Inorganic polyphosphate (polyP) is obtained by the polymerization of the terminal phosphate of ATP through the action of the enzyme polyphosphate kinase (PPK). Despite the presence of polyP in every living cell, a gene homologous to that of known PPKs is missing from the currently sequenced genomes of Eukarya, Archaea, and several bacteria. To further study the metabolism of polyP in Archaea, we followed the previously published purification procedure for a glycogen-bound protein of 57 kDa with PPK as well as glycosyl transferase (GT) activities from Sulfolobus acidocaldarius (R. Skórko, J. Osipiuk, and K. O. Stetter, J. Bacteriol. 171:5162-5164, 1989). In spite of using recently developed specific enzymatic methods to analyze polyP, we could not reproduce the reported PPK activity for the 57-kDa protein and the polyP presumed to be the product of the reaction most likely corresponded to glycogen-bound ATP under our experimental conditions. Furthermore, no PPK activity was found associated to any of the proteins bound to the glycogen-protein complex. We cloned the gene corresponding to the 57-kDa protein by using reverse genetics and functionally characterized it. The predicted product of the gene did not show similarity to any described PPK but to archaeal and bacterial glycogen synthases instead. In agreement with these results, the recombinant protein showed only GT activity. Interestingly, the GT from S. acidocaldarius was phosphorylated in vivo. In conclusion, our results convincingly demonstrate that the glycogen-protein complex of S. acidocaldarius does not contain a PPK activity and that what was previously reported as being glycogen-bound PPK is a bacterial enzyme-like thermostable glycogen synthase.