1. This study compares the effect of three chemically unrelated cystic fibrosis transmembrane conductance regulator (CFTR) activators on epithelial cell monolayers expressing the G551D-CFTR mutant. 2. We measured Cl(-) transport as the amplitude of short-circuit current in response to the membrane permeable cAMP analogue 8-(4-chlorophenylthio)adenosine-3'-5'-cyclic monophosphate (CPT-cAMP) alone or in combination with a CFTR opener. The correction of G551D-CFTR defect was quantified by comparison with maximal activity elicited in cells expressing wild type CFTR. To this end we used Fisher rat thyroid (FRT) cells transfected with wild type or G551D CFTR, and primary cultures of human nasal epithelial cells. 3. In both types of epithelia, cAMP caused activation of Cl(-) transport that was inhibited by glibenclamide and not by 4,4'-diisothiocyanato-stilbene-2,2'-disulfonic acid. After normalising for CFTR expression, the response of FRT-G551D epithelia was 1% that of wild type monolayers. 4. Addition of genistein (10-200 micro M), but not of 8-cyclopentyl-1,3-dipropylxanthine (CPX, 1-100 micro M) or of the benzo[c]quinolizinium MPB-07 (10-200 micro M) to FRT-G551D epithelia pre-treated with cAMP, stimulated a sustained current that at maximal genistein concentration corresponded to 30% of the response of wild type epithelia. 5. The genistein dose-response curve was bell-shaped due to inhibitory activity at the highest concentrations. The dose-dependence in G551D cells was shifted with respect to wild type CFTR so that higher genistein concentrations were required to observe activation and inhibition, respectively. 6. On human nasal epithelia the correction of G551D-CFTR defective conductance obtained with genistein was 20% that of wild type. The impressive effect of genistein suggests that it might correct the Cl(-) transport defect on G551D patients.