Glycoproteomics analysis of human liver tissue by combination of multiple enzyme digestion and hydrazide chemistry.

Rui Chen, Xinning Jiang, Deguang Sun, Guanghui Han, Fangjun Wang, Mingliang Ye, Liming Wang, Hanfa Zou

Index: J. Proteome Res. 8(2) , 651-61, (2009)

Full Text: HTML

Abstract

The study of protein glycosylation has lagged far behind the progress of current proteomics because of the enormous complexity, wide dynamic range distribution and low stoichiometric modification of glycoprotein. Solid phase extraction of tryptic N-glycopeptides by hydrazide chemistry is becoming a popular protocol for the analysis of N-glycoproteome. However, in silico digestion of proteins in human proteome database by trypsin indicates that a significant percentage of tryptic N-glycopeptides is not in the preferred detection mass range of shotgun proteomics approach, that is, from 800 to 3500 Da. And the quite big size of glycan groups may block trypsin to access the K, R residues near N-glycosites for digestion, which will result in generation of big glycopeptides. Thus many N-glycosites could not be localized if only trypsin was used to digest proteins. Herein, we describe a comprehensive way to analyze the N-glycoproteome of human liver tissue by combination of hydrazide chemistry method and multiple enzyme digestion. The lysate of human liver tissue was digested with three proteases, that is, trypsin, pepsin and thermolysin, with different specificities, separately. Use of trypsin alone resulted in identification of 622 N-glycosites, while using pepsin and thermolysin resulted in identification of 317 additional N-glycosites. Among the 317 additional N-glycosites, 98 (30.9%) could not be identified by trypsin in theory because the corresponding in silico tryptic peptides are either too small or too big to detect in mass spectrometer. This study clearly demonstrated that the coverage of N-glycosites could be significantly increased due to the adoption of multiple enzyme digestion. A total number of 939 N-glycosites were identified confidently, covering 523 noredundant glycoproteins from human liver tissue, which leads to the establishment of the largest data set of glycoproteome from human liver up to now.