We report the cloning and partial characterization of the fourth member of the vertebrate heparan sulfate/ heparin: GlcNAc N-deacetylase/GlcN N-sulfotransferase family, which we designate NDST4. Full-length cDNA clones containing the entire coding region of 872 amino acids were obtained from human and mouse cDNA libraries. The deduced amino acid sequence of NDST4 showed high sequence identity to NDST1, NDST2, and NDST3 in both species. NDST4 maps to human chromosome 4q25-26, very close to NDST3, located at 4q26 -27. These observations, taken together with phylogenetic data, suggest that the four NDSTs evolved from a common ancestral gene, which diverged to give rise to two subtypes, NDST3/4 and NDST1/2. Reverse transcriptionpolymerase chain reaction analysis of various mouse tissues revealed a restricted pattern of NDST4 mRNA expression when compared with NDST1 and NDST2, which are abundantly and ubiquitously expressed. Comparison of the enzymatic properties of the four murine NDSTs revealed striking differences in N-deacetylation and N-sulfation activities; NDST4 had weak deacetylase activity but high sulfotransferase, whereas NDST3 had the opposite properties. Molecular modeling of the sulfotransferase domains of the murine and human NDSTs showed varying surface charge distributions within the substrate binding cleft, suggesting that the differences in activity may reflect preferences for different substrates. An iterative model of heparan sulfate biosynthesis is suggested in which some NDST isozymes initiate the N-deacetylation and N-sulfation of the chains, whereas others bind to previously modified segments to fill in or extend the section of modified residues.Heparan sulfate and heparin bind a variety of growth factors, enzymes, and extracellular matrix proteins (1). These interactions depend on specific arrangements of variably sulfated glucosaminyl residues (GlcNAc, GlcN, 1 and GlcNS) and glucuronic (GlcA) and iduronic acids. The assembly of these sequences proceeds in a stepwise manner as follows. (i) The chains initiate by formation of the linkage tetrasaccharide, GlcA1,3Gal1,3Gal1,4Xyl, on serine residues of core proteins; (ii) the chains elongate by alternating the additions of GlcNAc␣1,4 and GlcA1,4 residues; (iii) the chains are modified initially by N-deacetylation and N-sulfation of subsets of GlcNAc residues, (iv) adjacent D-GlcA residues undergo C5-epimerization to L-iduronic acid, and (v) sulfation occurs at C2 of the uronic acid residues and at C6 and C3 of glucosaminyl residues. In this scheme, GlcNAc N-deacetylation and N-sulfation creates the prerequisite substrate needed for the later modification reactions (reviewed by Rodén (2)). These reactions are catalyzed by a family of enzymes designated the GlcNAc N-deacetylase/N-sulfotransferases (NDSTs).Three NDST isozymes have been identified in vertebrates, whereas only single orthologs are known in Drosophila melanogaster and Caenorhabditis elegans (3-11). Mutations in these genes can have profound effects on development. In D. melanogas...