We have characterized the S-antigen genes from human and mouse, 33K protein genes from mouse and human. The S-antigen genes were approximately 50 kbp in length, contained 16 exons and 15 introns, and were composed of 97 percent intron and 3 percent exon. The 5~-flanking regions of the genes, approximately 1.5 kbp long, had no known regulatory elements for transcription such as TATA, GC, or CCAAT boxes. Regulatory sequences and nuclear factors governing tissue-restricted expression of the mouse arrestin gene were investigated. The results showed that while proximal promoter sequence -38 to +304 are sufficient to direct low levels of retina-specific gene expression, sequences extending upstream to position -209 support higher levels of expression in the retina as well as detectable expression in the lens, pineal gland, and brain. Wit n the interval between positions -205 and -185 is a region that contains two direct repeats of the hexamer, TGACCT. The proximal promoter binds three apparently retina-specific nuclear factors, Bp1, Bp2, and Bp3, through overlapping sequences centered between positions -25 and -15. Bp1 and Bp3 also recognize a closely related sequence found in the promoter regions of several other vertebrate photoreceptor-specific genes. Moreover, the consensus binding site for Bp1, designated PCE1, is identical to RCS1, an element known to play a critical role in eliciting photoreceptor-specific g e expression in Drosophila melanogaster. The results suggest that PCE1 and RCS1 are functionally, as well as structurally, similar and, that despite marked differences in the fly and vertebrate visual system, the transcriptional machinery involved in photoreceptor-specific gene expressio has been strongly evolutionarily conserved.
