I have written a few blogs over the past months where I have discussed the importance glycan analytics and how we can use a combination of chromatography, mass spectrometry and a knowledge of carbohydrate chemistry to study glycan structures, both N- and O-linked. This covers the main considerations for a glycan characterization study but one important question is still outstanding: where on the molecule is/are my glycans attached?
The ICH Q6B document which details the expectations for a comprehensive glycoprotein structural characterization, makes clear the importance of this question, stating in Appendix 6 that “In addition, the structure of the carbohydrate chains, the oligosaccharide pattern (antennary profile) and the glycosylation site(s) of the polypeptide chain is analyzed, to the extent possible.”
So, we must consider not only what the glycans are like structurally but where they are attached if we are to give a full evaluation of the glycosylation profile of the molecule. As you may well imagine, N-glycans and O-glycans being very different forms of glycosylation, also differ in their attachment points within a protein chain and we must use different procedures to identify their locations as a result.
N-glycans are attached to the side chain of the amino acid Asparagine, but only when that Asparagine residue (N) is present in the tripeptide consensus sequence N-X-S/T where X is any amino acid except Proline and S and T are Serine and Threonine respectively. The presence of this consensus sequence is no guarantee of N-glycosylation but if it happens then this is where it will be located. So how can we find this?
Well, if there is only one potential N-glycosylation site in the molecule then there is nowhere else it can be! However, for those molecules that have more than one consensus site we have some options at the protein analysis level we can use.
Firstly, peptide mapping can be used to identify glycosylation sites, both occupied and unoccupied. With the careful selection of proteolytic enzyme or digestion strategy, the glycopeptides (or predicted glycopeptides) can be identified individually in an LC-MS experiment. The same chromatography can then also be used as a fraction collection process to collect the separated glycopeptides for subsequent identification of what N-glycosylation populations are where, because it is highly likely that the distribution of N-glycans will not be the same at each glycosylation site.
So, the above helps us identify the sites of glycosylation, but we can actually go a step further and assess the degree of occupancy at the individual glycosylation sites. This is achieved using the enzyme PNGase F, which releases N-glycans from the protein backbone and in so doing converts the glycosylated Asparagine residue to Aspartic acid. This amino acid is 1Da heavier than Asparagine and so the modified peptide can readily be identified in a mass spectrometer during an on-line LC/MS peptide mapping analysis where PNGase F has been used to release N-glycans.
A comparison of the peak areas of the peptide with Asparagine vs Aspartic acid then allows, in a relative abundance sense, a determination of the degree of occupancy of the glycosylation site. This doesn’t take into account any differences in ionization efficiency between the two peptides however, but for comparative purposes this effect is the same for all samples and thus cancels out.
This is all well and good for N-glycans but can we take the same approach for O-glycans? Well, not quite unfortunately!
There is no broad specificity enzyme for O-glycan release, let alone one that modifies the site of O-glycan attachment to the protein as it releases the sugar. So we can’t rely on an enzyme based method to help us. However, the idea of the release mechanism being used to shape the subsequent glycan attachment site is still valid.
We can use the chemical technique of reductive elimination, a standard chemical method for non-selective O-glycan release, to label the site of O-glycan attachment through production of a dehydroamino acid. This dehydropeptide can then be identified by LC/MS and, using fragment ion assessment, the site of O-glycosylation can be found. This is particularly useful since there is no O-glycosylation consensus sequence at the primary amino acid level apart from the fact that O-glycans are frequently found in Serine/Threonine rich regions often close to Prolines, making identification of the precise site of attachment challenging.
It is also possible, through the use of specific types of fragmentation techniques available on some mass spectrometers, to identify the site(s) of O-glycan attachment directly through fragmentation of the O-glycopeptide. The mechanism of fragmentation in this case favors cleavage along the peptide backbone and preserves the glycan group at its site of attachment.
Thus, the chromatographic and mass spectrometric techniques available at BioPharmaSpec, coupled with appropriate protein and carbohydrate chemical preparative methods, can shine a spotlight on the site(s) of glycan attachment.
Contact us to discuss how our scientists can use their expert knowledge of glycosylation pathways to generate glycan data for your molecule as part of a structural characterization or analytical comparability package.
