Extended Data Fig. 1: Optimization of N-glycoproteomic method using the synthesized long chain biotin-hydrazide probe.

a, Chemical structure of in-house synthesized biotin-hydrazide for glycoprotein labelling and enrichment. b, Labelling and enrichment efficiency of the biotin-hydrazide probe. Mouse pancreas proteins were extracted and labelled with 2 mM probe for 1 h (lane 2). Lysate without biotin-hydrazide labelling was subjected for streptavidin pull down as control (lane 3). Labelled proteins were pulled down by streptavidin beads (lane 4). Supernatant was collected from the streptavidin pull down (lane 5). c, Comparison of biotin-hydrazide probe-based method with conventional hydrazide bead-based method. Either biotin-hydrazide probe or hydrazide beads were incubated with 500 μg lysate for different times. After on-bead trypsin digestion to remove non-glycopeptides, glycopeptides were released from the streptavidin beads or hydrazide beads by PNGase F for LC–MS/MS analysis. Data are mean of two technical replicates. d,e, Subcellular ___location of proteins identified by glycopeptides (d) and sequence coverage comparison of 1,313 S–PM proteins commonly identified by glycopeptides and non-glycopeptides (e) using 6 pairs of tumour and NT samples. f, The comparison of analysis by non-glycopeptides and glycopeptides in terms of identification and quantification performance of S–PM proteins. The left panel is the number of S–PM proteins identified by glycopeptides and non-glycopeptides. Data are mean ± s.d. of 6 pairs of tumour and NT samples. The middle panel is the sum peptide intensities of all identified S–PM proteins from these 12 tissue samples. The right panel is the CV of 1,313 overlapped S–PM proteins quantified with LFQ intensities by glycopeptides and non-glycopeptides. g, The percentage of secreted (S), plasma membrane (PM), and other transmembrane (TM) proteins, non-S/PM/TM proteins, and nonspecifically labelled proteins reported by Matthews et al.⁷⁰ to total proteins in terms of protein number (left panel) and sum LFQ intensity (right panel) as identified by glycopeptides and non-glycopeptides. h, The percentage of proteins nonspecifically absorbed on streptavidin beads to total proteins identified by glycoprotein enrichment. The experimental procedure for identification of nonspecifically absorbed proteins was the same as glycoprotein enrichment, except that no biotin-hydrazide probe was added. Proteins identified only in the streptavidin pull down sample without biotin-hydrazide probe or having 2-fold higher LFQ intensity as compared with glycoprotein enrichment were defined as nonspecifically absorbed proteins. i,j, Comparison of quantification performance of S–PM proteins by global proteomic method (Unenriched) and N-glycoproteomic method (Enriched) using 6 pairs of tumour and NT samples. S–PM proteins quantified by Unenriched and Enriched were compared in terms of protein number and intensities (i), and CV of 1,546 S–PM proteins commonly quantified by Unenriched and Enrich (j). k,l, Correlation of the average tumour/NT ratios of commonly quantified S–PM proteins between the two methods. Bio-replicates denote correlation of the 6 pairs of tumour and NT samples; correlations of glycoproteins identified with different number of N-glycosites are independently displayed (k); technical replicates denote correlation of 1 pair of tumour and NT samples for three technical replicates (l).