Gating currents from a Kv3 subfamily potassium channel: charge movement and modification by BDS-II toxin

Abstract

Kv3 channels have a major role in determining neuronal excitability, and are characterized by ultra-rapid kinetics of gating and a high activation threshold. However, the gating currents, which occur as a result of positional changes of the charged elements in the channel structure during activation, are not well understood. Here we report a study of gating currents from wild-type Kv3.2b channels, expressed in human embryonic kidney (HEK) cells to facilitate high time-resolution recording. On-gating currents (I_g,on) had extremely rapid kinetics such that at +80 mV, the time constant for the decay of I_g,on was only ∼0.3 ms. Decay of I_g,on appeared mono-exponential at all potentials studied, and in support of this, the charge–voltage (Q–V) relationship was fitted with a single Boltzmann function, supporting the idea that only one charge system is required to account for the time course of I_g,on and the voltage dependence of Q_on. The voltage (V_½) for half movement of gating charge was −8.4 ± 4.0 mV (n = 6), which closely matches the voltage dependence of activation of Kv3.2b ionic currents reported before. Depolarizations to more positive potentials than 0 mV decreased the amplitude and slowed the decay of the off-gating currents (I_g,off), suggesting that a rate-limiting step in opening was present in Kv3 channels as in Shaker and other Kv channels. Return of charge was negatively shifted along the potential axis with a V_½ of Q_off of −80.9 ± 0.8 mV (n = 3), which allowed ∼90% charge return upon repolarization to −100 mV. BDS-II toxin apparently reduced I_g,on, and greatly slowed the kinetics of I_g,on, while shifting the Q–V relationship in the depolarizing direction. However, the Q–V relationship remained well fitted by a single Boltzmann function. These data provide the first description of Kv3 gating currents and give further insight into the interaction of BDS toxins and Kv3 channels.

Many neuronal types throughout the mammalian CNS, such as interneurones, are able to fire action potentials at high frequencies or follow high frequency inputs. Kv3 potassium channels have been identified as crucial components of the neuronal circuits capable of producing such responses, with their high activation threshold and remarkably fast activation and deactivation properties (Rudy et al. 1999; Rudy & McBain, 2001). These biophysical gating properties mirror those of certain endogenous neuronal K⁺ currents (Brew & Forsythe, 1995; Perney & Kaczmarek, 1997; Southan & Robertson, 2000; Lien & Jonas, 2003) and thereby suggest that Kv3 channels can confer a ‘fast spiking’ phenotype to many CNS neurons (Wang et al. 1998; Martina et al. 1998; Southan & Robertson, 2000; Lien & Jonas, 2003). Some previous gating current recordings suggest that fundamental differences in the kinetics of voltage-dependent charge movement might underlie the rapid kinetics of Kv3 ionic currents. Comparison between Kv2.1 and Kv3.1 gating currents recorded using macropatch methods in Xenopus oocytes demonstrated a remarkable acceleration of the kinetics of charge movement during activation and deactivation in Kv3.1 channels, such that the onset of on-gating currents (I_g,on) could not be well resolved by patch clamp recording (Shieh et al. 1997).

In addition, a reduction in the amplitude and slowing in the kinetics of off-gating currents (I_g,off) after larger depolarizations, that is a general phenomenon reported in Shaker and Kv channel gating current studies (Perozo et al. 1992, 1993; Chen et al. 1997; Yang et al. 1997; Wang et al. 1999; Wang & Fedida, 2001), was not present in the Kv3.1 gating current recordings (Shieh et al. 1997). Rapid kinetics of gating charge movement were also apparent in previous gating current recordings from HEK cells expressing a high density of Kv3.2b P468W mutant channels, as was a lack of gating charge slowing on repolarization (Yeung et al. 2005). Although Kv3 channels apparently show these differences in the gating kinetics from Shaker and Kv1 channels, the molecular mechanisms have not been fully explored, partly due to the difficulty in making such measurements.

BDS (blood-depressing substance) toxins are sea anemone peptides from Anemonia sulcata (Diochot et al. 1998) which inhibit Kv3 channels with high affinity when applied to the extracellular face of the membrane (Diochot et al. 1998; Yeung et al. 2005). BDS toxins bound to the S3b and S4 segments hinder voltage sensor movement (Yeung et al. 2005), and shift the voltage dependence of activation of Kv3 channels to more depolarizing potentials, analogous to the interaction of hanatoxin and Kv2.1 channels (Swartz & MacKinnon, 1997a, b; Lee et al. 2003). Unlike pore-blocking toxins that directly interfere with ion conduction (Miller, 1995), BDS toxins may inhibit Kv3 channels by modifying the energetics of activation (Swartz & MacKinnon, 1995; Lee et al. 2003). This mechanism is supported by our prior gating current study of Kv3.2b P468W mutant channels where we demonstrated that the charge–voltage relationship is shifted +25 mV by 1 μm BDS-II toxin, which matches the shift of the voltage dependence of activation (Yeung et al. 2005).

Here we have characterized the gating currents of Kv3.2b wild-type channels at high time resolution in order to compare these data with the accelerated kinetics of gating current and the lack of gating charge slowing observed in Kv3.1 channels expressed in Xenopus oocytes (Shieh et al. 1997), and in Kv3.2b P468W channels expressed in HEK cells (Yeung et al. 2005). The successful recording of gating currents has allowed the measurement of the voltage dependence and the kinetics of gating charge movement during activation and deactivation, and provided the first systematic description of Kv3 gating currents. Moreover, changes in the kinetics of gating charge movement as well as the voltage dependence of charge distribution after application of BDS-II offer insight into interaction of the channel proteins and the toxin.

Methods

Molecular biology and cell culture

The mammalian expression vector pcDNA3 was used for expression of rat Kv3.2b channels (Luneau et al. 1991) and was sequenced to check for errors before being used in transient transfections of HEK 293 cells. The cells were grown in minimal essential medium (MEM) with 10% fetal bovine serum, at 37°C in an air–5% COB₂ incubator. For transient transfection, HEK cells were plated at 20–30% confluence on sterile glass coverslips in 25 mm Petri dishes overnight. Kv3.2b channel DNA was incubated with enhanced green fluorescent protein (eGFP) in order to identify the transfected cells efficiently (2 μg of eGFP, 2 μg of channel DNA) and 3 μl of Lipofectamine 2000 (Invitrogen) in 100 μl of serum-free media for 30 min, and then added to the dishes containing HEK 293 cells in 2 ml of MEM with 10% fetal bovine serum. After 5 h incubation, the culture medium was changed and then cells were incubated overnight before recording. Cells that expressed GFP were selected for patch clamp experiments.

Electrophysiology

Coverslips containing adherent cells were removed from the incubator before experiments and placed in a superfusion chamber (volume 250 μl) containing the control bath solution at ambient temperature (22–23°C). The bath solution was exchanged by switching the perfusates at the inlet of the chamber, with complete bath solution changes taking 5–10 s. Whole-cell current recording and data analysis were done using an Axopatch 200B amplifier and pCLAMP 8 software (Axon Instruments, Union City, CA, USA). Patch electrodes were fabricated using thin-walled borosilicate glass (World Precision Instruments, FL, USA), routinely coated with Sylgard (Dow Corning) and fire polished to reduce electrode capacitance and improve seal resistance. Electrodes had resistances of 1–3 MΩ when filled with the filling solutions. Capacity compensation was routinely used in all whole-cell recordings. Measured series resistance was between 1 and 3 MΩ for all recordings. When the series resistance changed during the course of an experiment, data were discarded. In gating current recording a ‘P/−6’ protocol was routinely used for the online subtraction of the leakage and capacitive currents. The potential used for delivery of leak subtraction pulses was −90 to −110 mV. Data were filtered using a 4-pole Bessel filter with a cut-off frequency (f_c) of 10 kHz, and sampled at 100 kHz. For some experiments, as indicated in the text and figure legends, the gating current was recorded using 60% series resistance compensation, a 100 kHz filter and a 330 kHz sampling rate. Membrane potentials have not been corrected for small junction potentials that arose between bath and pipette solutions.

All charge measurements (Q_on and Q_off) were obtained by integrating the on- (I_g,on) and off-gating (I_g,off) current over sufficient time to allow gating currents to return to the baseline. The charge–voltage (Q–V) relationships were determined from Q_on or Q_off elicited during or after 12 ms pulses to a variable test potential. The Q_on or Q_off for each test potential was normalized to the maximum Q_on or Q_off, respectively. The voltage dependence of gating charge movement was determined by fitting a single Boltzmann function to the normalized gating charge (Q) values. The equation used to fit the data was: Q =Q_max/{1 + exp [−(V−V_½)/k]}, where Q_max is the maximum gating charge, V the membrane potential, V_½ the voltage for half charge movement, and k the slope factor.

The time course of the decaying phase of on- and off-gating currents was fitted with a single exponential function: a + b exp(−x/t), where a represents an offset, b is the current amplitude and t is the time constant. During the present study we report time constants of off-gating current decay (τ_off) of ∼0.8 ms, much slower than the mean capacity transient decay rates 58 ± 0.3 μs, so that recording bandwidth does not appear to be a limiting factor in the measurements. Throughout the text data are presented as mean ±s.e.m.

For recordings of gating currents, patch pipettes contained (mm): N-methyl d-glucamine (NMG⁺), 140; MgCl₂, 1; EGTA, 10; Hepes, 10; and was adjusted to pH 7.2 with HCl. The bath solution contained (mm): tetraethylammonium (TEA⁺), 140; Hepes, 10; dextrose, 10; MgCl₂, 1; CaCl₂, 1; and was adjusted to pH 7.4 with HCl. All chemicals were from Sigma Aldrich Chemical Co. BDS-II toxin, isolated from Anemonia sulcata, was purchased from Alomone Laboratories (Jerusalem, Israel). BDS-II was applied externally via bath superfusion.

Results

Gating currents from HEK cells expressing Kv3.2b channels

Rat Kv3.2b channels (Luneau et al. 1991) were transiently expressed in HEK cells and the non-permeant cations, NMG⁺ and TEA⁺, were used in the intracellular and extracellular solution, respectively, to replace permeant K⁺ and Na⁺ ions. Given that Kv3 channels are characterized by ultra-rapid kinetics of gating (Rudy et al. 1999; Rudy & McBain, 2001), the initial experiments examined the properties of the recording system. A typical uncorrected capacity transient recorded from an untransfected HEK cell during a 150 mV step pulse decayed with a time constant of 54 μs (58 ± 3 μs, n = 5; Fig. 1A), similar to values reported before (Hesketh & Fedida, 1999). Data recorded from a HEK cell expressing Kv3.2b channels are shown in Fig. 1B. The cell was clamped at −100 mV and depolarized to +80 mV, the leak current was subtracted using a standard (P/−6) protocol, and the data were sampled at 100 kHz through a 10 kHz f_c filter (Wang et al. 1999; Wang & Fedida, 2001, 2002). The continuous black line is the I_g,on during the depolarization, and has a time to peak (t_peak) of 240 μs. The grey line shows the current after application of 60% series resistance compensation, and final filtering at 20 kHz. This reduced t_peakto < 170 μs, but did not significantly affect the amount of charge moved during the depolarization or the time course of I_g,on decay to the baseline, as seen from the two current recordings (Fig. 1B, n = 4).

Figure 1.

Recording gating currents from HEK cells expressing Kv3.2b channels A, an uncorrected capacity transient from an untransfected HEK cell during a voltage step from −100 mV to +50 mV as shown at the top. A monoexponential decay function is superimposed as a dotted line (τ= 54 μs). B, on-gating current tracings (I_g,on) from Kv3.2b channels elicited by the voltage pulse protocol above. The cell was held at −100 mV and depolarized to +80 mV, and currents were recorded with both capacity compensation and leak current subtraction as described in Methods. Grey: a current trace recorded using 60% series resistance compensation, with a 100 kHz f_c filter at a 330 kHz sampling frequency. For presentation, the data were further filtered at 20 kHz offline. Black: without series resistance compensation, filtered at 10 kHz, and sampled at 100 kHz. Both current tracings were obtained from the same cell. C and D, gating currents recorded from the same cell before (C), and after (D) the application of leak-current subtraction. The pulse protocol used to elicit gating currents is shown at the top of C. The cell was held at −100 mV and depolarized at 0.5 Hz to potentials between −80 and +40 mV in 20 mV steps. The short dashed lines denote the zero current level in this and subsequent figures.

Current records in Fig. 1C and D show Kv3.2b gating currents before and after P/−6 online leak subtraction, respectively. It can be seen that the small-maintained current component present during the depolarization and repolarization in Fig. 1C was reduced. The transient current components were otherwise little changed by the leak subtraction protocol.

The kinetics of gating currents recorded from Kv3.2b channels

On-gating currents appeared as transient outward current deflections during depolarizations positive to −40 mV which increased in amplitude with larger depolarizations, and decayed more quickly (Fig. 2A). The I_g,on exhibited a rising phase that was barely visible for small depolarizations but became pronounced at larger depolarizations, similar to that seen in Shaker and other Kv channels (Bezanilla et al. 1991, 1994; Perozo et al. 1992; Chen et al. 1997). Clearly, the on-gating currents decayed very rapidly during depolarizing pulses. At +80 mV the I_g,on decayed to the zero current level in only ∼1.2 ms, a much briefer transient current than that reported in Shaker and other Kv channels, such as Kv1.5 (Wang et al. 1999; Claydon et al. 2007). When pulsed back to −100 mV, the off-gating current (I_g,off) represents the return of gating elements as channels deactivate. In Fig. 2A, the I_g,off are seen as downward current deflections after depolarizing pulses. In contrast to I_g,on, the I_g,off time course was similar to that described in Shaker and Kv channels. After small depolarizations (at potentials < 0 mV), the charge return was fast and the I_g,off increased in amplitude with the enhancement of depolarization. For intermediate depolarizations, I_g,off had a fast component followed by a smaller slower component. At depolarizations > +20 mV, the peak of I_g,off became smaller, with a rising phase and a very slow decay. However, the rising phase of I_g,off after the larger depolarizations was not as prominent as in Shaker (Bezanilla et al. 1994) or Kv1.5 channels (Wang et al. 1999).

Figure 2.

Measurement of the gating currents from Kv3.2b channels A, current tracings recorded from a HEK cell expressing Kv3.2b channels. The cell was held at −100 mV and pulsed to the potentials as indicated above the current trace. B, time course of the gating charge movement in Kv3.2b channels during depolarization (top) to potentials indicated and repolarization (bottom). The amount of charge displaced during depolarization and repolarization (Q_on and Q_off) was obtained from the time integral of the I_g,on and I_g,off shown in A. C, averaged ratio of Q_off/Q_on as a function of the pulse potentials, from data as in A (n = 8). D, time constants (τ_off) for the relaxation of I_g,off plotted against the pulse potentials, from data as in A (n = 8). A mono-exponential function was used to fit the relaxation of I_g,off except for depolarizations between −10 mV and +10 mV, where the relaxation phase of I_g,off was best fitted with a bi-exponential function.

The time- and voltage-dependent movement of gating charge (Q_on) as channels progress towards the open state, and the subsequent relaxation of charge upon channel closing (Q_off) are shown in Fig. 2B. As expected, Q_on increased in amplitude and moved faster at +80 than 0 mV, but for Q_off, although the amplitude was larger at +80 mV, less change was found in the already slowed kinetics. The averaged ratio of Q_off over Q_on was close to 1.0 over the range of potentials from −40 mV to −10 mV and ∼0.9 when the potential was 0 mV and above (Fig. 2C). The relaxation of I_g,off (τ_off) was mono-exponential between −40 and −20 mV and from +20 to +80 mV. Over the range of intermediate depolarizations (from −10 to +10 mV), the relaxation of I_g,off was fitted with a bi-exponential function (Fig. 2D). The relaxation of off-gating current became significantly slower with increasing depolarization, as τ_off increased from ∼0.7 ms to ∼5 ms over the range of potentials studied (Fig. 2D). These results are similar to those from wild-type and pore mutants of Shaker and Kv channels (Perozo et al. 1992, 1993; Chen et al. 1997; Yang et al. 1997), where Q_off is smaller and the charge return becomes dramatically slowed when the depolarization crosses the threshold of activation, but clearly different from the previously reported gating currents of Kv3.1 expressed in oocytes (Shieh et al. 1997) and Kv3.2b P468W mutant channels expressed in HEK cells (Yeung et al. 2005) that lack this slowing.

Voltage dependence of the gating charge movement in Kv3.2b channels

The voltage dependence of gating charge movement during activation and deactivation in Kv3.2b channels was analysed as shown in Fig. 3. The cell was held at −150 mV and depolarized to potentials between −60 and +80 mV for 12 ms to record I_g,on. Then the membrane potential was stepped back to −150 mV and the I_g,off was also recorded (Fig. 3A). The Q_on was normalized to the maximum value and plotted as a function of potential in Fig. 3C (filled circles). It is apparent that the charge distribution is symmetrical in Kv3.2b channels, and that the Q_on–V relationship can be fitted with a single Boltzmann function. The voltages for half movement of the gating charge (V_½) and the slope factor (k) were −8.4 ± 4.0 mV and 12.2 ± 1.6 mV (n = 6), respectively. At −150 mV, the return of gating charge was rapid and complete in 10 ms (Fig. 3A), and when the normalized Q_off values were fitted with a Boltzmann function, the parameters were almost identical to the Q_on–V curve data (Fig. 3C, filled triangles). The V_½ and slope factor (k) were −6.0 ± 3.5 mV and 13.8 ± 1.3 mV (n = 6), respectively. This symmetrical distribution of the gating charge is unlike that found in Shaker and Kv1.5 channels, where the charge distribution is asymmetrical, and usually fitted to a double-Boltzmann function that indicates the existence of two activating charge systems (Bezanilla et al. 1994; Hesketh & Fedida, 1999). In addition, these data indicate that the voltage dependence of the gating charge movement during depolarization can be analysed equivalently by the measurement of Q_on during depolarization or Q_off at very negative repolarizing potentials.

Figure 3.

Voltage dependence of gating charge movement in Kv3.2b channels A, gating currents recorded during 12 ms depolarizations from −150 mV to potentials between −60 mV and +80 mV in 20 mV steps applied at 0.5 Hz. B, gating currents recorded after depolarizations to +80 mV upon repolarization to potentials between −40 mV and −150 mV in 10 mV steps (protocol at top), pulses applied at 0.5 Hz. C, top, voltage dependence of gating charge movement during depolarization. The Q_on and Q_off from the experiments shown in A were normalized to the values obtained with a +80 mV depolarization. The normalized Q_on (•) and Q_off (▴) was plotted against the pulse potential and fitted with a single Boltzmann function (n = 6). The V_½ was −8.4 ± 4.0 mV and −6.0 ± 3.5 mV and slope factor (k) was 12.2 ± 1.6 mV and 13.8 ± 1.3 mV, for Q_on and Q_off, respectively. Bottom, voltage dependence of the charge movement during repolarization. The Q_off from the experiments as shown in B was normalized to the value obtained at −150 mV. The normalized Q_off (▪) was plotted against the membrane potential and fitted with a single Boltzmann function (n = 3) with a V_½ of −80.9 ± 0.8 mV and a slope factor (k) of 8.7 ± 0.4 mV. D, time constants (τ_off) for the relaxation of I_g,off plotted as a function of repolarizing potential, from data as in B (n = 6). The relaxation phase of I_g,off was fitted with a mono-exponential function.

The voltage dependence of charge movement during repolarization was measured using the protocol illustrated in Fig. 3B. After a depolarization to +80 mV for 12 ms, to fully move the gating charge associated with channel activation (cf. Fig. 3C), the membrane was repolarized to various potentials between −150 and −40 mV for long enough to allow the current decay to reach an apparent baseline. The Q_off was obtained by integrating the I_g,off at these different potentials and, after normalization, was plotted in Fig. 3C (bottom panel, filled squares) and fitted with a Boltzmann function. The Q_off–V curve was shifted to hyperpolarizing potentials compared with the charge relationships plotted in the upper panel, and this was caused by a failure of complete charge return at potentials positive to −100 mV during the repolarizing pulse, due to the voltage-dependent kinetics of I_g,off. The time constant of decay (τ_off) increased from ∼0.8 ms at −190mV to ∼5 ms at −70 mV (Fig. 3D). This can be clearly seen in the current records in Fig. 3B, where, at the end of the protocol, a return to the holding potential of −150 mV elicited small transient off-gating currents as the remaining charge was allowed to return to rest.

Kinetics of the gating charge movement upon depolarization in Kv3.2b channels

On-gating currents measured from Shaker channels decay in a double exponential manner at intermediate depolarizations (Bezanilla et al. 1994). In Kv1.5 channels, I_g,on also shows complex kinetics, having more rounded peaks and plateaus at the intermediate depolarizations (Hesketh & Fedida, 1999). Separation of the gating currents into two charge systems fitted with a double Boltzmann distribution, and subsequent modelling allowed the waveforms to be well described in both systems. When Kv3.2b gating currents were inspected (Fig. 2A), however, it was apparent that there was only a single decay phase to I_g,on regardless of the amplitude of depolarization. The current data in Fig. 4A show mono-exponential fits to the decaying phase of Kv3.2b I_g,on at small (−10 mV), intermediate (0 and +20 mV) and large depolarizations (+60 mV). The time constants are plotted against voltage in Fig. 4B, and between −30 mV and +20 mV the currents tend to decay more slowly with time constants between 1 and 2 ms, whereas for larger depolarizations (> +20 mV) the current decay becomes much faster. So that, in addition to the monophasic nature of the decay of I_g,on, the data also show that Kv3.2b gating charge moves quite rapidly, especially during larger depolarizations (Shieh et al. 1997; Yeung et al. 2005). For example, at +80 mV, the time constant for the decaying phase of I_g,on is ∼0.3 ms (Fig. 4B). The slowest charge movement occurs around 0 mV with a time constant of ∼2 ms.

Figure 4.

Kinetic properties of I_g,on in Kv3.2b channels A, I_g,on at −10, 0, +20 and +60 mV from a holding potential of −100 mV. Capacity compensation, leak subtraction and 60% series resistance compensation were applied, and current was sampled at 330 kHz after filtering at 100 kHz. For presentation, the data were further filtered at 20 kHz offline. Superimposed continuous lines are mono-exponential fits to the decaying phase of I_g,on. B, averaged time constants (τ_decay) for the decay of I_g,on as a function of pulse potential (n = 9). C, the Q–V curve obtained by fitting the integral of the exponential decay of I_g,on to a single Boltzmann function. The V_½=−8.2 ± 1.2 mV, and the slope factor (k) = 6.7 ± 0.7 mV (n = 4).

The charge distribution was also estimated from these data by multiplying the fitted amplitude by the time constants (Bezanilla et al. 1994) to obtain the relationship plotted in Fig. 4C. The experimental points can be well fitted to a single Boltzmann function, with a V_½ and k of −8.2 ± 1.2 mV and 6.7 ± 0.7 mV (n = 4), respectively. The values are close to those obtained by integration of the entire transients (Fig. 3C) which supports the validity of the exponential analysis in Fig. 4A, and the idea of a single activating charge system in Kv3.2b gating.

Modulation of Kv3.2b gating currents by BDS-II toxin

As BDS toxins bind to the S3b and S4 segments in the Kv3 channel α-subunit (Yeung et al. 2005), the movement of ‘voltage sensor’ domains during activation and deactivation should be directly affected, and potentially could give further insight into the mechanisms of Kv3.2b gating. Gating currents recorded from Kv3.2b in the presence of 1 μm BDS-II in the bathing solution are shown in Fig. 5A. Compared with control data in Fig. 2A, several changes in the on-gating currents are clear. First, the I_g,on does not appear until the membrane is depolarized to −20 mV. Second, a slowly decaying component of I_g,on can be seen after the initial spike during larger depolarizations. Third, the I_g,off becomes clearly larger than the I_g,on during the prepulse. The appearance of the slowly decaying component in I_g,on induced by BDS-II toxin was studied by fitting the relaxation time course of depolarizations below +50 mV to a mono-exponential, and depolarizations at +50 mV and above, where the slow decaying component becomes apparent, to a double exponential function (Fig. 5B). Both fast and slow time constants are voltage sensitive (Fig. 5C), with fast time constants similar to those in control (Fig. 4B), and the slow time constants almost 10-fold slower. It is notable that the slowly decaying component remains separated from the fast component, even at very large depolarizations (+100 mV).

Figure 5.

Kv3.2b gating currents in the presence of BDS-II A, current tracings from Kv3.2b channels with 1 μm BDS-II toxin in the bath. The cell was held at −100 mV and pulsed to the potentials as indicated above the current traces. B, I_g,on at +40 and +80 mV from a holding potential of −100 mV. Capacity compensation, leak subtraction and 60% series resistance compensation were applied and currents were filtered at 100 kHz and sampled at 330 kHz. For presentation, the data were further filtered at 20 kHz offline. Superimposed continuous lines are mono-exponential (+40 mV) or bi-exponential (+80 mV) fits to the decay of I_g,on with fast and slow time constants (τ_fast and τ_slow). C, averaged τ_fast and τ_slow of the decaying phase of I_g,on as a function of pulse potential (n = 5–8). D, averaged time to peak I_g,on from the beginning of the pulse in the presence (•) and absence (○) of 1 μm BDS-II plotted as a function of potential, from data as in A (n = 4). E, the averaged ratio of Q_off/Q_on as a function of pulse potential, from data as in A (n = 5). F, the time constants (τ_off) for the relaxation of I_g,off plotted as a function of pulse potential, from data as in A (n = 7). A mono-exponential function was used to fit I_g,off except between +20 mV and +80 mV, where a bi-exponential function was used.

The rising phase of I_g,on remained very rapid in the presence of 1 μm BDS-II. The time to peak I_g,on, i.e. the time from the beginning of the pulse to the peak of I_g,on, was measured in the absence and presence of 1 μm BDS-II and plotted against potential (Fig. 5D). The voltage dependence of the time to peak I_g,on was not altered significantly by BDS-II toxin.

The average ratio of Q_off to Q_on was significantly greater than 1.0 at potentials more positive than +10 mV (Fig. 5E). The largest ratio value (∼3) was observed at +50 mV, and then the ratio decreased again with larger depolarizations. This voltage-dependent inhibition of I_g,on is reminiscent of the inhibition of ionic current by BDS-II, which is more marked between +20 and +80 mV (Yeung et al. 2005). Few significant changes in I_g,off were observed in the presence of the toxin, except for a more clearly bi-exponential relaxation of I_g,off over a wider potential range than in control (Fig. 5F).

The voltage dependence of charge movement in the presence of 1 μm BDS-II toxin is shown in Fig. 6, in the same format as control data (Fig. 3). Because of the minor effect of BDS-II on deactivation and I_g,off (see below), the measurement of Q_off becomes the only way to measure the charge moved by depolarizing pulses (Lee et al. 2003; Yeung et al. 2005). Normalized Q_off (Fig. 6A) values were plotted versus pulse potential and fitted with a single Boltzmann function (Fig. 6C). The V_½ and slope factor (k) for the Q_off–V curve was 15.0 ± 5.8 mV and 13.0 ± 0.6 mV (n = 4), respectively, a 23 mV depolarizing shift from the control. The unchanged slope factor (k) implied that the equivalent gating charge sensed by the channels in the presence of BDS-II was unaltered from control (Fig. 3C). Neither the voltage dependence of charge return upon repolarization, nor the time constants for the relaxation of I_g,off (τ_off) were significantly changed by BDS-II (Fig. 6C and D), further suggesting that the toxin has little effect on the deactivation process. The Q_off–V curve measured from experiments such as that shown in Fig. 6B had a V_½ of −78.6 ± 0.7 mV and a slope factor (k) of 10.8 ± 1.0 mV (n = 5) (Fig. 6C, bottom).

Figure 6.

Voltage dependence of charge movement in the presence of 1 μm BDS-II A, gating currents recorded during 12 ms depolarizations from −150 mV to between −60 mV and +100 mV in 20 mV steps (top) applied at 0.5 Hz. B, gating currents recorded during a depolarization to +80 mV for 12 ms followed by repolarization to between −40 mV and −150 mV in 10 mV steps. The pulses were applied at 0.5 Hz. C, top, voltage dependence of gating charge movement during depolarizations in the presence of 1 μm BDS-II toxin. Q_off from the experiments shown in A was normalized, plotted, and fitted with a single Boltzmann function, with a V_½ of 15.0 ± 5.8 mV and a slope factor (k) of 13.0 ± 0.6 mV (n = 4). The control Q_on–V curve from Fig. 3C is duplicated here as the dashed line. Bottom, voltage dependence of the charge movement during repolarization in the presence of 1 μm BDS-II toxin. The Q_off from the experiment shown in B was normalized to the value obtained at −150 mV and plotted against the pulse potential during repolarization. The V_½ was −78.6 ± 0.7 mV and the slope factor (k) was 10.8 ± 1.0 mV (n = 5). D, mono-exponential time constants (τ_off) for the relaxation of I_g,off plotted as a function of repolarizing potential, from data as in B (n = 6).

Discussion

Kv3.2b on-gating currents

On-gating current waveforms from rat Kv3.2b channels share several features in common with gating currents from Shaker (Perozo et al. 1992, 1993; Bezanilla et al. 1994; Yang et al. 1997) and other Kv channels, such as Kv1.5 (Fedida et al. 1996; Chen et al. 1997; Wang et al. 1999) and Kv2.1 channels (Taglialatela & Stefani, 1993; Shieh et al. 1997; Lee et al. 2003). Kv3.2b I_g,on waveforms consistently have a rising phase followed by a decaying phase that becomes faster with depolarization (Figs 4A and 5B), highlighting a conserved voltage-dependent gating mechanism in which initial transitions are slower and/or carry less charge than subsequent gating transitions (Bezanilla & Stefani, 1994). In addition, the decay of I_g,on is fast for small depolarizations, it slows for intermediate depolarizations, and for larger depolarizations (> 0 mV), the decay becomes fast again (Fig. 4). However, several differences between I_g,on from Kv3.2b and those from other channels are also clear. The kinetics of the gating current transients appear much faster than previously observed in Shaker (Claydon et al. 2007) and other Kv channels (Fedida et al. 1996; Shieh et al. 1997; Wang et al. 1999; Wang & Fedida, 2001; Lee et al. 2003). The time to peak I_g,on was very brief, ∼0.3 ms at +10 mV, and it decreased to ∼0.2 ms at +80 mV (Fig. 5D). The decay of I_g,on was also fast, ∼2 ms at 0 mV and ∼0.3 ms at +80 mV (Fig. 4B).

The most striking difference between Kv3.2b and other channels is that only one charge system is required to account for the time course of I_g,on (Figs 2–4) and the voltage dependence of Q_on (Fig. 3C). The existence of two gating charge systems in Shaker channels is necessitated by the presence of a slowly decaying component of I_g,on at intermediate depolarizations and the double Boltzmann distribution of the Q_on–V relationship (Bezanilla et al. 1994). In Kv1.5 channels, the charge–voltage relationship was also best fitted with a double Boltzmann function, and the decaying phase of I_g,on at intermediate depolarizations exhibits clearly biphasic waveforms due to the separation of the voltage-dependent parameters of two charge systems (Hesketh & Fedida, 1999). However, Kv3.2b demonstrated a monophasic single exponential decay of I_g,on at all tested potentials (Fig. 4A) and a symmetrical Q_on–V relationship, well fitted to a single Boltzmann function (Figs 3C and 4C). Although these data can be described using a single charge system, it is possible that the voltage-dependent parameters are such that two charge systems with sufficiently similar half-activation potentials and forward kinetics overlap each other sufficiently to produce a smooth single Boltzmann charge distribution.

Kv3.2b off-gating currents

The I_g,off of Kv3.2b channels remains qualitatively similar to that observed in Shaker and other Kv channels (Perozo et al. 1992, 1993; Chen et al. 1997; Yang et al. 1997; Wang et al. 1999; Wang & Fedida, 2001), and specifically the slowing of gating charge return at more positive potentials was retained. In these channels a slow gating step associated with channel opening, but carrying little charge, is responsible for slowed off-gating currents on repolarization as this step is traversed in the opposite direction. Consistent with this idea, the Q_off–V curve for the return of gating charge on repolarization in Kv3.2b channels is negatively shifted by about −70 mV in respect to the Q–V curve during depolarization (Fig. 3C). However, the V_½ of the Kv3.2b Q_off–V curve is at −81 mV about 15 mV less negative than that of Kv1.5 channels (Wang & Fedida, 2001), so that when the membrane was repolarized to −100 mV, only about 10% of the charge was not captured during the 12 ms repolarizing pulse, much less than the ∼40% charge ‘loss’ observed in Kv1.5 channels. Thus, the lack of gating charge slowing seen in Kv3.1 expressed in Xenopus oocytes (Shieh et al. 1997) and Kv3.2b P468W expressed in HEK cells (Yeung et al. 2005) is probably due to the nature of the expression system, combined with the voltage-dependent kinetics of the mutant channels and the potentials used to study them, rather than an inability of the charge to return slowly in Kv3 channels. In the case of Kv3.2b P468W, this non-conducting mutant is analogous to Shaker P473W (Hackos et al. 2002b). As P468 is one of the components of the PVP motif that forms the ‘gating hinge’ in Kv channels (Hackos et al. 2002a; Labro et al. 2003; Sukhareva et al. 2003; Long et al. 2005), the mutation from P to W prevents channel opening, which is required to see charge slowing on repolarization. Preliminary gating current data obtained from Kv3.1 channels expressed in HEK cells appear very similar to Kv3.2b gating currents in showing charge slowing (authors' unpublished data), which suggests that slowed off-gating currents upon repolarization are generally observed in Kv channels.

Structural considerations in Kv3 gating

The number of charged residues in the S4 segment is an obvious difference in the structure between Kv3 and Shaker and its mammalian homologous channels. A valine residue in Kv3.2b channel replaces the lysine at the cytoplasmic end of S4 (K380) in Shaker. However, the effect of mutating lysine to valine at this residue on the Shaker gating current has not been tested. Exchange of the cytoplasmic half of S5 between Kv3.1 and Kv2.1 channels significantly affected the kinetics of deactivation and I_g,off, suggesting that S5 participates in the conformational rearrangements associated with the concerted step(s) in channel activation and early step(s) in deactivation (Shieh et al. 1997), but no effect on I_g,on was observed. Our data from Kv3.2b P468W showed little change in I_g,on (Yeung et al. 2005) when compared with the data here from wild-type channels, suggesting that charge movement during activation is not influenced significantly by alterations in the gate region of the pore. Interestingly, a recent study of the interaction of S4 with S5 and S6 revealed that a group of mutants located at the interface between the voltage-sensing and pore domains in the Kv1.2 structure perturbed the voltage dependence of activation and induced a pronounced slowing of the relaxation of I_g,on in Shaker channels (Soler-Llavina et al. 2006), suggesting the importance of the interface between the voltage-sensing and pore domains in influencing the kinetics of charge movement during activation.

Mechanism of action of BDS-II

BDS toxins apparently bind to the voltage-sensing domains, S3b and S4, and stabilize them in their resting state, which results in a positive shift of the voltage dependence of activation and slowing of the activation kinetics (Yeung et al. 2005). The gating current results from Kv3.2b further support this mechanism by showing that the voltage dependence of the charge movement was shifted by about +23 mV in the presence of 1 μm BDS-II (Fig. 6), without a change in the slope of the relationship. The unchanged slope factor (k) indicated that the effective valency of gating was not altered by BDS-II. Also, a single Boltzmann function was still adequate to fit the Q–V relationship in the presence of BDS-II (Fig. 6), implying that this concentration of BDS-II was sufficient to affect all the apparent charge movement. This result also suggests that BDS-II inhibited the displacement of gating charges in an independent fashion, as suggested for hanatoxin (Lee et al. 2003; Swartz, 2007).

One reason for the use of BDS-II in the present experiments was the possibility that it might affect multiple charge systems in Kv3.2 channels differentially, and convert the single Boltzmann charge distribution seen in control to a distribution best fit with a double, or even triple Boltzmann function. This proved not to be the case and supports the idea that a single charge system can explain Kv3.2 gating under the conditions of our experiments and may contribute to the very rapid activation of channels seen physiologically.

Despite this, there were complex changes observed in the kinetics of I_g,on, with an initial fast phase of on-gating current and the appearance of a second, slower phase. As the voltage dependence of the time constant for the fast decaying component of I_g,on was also shifted by BDS-II (Fig. 5C), it is hard to explain the fast decaying component as the movement of unbound subunits. The slow phase that appeared in the on-gating currents was more pronounced over the potential range between +20 and +80 mV, and this feature may account for the voltage-dependent slowing of ionic current activation (Yeung et al. 2005). Interestingly, BDS-II affected I_g,on elicited by relatively weak depolarizations (≤+10 mV), although I_g,on transients were small, and at these potentials, the ratio of Q_off over Q_on was almost equal to 1 (Fig. 5E). In the presence of BDS-II the Q_off became significantly larger than Q_on over a broad range of voltages (Fig. 5E), with a maximum ratio of ∼3. This ratio diminished again at the most positive potentials studied (+100 mV). The ultimate fate of the I_g,on in the presence of BDS-II at intermediate potentials is unclear. The fact that the charge return was complete upon repolarization (Fig. 6) suggests that there was a slowed movement of voltage sensor domains in response to depolarization that was not differentiated from the baseline currents. This is reminiscent of the inability to record immobilized gating charge upon repolarization in Kv1 channels (Fedida et al. 1996; Olcese et al. 1997) which nevertheless becomes re-available for the next depolarization.

In contrast to the effect on I_g,on, the I_g,off was unaffected by BDS-II. Both the voltage dependence of charge return and the time constant for the relaxation of I_g,off remained almost unchanged from the control (Fig. 6), except for a positive potential shift coupled with I_g,on. This is in good agreement with ionic tail current data from Kv3.1 with BDS-I and BDS-II which shows little effect on the rate of deactivation. But it is different from the reported action of hanatoxin on Kv2.1 ionic current deactivation, which suggests a re-association to channels as they deactivate (Swartz & MacKinnon, 1997a, b; Lee et al. 2003). Two possible explanations for BDS toxins are that they might remain bound to the channel protein during closing, but cannot influence the deactivation process due to the independence of their binding site(s) from the structural elements that regulate deactivation. Alternatively, the toxins may dissociate from the channel after activation, at some time before closing. It seems unlikely that the toxin is actually expelled during closing, since, for most agents like drugs that block open ion channels, the expulsion of an exogenous agent usually slows the process of closing somewhat. These hypotheses remain to be tested by further experimentation.