Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Nifedipine is a Ca⁺⁺ channel antagonist widely used for the treatment of a variety of cardiovascular disorders. Normally, nifedipine blocks voltage-gated calcium channels with a high affinity (K_d = 310 nM in rabbit right atrium, Mecca and Love, 1992; and 200 nM in myocardium, Charnet et al., 1987). However, voltage-dependent K⁺ channels belong to the same supergene family as Ca⁺⁺ and Na⁺ channels (Catterall, 1988; Jan and Jan, 1989), with areas of homology in the pore and around the carboxyl terminal of S6 in Ca⁺⁺ and K⁺ channels (Rampe et al., 1993; Nakayama et al., 1991), and it is known that three types of Ca⁺⁺ antagonists, verapamil, nifedipine and diltiazem, all block cloned K⁺ channels (Rampe et al., 1993; Grissmer et al., 1994). Verapamil and nifedipine produced a marked block of the transient outward current, I_to in rat ventricular myocytes (Jahnel et al., 1994). Because I_to may contain multiple components of rapidly inactivating and slowly inactivating K⁺ channels, it is unclear which specific Kv channel or channels may be blocked.

One component of I_to may be the rapidly activating delayed rectifier K⁺ channel, hKv1.5 (Van Wagoner et al., 1996), cloned from human heart (Tamkun et al., 1991; Fedida et al., 1993), which is important in determining the duration of the plateau phase of the cardiac action potential. Data exist showing that hKv1.5 is blocked by all three types of Ca⁺⁺ antagonists. Verapamil block of hKv1.5 was described in detail (Rampe et al., 1993) and a mechanism of open channel block from the inner pore was suggested. Diltiazem and nifedipine (Grissmer et al., 1994) have also been shown to block hKv1.5, but the detailed characteristics of nifedipine's effects on hKv1.5 have not been studied. Our study was undertaken to examine the effects of nifedipine on hKv1.5 and to explore its mechanisms of action. Our data suggested that nifedipine was an open channel blocker, which acted predominantly at the external pore of hKv1.5 channels. The effects of nifedipine were concentration and voltage dependent, but the block was somewhat relieved at more positive potentials. This effect is in contrast to the block of hKv1.5 by verapamil (Rampe et al., 1993), the inhibition of native K⁺ channels by nifedipine (Jacobs and DeCoursey, 1990) or D600 and related phenylalkylamines (DeCoursey, 1995), and the block of Ca⁺⁺ channels by organic Ca⁺⁺ channel antagonists (Sanguinetti and Kass, 1984; Hume, 1985; Uehara and Hume, 1985), where block increased with potential.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Cell culture. The methods used to establish stable HEK cell lines expressing the hKv1.5 K⁺ channel and those used for electrophysiological measurement of hKv1.5 currents have been described in detail previously (Fedida et al., 1993). Alternatively, hKv1.5 was transiently transfected into HEK cells using the mammalian expression vector pRc/CMV. Cells expressing hKv1.5 were detected by cotransfecting cells with the vector pHook-1 (Invitrogen, San Diego, CA). This plasmid encoded the production of an antibody to the hapten phOX, which when expressed is displayed on the cell surface. Transfected cells were maintained in modified Eagle's medium at 37°C in an air/5% CO₂ incubator in 25-mm Petri dishes plated on glass coverslips until use. One hour before experiments, cells were treated with beads coated with phOX. After 5 min, excess beads were washed off with cell culture medium and cells which had beads stuck to them were used for electrophysiological tests. The efficiency of dual transfection was observed to be better than 80%, so the beads provided a good means of identifying those cells that expressed hKv1.5. No difference was observed from data obtained using stable cell lines or transient expression of hKv1.5, so all results have been included in the analysis.

Solutions. For W/C and O/O macropatches, the control pipette filling solution contained (in mM): KCl, 130; EGTA, 5; MgCl₂, 1; HEPES, 10; Na₂ATP, 4; GTP, 0.1; and was adjusted to pH 7.2 with KOH. The control bath solution contained (in mM): NaCl, 135; KCl, 5; sodium acetate, 2.8; MgCl₂, 1; HEPES, 10; CaCl₂, 1; and was adjusted to pH 7.4 with NaOH. For C/A and I/O macropatches, the pipette filling solution was the extracellular control solution used in W/C recording, and the bath solution was a 135 mM K⁺ solution designed to zero the membrane potential (in C/A recording). It contained (in mM): KCl, 135; HEPES, 10; MgCl₂, 1; dextrose, 10; and was adjusted to pH 7.4 with KOH. For gating current experiments, cells were superfused with a solution containing (in mM): NMG, 140; HEPES, 10; CaCl₂, 1; MgCl₂, 1; dextrose, 10; pH 7.4 with HCl. The pipette solution contained (in mM): NMG, 140; HEPES, 10; MgCl₂, 1; EGTA, 10; pH 7.2 using HCl. Nifedipine was dissolved in alcohol at a stock concentration of 1, 10 or 100 mM, and was protected from the light during all experiments. After control data were collected, the bathing solution was changed to include nifedipine. Nifedipine effects were very rapid on cells, apparently reaching a steady-state within three pulses (30 sec). Steady-state measurements made in the presence of nifedipine were obtained after at least 3 min exposure to the drug. All chemicals were from Sigma Chemical Co. (St. Louis, MO).

Electrophysiological procedures. Coverslips containing cells were removed from the incubator before experiments and placed in a superfusion chamber (volume 250 µl) containing the control bath solution at 22 to 23°C. W/C, C/A, O/O and I/O macropatch recordings were made via the variations of the patch-clamp technique (Hamill et al., 1981), using an Axopatch 200A amplifier (Axon Instruments, Foster City, CA). Patch electrodes were pulled from thin-walled borosilicate glass (World Precision Instruments, Sarasota, FL) on a horizontal micropipette puller, fire-polished, and filled with appropriate solutions. Electrodes had resistances of 1.5 to 3.0 m when filled with control filling solution. Analog capacity compensation and 75 to 85% series resistance compensation were used in all W/C measurements. In some experiments, leak subtraction was applied to data. Membrane potentials have been corrected, where appropriate, for junctional potentials that arose between the pipette and bath solution. Data were filtered at 5 to 10 kHz before digitization and stored on a microcomputer for later analysis using the pClamp6 software (Axon Instruments). In experiments where gating currents were recorded, the sample rate was 330 kHz and currents were usually leak-subtracted using a P/6 protocol similar to that described previously (Stühmer et al., 1991; McCormack et al., 1994; Bouchard and Fedida, 1995). Currents were low-pass filtered at 10 to 50 kHz during data collection or later at 10 kHz for data presentation. Pipettes were routinely sylgarded and fire polished to reduce electrode capacitance and improve seal resistance. The system that we used for expression of hKv1.5 conferred certain advantages for the measurement of gating current. Due to the high level of expression of hKv1.5 channels in HEK cells, there was no need for signal averaging, and single gating current transients could be observed. The average cell capacitance was quite small, and the absence of ionic current at negative membrane potentials allowed faithful leak subtraction of data.

Data analysis. The concentration-response curve (figs. 2C and 7B) for changes in peak and steady-state current produced by nifedipine were computer-fitted to the Hill equation:

(1)

where f is the fractional current block (f = 1-I_drug/I_control) at drug concentration [D]; K_d is the concentration producing half-maximal inhibition and n is the Hill coefficient. The rapid component of inactivation induced by nifedipine was much faster than that observed in the absence of drug. Therefore, we used this drug induced time-constant (₂) as an approximation of the drug channel interaction kinetics, as described previously (Snyders et al., 1992; Slawsky and Castle, 1994), according to the equations:

(2a)

and

(2b)

in which ₂ is the current decay time constant caused by the drug; [D] is the concentration of drug; k₊₁ and k₁ are the apparent rate constants of binding and unbinding for the drug, respectively.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Concentration-dependent block of hKv1.5 by nifedipine. A and B, W/C current recording during voltage pulses to +40 mV from the holding potential of -80 mV. Currents were recorded in the steady-state at a large range of nifedipine concentrations. Current traces in A and B were from two different cells. The cell in A was exposed to low concentrations of nifedipine from 0.2 to 10 µM and the cell in B was exposed to high concentrations of nifedipine from 20 to 200 µM. C: Dose-response curve for nifedipine block of hKv1.5. Peak and steady-state reduction of current (relative to control) at a test potential of +40 mV was plotted against the concentration of nifedipine. Data were averaged from 4 to 10 experiments. Solid lines were fit to the data using a Hill equation (see "Materials and Methods," equation 1). For the reduction of peak current, the Kd value was 18.6 ± 2.7 µM, Hill coefficient was 0.75 ± 0.04; for the reduction of steady-state current, the Kd value was 6.3 ± 0.5 µM, and the Hill coefficient was 0.93 ± 0.03.

View larger version (22K): [in this window] [in a new window]   Fig. 7.   Comparison of nifedipine-induced block of hKv1.5 in different recording configurations. A, For O/O macropatch (left panel), current traces were elicited from the holding potential of -80 mV during a 400-msec pulse to the test potential of +40 mV. For the C/A macropatch (right panel), traces were obtained from the holding potential of +80 mV to a pipette test potential of -40 mV in a 200-msec pulse. Traces shown were from two different cells in which currents were recorded in the absence and presence of 10 and 100 µM nifedipine. Note that nifedipine actions on hKv1.5 were stable after only two or three pulses after nifedipine was washed into the bath. However, all steady-state data shown here, and in B were obtained after more than 3 min exposure to nifedipine. B, The relative steady-state currents (Inif/Ictl) in the presence of different concentrations of nifedipine have been obtained from four modes of patch clamp recording. These were C/A (), I/O () and O/O () macropatches, and W/C () recording. Current were obtained during the same voltage protocol as illustrated in A. Solid lines were the best fits to the data using Hill equation (see "Materials and Methods," equation 1). For the data from C/A, I/O, O/O macropatches and W/C recording, the resultant Kd were 47.3 ± 10.1, 7.5 ± 1.5, 6.9 ± 1.0, 6.7 ± 0.6 µM and Hill coefficients were 0.88 ± 0.05, 0.70 ± 0.06, 0.90 ± 0.06, 0.95 ± 0.04, respectively. Data were means ± S.E. from three to seven experiments. The asterisk represents a statistically significant (P < .05) difference between the C/A or I/O macropatch data and O/O macropatch or W/C data.

(3)

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Concentration-dependent and reversible block. The data in Figure 1 show the effects of nifedipine on hKv1.5 currents expressed in HEK cells under W/C recording conditions. The cell was held at -80 mV, and membrane currents were elicited before exposure to nifedipine, during exposure to 10 µM and 50 µM nifedipine and after wash out. These are currents in response to a series of step depolarizing pulses from -30 mV to +40 mV. Extracellular application of 10 µM and 50 µM nifedipine resulted in a reduction of both peak and steady-state hKv1.5 currents with a marked increase in the rate of outward current relaxation in the presence of nifedipine (Fig. 1B, 1C). This meant that steady-state currents recorded at the end of 250 ms pulses, were much more reduced by nifedipine than the peak currents, in a concentration-dependent manner. After 5 min wash out to control bath solution, the effect of nifedipine was largely reversed. Figure 1D showed that the current was restored to 85% of control current level. Current-voltage (I-V) relationships for the peak and steady-state outward current in the absence, presence of 10 and 50 µM and washout of nifedipine are shown in Figure 1E and F. Here it can be seen that both peak and steady-state current were reduced in a concentration-dependent manner, but that the reduction of steady-state current was much more than that of peak current. Native HEK cells also possess a small outwardly rectifying K⁺ current. The mean amplitude of this current was 204 ± 19.7 pA at +40 mV (n = 10), which is less than 2% of the total current in transfected cells on depolarization (compare control hKv1.5 current amplitudes in figs. 1E and 2). Nifedipine also reduced this current, with 50% block at >20 µM (n = 3). However, due to the small current size, we have not considered contamination by this endogenous current to be significant.

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Nifedipine block of hKv1.5 current. A to D, Currents were elicited by 250-msec depolarizing pulses from the holding potential of -80 mV to voltages between -30 mV and +40 mV in increments of +10 mV. W/C currents were recorded in the absence of nifedipine (A) and the presence of 10 µM (B), 50 µM (C) and after 5 min wash out (D) nifedipine, respectively. E and F, Current-voltage (I-V) relationships for peak and steady-state outward current in the absence () and presence of 10 (), 50 ()µM and wash out () of nifedipine are shown in E and F, respectively. The dotted line marks the zero current level. Data were from the same cell as panels A to D. Note that after exposure to high concentrations of nifedipine, wash out was often incomplete as in D.

Voltage-dependent block. To examine the voltage-dependence of block, the relative steady-state current I_nif/I_ctl at the end of 400 msec voltage clamp pulses was plotted as a function of potential. The data in figure 3 show normalized hKv1.5 current-voltage(I-V) relationships for different concentrations of nifedipine. The dotted line is the normal activation curve of hKv1.5. This was obtained from the deactivating tail current amplitude at -20 mV after 15 msec depolarizing steps to potentials between -80 to +100 mV from a holding potential of -100 mV. In the presence of 5, 10, 20 and 50 µM nifedipine, block increased rapidly between -10 and +10 mV, coinciding with the voltage range of channel opening. These data suggest that nifedipine binds primarily to the open state of hKv1.5 channels. Over the voltage range between +20 and +90 mV, almost all channels are open, but block was slightly relieved with depolarizing test potentials and showed a shallow voltage dependence. At different concentrations of nifedipine, block at +90 mV compared with +20 mV was reduced from 0.50 ± 0.01 to 0.40 ± 0.03 (5 µM), from 0.57 ± 0.04 to 0.49 ± 0.06 (10 µM), from 0.70 ± 0.05 to 0.60 ± 0.05 (20 µM), and from 0.89 ± 0.02 to 0.84 ± 0.02 (50 µM); (n = 4-6). Over the potential range where channels were fully open, the relationship for block by nifedipine at depolarizing voltages was well fitted to a linear equation. The slopes of fit were 1.53 × 10³, 1.16 × 10³, 1.38 × 10³ and 7.35 × 10⁴ mV¹ with 5, 10, 20 and 50 µM nifedipine, respectively.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Voltage-dependence of current block by nifedipine. The dotted line represents the activation curve of control hKv1.5. This was obtained from deactivating tail current amplitude at -20 mV after 15-msec depolarizing steps to potentials between -80 to +100 mV, from a holding potential of -100 mV. The symbols denote ratios of steady-state current in nifedipine to the control current value at each potential. Open symbols show the data between -10 and +10 mV. Filled symbols show the data at voltages positive to +20 mV. The solid lines were the best fit to a linear function in the form of y = ax + b. With 5, 10, 20 and 50 µM nifedipine, the slopes from this linear fit were 1.53 × 103, 1.16 × 103, 1.38 × 103 and 7.35 × 104 mV1. Data were mean ± S.E. from 4 to 10 experiments.

Effects of nifedipine on current decay of hKv1.5. In the absence of nifedipine, current decayed slowly (fig. 2, A and B) and during the relatively short voltage pulses used here, was fitted to a single exponential function with a decay time constant of 230 ± 8 msec (n = 14) at +40 mV. After addition of nifedipine, the rate of current decay increased in a concentration-dependent manner (fig. 2, A and B) and could be well fitted with a double exponential function. The nifedipine-induced fast time constant, ₂, was used as an index of the rate of block of hKv1.5. Figure 4A shows the effects of different concentrations of nifedipine and different voltages on the mean ₂ values. At each voltage, ₂ decreased in a concentration-dependent manner, which indicated a more rapid current decay. At each concentration of nifedipine, ₂ became faster at voltages between 0 and +30 mV, then ₂ showed a shallow decrease at more positive potentials.

View larger version (12K): [in this window] [in a new window]   Fig. 4.   Time constants of hKv1.5 current decay in the presence of nifedipine. A, Voltage-dependence of hKv1.5 current decay in different concentrations of nifedipine. In the presence of nifedipine, 2 was the fast component obtained from biexponential fits to the falling phase of the currents obtained during 250-msec voltage steps to between 0 and +90 mV from the holding potential -80 mV. Data were mean ± S.E. from 4 to 10 experiments. B, Kinetics of nifedipine block of hKv1.5. The reciprocal of the nifedipine-induced fast time constant (1/2) at +40 mV has been plotted against the nifedipine concentration. The best fit to the data (solid line) using the equation 1/2=k+1. [D] + k1 resulted in an apparent association rate constant (k+1) of 5.64 × 106 M1 sec1 and an apparent dissociation rate constant (k1) of 37.5 sec1. The Kd value (k1/k+1) was 6.65 µM. Error bars, S.E. from four to six experiments.

Open or closed channel block? Two methods are often used to decide whether drugs predominantly interact with open or closed channels. For open channel blockers that have a slow block rate and that dissociate from closed channels, currents peak in the presence of drug at a constant level, before a rapid decay occurs with drug-channel interaction. Such an effect is seen when K⁺ channels are exposed to 4-aminopyridine for the first time (Choquet and Korn, 1992; Bouchard and Fedida, 1995). As a result, an acceleration of current inactivation is generally inferred to mean open-channel binding (Slawsky and Castle, 1994). However, if the peak current is reduced due to rapid drug-channel interaction, as for tetrapentylammonium block of hKv1.5 (Snyders and Yeola, 1995) or in our experiments with nifedipine, it can be more difficult to exclude a resting channel block. One method to analyze the onset of block is to fit the fractional block of current back to the start of the depolarizing pulse (Slawsky and Castle, 1994). If the fit intersects zero block after the start of the pulse, a purely open channel block mechanism is favored.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Concentration-dependent rate of onset of hKv1.5 block by nifedipine. Original currents were obtained from a holding potential of -80 mV during 200-msec pulses to a test potential of +40 mV. Nifedipine sensitive current was calculated from the subtraction of steady-state currents in control and the presence of nifedipine, divided by the control current level (Ictl-Inif)/Ictl. These have been plotted against the time from the start of the voltage clamp pulse. The times of onset of block were 3.0, 2.4, 1.3, 1.1, 0.75 and 0.35 msec with 5, 10, 20, 50, 100 and 200 µM nifedipine. The solid lines were the best fits to a biexponential equation of the form [y = A1*exp(-t/1) + A2*exp(-t/2) + C]. Data shown were from two different cells.

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Deactivation tail current crossover caused by nifedipine. A, Under W/C recording conditions, cell was held at -80 mV and stepped to +40 mV for 100 msec, then stepped back to -50 mV as shown by the protocol inset in A. Tail currents were recorded during the last voltage step to -50 mV. Superimposed currents are shown in control and the presence of 2, 5 and 10 µM nifedipine. B, Deactivation tail currents from panel A (dotted box) were enlarged to illustrate cross-over of control and in the presence of 2, 5 and 10 µM nifedipine. C, Tail currents in B were enlarged again to show crossover clearly (upper panel, in the presence of 5 µM nifedipine) and the rising phase at the beginning of tail currents (lower panel) in the presence of 2, 5 and 10 µM nifedipine, with the control tracing for comparison.

Site of action. Nifedipine is a 1,4-dihydropyridine with a pKa  1.0 (Uehara and Hume, 1985). Thus, at a physiological pH of 7.4, almost all of the nifedipine will be in its neutral form. As an uncharged moiety it should rapidly cross biological membranes, and thus it has a possible site of action at the outer or inner mouth of open hKv1.5 channels. It has been reported that nifedipine can block Ca⁺⁺ channels from the extracellular side or penetrate the membrane to approach its binding site in the hydrophobic domain near to the extracellular side. To attempt to determine on which side of the membrane that hKv1.5 channels could be blocked by nifedipine, the efficacy of block under different macropatch recording conditions was compared. The data in figure 7A illustrate the effects of 10 and 100 µM nifedipine on the current recorded from O/O and C/A macropatches. In both cases the currents were recorded in the steady state in nifedipine, after at least 3 min exposure. In the presence of nifedipine, at similar concentrations, current was much reduced in the O/O patch mode of recording compared with the C/A patch. Summary steady-state data from two to eight cells in figure 7B shows the effect of different concentrations of nifedipine on hKv1.5 in C/A, I/O, O/O macropatches and during W/C recording. Data points were obtained by comparing the steady-state current in control and different concentrations of nifedipine under the different recording conditions. Compared with O/O macropatches and W/C recording, the current in I/O and C/A macropatches was relatively less sensitive to nifedipine. At concentrations of 1 to 5 µM nifedipine, very little block was observed in C/A macropatches, whereas at 5 µM nifedipine, current was 50% blocked in W/C and O/O macropatch recordings. The rank order of block of hKv1.5 by nifedipine was then, whole-cell  outside-out > inside-out  cell-attached macropatches. Analysis of variance of the results with 10, 20, 50 and 100 µM nifedipine showed that the difference of relative current between C/A or I/O macropatches and O/O macropatches or W/C recordings was statistically significant (P < .05), the difference between O/O macropatches and W/C recording was not statistically significant (P > .05).

View larger version (23K): [in this window] [in a new window]   Fig. 8.   Effect of nifedipine on gating currents of hKv1.5 channel. A and B, Data in panel A and B are the gating currents of hKv1.5 channel in the absence (A) and presence (B) of 5 µM nifedipine. Cell was depolarized from -100 to +70 mV for 12 msec from a holding potential of -100 mV to initiate the on-gating current, then repolarized to -100 mV to record off-gating current. Traces shown are from -70 to +30 mV in increments of +20 mV. C and D, Data from a different cell in the absence (C) and presence (D) of 50 µM nifedipine. Traces were elicited using the same voltage protocol as in panels A and B. E and F, Integrated on- () and off- () gating charge movement in the presence of 5 µM (E) and 50 µM (F) nifedipine expressed relative to the maximum on- () and off- () gating charge movement in controls. Data were means ± S.E. from four (F) or six (E) experiments. The asterisk represents a statistically significant (P < .05) difference between off-gating charge movement in control and in the presence of nifedipine.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

We have investigated the mechanism of action of nifedipine on the hKv1.5 channel expressed in a human cell system. Nifedipine is a tissue-specific Ca⁺⁺ channel antagonist that has a major role in the vascular bed where it is an effective vasodilator. In our study nifedipine also produced a strong block of a human heart Kv channel.

Nifedipine blocks hKv1.5 with a low K_d. Nifedipine is a widely used Ca⁺⁺ antagonist and has been shown to block native Ca⁺⁺ channels with K_d in the range of 200 to 310 nM (Charnet et al., 1987; Mecca and Love, 1992). Cloned Ca⁺⁺ channels are also blocked by dihydropyridines and phenylalkylamines with K_ds in the 100s of nM to low µM range (Schuster et al., 1996). Nifedipine blockade of native cardiac potassium channels has also been reported. Transient outward potassium currents I_t in rabbit atrium (Gotoh et al., 1991) and I_to in rat ventricular myocytes (Jahnel et al., 1994) were largely blocked by 30 µM nifedipine and the blockade was concentration-dependent. Besides the block of potassium channels in intact myocytes, nifedipine also caused the block of cloned potassium channel hKv1.5 with a K_d of 81 µM (Grissmer et al., 1994). Our data in figures 1 and 2 demonstrated that nifedipine accelerated the time course of decay of hKv1.5 currents. Currents reached a peak that was less than corresponding control currents at concentrations greater than the K_d. Subsequently nifedipine caused a rapid current decay that was concentration dependent. Such effects of nifedipine allowed us to calculate dose-response curves for both the peak and steady-state outward currents (fig. 2). Fits of the Hill equation to these data gave K_d values of 18.6 ± 2.7 and 6.3 ± 0.5 µM for peak and steady-state currents, respectively, with Hill coefficients of 0.75 ± 0.04 and 0.93 ± 0.03. The peak current measurement represented the partial block of current by nifedipine, and the interaction between rates of channel opening and the onset of block at different pulse potentials and nifedipine concentrations (see below). For this reason, the K_d value (18.6 µM) was expected to be lower than for steady-state block of open channels (6.3 µM). The non-steady-state value for the Hill coefficient that was measured (0.75) from the peak current dose-response cannot then be used to indicate cooperativity. The measurement of steady-state current block reflected the equilibrated interaction of hKv1.5 channels with nifedipine at different concentrations and potentials and was of more interest in the present study. The Hill coefficient close to 1.0 for steady-state current block suggests that binding of one nifedipine molecule per channel is sufficient to block the hKv1.5 channel. Although nifedipine also showed some effects on endogenous current of HEK cells, the average amplitude was about 2 to 3% of the hKv1.5 current. The endogenous current was also less sensitive to nifedipine, so was unlikely to significantly distort our quantitation of hKv1.5 current block by nifedipine in HEK cells. The K_d of 6.3 ± 0.5 µM for hKv1.5 expressed in HEK cells was an order of magnitude lower than for an isoform of hKv1.5 (HPCN1) (Grissmer et al., 1994) expressed in MEL cells. Possible reasons for this difference are the different mammalian expression systems (MEL vs. HEK cells) and small structural differences between the two isoforms of hKv1.5 used here (fHK vs. HPCN1), although it should be noted that these isoforms are identical throughout the S4-S6 gating and pore regions. We found that photoinactivation of the drug occurred readily in the cloned cell system and when making measurements with nifedipine it was necessary to carry out all experiments in the dark. A K_d in the low micromolar range, and threshold effects in the 100 nM range make nifedipine a potent blocker of hKv1.5. No studies have addressed block of specific components of cardiac K⁺ current in intact human myocytes by nifedipine, so at the present time it is not possible to relate our observations directly to currents in human heart.

Time dependence of block and open channel block. Our data indicated that nifedipine block of hKv1.5 showed marked time-dependence. Block increased in an exponential manner during the depolarizing pulses (fig. 2), and the onset of block occurred sharply after current activation (figs. 3 and 5). Nifedipine also modified the tail current (fig. 6). Upon repolarization, control channel deactivation was fast and virtually irreversible (fig. 6A). Open-channel models predict that if a large fraction of the channels is blocked at the start of repolarization and the unbinding rate (k₁) is fast enough, then the tail may display a rising phase reflecting the unblocking from blocked to open state. Subsequently, the tail should deactivate more slowly than in control, because some unblocked channels become blocked again, depending on the relative rate constants for the open to blocked state and the open to closed state. Current traces in figure 6C (lower panel) show that a rising phase was prominent with 10 and 5 µM, but less so with 2 µM nifedipine and absent in control. Subsequently tail currents cross-over as channels in the presence of nifedipine move more slowly from the blocked state to open and closed states than in control (fig. 6, B and C, upper panel) (Snyders et al., 1992; Fedida, 1997). From the results discussed above we conclude that nifedipine binds to the open state of the channel.

Voltage-dependence of block. In some K⁺ channels voltage-dependence of nifedipine block has been described. In rat alveolar-epithelial cells an inactivating delayed rectifier K⁺ channel demonstrated a voltage-dependent _block (Jacobs and DeCoursey, 1990). In frog atrial myocytes, block of I_k by nisoldipine, a nifedipine analogue was voltage dependent (Hume, 1985). Our data in figure 3 showed that current was blocked quickly by nifedipine in the voltage range of channel opening. After channels were fully open, block was slightly relieved at more positive voltages. This relief of block was independent of the concentration of nifedipine between 5 and 50 µM (fig. 3), and was well fitted to a linear function (fig. 3). The slopes of the fits were approximately similar, between 0.75 × 10³ and 1.5 × 10³ mV¹, and these quantify the increase in normalized current with depolarizations in the presence of the drug. At the same time, data in figure 4A indicate that once channels were fully open (>+30 mV), there was a shallow decrease in the time constants of block with potential. From the above and a consideration of equation 2a and 2b, these data indicate that the voltage-dependence to the block must be given by an increase in the off-rate (k₁) rather than a decrease in the on-rate (k₊₁) at more positive potentials. There are a number of possible explanations for this. First, although nifedipine is uncharged, its binding site may be coupled to some voltage-dependent process, such as conformational changes in the pore with potential or movement of the voltage sensor (Jacobs and DeCoursey, 1990), which can sense the transmembrane voltage change and confer voltage-dependence to block by an uncharged drug. It has been proposed by De Coursey that neutral phenylalkylamine drugs may have rapid access to their receptors, where block is then stabilized by protonation of the drugs (DeCoursey, 1995). Although nifedipine has a much lower pK_a, such a mechanism could explain the voltage-dependence observed in the present experiments.

Nifedipine site of action. An alternative explanation for the voltage-dependence of block could be that nifedipine blocks open K⁺ channels from an external site. It is possible that at more positive voltages, potassium permeation increases and hinders in some manner the binding of nifedipine to its site with a resultant relief of block. Analogous with this, it is known that K⁺ occupancy of sites in the external mouth of the K⁺ channel pore affects the rate at which charged blockers (ions) can interact with the channel (Baukrowitz and Yellen, 1996). We have established that nifedipine induced open-channel block of hKv1.5 and two additional lines of evidence suggest that the site is in the external mouth of the pore. The rank order of block described in figure 7 was W/C  O/O > I/O  C/A macropatch. This suggested a preferential nifedipine block of hKv1.5 channels from the extracellular side or at a hydrophobic domain accessible from the extracellular surface. A similar conclusion has been drawn for native K⁺ channels (Jacobs and DeCoursey, 1990; although see DeCoursey, 1995) and for the dihydropyridine binding site in Ca⁺⁺ channels (Nakayama et al., 1991; Schuster et al., 1996), where it is thought that the dihydropyridines block the channel from the extracellular side and mutations in repeat IVS6 affect binding (Schuster et al., 1996). The second line of evidence is based on the gating current measurements (fig. 8). At concentrations required to produce approximate 50% block of ionic current, nifedipine had only very small effects on hKv1.5 gating currents. This suggests a binding site distant from the intracellular mouth of the pore, where binding of numerous drugs immobilizes gating charge (Stühmer et al., 1991; Fedida, 1997). At high concentrations, on-gating currents of hKv1.5 did not change in the presence of nifedipine compared with the control, but off-gating currents in the presence of nifedipine were reduced at voltages positive to +10 mV (fig. 8F), in the voltage range that channels open. This also supports the idea that nifedipine can bind only to channels in the open state, i.e., at some site exposed when the pore opens.