In contrast to the widespread ability of bacteria, plants, and animals to incorporate selenium nonspecifically into proteins in the form of selenomethionine residues, the selenoamino acid selenocysteine occurs as a highly specific component of a few selenium-dependent enzymes. Selenocysteine has been identified in glycine reductase, formate dehydrogenase, and hydrogenase of bacterial origin and glutathione peroxidase from mammalian and avian sources. In these enzymes there is evidence that the selenol group, which is largely ionized at physiological pH, functions as a redox center. It now seems clear, from studies with both prokaryotes and eukaryotes, that the UGA opal stop codon is used to specify the cotranslational insertion of selenocysteine into proteins. The factors that allow this unusual use of the stop codon are, however, unknown. The occurrence of selenium as a normal constituent of several bacterial tRNA species has been established. The presence of a selenonucleoside, 5-methylaminomethyl-2-selenouridine, in the first or wobble position of the anticodons of certain glutamate and lysine iso-acceptor species influences codon-anticodon interaction and thus may serve to regulate translational processes. The biosynthesis of the selenonucleoside appears to involve the ATP-dependent activation of the sulfur in a preformed 5-methylaminomethyl-2-thiouridine residue in tRNA and replacement of the sulfur with selenium.