Voltage-controlled gating at the intracellular entrance to a hyperpolarization-activated cation channel

Abstract

Hyperpolarization-activated cation (HCN) channels regulate pacemaking activity in cardiac cells and neurons. Our previous work using the specific HCN channel blocker ZD7288 provided evidence for an intracellular activation gate for these channels because it appears that ZD7288, applied from the intracellular side, can enter and leave HCN channels only at voltages where the activation gate is opened (Shin, K.S., B.S. Rothberg, and G. Yellen. 2001. J. Gen. Physiol. 117:91-101). However, the ZD7288 molecule is larger than the Na(+) or K(+) ions that flow through the open channel. In the present study, we sought to resolve whether the voltage gate at the intracellular entrance to the pore for ZD7288 also can be a gate for permeant ions in HCN channels. Single residues in the putative pore-lining S6 region of an HCN channel (cloned from sea urchin; spHCN) were substituted with cysteines, and the mutants were probed with Cd(2+) applied to the intracellular side of the channel. One mutant, T464C, displayed rapid irreversible block when Cd(2+) was applied to opened channels, with an apparent blocking rate of approximately 3 x 10(5) M(-1)s(-1). The blocking rate was decreased for channels held at more depolarized voltages that close the channels, which is consistent with the Cd(2+) access to this residue being gated from the intracellular side of the channel. 464C channels could be recovered from Cd(2+) inhibition in the presence of a dithiol applied to the intracellular side. The rate of this recovery also was reduced when channels were held at depolarized voltages. Finally, Cd(2+) could be trapped inside channels that were composed of WT/464C tandem-linked subunits, which could otherwise recover spontaneously from Cd(2+) inhibition. Thus, Cd(2+) escape is also gated at the intracellular side of the channel. Together, these results are consistent with a voltage-controlled structure at the intracellular side of the spHCN channel that can gate the flow of cations through the pore.

Figure 1

Effects of Cd²⁺ on wild-type and S6 cysteine-substituted spHCN channels. Representative recordings from inside-out patches excised from HEK293 cells expressing wild-type (WT) or mutant spHCN before (o) or during (+) application of 20 μM intracellular Cd²⁺. Dashed lines represent zero current levels. Channels were held at +10 mV, and currents were elicited by a step to −110 mV, followed by a step to +30 mV. For 459C and D471C, channels were held at +50 mV and stepped to −90 mV, and then back to +50 mV. Currents were not leak-subtracted. Time scale bar (200 ms) applies to all currents except G461C (100 ms) and N465C (400 ms). Maximum inward currents for these traces: WT, 736 pA; F459C, 101 pA; I460C, 52 pA; G461C, 142 pA; H462C, 32 pA; T464C, 761 pA; N465C, 95 pA; L466C, 221 pA; I467C, 241 pA; Q468C, 826 pA; M470C, 178 pA; and S472C, 209 pA.

Figure 2

Cd²⁺ inhibits the 464C mutant rapidly only when the channels are open. Channels were held at +10 mV, and current was monitored using test pulses (400 ms) to −110 mV (closed circles). Cd²⁺ (20 μM, horizontal bar) was applied for 25 s to closed channels (at +10 mV), resulting in ∼10% inhibition of current, which was irreversible. The same concentration of Cd²⁺ was then applied for 2 s to open channels (during a 5-s pulse to −110 mV, arrow), resulting in ∼95% inhibition of current, which was also irreversible. Linear leak current was subtracted, and currents were normalized to the mean pre-Cd²⁺ control level.

Figure 3

Voltage dependence of the Cd²⁺ inhibition rate of 464C channels. Rate constants (closed squares) are plotted as a function of voltage (means of three to five experiments at each voltage; bars indicate ±SEM). Relative open probabilities (open circles) obtained from tail current measurements are shown for comparison (means of five experiments). The dotted line is a Boltzmann function fit to the relative open probabilities, assuming that open probability goes to zero with increasing depolarization. The dashed line is a Boltzmann function fit with the Cd²⁺ inhibition rates, with a minimum rate of ∼200 M⁻¹s⁻¹.

Figure 4

Cd²⁺ inhibition is reversible in channels containing only two cysteines at position 464. (A) Reversible Cd²⁺ inhibition in channels composed of TC tandem-linked subunit dimers (results). The TC/TC channels were ∼80% inhibited in 6 μM Cd²⁺, and the current recovered to near control levels in ∼2 min. The channels were ∼86% inhibited during a subsequent application of 20 μM Cd²⁺. (B) Cd²⁺ inhibition in patches containing a mixture of channels composed of TT and CC dimers (results). In this experiment, the initial application of 20 μM Cd²⁺ inhibited ∼90% of the current (CC/CC + TT/CC channels). Upon removal of Cd2+, ∼25% of the current recovered in ∼2 min (TT/CC channels). 65% of the current was blocked irreversibly (CC/CC channels). The recovered 25% could again be reversibly blocked by a subsequent application of 20 μM Cd²⁺. The remaining 10% of the current was unaffected by Cd²⁺ (mostly TT/TT channels). For both A and B, channels were held at +10 mV, and current was monitored using 400-ms test pulses to −110 mV (closed circles). Linear leak current was subtracted, and currents were normalized to the mean pre-Cd²⁺ control level.

Figure 5

Recovery of Cd²⁺-bound 464C channels is prevented by depolarization. 464C channels were held at +10 mV, and current was monitored using test pulses (400 ms) to −110 mV (closed circles). Channels were blocked with Cd²⁺ (20 μM, arrow) applied during a 5-s pulse to −110 mV. Little or no spontaneous recovery occurred with Cd²⁺ washout. DMPS (1 mM, horizontal bar) application was started while the channels were held at +10 mV. Little recovery was seen upon an initial test pulse in DMPS. Subsequent test pulses in DMPS speeded up recovery, and a long pulse (8 s) to −110 mV recovered the remaining blocked channels.

Figure 6

Voltage dependence of recovery rate for Cd²⁺-bound 464C channels in the presence of 1 mM DMPS. Recovery rates (closed squares) are compared with relative open probabilities (open circles) for 464C channels. Each rate represents the mean of three experiments at each voltage. The dotted line is a Boltzmann function fit to the relative open probabilities. There is little difference between Cd²⁺ blocked and unblocked 464C channels in the voltage dependence of gating.

Figure 7

Cd²⁺ can be trapped in closed TC/TC channels. Current from an inside-out patch containing TC/TC channels, which contain cysteines at 464 in two of four subunits (see results and Fig. 4 A). In the top trace, channels were held open at −90 mV and blocked with Cd²⁺ (20 μM, horizontal bar). Cd²⁺ was removed, and the current recovered over the next 30 s. A subsequent test pulse (500 ms at −90 mV) shows activation of the recovered channels. In the bottom trace, channels were held open at −90 mV and blocked with the same concentration of Cd²⁺. After allowing for a fast component of recovery (results), a brief test pulse (500 ms at −90 mV) was given to measure the initial level of block. We attempted to hold the channels closed at +10 mV for 30 s. A subsequent test pulse showed that almost no recovery occurred after the initial test pulse, indicating that Cd²⁺ had been trapped.