Comparison of the performance of secretome analysis based on metabolic labeling by three unnatural sugars

Many secreted proteins, including cytokines, growth factors and hormones, are crucial in processes like intercellular signaling. Dynamic changes in secreted proteins usually reflect the growth and pathological state of the cells. Many drug targets are secretory proteins. The proteins are also important biomarkers. Conditioned cell culture media are important samples for secretory proteomic studies. Biomass spectrometry-based proteomic analysis enables the systematic study of secretory proteins. The main problem in analyzing secretory proteins in conditioned culture media is the low concentration of these proteins and the presence of serum, amino acids, and additives in culture media that may interfere with the protein analysis. Conventional secretory proteome analysis uses serum-free cell culture to reduce sample complexity, and typically involves protein concentration, purification, and desalting using ultrafiltration, dialysis, lyophilization, and trichloroacetic acid (TCA) or acetone precipitation, followed by enzymatic digestion and mass spectrometry analysis. This analytical process does not allow specific enrichment of secreted proteins. Thus, only a few secreted proteins can be identified. In addition, prolonged serum-free incubation of cells also tends to lead to unexpected changes in their activity status. A bioorthogonal-based enrichment approach can effectively avoid this problem. In recent years, unnatural sugars containing bio-orthologous groups, such as azide groups, have been used to metabolically label glycosylated proteins, enabling cellular imaging or selective enrichment of glycoproteins and their use for proteomic analysis. The strategy is a two-step process. First, azide-based sugar analogues are added to the cell culture medium and introduced to glycoproteins via the intracellular glycan biosynthesis pathway. Second, they are specifically covalently labeled with imaging probes or affinity probes via click chemistry. Since secreted proteins are usually glycoproteins, this glycolytic labeling has been used to label and enrich secreted proteins. N-Azidoacetylgalactosamine (GalNAz), N-azidoacetylglucosamine (GlcNAz), and N-azidoacetylmannosamine (ManNAz) are classical azide-based sugar analogues. Their effects on cytoplasmic membrane proteins have been compared. However, only ManNAz has been used for metabolic labeling of secreted proteins. No other glyco-analogues that label secreted proteins have been reported. Here, the bio-orthogonal chemical biology technology achieved highly selective labeling and enriched secreted proteins. In combination with click chemistry, different sugar analogues were evaluated for metabolic labeling of secreted proteins. HeLa cells were metabolically labeled by ManNAz, GalNAz, and GlcNAz (the three most commonly used commercial sugar analogues). These glycolytic markers can selectively label specific types of glycosylation. For example, ManNAz, an analogue of the biosynthetic precursor of sialic acid, N-acetylmannosamine (ManNAc), can label sialylated N- or O-glycoproteins. GalNAz, an analogue of N-acetylgalactosamine (GalNAc), can replace GalNAc as a core residue of mucin-type O-glycans and thus label O-glycoproteins. In addition, the intracellular metabolic intermediate of GalNAz (pyrophosphate) UDP-GalNAz can be interconverted with UDP-GlcNAz catalyzed by UDP-galactose-4-differential isomerase (GALE) and thus can also label N-glycoproteins and O-GlcNAc glycoproteins instead of GlcNAc. The GlcNAz analogue is commonly used to label nuclear and cytoplasmic glycoproteins with β-O-GlcNAc residues, but can also label N-glycoproteins with mucin-type O-glycoproteins by converting GALE to GalNAz, followed by enrichment using a biotin-alkynyl probe. Label-free quantitative proteomic analysis was performed to evaluate their labeling efficiency. ManNAz-based secretory protein labeling identified 282 secretory proteins, 224 plasma membrane proteins, and 846 N-glycosites. Compared with GalNAz and GlcNAz, the enrichment of secreted proteins was increased 130% and 67.2%, respectively, and the enrichment of plasma membrane proteins was increased 273.3% and 148.7%, respectively. This study provides a useful comparative analysis and new strategies for highly selective enrichment and systematic secretome analysis.