Identification of protein O-glycosylation site and corresponding glycans using liquid chromatography-tandem mass spectrometry via mapping accurate mass and retention time shift

We reported an improved combinatorial approach for identifying site-specific O-glycosylation using both glycan cleaved and non-cleaved methods. In this approach, a non-reducing β-elimination kit coupled with non-specific enzymes performed efficient digestion, O-glycan cleavage, and partial dephosphorylation without significant side reactions, thus enabling an automatic database search for the cleaved O-glycosylation or serine/threonine (S/T) phosphorylation sites. From the same sample concurrently prepared without β-elimination, the corresponding intact O-glycopeptides were mapped by accurate precursor ion mass using an in-house glycan database majorly composed of GalNAc (mucin-type) core and the retention-time shift (δR_t). Each glycopeptide assignment was verified by the detection of glycan-specific fragments using collision-induced dissociation (CID) to estimate False Discovery Rate (FDR). Using fetuin as a model, all identified S/T elimination sites were matched to multiple intact glycopeptides with a 31% FDR. This considerably reduced to 0% FDR by δR_t filtering. This approach was then applied to a protein mixture composed of therapeutic Factor IX and Enbrel^® mixed with fetuin and kappa-casein. A total of 26 glycosylation sites each of which corresponds to 1-4 glycans were positively mapped and confirmed. The FDR decreased from 33% to 3.3% by δR_t filtering and exclusion of repeated peptide tags that covered the same glycosylation sites. Moreover, the phosphorylation and O-glycosylation on the same site such as T159 of Factor IX and T170 of kappa-casein were able to be unambiguously differentiated. Thus, our approach is useful for in-depth characterization of site-specific O-glycosylation of a simple mixture such as protein-based therapeutics.