Sensitive and Specific Global Cell Surface N-Glycoproteomics Shows Profound Differences Between Glycosylation Sites and Subcellular Components

Abstract

Cell surface glycans are essential for establishing cell communication, adhesion, and migration. However, it remains challenging to obtain cell surface-specific information about glycoconjugate structures. Acquiring this information is essential for unraveling the functional role of glycans and for exploiting them as clinical targets. To specifically analyze the N-glycoprotein forms expressed at the cell surface, we developed a C18 liquid chromatography (LC)-mass spectrometry (MS)-based glycoproteomics method in combination with highly specific cell surface protein labeling and enrichment using a biotin label. The surface-specificity of the method was validated by MS-based proteomics of subcellular component marker proteins. Using the human keratinocytes N/TERT-1 as a model system, we identified and quantified the glycosylation of hundreds of cell surface N-glycosylation sites. This approach allowed us to study the glycoforms present at the functional relevant cell surface, omitting immaturely glycosylated proteins present in the secretory pathway. Interestingly, the different stages of N-glycan processing at individual sites displayed at the cell surface were found to correlate with their accessibility for ER-residing processing enzymes, as investigated through molecular dynamics simulations. Using the new approach, we compared N-glycosylation sites of proteins expressed on the cell surface to their counterparts in a total cell lysate, showing profound differences in glycosylation between the subcellular components and indicating the relevance of the method for future studies in understanding contextual glycan functions.

Figure 1

Capturing the cell surface N-glycoproteome. (A) Schematic of the methods. (B) Depiction of cellular N-glycosylation types and features, including biosynthetic intermediates in the secretory pathway. (C, D) Proteomic validation of surface-specificity via secretory pathway and surface markers in N/TERT-1 cells. (C) Surface markers (yellow): Integrin α5 (ITA5) and E-cadherin (CADH1). Intracellular markers (yellow): GM130 (GOLGA2), mannosyl-oligosaccharide-1,2-alpha-mannosidaseIB (MAN1A2), and trans-golgi network glycoprotein 46 (TGN46). Cells were counterstained with the F-actin marker Phalloidin 594 to visualize the cell surface (pink) and DAPI (blue). (D) LC-MS/MS-based proteomics identified 6057 proteins in the total cell lysate (TCL) and 3541 proteins in the surface preparation. Marker proteins were relatively quantified based on the total protein abundance. (E) Comparison between the TCL and the surface glycome of N/TERT-1 cells using LC-MS/MS based released glycan analysis. Glycan traits were calculated based on 188 individual N-glycans (Table S2). *t test derived Benjamini–Hochberg FDR < 0.05.

Figure 2

Cell surface glycoproteomics. (A) Numeric description of the cell surface and total cell lysate (TCL) N-glycoproteomes. (B) Representative LC-MS/MS glycoproteomics data showing the microheterogeneity of 5′-nucleotidase N53, including the MS/MS data of two of the isomeric glycopeptides with the glycan composition Hex5HexNAc3Fuc1 indicating structural differences in the fucose location. (C) Cell surface- and glycosylation site-specific glycosylation of 5′-nucleotidase (5NTD), including a molecular model indicating the simulated accessibility of the individual glycosylation sites for ERManI. (D) Cell surface- and glycosylation site-specific glycosylation of integrin beta 1 (ITB1), integrin alpha 5 (ITAV), epidermal growth factor receptor (EGFR), and CD166 antigen (CD166).