The glycome—the dizzying array of carbohydrate structures found in the cell—is still very poorly understood. One of the fundamental questions still to be addressed in the glycomics field is whether the glycan sugar chains that are post-translationally attached to proteins, called proteoglycans, have a regular, deterministic sequence. This question has interested Robert Linhardt of Rensselaer Polytechnic Institute for three decades.

“Our working hypothesis has always been that nature invests so much energy in biosynthesizing these very complex molecules, that they must have some defined structure,” says Linhardt. However, experimentally testing this hypothesis has proved extremely difficult. Unlike nucleic-acid or protein synthesis, glycan synthesis is not template-driven; despite much work put into understanding how glycans are biosynthesized in the Golgi, a template-driven process has never been found. Moreover, some glycans, such as starch, serve mainly as energy storage molecules and may not require a specific sequence to carry out their biological function. Proteoglycans, however, are mainly found on the cell surface and are thought to be responsible for helping to assemble signaling complexes; such an important function could suggest the need for proteoglycan sequence, Linhardt notes.

The glycomics field has also been challenged by the lack of available methods for sequencing glycans. The first hurdle is just obtaining enough material for mass spectrometry–based sequencing. Linhardt's collaborator, Toshihiko Toida of Chiba University in Japan, was instrumental in helping the team to obtain sufficient quantities of the simplest known proteoglycan, bikunin, which is used as a drug for treating acute pancreatitis in Japan.

Second, proteoglycans consist of one or more linear, sulfated glycosaminoglycan (GAG) chains linked to a core protein, and most core proteins contain multiple different GAG chains of different lengths, presenting a potentially major separation headache. Whereas bikunin contains only a single GAG chain, still, this GAG chain is heterogeneous and was challenging to purify. Linhardt's team first used proteolysis to generate a simple mixture of peptidoglycosaminoglycan chains and then applied a very efficient separation method, continuous elution preparative polyacrylamide gel electrophoresis, to fractionate the chains into mixtures of similar sizes and charges to simplify the analysis.

The third component to analyzing whether bikunin contained sequence required the expertise of collaborator Jonathan Amster of the University of Georgia. At Amster's mass spectrometry facility, the team analyzed the peptidoglycosaminoglycan fractions using two different instruments: the extremely sophisticated Fourier transform ion cyclotron resonance instrument and the more common Orbitrap instrument. Still, the analysis was not straightforward; one challenge the team had to overcome is that not all glycan bonds are equally sensitive to fragmentation by collision-induced dissociation. The sulfate bonds tend to break first, resulting in the loss of information before the glycan backbone can be sequenced. The team came up with a breakthrough trick to keep the charge state of the glycans higher than the number of sulfate groups, which prevented loss of the sulfate group and allowed them to obtain sequence information, using both mass spectrometry platforms.

In the end, this overall strategy led to the finding that bikunin indeed contains a regular sequence. In particular, the researchers identified two well-ordered domains in the GAG sequence: a 12-residue sulfated domain near the reducing end and a 6–22-residue nonsulfated domain at the nonreducing end.

Still, before a sweeping conclusion of whether all proteoglycans have sequence can be made, much more work remains to be done. “I would say that when we started this work, 20 years ago, our hypothesis that GAG chains had sequence was deeply contested,” says Linhardt. “The hypothesis is now stronger, but to generalize it to other proteoglycans would help prove it.” But as all other proteoglycans have more complex structures than bikunin, additional advances both in separation and in mass spectrometry technologies will likely be necessary to make glycan sequencing routine.