We showed previously that the synthesis of heparan sulfate on betaglycan occurs at a Ser-Gly dipeptide flanked by a cluster of acidic residues and an adjacent tryptophan (Zhang, L., and Esko, J.D. (1994) J. Biol. Chem. 269, 19295-19299). A survey of the protein data base revealed that most heparan sulfate proteoglycans contain repetitive (Ser-Gly)n segments (n = 2) and a nearby cluster of acidic residues. To study the role of these amino acid sequences in controlling heparan sulfate synthesis, we have examined the assembly of glycosaminoglycans on Chinese hamster ovary (CHO) cell syndecan-1. The glycosylation sites were mapped by making chimeric proteoglycans containing segments of CHO syndecan-1 cDNA fused to Protein A. Two sites near the transmembrane domain (-EGS205GEQ- and -ETS215GEN-) were used solely for chondroitin sulfate synthesis, whereas three sites near the N terminus (-DGS35GDDSDNFS45GS47GTG-) supported both heparan sulfate and chondroitin sulfate synthesis. The strongest sites for heparan sulfate synthesis consisted of the repeat unit, -S45GS47G-. An unusual coupling phenomenon occurred across the adjacent SG dipeptides, leading to a greater proportion of heparan sulfate than predicted by the behavior of each site acting independently. The clusters of acidic residues adjacent to the heparan sulfate sites play important roles as well. These sequence motifs suggest a set of rules for predicting whether heparan sulfate assembles at glycosylation sites in proteoglycan core proteins.