Site-specific ion occupation in the selectivity filter causes voltage-dependent gating in a viral K+ channel

Figure 1. Gating of KcvNTS. (a,b) Typical time series of current measured in symmetrical KCl solutions of (a) 100 mM and (b) 1500 mM. Closed “C” and apparent open level “A” are labelled. (c) Representative amplitude histograms in 100 (black) and 1500 mM KCl (grey), recorded at −120 mV. (d) Markov model of gating comprising one open state O, and three closed states (S = slow, M = medium, F = fast). The labels in (a) correspond to the states in (d).

Figure 2. Results from fitting extended beta distributions to amplitude histograms like those in Fig. 1c in symmetrical KCl concentrations of 500 mM. (a,b) Rate constants of the gating processes in Fig. 1d (c) IV curves of Iapp (black), IOF (open circles) and Itrue (grey). Mean and standard deviation from three different bilayers. If not visible, error bars are smaller than the symbols and below 25% for the rate constants (a,b) and below 1 pA for the currents (c), respectively.

Figure 3. Mutational studies for the identification of the location of the two different processes found in O-M gating. (a) Sequence of KcvNTS as compared to KcsA. Asterisks indicate conserved residues, colons and periods strongly and weakly similar properties, respectively. (b) Structural model of KcvNTS constructed with Swissmodel79 with KirBac1.1 (PDB 1P7B54) as a template and drawn with UCSF Chimera80. The positions of mutations used in this study are indicated in red and blue, respectively. (c) Representative current traces of the mutants KcvNTS G77S, G77T, and S42T measured in symmetrical 100 mM KCl at −120 mV (top) and +120 mV (bottom). C labels the closed state. (d,e) Amplitude histograms (grey) of KcvNTS (d) and KcvNTS S42T (e) at −160 mV. The fit by the model in Fig. 1d with extended beta distributions44 is given in black or red. (f,g) Voltage dependence of the rate constants (f) kOM and (g) kMO of KcvNTS wt (black), KcvNTS S42T (red), KcvNTS G77S (blue) and G77T (blue open symbols). For comparison, the rate constants for KcvNTS wt and NTS G77S in (f) and (g) have been taken from a previous publication14. Numbers of experiments = 3 (6 for KcvNTS wt). If the standard deviations…

Figure 4. Results of fitting ca. 1000 amplitude histograms from KcvNTS like those in Figs 1c and 3d from recordings in symmetrical and asymmetrical KCl solutions. Voltage and concentration dependence of those three parameters are shown, which describe the O-M gating: (a,b,c) current Itrue, (d,e,f) rate constant kOM of channel closure, and (g,h,i) rate constant kMO of channel opening. (a,d,g) Effect of symmetrical KCl concentrations ranging from 50 to 1500 mM. (b,e,h) Effect of internal K+: The external concentration was constant = 100 mM and internal concentration was varied. (c,f,i) Effect of external K+: internal KCl concentration was constant = 100 mM and external concentration was varied. 50 mM and 75 mM: green; 100 mM and 250 mM: black; 500 and 750 mM: orange; 1000 mM and 1500 mM: blue. The first concentration of a set of two with the same colour is given by closed circles, the second one by open circles. The inset in (e) shows the dependence on driving force Kmf = V − EK for internal K+ concentrations, indicated by the congruence of the curves when plotted over Kmf. Mean and standard deviations are obtained from 3–4 independent experiments for each concentration (6 for 100 mM…

Figure 5. Different ion hopping models of the permeation process in the selectivity filter. (a) 5-state model of Roux42 (b) 4-state model generated by removing the outer binding site of the model in (a) (c) Direct knock-on model43. S0 to S4 label the binding sites of K+ ions in the selectivity filter, as defined in the papers mentioned above. The numbers in (a,b) label the ion configurations as used in Eqs 1–4. In (c), the labels “letter/number” indicate the correspondence between the labels of the configurations of Köpfer et al.43 and those used in the Eqs S15 to S17 in the Supplementary material. In the model in (c), two reactions from the original scheme43 are merged according to the theory of reserve factors36 because they are not affected separately by the experimental conditions. All models in (a–c) are permeation models representing the open state. This is illustrated in (d) showing that the O-M gating in Fig. 1d occurs between the kinetic closed state M and the kinetic open (conducting) state O, which includes all permeation states of one of the models in (a–c). The rate constants in those cycles are in the range of 107 to 108 / s, much faster than the rate constants of O-M gating.…

Figure 6. Global fits (a–c) of the IV curves by Eq. 1 and (d–f) of ln(kOM) = ln(wP4/P3) (only at negative voltages as justified by Fig. 3f,g) by Eq. 6 for different (a,d) symmetrical, (b,e) internal and (c,f) external K+ concentrations. All curves have been fitted with the same set of parameters as given in the last column of Supplementary Table S2. The concentrations are indicated by the colours: 50 mM and 75 mM: green; 100 mM and 250 mM: black; 500 and 750 mM: orange; 1000 mM and 1500 mM: blue. Circles represent the data points, lines the theoretical curve obtained from the global fit. The first concentration of a set of two with the same colour is given by closed circles and a continuous line, the second one by open circles and a dashed line. Error bars results from three experiments (6 for symmetrical 100 mM). In the case of the IV curves, the error bars are so small (less than 1 pA) that they are mostly hidden behind the symbols. The current and rate constants for 100 mM, the current for 1000 mM and the rate constants for 1500 mM have been taken from a previous publication14.

Figure 7. Occupation probabilities P1 to P5 of the states of the model in Fig. 5a for symmetrical solutions of K+ (a,d,g,j,m), 100 mM external and internal concentration varied (b,e,h,k,n) and 100 mM internal and external concentration varied (c,f,i,l,o). The Pm are calculated from the rate constants kij of the transitions between state i and j (Eq. 4, Fig. 5a and Table S2) as obtained from global fitting of the IV curves (Eq. 1) and kOM = w P4/P3 (Eq. 6) as shown in Fig. 6. The colours give the K+ concentrations: green 50 and 75 mM, black 100 and 250 mM, orange 500 mM and 750 mM, blue 1000 and 1500 mM. In each pair with the same colour, the dashed lines give the higher concentration.