Mutually exclusive locales for N-linked glycans and disorder in glycoproteins

Abstract

Several post-translational modifications of proteins lie within regions of disorder, stretches of amino acid residues that exhibit a dynamic tertiary structure and resist crystallization. Such localization has been proposed to expand the binding versatility of the disordered regions, and hence, the repertoire of interacting partners for the proteins. However, investigating a dataset of 500 human N-linked glycoproteins, we observed that the sites of N-linked glycosylations, or N-glycosites, lay predominantly within the regions of predicted order rather than their unstructured counterparts. This mutual exclusivity between disordered stretches and N-glycosites could not be reconciled merely through asymmetry in distribution of asparagines, serines or threonines residues, which comprise the minimum-required signature for conjugation by N-linked glycans, but rather by a contextual enrichment of these residues next to each other within the ordered portions. In fact, N-glycosite neighborhoods and disordered stretches showed distinct sets of enriched residues suggesting their individualized roles in protein phenotype. N-glycosite neighborhood residues also showed higher phylogenetic conservation than disordered stretches within amniote orthologs of glycoproteins. However, a universal search for residue-combinations that are putatively domain-constitutive ranked the disordered regions higher than the N-glycosite neighborhoods. We propose that amino acid residue-combinations bias the permissivity for N-glycoconjugation within ordered regions, so as to balance the tradeoff between the evolution of protein stability, and function, contributed by the N-linked glycans and disordered regions respectively.