Charge movement in gating-locked HCN channels reveals weak coupling of voltage sensors and gate

Abstract

HCN (hyperpolarization-activated cyclic nucleotide gated) pacemaker channels have an architecture similar to that of voltage-gated K(+) channels, but they open with the opposite voltage dependence. HCN channels use essentially the same positively charged voltage sensors and intracellular activation gates as K(+) channels, but apparently these two components are coupled differently. In this study, we examine the energetics of coupling between the voltage sensor and the pore by using cysteine mutant channels for which low concentrations of Cd(2+) ions freeze the open-closed gating machinery but still allow the sensors to move. We were able to lock mutant channels either into open or into closed states by the application of Cd(2+) and measure the effect on voltage sensor movement. Cd(2+) did not immobilize the gating charge, as expected for strict coupling, but rather it produced shifts in the voltage dependence of voltage sensor charge movement, consistent with its effect of confining transitions to either closed or open states. From the magnitude of the Cd(2+)-induced shifts, we estimate that each voltage sensor produces a roughly three- to sevenfold effect on the open-closed equilibrium, corresponding to a coupling energy of ∼1.3-2 kT per sensor. Such coupling is not only opposite in sign to the coupling in K(+) channels, but also much weaker.

Figure 1.

Equilibrium models of coupling between voltage sensors and gating. (A) Standard allosteric model for coupling of four voltage sensors to simple channel opening. (B) Simplified version of A, with equilibrium constants summarized for fully activated (A) and deactivated (D) voltage sensors and open (O) and closed (C) channels. The diagram is heuristic and does not indicate a direct connection between C₀ and C₄ or between O₀ and O₄; the overall equilibrium constants shown between these extreme states are simply equal to the product of the intervening equilibrium constants. (C) Two approaches to estimating the coupling constant θ.

Figure 2.

Effect of Cd²⁺ on the 462C466C mutant or 462Y468C mutant. (A and B) Representative recordings from inside-out patches excised from Xenopus oocytes expressing either spHCN 462C466C (A) or spHCN 462Y468C (B) before and during the application of 130 nM free Cd²⁺. Channels were held at 10 mV, and currents were elicited by a step to −120 (A) or −130 mV (B), followed by a step to 60 mV. Currents were not leak subtracted. (C and D) To test whether the expressed channels still represent mutant phenotype in the condition of measuring the gating current (160 mM NMDA + 160 mM MeSO₃ for pipette and bath), a 160 mM K⁺ solution was applied rapidly to inside-out patches expressing spHCN 462C466C (C) or spHCN 462Y468C (D). The black trace in C indicates the response with NMDA on both sides and 130 nM free Cd²⁺, whereas the red trace shows the current response with the switch to intracellular K⁺ immediately before the voltage step from −120 to 60 mV; the sustained outward current indicates the locked-open effect. The gray arrow in C indicates the time when the solenoid was engaged; the solution switch at the patch occurred ∼50 ms later, and the actual switching speed was <1 ms. (D) The black trace shows a normal tail current for the locked-closed mutant in the absence of Cd²⁺ when intracellular K⁺ was applied immediately before the voltage step from −120 to 60 mV. The red trace again shows the response with fast perfusion of K⁺ solution, but in the constant presence of 130 nM free Cd²⁺, and the blue trace indicates the gating current alone (no K⁺ perfusion) with 130 nM free Cd²⁺ present.

Figure 3.

Gating current and Q-V relations for the spHCN 462C466C mutant. (A) Schematic description of the locked-open mutant. The state of the channel with Cd²⁺ is shaded, whereas the control condition without Cd²⁺ is in the dashed area. (B and C) Traces of gating current records for spHCN 462C466C without (B) or with 130 nM free Cd²⁺ (C), from the same patch. Holding potential was 10 mV, and test pulses were from −30 to −170 mV in 10-mV increments. A −P/4 protocol was used to subtract leak and linear capacitive currents. (D and E) Normalized Q-V relations for spHCN 462C466C without Cd²⁺ (open circles) or with Cd²⁺ (closed circles). Both Q_ON (D) and Q_OFF (E) were obtained by integrating components. Smooth curves are single Boltzmann fits to the data with these parameters: Q_ON control: V_1/2 = −132.7 ± 2.1 mV, slope factor (e-fold) = 16.5 ± 1.2; Cd²⁺: V_1/2 = −103.9 ± 5.3 mV, slope factor (e-fold) = 18.2 ± 1.2; n = 5. Q_OFF control: V_1/2 = −126.7 ± 5.1 mV, slope factor (e-fold) = 21.1 ± 1.2; Cd²⁺: V_1/2 = −99.9 ± 5.3 mV, slope factor (e-fold) = 19.5 ± 1.1; n = 9. Error bars represent SEM.

Figure 4.

Gating current and Q-V relations for the spHCN 462Y468C mutant. (A) Schematic description of the locked-closed mutant. The state of the channel with Cd²⁺ is shaded, whereas the control condition without Cd²⁺ is in the dashed area. (B and C) Traces of gating current records for spHCN 462Y468C without (B) or with Cd²⁺ (C), from the same patch. Holding potential was 10 mV, and test pulses were from −30 to −170 mV in 10-mV increments. A −P/4 protocol was used to subtract leak and linear capacitive currents. (D and E) Normalized Q-V relations for spHCN 462Y468C without (open squares) or with Cd²⁺ (closed squares). Both Q_ON (D) and Q_OFF (E) were obtained by integrating components. Smooth curves are single Boltzmann fits to the data with these parameters as follows: Q_ON control: V_1/2 = −116.4 mV ± 8.4, slope factor (e-fold) = 21.2 ± 4.7; 100 µM cAMP + 130 nM free Cd²⁺: V_1/2 = −131.8 ± 7.5 mV, slope factor(e-fold) = 18.1 ± 0.9; n = 5. Q_OFF control: V_1/2 = −96.7 ± 5.3 mV, slope factor (e-fold) = 16.0 ± 1.9; Cd²⁺: V_1/2 = −128.0 ± 2.3 mV, slope factor (e-fold) = 21.1 ± 2.0; n = 7. Error bars represent SEM.

Figure 5.

Comparison between the experimental Q-V and G-V characteristics of mutant channels and their respective fitting results with the 10-state allosteric model. (A) 462C466C; (B) 462Y468C. (A) Q-V (open and closed circles) and G-V (triangles) plots from spHCN 462C466C mutant channels (from Fig. 3 A), overlaid with the best fit from the 10-state model (black, red, and green lines; see parameter values in Table 2). (B) Q-V (open and closed squares) and G-V (triangles) plots recorded from 462Y468 mutant channels (from Fig. 4 A), overlaid with the best fit from the 10-state model (black, red, and green lines; see parameter values in Table 2). (A and B) The dotted outline shown in the simple four-state model on the right represents the states of the channel without Cd²⁺, whereas the shaded outline represents the states of the channel with Cd²⁺. Error bars represent SEM.