Ion coupled fluid transport by the corneal endothelium is stimulated by adenosine through a cAMP dependent mechanism. This study examines if anion conductance is enhanced by cAMP and, hence by adenosine. Cl- fluxes, measured by changes in fluorescence of the Cl- sensitive dye SPQ, following removal or re-addition of Cl- Ringer, could be accelerated by 20 mu M forskolin or 10 mu M adenosine. The cAMP cocktail (20 mu M forskolin + 100 mu M IBMX + 100 mu M cpt-cAMP) had no effect on resting [Cl-](i). However, when Cl- influx was inhibited by 100 mu M furosemide, net Cl- efflux was observed in response to the cAMP cocktail, Exposure to the cAMP cocktail alone depolarized the resting membrane potential. Conversely, the cAMP cocktail caused a relative hyperpolarization in cells which had been previously depolarized beyond the equilibrium potential for Cl- (ECl-), by application of 1 mu M Gramicidin D, cAMP dependent changes in membrane potential could be inhibited by 50 mu M NPPB, but not by 200 mu M DPC, 100 mu M H2DIDS or 50 mu M glibenclamide. Taken together, these results are consistent with NPPB-sensitive, cAMP activated Cl- channels. To examine if these channels are permeable to HCO3-, changes in pHi in response to the cAMP cocktail were measured in acidified and depolarized cells in the absence of Na+. The cAMP cocktail caused an increase in pHi only when HCO3- was present, consistent with HCO3- influx. In control HCO; Ringer, the cAMP cocktail caused a transient decrease in pHi, which could not be accounted for by inhibition of Na+:nHCO(3)(-), cotransport or stimulation of Cl-/HCO3-, exchange. These results are consistent with conductive HCO3- efflux through cAMP activated channels. We conclude that cultured bovine corneal endothelial cells possess cAMP activated anion channels. (C) 1997 Academic Press Limited.