O-linked glycoproteins begin their glycosylation with the action of the Golgi-specific enzyme, GalNAc transferase, which attaches an N-acetylgalactosamine to the hydroxyl group of a serine or threonine. The determination of which residue to glycosylate appears to be directed by secondary and tertiary structure as previously mentioned, and often occurs in dense clusters of glycosylation. Despite being fairly small additions (usu. <5 residues), the combined oligosaccharide chains attached to an O-linked glycoprotein can contribute over 50% of the mass of a glycoprotein. Two of the better known O-linked glycoproteins are mucin, a component of saliva, and ZP3, a component of the zona pellucida (which protects egg cells). These two examples also illustrate a key property of glycoproteins and glycolipids in general: the sugars are highly hydrophilic and hold water molecules to them, greatly expanding the volume of the protein.
Figure 15. O-linked glycosylation in the Golgi involves attachment of only a few sugars to serine or threonine.
Interestingly, this protective waterlogged shell can mask parts of the protein core. In the case of the cell adhesion molecule, NCAM, which is a highly polysialylated glyco- protein at certain developmental stages and locations, and unglycosylated in others, the naked protein can be recognized as an adhesive substrate while the glycosylated protein can be recognized as a repulsive substrate to other cells. Even in highly glycosylated proteins though, the sugar residues often acts as recognition sites for other cells. For instance, the zona pellucida is very important as a physical barrier that protects the egg, but glycosylated ZP3 also acts as a sperm receptor.