The counter-regulatory effect of adenosine, isoprenaline and selected cyclic AMP analogues on insulinstimulated 3-O-methylglucose transport and insulin binding were studied in rat fat-cells. Isoprenaline alone had no consistent effect on glucose transport in the presence of maximally effective insulin concentrations.However, it decreased insulin binding by approx. 20% and increased EC50 (concn. giving 50% of maximal stimulation) for insulin from 8 + 1 to 17 + 2 #sunits/ml. Adenosine deaminase (ADA) alone only exerted a slight effect, whereas isoprenaline and ADA in combination consistently decreased the maximal effect of insulin on glucose transport, decreased insulin binding by approx. 30% and markedly decreased insulinsensitivity (EC50 61 + 8 ,units/ml). In cells from pertussis-toxin-treated animals, isoprenaline alone decreased the insulin response by approx. 75%, decreased insulin binding by approx. 45% and caused a marked rightward shift in the dose-response curve for insulin (EC50 103 + 34 #sunits/ml). The importance of cyclic AMP for these effects was evaluated with the analogue M6-monobutyryl cyclic AMP, which is resistant to hydrolysis by the phosphodiesterase. The importance of phosphodiesterase activation by insulin was studied with 8-bromo cyclic AMP, which is an excellent substrate for this enzyme. N6-Monobutyryl cyclic AMP, in contrast with 8-bromo cyclic AMP, markedly impaired insulin-sensitivity (EC50 approx. 100,uunits/ml).However, the maximal effect of insulin was only slightly attenuated. In conclusion: (1) fl-adrenergic stimulation and cyclic AMP markedly alter insulin-sensitivity, but not responsiveness, mainly through post-receptor perturbations; (2) when cyclic AMP is increased phosphodiesterase activation by insulin is a critical step to elicit insulin action; (3) adenosine modulates the insulin-antagonistic effect of ,-adrenergic stimulation via Ni (inhibitory nucleotide-binding protein) through both cyclic-AMP-dependent and -independent mechanisms.