Abstract
The inwardly rectifying potassium current of the cardiomyocyte (IK1) is the main determinant of the resting potential. Ion channels Kir2.1, Kir2.2, and Kir2.3 form tetramers and are the molecular correlate of macroscopic IK1 current. Verapamil is an antiarrhythmic drug used to suppress atrial and ventricular arrhythmias. Its primary mechanism of action is via blocking calcium channels. In addition, it has been demonstrated to block IK1 current and the Kir2.1 subunit. Its effect on other subunits that contribute to IK1 current has not been studied to date. We therefore analyzed the effect of verapamil on the Kir channels 2.1, 2.2, and 2.3 in the Xenopus oocyte expression system. Kir2.1, Kir2.2, and Kir2.3 channels were heterologously expressed in Xenopus oocytes. Respective currents were measured with the voltage clamp technique and the effect of verapamil on the current was measured. At a concentration of 300 µM, verapamil inhibited Kir2.1 channels by 41.36% ± 2.7 of the initial current, Kir2.2 channels by 16.51 ± 3.6%, and Kir2.3 by 69.98 ± 4.2%. As a verapamil effect on kir2.3 was a previously unknown finding, we analyzed this effect further. At wash in with 300 µM verapamil, the maximal effect was seen within 20 min of the infusion. After washing out with control solution, there was only a partial current recovery. The current reduction from verapamil was the same at − 120 mV (73.2 ± 3.7%), − 40 mV (85.5 ± 6.5%), and 0 mV (61.5 ± 10.6%) implying no voltage dependency of the block. Using site directed mutations in putative binding sites, we demonstrated a decrease of effect with pore mutant E291A and absence of verapamil effect for D251A. With mutant I214L, which shows a stronger affinity for PIP2 binding, we observed a normalized current reduction to 61.9 ± 0.06% of the control current, which was significantly less pronounced compared to wild type channels. Verapamil blocks Kir2.1, Kir2.2, and Kir2.3 subunits. In Kir2.3, blockade is dependent on sites E291 and D251 and interferes with activation of the channel via PIP2. Interference with these sites and with PIP2 binding has also been described for other Kir channels blocking drugs. As Kir2.3 is preferentially expressed in atrium, a selective Kir2.3 blocking agent would constitute an interesting antiarrhythmic concept.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
Among the potassium currents of the heart, the inwardly rectifying current IK1 is the main determinant of the resting potential. It is active during the late phase 3 and during the phase 4 of the action potential (Hibino et al. 2010).
Kir2.1 channel subunit together with Kir2.2 and Kir2.3 are the molecular correlates of cardiac Ik1 current (Nerbonne and Kass 2005). Kir2.1, 2.2, and 2.3 form heterotetramers. Four Kir subunits are necessary for the assembly of one channel pore. There is evidence of regional dispersion of expression of these subunits in different parts of the heart. Kir2.1 has been shown to be expressed in the whole myocardial tissue, whereas Kir2.3 is predominantly expressed in the atria (Gaborit et al. 2007).
Ik1 current is of particular interest as it has been involved in the pathogenesis of arrhythmias (Nattel et al. 2007). In both simulation and experimental models, an increase of Ik1 currents has been demonstrated to attenuate the occurrence of focal arrhythmia and provoke reentry based arrhythmias (Noujaim et al. 2007). An overexpression of Kir2.1 has been demonstrated to be a part of the electrical remodeling that occurs in the setting of atrial fibrillation (Luo et al. 2013, Girmatsion et al. 2009). A gain of function mutation in human Kir2.1 channel has been identified in a genetic syndrome of familial atrial fibrillation (** protocol before (A) and after (B) incubation in verapamil. C The current–voltage relationships before and after verapamil application. D Overview of the current change from baseline for Kir2.1, 2.2, and 2.3 at − 120 mV in control solution and verapamil solution, respectively. For Kir2.1 expressing oocytes, verapamil causes a current reduction by 41.36% ± 4.2% (N = 7) of the initial current. With Kir2.2 expressing oocytes, the current reduction is 16.51 ± 3.6% (N = 9) of the initial current. In Kir2.3 expressing oocytes, the current reduction reached 69.98 ± 4.2% of the initial current (N = 7). In control solution, the current showed a small increase compared to the initial current of 12.3 ± 2.7% (N = 7) for Kir2.1, 22.8 ± 4.9% for Kir2.2, and 15 ± 2.9% for Kir2.3
Concentration-dependent inhibition of Kir2.3 currents
As the effect was most prominent on Kir2.3 channels, we focused the further analysis on these channels. We therefore measured the effect of verapamil at different concentrations. Kir2.3 expressing oocytes were incubated for 30 min either in verapamil at the respective concentration or in control solution for 30 min. The currents were measured as a % change normalized to the current after incubation for the same time in control conditions. At a concentration of 0.1 µM verapamil, we measured no effect with the normalized current unchanged at 104.2 ± 4.9% (N = 6). At concentration of 1 µM, no effect was observed either with a normalized current at 105.3 ± 3.02% (N = 6). At a substance concentration of 6 µM, no effect was observed with a respective normalized current of 107 ± 4.6% (N = 6). At a concentration of 50 µM, there was a reduction to 87.5 ± 5% (N = 6) of the control current, whereas at a concentration of 100 µM there was a reduction to 50.8 ± 5.4% (N = 5). At a concentration of 150 µM, the current was reduced to 30 ± 4.2% (N = 6), and at 300 µM the current was reduced to 27.5 ± 4.8% (N = 6). These results were then fitted to the Hill equation to obtain the concentration/effect relation. This result is shown as plot on Fig. 2. The results yielded an IC50 of 58.10 µM, which represents a low affinity of the drug on the channel in this expression system (supplement2.xls).
Dose response curve of Kir2.3 channels at different verapamil concentrations. After fitting the results to the Hill equation, we obtained an IC50 of 58.10 µM which represents a low affinity of the drug in the oocyte expression system
The verapamil effect initiates rapidly and is only partially reversible upon wash-out
To further elucidate the blocking properties, we applied a wash-in and wash-out protocol. Oocytes expressing Kir2.3 channels were clamped at a holding potential of − 80 mV in control solution. After that an infusion of 300 µM verapamil was initiated (wash-in). At intervals of 5 s, a voltage command to − 120 mV was given and the inward currents were measured. After the maximum current reduction was observed in form of a steady state of current, the infusion was changed again to control solution (wash-out) and the same voltage command protocol was applied. These data are shown in Fig. 3. After 20 min, we observed the maximal effect of the verapamil infusion with a relative current of 25.8 ± 3% of the initial current, so we conclude that the effect initiates relatively quickly. At the wash-out phase, we observed only a slight recovery of the current to a level of 46.09% ± 7 relative to the initial current (p = 0.03). We therefore concluded only a partially reversible verapamil effect (N = 6, Fig. 3, supplement3.xls).
Wash-in: relative current change over time after verapamil infusion in the bath solution. At intervals of 5 s, a voltage command to − 120 mV was given and the inward currents were measured. A steady state inhibition is seen after approximately 20 min (relative current of 25.8 ± 3% of the initial current). Wash-out: relative current change over time after change of the infusion of the bath solution with control solution. We observe a partial reversal of verapamil effect to a level of 46.09 ± 7% relative to the initial current
The block is not voltage dependent
Further we measured the effects of verapamil at different clam** potentials at − 120, − 40, and + 0 mV. We observed a reduction of the current of 73.2% ± 3.7 (N = 6) at − 120 mV, 85.5 ± 6.5% at − 40 mV, and 61.5 ± 10.6% at 0 mV. The ANOVA test showed no significant difference at these potentials (p = ns for the comparison). Hence, the block is not voltage dependent. The results are summarized in Fig. 4 (supplement4.xls).
Amount of block of verapamil on Kir2.3 channels at different clam** voltages. There was no significant difference at the amount of block at clam** potentials − 120 mV, − 40 mV, and 0 mV with current reduction (relative to the initial current) of 73.2 ± 3.7%, 85.5 ± 6.5%, and 61.5 ± 10.6%, respectively. ANOVA F value 2.58, p = 0.11
Delineating the blocking mechanism-pore mutants
A well described mechanism of block of Kir2.1 channels is via electrostatic interaction with residues forming the inner lining of the pore. These sites include for Kir2.1 channels, among other residues E224, F254, D259, and E299. These negatively charged amino acids serve as residues for electrostatic interaction with drugs. Drug binding in these regions narrows the channel pore and blocks the flow of potassium ions. This is the case for drugs chloroquine (Rodríguez-Menchaca et al. 2008; Noujaim et al. 2010) and pentamidine (de Boer et al. 2010). Equivalent residues on Kir2.3 channels have been identified via comparative alignment and have been used to screen for putative binding sites of quinidine on Kir2.3 channels (Koepple et al. 2017; de Boer et al. 2010). These residues include E291, D251, E216, and D247. Pore mutants E291A, D251A, E216A, and D247A (Koepple et al. 2017) were used to scan for putative molecular determinants of quinidine block.
Equivalently we generated pore mutants E291A, D251A, E216A, and D247A of Kir2.3 channels and measured the effect of verapamil as compared to the effect on wild type channels. The same voltage protocol was applied as with wild type channels. The currents were normalized to the control current for the comparison.
With pore mutant E291A (N = 6 for verapamil and N = 8 for controls), we observed a decrease of the verapamil effect compared to the wild type current. Normalized current was reduced to 27.05 ± 0.04% of the control current in wild type channels (N = 19 for control experiments and N = 6 for verapamil incubation) after verapamil treatment, whereas in mutant E291A it was reduced only to 73.46 ± 0.08% of the control current (p = 3.03 × 10−4).
With mutant D251A, we observed an abolishment of verapamil effect. Here the normalized current was minimally reduced to 94.9 ± 0.06% (N = 7 for verapamil and N = 4 for control) (p = 3.48 × 10−7 in comparison to the wild type channels). So we can conclude that these pore residues (E291 and D251) seem to be essential for the verapamil effect.
On the contrary the effect of verapamil was not different among wild type, D247A (N = 7 for verapamil and N = 8 for controls), and E216A (N = 7 for verapamil and controls respectively) mutant channels. The relative effect was 28.2 ± 0.06% with mutant D247A and 22.05 ± 0.02% with mutant E216A (p = ns for each comparison). The results are summarized in Fig. 5 (supplement6.xls).
Verapamil effect (displayed as normalized current) of different Kir2.3 mutants compared to wild type channels. The effect on mutants E216A and D247A was similar to wild type (WT) channels. On pore mutants E219A and D251A, there is a reduction of verapamil effect. The effect of verapamil is also reduced on PIP2 mutant I214L. * marks statistical significance
PIP2interference
Another mechanism of block on Kir2.3 channels is interference with regulating PIP. These phospholipids are essential for channel function probably by stabilizing the open conformation. For example, the drug mefloquine and carvedilol have been shown to interfere with Kir2.3 channel-PIP2 interaction, thus reducing the current. Mutant channel I214L was used to delineate this effect (Ferrer et al. 2011). For this mutant, a stronger PIP2 affinity has been described. If an involvement of PIP2 interference mechanism of verapamil action is present, we would expect a weaker effect of verapamil with this mutant (Du et al. 2004).
We therefore also generated the same mutant and compared the effect of verapamil on oocytes expressing the mutant channel and oocytes expressing WT channels. The oocytes were incubated in a bath solution of 300 µM verapamil or control for 30 min as described above. Measurements were performed before and after incubation. The voltage protocol applied was the same as aforementioned (incremental voltage steps from − 120 to + 40 V, holding potential − 80 mV). The measurements were made on the maximal inward current at − 120 mV. The currents were normalized to the control current for the comparison. In this case, the observed effect was indeed less pronounced. I214L mutant channels displayed a normalized current reduction to 61.9 ± 0.06% (N = 7 for verapamil and N = 14 for controls) of the control current, whereas wild type channels had a reduction of 27.36 ± 0.04% (p = 0.04 compared to the control current). Results are also shown in Fig. 5 (supplement6.xls).
