Our universal glycomics approach characterises and quantifies glycan structures, from any source, without labels or derivatisation. To separate these glycan mixtures, we use porous graphitised carbon (PGC) as a chromatographic stationary phase, to resolve structurally similar glycan isomers. Previous inter-lab studies have identified that glycomics can benefit from a reduction in technical variability in glycan LC-MS analysis, extending the value of glycomics for probing biology[1].
Following separation, detected glycan structures have associated retention times, however these values are specific to LC conditions and can vary particularly when using PGC due to the column’s capacity for redox reactions[2]. As a result, although order of elution is reproducible, actual retention time values are not widely comparable. This challenge has been addressed in fields such as proteomics through the addition of internal retention time standards but standards are not widely available in glycomics due to the structural complexity of carbohydrate chemistry.
A dextran oligosaccharide ladder, similar to that used in reversed phase chromatography of fluorescently labelled glycans, was produced by acid hydrolysis and was characterised by LC-PGC-MS/MS. Eight subunits from 3-10 glucose units (GU) were separated and an equation was fitted to the elution pattern (R2 >0.999). Analysing N-glycans released from eight purified glycoproteins resulted in 212 unique glycan structures being characterised and assigned GU values. These values were found to be more specific for each structure than their retention time value and successfully discriminated closely eluting isomers.
To assess the technical variation of glycan analysis using PGC, a pre-defined mixture of released glycans was analysed including a simulated reduction in retention. Following alignment with the dextran ladder, retention time variation was significantly reduced. The ladder was also used to normalise instrument variation for glycan quantitation, resulting in an average peak area variation (CV) reduction of 30%.