to label glycan chains.
Glutamine is the sole nitrogen source for all amino sugars, including those in glycan chains. So by providing N15-glutamine, all glycoprotein glycans will be isotopically labeled. Best of all, Orlando says, most growth media comes glutamine-depleted, meaning no special media is required to do this work.
Orlando's iTRAQ analog is a pair of isobaric tags (C13-methyliodide and 1-deutero-C12-methyliodide) that can tag glycan structures in vitro for relative quantitation.
Perhaps one of the biggest deficiencies in glycobiology has been the lack of good analytical tools to analyze sugar spectral data. Protein biologists have a wealth of software available that can identify and sequence peptides based on mass spectra. But glycobiologists have largely been out of luck. Larsen does all his analyses manually, which necessarily limits the scope of his research.
But that could soon change. Before joining the faculty of Macquerie University, Packer cofounded (and worked for) the Australian firm Proteome Systems, where she helped develop a proprietary glycan database called GlycoSuite. Now Proteome Systems is releasing GlycoSuite into the public domain, Packer says; it should be available in early 2009 as part of the Swiss Institute of Bioinformatics ExPASy server toolset.
The result, she says, could be the same sort of growth spurt that characterized proteomics a decade or so ago.
"Proteomics only really took off when the software was there to analyze the mass spec data by everybody, rather than just the experts," she says. "Similarly, glycomics will take off when the MS/MS spectra can be analyzed easily. And to do that, you need to have the tools and the database to match with."
Programmers, she explains, can use GlycoSuite to develop the glycan equivalents of SEQUEST and MASCOT, thereby enabling researchers to match spectral data against "a theoretical breakdown of the known glycan structures." And that is important, she says, because glycan modifications "are obviously important, but they've largely been overlooked because people perceive it as hard. I think you'll see in the next few years that that perception will change."