Cystic fibrosis is an autosomal recessive disorder affecting chloride transport in pancreas, lung, and other tissues, which is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR). Certain alkyl xanthines such as CPX (8-cyclopentyl-1,3-dipropylxanthine) stimulate Cl- efflux from cells bearing the delta F508 genotype common to most cases of cystic fibrosis. We have hypothesized that the CFTR molecule itself might be the site for CPX action, perhaps in the region of the first nucleotide binding fold (NBF-1) domain. Therefore, to test this hypothesis directly we have used a rapid membrane filtration assay to measure the kinetics of association and dissociation of [3H]CPX to both recombinant NBF-1 and recombinant NBF-1 bearing the delta F508 mutation. We report that [3H]CPX binds with higher affinity to the delta F508-NBF-1 of CFTR (Kd = 1.0 nM) than to the wild-type NBF-1 of CFTR (Kd = 17.0 nM). These Kd values were calculated from direct measurements of the association and dissociation rate constants. The rate constants for the dissociation reaction of the wild-type NBF-1 and delta F508-NBF-1 of CFTR were not different from each other. However, the corresponding rate constants for the association reaction were k(+1) (NBF-1) = 4.7 +/- 0.9 x 10(4) M(-1) s(-1) and k(+1) (delta F508-NBF-1) = 1.6 +/- 0.3 x 10(5) M(-1) s(-1), respectively. These Kd values were corroborated by equilibrium-binding experiments, which gave very similar values. We have also measured the relative displacement of various xanthines from both wild-type NBF-1 and delta F508-NBF-1, in anticipation that the order of potencies for binding might parallel the action of the different xanthines on CF cells. For wild-type NBF-1, the rank order was DA-CPX > DAX > CPX > caffeine > adenosine >> IBMX > 2-thioCPX. For delta F508-NBF-1, the rank order was DAX > CPX > caffeine > DA-CPX > adenosine >> IBMX > 2-thioCPX. These relative potencies show close parallels with previously observed relative potencies of these drugs on CF cells, and thus lend strong support to the hypothesis that the mechanism of action on CF cells may involve direct interaction with the CFTR molecule itself.