Asparagine (N)-linked glycosylation occurs on 90% of membrane and secretory proteins and drives folding and trafficking along the secretory pathway. The N-glycan can be attached to an N-X-T/S-Y (X,Y ≠ P) consensus site by one of two oligosaccharyltransferase (OST) STT3 enzymatic isoforms either during protein translation (co-translational) or after protein translation has completed (post-translational). While co-translational N-glycosylation is both rapid and efficient, post-translational N-glycosylation occurs on a much slower time scale and, due to competition with protein degradation and forward trafficking, could be detrimental to the success of a peptide heavily reliant on post-translational N-glycosylation. In evidence, mutations in K⁺ channel subunits that shift N-glycosylation kinetics have been directly linked to cardiac arrhythmias. My thesis work focuses on identifying primary sequence factors that affect the rate of N-glycosylation.

To identify the molecular determinants that dictate whether a consensus site acquires its initial N-glycan during or after protein synthesis, I used short (~ 100-170 aa) type I transmembrane peptides from the KCNE family (E1-E5) of K⁺ channel regulatory subunits. The lifetime of these small membrane proteins in the ER translocon is short, which places a significant time constraint on the co-translational N-glycosylation machinery and increases the resolution between co- and post-translational events. Using rapid metabolic pulse-chase experiments described in Chapter II, I identified several molecular determinants among native consensus sites in the KCNE family that favor co-translational N-glycosylation: threonine containing-consensus sites (NXT), multiple N-terminal consensus sites, and long C-termini. The kinetics could also be shifted towards post-translational N-glycosylation by converting to a serine containing-consensus site (NXS), reducing the number of consensus sites in the peptide, and shortening the C-termini.

In Chapter III, I utilized an E2 scaffold peptide to examine the N-glycosylation kinetics of the middle X residue in an NXS consensus site. I found that large hydrophobic and negatively charged residues hinder co-translational N-glycosylation, while polar, small hydrophobic, and positively charged residues had the highest N-glycosylation efficiencies. Poorly N-glycosylated NXS consensus sites with large hydrophobic and negatively charged X residues had a significantly improved co-translational N-glycosylation efficiency upon conversion to NXT sites.

Also in Chapter III, I adapted a siRNA knockdown strategy to definitively identify the OST STT3 isoforms that perform co- and post-translational N-glycosylation for type I transmembrane substrates. I found that the STT3A isoform predominantly performs co-translational N-glycosylation while the STT3B isoform predominantly performs post-translational N-glycosylation, in agreement with the roles of these enzymatic subunits on topologically different substrates.

Taken together, these findings further the ability to predict the success of a consensus site by primary sequence alone and will be helpful for the identification and characterization of N-glycosylation deficiency diseases.