Abstract
A 2.4-kilobase cDNA clone for human steroid-sulfatase (STS) was isolated and sequenced, which encoded an enzymatically active protein. The deduced amino acid sequence comprises 583 amino acids with an N-terminal signal peptide of 21 or 23 residues and four potential N-glycosylation sites. Two of the N-glycosylation sites are utilized and were localized to the asparagine residues 47 and 259. STS has the solubility properties of an integral membrane protein. The resistance of STS toward proteinase K after translocation into microsomes suggests that most, if not all, sequences of STS are exposed at the luminal side of microsomes. The deduced amino acid sequence predicts two membrane-spanning domains (amino acids 185-211 and 213-237) separated by a helix-breaking proline residue. We propose for STS a three-domain model. Two glycosylated luminally oriented domains of 161 and 346 residues are separated by a hydrophobic domain spanning the membrane twice in opposite directions. STS expressed in BHK-21 cells is located predominantly in the endoplasmic reticulum; smaller fractions are found in the Golgi, at the cell surface, multivesicular endosomes, as well as in lysosomes. The stability of STS in lysosomes may be related to the high homology of the two luminal domains of STS with the lysosomal sulfatases, arylsulfatase A, and arylsulfatase B. In spite of its similarity with these two lysosomal sulfatases, STS does not contain mannose 6-phosphate residues and is transported to lysosomes by a mannose 6-phosphate receptor-independent mechanism.