INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

LIFE-THREATENING VENTRICULAR fibrillation is most often associated with structural heart disease or ischemia, but sudden cardiac death also can be triggered by idiopathic ventricular fibrillation (IVF) that has its origin in ion channel defects. One form of IVF (Brugada syndrome), characterized by right bundle-branch block with S-T segment elevation in the electrocardiogram (ECG; see Refs. 3 and 13), is linked to mutations of SCN5A (5), the gene that encodes the cardiac Na⁺ channel -subunit. The abnormal ECG pattern may be intermittent but can be provoked by class Ia Na⁺ channel blockers (4, 14). These observations and results from a canine model (21) suggest that decreased Na⁺ conductance may be the underlying cause. Consistent with this notion, we identified IVF mutations in SCN5A (5) that would result in truncated, nonfunctional Na⁺ channel -subunits (an insertion that disrupts an exon splice site and a deletion that introduces a premature stop codon) and missense mutations in the coding region (T1620M and a benign polymorphism R1232W). The T1620M mutant channels, when expressed heterologously in mammalian cells and recorded at near-physiological temperatures, have been reported to show accelerated inactivation compared with the wild type (WT; see Ref. 7) and therefore produce less Na⁺ current. Coexpression of mutant -subunits with auxiliary ₁-subunits has been reported to alter the phenotype (1, 12).

Recently, several additional SCN5A mutations have been linked to IVF: four missense mutations that result in amino acid substitutions (R1432G, R1512W, A1924T, and L567Q; see Refs. 6, 16, 17) and an insertion mutation that adds an aspartic acid residue (1795inD; see Ref. 2). R1512W, A1924T, and 1795inD (2, 17, 19) cause shifts in the voltage dependence of gating, whereas R1432G (6) produced no detectable current when expressed in Xenopus oocytes. The L567Q mutation is particularly interesting because it is associated with IVF and sudden infant death syndrome (16), but the L567Q phenotype has not been reported. Here we have analyzed the functional characteristics of this mutant expressed in human embryonic kidney cells to test the hypothesis that Brugada mutations generally suppress Na⁺ channel reduced function. The channels were coexpressed with human ₁-subunits to determine whether - and -subunit interactions alter the phenotype (12), and experiments were conducted at both 22 and 32°C to test the possibility of temperature-sensitive alterations in gating (7).

Coexpression of L567Q mutant -subunits with ₁-subunits resulted in robust Na⁺ currents that differed from WT primarily by a marked acceleration of the onset of inactivation with no change in the rate of recovery from inactivation. Mutant channels differed secondarily by a negative shift in the voltage dependence of steady-state inactivation. The acceleration of inactivation was unchanged by the absence of ₁-subunits. We found no differences in the temperature sensitivity of the inactivation time constant between mutant and WT channels.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Na⁺ channel clone and mutagenesis. WT human heart SCN5A cDNA (9) was cloned into the pcDNA3 plasmid vector (Invitrogen, Carlsbad, CA) for expression in human embryonic kidney cells (HEK-293). The L567Q mutation was created by site-directed mutagenesis using the mega-primer PCR-based method (18) with verification by DNA sequencing. Human Na⁺ channel ₁-subunit subcloned into the pRc/CMV vector (Invitrogen) for mammalian expression was the generous gift of Dr. A. L. George (Vanderbilt University, Nashville, TN).

Mammalian cell transfection and expression. A cell line that stably expressed Na⁺ channel ₁-subunits was established by the following method. HEK-293 cells were maintained in MEM containing 10% BSA, 1% penicillin, and streptomycin (MEM/FBS/P-S). Digested human Na⁺ channel ₁-subunit cDNA (5 µg) and 25 µg lipofectamine (Life Technologies, Rockville, MD) were mixed with 0.2 ml MEM and incubated at room temperature for 15 min. MEM-washed cells were incubated with the transfection mixture at 37°C for 4 h. After replacement of the transfection mixture with 10 ml MEM/FBS/P-S, cells were cultured for another 2 days. Positive colonies were then selected by G418 resistance and were identified by the presence of ₁-subunit mRNA detected by RT-PCR.

Electrophysiological recording. Macroscopic Na⁺ currents were recorded using the patch-clamp technique in the whole cell mode. Patch pipettes were pulled from borosilicate capillary glass, lightly fire-polished to resistance 0.9-1.4 M when filled with pipette solution, and connected to the head stage of a patch-clamp amplifier (Axopatch 200; Axon Instruments, Foster City, CA). Cells were transferred on glass coverslips to the recording chamber on the stage of an inverted microscope and superfused continuously with Tyrode solution at a rate of 1-2 ml/min. Unless otherwise noted, recordings were made at 22°C. However, in some experiments temperature was raised to 32°C using an electronically controlled heated chamber infused with preheated bathing solution (TC2bip; Cell MicroControls, Virginia Beach, VA). Bath temperature, measured by a thermistor placed within 5 mm of the recording pipette, served as the input to feedback control of the resistive heater that formed the bottom of the recording chamber. Under these conditions, temperature variations of <0.2°C were measured by a roving thermocouple in the bath.

Solutions. External Tyrode bathing solution consisted of (in mM) 137 NaCl, 5.4 KCl, 1 MgCl₂, 2 CaCl₂, 10 glucose, and 10 HEPES, pH 7.3. Internal pipette solution contained (in mM) 120 CsF, 10 CsCl, 10 EGTA, and 10 HEPES, pH 7.3.

Data acquisition and analysis. Generation of voltage commands, data acquisition, data analysis, and curve fitting was accomplished with pClamp6 software (Axon Instruments). The choice of biexponential over monoexponential functions to fit kinetic data was based on an F-ratio test (P < 0.05) that takes into account the increased number of free parameters. Where appropriate, data are expressed as means ± SE. A two-tailed Students t-test was used to evaluate the significance of the difference between means (P < 0.05 or 0.01).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

After transient transfection with -subunits, we recorded large Na⁺ currents from both HEK cells and the stable ₁-cell line. No detectable currents were observed in the absence of -subunit transfection. The mechanism of IVF has been attributed to a reduction in Na⁺ conductance (20), but we observed no obvious difference in levels of expression between WT and mutant channels. Typical maximum current densities of 492 ± 99 (n = 16) and 323 ± 65 (n = 12) pA/pF, respectively, were obtained in WT and L567Q channels (in the presence of ₁-subunit) and 205 ± 26 (n = 14) and 174 ± 29 (n = 13) pA/pF, respectively, in WT and L567Q channels (in the absence of ₁) from parallel experiments under identical conditions. Thus, although expression levels in both channel types were elevated by coexpression with ₁, the L567Q mutation in the -subunit did not have a gross effect on peak currents. Therefore, we tested for alterations in gating kinetics.

Current-voltage families recorded in the ₁-cell line expressing WT Na⁺ channels (Fig. 1A) showed the characteristic pattern of transient inward currents that inactivated with a voltage-dependent time course from ~12 to 2 ms as test potential varied from 40 to +20 mV. At more positive test potentials, the decay time showed little voltage dependence. By contrast, Na⁺ currents in mutant L567Q channels (Fig. 1B) decayed to baseline within the first 4 ms of step depolarizations in the range 40 to +20 mV. This apparent acceleration of the time course of inactivation was observed throughout the voltage range of activation. Thus, at a test potential of 40 mV (near the activation midpoint, Fig. 1C), WT currents decayed to 10% of the peak amplitude within 10 ms, whereas L567Q currents subsided to the same level in about one-half the time. A similar acceleration also was observed at test potentials in the activation plateau voltage range. Thus, at a test potential of +20 mV, WT currents reached the 10% level within 2 ms, whereas L567Q decayed to the same level in ~1.3 ms.

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Typical current-voltage records in HEK cells stably expressing 1-subunits after transient transfection with either wild-type (WT; A) or L567Q mutant (B) -subunits. A and B: superimposed currents (Im, membrane current) evoked by test pulse potentials 60 to +20 mV (20-mV increments) from a holding potential of 120 mV, temperature 22°C. Horizontal line shows 0 current level. C and D: time course of decay at test potentials 40 mV (C) and +20 mV (D) from the same experiment as in A and B. Thick and thin traces, respectively, are from either WT or L567Q mutant channels. Current traces were normalized to the maximum peak for kinetic comparison and were plotted semilogarithmically.

Semilogarithmic plots of current decay for both groups of channels (Fig. 1, C and D) show curvature that could be accurately fit to a biexponential decay function as previously reported (20). The time constants from biexponential fits to the decay phase of currents evoked by test pulses 40 to +40 mV provide a quantitative comparison of the onset of inactivation (Table 1). For both types of channels, inactivation time constants became progressively shorter with increased test potential (Fig. 2B). However, throughout this range, L567Q time constants were nearly twofold shorter than WT. Thus the difference in the major (fast) time constant provided an accurate representation of changes in the overall time course of decay (Fig. 1, C and D). At more negative test potentials (80 to 50 mV), inactivation time course was measured using a standard double-pulse protocol. Figure 2A shows the average time course of inactivation at a test potential of 60 mV. The onset of inactivation was clearly faster in normal vs. mutant channels coexpressed with ₁-subunits. When fitted to biexponential decay functions, the fast time constant was significantly shorter in mutant compared with WT channels (Table 1), which is indicative of a three- to fourfold acceleration. Thus an obvious characteristic of the mutant phenotype was an accelerated rate of inactivation throughout the voltage range.

View larger version (28K): [in this window] [in a new window]   Fig. 2.   Inactivation time course in WT () and L567Q () mutant channels coexpressed with 1-subunits. A: time course of inactivation at a test potential of 60 mV was obtained using a conditioning potential of varying duration, and the envelope of a test pulse current amplitude (at a test potential of 10 mV, holding potential 120 mV) was fit to a biexponential decay function. Data were pooled from 8 WT and 5 L567Q cells. A similar analysis was performed at test potentials 80 to 50 mV (except that at 80 mV only a single decay component was detected). B: fast time constants comprising the major fraction of inactivation time course were plotted semilogarithmically vs. test potential. In the range 40 to +40 mV, the time course was analyzed by fitting the decay phase of current records to biexponential decay functions. The fast time constant at each potential was averaged from 5-8 cells in each group. C: average time course of recovery from inactivation was measured using pulse protocol consisting of a fixed conditioning pulse (500-ms duration, 10-mV amplitude), variable-duration recovery potential (120 mV), and test pulse (10 mV, 5 ms) to assess the amount of recovery. The envelope of test pulse current peak amplitudes was plotted vs. recovery interval and was fit to a biexponential decay function (n = 11 WT and 9 mutant). D: average time constants of the fast (major) component of recovery vs. recovery potential. Em, membrane voltage.

Similar results were obtained in the absence of the ₁-subunit. For instance, mean fast time constants at a test potential of 20 mV were approximately twofold shorter in the mutant channels [time constants 1.34 ± 0.27 and 0.63 ± 0.04 ms, respectively, were obtained in WT and L567Q (n = 13 cells/group)]. Therefore, the phenotype does not depend on the presence or absence of ₁.

T1620M, a previously reported Brugada mutation of SCN5A, has been shown to accelerate the time course of recovery from inactivation (5, 12). Therefore, we analyzed the recovery time course in the L567Q mutant using a standard pulse protocol consisting of a 500-ms conditioning pulse (10 mV amplitude) to completely inactivate the channels, followed by a variable-duration interval (150 to 110 mV amplitude) to allow recovery and a fixed 5-ms test pulse (10 mV amplitude) to assess the amount of recovery. Time constants were obtained from fits of the envelope of test pulse current peaks to biexponential decay functions (Table 2). Typical results obtained at 120 mV (Fig. 2C) show nearly complete superposition between the time course of recovery in normal and mutant channels. Analysis of time constants at recovery potentials 150 to 110 mV (Fig. 2D and Table 2) indicated that, unlike the T1620M mutation, L567Q had no statistically significant effect on either time constants or amplitudes obtained from exponential curve fitting.

We also tested whether the L567Q inactivation phenotype might be influenced by differences in temperature sensitivity (7). Decay time constants at a test potential of 10 mV recorded at temperatures of 22 and 32°C were compared (Fig. 3). Increased temperature shortened the decay times and increased peak amplitude with no significant difference in temperature sensitivity between the two groups of channels. At a test potential of 10 mV, temperature coefficient (Q₁₀) values of 2.60 ± 0.31 (n = 8) and 2.61 ± 0.17 (n = 10), respectively, were obtained for the fast time constant in WT and L567Q channels expressed in the ₁-cell line.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Effect of temperature on Na+ currents in normal and mutant channels. Typical current waveforms evoked by 10-mV, 20-ms test pulses from a holding potential 120 mV in cells expressing either WT + 1 (A) or L567Q + 1 (B) subunits. Records were obtained in each cell at ambient temperatures of 22 and 32°C. Current signals were filtered at 10 kHz and digitized at 100 kHz. On-line leakage and capacitance subtraction was omitted (capacity transients were blanked for clarity). Only the first 5 ms of the records are shown.

In principle, reduced macroscopic conductance at physiological potentials also could result from a shift in the voltage range of activation gating. We quantified the steady-state voltage dependence of activation (Fig. 4A) by fitting conductance-voltage (G-V) curves derived from peak current amplitudes to Boltzmann distributions. For WT channels in the presence of ₁-subunits, the voltage range for a peak G-V relationship centered on a midpoint potential of 41.4 mV (slope factor 7.6 mV). By contrast, the G-V curve in L567Q channels was shifted to more depolarized potentials by ~5 mV (midpoint 35.3 mV, slope factor 10.1 mV). Activation curves obtained in a total of three sets of transfections yielded an average +6.2-mV shift in midpoint potential and a +1.5-mV increase slope factor. Although statistically significant, these small shifts may be secondary to the mutation-induced acceleration of inactivation rather than a true indication of altered voltage dependence of activation.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Voltage dependence of steady-state activation (A) and inactivation (B). A: peak current amplitude obtained from current-voltage families at test potentials 70 to 20 mV (holding potential 120 mV). Currents were recorded after a delay of at least 5 min from establishment of the whole cell mode to allow complete exchange of the cytosol and pipette solution and stabilization after initial time-dependent shifts. Current amplitudes were converted to conductances assuming an Ohmic open channel relationship. Conductances were normalized to the maximum obtained at strongly depolarized test potentials. Data were pooled from parallel experiments in the same batch of cells (n = 15 and 12 cells, respectively, for WT + 1 and L567Q + 1 groups) recorded under identical experimental conditions. Smooth curves represent fits to Boltzmann distributions with midpoint potentials 42.4 mV (slope factor 7.6 mV) and 35.3 mV (slope factor 10.1 mV), respectively, in WT and L567Q channels. B: voltage dependence of steady-state inactivation measured using a standard double-pulse protocol consisting of a 500-ms conditioning potential (150 to 0 mV, 10-mV increments) followed by a fixed test pulse (10 mV, 5 ms) to assess the fraction of activatable channels. Recordings were made in the same cells as described in A. Smooth curves represent fits to Boltzmann distributions with midpoint potentials 86.2 mV (slope factor 7.6 mV) and 97.5 mV (slope factor 7.3 mV), respectively, in WT and L567Q channels.

Negative shifts in the voltage dependence of steady-state inactivation also would be predicted to reduce Na⁺ conductance by reducing the availability of channels from diastolic potentials. We measured steady-state inactivation using the standard two-pulse method (500-ms conditioning pulse, 5-ms test pulse) and fit the data to Boltzmann distributions. For WT -subunits expressed in the ₁-cell line (Fig. 4B) one-half maximum availability occurred at a conditioning potential 86.2 mV, whereas for L567Q channels the midpoint was shifted to 97.5 mV (no change in slope factor). In a total of three sets of transfection experiments, we observed a statistically significant average shift of 8.8 ± 1.3 mV (n = 28 WT and 31 L567Q cells, P < 0.01). Similar results were obtained in parallel experiments performed in the absence of ₁-subunit, where the midpoint shifted by an average of 6.5 ± 1.7 mV (n = 5 WT and 6 L567Q cells, P < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Brugada syndrome is characterized by an ECG signature of S-T segment elevation, right bundle branch block, episodic ventricular tachycardia, and fibrillation. The most comprehensive explanation is that the arrhythmia originates from an abnormally large regional variation in action potential waveform, leading to a dispersion of repolarization and phase 2 reentry (8, 21). Regional variation in normal hearts may reflect nonuniform expression of I_to, the Ca-independent, transient outward K⁺ current (10, 11). In Brugada syndrome, however, reduced Na⁺ conductance may alter the balance of ionic currents (primarily repolarizing K⁺ currents, I_to, and depolarizing Na⁺ currents) during phase 1 (early repolarization). In normal hearts the activation of Na⁺ conductance during phase 0 prevents all-or-none repolarization during phase 1 and allows activation of Ca²⁺ channels, which generate the phase 2 plateau. A shift in the balance in favor of K⁺ conductance can lead to selective loss of the plateau and premature termination of the action potential, particularly in those regions of the heart where I_to is most prominent (e.g., right ventricular epicardium; see Ref. 10). Such a shift has been demonstrated in an in vitro canine model using Na⁺ channel blockers or K⁺ channel openers (21). Similarly, the unmasking of ECG abnormalities by administration of Na⁺ channel blockers in Brugada patients (4, 14) can be explained by an exacerbation of the mutation-induced reduction in Na⁺ conductance. Our present results provide clear evidence for the hypothesis that IVF mutations of SCN5A reduce Na⁺ channel function. Our most striking observation was that the L567Q channels inactivate more rapidly than WT, thus decreasing the availability of Na⁺ conductance. Second, a hyperpolarizing shift in the voltage dependence of inactivation may further reduce the pool of available Na⁺ channels at physiological potentials.

Potentiation of a slow phase of inactivation at negative potentials (50- to 150-ms range), deceleration of recovery, and destabilization of the fast phase at depolarized potentials have been implicated previously for 1795insD, a mutation associated with electrocardiographic features of both Brugada and long QT syndromes (19). Our results suggest that the Brugada mutation, L567Q, accelerates inactivation over a broad range of potentials encompassing both fully and partially activated channels. Unlike 1795insD, which alters the distribution between the amplitudes of the fast and slow phases of inactivation, L567Q clearly shortens time constants associated with the fast phase of inactivation (Table 1), with little effect on the distribution of amplitudes and little effect on the time course of recovery (Fig. 2, C-D, and Table 2). Nonetheless, both phenotypes are consistent with a loss-of-function etiology.

By contrast, some previous studies of Brugada SCN5A mutant channels in heterologous expression have not been entirely consistent with the reduced function hypothesis. Our initial characterization of T1620M expressed in Xenopus oocytes (5) and that of another laboratory (12) found a positive shift in the voltage dependence of inactivation and an acceleration of the time course of recovery, features that were not consistent with reduced Na⁺ channel function. However, a more recent report (7) suggests that, when analyzed in a mammalian expression system at near-physiological temperature, the T1620M mutant phenotype is qualitatively similar to that of L567Q, i.e., an acceleration of inactivation. A major difference, however, is that for T1620M the difference in inactivation rate disappeared at room temperature. Also, we observed a greater temperature sensitivity for WT + ₁ channels (Q₁₀ = 2.6) compared with that reported by Dumaine et al. (Q₁₀ = 1.2; see Ref. 7) for WT alone. Whether these differences are a reflection of mutant phenotypes or experimental conditions (e.g., presence of ₁-subunits) remains to be determined.

The L567Q mutation is located in the cytoplasmic linker between homologous transmembrane domains I and II, a highly variable region among vertebrate -subunit isoforms. The fact that Leu⁵⁶⁷ is not conserved and is located in a variable region suggests that it may play a unique role in modulating inactivation of the cardiac Na⁺ channel isoform.