Selenium has been implicated in cancer prevention, but the mechanism and possible involvement of selenoproteins in this process are not understood. To elucidate whether the 15-kDa selenoprotein may play a role in cancer etiology, the complete sequence of the human 15-kDa protein gene was determined, and various characteristics associated with expression of the protein were examined in normal and malignant cells and tissues. The 51-kilobase pair gene for the 15-kDa selenoprotein consisted of five exons and four introns and was localized on chromosome 1p31, a genetic locus commonly mutated or deleted in human cancers. Two stemloop structures resembling selenocysteine insertion sequence elements were identified in the 3-untranslated region of the gene, and only one of these was functional. Two alleles in the human 15-kDa protein gene were identified that differed by two single nucleotide polymorphic sites that occurred within the selenocysteine insertion sequence-like structures. These 3-untranslated region polymorphisms resulted in changes in selenocysteine incorporation into protein and responded differently to selenium supplementation. Human and mouse 15-kDa selenoprotein genes manifested the highest level of expression in prostate, liver, kidney, testis, and brain, and the level of the selenoprotein was reduced substantially in a malignant prostate cell line and in hepatocarcinoma. The expression pattern of the 15-kDa protein in normal and malignant tissues, the occurrence of polymorphisms associated with protein expression, the role of selenium in differential regulation of polymorphisms, and the chromosomal location of the gene may be relevant to a role of this protein in cancer.The essential trace element selenium occurs in several proteins found in bacteria (1), archaea (2), and eukaryotes (1, 3). All of the known selenoproteins described to date incorporate selenium into protein co-translationally as the amino acid selenocysteine (Sec) 1 in response to the UGA codon (1, 3, 4), with the exception of several bacterial molybdenum-containing enzymes such as Clostridium barkeri nicotinic acid hydroxylase that contains an active center dissociable selenium species (5-7). Among the selenoproteins thus far identified in mammals are four glutathione peroxidases, three thioredoxin reductases, three thyroid hormone deiodinases, selenophosphate synthetase, selenoprotein P, selenoprotein W, selenoprotein T, selenoprotein R (also called selenoprotein X), selenoprotein N, and a 15-kDa selenoprotein (1, 3, 8 -10).The 15-kDa selenoprotein was initially identified as a strongly labeled protein that was detected when human T cells were grown in the presence of [75 Se]selenite (11). The open reading frame within the human 15-kDa protein cDNA coded for 162 residues and contained an in-frame TGA codon that would result in the incorporation of Sec at codon position 93. A putative stem-loop structure, designated the Sec insertion sequence (SECIS) element, was predicted in the 3Ј-untranslated region (3Ј-UTR). SECIS elements are ...