Refinement and Evaluation of a CFTR Homology Model and Identification of Residues Controlling Channel Gating

Cystic Fibrosis, which affects nearly 1 in 2,500 births in the Caucasian population, is caused by various mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) protein. CFTR is a member of the diverse ATP-binding-cassette (ABC) protein superfamily. While most ABC proteins are transporters, CFTR alone has evolved to function as an ion channel. Currently no crystal structure of full length CFTR exists, but due to its biological relevance, many CFTR homology models have been proposed. Using a recently published and experimentally supported model of CFTR, we have carried out all-atom Molecular Dynamics (MD) in an attempt to further characterize the dynamics of this recalcitrant protein. First, we refined the model utilizing MD flexible fitting (MDFF) and structural data from a 9-A cryo-EM map of CFTR (Rosenberg et al., 2011). Next, equilibrium MD simulations of the model within a membrane were used to probe 3 physically relevant states: the apo state, the proposed semi-apo state (1 ATP bound), and the bound state (2 ATP bound). Initial results show a difference in dimerization in the nucleotide-binding domains (NBDs) between the three states. In other experiments, residues in CFTR's NBDs indicating sequence divergence from the canonical ABC protein sequence (suggesting that they may be involved in slowing the ATP hydrolysis rate, which would likely support longer open durations) were returned to the canonical sequence. Burst durations (in ms) measured for single-channels expressed in oocytes were WT-CFTR (683 ± 59), S573E-CFTR (427 ± 83)∗, G576S/Y577A-CFTR (431 ± 46)∗, (∗ = p < 0.05). These results suggest that CFTR's NBDs evolved in a manner that would support the shift from transporter to channel function. (Support: 5R01-DK056481).