CFTR channel opening by ATP-driven tight dimerization of its nucleotide-binding domains

Abstract

ABC (ATP-binding cassette) proteins constitute a large family of membrane proteins that actively transport a broad range of substrates. Cystic fibrosis transmembrane conductance regulator (CFTR), the protein dysfunctional in cystic fibrosis, is unique among ABC proteins in that its transmembrane domains comprise an ion channel. Opening and closing of the pore have been linked to ATP binding and hydrolysis at CFTR's two nucleotide-binding domains, NBD1 and NBD2 (see, for example, refs 1, 2). Isolated NBDs of prokaryotic ABC proteins dimerize upon binding ATP, and hydrolysis of the ATP causes dimer dissociation. Here, using single-channel recording methods on intact CFTR molecules, we directly follow opening and closing of the channel gates, and relate these occurrences to ATP-mediated events in the NBDs. We find that energetic coupling between two CFTR residues, expected to lie on opposite sides of its predicted NBD1-NBD2 dimer interface, changes in concert with channel gating status. The two monitored side chains are independent of each other in closed channels but become coupled as the channels open. The results directly link ATP-driven tight dimerization of CFTR's cytoplasmic nucleotide-binding domains to opening of the ion channel in the transmembrane domains. This establishes a molecular mechanism, involving dynamic restructuring of the NBD dimer interface, that is probably common to all members of the ABC protein superfamily.

Figure 1

Open CFTR channels correspond to dimerized NBDs. a, Diagram illustrating the proposed mechanism coupling the opening of the Cl⁻ channel pore (C_n, closed states; O, open) in the transmembrane domains (converging, or semi-parallel, straight lines) to the hydrolysis cycle through the dimerization of NBDs (green, NBD1; blue, NBD2). The dynamic formation and disruption of a tight NBD dimer interface are represented by major changes in shape and position simply for clarity (see text). b, Mutating the ‘Walker B’ glutamate, Glu 1371, in NBD2 markedly increases the stability of the Cl⁻ channel’s open burst state. Records from patches containing hundreds of channels, activated by exposure to 5mM ATP and 300 nM cAMP-dependent protein kinase (PKA, red). Time constants for current decay fit lines (blue): WT, τ = 0.45 s; E1371Q, τ = 476 s. Note the fivefold expanded timescale for the WT record.

Figure 2

Statistical coupling analysis and electrophysiological recordings position CFTR’s Arg 555 in the composite NBD2 catalytic site. a, Representation of hydrogen-bonded arginine–threonine (R—T) and lysine–asparagine (K—N) pairs. In serine and threonine (the side chain present in CFTR) the acceptor oxygen atom is positioned at the same distance from the peptide backbone. b, Amino acid frequencies at the ‘head’ position, equivalent to Thr 1246, in alignment subsets having Arg (blue bars) or Lys (cyan bars) at the ‘tail’ position. The total alignment contained 10,194 sequences (http://www.sanger.ac.uk/cgi-bin/Pfam/getacc?PF00005). c, Mutations at Arg 555 in NBD1 ‘tail’ affect mean open burst (R555Q) and closed interburst (R555K) dwell times. R555Q opened with rates comparable to WT (R555Q, τ_ib = 2.84 ± 0.53 s (15)) despite expected loss of the interfacial hydrogen bond, suggesting possible compensation by other factors, for example the removal of repulsive electrostatic forces.

Figure 3

Arg 555 and Thr 1246 are not energetically coupled in channel closed states. a, Thermodynamic mutant cycle (each corner defined by the side chains at position 555 and 1246) using changes in apparent dissociation constant (K_0.5) as an estimate of the equilibrium constant for the reaction C₂⇌ C₁ + ATP (Fig. 1a). Values show changes (means ± s.d.) in free energy difference driving this reaction under standard conditions, ΔG°_{(C1+ATP) − C2} = −kTlnK _0.5 (k is Boltzmann’s constant, T is absolute temperature). b, Representative records from WT, single mutants R555K and T1246N, and double mutant R555K T1246N. Exposure to the test [ATP] (here 50 µM) was bracketed by exposures to 5mM ATP. c, Mean relative opening rates (±s.e.m., with 2 ≤ n ≤ 25).

Figure 4

Energetic coupling between Arg 555 and Thr 1246 accompanies channel opening. a, Thermodynamic cycle showing changes (means ± s.d.) in the activation energy barrier for opening (ΔΔG^‡). b, Representative records. c, Mean closed interburst duration (±s.e.m.). d, Thermodynamic cycle showing changes (means ± s.d.) in stability of open state with respect to the closed state, ΔΔG _{(open–closed)}, calculated from P _o. e, Representative records. Current levels of the triple mutant R555K T1246N K1250R did not change when [ATP] was increased to 10 mM, indicating that 5mM [ATP] was saturating. f, Mean P _o (±s.e.m.).