On the Coupling Mechanism of CFTR Gating by ATP Hydrolysis

Abstract

Once phosphorylated, CFTR, a member of the ABC protein superfamily, functions as an ATP-gated chloride channel. It is generally held that ATP-induced NBD dimerization opens the gate located in the transmembrane domains (TMDs) and subsequent ATP hydrolysis in composite site 2 causes the NBD dimer to separate partially and subsequently closes the gate. Using pyrophosphate (PPi) or AMP-PNP as a bait, we have previously identified a stable post-hydrolytic closed state (C2 state), which can be locked-open by these non-hydrolyzable analogs with a slow rate. Here we captured, upon channel closing, another novel state, which distinguishes itself by its prompt response to PPi or AMP-PNP. Nonetheless, the locked-open time from this newly identified state is no different from that of locked-open channels from the C2 state, suggesting an identical locked-open configuration. Single-channel ligand-exchange experiments revealed an open-to-locked-open transition indicating that this new state with a vacated composite site 2 is an open state (O2 state). Although a [ATP]-dependent open time was not observed with wild-type CFTR due to a limited signal noise ratio, the open time for a conserved mutant, W401F-CFTR, increases with increasing [ATP] suggesting that ATP can also bind to the O2 state and go through another hydrolysis reaction within an opening burst - thus a violation of one-to-one stoichiometry between the gating cycle and the ATP hydrolysis cycle. Interestingly, for both WT- and W401F-CFTR, experimental data as well as computer simulations based on the new gating scheme we proposed show a bimodal distribution of the open time histograms with a paucity of short events. In conclusion, our results suggest that the gating signal is transmitted from NBDs to the gate with a delay and that TMDs and NBDs do not move synchronously during a gating cycle.