N- and O-linked glycosylation of proteins
Glycosylation of viral envelope glycoproteins depends on the host glycosylation machinery. Viral glycoproteins, like cellular proteins, possess signal peptides directing them to the secretory pathway. Proteins targeted to the secretory pathway can be decorated with different kinds of post-translational modifications en route to the cell surface or another cellular compartment. Glycosylation is one of the most widespread and versatile modifications of proteins and is classified as N-, O- and C-linked, named after the acceptor amino acid atom to which the sugar moiety is attached. While there is only one type of N-linked glycosylation in terms of initiation, many types of O-glycosylation exist, classified based on the identity of initiating monosaccharides (Stanley 2011; Moremen et al. 2012). Most types of O-glycosylation are carried out in the secretory pathway, but cytosolic and nuclear proteins can also be glycosylated with N-acetylglucosamine (GlcNAc) by a cytosolic glycosyltransferase (Moremen et al. 2012; Yang and Qian 2017). Some types of O-glycosylation are specific to distinct classes of proteins or domains (Moremen et al. 2012; Haltom and Jafar-Nejad 2015; Larsen et al. 2017). In this review the term O-glycosylation will refer to GalNAc- or mucin-type O-glycosylation, which is one of the most widespread forms of protein O-glycosylation (Bennett et al. 2012). N-linked glycosylation of proteins is acquired cotranslationally as they fold in the endoplasmic reticulum (Rothman and Lodish 1977). The initial glycan moiety is comprised of nine mannose (Man), three glucose (Glc) and two GlcNAc residues that are transferred on proteins from a dolichol-phosphate-linked precursor to asparagine residues within NXS/T consensus sequons (X—all amino acids except P) (Robbins et al. 1977; Bause 1983; Kornfeld and Kornfeld 1985). The initial glycan structure is trimmed to Man8GlcNAc2 in the endoplasmic reticulum, with intermediates assisting in protein quality control (Hammond et al. 1994; Helenius and Aebi 2004). In the Golgi apparatus, the glycans can be further trimmed, elongated and branched by a differentially expressed set of glycosidases and glycosyltransferases to yield hybrid or complex-type N-glycans (Hunt et al. 1978; Rabouille et al. 1995; Schachter 2000). As proteins pass through the Golgi apparatus, they can also be modified with mucin-type O-glycans, initiated by a family of 20 GalNAc-transferases that add the initial GalNAc monosaccharide to Ser, Thr and possibly Tyr residues (Rottger et al. 1998; Halim et al. 2011; Bennett et al. 2012). The GalNAc-transferases exhibit somewhat overlapping, but also distinct substrate specificities and the expression of the isoenzymes is regulated in a tissue- and differentiation-specific manner (Sutherlin et al. 1997; Wandall et al. 1997, 2007; Mandel et al. 1999; Young et al. 2003; Tian and Ten Hagen 2006). Since there are no conserved protein sequence motifs for general or isoform-specific O-glycosylation, it is much more difficult to predict this modification (Gerken et al. 2011). However, prediction algorithms exist and are rapidly improving due to the vast increase in the number of identified O-glycosites using proteome-wide strategies. Nevertheless, experimental evidence is still required for reliable site identification (Steentoft et al. 2013). O-linked glycans can be further elongated by competing glycosyltransferases to form up to eight core structures, which adds to the heterogeneity of O-glycosylation and impedes its analysis (Brockhausen et al. 1990; Yeh et al. 1999; Dalziel et al. 2001; Schachter 2000). Although many types of protein glycosylation exist, N- and mucin-type O-linked glycosylation are the most studied in viral research. These types of glycosylation will also be the focus in the following sections.