Structure of human osteonectin based upon analysis of cDNA and genomic sequences.

Overlapping human bone osteonectin cDNAs were obtained by screening two independent human SaOS-2 lambda gt11 libraries using antibovine osteonectin monoclonal antibodies. One clone contains a 0.54-kb insert and the other a 1.9-kb insert. Insertion fragments from lambda clones were liberated by restriction digestion and subcloned into pUC19 for sequencing. Digestion of the 1.9-kb insert with EcoRI released 0.4- and 1.5-kb fragments. Sequencing analysis revealed that the 0.54- and 0.4-kb fragments are identical except for 150 nucleotides missing at the 5' region of the 0.4-kb fragment. The composite nucleotide sequence of human osteonectin has a total length of 2091 nucleotides and is comprised of 50 nucleotides of 5'-noncoding sequence, a coding segment for 303 amino acids, a termination codon, and 1114 nucleotides of 3'-noncoding sequence. The primary transcript codes for 286 amino acids of mature protein and a 17-residue amino-terminal hydrophobic signal peptide. Outstanding properties inferred from the primary structure are putative Ca2+ binding domains located in the glutamic acid rich NH2 terminus (residues 1-52) and two "EF"-hand domains in the C-terminal half of the protein (residues 165-176 and 257-286). The mature protein also contains a cysteine-rich, highly hydrophilic region homologous to the ovomucoid serine protease inhibitors (residues 76-132). Overlapping human genomic clones in lambda EMBL3 for osteonectin have been isolated and characterized. Intron/exon junction sequencing of the human osteonectin gene shows the presence of 10 exons and 9 introns. The mature protein is encoded by nine exons separated by eight introns.(ABSTRACT TRUNCATED AT 250 WORDS)