Characterization of a gel-separated unknown glycoprotein by liquid chromatography/multistage tandem mass spectrometry: analysis of rat brain Thy-1 separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

We developed an efficient and convenient strategy for protein identification and glycosylation analysis of a small amount of unknown glycoprotein in a biological sample. The procedure involves isolation of proteins by electrophoresis and mass spectrometric peptide/glycopeptide mapping by LC/ion trap mass spectrometer. For the complete glycosylation analysis, proteins were extracted in intact form from the gel, and proteinase-digested glycoproteins were then subjected to LC/multistage tandem MS (MSn) incorporating a full mass scan, in-source collision-induced dissociation (CID), and data-dependent MSn. The glycopeptides were localized in the peptide/glycopeptide map by using oxonium ions such as HexNAc+ and NeuAc+, generated by in-source CID, and neutral loss by CID-MS/MS. We conducted the search analysis for the glycopeptide identification using search parameters containing a possible glycosylation at the Asn residue with N-acetylglucosamine (203 Da). We were able to identify the glycopeptides resulting from predictable digestion with proteinase. The glycopeptides caused by irregular cleavages were not identified by the database search analysis, but their elution positions were localized using oxonium ions produced by in-source CID, and neutral loss by the data-dependent MSn. Then, all glycopeptides could be identified based on the product ion spectra which were sorted from data-dependent CID-MSn spectra acquired around localized positions. Using this strategy, we successfully elucidated site-specific glycosylation of Thy-1, glycosylphosphatidylinositol (GPI)-anchored proteins glycosylated at Asn23, 74, and 98, and at Cys111. High-mannose-type, complex-type, and hybrid-type oligosaccharides were all found to be attached to Asn23, 74 and 98, and four GPI structures could be characterized. Our method is simple, rapid and useful for the characterization of unknown glycoproteins in a complex mixture of proteins.