Glycosylation

Glycosylation is the enzymatic attachment of carbohydrate chains to a peptide or protein in the cell. N-linked glycans attach to an asparagine in the sequon Asn-X-Ser or Asn-X-Thr; O-linked glycans attach to the hydroxyl of serine or threonine. The sugars are branched, often capped with sialic acid, and form a population of glycoforms rather than one structure; in a written sequence the glycan is a side-chain group on the residue that carries it.

Erythropoietin carries three N-linked and one O-linked chain amounting to roughly 40 percent of its mass, and darbepoetin alfa was engineered with five substitutions that add two N-glycosylation sites, tripling its half-life. None of this applies to a peptide made by solid-phase synthesis: glycosylation is a product of cellular expression, so a synthetic GLP-1 agonist or GnRH analogue carries no sugar at all.

For a glycoprotein, sequence identity is not product identity. The glycan pattern sets clearance, since desialylated chains are pulled from circulation by the liver, and immunogenicity, since non-human sugars from some cell lines are recognised by pre-existing antibodies. Biosimilar comparability therefore rests on glycan profiles, and an identical sequence from a different cell line can behave measurably differently.

Two errors recur. Glycosylation is not glycation: glycated haemoglobin, HbA1c, forms by non-enzymatic reaction with glucose and is a marker of exposure, not a designed modification. And a mass check that treats a glycoprotein like a peptide fails: the sequence-implied mass undercounts by the weight of the sugars, and the spectrum shows a spread of glycoforms rather than one peak; a product sold as a synthetic powder cannot be a glycosylated biologic.