Riding the Conformational Wave to the Open Channel State in the CFTR Chloride Channel

Abstract

The pore structure of the CFTR chloride channel is unknown. We showed previously that R352 in TM6 forms a salt bridge with D993 in TM9; charge-destroying mutations at either site destabilized the open state, affecting conductance, selectivity, and pore blockade. Other pairs of interacting residues also contribute to stabilizing the open state. We continued these experiments to determine how steps leading to the dimerization of the NBDs upon binding of nucleotide relate to the steps leading to pore opening, using single-channel recordings of WT-CFTR and channels bearing a cysteine or alanine at 352, 993, or both. In R352C-CFTR, but not R352A-CFTR, modification of the cysteine by positively-charged MTSET+ and MTSEA+ recovered the stability of the open state. In D993C-CFTR, but not D993A-CFTR, negatively-charged MTSES- recovered the stability of the open state. In contrast, D993C-CFTR modified by MTSET+ retained the instability of the open state. The R352C/D993C-CFTR double mutant exhibited instability of the open state in both the absence and presence of DTT, suggesting that R352C did not form a disulfide with D993C. In WT-CFTR, exposure to AMP-PNP led to greatly prolonged channel openings, as expected. However, this response was not found for R352A-CFTR. Surprisingly, R352C/D993C-CFTR could be latched open by the bifunctional crosslinker, MTS-2-MTS, such that channels could not close upon washout of ATP. MD simulations based on CFTR homology models (see Dawson Lab abstract) predict conformational states in which R352 and D993 approach each other to within van der Waals distances. These results suggest that the binding of ATP at CFTR's NBDs initiates a conformational wave, which leads to a change in pore structure from the closed to the open state, the latter being stabilized by inter-TM interactions including the R352-D993 salt bridge. (Support: NIH-2R56DK056481-07)