INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Potassium channels play an important role in the maintenance of resting potential and in determining the action potential duration of cardiac cells. For example, the relatively negative resting potential of ventricular myocardial cells is attributable to the background potassium ion current, I_K1, which also contributes to the late phase of repolarization of the action potential (Giles and Imaizumi, 1988; Shimoni et al., 1992). The I_K1 channel is open at the resting potential, but strongly rectifies with depolarization, so that it contributes essentially no current for potentials positive to ~20 mV. The mechanism underlying I_K1 rectification is thought to involve voltage-dependent block of the channel pore by intracellular Mg²⁺ (Matsuda et al., 1987; Vandenberg, 1987) and/or the polyamines spermine and spermidine (Ficker et al., 1994). A second inwardly rectifying current in cardiac muscle, the E-4031-sensitive current, I_Kr, also plays an important role in repolarization (Noble and Tsien, 1969; Shrier and Clay, 1986; Sanguinetti and Jurkiewicz, 1990a). This component is rapidly activated during the plateau phase of the action potential, but contributes relatively little current until the membrane potential is repolarized below 0 mV. The mechanism for this effect was originally attributed to inward rectification of the I_Kr channel, similar to that for I_K1 (Shrier and Clay, 1986). However, recent work on heterologous expression of the HERG channel has clearly demonstrated that rectification of I_Kr is, instead, attributable to rapid, voltage-dependent inactivation similar to that of other kinds of K⁺ channels (Smith et al., 1996; Spector et al., 1996).

Divalent cations have been one of the classical tools used to probe potassium ion channel gating and permeation (Hille, 1992). For example, Standen and Stanfield (1978) found a time- and voltage-dependent block of I_K1 in skeletal muscle fibers by Ba²⁺ and Sr²⁺, which they attributed to competition between Ba²⁺ and Sr²⁺ with K⁺ for a site deep within the aqueous pore of the I_K1 channel. A similar result was reported for Sr²⁺ with the I_K1 channel in guinea pig ventricular myocytes by Shioya et al. (1993). Divalent cations have also been shown to block the delayed rectifier potassium ion current in squid giant axons (Eaton and Brodwick, 1980; Armstrong and Taylor, 1980; Clay, 1995). In guinea pig cardiac cells, I_Kr and I_Ks have been reported to have differing sensitivities to divalent cations (Sanguinetti and Jurkiewicz, 1990a,b). However, the effects of divalent cations on I_Kr have not been elucidated. Follmer et al. (1992) found that the addition of 0.2 mM Cd²⁺ to the extracellular solution, a condition that has often been used to block I_Ca, actually increased I_Kr amplitude in cat ventricular myocytes. A similar result has recently been reported in guinea-pig ventricular myocytes by Daleau et al. (1997). We have expanded upon this work, using rabbit ventricular myocytes and various other divalent cations in addition to Cd²⁺. We found that Ni²⁺, Mn²⁺, and Co²⁺ all produced potentiation of I_Kr similar to the Cd²⁺ result, which we have attributed to the alteration of I_Kr inactivation. In contrast, Zn²⁺ (0.1-1 mM) blocked I_Kr. The effects of Ba²⁺ on I_Kr could not be readily characterized because of a Ba²⁺-induced time- and voltage-dependent current attributed to an interaction between Ba²⁺ and the I_K1 channel, as reported previously (Standen and Stanfield, 1978).

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Isolation of ventricular myocytes

Single ventricular cells from rabbit hearts were prepared by using a modification of techniques described by Mitra and Morad (1985). New Zealand white rabbits weighing 1.6-2.5 kg were anesthetized with a combination of ketamine (70 mg/kg) and xylazine (10 mg/kg), and exsanguinated via the carotid artery. The hearts were excised, mounted on a Langendorff reperfusion apparatus, and perfused at 37°C with normal Tyrode's solution (see below) for 5 min followed by an additional 10 min of perfusion with nominally Ca²⁺-free Tyrode's solution. The hearts were then perfused for 5-10 min with Ca²⁺-free Tyrode's solution containing collagenase (type IA 33 units/ml; Sigma Chemical Co., St. Louis, MO), followed by an additional 5-10 min of perfusion with a solution containing collagenase and protease (type XIV, 0.14 units/ml; Sigma Chemical Co.). Pieces of the ventricle were cut and placed in a K-B solution (Isenberg and Klockner, 1982) and agitated gently. The resulting single-cell suspension was stored in K-B solution at room temperature. Cells were used for electrophysiological recordings within 1-8 h.

Solutions

The Tyrode's solution used in the cell isolation procedure contained (in mM) 121 NaCl, 5 KCl, 15 NaHCO₃, 1 Na₂HPO₄, 2.8 sodium acetate, 1 MgCl₂, 2.2 CaCl₂, and 5.5 glucose, gassed with 95% O₂/5% CO₂ mixture. For the Ca²⁺-free Tyrode's solution, the 2.2 mM CaCl₂ was omitted. The K-B solution was a modification from Isenberg and Klockner, (1982) which contained (in mM) 85 KCl, 30 K₂HPO₄, 5 MgSO₄, 5 K₂ATP, 5 sodium pyruvate, 5 -OH-butyric acid, 5 creatine, 20 taurine, 20 glucose, 11 succinic acid, 0.62 polyvinylpyrolidone, and 0.5 EGTA. The pH was adjusted to 7.2 with KOH. The extracellular solution used during electrophysiological recordings contained (in mM) 121 NaCl, 5 KCl, 2.8 sodium acetate, 1 MgCl₂, 2.2 CaCl₂, 10 HEPES, and 10 glucose with the pH adjusted to 7.4 with NaOH. The extracellular solution was continuously gassed with 100% O₂. The solution in the recording pipette contained (in mM) 140 KCl, 5 ATP disodium salt, 5 creatine phosphate disodium salt, 1 MgCl₂, 5 HEPES, 5 EGTA, and 1.54 CaCl₂ with the pH adjusted to 7.2 with KOH (pCa 7.2; pMg 4.2). The calcium current was completely blocked with nifedipine (5-10 µM) (Sigma Chemical Co.). E-4031 was kindly provided by Eisai Co. (Tsukuba Research Laboratories, Japan). Various concentrations (10 µM to 5 mM) of divalent cations (CdCl₂, ZnCl₂, BaCl₂, NiCl₂, CoCl₂, CaCl₂, MnCl₂, SrCl₂, and MgCl₂) were added to the extracellular solution from stock solutions of 1 M concentration. In a few experiments with 5 mM CdCl₂, the NaCl concentration was reduced by 10 mM to keep the ionic strength of the extracellular solution constant.

Electrical recording and data analysis

The whole-cell patch-clamp technique was used to record membrane currents in these experiments. Ventricular cells were placed in a chamber mounted on the stage of an inverted microscope (Ziess IM35, OberKochen, Germany) and allowed to settle for ~5-10 min. The cells were continuously superfused with extracellular solution (see above) at 33-35°C. Rod-shaped myocytes with clear striations were selected for electrical recordings using an Axopatch amplifier (Axopatch-1D; Axon Instruments Corp., Foster City, CA). The pipette tip resistance was 2-4 M. The input capacitance of the cells was in the 50-95-pF range. Voltage clamp pulses were delivered from a custom-designed software package (Alembic Software Co., Montréal, Canada) implemented on a personal computer equipped with an analog-to-digital card (Omega Corp., Stanford, CT). The holding potential used throughout was 40 mV, which effectively inactivates both the sodium ion current, I_Na, and the transient outward current, I_to (Ogbaghebriel and Shrier, 1994). Membrane currents and voltages were filtered at 10 kHz, digitized at 22 kHz via a pulse code modulation unit (Neurocorder DR-390; NeuroData Corp., New York, NY), and recorded on a Betamax VCR (SL-HF 450; Sony Corp., New York, NY). Currents were analyzed off-line at 5 kHz and digitized by a 12-bit analog-to-digital converter at 10 kHz.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Properties of I_Kr in control conditions

Representative recordings of membrane currents obtained in this study are illustrated in Fig. 1. The steady-state holding current at 40 mV was 0.39 nA, which was attributable to I_K1 (Carmeliet, 1993). Moreover, 40 mV lies on the negative slope region of the I_K1 current-voltage relation, as shown by the instantaneous current jumps in the inward direction in all steps. (The steady state at the end of 0 and +20 mV steps is probably attributable to a "leak" current component.) With increasing depolarizations, a time-dependent outward going current, I_Kr, was elicited as indicated by the arrows in Fig. 1. These results are shown in more detail in Fig. 2, in which the steady-state current has been subtracted and the current scale amplified. (The top four records in Fig. 2 were taken from the same preparation as in Fig. 1.) These records indicate that the threshold for activation of I_Kr was between 20 and 10 mV and the conductance saturated at ~+10 to +20 mV. Time-dependent current during the voltage steps themselves was evident at ~20 to 10 mV, and became smaller with increasing depolarizations, so that it was essentially nil at +20 mV, as expected for I_Kr (Sanguinetti and Jurkiewicz, 1990a; Clay et al., 1995). The I_Kr component was completely blocked by E-4031 (1 µM), as shown in Fig. 2. A complete biophysical analysis of these results is given in Clay et al. (1995).

View larger version (5K): [in this window] [in a new window]   FIGURE 1   Control recordings of membrane current from a rabbit ventricular myocyte. These results were obtained by stepping the membrane potential to the voltages indicated for 2 s, followed by a return to the holding level (40 mV). The holding current was 0.39 nA.

View larger version (10K): [in this window] [in a new window]   FIGURE 2   The time-dependent currents (IKr). The top four recordings were obtained from the corresponding results of the preparation in Fig. 1 with the steady-state current subtracted. The step potential (VS) is indicated above each recording. The bottom panel illustrates the blocking effect of 1 µM E-4031 on IKr. These results were taken from a preparation different from that of the other recordings shown here. Calibrations are 0.5 s and 0.05 nA.

Effect of divalent cations on I_Kr

Follmer et al. (1992) previously reported that Cd²⁺, at a concentration used to block the calcium ion current (0.2 mM), modified the kinetic and ion transfer properties of I_Kr. We have expanded upon their work by using a larger concentration of Cd²⁺ (5 mM), as well as testing the effects of other divalent cations. Fig. 3 illustrates our results with 5 mM Cd²⁺. Cadmium shifted the activation curve to more positive potentials (Fig. 3 A) and increased the rate of I_Kr deactivation (Fig. 3 B, inset), which is similar to the effects of divalent cations on voltage-gated conductances in nerve axon (Frankenhaeuser and Hodgkin, 1957; Gilly and Armstrong, 1982a,b). Moreover, the amplitude of I_Kr tail current was significantly increased by 5 mM Cd²⁺, especially at 20 mV (Fig. 3 B, and records to the left of Fig. 3 A). The latter effect appeared to be entirely attributable to I_Kr, inasmuch as the tail currents with 5 mM Cd²⁺ were completely blocked by 1 µM E-4031 (Fig. 3, inset). The activation curves in Fig. 3 A were fit with the Boltzmann relation, f_o(1 + exp((V  V_1/2)/V_S)¹, where V_1/2 = 12 or 36 mV; V_S = 9.1 or 10.0 mV; and f_o = 1 or 1.75, respectively, for control and 5 mM Cd²⁺ conditions. The fit to the Cd²⁺ results is also shown with f_o = 1 (Fig. 3 A, dashed curve) to better illustrate the voltage shift. The results in Fig. 3 B were obtained with a 1-s prepulse to +60 mV to fully activate the I_Kr conductance in both control and 5 mM Cd²⁺ conditions, followed by a step to various potentials of less than +60 mV. The amplitudes of the current obtained by the second step in this protocol are shown in Fig. 3 B for control and test conditions. All of these results have been normalized to the maximum outward current in control, which occurred at ~ 40 mV. We previously modeled the rectification of the fully activated current-voltage relation for I_Kr by assuming that a blocking particle (either membrane bound or in the cytoplasm) moved some distance into the channel with membrane depolarization, thereby reducing outward current in a voltage-dependent manner without significantly altering inward current (Clay et al., 1995). Alternatively, the effect can also be modeled by a rapid, voltage-gated inactivation process that has been shown to underlie the apparent rectification of I_Kr (Smith et al., 1996; Spector et al., 1996). That is, the I_Kr channel is assumed to make transitions between its open and inactivated states, i.e., so that the probability that the open state is occupied (p_o) immediately after a prepulse to +60 mV is p_o = (1 + k₁/k₂)¹, where k₁ and k₂ are functions of membrane potential. Consequently, the fully activated current-voltage relation for I_Kr is given by g(V  E_Kr)/(1 + k₁/k₂), where g is its limiting slope conductance (V  ) and E_Kr is its reversal potential. This equation was fit, by eye, to the results in Fig. 3 B by using k₁/k₂ = 6 exp (0.05V) and E_Kr = 70 mV. In other words, the experimentally observed rectification of I_Kr is sufficient to determine the ratio of k₁ and k₂. Direct measurements of the inactivation kinetics themselves, which we have been unable to carry out because they are so fast (even at room temperature), would be required to determine k₁ and k₂ separately. We have assumed that Cd²⁺ alters these kinetics by binding to the inactivation gate, so that the rate constant k₁ is reduced, i.e., k₁ = k₁^o(1 + [Cd²⁺]/K_D)¹, where [Cd²⁺] is the cadmium concentration, and K_D is the dissociation constant for the binding of Cd²⁺ to the gate. For the Cd²⁺ results in Fig. 3 B, we used k₁/k₂ = 0.8 exp(0.085V), which is consistent with K_D = 0.74 mM. (We have not yet determined the significance of the slight change in electrical distance of the inactivation gating in the presence of Cd²⁺, which this result implies.) This analysis illustrates one way in which I_Kr amplitude can be increased by Cd²⁺. Other models of the Cd²⁺ effect may be equally likely (see Discussion).

View larger version (22K): [in this window] [in a new window]   FIGURE 3   Effects of cadmium on activation and current-voltage relation of IKr. (A) Activation curve in control () and in the presence of 5 mM Cd2+ in the extracellular solution (). These results were obtained from tail current amplitudes at the holding potentials (40 mV) after 2-s duration voltage steps to the potential on the abscissa. Each symbol represents the  ± SD from five different cells. All results were normalized by the maximum tail current obtained in the control. The test results illustrate the significant increase in IKr induced by Cd2+ and a few other divalent cations (Table 1). This effect is shown explicitly by the records in the inset alongside A. The results in A also illustrate the voltage shift of IKr activation, which is further described by the theoretical curves in A. The curve describing the control results is a best fit (least-squares minimization) of the Boltzmann relation, (1 + exp((V  V1/2)/VS)1, where V1/2 = 12 mV and the slope factor VS = 9.1 mV. The curve describing the Cd2+ results represents fo(1 + exp((V  V1/2)/VS)1, where fo = 1.75, V1/2 = 36 mV, and VS = 10.0 mV. This result is also shown with fo = 1 (dashed curve) to better illustrate the voltage shift of the activation curve. (B) Effect of 5 mM Cd2+ on the fully activated current-voltage (I-V) relation of IKr ( ± SD; n = 6). All results were normalized to the maximum outward current obtained in the control. These results correspond to amplitudes of the time-dependent current elicited by a voltage step to the potentials indicated on the abscissa after a 1-s-duration prepulse to +60 mV. The curve describing the control results () is given by 0.043(V + 70)/(1 + 6 exp(0.05V)). The curve describing the Cd2+ results () is given by 0.043(V + 70)/(1 + 0.8exp (0.085V)). The theoretical basis for these relationships is given in the Results. (Inset) Control and test results are shown superimposed to the right for 20 and 50 mV. Block of the time-dependent current with 5 mM Cd2+ by 1 µM E-4031 is shown by the records in the bottom panel of the inset. Calibrations are 0.5 s and 0.1 nA.

As shown in Table 1, Ni²⁺, Co²⁺, and Mn²⁺ all had effects similar to those of Cd²⁺ in terms of the change in rectification (increase in maximum outward current), with Cd²⁺ being more potent than any other divalent cation (Cd²⁺ > Ni²⁺  Co²⁺  Mn²⁺), whereas Ni²⁺ produced approximately the same shift in the I_Kr activation curve as Cd²⁺, with Co²⁺ and Mn²⁺ being considerably less potent (Cd²⁺  Ni²⁺ > Co²⁺  Mn²⁺). The fully activated current-voltage relation in control (2 mM Ca²⁺) was essentially unchanged with 5 mM Ca²⁺ (results not shown). Magnesium and strontium had a similar lack of effect on I_Kr. In these experiments we simply added divalent cations to the extracellular solution, thereby slightly increasing ionic strength. We confirmed that the increase in ionic strength was not a contributing factor in a series of control experiments in which ionic strength was maintained constant (see Materials and Methods). In particular, the voltage shift with 5 mM Cd²⁺ was +23.5 mV (mean from two experiments, 22 and 25 mV, respectively), which is similar to the results obtained when ionic strength was not kept constant.

The only divalent cation that blocked I_Kr was Zn²⁺, as shown in Fig. 4 A. The I_Kr amplitude was reduced by ~20% with 0.1 mM Zn²⁺ and by ~75% with 1 mM Zn²⁺. The records in Fig. 4 B illustrate a preparation in which I_Kr was completely blocked by 1 mM Zn²⁺, as indicated by the tail currents after a voltage step to +20 mV. These results also show a significant increase in time-dependent current during the voltage step itself, which we have attributed to a voltage shift of inactivation of the transient outward current, I_A, by Zn²⁺, as has previously been shown (Agus et al., 1991). The results in Fig. 4 A indicate that block of I_Kr occurred without a voltage shift of I_Kr activation. The curve describing the control results in Fig. 4 A corresponds to the Boltzmann equation, with V_1/2 = 12 mV and V_S = 9 mV. The other two curves in Fig. 4 A are the same as the control, but are scaled by (1 + [Zn²⁺]/K_D)¹, with K_D = 0.4 mM. In other words, the I_Kr channel appears to have a binding site on its external surface for Zn²⁺ with a dissociation constant of 0.4 mM. The channel is blocked when the site is occupied by Zn²⁺.

View larger version (18K): [in this window] [in a new window]   FIGURE 4   Block of IKr by Zn2+. (A) Activation curves for IKr obtained according to the protocol in Fig. 3 A in the control () and in saline containing 0.1 mM Zn2+ (n = 6) () and 1 mM Zn2+ (n = 5) (). All results were normalized by the maximum tail current obtained in control. The curves correspond to (1 + exp((V  12)/9)1 (1 + [Zn2+]/0.4)1, where [Zn2+] is either 0, 0.1, or 1.0 mM. (B) Records obtained with a step potential to +20 mV in control and with 1 mM Zn2+. The tail current in the control (indicated by the arrow) was essentially absent from the test results. We have attributed the significant time-dependent current during the step to +20 mV in the test record to IA, as indicated in the Results. Calibrations are 0.5 s and 0.1 nA.

Barium (1.0-5.0 mM) has often been used with cardiac preparations to remove I_K1 during voltage clamp steps (DiFrancesco, 1981; Brochu et al., 1992), so our expectation was that it would potently block I_K1 in our experiments as well. We did observe significant block of I_K1 by Ba²⁺ (0.5-2 mM) in steady state. However, the block was time-dependent under these conditions (even in the presence of 1 µM E-4031). To our knowledge, similar results have not been reported for cardiac cells, although Standen and Stanfield (1978) observed time-dependent block of I_K1 by Ba²⁺ in skeletal muscle, and Tang and Yang (1994) reported similar results with Ba²⁺ in hIRK2, an inward rectifier channel cloned from human brain that is also found in cardiac tissues.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

We have investigated the effects of various divalent cations on I_Kr in a mammalian cardiac ventricular myocyte preparation. Our results concerning the shift of the I_Kr activation curve and the acceleration of the I_Kr channel closing rate are qualitatively similar to the effects of divalent cations on several other preparations, beginning with the original work on this topic by Frankenhaeuser and Hodgkin (1957) on squid giant axons. Those results were originally attributed to a surface charge mechanism, although several reports have demonstrated that this theory, at least in its simplest form, does not account for many of the relevant experimental observations (Gilly and Armstrong, 1982a,b; Armstrong and Cota, 1990). The effects of divalent cations on I_Kr amplitude, specifically our results with Cd²⁺, Ni²⁺, Co²⁺, and Mn²⁺, are qualitatively different from the effects of these ions on channel activation. The observation that Cd²⁺ increased I_Kr amplitude significantly more than Ni²⁺, whereas the two ions shifted channel activation by approximately the same amount, further supports the idea that the loci on the I_Kr channel for these two types of effects are different. The effect of Cd²⁺ on I_Kr amplitude was originally reported by Follmer et al. (1992) for cat ventricular myocytes. They attributed their result to an interaction between Cd²⁺ and the mechanism responsible for inward rectification of the I_Kr channel, which they assumed was either an ion in the cytoplasm, or a membrane-bound, positively charged particle, similar to our original hypothesis concerning I_Kr rectification (Shrier and Clay, 1986). This view requires revision based on the work of Smith et al. (1996) and Spector et al. (1996), who have clearly shown that the apparent rectification of I_Kr is attributable to a very rapid, voltage-gated inactivation mechanism. We have proposed one way in which Cd²⁺, Mn²⁺, Ni²⁺, and Co²⁺ might alter inactivation gating to produce an increase in I_Kr amplitude: a direct interaction between divalent cations and the inactivation gate. Alternatively, the effect could be attributable to a surface charge mechanism, i.e., a voltage shift of the inactivation curve similar to the shift of the activation curve produced by these cations, as shown for Cd²⁺ in Fig. 3 A. This mechanism appears not to be applicable to our results, as illustrated in Fig. 5 A. The results in Fig. 3 Bthe fully activated current-voltage relations in control and in the presence of 5 mM Cd²⁺have been reproduced in Fig. 5 A. The theoretical curve describing the control results in Fig. 5 A is the same as in Fig. 3 B, i.e., I_Kr = 0.043(V + 70)/(1 + 6 exp(0.05V)) (normalized as described in the Fig. 3 B legend). From this analysis the inferred inactivation curve is given by 1/(1 + 6 exp(0.05V)), which is shown in Fig. 5 B (solid line). A rightward shift of this curve by 30 mV is illustrated in Fig. 5 B by the dashed line. The current-voltage relation that we would have obtained in these experiments if this simple shift of inactivation had occurred is shown in Fig. 5 A (dashed line). This prediction deviates significantly from the experimental results. To describe these results, the inactivation curve would have to not only be shifted rightward along the voltage axis, but also steepened considerably, which is effectively what occurs in our model of the Cd²⁺ results given above (Results). Measurements of inactivation kinetics may help to distinguish between the two mechanismsa direct interaction between divalent cations and the inactivation gate, or a mechanism more closely related to a surface charge effect.

View larger version (17K): [in this window] [in a new window]   FIGURE 5   Predictions of a simple shift of the IKr inactivation for the fully activated IKr current-voltage relation. The data in A were taken from Fig. 3 B. The theoretical curve that described the control results is 0.043(V + 70)/(1 + 6 exp(0.05V)), from which the steady-state inactivation curve can be inferred to be (1 + 6 exp(0.05V))1, shown in B by the solid line. The dashed curve in B illustrates a 30-mV rightward shift of this curve along the voltage axis, i.e., (1 + 6 exp(0.05(V  30))1. The corresponding prediction for the fully activated current-voltage relation is shown in A by the dashed line, i.e., 0.043(V + 70)/(1 + 6 exp(0.05(V  30))). This curve does not provide a good description of the 5 mM Cd2+ results, as described in the text.

Our results with Co²⁺ appear to be inconsistent with the work of Baro and Escande (1989) and Sanguinetti and Jurkiewicz (1991), who found that Co²⁺ eliminated a small component or "hump" of outward current with a voltage ramp, which the latter authors attributed to block of I_Kr. The analysis in Fig. 6 demonstrates that this result is attributable instead to a voltage shift of the I_Kr activation curve into a voltage range where rectification of its fully activated current-voltage relation is even steeper than in control. We have used our Cd²⁺ results to illustrate this point. The theoretical description of steady-state activation of I_Kr in control and with 5 mM Cd²⁺ (Fig. 3 A) is reproduced in Fig. 6 A. Similarly, the fully activated I-V relations in control and with 5 mM Cd²⁺ in Fig. 3 B are reproduced in Fig. 6 B. The contribution of I_Kr to net current during a slow voltage ramp is approximately given by its steady-state amplitude, because I_Kr activation is relatively rapid. These results correspond to the product of the steady-state activation and the fully activated I-V curves, which are given in Fig. 6 C for the control and for 5 mM Cd²⁺ (curves a and b, respectively). The difference between these results, given in Fig. 6 D, is to be compared with the experimental results for 3 mM Co²⁺ in figure 2 C of Baro and Escande (1989), and Figure 1 B of Sanguinetti and Jurkiewicz (1991). (Note the block of I_K1 by Co²⁺ in the latter result.) That is, millimolar concentrations of Cd²⁺ and Co²⁺ (and Mn²⁺ and Ni²⁺) reduce outward steady-state current while at the same time increasing peak outward I_Kr. The analysis in Fig. 6 provides a resolution to this paradox. Moreover, it further illustrates that divalent cations do not simply shift the I_Kr inactivation curve along the voltage axis. If this mechanism applied, then the apparent "block" of I_Kr by Co²⁺ would not have been observed by Sanguinetti and Jurkiewicz (1991). Rather, the difference current from the voltage ramps with 3 mM Co²⁺ and in control would have been biphasic.

View larger version (24K): [in this window] [in a new window]   FIGURE 6   Analysis of the apparent "block" of IKr by Co2+ during voltage ramps (Baro and Escande, 1989; Sanguinetti and Jurkiewicz, 1991). (A) Theoretical representation of the steady-state activation of IKr in control and 5 mM Cd2+ conditions (taken from Fig. 3 A). (Because the effects of Co2+ are quantitatively similar to those of Cd2+, we have used our Cd2+ results in this analysis.) (B) Fully activated current-voltage relations in control and 5 mM Cd2+ (taken from Fig. 3 B). (C) Steady-state IKr. Because of the rapid activation of IKr, its contribution to net current during a slow ramp is essentially given by the product of its steady-state activation and fully activated I-V curves, as shown here. (D) Apparent block of IKr during a ramp by Cd2+ (or Co2+). The result shown here is the difference between curves a and b in C. This is the current that would be obtained in a full ionic model of the preparation by subtracting the steady-state I-V curve after the addition of 5 mM Cd2+ to the external solution from the control I-V curve. That is, an outward current appears to be blocked by Cd2+, whereas the analysis in A-C demonstrates that this result is attributable to a voltage shift of the activation (A) into a voltage range where the rectification of the fully activated I-V curve is even steeper than in the control (B).

The effects of divalent cations on K channels have been investigated by the use of reducing agents that neutralize the positive charge of the amino group of lysines and histidines, thereby altering the sensitivity of the channel to divalent cations, such as Zn²⁺ and Cd²⁺ (Spires and Begenisich, 1994, and references therein). Similar experiments with I_Kr might help to further define the effects of divalent cations on its inactivation gate. Site-directed mutagenesis has been used to localize the amino acid residue where Cd²⁺ binds to the sodium channel in cardiac and skeletal muscle (Backx et al., 1992). Similar experiments with the HERG K channel may help to localize the Cd²⁺ binding site on its inactivation gate and provide insight into the relative sensitivity of this site for the various different divalent cations.