br Discussion In response to an increase in

Discussion In response to an increase in glucose or DHA concentration, FBPase showed nuclear import and recruitment in a peripheral cell compartment. Interestingly, FBPase and GS were co-recruited to the periphery of the cells. This is precisely the site where glycogen is initially synthesized and the deposits of the polysaccharide grow from the periphery of the hepatocytes towards the interior [15]. This finding supports the notion of the formation of a complex between FBPase and the glycogen synthesis enzymes. This complex may allow the channeling of substrates and the enhancement of glycogen synthesis. In this context, the importance of the compartmentation of palmitic acid and the concomitant channeling of substrates has been discussed by Ovadi and Srere [20], [21]. This micro-compartmentation, together with other factors involved in channeling, provides many potential biological advantages, such as the isolation of intermediates from competing reactions and a new means of metabolic regulation by modulating enzyme associations, which ensure the proper function and integration of metabolic pathways [20], [21]. We observed similar capacities of glucose and DHA to stimulate the redistribution of FBPase and GS to the cell periphery. This result indicates that the direct pathway of glycogen synthesis from glucose and the indirect pathway from gluconeogenic precursors are located in the same subcellular functional compartment. By contrast, DHA has no effect on the subcellular localization of GK (data not shown). We measured Glu-6-P as an intermediate precursor of glycogen synthesis, and as previously reported [19], we found that part of the Glu-6-P produced by gluconeogenesis from DHA was channeled to glycogen deposition. Moreover, the colocalization of FBPase with GS was not inhibited by the action of insulin. This hormone stimulates glycogen synthesis, inhibits glycogen breakdown, and suppresses gluconeogenesis [22]. However, our data reveal that the synthesis of Glu-6-P and glycogen induced by DHA was not inhibited by insulin. These results show that the major effect of insulin is on neither Glu-6-P synthesis nor on the synthesis of glycogen, indicating that under these conditions the indirect pathway of glyconeogenesis is active in hepatocytes. These results are consistent with current data that demonstrate that liver glycogenolysis is acutely sensitive to small changes in plasma insulin, whereas gluconeogenic flux is not [23]. On the basis of the observation of the recruitment of FBPase and GS in the cell periphery induced by high glucose, we hypothesize that, even in this condition, FBPase participates in glycogen synthesis. This notion is also supported by studies which show that the postprandial synthesis of glycogen is dramatically decreased when the gluconeogenic pathway is blocked by specific inhibitors [24], [25]. Glucose utilization raises Glu-6-P and glycogen in hepatocytes, but also induces the increase of gluconeogenic/glycolytic intermediates. Thus, the recruitment and binding of FBPase to the GS compartment may prevent the inhibition of FBPase by the allosteric effector (AMP), thereby allowing the recycling and channeling of glycolytic intermediates to glycogen. Studies by Dzugaj's laboratory support this hypothesis and demonstrate that FBPase in a complex with aldolase is insensitive to AMP inhibition [26]. Furthermore, we found that insulin strongly induced a reversible FBPase translocation into the nucleus. This result confirms our previous data that showed the capacity of liver FBPase to translocate into the nuclei of hepatocytes and reveals, for the first time, a hormonal regulation mechanism for the nuclear localization of a gluconeogenic enzyme [11], [27]. The entry of proteins of a size greater than the exclusion limit of the nuclear pore complex depends on the presence of a nuclear localization signal (NLS) sequence within the protein. In the absence of this sequence, large proteins enter the nucleus via interactions with other proteins by a piggy-back mechanism. Rat liver FBPase is an homotetrameric enzyme of approximately 155 KDa, whereas the exclusion limit of the nuclear pore permits only the passive diffusion of proteins of less than 45 KDa [28]. Similarly, the export of proteins from the nucleus is an active process that depends on a distinct motif called a nuclear export signal (NES) sequence. Using neural networks that predict subcellular localization, we found that liver FBPase contains a classical NLS [11], [29], [30]. Nevertheless, we cannot rule out a piggy-back mechanism to explain FBPase nuclear import.