Differential calcium sensitivity in Na_V1.5 mixed syndrome mutants

Abstract

Key points

• SCN5a mutations may express gain‐of‐function (Long QT Syndrome‐3), loss‐of‐function (Brugada Syndrome 1) or both (mixed syndromes), depending on the mutation and environmental triggers. One such trigger may be an increase in cytosolic calcium, accompanying exercise.

• Many mixed syndromes mutants, including ∆KPQ, E1784K, 1795insD and Q1909R, are found in calcium‐sensitive regions.

• Elevated cytosolic calcium attenuates gain‐of‐function properties in ∆KPQ, 1795insD and Q1909R, but not in E1784K. By contrast, elevated cytosolic calcium further exacerbates gain‐of‐function in E1784K by destabilizing slow inactivation.

• Action potential modelling, using a modified O'Hara Rudy model, suggests that elevated heart rate rescues action potential duration in ∆KPQ, 1795insD and Q1909R, but not in E1784K.

• Action potential simulations suggest that E1784K carriers have an increased intracellular sodium‐to‐calcium ratio under bradycardia and tachycardia conditions.

• Elevated cytosolic calcium, which is common during high heart rates, ameliorates or exacerbates the mixed syndrome phenotype depending on the genetic signature.

Abstract

Inherited arrhythmias may arise from mutations in the gene for SCN5a, which encodes the cardiac voltage‐gated sodium channel, Na_V1.5. Mutants in Na_V1.5 result in Brugada Syndrome (BrS1), Long‐QT Syndrome (LQT3) or mixed syndromes (an overlap of BrS1/LQT3). Exercise is a potential arrhythmogenic trigger in mixed syndromes. We aimed to determine the effects of elevated cytosolic calcium, which is common during exercise, in mixed syndrome Na_V1.5 mutants. We used whole‐cell patch clamp to assess the biophysical properties of Na_V1.5 wild‐type (WT), ∆KPQ, E1784K, 1795insD and Q1909R mutants in human embryonic kidney 293 cells transiently transfected with the Na_V1.5 α subunit (WT or mutants), β1 subunit and enhanced green fluorescent protein. Voltage‐dependence and kinetics were measured at cytosolic calcium levels of approximately 0, 500 and 2500 nm. In silico, action potential (AP) model simulations were performed using a modified O'Hara Rudy model. Elevated cytosolic calcium attenuates the late sodium current in ∆KPQ, 1795insD and Q1909R, but not in E1784K. Elevated cytosolic calcium restores steady‐state slow inactivation (SSSI) to the WT‐form in Q1909R, but depolarized SSSI in E1784K. Our AP simulations showed a frequency‐dependent reduction of AP duration in ∆KPQ, 1795insD and Q1909R carriers. In E1784K, AP duration is relatively prolonged at both low and high heart rates, resulting in a sodium overload. Cellular perturbations during exercise may affect BrS1/LQT3 patients differently depending on their individual genetic signature. Thus, exercise may be therapeutic or may be an arrhythmogenic trigger in some SCN5a patients.

Key points

• SCN5a mutations may express gain‐of‐function (Long QT Syndrome‐3), loss‐of‐function (Brugada Syndrome 1) or both (mixed syndromes), depending on the mutation and environmental triggers. One such trigger may be an increase in cytosolic calcium, accompanying exercise.

• Many mixed syndromes mutants, including ∆KPQ, E1784K, 1795insD and Q1909R, are found in calcium‐sensitive regions.

• Elevated cytosolic calcium attenuates gain‐of‐function properties in ∆KPQ, 1795insD and Q1909R, but not in E1784K. By contrast, elevated cytosolic calcium further exacerbates gain‐of‐function in E1784K by destabilizing slow inactivation.

• Action potential modelling, using a modified O'Hara Rudy model, suggests that elevated heart rate rescues action potential duration in ∆KPQ, 1795insD and Q1909R, but not in E1784K.

• Action potential simulations suggest that E1784K carriers have an increased intracellular sodium‐to‐calcium ratio under bradycardia and tachycardia conditions.

• Elevated cytosolic calcium, which is common during high heart rates, ameliorates or exacerbates the mixed syndrome phenotype depending on the genetic signature.

Abbreviations

α₁

fast amplitude

α₂

slow amplitude

τ₁

fast time constant

τ₂

slow time constant

AP

action potential

APD

action potential duration

BrS1

Brugada syndrome 1

CaM

calmodulin

CaMKII

calmodulin‐dependent protein kinase II

cDNA

complementary DNA

e₀

elementary charge

E_Na

Na⁺ Nernst equilibrium potential

ERC

electrical restitution curve

G/G_max

normalized conductance

G_Na

sodium channel conductance

GV

conductance

HEK293

human embryonic kidney 293

I

current amplitude

I/I_max

normalized current

I_ss

plateau current amplitude

I_Na

Na⁺ current

LQT3

long QT syndrome 3

late I_Na

late Na⁺ current

Na_V

voltage‐gated sodium channel

Na_V1.5

cardiac voltage‐gated sodium channel

NCX

sodium‐calcium exchanger

QT_C

corrected QT interval

RMP

resting membrane potential

SSFI

steady‐state fast inactivation

SSSI

steady‐state slow inactivation

t

time

V_1/2

midpoint of voltage‐dependence

V_m

command potential

WT

wild‐type

z

apparent valence

Introduction

Inherited arrhythmias may arise from mutations in the the gene for SCN5a, which encodes the cardiac voltage‐gated sodium channel (Na_V1.5) (Jones et al. 2011; Priori et al. 2013). These mutations result in gain‐of‐function (Long QT Syndrome; LQT3) or loss‐of‐function (Brugada Syndrome; BrS1) in Na_V1.5. Although BrS1 and LQT3 are clinically distinct, a subset of mutations simultaneously trigger both syndromes (BrS1/LQT3, mixed syndromes) (Rivolta et al. 2001; Clancy & Rudy, 2002; Makita et al. 2008).

BrS1 and LQT3 patients have differential expressivities depending on the physiological triggers, such as exercise (Shimizu & Antzelevitch, 2000; Veldkamp et al. 2000; Schwartz et al. 2001; Masrur et al. 2015). BrS1 patients manifest the diagnostic ST‐elevation during both exercise and recovery from exercise (Amin et al. 2009; Masrur et al. 2015). By contrast, LQT phenotype manifestation during exercise is genotype‐dependent (Schwartz et al. 2001). LQT1 patients have cardiac events during exercise, whereas LQT3 is mainly associated with arrhythmia during sleep (Schwartz et al. 1995; Shimizu & Antzelevitch, 2000). Although sleep is primarily a restful physiological state, the rapid eye movement phase includes exercise‐like physiological properties (Somers et al. 1993). A small proportion of LQT3 patients have lethal cardiac events during exercise (Schwartz et al. 1995).

Studies showing exercise‐induced QT_C shortening in LQT3 patients have focused on cases positive for the ∆KPQ mutation (Schwartz et al. 1995; Chandra et al. 1998). In ∆KPQ patients, the exercise‐induced QT_C shortening was linked to reductions in late sodium current (late I _Na) caused by elevated cytosolic calcium (Potet et al. 2015). The assumption that the LQT3 phenotype is rescued by exercise is highly controversial because SCN5a mutation responses to exercise‐induced triggers vary (Chen et al. 2015; Abdelsayed et al. 2015). For example, the C‐terminal mutant, E1784K, has greater channel availability with elevated temperature and stimulation frequencies (Abdelsayed et al. 2015), both of which are associated with exercise. E1784K carriers generally have a high phenotypic gain‐of‐function (LQT3) expressivity, as in the Okinawa islands, where the mutation is most prevalent, whereas a minority of patients display sinus node dysfunction, BrS1 or mixed syndrome phenotypes (Makita et al. 2008; Takahashi et al. 2014). Another C‐terminal mutant, V2016M, expresses protein kinase‐induced gain‐ and loss‐of‐function translating to epinephrine‐induced LQT3 and sinus node dysfunction (Chen et al. 2015). Unfortunately, little is known regarding the response of SCN5a mutations to physiological factors accompanying exercise. To better understand the variability in response to exercise among SCN5a‐mutation patients, it is critical to study Na_V1.5 mutants with cellular perturbations that mimic exercise.

With elevated heart rates, a rise in sympathetic tone triggers multiple cellular cascades that elevate cytosolic calcium (Song et al. 2001; Baartscheer et al. 2003). The Na_V1.5 C‐terminus contains a pair of EF‐hand domains and an IQ motif (Kim et al. 2004). Calcium effects are mediated via calmodulin (CaM) binding to the IQ motif (Chagot et al. 2009; Miloushev et al. 2009; Van Petegem et al. 2012). The Ca²⁺‐CaM complex interacts with the Domain III–IV linker, destabilizing inactivation and increasing Na_V1.5 channel availability during elevated heart rates (Shah et al. 2006). The presence of intact intra‐ and inter‐C‐terminal interactions is essential for calcium binding (Shah et al. 2006). We hypothesized that mixed syndrome mutants located in intracellular linkers or the C‐terminus may either strengthen or weaken the Ca²⁺‐CaM modulation of Na_V1.5. Thus, elevated cytosolic calcium may alleviate or exacerbate gain/loss‐of‐function properties in mutants, consequently rescuing or unmasking the phenotype. The BrS1/LQT3 mutants that we investigated include the Domain III‐IV linker mutant, ∆KPQ, and C‐terminal mutants, E1784K, 1795insD and Q1909R. All of these mutants are located in calcium‐sensitive regions known to modulate Na_V1.5 fast and slow inactivation.

Methods

Ethical approval

The research was approved by Biohazards review 251‐2012 issued by the office of the Environmental Health and Safety at Simon Fraser University, Burnaby, BC, Canada.

Cell culture and transfection

Human embryonic kidneys 293 (HEK293) cells were grown at pH 7.4 in a DMEM (1 ×) nutrient medium (Life Technologies, Grand Island, NY, USA), supplemented with 10% fetal bovine serum and maintained in a humidified environment at 37°C with 5% CO₂. To preserve Ca²⁺‐CaM effects on Na_V1.5, HEK293 cells were selected for the present study because they contain a relatively high [CaM]_free level compared to other cell lines (Black et al. 2004).

Transfection followed the procedures suggested by the manufacturer (Qiagen, Valancia, CA, USA). Briefly, 1.5 μg of the Na_V1.5 α subunit, 0.75 μg of the sodium channel β₁ subunit and 1.5 μg of enhanced green fluorescent protein were allowed to incubate with 15 μl of polyfect transfection reagent (Qiagen) and 146 μl of unsupplemented medium for 10 min. The cDNA mixture was then allowed to incubate with the HEK293 cells for 8 h before plating on coverslips.

Electrophysiology

Whole‐cell patch clamp recordings were performed in extracellular solution containing (mm): 96 NaCl, 4 KCl, 2 CaCl₂, 1 MgCl₂ and 10 Hepes (pH 7.4). Solutions were titrated with CsOH. Pipettes were fabricated with a P‐1000 puller using borosilicate glass (Sutter Instruments, Novato, CA, USA), dipped in dental wax to reduce capacitance, then thermally polished to a resistance of 1.0–1.5 MΩ. Voltage measurement error were minimized by using low resistance electrodes to limit series resistance between pipette and intracellular solution to 3.5 MΩ or less. Pipettes were filled with intracellular solution. For minimal cytosolic calcium levels, reported below as 0 nm, pipettes contained (mm): 130 CsF, 9.6 NaCl, 10 Hepes and 10 EGTA titrated to pH 7.4. To mimic average and peak systolic cytosolic calcium, we calculated, using the Ca‐EGTA Calculator TS, version 1.3 (http://maxchelator.stanford.edu), the amount of CaCl₂ added to bring cytosolic calcium to 500 and 2500 nm: 8.04 and 9.5 mm CaCl₂, respectively (Steenbergen et al. 1987; Kirschenlohr et al. 2000).

All recordings were made using an EPC‐9 patch clamp amplifier (HEKA Elektronik, Lambrecht, Germany) digitized at 20 kHz using an ITC‐16 interface (HEKA Elektronik). Data were acquired and low‐pass‐filtered (5 kHz) using PatchMaster/FitMaster software (HEKA Elektronik) running on an Apple iMac (Apple Computer, Cupertino, CA, USA). Leak subtraction was performed online using a P/4 procedure. Recordings were performed at room temperature (22°C). After a giga ohm resistance seal was achieved, the whole‐cell configuration was attained. The holding potential between protocols was −110 mV. We recorded I _Na from cells that expressed currents no greater than –5 nA. The average voltage error calculated for all cells used in this study (n = 362) is 5.16 ± 0.24 mV (Table 1). There are no differences between the voltage‐errors in the different conditions (P > 0.05).

Table 1.

Calcium: voltage error

NaV1.5 channel	[Ca2+]cytosolic	Voltage error (mV)	n
Wild‐type	0 nm Ca2+	4.45 ± 0.45	26
	500 nm Ca2+	5.10 ± 0.81	14
	2500 nm Ca2+	7.17 ± 1.15	13
∆KPQ	0 nm Ca2+	6.68 ± 0.67	33
	500 nm Ca2+	5.98 ± 0.83	24
	2500 nm Ca2+	7.67 ± 0.99	12
E1784K	0 nm Ca2+	6.63 ± 0.66	33
	500 nm Ca2+	4.49 ± 0.55	31
	2500 nm Ca2+	5.98 ± 0.56	20
1795insD	0 nm Ca2+	2.56 ± 0.32	28
	500 nm Ca2+	2.97 ± 0.51	21
	2500 nm Ca2+	2.98 ± 0.29	23
Q1909R	0 nm Ca2+	5.35 ± 0.46	36
	500 nm Ca2+	4.33 ± 0.83	27
	2500 nm Ca2+	5.02 ± 0.69	21

TTX‐subtraction experiments

To confirm the mutant‐induced increases in late I_Na, we performed TTX‐subtraction experiments at 0 nm cytosolic calcium, where the late I _Na in mutant channels was the largest. The concentration of TTX used was 40 μm in extracellular solution. The 40 μm TTX current trace was subtracted from the control (non‐TTX) current trace in all conditions to calculate the TTX‐sensitive late I _Na.

Analysis and statistics

Analysis and graphing were done using FitMaster software (HEKA Elektronik) and Igor Pro (Wavemetrics, Lake Oswego, OR, USA) with statistical information derived using JMP statistical software (SAS Institute Inc., Cary, NC, USA). Statistical significance was accepted at P < 0.05 using a two‐factor completely randomized design ANOVA test followed by a post hoc Tukey test. Statistical results were obtained for the channel variant, calcium, and channel variant × calcium factors. We report the statistical results for the mutants and the interaction between mutants and calcium. All values reported are given as the mean ± SEM.

Voltage protocols

Peak sodium current density (I _Na)

We measured current density from the ratio of peak current amplitude in pA to the cell membrane capacitance in pF.

Maximal peak and late sodium conductance density (G _Na)

Maximal conductance density relates to membrane channel trafficking and expression. We could not directly measure maximal peak or late conductance density, rather we indirectly calculated it from the conductance at 0 mV, measured by a test pulse described with respect to conductance (GV) (see below):

$$G_{\mathrm{Na}}=g_{\mathrm{Na}}/C_{\mathrm{M}}$$ (1)

where G _Na is the conductance at 0 mV, g_Na is the conductance at 0 mV (calculated using eqn (2)) and C _M is the membrane capacitance (pF).

Conductance (GV)

To determine the voltage dependence of activation, we measured the peak current amplitude at test pulse potentials ranging from −100 mV to +80 mV in increments of +10 mV for 19 ms. Prior to the test pulse, channels were allowed to recover from fast inactivation at −130 mV for 197 ms. Channel conductance was calculated from peak I _Na.

$$g_{\mathrm{Na}}=I_{\mathrm{Na}}/V-E_{\mathrm{rev}}$$ (2)

where g _Na is sodium channel conductance, I _Na is peak sodium current in response to the command potential V, and E _rev is the reversal potential. Calculated values for conductance were fit with the Boltzmann function:

$$G/G_{\max }=1/(1+\exp [-z{e}_0[V_{\mathrm{m}}-V_{1/2}]/\mathrm{kT}])$$ (3)

where G/G _max is the normalized conductance amplitude, V _m is the command potential, z is the apparent valence, e ₀ is the elementary charge, V _1/2 is the midpoint voltage, k is the Boltzmann constant and T is temperature (K).

Steady‐state fast inactivation (SSFI)

The voltage‐dependence of SSFI was measured by preconditioning the channels to a hyperpolarizing potential of −130 mV and then eliciting prepulses between −150 mV and +10 mV in increments of 10 mV for 500 ms. Channel availability was assessed during a test pulse to 0 mV. Different hyperpolarizing prepulse potentials (−130 mV or −150 mV) were used to obtain a plateau for the SSFI curve. Normalized current amplitude as a function of voltage was fit using the Boltzmann function:

$$I/I_{\max }=1/(1+\exp [-z{e}_0[V_{\mathrm{m}}-V_{1/2}]/\mathrm{kT}])$$ (4)

where I/I _max is the normalized current amplitude, z is apparent valence, e ₀ is the elementary charge, V _m is the prepulse potential, V _1/2 is the midpoint voltage of SSFI, k is the Boltzmann constant and T is temperature (K).

Fast inactivation onset

Time constants for open‐state fast inactivation were derived by fitting a double exponential function to the decay of current obtained from the activation protocol. To measure closed‐state fast inactivation onset, channels were preconditioned at −130 mV prior to a prepulse at −50 mV, −70 mV or −90 mV for 0 – 0.256 s. Current amplitude was measured during a test pulse to 0 mV for 20 ms. Normalized current amplitudes as a function of time were fit using a double exponential equation:

$$I=I_{\mathrm{ss}}+{\alpha }_1\exp (-t/{\tau }_1)+{\alpha }_2\exp (-t/{\tau }_2)$$ (5)

where I is current amplitude, I _ss is the plateau amplitude, α₁ and α₂ are the amplitudes at time 0 for time constants τ₁ and τ₂, and t is time.

Fast inactivation recovery

Channels were fast‐inactivated during a 500 ms depolarizing step to 0 mV. Recovery was measured during a 19 ms test pulse to 0 mV following 0–1.024 s conditioning pulses at −130 mV, −110 mV or −90 mV. Time constants of fast inactivation recovery as a function of time were fit using a double exponential equation, as above.

Persistent current

Persistent current was measured between 450 and 500 ms during a 500 ms depolarizing pulse to 0 mV from a holding potential of –130 mV. An average of 30 pulses was used to increase the signal‐to‐noise ratio.

Steady‐state slow inactivation (SSSI)

The voltage‐dependence of SSSI was measured by preconditioning the channels to −150 mV for 30 s and then eliciting prepulse potentials that range from −150 to −10 mV in increments of 20 mV for 60 s. Channel availability was assessed during a test pulse to 0 mV following a −130 mV recovery pulse from fast inactivation at 20 ms. Normalized current amplitude as a function of voltage was fit using a modified Boltzmann function:

$$\begin{array}{ccc}\hfill I/I_{\max }& =& (I_1-I_2)/(1+\exp [-z{e}_0[V_{\mathrm{m}}-V_{1/2}]/\mathrm{kT}])\hfill \\ & & +I_2\hfill \end{array}$$ (6)

where I ₁ and I ₂ are maximum and minimum values of fit. The other symbols are as defined previously.

Slow inactivation onset

To measure onset into slow inactivation, channels were preconditioned at −130 mV for 30 s prior to a prepulse at 0 mV for 0–64 s. A test pulse to 0 mV followed a −130 mV fast inactivation recovery pulse for 20 ms. Normalized current amplitude as a function of time was fit with a double exponential.

Slow inactivation recovery

To measure recovery from slow inactivation, channels were preconditioned at −130 mV for 30 s prior to a prepulse at 0 mV for 60 s, followed by series of test pulses to 0 mV for 20 ms between increasing incremental recovery durations at −130 mV for 0 – 32s. Normalized current amplitudes as a function of time were fit using a double exponential equation with the plateau equal to 1.00.

Weighted slow inactivation time constants

The weighted slow inactivation time constant at each voltage was calculated using the double τ values obtained from eqn (5):

$$\begin{array}{c}\hfill {\tau }_{\mathrm{slow}-\mathrm{weighted}}=(({\alpha }_1\times {\tau }_1)+({\alpha }_2\times {\tau }_2))/({\alpha }_1+{\alpha }_2)\end{array}$$ (7)

Slow inactivation α_j and β_j rates

The recovery from slow inactivation (forward, α_j) and the onset into slow inactivation (reverse, β_j) rates were calculated using Hodgkin–Huxley formulations:

$${\alpha }_{\mathrm{j}}={\mathrm{j}}_{\infty }/{\tau }_{\mathrm{j}}$$ (8)

$${\beta }_{\mathrm{j}}=(1-{\mathrm{j}}_{\infty })/{\tau }_{\mathrm{j}}$$ (9)

where j _∞ is channel availability obtained from eqn (6) and τ_j is the weighted slow inactivation time constant obtained from eqn (7), at a given voltage. The slow inactivation α_j rates vs. voltage were fit using a single exponential curve and β rates were fit with a line with a slope of 0.

Myocardial action potential (AP) modelling

APs were simulated using a modified version of the O'Hara Rudy (ORd) model at 37°C programmed in Python (O'Hara et al. 2011). The sodium data were extrapolated to physiological temperatures using previously reported Q ₁₀ values for WT, ∆KPQ and E1784K variants (Nagatomo et al. 1998; Abdelsayed et al. 2015). The thermosensitivity of 1795insD and Q1909R mutants was never reported; thus, we used the WT temperature‐dependence for these mutants. The maximal G_Na density was 75 mS μF⁻¹ across all channel variants in the ORd model. We modified the gating I _Na parameters data in accordance with our biophysical data for the various channel variants. The GV, SSFI and SSSI midpoints and slopes of the channel variants extrapolated to 37°C were normalized to the original ORd parameters. The normalized midpoint and slope values for GV and SSFI were then incorporated into the Boltzmann curve (eqn (4)) and SSSI into the modified Boltzmann curve (eqn (6)). The original ORd SSSI (j_∞) curve was equated to the SSFI curve (h _∞), which had no plateau. Our experimental SSSI plateaus were subtracted from the original ORd value of 0 to compensate for this caveat. The phosphorylated steady‐state fast inactivation midpoints in all channel variants were equally hyperpolarized by 6.2 mV. Late I _Na density was normalized to the original ORd value and multiplied by the percentage of late to peak I _Na calculated above. The relationship between fast inactivation time constants vs. voltage was calculated using an inverse double exponential distribution at 37°C. The ORd slow inactivation time constant equation was used for all the channel variants since it encompassed a comprehensive range of voltages.

To model the calcium‐dependence of our I _Na data, we fit the extrapolated biophysical parameters to 37°C with a Hill equation:

$$Z=Y_0+(Y_{\mathrm{M}}-Y_0)/\left(1+{\left(X_{1/2}/X\right)}^{\mathrm{b}}\right)$$ (10)

where Z is the biophysical parameter of interest, Y ₀ is the minimum value, Y _M is the maximum value, X _1/2 is the midpoint of the curve, X is the intracellular cytosolic calcium and b is the rate.

Subspace calcium was not accounted for as a result of the lack of experimental data. Thus, the modified ORd model is a dynamic simulation of the calcium‐induced shifts that are observed with increasing intracellular calcium levels as a function of pacing frequency, comprising the positive staircase phenomenon.

Simulations were run on endocardial and epicardial ventricular myocytes using a 0.5 ms stimulus pulse with an amplitude of −80 μA μF⁻¹. The stimulus protocol was designed accordingly to step up the frequency gradually from 0.5 Hz to 2.5 Hz. The stimulus protocol had a cycle length of 2000 ms for the first 75 APs, 1000 ms for the following 75 APs, 667 ms for the following 75 APs, 500 ms for the following 75 APs and 400 ms for the following 100 APs.

Analysis of APs only included those that fully recovered and were restored to baseline. Action potential duration (APD) was measured at 30%, 60% and 90% of repolarization by multiplying these percentages by the resting membrane potential (RMP) value prior to the current stimulus pulse. The APD_X (X = 30, 60, 90) values were plotted vs. the diastolic interval (DI = BCL – APD_X), where BCL is the basic cycle length, creating electrical restitution curves. The curves were fitted with a double‐exponential eqn (5). AP velocity was calculated from initiation of the current stimulus to the maximal peak.

Results

Peak current amplitude and conductance density

Raw I _Na traces are shown in Fig. 1 for all the channel variants at 0 and 2500 nm. The current density of 1795insD was smaller (P < 0.0001) (Table 2) compared to WT and the rest of the mutants. Elevations in cytosolic calcium had no effect on current density in any of the channel variants (P > 0.05) (Fig. 1 A) (Table 2). Peak conductance density was also reduced in 1795insD compared to WT and the mutants (P < 0.0001) (Fig. 1 B and Table 2). In ∆KPQ, peak conductance density was larger (P = 0.0158) (Table 2) by 4.81 ± 1.24 pA/pF when cytosolic calcium was elevated from 0 to 2500 nm. No other mutants were affected by elevations in cytosolic calcium (P > 0.05) (Table 2).

Figure 1. Current traces and properties.

Raw current traces for the wild‐type and mutant channels at 0 and 2500 nm cytosolic calcium. A–C, peak I _Na current density, peak I _Na conductance density and late I _Na conductance density, respectively, as a function of the three cytosolic calcium levels. The asterisk in (B) indicates a rise in ∆KPQ conductance density with elevations in cytosolic calcium from 500 to 2500 nm. [Color figure can be viewed at wileyonlinelibrary.com]

Table 2.

Calcium: absolute conductance and current density

NaV1.5 channel	[Ca2+]cytosolic	Peak G max (nS/pF)	n	Current density (pA/pF)	n
Wild‐type	0 nm Ca2+	4.84 ± 0.57	18	−258.9 ± 23.8	17
	500 nm Ca2+	6.02 ± 0.69	11	−351.7 ± 44.1	13
	2500 nm Ca2+	8.58 ± 1.27	6	−452.5 ± 67.6	5
∆KPQ	0 nm Ca2+	6.60 ± 0.86	29	−386.3 ± 49.7	30
	500 nm Ca2+	5.27 ± 0.64	20	−308.5 ± 37.6	20
	2500 nm Ca2+	10.08 ± 1.27*1	8	−427.8 ± 68.1	8
E1784K	0 nm Ca2+	6.70 ± 0.59	21	−326.1 ± 27.4	16
	500 nm Ca2+	5.24 ± 0.60	29	−361.1 ± 20.5	19
	2500 nm Ca2+	6.55 ± 0.77	11	−419.1 ± 45.2	13
1795insD#1	0 nm Ca2+	2.85 ± 0.46	17	−209.8 ± 30.3	12
	500 nm Ca2+	3.27 ± 0.42	19	−205.9 ± 9.39	11
	2500 nm Ca2+	3.62 ± 0.68	12	−245.7 ± 39.6	10
Q1909R	0 nm Ca2+	4.62 ± 0.57	23	−311.4 ± 33.2	19
	500 nm Ca2+	5.00 ± 0.64	20	−283.6 ± 29.0	16
	2500 nm Ca2+	4.33 ± 0.59	15	−245.4 ± 30.1	13

^#1 P < 0.0001 vs. WT (current density).

^*1 P < 0.05 vs. ∆KPQ (0 and 500 nm).

Activation voltage‐dependence

Normalized conductance is plotted as a function of membrane potential in Fig. 2. Activation midpoint (GV‐V _1/2) was depolarized in all mutants compared to WT (P < 0.0001) (Table 3). The mutants depolarized GV‐V _1/2 in an increasing order compared to WT: Q1909R, 1795insD, E1784K and ∆KPQ, respectively. In all channel variants, GV‐V _1/2 was not affected by elevations in cytosolic calcium (P > 0.05) (Table 3). Activation slope (GV‐z) was reduced in the same order as the mutant‐induced GV‐V _1/2 shifts (P < 0.0001) (Table 3). Elevations in cytosolic calcium did not affect GV‐z in any channel variant (P > 0.05) (Table 3).

Figure 2. Effects of cytosolic calcium on conductance.

A and B, mutant effects at 0 and 2500 nm cytosolic calcium. B, voltage pulse inset. C–F, calcium effect on each individual mutant compared to WT. Normalized conductance vs. membrane potential was fit with a Boltzmann fit. [Color figure can be viewed at wileyonlinelibrary.com]

Table 3.

Calcium: activation

NaV1.5 channel	[Ca2+]cytosolic	GV‐V 1/2 (mV)	GV‐z	n
Wild‐type	0 nm Ca2+	−43.8 ± 1.10	4.92 ± 0.26	18
	500 nm Ca2+	−43.7 ± 1.55	3.94 ± 0.30	10
	2500 nm Ca2+	−44.5 ± 0.66	5.02 ± 0.37	6
∆KPQ#3	0 nm Ca2+	−36.2 ± 0.99	2.61 ± 0.10	25
	500 nm Ca2+	−37.3 ± 1.01	3.06 ± 0.14	10
	2500 nm Ca2+	−36.5 ± 0.45	3.45 ± 0.13	7
E1784K#2	0 nm Ca2+	−39.2 ± 0.96	3.35 ± 0.12	21
	500 nm Ca2+	−39.7 ± 0.80	3.60 ± 0.16	20
	2500 nm Ca2+	−36.8 ± 0.93	3.67 ± 0.25	10
1795insD#1	0 nm Ca2+	−42.1 ± 0.94	3.80 ± 0.21	16
	500 nm Ca2+	−38.9 ± 0.92	3.68 ± 0.23	17
	2500 nm Ca2+	−36.9 ± 1.11	3.36 ± 0.21	11
Q1909R#1	0 nm Ca2+	−42.6 ± 1.01	4.17 ± 0.18	24
	500 nm Ca2+	−40.8 ± 0.90	3.85 ± 0.22	17
	2500 nm Ca2+	−39.8 ± 0.75	4.30 ± 0.21	12

^#1 p < 0.0001 vs. WT.

^#2 p < 0.0001 vs. WT and Q1909R.

^#3 p < 0.0001 vs. WT, Q1909R, and 1795insD.

Fast and intermediate inactivation voltage‐dependence

Normalized current vs. membrane potential is shown in Fig. 3. The steady‐state fast inactivation midpoint (SSFI‐V _1/2) was hyperpolarized by ∆KPQ, E1784K and 1795insD compared to WT (P < 0.0001) (Table 4). The SSFI‐V _1/2 of ∆KPQ was depolarized (P < 0.0001) by elevations in cytosolic calcium compared to the other channel variants (P > 0.05) (Table 4). When cytosolic calcium was elevated from 0 to 500 nm, the SSFI‐V _1/2 of ∆KPQ depolarized by 8.03 ± 1.82 mV (P < 0.0001) (Table 4). Additional elevations in cytosolic calcium to 2500 nm did not further depolarize the SSFI‐V _1/2 of ∆KPQ (P > 0.05) (Fig. 3 C and Table 4). The steady‐state fast inactivation slope (SSFI‐z) was reduced in E1784K, 1795insD and Q1909R compared to WT (P = 0.0021) (Table 4). Elevations in cytosolic calcium had no effects on SSFI‐z in any channel variant (P > 0.05) (Table 4). Steady‐state intermediate inactivation (SSII) was measured using a similar protocol to SSFI, except with 1000 ms prepulse durations (Table 4). The SSII‐V _1/2 of E1784K and 1795insD are hyperpolarized compared to WT (P < 0.0001) (Table 4). Both mutants also reduced SSII‐z (P < 0.0001) (Table 4) compared to WT. Elevations in cytosolic calcium had no effect on both SSII‐V _1/2 and SSII‐z in any channel variant (P > 0.05) (Table 4).

Figure 3. Effects of cytosolic calcium on steady‐state fast inactivation.

A and B, mutant effects at 0 and 2500 nm cytosolic calcium. B, voltage pulse inset. C–F, calcium effect on each individual mutant compared to WT. Normalized current vs. membrane potential was fit with a Boltzmann fit. [Color figure can be viewed at wileyonlinelibrary.com]

Table 4.

Calcium: steady‐state fast and intermediate inactivation

NaV1.5 channel	[Ca2+]cytosolic	SSFI‐V 1/2 (mV)	SSFI‐z	n	SSII‐V 1/2 (mV)	SSII‐z	n
Wild‐type	0 nm Ca2+	−89.2 ± 1.97	−3.72 ± 0.28	6	−88.7 ± 1.76	−4.31 ± 0.07	7
	500 nm Ca2+	−89.7 ± 0.79	−3.91 ± 0.23	8	−88.9 ± 2.66	4.05 ± 0.09	5
	2500 nm Ca2+	−88.2 ± 0.89	−3.92 ± 0.17	8	−90.6 ± 0.68	3.90 ± 0.14	8
∆KPQ#1	0 nm Ca2+	−100.1 ± 1.64*1	−3.80 ± 0.10	13	−97.2 ± 2.68	−3.70 ± 0.19	7
	500 nm Ca2+	−92.1 ± 0.86	−3.63 ± 0.21	11	−90.7 ± 1.79	−4.12 ± 0.15	6
	2500 nm Ca2+	−89.8 ± 0.76	−3.91 ± 0.08	8	−91.0 ± 0.80	−3.86 ± 0.09	8
E1784K#1	0 nm Ca2+	−95.9 ± 0.77	−3.36 ± 0.09	6	−99.3 ± 1.34	−3.08 ± 0.11	8
	500 nm Ca2+	−95.6 ± 2.28	−3.28 ± 0.12	7	−98.5 ± 1.57	−3.17 ± 0.13	6
	2500 nm Ca2+	−99.0 ± 1.59	−3.34 ± 0.15	6	−100.9 ± 1.86	−3.38 ± 0.15	6
1795insD#1	0 nm Ca2+	−99.0 ± 1.95	−3.71 ± 0.11	7	−101.9 ± 1.50	−3.74 ± 0.11	7
	500 nm Ca2+	−94.3 ± 1.63	−3.36 ± 0.23	12	−100.0 ± 1.38	−3.44 ± 0.14	8
	2500 nm Ca2+	−95.4 ± 1.42	−3.34 ± 0.25	7	−98.3 ± 1.52	−3.20 ± 0.19	8
Q1909R	0 nm Ca2+	−87.6 ± 1.99	−3.15 ± 0.14	6	−92.3 ± 1.76	−3.58 ± 0.18	7
	500 nm Ca2+	−89.3 ± 1.89	−3.46 ± 0.16	7	−86.3 ± 1.94	−4.03 ± 0.44	6
	2500 nm Ca2+	−89.1 ± 1.80	−3.57 ± 0.20	7	−87.1 ± 1.63	−3.56 ± 0.12	5

^#1 P < 0.0001 vs. WT.

^*1 P < 0.01 vs. ∆KPQ (500 and 2500 nm).

Fast inactivation recovery and onset kinetics

Double‐pulse protocols were used to measure onset (τ_on) and recovery (τ_off) kinetics of fast inactivation. The time constant (τ) obtained from the fits to the recovery and onset curves equals the inverse of the sum of both the forward (recovery) and reverse (onset) rates. At voltages hyperpolarized relative to SSFI‐V _1/2, recovery from fast inactivation predominates. At voltages depolarized relative to SSFI‐V _1/2, onset into fast inactivation predominates. At SSFI‐V _1/2, both onset and recovery are in equilibrium.

Fast inactivation time constants as a function of voltage are reported in Tables 5 and 6. At −10 mV, the fast time constant (τ₂) ∆KPQ decreased when cytosolic calcium was elevated from 0 to 500 nm. 1795insD slow time constant (τ₂) is increased at −30 mV when cytosolic calcium was elevated from 500 to 2500 nm. Q1909R fast inactivation kinetics were decelerated at both open‐state and closed‐state voltages with elevations in cytosolic calcium.

Table 5.

Calcium: fast inactivation time constants (−130 mV to −70 mV)

		−130 mV			−110 mV			−90 mV			−70 mV
NaV1.5	[Ca2+]cytosolic	τ1 (ms)	τ2 (ms)	n	τ1 (ms)	τ2 (ms)	n	τ1 (ms)	τ2 (ms)	n	τ1 (ms)	τ2 (ms)	n
Wild‐type	0 nm Ca2+	3.941 ± 0.576	113.207 ± 36.033	7	13.922 ± 2.042	288.737 ± 83.6	5	70.367 ± 11.48	91.753 ± 5.5	10	43.208 ± 10.94	155.028 ± 27.722	10
	500 nm Ca2+	4.329 ± 0.49	200.897 ± 89.522	7	12.504 ± 0.449	170.052 ± 73.005	5	56.259 ± 7.922	177.285 ± 37.913	9	20.699 ± 4.161*2	91.376 ± 17.83	9
	2500 nm Ca2+	5.275 ± 0.435	115.344 ± 24.523	7	17.254 ± 0.855	119.095 ± 14.303	7	75.668 ± 5.818	94.142 ± 8.632	7	17.927 ± 1.542	82.965 ± 15.411	7
∆KPQ	0 nm Ca2+	3.095 ± 0.264	77.184 ± 6.612	7	7.956 ± 0.711	203.226 ± 17.348	7	8.473 ± 1.045	29.45 ± 8.374	8	3.803 ± 0.645	24.974 ± 12.374	6
	500 nm Ca2+	2.995 ± 0.249	76.082 ± 13.389	7	7.158 ± 0.452	167.462 ± 16.227	8	8.666 ± 1.042	29.207 ± 10.683	8	3.440 ± 0.408	10.512 ± 6.883	8
	2500 nm Ca2+	3.333 ± 0.229	74.019 ± 12.928	7	8.367 ± 0.714	207.649 ± 36.71	7	10.019 ± 0.882	51.999 ± 19.538	6	6.414 ± 0.383	28.552 ± 15.477	6
E1784K	0 nm Ca2+	3.036 ± 0.303	158.209 ± 70.225	7	7.484 ± 0.469	105.837 ± 23.739	5	9.172 ± 1.016	31.887 ± 7.281	7	3.763 ± 0.483	16.854 ± 5.464	7
	500 nm Ca2+	3.252 ± 0.53	86.18 ± 24.568	7	7.888 ± 1.249	211.048 ± 58.341	6	7.449 ± 1.056	20.339 ± 5.405	7	3.599 ± 0.821	20.192 ± 7.544	7
	2500 nm Ca2+	5.060 ± 0.86	159.223 ± 56.861	7	12.899 ± 1.882	60.282 ± 17.81	7	11.018 ± 2.062	61.372 ± 22.898	7	3.814 ± 0.555	13.382 ± 2.915	7
1795insD	0 nm Ca2+	8.602 ± 1.234	248.962 ± 32.044	7	24.898 ± 3.652	227.888 ± 84.097	6	21.872 ± 5.41	79.799 ± 12.426	5	10.436 ± 2.097	29.186 ± 5.885	5
	500 nm Ca2+	6.367 ± 0.997	484.911 ± 156.726	12	19.949 ± 3.543	225.552 ± 39.381	7	16.374 ± 2.674	78.855 ± 9.549	15	10.949 ± 1.983	37.084 ± 5.399	14
	2500 nm Ca2+	6.443 ± 0.437	130.277 ± 27.897	10	17.676 ± 2.286	170.472 ± 46.222	8	37.030 ± 15.64	91.949 ± 21.144	6	3.537 ± 1.082	22.991 ± 5.187	5
Q1909R	0 nm Ca2+	3.434 ± 0.395	68.027 ± 11.146	6	9.918 ± 2.101	145.741 ± 26.439	5	58.390 ± 12.43	80.104 ± 14.21	6	26.721 ± 6.316	89.129 ± 11.004	5
	500 nm Ca2+	4.723 ± 0.836	116.142 ± 31.404	7	10.855 ± 2.233	227.291 ± 39.757	10	51.694 ± 19.717	48.621 ± 6.979	9	18.961 ± 5.032	74.539 ± 10.697	9
	2500 nm Ca2+	5.852 ± 0.586	128.702 ± 28.236	6	20.736 ± 1.817	160.045 ± 45.202	6	83.215 ± 35.211	136.728 ± 35.407*1	7	33.252 ± 10.007	102.469 ± 27.953	8

^*1 P < 0.001 vs. 0 and 500 nm Q1909R.

^*2 P < 0.05 vs. 0 nm WT.

Table 6.

Calcium: fast inactivation time constants (−50 mV to +10 mV)

		−50 mV			−30 mV			−10 mV			+10 mV
NaV1.5	[Ca2+]cytosolic	τ1 (ms)	τ2 (ms)	n	τ1 (ms)	τ2 (ms)	n	τ1 (ms)	τ2 (ms)	n	τ1 (ms)	τ2 (ms)	n
Wild‐type	0 nm Ca2+	6.316 ± 0.96	14.054 ± 6.117	10	0.879 ± 0.069	3.229 ± 0.23	21	0.502 ± 0.022	2.561 ± 0.443	21	0.392 ± 0.015	4.185 ± 1.796	18
	500 nm Ca2+	3.086 ± 0.16	3.449 ± 0.472	5	0.926 ± 0.104	3.359 ± 0.352	16	0.548 ± 0.049	3.248 ± 0.355	15	0.393 ± 0.046	1.973 ± 0.372	14
	2500 nm Ca2+	4.405 ± 0.467	4.512 ± 0.576	7	0.929 ± 0.098	6.346 ± 2.845	8	0.576 ± 0.028	1.991 ± 0.46	8	0.442 ± 0.018	2.684 ± 0.751	8
∆KPQ	0 nm Ca2+	1.412 ± 0.222	1.415 ± 0.224	6	0.482 ± 0.021	3.678 ± 0.617	34	0.452 ± 0.016*3	5.461 ± 1.354	33	0.457 ± 0.018	4.946 ± 1.984	32
	500 nm Ca2+	1.004 ± 0.09	2.159 ± 0.848	8	0.579 ± 0.054	3.384 ± 0.676	19	0.501 ± 0.033	3.992 ± 1.008	19	0.449 ± 0.017	3.893 ± 0.702	14
	2500 nm Ca2+	1.468 ± 0.302	1.932 ± 0.238	6	0.605 ± 0.018	4.998 ± 0.904	11	0.577 ± 0.021	4.341 ± 0.804	11	0.590 ± 0.025*5	3.15 ± 0.712	11
E1784K	0 nm Ca2+	0.753 ± 0.301	1.545 ± 0.348	5	0.460 ± 0.028	6.617 ± 3.339	22	0.334 ± 0.016	4.171 ± 1.235	22	0.273 ± 0.019	2.309 ± 0.747	21
	500 nm Ca2+	1.086 ± 0.24	1.212 ± 0.147	7	0.503 ± 0.026	2.13 ± 0.476	29	0.340 ± 0.012	3.294 ± 1.427	29	0.262 ± 0.016	2.014 ± 0.671	27
	2500 nm Ca2+	1.589 ± 0.193	1.789 ± 0.217	7	0.534 ± 0.033	3.802 ± 1.375	13	0.404 ± 0.014	2.13 ± 0.337	13	0.341 ± 0.014	1.464 ± 0.26	14
1795insD	0 nm Ca2+	3.534 ± 0.365	3.537 ± 0.402	6	0.810 ± 0.033	4.621 ± 0.846	15	0.455 ± 0.015	3.148 ± 0.671	15	0.309 ± 0.018	2.49 ± 0.842	15
	500 nm Ca2+	4.735 ± 0.645	4.685 ± 0.616	5	0.900 ± 0.061	9.175 ± 4.162	19	0.510 ± 0.026	5.503 ± 1.91	18	0.407 ± 0.018	3.926 ± 1.192	15
	2500 nm Ca2+	4.221 ± 0.383	4.233 ± 0.401	6	0.960 ± 0.101	18.716 ± 5.981*2	13	0.557 ± 0.03	4.633 ± 1.146	12	0.407 ± 0.022	5.987 ± 1.477	11
Q1909R	0 nm Ca2+	5.902 ± 0.893	21.729 ± 4.744	5	0.991 ± 0.076	9.119 ± 2.809	27	0.597 ± 0.021	3.055 ± 0.307	28	0.461 ± 0.023	2.45 ± 0.261	27
	500 nm Ca2+	5.374 ± 1.236	20.443 ± 4.78	7	0.884 ± 0.065	4.627 ± 0.544	24	0.532 ± 0.024	3.153 ± 0.288	23	0.378 ± 0.017	2.23 ± 0.343	22
	2500 nm Ca2+	7.864 ± 0.742	34.005 ± 9.824	7	1.333 ± 0.125*1	7.178 ± 1.547	14	0.753 ± 0.042*4	5.112 ± 1.272	14	0.616 ± 0.024*6	4.744 ± 0.925	12

^*1 P < 0.05 vs. 0 and 500 nm Q1909R.

^*2 P < 0.05 vs. 0 nm 1795insD.

^*3 P < 0.05 vs. 500 and 2500 nm ∆KPQ.

^*4 P < 0.05 vs. 0 and 500 nm Q1909R.

^*5 P < 0.05 vs 0 and 500 nm ∆KPQ.

^*6 P < 0.05 vs 0 and 500 nm Q1909R.

Late current amplitude and conductance density

Late I _Na conductance density was measured similarly to peak conductance density. Elevations in cytosolic calcium had no effect on late conductance density in any of the channel variants (P > 0.05) (Fig. 1 C and Table 7).

Table 7.

Calcium: late sodium current

NaV1.5 channel	[Ca2+]cytosolic	Late I Na (%)	n	Late G Max (nS/pF)	n
Wild‐type	0 nm Ca2+	0.49 ± 0.13	13	0.060 ± 0.013	9
	500 nm Ca2+	0.67 ± 0.20	5	0.077 ± 0.010	5
	2500 nm Ca2+	0.62 ± 0.13	6	0.061 ± 0.019	7
∆KPQ#1	0 nm Ca2+	1.94 ± 0.35*1	9	0.110 ± 0.013	11
	500 nm Ca2+	1.48 ± 0.36	5	0.125 ± 0.026	7
	2500 nm Ca2+	0.36 ± 0.14	6	0.077 ± 0.020	5
E1784K#2	0 nm Ca2+	0.95 ± 0.15	11	0.067 ± 0.016	11
	500 nm Ca2+	2.05 ± 0.30	12	0.138 ± 0.021	15
	2500 nm Ca2+	2.06 ± 0.43	10	0.121 ± 0.034	16
1795insD#3	0 nm Ca2+	3.68 ± 0.31*1	6	0.088 ± 0.018	9
	500 nm Ca2+	2.72 ± 0.48	5	0.116 ± 0.017	6
	2500 nm Ca2+	1.74 ± 0.18	9	0.047 ± 0.004	9
Q1909R#4	0 nm Ca2+	3.67 ± 0.44*2	7	0.054 ± 0.007	9
	500 nm Ca2+	1.30 ± 0.40	5	0.053 ± 0.007	4
	2500 nm Ca2+	1.21 ± 0.16	5	0.089 ± 0.011	6

^#1 P < 0.0001 vs. WT, 1795insD, Q1909R.

^#2 P < 0.0001 vs. WT, 1795insD.

^#3 P < 0.0001 vs. WT, ∆KPQ, E1784K.

^#4 P < 0.0001 vs. WT, ∆KPQ.

^*1 P < 0.0001 vs. same mutant (2500 nm).

^*2 P < 0.0001 vs. Q1909R (500 and 2500 nm).

Raw I _Na current records emphasizing late I _Na are shown in Fig. 4 A–D. Traces of TTX‐sensitive late I _Na in all mutants compared to WT are shown in Fig. 4 E–H. The current traces were normalized to peak I _Na. Late I _Na was larger in the mutants compared to WT (P < 0.0001) (Table 7). Late I _Na was lower by 1.58% ± 0.44% in ∆KPQ and by 1.95% ± 0.44% in 1795insD when cytosolic calcium was elevated from 0 to 2500 nm (P < 0.0001) (Fig. 4 I and Table 7). In Q1909R, late I _Na was lower by 2.36% ± 0.48% when cytosolic calcium was elevated from 0 to 500 nm (P < 0.0001) (Fig. 4 I and Table 7). Further elevations in cytosolic calcium did not affect late I _Na of Q1909R (P > 0.05) (Table 7). Late I _Na of E1784K was not affected by elevations in cytosolic calcium (P > 0.05) (Table 7).

Figure 4. Effects of cytosolic calcium on late I _Na .

A–D, effects of cytosolic calcium on the mutants compared to WT. The voltage pulse used to measure late I _Na is shown in (A). E–H, TTX‐subtracted late I _Na for all channel variants at 0 nm cytosolic calcium. I, late I_Na as a percentage of peak I _Na vs. the three cytosolic calcium concentrations. [Color figure can be viewed at wileyonlinelibrary.com]

Slow inactivation voltage‐dependence

Normalized current vs. membrane potential is shown in Fig. 5. The steady‐state slow inactivation midpoint (SSSI‐V _1/2) was not different in any of the channel variants (P > 0.05) (Table 8). The SSSI‐V _1/2 of E1784K was depolarized by 23.2 ± 5.20 mV when cytosolic calcium was elevated from 500 to 2500 nm (P = 0.0005) (Fig. 5 D and Table 8). The steady‐state slow inactivation slope (SSSI‐z) was not affected by the mutants (P > 0.05) (Table 8). SSSI‐z of 1795insD was larger by 0.79 ± 0.18 when cytosolic calcium was elevated from 500 to 2500 nm (P = 0.0035) (Table 8). The steady‐state slow inactivation plateau (SSSI‐Y ₀) was higher in E1784K by 21.7% ± 5.3% when cytosolic calcium was elevated from 0 to 2500 nm (P < 0.0001) (Fig. 5 D and Table 8). The SSSI‐Y ₀ of Q1909R was higher by 18.0% ± 0.05% when cytosolic calcium was elevated from 500 to 2500 nm (P < 0.0001) (Fig. 5 F and Table 8).

Figure 5. Effects of cytosolic calcium on steady‐state slow inactivation.

A and B, mutant effects at 0 and 2500 nm cytosolic calcium. B, voltage pulse inset. C–F, calcium effect on each individual mutant compared to WT. Normalized current vs. membrane potential was fit with a modified Boltzmann fit. [Color figure can be viewed at wileyonlinelibrary.com]

Table 8.

Calcium: steady‐state slow inactivation

NaV1.5 channel	[Ca2+]cytosolic	SSSI‐V 1/2 (mV)	SSSI‐z	SSSI‐Y 0	n
Wild‐type	0 nm Ca2+	−70.8 ± 2.28	−1.34 ± 0.07	0.30 ± 0.03	6
	500 nm Ca2+	−71.6 ± 2.48	−1.38 ± 0.15	0.38 ± 0.03	4
	2500 nm Ca2+	−68.2 ± 2.74	−1.20 ± 0.06	0.36 ± 0.03	7
∆KPQ	0 nm Ca2+	−66.2 ± 2.18	−1.50 ± 0.20	0.32 ± 0.03	5
	500 nm Ca2+	−76.2 ± 6.74	−1.67 ± 0.26	0.38 ± 0.05	5
	2500 nm Ca2+	−73.2 ± 2.05	−1.57 ± 0.15	0.48 ± 0.03	6
E1784K	0 nm Ca2+	−73.4 ± 2.41	−1.22 ± 0.10	0.21 ± 0.02	6
	500 nm Ca2+	−79.9 ± 4.72	−1.77 ± 0.13	0.37 ± 0.05	5
	2500 nm Ca2+	−56.7 ± 4.48*1	−1.02 ± 0.06	0.43 ± 0.05*1	5
1795insD	0 nm Ca2+	−84.3 ± 3.22	−1.86 ± 0.15	0.28 ± 0.03	6
	500 nm Ca2+	−69.8 ± 4.91	−1.25 ± 0.26	0.30 ± 0.03	5
	2500 nm Ca2+	−70.2 ± 2.91	−1.01 ± 0.06*2	0.27 ± 0.05	7
Q1909R	0 nm Ca2+	−76.7 ± 2.82	−1.51 ± 0.14	0.20 ± 0.02*3	7
	500 nm Ca2+	−73.1 ± 1.55	−1.40 ± 0.10	0.38 ± 0.02	5
	2500 nm Ca2+	−76.4 ± 2.88	−1.23 ± 0.09	0.35 ± 0.02	6

^*1 P < 0.01 (V _1/2) and P < 0.05 (Y₀) vs. E1784K (0 and 500 nm).

^*2 P < 0.05 vs. 1795insD (0 nm).

^*3 P < 0.05 vs. Q1909R (500 and 2500 nm).

Slow inactivation recovery and onset kinetics

The weighted slow inactivation time constants are plotted as a function of the membrane potential in Fig. 6 A and B. Slow inactivation recovery (α) rates are plotted against voltage in Fig. 6 C–F and slow inactivation onset (β) rates in Fig. 6 G. Mutant effects on τ are reported in Tables 9 and 10. At −130 and −110 mV, recovery kinetics of Q1909R are accelerated (P < 0.0001) (Table 9) when cytosolic calcium was elevated from 0 to 500 nm. Slow inactivation recovery kinetics were also accelerated in E1784K at −110 mV (P < 0.0001) (Table 9). The other channel variants had no significant changes in recovery. At voltages greater than −70 mV, slow inactivation kinetics was not affected by elevations in cytosolic calciumin any channel variant (Tables 9 and 10).

Figure 6. Effects of cytosolic calcium on slow inactivation time constants.

A and B, mutant effects on the weighted slow inactivation time constant at 0 and 2500 nm cytosolic calcium. C–F, calcium effects on the slow inactivation α rate for the mutants compared to WT. G, the bar graph below shows the slow inactivation β rate as a function of calcium and channel variant. Voltage pulse protocols of recovery and onset are not shown for clarity. [Color figure can be viewed at wileyonlinelibrary.com]

Table 9.

Calcium: slow inactivation time constants (−130 mV to −70 mV)

NaV1.5 Channel	[Ca2+]cytosolic	−130 mV (s)	n	−110 mV (s)	n	−70 mV (s)	n
Wild‐type	0 nm Ca2+	0.43 ± 0.06	6	1.22 ± 0.19	5	8.03 ± 1.69	8
	500 nm Ca2+	0.40 ± 0.07	5	1.41 ± 0.12	5	9.25 ± 0.99	5
	2500 nm Ca2+	0.70 ± 0.05	7	2.06 ± 0.15	7	6.32 ± 2.67	5
∆KPQ	0 nm Ca2+	0.97 ± 0.12	7	1.64 ± 0.29	6	9.31 ± 2.56	4
	500 nm Ca2+	0.58 ± 0.04	7	0.98 ± 0.08	6	6.41 ± 0.96	4
	2500 nm Ca2+	0.54 ± 0.07	5	1.25 ± 0.18	6	11.4 ± 1.35	5
E1784K	0 nm Ca2+	1.15 ± 0.27	5	3.50 ± 0.89*1	6	14.3 ± 2.23	5
	500 nm Ca2+	0.95 ± 0.24	6	1.37 ± 0.23	6	18.9 ± 4.31	5
	2500 nm Ca2+	0.60 ± 0.09	8	1.00 ± 0.12	5	16.7 ± 2.22	4
1795insD	0 nm Ca2+	1.30 ± 0.42	3	1.50 ± 0.37	2	11.8 ± 0.69	5
	500 nm Ca2+	1.78 ± 0.28	5	2.92 ± 0.55	4	11.2 ± 0.53	4
	2500 nm Ca2+	0.83 ± 0.17	5	1.40 ± 0.16	5	13.4 ± 0.94	4
Q1909R	0 nm Ca2+	2.49 ± 0.56*1	5	5.72 ± 0.84*1	4	12.6 ± 0.94	5
	500 nm Ca2+	0.68 ± 0.08	6	1.36 ± 0.25	6	6.98 ± 1.29	6
	2500 nm Ca2+	0.68 ± 0.10	5	1.76 ± 0.36	5	5.86 ± 0.38	5

^*1 P < 0.0001 vs. same mutant (500  and 2500 nm).

Table 10.

Calcium: slow inactivation time constants (−50 mV to −10 mV)

NaV1.5 Channel	[Ca2+]cytosolic	−50 mV (s)	n	−30 mV (s)	n	−10 mV (s)	n
Wild‐type	0 nm Ca2+	9.38 ± 1.00	5	6.32 ± 0.27	4	7.01 ± 1.15	5
	500 nm Ca2+	10.9 ± 0.83	5	7.17 ± 0.30	6	5.91 ± 0.65	6
	2500 nm Ca2+	5.62 ± 1.83	5	5.88 ± 1.77	5	5.53 ± 1.19	4
∆KPQ	0 nm Ca2+	9.15 ± 1.90	6	8.48 ± 1.59	5	5.86 ± 1.44	5
	500 nm Ca2+	14.6 ± 1.77	5	10.4 ± 1.39	4	10.7 ± 1.76	5
	2500 nm Ca2+	12.3 ± 1.13	5	12.1 ± 0.56	5	10.0 ± 0.72	5
E1784K	0 nm Ca2+	9.12 ± 0.98	5	7.10 ± 0.39	5	5.27 ± 1.02	4
	500 nm Ca2+	15.5 ± 2.81	5	9.36 ± 1.06	5	9.23 ± 1.07	5
	2500 nm Ca2+	14.5 ± 1.48	5	9.76 ± 1.83	5	10.2 ± 1.58	4
1795insD	0 nm Ca2+	8.05 ± 1.72	6	5.96 ± 1.30	5	6.48 ± 1.35	5
	500 nm Ca2+	7.82 ± 1.93	5	5.56 ± 1.59	5	6.48 ± 1.52	5
	2500 nm Ca2+	8.11 ± 1.34	5	4.62 ± 1.00	5	5.61 ± 1.89	5
Q1909R	0 nm Ca2+	7.17 ± 0.43	5	6.69 ± 0.49	5	8.68 ± 0.76	5
	500 nm Ca2+	7.56 ± 0.88	5	5.17 ± 0.30	5	5.48 ± 0.78	5
	2500 nm Ca2+	7.47 ± 1.35	6	7.02 ± 0.44	5	5.67 ± 0.97	5

Ventricular AP simulations

With increasing pacing frequency stimulations, intracellular calcium naturally increases inside the cell, known as the positive staircase phenomenon. To simulate dynamic properties of the experimentally observed calcium‐induced I _Na shifts, we used a modified version of the ORd model to generate APs in the endocardial and epicardial cells. The I _Na gating parameters were held uniform across all the three cell types. The last AP simulated at each frequency is in Fig. 7 for all the channel variants at bradycardia, normal and tachycardia heart rates. In E1784K, AP is lost with every other beat at elevated heart rates. Thus, both the final two E1784K APs at each frequency are shown in Fig. 7. AP upstroke velocity and APD at 30%, 60% and 90% of repolarization were measured and are shown in Fig. 8.

Figure 7. Ventricular action potential simulations.

AP simulations are shown for endocardial and epicardial cells as a function of frequencies (0.5, 1.5 and 2.5 Hz). [Color figure can be viewed at wileyonlinelibrary.com]

Figure 8. AP velocity and electrical restitution curves (ERC).

A and B, AP velocity as a function of BCL. C, ERC curve in the endocardium. D, difference (∆) in APD₉₀ as a function of frequency. [Color figure can be viewed at wileyonlinelibrary.com]

AP upstroke velocity was substantially reduced in ∆KPQ, E1784K and 1795insD compared to WT and Q1909R (Fig. 8 A and B). With decreasing BCLs, particularly below 600 ms, the Q1909R upstroke velocities are reduced. AP upstroke velocity in the remaining mutants appeared to be insensitive to any BCL shortening.

APDs (APD_X, X = 30,60,90) are plotted vs. the diastolic interval in Fig. 8 C to create electrical restitution curves (ERCs). The plateaus of the ERC curves were increased with increasing APD_X. The greatest change in ERC as function of repolarization time was evident in E1784K compared to other variants, which displays prolonged APD at reduced diastolic intervals. In addition, the ERC slope is greatest in E1784K (APD₉₀), suggesting greater arrhythmogenicity.

The shifts observed in APD as a function of heart rate in all the mutants can be ascribed to the positive‐staircase rise in cytosolic calcium. When the calcium‐dependent gating in I _Na was disabled (results not shown), APD₉₀ was less sensitive to higher pacing frequencies, suggesting a calcium‐mediated effect on AP morphology in the mutants that were studied.

To further investigate the frequency‐dependent shortening of APD, we calculated the difference in APD₉₀ between successive frequencies shown Fig. 8 D. In the endocardial cell, the APD₉₀ shortening displays a relativelyU‐type frequency dependence. The inflection point is evident especially at 1.5 Hz. However, E1784K exacerbates the U‐type frequency dependence by prolonging greatly (Fig. 8 D). This result suggests a frequency‐dependent APD prolongation in E1784K.

Each of the mutants characterized in the present study increases the late I _Na compared to WT. Increases in intracellular sodium, [Na]_i, known as sodium overload, causes the sodium‐calcium exchanger (NCX) to function in reverse mode, inducing a calcium‐overload, which underlies cardiac diastolic dysfunction. The rise in intracellular calcium ([Ca²⁺]_i) during the AP time course is shown Fig. 9. All mutants affect [Ca²⁺]_i similar to WT, except in E1784K, which has relatively suppressed [Ca²⁺]_i during AP plateau and elevated [Ca²⁺]_i during refractory period (Fig. 9 B, inset). The [Na]_i ³:[Ca²⁺]_i ratio was calculated for all channel variants at all frequencies and plotted against the APD₉₀ for the endocardial (Fig. 9 E). The [Na]_i ³:[Ca²⁺]_i ratio decreased in WT, ∆KPQ and Q1909R as a function of frequency. E1784K displays differential sensitivity to frequency‐induced [Na]_i ³:[Ca²⁺]_i shifts. In E1784K, the perturbations in [Na]_i ³:[Ca²⁺]_i follow a helter‐skelter manner, resulting in a sodium overload at bradycardia and near tachycardia frequencies. Analysis of sodium‐calcium levels at frequencies higher than 2.0 Hz in E1784K was difficult as a result of the presence of alternans. Elevated [Na]_i ³:[Ca²⁺]_i were sufficient to induce NCX reverse mode in the presence of alternans in E1784K (results not shown).

Figure 9. Sodium‐calcium overload.

A–D, intracellular calcium in all the channel variants at 0.5, 1.5 and 2.5 Hz simulated in endocardial cells. E, intracellular sodium‐calcium ratio as function of APD₉₀ in the endocardium. [Color figure can be viewed at wileyonlinelibrary.com]

Discussion

Elevated cytosolic calcium, induced by sympathetic stimulation during exercise, may serve as a potential arrhythmogenic trigger in patients with SCN5a mutations. The present study aimed to explore the response of mixed syndrome mutants to elevated cytosolic calcium. The mutants that were investigated have differential cytosolic calciumsensitivities. We found that biophysical defects were either ameliorated or exacerbated by elevated cytosolic calcium in a mutant‐dependent fashion.

Elevated cytosolic calcium clearly attenuates late I _Na in ∆KPQ, 1795insD and Q1909R. Because late I _Na is a key pathophysiological substrate in LQT3, elevated cytosolic rescues the gain‐of‐function phenotype in those Na_V1.5 mutants. By contrast, the therapeutic effect of cytosolic calcium is hampered in E1784K, suggesting that carriers may express LQT3 properties under elevated heart rates. Although our experiments were performed at room temperature, we predict that our results, and their phenotypic sequelae, may be exacerbated at higher temperatures because E1784K is also highly thermosensitive (Abdelsayed et al. 2015). Our AP modelling confirms that the most arrhythmogenic mutant is E1784K and it also predicts a large dispersion of repolarization across the ventricular wall. E1784K was the only mutant that was predicted to cause a [Na⁺]_i‐overload under both bradycardia and tachycardia by our model. The other mutants characterized showed decreases in [Na⁺]_i with elevated heart rates. The model also predicts [Na⁺]_i‐overload, exacerbating E1784K arrhythmogenicity by driving NCX to run in reverse mode, further increasing cytosolic calcium. Increased cytosolic sodium and calcium results in electrical and mechanical cardiac abnormalities (Antzelevitch et al. 2014).

Biophysical implications

The Domain III‐IV linker and C‐terminus are crucial for both fast and slow inactivation in sodium channels. Mutations in these regions perturb their normal interaction both under diastolic calcium levels and upon a calcium signal, causing a defect in Na_V1.5 inactivation. The C‐terminus contains indirect binding sites for calcium (Mori et al. 2003; Biswas et al. 2009), including a pair of EF‐hand domains composed of a helix‐loop‐helix, from Glu1773 to Asp1852 (comprising helices 1–4, H1–H4) (Wingo et al. 2004; Shah et al. 2006). Approximately 120 residues downstream, an IQ domain binds to CaM (Cormier et al. 2002; Tan et al. 2002; Wingo et al. 2004). The influence of calciumon inactivation has been studied using electrophysiology and isothermal titration calorimetry experiments (Shah et al. 2006; Sarhan et al. 2012). Under low calcium conditions, CaM binds via the C‐lobe to the IQ domain of the C‐terminus (Chagot et al. 2009). When cytosolic calcium levels rise, the CaM N‐lobe binds to the IQ domain and the C‐lobe binds to the Domain III‐IV linker, as in the tripartite complex (Shah et al. 2006; Biswas et al. 2009; Gaudioso et al. 2011; Sarhan et al. 2012).

Previous studies report intra C‐terminal interactions, observed by fluorescence resonance energy transfer. The IQ motif (H6) in the C‐terminus interacts with the aromatic residues of H1 (F1791 and Y1795) and the N‐terminal residues of H6 interact with residues in H1‐H2 and H2‐H3 linkers (Glaaser et al. 2006; Chagot et al. 2009). Proximity of the EF‐hand domain to the IQ motif is required for proper inactivation (Glaaser et al. 2012). The intra C‐terminal interaction associates with the Domain III‐IV linker, modulating inactivation (Bankston et al. 2007).

Studies supporting the tripartite complex show that residues 1498–1501 in the Domain III‐IV linker form an α helix that interacts with the C‐lobe of CaM, whereas the rest of the residues within the linker (1502–1522) are disordered (Sarhan et al. 2012). Deletion of residues 1505–1507 in ∆KPQ normally increases late I _Na and stabilizes steady‐state fast inactivation at RMP (Chandra et al. 1998). Stabilized inactivation may occur as a result of a tighter association between the truncated III‐IV linker and its binding site in the inner vestibule of Na_V1.5. Shortening the Domain III‐IV linker in ∆KPQ may increase the affinity for CaM binding upon a calcium signal. This further destabilizes fast inactivation and increases channel availability as is evident in the calcium‐induced depolarized shift in SSFI midpoint.

We predict the charge‐reversal in E1784K disrupts crucial electrostatic interactions linking the IQ motif to the EF‐hand domain, resulting in an abnormally structured C‐terminus. The positive lysine may form electrostatic interactions with downstream negative charges in H1, altering the IQ‐EF‐hand interaction, and resulting in a loosely bound IQ. Consequently, the IQ‐CaM complex may be more mobile and unable to modulate fast inactivation via the Domain III‐IV linker because IQ‐CaM requires anchorage to H1 (Pitt & Lee, 2016). Because the IQ‐EF‐hand interaction is destabilized, calcium affinity might decrease (Shah et al. 2006; Chagot et al. 2009). As a result, elevations in cytosolic calcium may be unable to attenuate increases in late I _Na in E1784K. The fast inactivation mechanism elicited by Domain III‐IV linker is assumed to function in synchrony with the C‐terminus, affecting slow inactivation (Motoike, 2004). A charge‐reversal mutant may uncouple both structures and thus affect the interaction between fast and slow inactivation. Consequently, the increase in E1784K late I _Na with elevations in cytosolic calcium accompanies a destabilized slow inactivation at depolarized potentials (Featherstone et al. 1998; Richmond et al. 1998).

The 1795insD mutant is in the EF‐hand domain (H1) adds a negative charge to the region. Chagot et al. (2009) show that Y1795 is a key residue, establishing hydrophobic interactions with the IQ motif. Adding a negative charge may stabilize the EF‐hand domain and preserve the intactness of the C‐terminal interactions, yielding greater sensitivity by Na_V1.5 to effects of Ca²⁺‐CaM. The Q1909R mutation is in the initial segment of the IQ motif, which is the region that CaM associates with via its N‐terminus during a calcium signal and via its C‐lobe during diastolic calcium. In the IQ motif, Arg3 and Arg6 are bound by hydrogen bonds to different CaM residues. It is thus expected that the Q1909R mutation should enhance CaM binding. Fast inactivation τ_onset in both 1795insD and Q1909R was decelerated at depolarized potentials in a calcium‐dependent fashion, agreeing with past data showing calcium‐induced deceleration of fast inactivation τ_onset in WT Na_V1.5 (Biswas et al. 2009).

Physiological implications

Accentuated gain‐of‐function in E1784K by elevated cytosolic calcium increases [Na⁺]_i‐overload. Our ORd model simulations suggested that prolonged APDs are ameliorated by elevated pacing in ∆KPQ, 1795insD and Q1909R. E1784K APD is relatively prolonged, even during tachycardia. Furthermore, in all mutants except E1784K, the [Na⁺]_i:[Ca²⁺]_i ratio is adjusted or minimized. In E1784K, a discordant curve develops, in which the sodium overload is observed at different heart rates. Sodium overload induces multiple cellular cascades, including the reverse mode in NCX, a factor known to shorten APD at elevated heart rates (Faber & Rudy, 2000; Moreau et al. 2013). Reverse mode NCX causes a rise in intracellular calcium, which threatens diastolic stability, resulting in diastolic dysfunction. In addition to those mechanical abnormalities, calcium overload disturbs the electrical stability of myocardial tissue by inducing DADs, increasing the likelihood of lethal cardiac events including ventricular tachycardia/fibrillation.

Clinical implications

The effect of exercise on LQT3 remains unsettled and controversial. Schwartz et al. (1995) reported a marked reduction in QT_C during exercise in LQT3 patients. This effect is lost during recovery from exercise. Most of the patients in that study carried the ∆KPQ mutation. A reduced QT_C in ∆KPQ may be explained by frequency‐ or calcium‐dependent attenuation of late I _Na. Nevertheless, a relatively minor percentage of LQT3 patients develop cardiac events during exercise. It remain to be determined whether sleep or rest without arousal (i.e. low heart rates) fully explains arrhythmogenesis in LQT3 because the rapid eye movement phase of sleep is accompanied by high sympathetic tone. Thus, both sleep and exercise share some common physiological properties and may share a mechanism of arrhythmogenesis in some LQT3 cases. These commonalities raise questions about the utility of β‐blockers as potential therapeutics for all LQT3 patients, regardless of the causative mutation. The variance seen in BrS1/LQT3 response to exercise depends, at least in part, on the exact SCN5a mutation. Thus, genetic screening is a critical diagnostic tool for determining whether exercise may be therapeutic or an arrhythmogenic trigger in SCN5a patients.

Conclusions

Greater calcium sensitivity in E1784K is reflected by the AP model, in which E1784K prolongs the APD. Compared to the other mutants, E1784K results in the largest transmural voltage gradient. The results of the present study may lead to further refined treatments for BrS1/LQT3. Anti‐arrhythmics, potent for LQT3, should be tested in mixed syndrome cases because the propensity to be phenotypically LQT3 and/or BrS1 may be affected by triggers such as exercise. Cellular cascades, such as activation of the Ca²⁺/CaMKII and other protein kinases, may determine the mutant‐specific biophysical effects of arrhythmogenic mutations (Herren et al. 2013). The present study is probably biased to the effects of calcium on predominantly phosphorylated Na_V1.5 as a result of the use of internal fluoride, which inhibits phosphatases (Proud, 1994). Future assays should measure phosphorylation‐induced shifts on gating and channel expression to further clarify how physiological changes impact the arrhythmogenicity of these mutants. The results of the present study support the need for a more detailed biophysical analysis of all mutants underlying sudden cardiac death.

Additional information

Competing interests

The authors declare that they have no competing interests.

Author contributions

MA collected, assembled, analysed and interpreted the data, designed the experiments and drafted the manuscript. PCR, along with MA, A‐EB, KG, SS, ADK and VP, conceived the experiments and revised the manuscript critically for important intellectual content. All authors approved the final version of the manuscript and qualify for authorship.

Funding

PCR and MA receive funding from Simon Fraser University, the Natural Sciences and Engineering Research Council of Canada, and the Canadian Foundation for Innovation. AK receives support from the Heart and Stroke Foundation of Canada, the Sauder Family and Heart and Stroke Foundation Chair in Cardiology and the Paul Brunes Chair in Heart Rhythm Disorders (G‐13‐0002775 and G‐14‐0005732).