Bacterial selenocysteine synthase converts seryl-tRNA^Sec to selenocysteinyl-tRNA^Sec for selenoprotein biosynthesis. The identity of this enzyme in archaea and eukaryotes is unknown. On the basis of sequence similarity, a conserved open reading frame has been annotated as a selenocysteine synthase gene in archaeal genomes. We have determined the crystal structure of the corresponding protein from Methanococcus jannaschii, MJO158. The protein was found to be dimeric with a distinctive domain arrangement and an exposed active site, built from residues of the large domain of one protomer alone. The shape of the dimer is reminiscent of a substructure of the decameric Escherichia coli selenocysteine synthase seen in electron microscopic projections. However, biochemical analyses demonstrated that MJ0158 lacked affinity for E. coli seryl-tRNA^Sec or M. jannaschii seryl-tRNA ^Sec, and neither substrate was directly converted to selenocysteinyl-tRNA^Sec by MJ0158 when supplied with selenophosphate. We then tested a hypothetical M. jannaschii O-phosphoseryl-tRNA^Sec kinase and demonstrated that the enzyme converts seryl-tRNA^Sec to O-phosphoseryl-tRNA^Sec that could constitute an activated intermediate for selenocysteinyl-tRNA^Sec production. MJ0158 also failed to convert O-phosphoseryl-tRNA^Sec to selenocysteinyl-tRNA ^Sec. In contrast, both archaeal and bacterial seryl-tRNA synthetases were able to charge both archaeal and bacterial tRNA^Sec with serine, and E. coli selenocysteine synthase converted both types of seryl-tRNA ^Sec to selenocysteinyl-tRNA^Sec. These findings demonstrate that a number of factors from the selenoprotein biosynthesis machineries are cross-reactive between the bacterial and the archaeal systems but that MJ0158 either does not encode a selenocysteine synthase or requires additional factors for activity.