CORE PROTEIN-STRUCTURE AND SEQUENCE DETERMINE THE SITE AND PRESENCE OF HEPARAN-SULFATE AND CHONDROITIN SULFATE ON SYNDECAN-1

Most proteoglycans bear either chondroitin sulfate or heparan sulfate chains linked to serine residues at Ser-Gly attachment sites on the co re protein. However, only a fraction of proteins with Ser-Gly sites ex hibit glycosaminoglycan chains. A variable proportion of these sites m ay be glycanated, and an unknown mechanism distinguishes whether these sites are for chondroitin sulfate or heparan sulfate. To evaluate the core protein features that determine whether and where chondroitin su lfate or heparan sulfate will be linked, we have studied mouse syndeca n-1, a transmembrane proteoglycan that is invariably glycanated and ca n contain both chondroitin sulfate and heparan sulfate chains. The ext racellular domain of the syndecan-1 core protein contains five Ser-Gly sites, three clustered near its N terminus and two adjacent to the tr ansmembrane domain near its C terminus. We have established the distri bution of glycosaminoglycans on these attachment clusters. In contrast to the C-terminal cluster, the N terminal cluster was always glycanat ed, suggesting that this domain of the core protein contains sequences responsible for the invariable attachment of glycosaminoglycan chains . Solely chondroitin sulfate was found on the C-terminal cluster. This cluster contains the sequences EGSGE and ETSGE, both estimated to be on the protein surface in a hydrophilic environment. Heparan sulfate w as found solely on the N-terminal cluster, which also bears some chond roitin sulfate. This cluster contains the sequences FSGSGTG and DGSGD, the former estimated to be in a hydrophobic pocket and the latter, si milar to the sequence on the C-terminal cluster, in an exposed hydroph ilic region. This glycosaminoglycan distribution was identical on mous e syndecan-1 produced by either mouse epithelial (NMuMG) or hamster me senchymal (CHO) cells, suggesting that site-specific attachment of gly cosaminoglycans is independent of cell type. These results implicate a cellular mechanism that distinguishes among the potential sites and a ttaches the correct glycosaminoglycan type unambiguously. Thus, struct ural elements of the core protein other than the Ser-Gly attachment si tes determine if a site will be glycanated and, if so, whether with ch ondroitin sulfate or heparan sulfate.