Fructose-1,6-bisphosphatase (FBPase) is one of the key enzymes in gluconeogenesis. Although FBPase activity has been detected in several hyperthermophiles, no orthologs corresponding to the classical FBPases from bacteria and eukaryotes have been identified in their genomes. An inositol monophosphatase (IMPase) from Methanococcus jannaschii which displayed both FBPase and IMPase activities and a structurally novel FBPase (FbpTk) from the hyperthermophilic archaeon Thermococcus kodakaraensis KOD1 have been proposed as the "missing" FBPase. For this study, using T. kodakaraensis, we took a genetic approach to elucidate which candidate is the major gluconeogenic enzyme in vivo. The IMPase/FBPase ortholog in T. kodakaraensis, ImpTk, was confirmed to possess high FBPase activity along with IMPase activity, as in the case of other orthologs. We therefore constructed Deltafbp and Deltaimp strains by applying a gene disruption system recently developed for T. kodakaraensis and investigated their phenotypes. The Deltafbp strain could not grow under gluconeogenic conditions while glycolytic growth was unimpaired, and the disruption resulted in the complete abolishment of intracellular FBPase activity. Evidently, fbpTk is an indispensable gene for gluconeogenesis and is responsible for almost all intracellular FBPase activity. In contrast, the endogenous impTk gene could not complement the defect of the fbp deletion, and its disruption did not lead to any detectable phenotypic changes under the conditions examined. These facts indicated that impTk is irrelevant to gluconeogenesis, despite the high FBPase activity of its protein product, probably due to insufficient transcription. Our results provide strong evidence that the true FBPase for gluconeogenesis in T. kodakaraensis is the FbpTk ortholog, not the IMPase/FBPase ortholog.