Inhibition of hormone-stimulated hepatic glycogenolysis by fructose (Fru) has been attributed to accumulation of the competitive inhibitor Fru1P and/or to the associated depletion of the substrate phosphate (Pi). To evaluate the relative importance of either factor, we used the Fru analogue 2,5-anhydro-D-mannitol (aHMol). This analogue is avidly phosphorylated, traps Pi, and inhibits hormone-stimulated glycogenolysis, but it is not a gluconeogenic substrate, and hence does not confound glycogenolytic glucose production. Livers were continuously perfused with dibutyryl-cAMP (100 microM) to clamp phosphorylase in its fully activated a form. We administered aHMol (3.8 mM), and studied changes in glycogenolysis (glucose, lactate and pyruvate output) and in cytosolic Pi and phosphomonoester (PME), using in situ 31P-NMR spectroscopy (n = 4). Lobes of seven livers perfused outside the magnet were extracted for evaluation, by high-resolution 31P-NMR, of the evolution of aHMol1P and of aHMol(1,6)P2. After addition of aHMol, both glycogenolysis and the NMR Pi signal dropped precipitously, while the PME signal rose continuously and was almost entirely composed of aHMol1P. Inhibition of glycogenolysis in excess of the drop in Pi could be explained by continuing accumulation of aHMol1P. A subsequent block of mitochondrial ATP synthesis by KCN (1 mM) caused a rapid increase of Pi. Despite recovery of Pi to values exceeding control levels, glycogenolysis only recovered partially, attesting to the Pi-dependence of glycogenolysis, but also to inhibition by aHMol phosphorylation products. However, KCN resulted in conversion of the major part of aHMol1P into aHMol(1,6)P2. Residual inhibition of glycogenolysis was due to aHMol1P. Indeed, the subsequent withdrawal of aHMol caused a further gradual decrease in the proportion of aHMol1P (being converted into aHMol(1,6)P2, in the absence of de novo aHMol1P synthesis), and this resulted in a gradual de-inhibition of glycogenolysis, in the absence of marked changes in Pi. Glycogenolytic rates were consistently predicted by a model assuming non-saturated Pi kinetics and competition by aHMol1P exclusively: In conclusion, limited Pi availability and the presence of competitive inhibitors are decisive factors in the control of the in situ catalytic potential of phosphorylase a.