We perfused livers from fed rats with a balanced salt solution containing 1 mmol/L glucose. Under these conditions a low steady rate of glycogenolysis was observed (approximately 1.7 micromol glucose equivalents/g/min; 20% of the maximal glycogenolytic activity). Nitric oxide (NO) transiently stimulated hepatic glucose production. A maximal response (on average doubling basal glucose output) was observed with 34 micromol/L NO. The same concentration of nitrite (NO2-) was ineffective. Half-maximal effects were seen at 8 to 10 micromol/L NO, irrespective of the flow direction (portocaval or retrograde). This glycogenolytic response to NO corresponded to a partial activation of phosphorylase. The NO effect was not additive to maximal stimulation of glycogenolysis (7.7 +/- 0.2 micromol hexose equivalents/g/min; n = 4) by 100 micromol/L dibutyryl cyclic adenosine monophosphate (Bt2cAMP). The requirement for activation of phosphorylase was also evidenced by the ineffectiveness of NO in phosphorylase-kinase-deficient livers of gsd/gsd rats. The NO effect was blocked by co-administration of cyclooxygenase inhibitors (50 micromol/L ibuprofen, 50 micromol/L indomethacin, or 2 mmol/L aspirin), suggesting a mediatory role of prostanoids from nonparenchymal cells. This conclusion was confirmed by the fact that NO did not activate phosphorylase in isolated hepatocytes. Moreover, NO was no longer glycogenolytic in livers perfused with Ca2+-free medium, in agreement with the known mediatory role of Ca2+ in prostanoid-mediated responses. Surprisingly, in Ca2+-free medium NO inhibited the basal glucose production. This coincided with an increased elution of cyclic guanosine monophosphate (cGMP). Inhibition of glycogenolysis by NO under these conditions was blocked by 1 mmol/L theophylline, suggestive for involvement of cGMP-stimulated cAMP phosphodiesterase. However, we could not confirm that an increase in cGMP resulted in a drop in cAMP. In conclusion, NO recruits opposing mechanisms with respect to modulation of basal hepatic glycogenolysis. In the presence of Ca2+, activation of phosphorylase with stimulation of glycogenolysis dominates. Cyclooxygenase inhibitors abolish this effect. Activation by NO of the cyclooxygenase in nonparenchymal cells is a distinct possibility. In the absence of Ca2+, inhibition of basal glycogenolysis becomes observable. It remains to be established whether this results from cGMP-mediated stimulation of hydrolysis of cAMP.