Structural Requirements of Human Ether-a-go-go–  related Gene Channels for Block by Bupivacaine

All animal experiments were conducted in accordance with institutional guidelines and approved by local authorities (Behörde für Soziales, Familie, Gesundheit und Verbraucherschutz, Freie und Hansestadt Hamburg, Germany).

The HERG mutants T623A, S624A, V625A, G648A, Y652A, Y652T, F656A, and F656T were created by site-directed mutagenesis.17All constructs were cloned in the pGEM expression vector for complementary RNA (cRNA) synthesis. After linearization of the plasmid with Not  I (Fermentas, St. Leon-Rot, Germany), the cRNA was synthesized in vitro  with the mMESSAGE mMACHINE Kit (Ambion, Austin, TX) according to the manufacturer’s protocol. The cRNA was purified with a phenol/chloroform extraction. The integrity of the cRNA was analyzed in a denaturating gel. The concentration was determined with the RiboGreen method (RiboGreen RNA Quantification Reagent, Molecular Probes, Eugene, OR).

For preparation of oocytes, female Xenopus laevis  frogs (Nasco, Fort Atkinson, WI) were anesthetized with tricaine solution (0.6 g/500 ml) for approximately 20 min, and ovarian lobes were removed surgically. The oocytes were separated mechanically, and the connective tissue was digested by collagenase A (Roche Diagnostics GmbH, Mannheim, Germany) for 2–4 h. Afterward, good oocytes (stages V and VI) were selected and incubated at 17°C in gentamicin containing oocyte ringer solution (75 mm NaCl, 2 mm KCl, 2 mm CaCl², 1 mm MgCl², 5 mm Na-pyruvate, and 5 mm HEPES, pH adjusted to 7.5 with NaOH, 50 μg/ml gentamicin; all from Sigma, Deissenhofen, Germany). The cRNA was injected (50 nl per oocyte) by using a Nanoliter injector (World Precision Instruments, Saratoga, FL). The total concentration of cRNA was 80 ng/μl. Two electrode voltage clamp experiments were performed 2–5 days after the cRNA injection.

Two-electrode voltage clamp recordings were performed with an Oocyte-Clamp OC-725C amplifier (Warner Instrument Corporation, Hamden, CT) and Pulse software version 8.50 (HEKA Electronik, Lambrecht, Germany). Sharp electrodes were pulled from borosilicate glass capillary tube (World Precision Instruments) on a vertical puller (Narishige, Tokyo, Japan) and filled with 3 m KCl and 3% agar. The standard extracellular recording solution contains 79.5 mm NaCl, 2 mm KCl, 2 mm CaCl², 1 mm MgCl², and 5 mm HEPES (all from Sigma, Deissenhofen, Germany), pH adjusted to 7.5 with NaOH. For some mutants that showed small currents under standard conditions, an extracellular solution with 20 mm KCl (containing 97 mm NaCl, 20 mm KCl, 1.8 mm CaCl², and 10 mm HEPES) was used. Different pulse protocols were used for characterization of the channels and to establish their pharmacologic sensitivities. The holding potential was −80 mV for all experiments. For the activation protocol, depolarizing pulses were applied from −80 to +60 mV in 10-mV steps. The duration of the depolarization was 1 s, and tail currents were recorded at −120 mV. The steady state inactivation was assessed by a three-step protocol: First, the membrane was depolarized for 2 s from −80 to +40 mV; then, pulses from −120 to +60 mV in 10-mV steps were applied for 30 ms; and finally, the membrane potential was +40 mV for 0.5 s. For pharmacologic experiments, single square pulses to +0 or +40 mV for 2 s were used. Tail currents were recorded at −40 or −120 mV. Repetitive pulses were applied to determine that steady state inhibition was reached.

Bupivacaine HCl (Sigma) was dissolved in the respective extracellular recording solution. A hydrostatically driven perfusion system was used to apply the drugs onto the oocytes and to exchange the extracellular solutions. All experiments were performed at room temperature.

Data were analyzed with Pulse Fit software (HEKA Electronik) and with KaleidaGraph software (Synergy Software, Reading, PA). The normalized tail currents during the activation protocol were fitted by a modified Boltzmann equation: I = I^max/[1 + exp((V^0.5− V^m)/k), where V^0.5is the voltage of half-maximal activation, V^mis the membrane potential, and k is the slope factor. For the analysis of the steady state inactivation, maximal current amplitudes during the tail were corrected for the closing during the hyperpolarizing step at very negative potentials as described previously.18The corrected curves were fitted with a Boltzmann function. The time course of inactivation was fitted with a monoexponential function after test pulse to −120 mV: y = C + Aⁱ(−t/τ^inact), with the time constants of inactivation τ^inactand the amplitude Aⁱ. The inhibition of currents by local anesthetics was quantified by the reduction of the maximal current during the test pulse. The fractional block f was calculated by the following formula: f = 1 − (I^{max, drug}/I^{max, control}). Concentration–response curves were fitted by a Hill function: f = 1/[1 + (IC⁵⁰/c)^h], where IC⁵⁰is the concentration of half-maximal inhibition, c is the concentration of the local anesthetic, and h is the Hill coefficient. The time course of the onset of block was analyzed using the following equation: y = I^bupi/I^con, where I^conis the current at any given time under control conditions and I^bupithe current at the same time after application of bupivacaine. An exponential function with one or two time constants was used to fit the time course of the onset of block: y = C + A¹exp(−t/τ¹) + A²exp(t/τ²), with the time constants τ¹and τ², and the amplitudes A¹and A². Statistical significance was tested using a two-sided Student t  test (Excel, Microsoft, Redmond, WA). Data are presented as mean ± SD unless stated otherwise; n values indicate the number of experiments.

Before evaluating the effect of the mutations on local anesthetic sensitivity of HERG channels, the electrophysiologic properties of the mutant channels had to be characterized first. Representative current traces evoked with the activation protocol are shown in figure 1. Tail currents were recorded at a potential of −120 mV because some of the mutants (HERG^T623A, HERG^V625A, and HERG^F656A) gave rise to only small tail currents at −40 mV. The mutants HERG^G648Aand HERG^Y652Twere recorded at an extracellular potassium concentration of 20 mm because small currents possibly due to poor membrane expression did not allow measurements under standard conditions. This potassium concentration allowed recording larger inward tail currents at negative potentials without profoundly affecting inhibition of HERG channels.19,20 

The mutations S624A, Y652A, and F656T had only a small influence on voltage-dependent activation (figs. 2A and B). The current–voltage relation determined at the end of the activating test pulse exhibited the typical bell-shaped relation that was similar to the current–voltage relation of HERG^wtchannels (figs. 2A and B). In contrast, the mutations T623A and V625A caused a loss of the inward rectification and showed a linear current–voltage relation (fig. 2A). Outward currents of HERG^F656Achannels were too small to be analyzed. The current–voltage relation of HERG^Y652Tchannels reached a plateau between membrane potentials of 0 to +50 mV (fig. 2C). The mutation G648A resulted in a linear current–voltage relation between membrane potentials of −120 and +40 mV (fig. 2C). The current–voltage relation of tail currents were described with Boltzmann functions (figs. 2D–F). Apart from HERG^Y652A, all mutations significantly shifted the midpoint of current activation either in the depolarizing or in the hyperpolarizing direction (figs. 2D–F). The parameters of the Boltzmann functions are summarized in table 1. The influence of the mutations on voltage dependence of inactivation was analyzed next (fig. 3). The mutant HERG^T623Agave rise to only small outward currents after recovery from inactivation, and the mutation V625A completely abolished inactivation (fig. 3A). The voltage dependence of inactivation was described with Boltzmann functions (fig. 4). Parameters of the Boltzmann functions are given in table 1. Whereas Y652A shifted the inactivation midpoint into the depolarizing direction and Y652T did not influence the inactivation midpoint, all other mutations shifted the midpoint of current inactivation into the hyperpolarizing direction. Time constants of inactivation at −120 mV remained unchanged by F656T, were slowed by S624A, and were accelerated by all other mutations (table 1).

Concentration dependence of HERG^wtinhibition was established first to determine a concentration suitable for comparing the sensitivities of mutant and wild-type channels. Inhibition was analyzed as inhibition of the current elicited by the depolarization to 0 mV (I^test) and as inhibition of the tail current elicited by a step to −40 mV (I^tail−40) and to −120 mV (I^tail−120), respectively (fig. 5A). The concentration response data were fitted with Hill functions (fig. 5B). The parameters of the Hill functions are given in table 2. Deactivating tail currents at −120 mV were less sensitive to inhibition by bupivacaine than tail currents elicited at −40 mV as well as currents elicited by the depolarizing pulse to 0 mV. The extent of HERG^wtchannel inhibition by 300 μm bupivacaine was significantly altered by extracellular K⁺when analyzed at a test potential of 0 mV (2 mm K⁺: 84 ± 3%, n = 7; 20 mm K⁺: 76 ± 5%, n = 6; P < 0.01). However, inhibition remained unchanged when analyzed as inhibition of tail currents at a membrane potential of −120 mV (2 mm K⁺: 62 ± 12%, n = 7; 20 mm K⁺: 58 ± 7%, n = 6; P > 0.05).

Because the IC⁵⁰value for inhibition of HERG^wtchannels was dependent on the test protocol (table 2), a concentration of bupivacaine (300 μm) was chosen that allowed comparing inhibitory effects on mutated and wild-type channels with all protocols (fig. 6). For the mutant channels HERG^S624A, HERG^Y652A, and HERG^F656T, inhibition of currents elicited during pulses to 0 mV (figs. 6A and E) and inhibition of tail currents at −40 and −120 mV was analyzed (figs. 6A, E, and F). All three mutations reduced inhibition of HERG channels at all protocols, with the effect being most pronounced for HERG^F656T. The mutant channels HERG^T623Aand HERG^V625Aexhibited only small outward currents. Therefore, inhibition of outward currents was analyzed at +40 mV, and tail current inhibition was analyzed at −120 mV (figs. 6B and F). The mutations T623A and V625A significantly reduced the inhibitory effect of bupivacaine at both test potentials. The mutation F656A only allowed measuring inhibition of tail currents at −120 mV (figs. 6C and F). The mutation significantly reduced the inhibitory effect of bupivacaine. Inhibition of HERG^G648Aand HERG^Y652Thad to be analyzed at an extracellular K⁺concentration of 20 mm (figs. 6D and F). Inhibition of HERG^wtchannels was analyzed under these conditions as well. Both G648A and Y652T significantly reduced inhibition when compared with wild-type channels (figs. 6D and F).

The onset of block was analyzed for HERG^wtand all mutants except for HERG^F656A. Outward currents conducted by this mutant were too small for analysis. Exemplary plots are shown in figure 7. Onset of block developed biexponentially for HERG^wtunder either extracellular K⁺concentrations (2 mm K⁺: τ^{onset 1}= 42 ± 11 ms, τ^{onset 2}= 300 ± 64 ms, n = 7; 20 mm K⁺: τ^{onset 1}= 44 ± 13 ms, τ^{onset 2}= 374 ± 54 ms, n = 6). The slower time constant τ^{onset 2}differed significantly between both conditions (P < 0.05). Both time constants were significantly slowed for HERG^S624A(fig. 8A). For HERG^V625A, τ^{onset 1}decreased and τ^{onset 2}remained unchanged compared with HERG^wt. For the mutants T623A, G648A, Y652A, Y652T, and F656T, the onset of block was best described with only a single time constant (fig. 8A). The onset of block was accelerated for HERG^Y652Aand slowed for HERG^T623AHERG^G648A, and HERG^Y652T. It remained unchanged for HERG^F656T. Whereas the onset of block did not significantly differ between HERG^Y652Aand HERG^F656T, it was slower for HERG^Y652Tthan for HERG^Y652Aand HERG^F656T. After the initial onset of block HERG^T623A, HERG^Y652A, HERG^Y652T, and HERG^F656Tpartially recovered from block (fig. 7). Time constants for partial recovery from block were not significantly different between these mutations (fig. 8B).

The aromatic residues F656 and Y652 in the S6 region of HERG channels have been identified to be important for local anesthetic interaction.8The results of this study confirmed the role of these aromatic amino acids in local anesthetic block and extended the analysis of molecular determinants of HERG channel block to residues residing deeper within the pore. The contribution of three residues near the selectivity filter, T623, S624, and V625, was analyzed in this work by mutating these residues to alanine. All three mutations altered the gating behavior of the channels. T623A and V625A eliminated inward rectification of HERG channels and altered the typical bell-shaped voltage dependence of outward currents into a linear current–voltage relation. The mutation V625A furthermore abolished inactivation of HERG channels. The mutation S624A shifted the voltage dependence of activation and inactivation without larger changes of the macroscopic current behavior. Although it cannot be excluded that the reduced inhibition by bupivacaine is related to the changes in channel gating, this seems less likely because the changes in gating are qualitatively different, but the extent of reduction in inhibition is similar for all three mutations.

The role of inactivation in HERG channel block is controversial: Lipka et al . demonstrated that abolishing C-type inactivation by introducing the double mutant HERG^G628C/S631Creduces inhibition by bupivacaine5similar to other HERG channel blockers such as dofetilide21and E-4031.19In contrast, other mutations affecting HERG inactivation yield inconsistent results concerning the inhibition by high-affinity4,22as well as by low-affinity blockers.23Also, inhibition of HERG channels by cocaine is only affected by some but not all mutations influencing inactivation.24The results of this study demonstrate that inhibition of HERG channels by bupivacaine is reduced to a similar extent by a mutation that accelerates inactivation (T623A) and by a mutation that abolishes inactivation (V625A). Reduced bupivacaine affinity of noninactivating HERG mutant channels (Lipka et al .,5this study) may thus result from inactivation gating-associated reorientation of residues in the S6 domain that comprise the bupivacaine binding site rather then resulting from a direct effect of the inactivation. Taken together, our findings may therefore indicate that inactivation is neither a sufficient nor a necessary prerequisite for low-affinity block of HERG channels by bupivacaine. The results furthermore suggest that these three residues at the base of the pore helix immediately adjacent to the GFG signature sequence of the selectivity filter are potential binding sites not only for some high-affinity blockers4,14,25but also for the low-affinity blocker bupivacaine. Mutations from threonine to alanine at position 623 and serine to alanine at position 624 both remove the hydroxyl group of the amino acid side chain. This group may be responsible for mediating polar interactions with hydrophilic parts of the bupivacaine molecule. Because loss of the hydroxyl group at the position T623 caused a more pronounced reduction of inhibition by bupivacaine than loss of the same group at position S624, it may be argued that position T623 is more important for interaction with bupivacaine. On the other site, it may be argued that the small difference in inhibition between these mutation points to a nonpolar rather than a polar interaction of the drug molecule with the ion channel protein that does not depend on the hydroxyl group of the amino acids.

The mutation G648A in the S6 helix poorly expresses under standard conditions12and was thus analyzed with an extracellular solution containing 20 mm potassium. These channels were continuously open, even at very negative membrane potentials. Voltage dependence of activation and inactivation were shifted to negative potentials. Because glycine does not possess a side chain that could account for bupivacaine interaction, it is surprising that introducing a methyl group at this position by the mutation G648A did not increase but decreased anesthetic sensitivity. The reduced inhibition by bupivacaine for this mutant may thus also be related to gating-associated reorientation of residues in the S6 domain rather then to a direct involvement of this amino acid in bupivacaine binding.

The aromatic amino acids Y652 and F656 have been implicated in the interaction of local anesthetics with HERG channels.8Mutating either aromatic amino acid to alanine significantly reduced local anesthetic sensitivity of HERG channels. However, these mutations also introduce gating changes that may explain altered local anesthetic sensitivity.8These gating changes are qualitatively and quantitatively different, and they are much more pronounced for F656A than for Y652A. The introduction of threonine in either position had the opposite effects on channel gating than the introduction of alanine. Albeit being again qualitatively and quantitatively different, they were more pronounced for Y652T than for F656T. The consequences of these mutations therefore allowed comparing the effects on bupivacaine sensitivity of two different mutations at either position causing different effects on channel gating. Regardless of the effects on channel gating, however, all four mutations at positions 652 and 656 reduced the inhibition by bupivacaine. These results, therefore, confirm and extent our previous observations8that aromaticity at positions 652 and 656 is an important prerequisite for local anesthetic sensitivity of HERG channels.

Analysis of the onset of block revealed a partial recovery from block for mutants at positions 623, 652, and 656. Partial recovery from block may explain why mutating the aromatic residues Y652 and F656 as well as T623 reduced inhibition of HERG channels by bupivacaine the most. This observation supports the notion that these three amino acids are pivotal in the interaction of local anesthetics with HERG channels. It may, furthermore, be suggested that—after a first contact with the pore—stable interactions with the bupivacaine molecule cannot be established without these amino acids. It may be hypothesized that interaction with only one of these amino acids is required for initial bupivacaine binding. Interaction with all three amino acids would, however, be required for establishing a stable drug–protein interaction. The time constants both for the onset of block and for the partial recovery from block did not differ between the mutants Y652A and F656T, the mutation that altered channel gating the least. This observation may suggest that local anesthetic interaction with either aromatic amino acid is mechanistically similar and does not depend on channel gating. A partial recovery from block for the mutation Y652A has been described for the inhibition of HERG channels by chloroquine, quinidine, and vesnarinone.26,27In contrast to Y652A, a partial recovery from block was not detectable for the mutant Y652F.26,27This result is compatible with the idea that aromaticity at this position is important for establishing a stable interaction between HERG channels and the drug molecule. In addition to this aromatic amino acid at position 652 the residues F656 as well as T623 near the selectivity filter seem a contributing factor for stable interaction of local anesthetics with HERG channels.

In summary, the results of this study support the hypothesis that aromaticity in positions 652 and 656 is an important feature in bupivacaine binding to HERG channels. The results furthermore define a significant contribution of amino acids T623 and S624 as well as V625 near the selectivity filter to the interaction of the low-affinity blocker bupivacaine with HERG channels. Moreover, our results indicate that inactivation is neither a sufficient nor a necessary prerequisite for inhibition of HERG channels by bupivacaine. Therefore, despite being low-affinity blockers, amino-amide local anesthetics seem to share a common mode of interaction with several high-affinity blockers of HERG channels.

The authors thank Olaf Pongs, Ph.D. (Director of the Institute of Neural Signal Transduction, University Medical Center Hamburg-Eppendorf, Germany), for his continuous support, and Alexander Schwoerer, M.D. (Department of Vegetative Physiology and Pathophysiology, University Medical Center Hamburg-Eppendorf), for critically reading the manuscript.