Sun et al.1 denatured and digested the mouse cell culture and captured the N-glycopeptides for chromatographic separation and MS/MS analysis on a linear-quadrupole ion-trap LTQ mass spectrometer (Thermo Scientific). The researchers described their method as “low cost shotgun glycopeptide capture” using ordinary MS in lieu of more sophisticated MS strategies and instrumentation. In this way, they deglycosylated N-linked peptides and used a consensus sequence (sequon) to determine the site of N-glycosylation. The drawbacks inherent in low-resolution, high-throughput MS, such as an indistinguishable mass shift, were compensated for with the high selectivity of the enrichment model used. Indeed, the enriched sample yielded a 90% abundance of sequon-containing peptides with a confidence probability of >0.9 compared with a less than 1% abundance for unenriched sample. Approximately 5.6% of the identified peptides contained two or more sequons on a single peptide, resulting in difficulty in determining the precise site of glycosylation. Overall, 468 IPI glycoproteins were identified, including 405 unique glycoproteins that were then translated to gene IDs using PIPE software.
The researchers cataloged the identified N-glycoproteins with several databases and found that 42% of the proteins were receptors, making this the most diverse protein class in the study. Other notable classes were transporters (17%) and enzymes (15%). However, on a quantitative level, transporters were identified to be the most abundant class with 19 spectra compared with receptors (10 spectra) and enzymes (11 spectra). Aside from transmembrane proteins, which accounted for 60% of the found proteins, other membrane-associated proteins that are generally not predictable by transmembrane analysis were identified. These included GPI-anchored membrane proteins, extracellular matrix components, secreted molecules, growth factors, and cytokines.
When the identified N-glycoproteins were compared to cell surface proteomes, researchers found a greater similarity among mES cells than between ES cells and RBCs. They cite that approximately 30% of E14.Tg2a glycoproteins are shared with the D3 surface proteome, while only 5% are shared with the RBC membrane proteome. The overlap between the N-glycoproteins identified in the study and the N-glycoproteins found in the surface proteome approaches 79%, including a 78% overlap in N-glycosylation sites. The researchers hypothesize that the discrepancies found between the two mES cell subtypes may indicate biomarker potential. They also assert that their identified glycoproteins include components of membrane-bound organelles, such as lysosomes, vacuoles, and endoplasmic reticulum, whereas the comparative surface glycoproteins list these cytosolic proteins as contaminants.
The study results were also compared to genomic and whole-cell proteomic results, with the finding that proteins that were identified as unique to this study are generally less abundant in genomic and whole-cell profiling than are those proteins that are commonly identified. Some examples include protein kinases, tyrosine kinases, peptidases, growth factors, and cell adhesion molecules. The researchers hypothesize that a relative dearth of transcripts compared with translated proteins may be a result of endosomal membrane-protein recycling.
Sun et al.1 found 1,182 glycosites in the 468 glycoproteins identified in the N-glycoproteome of mES E14.Tg2a cells. While the average expression was 10 spectra, a full 71% of the proteins demonstrated below-average spectra with 44% of the identified proteins demonstrating very low abundance with less than four spectra. When the glyco-proteome was grouped based on protein glycosylation for comparison purposes, the researchers found that 43% of the glycoproteins were monoglycosylated, while heavily glycosylated proteins were more highly expressed than less glycosylated proteins. This finding piqued researcher interest in this small number of high-abundance, highly glycosylated membrane proteins and led to the assertion that these glycoproteins are functionally important for stem cells. Examples of these proteins include Lrp2, Lrp1, Lama1, Lama5, and Tmem2. The researchers also note that increased glycosylation appears to accompany a decrease in transmembrane domains.
This study represents a large-scale profiling of the glycoproteins of mouse embryonic cells using standard yet sensitive instrumentation that increases the overall conceptual understanding of membrane proteins. In particular, the findings related to the labeling and identification of cytosolic proteins in the N-glycoproteome and the potential for greater understanding of plasma membrane recycling may prove useful for future research. In terms of intriguing findings, the researchers put forward the discovery of the purported evolutionary preference for N-glycosylation over transmembrane domains in the functional roles of membrane-associated proteins.
1. Sun, B., et al. (2013) ‘N-glycoproteome of E14.Tg2a mouse embryonic stem cells‘, PLoS One, 8 (2), (p. e55722)