The role of adenosine 3',5'-cyclic monophosphate (cAMP)-dependent protein kinase A (PKA) in mediating the hydrosmotic effect of vasopressin in in vitro microperfused rabbit cortical collecting ducts (CCDs) was examined. We measured PKA substrate phosphorylation and water permeability [hydraulic conductivity (Lp) = 10(-7) cm.atm-1.s-1], stimulated by substituted cAMP analogues selective for a unique cAMP binding site (site A or B) on PKA regulatory subunit (R). Synergy between site A- and site B-selective analogues suggests involvement of PKA, because both sites must be occupied for R to dissociate from the catalytic subunit (C), allowing phosphorylation to proceed. As single agents, the site B-selective analogues 8-(4-chlorophenylthio)-cAMP (8-CPT) and 8-thiomethyl-cAMP (8-SCH3) were at least two orders of magnitude more potent than the site A-selective analogues N6-monobutyryl-cAMP (N6-mono) or N6-benzoyl-cAMP (N6-benz). Combinations of subthreshold concentrations of two site A analogues (N6-mono+N6-benz) or two site B-selective analogues (8-CPT + 8-SCH3) failed to significantly increase protein phosphorylation or water permeability. In contrast, combination of a site A plus site B analogue synergistically stimulated both protein phosphorylation and Lp. Rp-cAMPS, an inhibitor of cAMP binding to PKA, reduced both vasopressin (41% inhibition)- and cAMP (56% inhibition)-stimulated water permeability. H-89 (50 microM), an inhibitor of PKA kinase activity, also blocked cAMP-stimulated water permeability (90% inhibition). These findings suggest that vasopressin-induced water permeability in the rabbit CCD is mediated by PKA.