CHARACTERIZATION OF THE HUMAN BONE SIALOPROTEIN (BSP) GENE AND ITS PROMOTER SEQUENCE

Bone sialoprotein (BSP) is a major structural protein of the bone matr ix that is specifically expressed by fully-differentiated osteoblasts. To characterize the gene and to study the tissue- and differentiation stage-specific regulation of BSP gene transcription we have isolated and partially sequenced two overlapping genomic fragments which span t he complete human BSP gene and its promoter region. The similar to 15 kb gene comprises seven exons of 82 bp, 68 bp, 51 bp, 78 bp, 63 bp, 15 9 bp and 2.5 kb (1-7, respectively), separated by six introns of simil ar to 3 kb, 92 bp, 95 bp, similar to 3 kb, similar to 0.5 kb and simil ar to 4.5 kb. All of the intron-exon boundaries defining the splice si tes conform to the consensus sequence of: AG at the 3' splice site; an d GT at the 5' splice site, except the 3' splice site of exon 1. The f irst exon encodes the 5'-UTR, the second exon the signal sequence and the first two amino acids, exons 3 and 4 the Tyr- and Phe-rich amino t erminus, and exon 5 the first segment of polyglutamic acid. Exon 7 enc odes over half of the protein including a second polyglutamic acid seg ment, the RGD cell attachment motif, the sulphated tyrosine-rich C-ter minus and the 3'-UTR. The promoter region is characterized by an inver ted TATA-like sequence (TTTATA), nts - 28 to - 23 from the transcripti onal start site (+ 1), and an inverted CCAAT box (ATTGG) at - 54 to - 50. Analysis of chimeric constructs fused to a CAT reporter gene indic ate that the presence of both the inverted TATA-like sequence and CCAA T elements are required for basal promoter activity. Comparison of the human BSP promoter with the rat BSP promoter (Li and Sodek, 1993) rev eal that the nature and position of the inverted TATA-like sequence an d CCAAT box together with an Apl (- 148 to - 142), CRE (- 122 to - 116 ) and a homeobox-binding site (- 200 to - 191) have been conserved. A putative Glucocorticoid Response Unit (GRU) consisting of a Glucocorti coid Response Element (GRE) and an overlapping direct repeat (DR5) sim ilar to the retinoic acid receptor element (RARE) is present at - 1038 to - 1022. These studies have defined the structure of the human BSP gene and have identified novel transcriptional elements in the promote r that may be involved in the developmentally regulated, tissue specif ic expression of this gene.