FIGURE 1. Assembly of heparan sulphate and formation of binding sites for ligands.

In mammals, as many as 26 enzymes participate in the formation of HS chains. HSPG core proteins are xylosylated co-translationally at specific serine residues (by XYLT1 and XYLT2). A tetrasaccharide primer (GlcA-Gal-Gal-Xyl, where Gal is galactose and Xyl is xylose) is assembled in the Golgi (by GALT1 and GALT2, which are galactosyltransferases, and by glucuronyltransferase 1 (GLCAT1)), and the first GlcNAc unit is added (by EXTL3). EXT1/EXT2 co-polymerize GlcA4 and GlcNAc4 units into linear chains of 40–100 residues, which then undergo extensive modification by GlcNAc N-deacetylation and N-sulphation (denoted NS, by NDST1–4), C5 epimerization of D-GlcA to L-IdoA (by HS glucuronyl C5 epimerase, HSGLCE) and O-sulphation at C2 of uronic acids (denoted 2S, by HS2ST), and at C6 (denoted 6S, by HS6ST1–3) and more rarely at C3 (denoted 3S, by HS3ST1–6) of glucosamine residues. The length of the sulphated and non-sulphated segments varies. Outside the cell, two endosulphatases (SULF1 and SULF2) catalyse the removal of specific 6-O-sulphate groups, and secreted heparanase (HPA) can fragment the chains (not shown). The pattern of negatively charged sulphates and uronic acids creates binding sites for various protein ligands, including growth factors (such as FGF), receptors (such as FGFR) and protease inhibitors (such as antithrombin), as well as other proteins (not shown).