We studied the effects of cyclic AMP (cAMP) on HCO-3 transport by rabbit cortical collecting tubules perfused in vitro. Net HCO-3 secretion was observed in tubules from NaHCO3- loaded rabbits. 8-Bromo-cAMP-stimulated net HCO-3 secretion, whereas secretion fell with time in control tubules. Both isoproterenol and vasopressin (ADH) are known to stimulate adenylate cyclase in this epithelium; however, only isoproterenol stimulated net HCO-3 secretion. The mechanism of cAMP-stimulated HCO-3 secretion was examined. If both HCO-3 and H+ secretion were to occur simultaneously in tubules exhibiting net HCO-3 secretion, cAMP might increase the net HCO-3 secretory rate by inhibiting H+ secretion, by stimulating HCO-3 secretion, or both. These possibilities were examined using basolateral addition of the disulfonic stilbene (4,4'-diisothiocyanostilbene-2,2'-disulfonate (DIDS). In acidifying tubules from NH4Cl-loaded rabbits, DIDS eliminated HCO-3 reabsorption, a result consistent with known effects of DIDS as an inhibitor of H+ secretion. In contrast, cAMP left acidification (H+ secretion) intact. DIDS applied to HCO-3 secretory tubules failed to increase the HCO-3 secretory rate, indicating minimal H+ secretion in HCO-3 secreting tubules. Thus, inhibition of H+ secretion by cAMP could not account for the cAMP-induced stimulation of net HCO-3 secretion. cAMP-stimulated HCO-3 secretion was reversibly eliminated by 0 Cl perfusate, whereas luminal DIDS had no effect. Bath amiloride (1 mM) failed to eliminate cAMP-stimulated HCO-3 secretion when bath [Na+] was 145 mM or 5 mM. cAMP depolarized the transepithelial voltage. The collected fluid [HCO-3] after cAMP could be accounted for by electrical driving forces, suggesting that cAMP stimulates passive HCO-3 secretion. However, cAMP did not alter HCO-3 permeability measured under conditions expected to inhibit transcellular HCO-3 movement (0 Cl- solutions and bath DIDS). This measured HCO-3 permeability was not high enough to account, by passive diffusion, for the HCO-3 fluxes observed in Cl-containing solutions. We conclude the following: cAMP increased net HCO3- secretion by stimulating HCO3- secretion and not by inhibiting H+ secretion; this HCO3- secretion may have occurred by Cl-HCO3- exchange; Na+-H+ exchange appeared not to play a role in basolateral H+ extrusion under these conditions; and the stimulation of HCO3- secretion by isoproterenol, but not ADH, suggests the existence of separate cell cAMP pools or cellular heterogeneity in this cAMP response.