Previous studies from this laboratory with a 67-amino acid synthetic peptide (P-67) demonstrated directly that the first predicted nucleotide binding fold of the cystic fibrosis transmembrane conductance regulator (CFTR) binds ATP (Thomas, P.J, Shenbagamurthi, P., Ysern, S., and Pedersen, P.L. (1991) Science 251, 555-557). Although mutational analysis within the predicted second nucleotide binding fold indicates that this domain may be functionally important also, direct evidence for nucleotide binding is lacking. Here, we report the design, chemical synthesis, and purification of a 51-amino acid segment (P-51) of the second predicted nucleotide binding fold of CFTR and demonstrate that this peptide binds ATP. P-51 consists of amino acid residues from glutamic acid 1228 through threonine 1278 and contains a motif, GX4GKS, very similar or identical to that found in many nucleotide-binding proteins. The freshly dissolved peptide moves predominantly as a single species upon molecular sieve chromatography and readily binds ATP without eliciting its hydrolysis. P-51 also readily binds the fluorescent ATP analogs TNP-ATP (2'(3')-0-(2,4,6-trinitrophenyl)-adenosine-5'-triphosphate) and TNP-ADP but exhibits much less capacity to bind TNP-AMP. ATP displaces TNP-ATP with a Kd (ATP) of 0.46 mM. In the presence of the denaturant urea, P-51 loses most of its binding capacity indicating that structure is important for binding. Consistent with this conclusion, circular dichroism spectroscopy revealed that P-51 has significant secondary structure. Elements of such structure calculated from deconvolution of the circular dichroism spectra compare favorably with those predicted from the program of Chou, P.Y., and Fasman, G.D. (1977) J. Mol. Biol. 115, 135-175. These experiments provide the first direct evidence that the second predicted nucleotide binding fold of CFTR binds ATP and define a 51-amino acid segment within the approximately 150-amino acid fold critical for this function. They also indicate that the beta and gamma phosphate groups of ATP may be important for binding and that the 51-amino acid region studied is not sufficient to catalyze ATP hydrolysis. Finally, as seven different mutations within P-51 are known to cause cystic fibrosis, these studies will be important in future efforts to understand the molecular basis of the disease.