Mutation of Walker-A lysine 464 in cystic fibrosis transmembrane conductance regulator reveals functional interaction between its nucleotide-binding domains

Abstract

The cystic fibrosis transmembrane conductance regulator (CFTR) chloride channel bears two nucleotide-binding domains (NBD1 and NBD2) that control its ATP-dependent gating. Exactly how these NBDs control gating is controversial. To address this issue, we examined channels with a Walker-A lysine mutation in NBD1 (K464A) using the patch clamp technique. K464A mutants have an ATP dependence (EC(50) approximate 60 microM) and opening rate at 2.75 mM ATP (approximately 2.1 s(-1)) similar to wild type (EC(50) approximate 97 microM; approximately 2.0 s(-1)). However, K464A's closing rate at 2.75 mM ATP (approximately 3.6 s(-1)) is faster than that of wild type (approximately 2.1 s(-1)), suggesting involvement of NBD1 in nucleotide-dependent closing. Delay of closing in wild type by adenylyl imidodiphosphate (AMP-PNP), a non-hydrolysable ATP analogue, is markedly diminished in K464A mutants due to reduction in AMP-PNP's apparent on-rate and acceleration of its apparent off-rate (approximately 2- and approximately 10-fold, respectively). Since the delay of closing by AMP-PNP is thought to occur via NBD2, K464A's effect on the NBD2 mutant K1250A was examined. In sharp contrast to K464A, K1250A single mutants exhibit reduced opening (approximately 0.055 s(-1)) and closing (approximately 0.006 s(-1)) rates at millimolar [ATP], suggesting a role for K1250 in both opening and closing. At millimolar [ATP], K464A-K1250A double mutants close approximately 5-fold faster (approximately 0.029 s(-1)) than K1250A but open with a similar rate (approximately 0.059 s(-1)), indicating an effect of K464A on NBD2 function. In summary, our results reveal that both of CFTR's functionally asymmetric NBDs participate in nucleotide-dependent closing, which provides important constraints for NBD-mediated gating models.

Figure 1. ATP dose-response relationships for CFTR-K464A

A, representative trace of macroscopic CFTR-K464A channel current stimulated by 2.75 mm and 25 μm ATP after steady-state activation by PKA phosphorylation (not shown). B, trace of CFTR-K464A channels exposed to different [ATP] after steady-state activation by PKA and ATP. C, macroscopic dose-response relationship for CFTR-K464A (▪; present study) and wild type CFTR (□; data taken from Zeltwanger et al. 1999). D, open probability versus [ATP] for CFTR-K464A (•; present study) and wild type (○; data taken from Zeltwanger et al. 1999). The maximum P_o for CFTR-K464A at 2.75 mm ATP is ∼ 0.37. E, superimposition of macroscopic (□) and microscopic (•) data shown in C and D. Arrows in A and B indicate the baseline current level (all channels closed) and downward deflections are channel openings. Numbers in parentheses indicate the number of experiments per data point. Dashed lines are fits of the Michaelis-Menten equation to the CFTR-K464A data (see Methods).

Figure 3. Dependence of CFTR-K464A closing and opening rates on [ATP]

A, plot of mean opening rates (reciprocals of mean closed times; •) and closing rates (reciprocals of mean opened times; □) versus[ATP] for CFTR-K464A. Dashed line is a fit of the Michaelis-Menten equation to the opening rate data (see Methods). B, comparison of opening and closing rates for CFTR-K464A and wild type at 2.75 mm ATP. Asterisks indicate a significant difference between closing rates for wild type and K464A (P < 0.005).

Scheme 1

Figure 2. Single-channel kinetics of CFTR-K464A

A, representative sweeps from experiments with single CFTR-wild type (left-hand sweeps; taken from Zeltwanger et al. 1999) and CFTR-K464A channels (right-hand sweeps) exposed to 2.75 mm (top sweeps) and 100 μm MgATP (bottom sweeps) subsequent to activation by PKA (40 U ml⁻¹) and ATP (2.75 mm). Arrows indicate the baseline and downward deflections indicate openings of the channel. B, survivor plots of closed (left-hand panels) and open dwell times (right-hand panels) at 2.75 mm (top panels) and 100 μm MgATP (bottom panels) for the single CFTR-K464A channel shown in A (cf. wild type distributions in Zeltwanger et al. 1999). Dashed lines indicate single exponential fits to the data; time constants (τ) for the fits are indicated.

Figure 4. AMP-PNP weakly locks open CFTR-K464A

A, representative sweeps from experiments with wild type (top trace) and CFTR-K464A single channels (bottom trace) exposed first to PKA (40 U ml⁻¹) and MgATP (250 μm), then with the addition of AMP-PNP (1 mm). Arrows indicate the baseline, downward deflections channel openings. B, survivor plot of open dwell times for CFTR-K464A. Dashed line is a double-exponential fit to the data. Time constants and their relative weights from the fit are indicated.

Figure 5. K464A enhances the dissociation of AMP-PNP

A, representative traces from experiments with macroscopic currents from CFTR-wild type (top trace) and CFTR-K464A channels (bottom trace) exposed to PKA (40 U ml⁻¹), MgATP (250 μm) and AMP-PNP (1 mm). Current relaxations upon withdrawal of PKA, ATP and AMP-PNP were fitted with single exponential curves to estimate mean open time (see Methods). For the traces shown, the mean relaxation time constant for wild type is 64.8 ± 0.1 s and for CFTR-K464A is 9.1 ± 0.1 s. B, mean relaxation time constants (± s.e.m.) for wild type and K464A channels. Asterisks indicate a significant difference between wild type and K464A (P < 0.01).

Figure 6. K464A reduces the apparent on-rate of AMP-PNP

A, Semilog plot of a probability function determined by measuring the cumulative time channels spent in the O state before entering the L state (see Methods). The dwell times are from individual experiments for CFTR-K464A (○) and CFTR-wild type (□). Dashed lines are single exponential fits for estimating the pseudo-first order locking rate (i.e. the O-L transition in Scheme 1). B, comparison of the mean locking rates for CFTR-wild type and CFTR-K464A. Asterisk indicates a significant difference between wild type and K464A (P < 0.05).

Figure 7. K464A shortens K1250A relaxation

A, representative trace of macroscopic current relaxations from CFTR-K1250A (top trace) and CFTR-K464A-K1250A double mutant channels upon withdrawal of PKA (40 U ml⁻¹) and ATP (1 mm). Mean relaxation time constant for the CFTR-K1250A trace shown is 110 ± 1 s and for CFTR-K464A-K1250A is 30 ± 1 s. B, few-channel traces of CFTR-K1250A and CFTR-K464A-K1250A at the steady state in 2.75 mm ATP. Dashed lines indicate baseline (all channels closed); marks at the left indicate open channel current levels (a total of 4 channels for K1250A and 3 for K464A-K1250A). C, comparison of steady-state P_o, mean open (relaxation) times and mean closed times for CFTR-K1250A and CFTR-K464A-K1250A. Asterisks indicate significant differences between CFTR-K1250A and CFTR-K464A-K1250A (**P < 0.01; ***P < 0.005).

Figure 8. K464A-K1250A gating at millimolar and micromolar [ATP]

A, representative sweep from experiments with CFTR-K464A-K1250A channels exposed first to 1 mm and then to 10 μm MgATP. Arrow indicates the baseline, downward deflections channel openings. B, survivor plot of open dwell times for CFTR-K464A-K1250A at 10 μm MgATP. Dashed line is a single-exponential fit to the data. Time constant for the fit is 241 ± 3 s (220 events pooled from 3 patches).