Novel data analysis tool for semiquantitative LC-MS-MS2 profiling of N-glycans

Despite recent technical advances in glycan analysis, the rapidly growing field of glycomics still lacks methods that are high throughput and robust, and yet allow detailed and reliable identification of different glycans. LC-MS-MS(2) methods have a large potential for glycan analysis as they enable separation and identification of different glycans, including structural isomers. The major drawback is the complexity of the data with different charge states and adduct combinations. In practice, manual data analysis, still largely used for MALDI-TOF data, is no more achievable for LC-MS-MS(2) data. To solve the problem, we developed a glycan analysis software GlycanID for the analysis of LC-MS-MS(2) data to identify and profile glycan compositions in combination with existing proteomic software. IgG was used as an example of an individual glycoprotein and extracted cell surface proteins of human fibroblasts as a more complex sample to demonstrate the power of the novel data analysis approach. N-glycans were isolated from the samples and analyzed as permethylated sugar alditols by LC-MS-MS(2), permitting semiquantitative glycan profiling. The data analysis consisted of five steps: 1) extraction of LC-MS features and MS(2) spectra, 2) mapping potential glycans based on feature distribution, 3) matching the feature masses with a glycan composition database and de novo generated compositions, 4) scoring MS(2) spectra with theoretical glycan fragments, and 5) composing the glycan profile for the identified glycan compositions. The resulting N-glycan profile of IgG revealed 28 glycan compositions and was in good correlation with the published IgG profile. More than 50 glycan compositions were reliably identified from the cell surface N-glycan profile of human fibroblasts. Use of the GlycanID software made relatively rapid analysis of complex glycan LC-MS-MS(2) data feasible. The results demonstrate that the complexity of glycan LC-MS-MS(2) data can be used as an asset to increase the reliability of the identifications.