Y1767C, a novel SCN5A mutation, induces a persistent Na⁺ current and potentiates ranolazine inhibition of Na_v1.5 channels

Abstract

Long QT syndrome type 3 (LQT3) has been traced to mutations of the cardiac Na⁺ channel (Na_v1.5) that produce persistent Na⁺ currents leading to delayed ventricular repolarization and torsades de pointes. We performed mutational analyses of patients suffering from LQTS and characterized the biophysical properties of the mutations that we uncovered. One LQT3 patient carried a mutation in the SCN5A gene in which the cysteine was substituted for a highly conserved tyrosine (Y1767C) located near the cytoplasmic entrance of the Na_v1.5 channel pore. The wild-type and mutant channels were transiently expressed in tsA201 cells, and Na⁺ currents were recorded using the patch-clamp technique. The Y1767C channel produced a persistent Na⁺ current, more rapid inactivation, faster recovery from inactivation, and an increased window current. The persistent Na⁺ current of the Y1767C channel was blocked by ranolazine but not by many class I antiarrhythmic drugs. The incomplete inactivation, along with the persistent activation of Na⁺ channels caused by an overlap of voltage-dependent activation and inactivation, known as window currents, appeared to contribute to the LQTS phenotype in this patient. The blocking effect of ranolazine on the persistent Na⁺ current suggested that ranolazine may be an effective therapeutic treatment for patients with this mutation. Our data also revealed the unique role for the Y1767 residue in inactivating and forming the intracellular pore of the Na_v1.5 channel.

long QT syndrome (LQTS) is a congenital cardiac disorder characterized by a prolonged corrected QT (QT_c) interval (QT_c > 440 ms) on surface ECGs and has been linked to life-threatening arrhythmias and sudden cardiac death (24, 33). LQTS type 3 (LQT3) mutations cause dysfunctions of the cardiac Na⁺ channel (Na_v1.5) encoded by the SCN5A gene. The Na_v1.5 channel is a critical determinant of cardiac excitability and plays a central role in the initiation and propagation of cardiac action potentials. The underlying mechanism of LQT3 has been attributed to persistent Na⁺ currents through mutant Na_v1.5 channels leading to prolonged repolarization of cardiac action potentials (9, 13, 26, 27, 51).

Previous studies (5, 41) have revealed that the class I antiarrhythmic agents mexiletine [I(B)] and flecainide [I(C)] restore normal QT_c intervals in some LQT3 patients by blocking the persistent Na⁺ current. Ranolazine, an antiangina agent, has been recently shown to inhibit persistent Na⁺ currents associated with cardiac ischemia and to suppress early afterdepolarization (EAD)-triggered arrhythmias in an experimental model of LQT3 by blocking the persistent Na⁺ current (2, 52). Ranolazine has recently been successfully used to treat LQTS patients harboring Na_v1.5 ΔKPQ mutations (14, 34). In some instances, ranolazine has been shown to increase the QT interval by suppressing the delayed rectifier K⁺ current, but it has not been linked to substantial dispersion of myocardial repolarization, EAD, or torsades de pointes (2, 21).

While ranolazine is not classified as an antiarrhythmic drug, its binding site is located near the cytoplasmic entrance of the Na_v1.5 channel and overlaps with that of class I antiarrhythmics (14). Ranolazine binding is state dependent and preferentially interacts with open rather than closed or inactivated Na⁺ channels (47).

A 17-yr-old LQTS patient diagnosed with episodic cardiac arrhythmia was found to harbor a Y1767C mutation of the SCN5A gene in transmembrane segment 6 of domain IV. Y1767 of Na_v1.5 is a highly conserved tyrosine residue located in the membrane-spanning S6 segment of homologous domain IV that contributes to rapid inactivation and the binding of local anesthetics and antiarrhythmic drugs (36, 40). Another LQT3 mutation, V1763M, four amino acids upstream from Y1767, was previously characterized by persistent Na⁺ currents that were insensitive to lidocaine blocking (10). This finding suggested that traditional class I antiarrhythmic drugs are not an optimal therapeutic for the treatment of LQT3 patients with mutations in this specific region.

In the present study, we investigated the biophysical properties of the Y1767C mutant of Na_v1.5 transiently expressed in tsA201 cells. Y1767C mutant channels displayed typical LQT3-like biophysical properties, including an increased persistent current, accelerated inactivation kinetics, an increased window current, and faster recovery from fast inactivation. The persistent current of the Y1767C channel was potently blocked by ranolazine but not by other class I antiarrhythmics, suggesting that ranolazine may be a highly selective therapeutic treatment for LQTS patients carrying this mutation. Further analysis showed that Y1767 plays a pivotal role in ranolazine blocking.

MATERIALS AND METHODS

Clinical evaluation.

LQTS symptoms were defined as syncope and cardiac arrest. The patient underwent a detailed clinical, particularly cardiovascular, examination, including a standard 12-lead ECG, 24-h Holter monitoring, an echocardiogram, and an exercise stress test. The QT interval was measured in lead 2 of the ECG, and the resulting QT_c was determined by correcting for the heart rate using Bazett's formula. The patient was diagnosed with LQTS because he 1) exhibited syncope accompanied with a QT_c of >440 ms, 2) had no obvious abnormal cardiovascular structures, and 3) was not taking any medication that could lengthen the QT_c.

Molecular genetics of LQTS patients.

The local ethics committee approved the study protocol in accordance with the standards set out in the Declaration of Helsinki. The patient gave written informed consent before the clinical and genetic investigations. Genomic DNA was extracted from peripheral lymphocytes. Coding exons of the KCNQ1, KCNH2, SCN5A, KCNE1, and KCNE2 genes were amplified by PCR using flanking intronic primers according to the gene sequence described by Wang et al. (50). PCR products were directly sequenced and analyzed as previously described (27).

Na ⁺ channel mutagenesis.

Mutant human Na_v1.5/Y1767C was generated using QuickChange TM site-directed mutagenesis kits according to the manufacturer's instructions (Stratagene, La Jolla, CA). Oligonucleotide primers containing the corresponding Y1767C mutation were synthesized using following sequences: forward 5′-CTCATCGTGGTCAACATGTGCATTGCCATCATCCTG-3′ and reverse 5′-CAGGATGATGGCAATGCACATGTTGACCACGAT-3′ (the mutated site is underlined). Mutant and wild-type (WT) Na_v1.5 channels were inserted in the pcDNA1 plasmid, amplified in Esherichia coli DH5α, and purified using Qiagen columns (Qiagen, Chatsworth, CA).

Heterologous expression of tsA201 cells.

The tsA201 cell line is a modified human embryonic kidney-293 cell line stably transfected by the simian virus 40 large T antigen that can promote the replication of viral promoter-containing constructs (30). Cells were grown in high-glucose DMEM supplemented with FBS (10%), l-glutamine (2 mM), penicillin G (100 U/ml), and streptomycin (10 mg/ml) (GIBCO-BRL Life Technologies, Burlington, ON, Canada). Cells were incubated in a 5% CO₂ humidified atmosphere after being transfected with WT or mutant human Na_v1.5 cDNA (2 μg) and the human β₁-subunit (2 μg) using the calcium-phosphate method as previously described (13). The human Na⁺ channel β₁-subunit and CD8 were inserted in the pIRES bicistronic vector in the form of pCD8-IRES-β₁. Using this strategy, transfected cells that bound beads also expressed the β₁-subunit protein. Transfected cells were incubated in the medium containing anti-CD8-coated beads (Dynabeads CD8, Dynal Biotech, Oslo, Norway) for 2 min before patch-clamp experiments were performed (30). Cells expressing CD8 were distinguished from nontransfected cells by visualizing beads fixed on the cell membrane by light microscopy.

Patch-clamp electrophysiology.

The whole cell configuration of the patch-clamp technique was used to record macroscopic Na⁺ currents from transfected tsA201 cells. Patch-clamp recordings were obtained using low-resistance, fire-polished electrodes (<1 MΩ) made from 8161 Corning borosilicate glass coated with Sylgard (Dow-Corning, Midland, MI) to minimize electrode capacitance. Currents were recorded with an Axopatch 200 amplifier (Molecular Devices, Sunnyvale, CA), and series resistance was >80% compensated. Command pulses were generated, and currents were acquired using a Pentium-based computer running pCLAMP software (version 8.0) equipped with a DigiData 1300 AD converter (Molecular Devices). P/4 leak subtraction was used to compensate for linear leaks and eliminate capacitative transients. Currents were filtered at 5 kHz and digitized at 10 kHz. All recordings were performed at room temperature (22–23°C). Cells were permitted to stabilize for 10 min after the whole-cell configuration was established before currents were recorded.

Solutions and reagents.

For the whole cell recordings, the patch pipettes were filled with a solution containing 35 mM NaCl, 105 mM CsF, 10 mM EGTA, and 10 mM Cs-HEPES. pH was adjusted to 7.4 using 1 N CsOH. The bath solution consisted of 150 mM NaCl, 2 mM KCl, 1.5 mM CaCl₂, 1 mM MgCl₂, 10 mM glucose, and 10 mM Na-HEPES. pH was adjusted to pH 7.4 using 1 N NaOH (final Na⁺: 152.4 mM). Fluoride did not significantly affect the activation or inactivation states of WT Na_v1.5 channels (data not shown). The liquid junction potential between the patch pipette and the bath solution was corrected by −7 mV. Stock 4 mM (ranolazine, quinidine, and flecainide) and 20 mM (mexiletine) solutions were prepared and were diluted in bath solution. Effects of the drugs were recorded 10 min after application to allow for drug equilibration.

Homology modeling of the DIVS6 segment.

The DIVS6 segment of Na_v1.5 was modeled based on the crystal structure of the S6 segment of the MthK channel (25) using the ZMM software package (7, 53). The alignment of Na_v1.5 with MthK S6 segments was performed using ClustalW (11). Model optimization by Monte Carlo energy minimization was performed as previously described with some variations (25). Briefly, the model was constructed by first assigning the backbone geometry of the template to model residues. Residues that were conserved in both proteins were assigned the same side chain geometry. At nonconserved positions, the side chain torsion angles χ₁ and χ₂ were taken from the template, and additional torsion angles of longer side chains were assigned a value of 180°. Bad contacts in the models were removed using a multistep relaxation method consisting of four consecutive short MCM trajectories, which were terminated after 50 energy minimization steps.

Statistical analysis.

Recorded data were analyzed using a combination of pCLAMP software (version 9.0, Axon Instruments), Microsoft Excel, and SigmaPlot 8.0. Results are presented as means ± SE. Statistical comparisons were performed using an unpaired Student's t-test in SigmaStat (Jandel Scientific Software, San Rafael, CA). Differences were considered significant at P < 0.05.

RESULTS

Phenotypic characterization and clinical findings of a unique case of LQTS.

A 17-yr-old LQTS patient (subject 1; Fig. 1B) was found to experience an episodic cardiac arrest during emotional stress. The patient was admitted to the emergency department, where normal sinus rhythm was restored by conventional therapies. The ECG revealed a prolonged QT_c interval of 507 ms (lead 2). Prolonged QT_c was also observed in Holter monitoring in the absence of arrhythmias. The echocardiogram of the patient was unremarkable. An exercise stress test confirmed the QT_c prolongation diagnosis, with no signs of reduced coronary reserve or arrhythmia. β-Blocker (propranolol) therapy was initiated (2.5 mg·kg⁻¹·day⁻¹), after which the patient remained symptom free for the following 15 yr. The proband's brother (subject 3; Fig. 1A) presented with of syncope at the age of 18 yr with QT prolongation at the baseline ECG (QT_c: 500 ms). After the diagnosis of LQTS, he refused any further evaluation as well as therapy and implantation of an implantable cardioverter defibrillator. He died suddenly at the age of 26 yr. Another brother and the patient's father were asymptomatic and presented QT_c prolongation at resting ECG (subjects 1 and 4; Fig. 1A). They also refused any further clinical evaluation and therapy (last followup: July 2010).

Fig. 1.

ECG and the sequencing analysis. A: pedigree of the family of the Y1767C patient. The mutation was identified in screened individuals that presented with QT prolongation. *Carriers of the mutation. The filled symbols represent QT prolongation and syncope; the half-filled symbols represent QT prolongation with no syncope. B: DNA sequence showing part of the forward strand of SCN5A exon 28 in the proband.

Genotyping of the LQTS patient.

The genomic DNA of the patient was screened for sequence changes in all 28 exons of the SCN5A gene. Sequence analysis revealed a heterozygous A-to-G base change at oligonucleotide position 5300 in exon 28 (Fig. 1B) that resulted in a tyrosine (Y) to cysteine (C) substitution at position 1767 of the Na_v1.5 channel. The same mutation was identified in one brother, in the patient's father, and in the deceased brother on DNA extracted from a postmortem sample. Y1767 is located in the S6 segment of homologous domain IV (DIVS6) and is highly conserved in the Na⁺ channels of fish, dogs, rats, mice, and humans.

The Y1767C mutation of Na_v1.5 induces a persistent Na⁺ current.

Whole cell patch-clamp recordings were used to study Na⁺ currents of WT and Y1767C channels heterologously expressed in mammalian tsA201 cells. Channels were coexpressed with the β₁-subunit, which is known to modulate the gating and trafficking of cardiac Na⁺ channels (19). Representative Na⁺ currents of heterologously expressed WT and Y1767C channels are shown in Fig. 2A. Peak Na⁺ currents at each voltage were normalized to the whole cell capacitance and plotted versus test voltages (Fig. 2B). The currents activated at voltages more depolarized than −80 mV peaked around −40 mV and had maximal current densities of 657 ± 49 pA/pF (n = 9) and 617 ± 34 pA/pF (n = 10) for WT and Y1767C channels, respectively. The time course of the current decay was biexponential with fast and slow time constants (τ_f and τ_s, respectively) at −30 mV of τ_f = 0.7 ± 0.02 ms and τ_s = 8.9 ± 0.3 ms for the WT channel (n = 9) and τ_f = 0.5 ± 0.01 ms and τ_s = 3.5 ± 0.4 ms for the Y1767C channel (n = 10). Over a range of voltages, the time constants of the Y1767C channel were significantly smaller than those of the WT channel, which is consistent with more rapid inactivation of the mutant channel (Fig. 2C).

Fig. 2.

Whole cell currents in tsA201 cells. A: representative whole cell traces of wild-type (WT; left) and Y1767C (right) channels. Na⁺ current was elicited with a holding potential of −140 mV, depolarizing from −100 to +50 mV in 10-mV increments lasting 50 ms for each step (see the protocol in the inset). B: current-voltage (I-V) relationship of WT (n = 9) and Y1767C (n = 10) channels. The current amplitude was normalized to the membrane capacitance to obtain a measure of Na⁺ current densities. C: voltage-dependent time constants of inactivation in WT [slow time constant (τ_slow), n = 9; fast fime constant (τ_fast), n = 9] and Y1767C (τ_slow, n = 10; τ_fast, n = 10) channels. Currents were fitted to the following two-exponential function: I = I_resid + A_fast×exp [−(t −k)/τ_fast] + A_slow × exp[−(t − k)/τ_slow], where I_resid is the residual current, A_fast and A_slow are fractions of recovery of the fast and slow components, t is time, and k is the slow factor for activation or inactivation. τ_fast and τ_slow were obtained from the equation with significant differences ranging from −60 to +20 mV between WT and Y1767C channels (*P < 0.05; **P < 0.01).

In addition to changes in inactivation kinetics, the Y1767C mutation appeared to increase the fraction of Na⁺ current that failed to inactivate during prolonged (>40 ms) depolarization. The residual Na⁺ current (Fig. 3, A and B) measured after 300 ms of depolarization was normalized to peak currents. The Y1767C mutation significantly increased the amplitude of the persistent current (1.2 ± 0.3%, n = 6) compared with the WT channel (0.2 ± 0.2%, n = 5; Fig. 6B). In both cases, the residual currents were completely inhibited by 10 μM TTX, indicating that the currents were mediated by Na⁺ channels. The fivefold increase in the persistent Na⁺ current of the Y1767C channel suggested that this mutation may impair rapid inactivation.

Fig. 3.

Persistent Na⁺ currents and enlarged window currents induced by Y1767C. A and B: persistent Na⁺ currents of WT (A) and Y1767C (B) channels. The dashed line represents zero current. Persistent Na⁺ currents were induced by a depolarizing step to −30 mV for 400 ms jumping from a holding potential of −140 mV (see the protocol in the inset). The currents in blue represent the blocking effects of 10 μM TTX. C: voltage dependence of steady-state activation and inactivation of WT (activation, n = 9; inactivation, n = 7) and Y1767C (activation, n = 10; inactivation, n = 6) channels. Activated curves were derived from the same data and their graphically determined reversal potentials as I-V curves (Fig. 2B) and were fitted to the following standard Boltzmann equations: G(V)/G_max = 1/{1 + exp[(V − V_1/2)/k] }, where G is conductance, G_max is maximal conductance, and V_1/2 is the midpoint for voltage activation or inactivation. The voltage dependence of inactivation was induced by applying conditioning prepulses to membrane potentials ranging from a holding potential of −140 to −10 mV for 500 ms with 10-mV increments and was then measured using a 20-ms test pulse to −30 mV for each step (see the protocol in the inset). The recorded inactivation data were fitted to the following standard Boltzmann equation: I(V)/I_max = 1/{1+ exp[(V − V_1/2)/k]} + C, where C is the value of the persistent Na⁺ current in steady-state inactivation. D: window currents. The window region represents an enlarged portion of the overlapping area between the activation and inactivation of WT and Y1767C channels.

Fig. 6.

Effects of class I Na⁺ channel blockers and ranolazine on persistent Na⁺ currents. A: effects on persistent Na⁺ current of channels treated with 150 μM quinidine (top left), 10 μM mexiletine (top right), 50 μM flecainide (bottom left), and 50 μM ranolazine (bottom right). Using the same protocol as shown in Fig. 3, 10 μM TTX was used to block the persistent currents after the unsuccessful treatment using class I Na⁺ channel blockers. B: histogram of persistent Na⁺ currents. The persistent Na⁺ current accounted for 0.2 ± 0.2% of the peak current amplitude for WT (n = 5) channels at −30 mV and 1.2 ± 0.3% for Y1767C (n = 6) channels (**P < 0.01). The percentage of persistent Na⁺ currents was reduced to 0.3 ± 0.3% after ranolazine treatment. The percentages were not statistically reduced by quinidine (1.8 ± 0.5%, n = 6), mexiletine (1.6 ± 0.4%, n = 5), or flecainide (1.5 ± 0.5%, n = 6) treatments compared with those of untreated Y1767C channels.

Y1767C has a residual Na⁺ current during steady-state inactivation.

Figure 3, C and D, shows a comparison of the gating properties of WT and Y1767C channels. The normalized conductance was calculated from peak Na⁺ currents and plotted versus the test voltage (Fig. 3C). The smooth curve was fitted to a Boltzmann function with midpoints [midpoint for voltage activation or inactivation (V_1/2)] of −54 ± 1 mV (n = 9) and −57 ± 1 mV (n = 10) for WT and Y1767C channels, respectively (Table 1). The steady-state inactivation of WT and Y1767C channels had midpoints of −101 ± 1 mV (n = 7) and −101 ± 1 mV (n = 6), respectively (Table 1). While the midpoints of steady-state inactivation were not significantly different between WT and Y1767C channels, the mutation produced a fivefold increase in the residual Na⁺ current that failed to inactivate at voltages more depolarized than −40 mV. This finding was consistent with the observed TTX-sensitive persistent Na⁺ current at depolarized voltages (Fig. 3, A and B).

Table 1.

Biophysical properties of WT and Y1767C channels with and without ranolazine treatment

	WT Channels		Y1767C Channels		WT Channels + Ranolazine		Y1767C Channels + Ranolazine
	Means ± SE	n	Means ± SE	n	Means ± SE	n	Means ± SE	n
Steady-state activation
V1/2, mV	−54.3 ± 0.9	9	−56.7 ± 0.6	10	−55.3 ± 0.9	15	−56.3 ± 0.5	16
k, mV	6.5 ± 0.3	9	6.8 ± 0.2	10	6.2 ± 0.2	15	6.2 ± 0.2	16
Steady-state inactivation
V1/2, mV	−100.9 ± 0.6	7	−100.6 ± 0.5	6	−103.8 ± 0.8	15	−103.1 ± 0.7	16
k, mV	5.5 ± 0.2	7	6.6 ± 0.2†	6	5.5 ± 0.1	15	6.5 ± 0.1‡	16
C, %	0.2 ± 0.1	7	1.0 ± 0.2†	6	0.2 ± 0.1	8	0.2 ± 0.2	8
Recovery from fast inactivation
τfast, ms	4.3 ± 0.5	7	2.6 ± 0.1†	7	4.4 ± 0.6	12	2.2 ± 0.3§	12
τslow, ms	57.3 ± 7.2	7	38.8 ± 4.4*	7	53.3 ± 6.6	12	35.2 ± 5.0§	12
Afast , %	95.6 ± 5.3	7	94.4 ± 3.0	7	93.7 ± 3.9	12	95.7 ± 2.1	12
Aslow, %	4.4 ± 0.6	7	5.2 ± 1.3	7	6.2 ± 0.8	12	4.8 ± 0.8	12
Recovery from frequency inhibition
τfast, ms	3.8 ± 0.5	7	4.6 ± 0.5	7	15.6 ± 3.4	8	28.8 ± 3.5§	8
τslow, ms	35.1 ± 3.9	7	40.8 ± 0.6	7	2,446.1 ± 415.1	8	3,986.5 ± 592.2§	8
Afast , %	7.8 ± 0.9	7	8.1 ± 0.1	7	10.2 ± 0.2	8	18.4 ± 0.3‡	8
Aslow, %	7.9 ± 0.8	7	8.9 ± 0.1	7	32.8 ± 2.0	8	52.5 ± 3.0§	8

WT, wild type; V_1/2, midpoint for activation or inactivation; k, slow factor for activation or inactivation; C, value of the persistent Na⁺ current in steady-state inactivation; τ_fast and τ_slow, fast and slow time constants; A_fast and A_slow, fractions of recovery of the fast and slow components.

^*P < 0.05 and

^†P < 0.01, WT vs. Y1767C channels;

^‡P < 0.05 and

^§P < 0.01, WT channels + ranolazine vs. Y1767C channels + ranolazine.

Y1767C increases the Na_v1.5 window current.

The overlap of the activation and steady-state inactivation of Na⁺ channels identifies a hyperpolarized range of voltages (i.e., window) where the channels have a small probability of being partially but not fully inactivated (4). Na⁺ channel opening at these voltages is significant because the resulting inward window currents tend to depolarize the resting membrane potential and may alter the excitability of cardiomyocytes. Figure 4A shows a plot of the probability of Na⁺ channel opening calculated from the product of the fitted activation and steady-state inactivation parameters (Fig. 3C). The resulting probability is a biphasic function of voltage with a relatively small component with a peak near −90 mV and a larger component at more depolarized voltages (greater than −70 mV). These increases in probability result from a combination of the activation, steady-state inactivation, and persistent (i.e., non-inactivating) components of Na⁺ channel gating. Figure 4, B and C, shows a separation of these biphasic functions into their individual window and persistent components. The window component is primarily determined by the overlap of Na⁺ channel activation and inactivation. Because of the low probability of Na⁺ channel opening within the window, the noninactivating component of the WT channel did not substantially alter the window amplitude, which was relatively small and reached a peak probability of 0.05% of the maximum at −90 mV (Fig. 4B). The Y1767C mutation enlarged the window, shifted the peak toward more depolarized voltages (−81 mV), and produced a threefold increase in amplitude (0.13%). The increased window predicted that a larger fraction of Y1767C channels would be activated near the resting membrane potential of cardiomyocytes (approximately −90 mV) and at voltages where Na⁺ channels activate (greater than −80 mV).

Fig. 4.

Effects of Y1767C on the Na_v1.5 window current. The overlap of the activation and inactivation of Na⁺ channels defines a range of voltages (i.e., window) where the channels are partially activated but not fully inactivated. The probability of being within this window was calculated from the product of the activation and steady-state inactivation parameters (Fig. 3C) through the following equation: (1/{1 + exp[(V_1/2act − V)/k_act]} × ((1 − C)/{1 + exp[(V − V_1/2inact)/k_inact]} + C). A: this probability is a biphasic function of voltage with a peak at −90 mV and at voltages more depolarized than −40 mV. This pattern resulted from a combination of the window and persistent components of Na⁺ channel gating. The individual contributions of the window and persistent components were evaluated and are plotted in B and C. B: the Y1767C mutation widened the window and increased the peak by threefold. C: Y1767C shifted the onset of the persistent component toward more hyperpolarized voltages and increased its amplitude by eightfold.

The second increase in probability occurred over a more depolarized range of voltages (greater than −70 mV) and was determined by Na⁺ channel activation and the persistent (noninactivating) component of Na⁺ channel gating (Fig. 4C). In the WT channel, the persistent component increased over the range of voltages where Na_v1.5 channels activate when they reach a peak of 0.13% of the maximum probability at voltages more depolarized than −40 mV. The Y1767C mutation increased the peak probability eightfold and caused a hyperpolarizing shift in the range of voltages where the persistent component activates. This apparent shift was not caused by a change in Na⁺ channel activation but rather reflected an increased overlap of the persistent and window components of the mutant channel. This overlap, coupled with the substantial increase in the amplitude of the persistent component, may act synergistically to produce subthreshold activating Na⁺ currents in the Y1767C channel.

Y1767C promotes a rapid recovery from inactivation.

At least two mechanisms could explain the persistent Na⁺ current produced by the Y1767C mutation. The mutation could simply decrease the entry of Na⁺ channels into the fast inactivated state. However, this mechanism is inconsistent with measurements of Na⁺ current decay, indicating that inactivation was more rapid in the mutant channel (Fig. 2C). Alternatively, the Y1767C mutation could induce persistent Na⁺ current by promoting a rapid recovery from inactivation. This might be expected if the Y1767C mutation destabilizes fast inactivation similar to what has been observed for other LQT3 mutations (6, 44, 45). The experiments shown in Fig. 5A used a double-pulse protocol to directly measure the recovery from fast inactivation. The smooth curves are fits to a biexponential equation with τ_f and τ_s of 4.3 ± 0.5 and 57.3 ± 7.2 ms for the WT channel (n = 7) and 2.6 ± 0.1 and 38.8 ± 4.4 ms for the Y1767C channel (n = 7). The Y1767C mutation significantly reduced both τ_f and τ_s, which is consistent with a more rapid recovery from inactivation (Table 1). Overall, the data suggested that the Y1767C mutation accelerates the entry into and recovery from the fast inactivated state. The rapid recovery appeared to predominate at depolarized voltages and accounted for the majority of the persistent Na⁺ current produced by this mutation.

Fig. 5.

Recovery from fast inactivation and ramp currents. A: recovery from fast inactivation of WT (n = 7) and Y1767C (n = 7) channels. Currents were induced by a two-pulse protocol with a variable recovery interval ranging from 0.1 to 200 ms. The conditioning prepulse lasting 40 ms was depolarized to −30 mV. Currents were measured using a test pulse at −30 mV lasting 20 ms (see the protocol in the inset). The time constants (values shown in Table 1) were obtained using a two-exponential function: Na⁺ current = Y₀ + A_fast × [1 − exp (−t/τ_f)] + A_slow × [1 − exp(−t/τ_s)], where Y₀ is the y-axis intercept. B: representative traces of ramp currents in WT (black) and Y1767C (red) channels. Traces were elicited by 0.32 mV/ms of ramp depolarizations from −140 to +20 mV for 500 ms (see the protocol in the inset). The ramp currents increased, with a slight shift of their peak voltages, to the depolarized direction for the Y1767C mutation.

Y1767C increases the subthreshold activating Na⁺ current.

The rapid recovery from inactivation (Fig. 5A) and enlarged window current (Fig. 3D) of the Y1767C channel may enable a small but significant fraction of Na⁺ channels (≤1%) to open at unusually hyperpolarized voltages (less than −60 mV). These subthreshold activating Na⁺ currents were further investigated using a voltage ramp (−140 to 20 mV) to activate Na⁺ currents (Fig. 5B). The ramp current of the WT channel began to activate at a relatively hyperpolarized voltage (−100 mV) and reached a maximum of 0.66 ± 0.15% (n = 7) of the peak Na⁺ current at −80 mV. The ramp current of the Y1767C channel displayed similar activation but peaked at a more depolarized voltage (−70 mV) and was 1.8-fold larger (1.21 ± 0.13%, n = 8) than the WT channel. Varying the duration over which the voltage ramp was applied (0.32–3.2 mV/ms) predictably altered the amplitudes of the ramp currents but did not change the relative differences between WT and Y1767C currents, indicating that the effects of the mutation did not depend on the rate of the voltage change (data not shown). These subthreshold activating Na⁺ currents are predicted to depolarize the resting membrane potential and reduce the threshold for initiating action potentials in myocytes carrying the Y1767C mutation.

Ranolazine is a selective inhibitor of the Y1767C mutant persistent Na⁺ current.

The rapid recovery from inactivation and resulting persistent Na⁺ current caused by the Y1767C mutation appeared to underlie the LQT3 phenotype in this patient. Previous studies (5, 41) have shown that class I antiarrhythmic drugs are effective inhibitors of persistent Na⁺ currents linked to LQT3 mutations and have proven to be beneficial in the treatment of patients carrying these mutations. The Y1767C mutation described here is unique because, unlike most LQT3 mutations, it is situated in the DIVS6 segment, a region of Na⁺ channels that is known to contribute to local anesthetic binding (39). A previous study (36) has shown that the Y1767C mutation weakens drug binding to Na_v1.5 channels. This raises the possibility that, in addition to inducing persistent Na⁺ currents, the Y1767C mutation may also alter the pharmacology of the mutant channel. This was further investigated by examining the effects of several class I antiarrhythmic drugs (quinidine, mexiletine, flecainide, and lidocaine) on the Y1767C-induced persistent Na⁺ current (Fig. 6A). At clinically relevant doses (10–150 μM), none of these drugs significantly reduced the persistent Na⁺ current of the Y1767C channel. This contrasts with other LQT3 mutations, where these drugs are potent inhibitors of persistent Na⁺ currents (5, 41). Because of its location in the pore and the role of this tyrosine in drug binding, we speculate that the Y1767C mutation weakens Na⁺ channel inhibition by disrupting the binding of these antiarrhythmics.

Recent studies (5, 14, 41) have shown that the antianginal agent ranolazine is an effective inhibitor of persistent Na⁺ currents resulting from LQT3 mutations. Unlike class I antiarrhythmics, ranolazine (50 μM) significantly reduced the persistent Na⁺ current of the Y1767C channel by 74 ± 0.25% (Fig. 6, A and B). The underlying mechanism was further investigated by applying a series of depolarizing pulses at frequencies between 1 and 50 Hz. In the absence of the drug, the pulsing protocols produced only a small reduction in Na⁺ current amplitude (Fig. 7B). Ranolazine (50 μM) significantly reduced the relative peak current amplitude (P₅₀/P₁) of the WT channel by 27 ± 3% (n = 8). Surprisingly, the Y1767C mutation potentiated the ranolazine inhibition, reducing currents by 43 ± 2% (n = 8). The results shown in Fig. 7A demonstrate that ranolazine was a more effective inhibitor of the Y1767C channel at pulsing frequencies of >10 Hz; higher frequencies of stimulations were also used for the purpose of the biophysical characterization. At lower frequencies (1–5 Hz), ranolazine inhibition of the Y1767C channel was not statistically different from that of the WT channel.

Fig. 7.

Frequency-dependent inhibition and recovery from frequency inhibition. A: frequency-dependent inhibition at 20 Hz of WT channels (n = 7), Y1767C channels (n = 7), WT channels with ranolazine (n = 8), and Y1767C channels with ranolazine (n = 8). A 50-pulse train was applied at −30 mV for 10 ms, jumping from a holding potential of −140 mV at 20 Hz (see the protocol in the inset). Each normalized peak current was plotted against the pulse number. B: relative amplitudes normalized to the 50th sweep (P₅₀/P₁) with and without ranolazine. Before ranolazine treatment, there were no differences in normalized currents between WT (n = 7) and Y1767C (n = 7) channels. After ranolazine treatment, normalized currents of both WT (n = 8) and Y1767C (n = 8) channels were inhibited. There was a further inhibition at high-frequency stimuli from 10 to 50 Hz for Y1767C channels (*P < 0.05 and **P < 0.01 vs. WT; #P < 0.05 and ##P < 0.01 vs. WT channels with ranolazine). C: recovery from frequency-dependent inhibition of WT channels (n = 7), Y1767C channels (n = 7), WT channels with ranolazine (n = 8), and Y1767C channels with ranolazine (n = 8). Currents were induced by applying a 50-pulse train of conditioning prepulses to −30 mV for 10 ms at 50 Hz from a holding potential of −140 mV. Test pulses were then applied to −30 mV for 20 ms to measure peak current amplitudes after a variable duration of 0.1–9,000 ms. The resulting curves were fitted to the following two-exponential function: Na⁺ current = Y₀ + A_fast × [1 − exp(−t/τ_f)] + A_slow × [1 − exp(−t/τ_s)].

Many class I antiarrhythmic drugs act by preferentially binding to inactivated states of Na⁺ channels, resulting in a hyperpolarizing shift in steady-state availability (8). To test this potential mechanism, the effects of ranolazine on the activation and steady-state inactivation of the channels was examined (Supplemental Material, Supplemental Fig. 1).¹ Ranolazine reduced the peak current densities of both WT (21.2%, n = 15) and Y1767C (19.5%, n = 16) channels but failed to produce any changes in activation or steady-state inactivation (Table 1). These data suggest that while ranolazine inhibited both the WT and mutant channels, it did not alter their gating properties. This is inconsistent with a state-dependent mechanism in which ranolazine preferentially binds to inactivated states. Rather, the strong frequency dependence of the inhibition suggests the ranolazine may principally act by blocking open channels (47). Ranolazine markedly reduced both the channel opening probability (Supplemental Fig. 3A) and window (Supplemental Fig. 3B) Na⁺ current of Y1767C channels but only caused a slight reduction for WT channels (Supplemental Fig. 2, A and B). We also tested the effect of ranolazine on another LQT3 mutation (V1763M), which was located at the same region as Y1767C and was characterized by increased persistent Na⁺ currents that are insensitive to lidocaine block (10). Ranolazine effectively blocked lidocaine-insensitive persistent Na⁺ currents (Supplemental Fig. 4) but had similar effects on frequency inhibition and recovery from frequency inhibition as WT channels (Supplemental Fig. 5).

Ranolazine slows the recovery from frequency inhibition of the Y1767C channel.

A prominent feature of many class I antiarrhythmics is that they slow the recovery of drug-modified Na⁺ channels. A repetitive pulsing protocol was used to induce ranolazine inhibition and assess the recovery time course of WT and Y1767C channels (Fig. 7C). In the absence of the drug, the recovery from frequency inhibition of WT and Y1767C channels was similar, indicating that the mutation did not directly alter the recovery from inactivation after repetitive stimulations. Ranolazine significantly increased the fraction of inhibited channels and slowed the recovery time course. The smooth curves are biexponential curve fits with τ_f and τ_s of 16 and 2,446 ms for the WT channel and 29 and 3,987 ms for the Y1767C channel (Table 1). The Y1767C channel displayed more pronounced ranolazine frequency-dependent inhibition, and its recovery time course was significantly slower than the WT channel. These data are consistent with the results of studies of persistent Na⁺ currents showing that ranolazine is a more potent inhibitor of the Y1767C channel (Fig. 7, A and B).

Y1767 is located in the pore region as a bulky aromatic residue.

Figure 8A shows the orientation of the residues predicted by a homology model of the S6 segments based on the M2 helix of the MthK channel (25). Figure 8B shows a top view of the model indicating that F1760, V1763, and Y1767 are prominently exposed within the central cavity of the channel.

Fig. 8.

Homology model of the D4S6 segment. A: side view of a homology model of the Na⁺ channel S6 segment based on the M2 segment of the MthK channel. D1S6 was removed for clarity. B: top view of the model showing that F1760, V1763, and Y1767 are exposed within the central cavity.

DISCUSSION

This study investigated a novel mutation (Y1767C) identified in a 17-yr-old male displaying a classic LQT3 phenotype in ECG recordings and documented episodes of cardiac arrest. Genomic sequencing identified an A-to-G base change at position 5300 of the SCN5A gene resulting in a tyrosine to cysteine (Y→C) substitution. The resulting mutation (Y1767C) was located in the S6 membrane-spanning region of homologous domain IV (DIVS6) of the cardiac Na⁺ channel (Na_v1.5). This region of the channel is known to play important roles in fast inactivation and the binding of local anesthetics and antiarrhythmic drugs (36, 40). The principal goal of the present study was to investigate the effects of the Y1767C mutation on Na⁺ channel gating and to correlate the observed changes with the ECG phenotype and clinical history of the patient.

Electrophysiological analyses of the heterologously expressed Y1767C mutant revealed an enhanced persistent Na⁺ current at depolarized voltages associated with accelerated inactivation (Fig. 2C) and a rapid recovery from inactivation (Fig. 5A). The rapid recovery caused an increase in the fraction of mutant channels remaining persistently open at depolarized voltages. The lack of an absorbing inactivated state in the Y1767C channel appeared to be the primary cause of the persistent Na⁺ current. In addition to the rapid recovery, the Y1767C mutation significantly increased the overlap of the activation and steady-state inactivation of the channel (Fig. 3D). The area between these curves defines a range of voltages (i.e., window) where Na⁺ channels may be activated but not inactivated (4). The larger window produced by the Y1767C mutation predicts an increase in the fraction of channels that open at hyperpolarized voltages resulting in an increase in inward Na⁺ current. This is consistent with the observed 1.8-fold increase in Y1767C ramp current (Fig. 5B). This current would serve to both amplify small depolarizations near the resting membrane potential and reduce the threshold for initiating cardiac action potentials. Myocardial hyperexcitability stemming from increased subthreshold activating Na⁺ currents coupled with prolongation of the plateau phase of the cardiac action potential form the basis of the LQT3 phenotype in this patient.

The Y1767C mutation is situated in the DIVS6 segment, a region of Na⁺ channels that has been implicated in both voltage-dependent gating and drug binding (36). Mutational analyses have suggested that residues situated near the COOH-terminus of DIVS6 disrupt inactivation by interfering with the binding of the native inactivation gate (31, 32). Y1767 is located adjacent to these sites and may contribute to inactivation gating by a similar mechanism. Impaired inactivation provides a plausible explanation for both the rapid inactivation (Fig. 2C) and recovery from inactivation (Fig. 5A) produced by the Y1767C mutation. Y1767 is bracketed by two LQT3 mutations that either produce persistent Na⁺ currents (M1766L) or accelerate recovery from inactivation (I1768V) (20, 43). These findings further support the idea that Y1767 is situated within a region of the DIVS6 segment that is important for Na⁺ channel inactivation and the generation of persistent Na⁺ current. In addition, DIVS6 is an important locus for the binding of class I antiarrhythmics. Two highly conserved aromatic amino acids (F1760 and Y1767) of DIVS6 are known to be critical determinants of drug binding (40). The Y1767C mutation examined in this study is known to weaken the anesthetic inhibition of Na_v1.5 channels, consistent with an important role for this residue in drug binding (36, 37). The data show that class I antiarrhythmic drugs (quinidine, mexiletine, lidocaine, and flecainide) are relatively weak inhibitors of the Y1767C-induced persistent Na⁺ current (Fig. 6A). This sharply contrasts with other LQT3 Na⁺ channel mutations, where class I antiarrhythmic drugs are generally found to be potent inhibitors of persistent Na⁺ currents (29, 35, 46). Unlike the Y1767C mutation described here, the majority of these LQT3 mutations are located outside the DIVS6 segment and are thus removed from sites that directly contribute to drug binding (12). Our data support the conclusion that the Y1767C mutation weakens class I antiarrhythmic binding resulting in persistent Na⁺ currents that are insensitive to these drugs.

Ranolazine is an antianginal agent that shares a similar structure with class I antiarrhythmics (14) and is an effective inhibitor of persistent Na⁺ currents of both LQT3 mutant and native cardiac Na⁺ channels (38, 42). Ranolazine reduced peak Na⁺ currents but did not alter the midpoints of activation or steady-state inactivation of WT or Y1767C mutant channels (Supplemental Fig. 1). These data indicate that at low stimulation frequencies (≤1 Hz), ranolazine is a relatively weak inhibitor of Na_v1.5 channels. However, increasing the stimulation frequency produced a potent use-dependent inhibition of the WT channel (Fig. 7B). Paradoxically, the Y1767C mutation potentiated the ranolazine-induced use-dependent inhibition (Fig. 7, A and B). These data indicate that, in contrast to class I antiarrhythmics, the Y1767C mutation enhances ranolazine binding.

The differential effect of the Y1767C mutation on ranolazine and class I antiarrhythmics reflects substantial differences in the underlying mechanisms of drug binding and Na⁺ channel inhibition. The strong dependence of ranolazine inhibition on rapid repetitive pulsing and the relative weak inhibition of inactivated channels suggest that this drug primarily inhibits Na_v1.5 channels by an open channel blocking mechanism. This contrasts with conventional class I antiarrhythmics, such as lidocaine, that inhibit Na⁺ channels by preferentially binding with high affinity to inactivated states of the channel (22, 23).

Recent work has demonstrated that mutations near the middle of the DIVS6 segment of the Na_v1.5 channel (e.g., F1760A) that weaken antiarrhythmic drug binding also attenuate ranolazine inhibition, suggesting that these drugs share overlapping binding sites (14). Y1767 is situated near the cytoplasmic end of the DIVS6 segment and is an important determinant of antiarrhythmic drug binding (36, 37). The aromatic side chain at this position appears to be exposed within the aqueous pore, where it is believed to directly interact with bound drugs (1). We speculate that the presence of an aromatic residue near the cytoplasmic entrance of the pore (Fig. 8) may impede the entry of bulky drugs like ranolazine (47). Replacing the tyrosine with the smaller cysteine appears to reduce this steric hindrance and facilitate ranolazine binding. This differs from class I antiarrhythmics, where drug binding is highly dependent on the presence of an aromatic amino acid at that position (28). The tyrosine at position 1767 appears to play a dual role in drug binding. Y1767 directly contributes to the local anesthetic binding site of Na_v1.5 channels and is critical for the high-affinity binding of many antiarrhythmic drugs. In addition, Y1767 appears to act as a barrier that prevents large drugs from gaining access to the cytoplasmic aqueous pore. Another LQT3 mutation located just upstream from Y1767 (V1763M) induces lidocaine-insensitive persistent Na⁺ currents, suggesting that this mutation may disrupt antiarrhythmic drug binding in a similar way (10). The V1763 residue is not exposed to the cytoplasmic pore (Fig. 8). Consequently, ranolazine has a lower accessibility to the V1763M mutation because of the barrier effect of Y1767. Indeed, ranolazine has different effects on the frequency and recovery from frequency inhibition of Y1767C (Fig. 7) and V1763M channels (Supplemental Fig. 5).

We identified a novel LTQ3 mutation in a patient suffering from arrhythmias and cardiac arrest. This mutation involved a single base change resulting in a tyrosine to cysteine substitution at position 1767 of the cardiac Na⁺ channel (Y1767C). Y1767 is located in the DIVS6 segment of the Na⁺ channel, a region that is known to contribute to both fast inactivation and antiarrhythmic drug binding. The Y1767C mutation produced changes in Na⁺ channel gating similar to other LQT3 mutations, including persistent Na⁺ current at depolarized voltages, increased window current, accelerated inactivation, and a rapid recovery from inactivation. The changes predicted an increase in the excitability of Na_v1.5 channels and enhanced persistent Na⁺ current during the plateau phase of the cardiac action potential. These effects appeared to underlie the LQT3 phenotype observed in this patient. Ranolazine is a highly selective inhibitor of Y1767C-induced persistent Na⁺ currents. Ranolazine restored the window and persistent Na⁺ currents to control levels, suggesting that this drug may be an effective therapeutic treatment for LQT3 patients harboring this mutation. Finally, these data indicate that Y1767 plays unique roles in both fast inactivation and drug binding to cardiac Na⁺ channels.

GRANTS

This work was supported by a grant from the Heart and Stroke Foundation of Québec and by Canadian Institutes of Health Research Grant MT-13181.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).