Calcium-Mediated Dual-Mode Regulation of Cardiac Sodium Channel Gating

Abstract

Intracellular Ca²⁺ ([Ca²⁺]_i) can trigger dual-mode regulation of the voltage gated cardiac sodium channel (Na_V1.5). The channel components of the Ca²⁺ regulatory system are the calmodulin (CaM)-binding IQ motif and the Ca²⁺ sensing EF hand–like (EFL) motif in the carboxyl terminus of the channel. Mutations in either motif have been associated with arrhythmogenic changes in expressed Na_V1.5 currents. Increases in [Ca²⁺]_i shift the steady-state inactivation of Na_V1.5 in the depolarizing direction and slow entry into inactivated states. Mutation of the EFL (Na_V1.5_4X) shifts inactivation in the hyperpolarizing direction compared with the wild-type channel and eliminates the Ca²⁺ sensitivity of inactivation gating. Modulation of the steady-state availability of Na_V1.5 by [Ca²⁺]_i is more pronounced after the truncation of the carboxyl terminus proximal to the IQ motif (Na_V1.5_Δ1885), which retains the EFL. Mutating the EFL (Na_V1.5_4X) unmasks CaM-mediated regulation of the kinetics and voltage dependence of inactivation. This latent CaM modulation of inactivation is eliminated by mutation of the IQ motif (Na_V1.5_4X-IQ/AA). The LQT3 EFL mutant channel Na_V1.5_D1790G exhibits Ca²⁺ insensitivity and unmasking of CaM regulation of inactivation gating. The enhanced effect of CaM on Na_V1.5_4X gating is associated with significantly greater fluorescence resonance energy transfer between enhanced cyan fluorescent protein–CaM and Na_V1.5_4X channels than is observed with wild-type Na_V1.5. Unlike other isoforms of the Na channel, the IQ-CaM interaction in the carboxyl terminus of Na_V1.5 is latent under physiological conditions but may become manifest in the presence of disease causing mutations in the CT of Na_V1.5 (particularly in the EFL), contributing to the production of potentially lethal ventricular arrhythmias.

Sodium channels play a central role in the electrophysiology of the heart, being essential for rapid impulse conduction and regulation of the action potential duration. Congenital or acquired defects in inactivation gating of the Na channel have important implications for maintenance of normal rhythmicity in the heart.^1–3 The carboxyl terminus (CT) of Na_V1.5 has been shown to play an important role in inactivation.^4–6 Ca²⁺, CaM, CaM kinase are known to modulate inactivation though interaction with structural motifs in the CT.^6,7 Ca²⁺ binding to an EF hand–like (EFL) sequence in the CT of Na_V1.5 has been shown to regulate channel gating.^8,9 The CT of Na_V1.5 contains a CaM binding motif of the IQ type; however, CaM apparently has minimal functional effect on the expressed current, unlike other voltage gated sodium channels.^6,10 The functional relationship between the EFL and IQ motifs remains uncertain but it has been suggested that a Ca²⁺ load displaces CaM from the IQ motif, freeing the IQ to interact with the EFL stabilizing inactivation.⁸

Mutations near the IQ and EFL motifs (Figure 1A) have been linked to the heritable cardiac rhythm disturbances long QT and Brugada syndromes.^11–13 A number of acquired pathological conditions (cardiac hypertrophy and heart failure) are associated with dysregulation of Ca²⁺ homeostasis, which may generate arrhythmias involving the cardiac Na channel through functional alterations mediated by the IQ and EFL motifs. To better understand the functional relationship between Ca²⁺, CaM, and Ca²⁺ binding directly to the channel via the EFL, we mutated the key Ca²⁺-chelating and IQ residues in the CT of Na_V1.5 and studied the interaction of Ca²⁺ and CaM-based signaling on channel gating.

Figure 1.

Ca²⁺-mediated modulation of gating. A, Schematic of the structured part of the CT of Na_V1.5. The predicted helices are labeled H1 through H6. The EFL is in the loop between H1 and H2 and harbors the 4× mutations (E1788A, D1790A, D1792A, E1799A) or the LQT3 mutation D1790G. The IQ is in H6, and the location of the truncation in Na_V1.5_Δ1885 is shown. B, Representative families of currents (protocol shown in inset) through Na_V1.5 and Na_V1.5_4x channels with Ca²⁺. Ca²⁺ does not alter the voltage of the peak IV relationships (C) or the activation curves (dotted lines in D) in the absence or presence of Ca²⁺. The V_1/2 of steady-state inactivation (solid lines in D) of Na_V1.5 is modestly shifted in the absence of Ca²⁺. The half-maximal inactivation of Na_V1.5_4x is shifted by nearly ≈15 mV in the hyperpolarizing direction compared with Na_V1.5 (Table). Currents were recorded at a test pulse of −20 mV after a 500-ms conditioning pulse from −140 to +30 mV. For Na_V1.5, increasing [Ca²⁺]_i decreases τ_rec (E) and increases τ_entry (F) into inactivated states compared with the Ca²⁺-free condition. Recovery of Na_V1.5_4X is slower than wild-type, and Ca²⁺ further slows recovery. In 10 μmol/L Ca²⁺, the entry of Na_V1.5_4X into inactivated states is faster and more extensive than Na_V1.5. The entry rates converge in the absence of Ca²⁺, but Na_V1.5_4X undergoes more extensive inactivation compared with Na_V1.5. Unless otherwise specified, the data in this and all other figures are in means±SE, and activation and steady-state inactivation data are fit with Boltzmann functions. The symbols are the same in C through F.

Materials and Methods

An expanded Materials and Methods section describing plasmid construction, solutions, and electrophysiology and fluorescence resonance energy transfer (FRET) measurements is available in the online data supplement at http://circres.ahajournals.org. These procedures were similar to those that we have published previously.¹⁰ The bath solution contained (in mmol/L): 145 NaCl, 4 KCl, 1.8 CaCl₂, 1 MgCl₂, 10 glucose, and 10 Na-HEPES (pH 7.4). The patch pipette solution for the Ca²⁺-free condition contained (in mmol/L): 10 NaF, 100 CsF, 20 CsCl₂, 20 BAPTA, and 10 HEPES, pH adjusted to 7.35 with CsOH. The 10 μmol/L Ca²⁺ pipette solution was similar to the Ca²⁺-free solution except that BAPTA was reduced to 1 mmol/L and 1 mmol/L CaCl₂ was added. Similarly, the 0.5 μmol/L Ca²⁺ solution contained 5 mmol/L BAPTA and 4 mmol/L CaCl₂. The osmolarity of the bath and pipette solutions were equilibrated with glucose.

Results

Ca²⁺ Directly Interacts With Na_V1.5

We first examined the effect of altering intracellular Ca²⁺ on gating of wild-type Na_V1.5 and channels with the Ca²⁺-sensing residues of the EFL mutated (Na_V1.5_4X). The pipette solution was either Ca²⁺-free or contained 10 μmol/L Ca²⁺ (Figure 1B). The peak current–voltage (IV) relationships and V_1/2 of the activation curves of Na_V1.5 were unchanged in absence or in presence of 10 μmol/L Ca²⁺ (Figure 1C and 1D and the Table). The voltage dependence of the steady-state availability of wild-type Na_V1.5 is shifted in the depolarizing direction by increases in [Ca²⁺]_i (Figure 1D). To determine whether the observed shift is attributable to direct Ca²⁺ interaction with the channel, we mutated the predicted Ca²⁺-sensing residues in the CT of Na_V1.5 EFL motif (E1788A, D1790A, D1792A, E1799A). Mutations in the EFL eliminate Ca²⁺ sensitivity of inactivation gating of Na_V1.5 and shift the V_1/2 almost ≈15 mV in the hyperpolarizing direction (Figure 1D) without affecting the IV relationship (Figure 1C) and V_1/2 of activation (Figure 1D). However in Ca²⁺-free conditions the reversal potential of Na_V1.5_4X is modestly shifted in the hyperpolarizing direction.

Table.

Effects of Intracellular Ca²⁺, CaM, and CaM₁₂₃₄ on Na_V1.5 Variant Channels

		Activation (mV)
Channel/Mutant	V1/2 Steady State Inactivation (mV)	V1/2	IPeak	EReversal	τRec (ms)	τEntry (ms)
NaV1.5
+Ca2+	−82.6±0.2 (12)	−39.7±1.7	−22.5±2	60±1.1 (10)	3.9±0.1 (10)	4523±184 (5)
−Ca2+	−87.5±0.2 (14)	−38±2	−23±3	45±3 (10)	5.5±0.1 (9)	2783±109 (5)
NaV1.54X
+Ca2+	−98.6±0.4 (15)	−41±2	−23.7±1.5	56±2 (11)	14.5±1.2 (7)	2594±209 (5)
−Ca2+	−99.3±0.1 (11)	−39±2.5	−21±3.5	46±3.5 (9)	21.3±3 (8)	3061±16 (5)
NaV1.5Δ1885
+Ca2+	−81.8±0.1 (10)	−36.8±1.7	−21.6±1.8	55.5±1.5 (9)	2.7±0.1 (5)	6173±1037 (5)
−Ca2+	−93.3±0.1 (18)	−34.1±0.8	−18.3±0.7	50±2.6 (18)	5.3±0.1 (9)	2318±231 (7)
NaV1.5IQ/AA
+Ca2+	−73.9±0.1 (7)	−31.9±1.3	−20±1.6	64±2 (5)	3.7±0.1 (6)	5416±568 (5)
−Ca2+	−82.5±0.4 (8)	−32.7±2.6	−21.3±1.3	49±5.2 (8)	7.8±1.3 (7)	2807±90 (6)
NaV1.5D1790G
+Ca2+	−98.2±0.1 (8)	−38.9±0.4	−20.8±1	56.7±1.8 (11)	5.4±0.07 (5)	3195±150 (6)
−Ca2+	−99.9±0.1 (9)	−37.6±0.4	−21.4±1	53.8±1.2 (13)	5.1±0.08 (6)	3051±144 (6)
NaV1.5 (Ca2+-Free)
−CaM	−87.5±0.2 (14)	−38±2	−23±3	45±3 (10)	5.5±0.1 (9)	2783±109 (5)
+CaM	−82.3±0.6 (6)	−35.1±1.1	−21.7±1.1	53±5 (6)	6±0.9 (5)	4309±89 (5)
+CaM1234	−77.4±0.2 (7)	−35±2	−20±2.5	58±5.3 (8)	7±2.3 (5)	3452±369 (6)
NaV1.5Δ1885 (Ca2+-Free)
−CaM	−93.3±0.1 (18)	−34.1±0.8	−18.3±0.7	50±2.6 (18)	5.3±0.1 (9)	2318±231 (7)
+CaM	−90.4±0.3 (8)	−32.4±2.7	−18±1.5	51±3.6 (7)	5.3±0.8 (7)	2994±315 (6)
+CaM1234	−89.3±0.2 (6)	−28.2±0.8	−18±1.3	62±3 (5)	4±0.7 (5)	3112±221 (5)
NaV1.5 (0.5 μmol/L Ca2+)
−CaM	−82.6±0.1 (5)	−36±2	−20.7±0.7	55±2.2 (7)	3.8±0.4 (5)	3413±470 (5)
+CaM	−80±0.08 (12)	−34.6±0.3	−20±2.2	53±2.2 (7)	3.3±0.6 (5)	3659±46 (5)
+CaM1234	−75.4±0.04 (14)	−37.8±2.7	−22.9±20.6	45±4.3 (7)	2.3±0.4 (5	3291±203 (5)
NaV1.54X (0.5 μmol/L Ca2+)
−CaM	−99.9±0.1 (6)	−37.4±1.7	−20.8±1.5	54.2±4 (6)	14.9±0.2 (7)	1704±195 (5)
+CaM	−106.8±0.1 (8)	−36±4.3	−21±4.9	54±6.2 (5)	15.5±0.3 (5)	1745±120 (5)
+CaM1234	−94.9±0.2 (15)	−35±1.1	−18.7±1.5	55±3.6 (11)	15.3±0.2 (8)	1734±142 (7)
NaV1.54X+AIP (0.5 μmol/L Ca2+)
−CaM	−101.6±0.5 (5)	−34.4±1.3	−18.8±2.4	52.5±6 (5)	15.8±0.3 (5)	1685±133 (5)
+CaM	−102±0.2 (11)	−34±0.7	−18±1.5	56.3±1.3 (5)	14.8±0.2 (5)	1509±150 (5)
+CaM1234	−103.7±0.1 (15)	−33±3.1	−20±3.1	52±2.8 (10)	18.1±0.3 (9)	1587±142 (7)
NaV1.54X-IQ/AA (0.5 μmol/L Ca2+)
−CaM	−100.8±0.3 (9)	−32.3±2	−18.4±1.7	55.8±3 (6)	18.2±0.5 (6)	1951±111 (5)
+CaM	−100.3±0.1 (10)	−32±3.9	−18±2.6	52±3 (5)	14.3±0.2 (5)	1942±129 (5)
+CaM1234	−103.1±0.2 (5)	−34±3	−17±1.2	52±5.9 (5)	15.8±0.2 (5)	2159±196 (5)
NaV1.5D1790G (0.5 μmol/L Ca2+)
−CaM	−96.5±0.1 (6)	−38.9±0.5	−22.5±1.3	62.5±1.3 (6)	3.5±0.3 (5)	3565±202 (6)
+CaM	−102±0.2 (6)	−39.7±0.4	−20.7±2	62.1±2.4 (6)	4.9±0.5 (5)	3122±198 (5)
+CaM1234	−92.3±0.2 (7)	−37.9±0.4	−21.5±1.5	68.5±2.6 (6)	4.5±0.3 (7)	3379±109 (7)

Values are means±SE (n). +Ca²⁺ indicates 10 μmol/L Ca²⁺; −Ca²⁺, Ca²⁺-free; −CaM, without CaM/CaM₁₂₃₄ overexpression.

Alterations in [Ca²⁺]_i affect both entry into and recovery from inactivated states. Increasing the [Ca²⁺]_i significantly shortens the fast recovery time constant (τ_rec) from 5.5±0.1 to 3.9±0.1 ms in wild-type Na_V1.5 (Figure 1E). However, recovery from slower inactivated states after a prepulse of 1000 ms is not affected by the [Ca²⁺]_i (Figure I in the online data supplement). The EFL mutant, Na_V1.5_4X exhibits slower recovery independent of the [Ca²⁺]_i with a fast τ_rec that is ≈3- to 4-fold larger than wild-type Na_V1.5. Reducing the [Ca²⁺]_i does not affect the current available immediately after the conditioning pulse (P₁) but significantly shortens the τ_rec of Na_V1.5_4X (Figure 1E). Conversely, increasing [Ca²⁺]_i significantly slows entry into inactivation of wild-type Na_V1.5 with τ_entry increasing to 4523±184 ms in the presence of 10 μmol/L Ca²⁺ from 2783±109 ms in Ca²⁺-free conditions (Figure 1F). Na_V1.5_4X hastens the rate of entry into inactivated states compared to wild-type Na_V1.5 in 10 μmol/L Ca²⁺ and is not associated with any Ca²⁺-dependent changes in entry into inactivated states (Figure 1F). Changes in [Ca²⁺]_i have no significant effect on the persistent current through Na_V1.5 or Na_V1.5_4X channels but in the absence of Ca²⁺, the persistent current through Na_V1.5_4X is greater than Na_V1.5 (supplemental Figure II).

IQ-CaM Interaction Does Not Modulate Ca²⁺-Dependent Channel Kinetics

The role of the Ca²⁺-binding protein, calmodulin (CaM) in mediating Ca²⁺-dependent regulation of cardiac Na channel gating remains controversial. On the one hand, CaM is a ubiquitous Ca²⁺ buffer; however, CaM or CAM without bound calcium (apo)-CaM directly binds to the IQ motif in the CT of Na_V1.5; thus, the IQ motif binds an extrinsic Ca²⁺ sensor (CaM) that regulates channel function. We deleted the IQ motif by truncation of the channel after amino acid1885 (Na_V1.5_Δ1885) and assessed the sensitivity of the truncation mutant to the [Ca²⁺]_i level (Figure 2A). Similar to wild-type Na_V1.5and Na_V1.5_4X the IV relationship and voltage dependence of activation of Na_V1.5_Δ1885 are not altered by changes in [Ca²⁺]_i (Figure 2A and 2B). Previous reports suggested a significant increase in late current with the Δ1885 truncation and the IQ/AA mutation^5,14; however, a change in [Ca²⁺]_i had no effect on the magnitude of the late current through IQ/AA mutation.⁷ The persistent current through Na_V1.5_Δ1885 was not influenced by [Ca²⁺]_i, although the magnitude of the persistent current was greater than that of wild-type Na_V1.5 (supplemental Figure II). However, the voltage dependence of steady-state inactivation of Na_V1.5_Δ1885 remained sensitive to [Ca²⁺]_i. In the presence of 10 μmol/L Ca²⁺, the V_1/2 of steady-state inactivation is −81.8±0.1mV, similar to wild-type Na_V1.5, removal of Ca²⁺ shifts the V_1/2 of Na_V1.5_Δ1885 approximately −11 mV (Figure 2C). The kinetics of inactivation gating of Na_V1.5_Δ1885 exhibited changes comparable to wild-type Na_V1.5 in response to changes in [Ca²⁺]_i. Increasing the [Ca²⁺]_i significantly shortens τ_rec of Na_V1.5_Δ1885 from 5.3±0.1 ms in Ca²⁺-free conditions to 2.7±0.1 ms (Figure 2D). In contrast, 10 μmol/L Ca²⁺ significantly increases the τ_entry of Na_V1.5_Δ1885 compared to the Ca²⁺-free condition, an effect similar to wild-type Na_V1.5 (Figure 2E). Point mutations in IQ region (Na_V1.5_IQ/AA) of the CT of Na_V1.5 that disable CaM binding exhibit comparable Ca²⁺-induced changes in inactivation gating (Figure 3 and Table). The preservation of the Ca²⁺ sensitivity of inactivation by the deletion mutant Na_V1.5_Δ1885 and Na_V1.5_IQ/AA suggests that the IQ motif is not required for Ca²⁺ dependent modulation of inactivation voltage dependence or kinetics.

Figure 2.

Ca²⁺ regulation of truncated Na_V1.5 channels. A, Representative family of Na_V1.5_Δ1885 currents in absence and presence of 10 μmol/L Ca²⁺. The normalized IV relationships (B) and the V_1/2 of activation (dotted lines in C) are not affected by changes in [Ca²⁺]_i. The steady-state inactivation is shifted in the hyperpolarizing direction in the absence of [Ca²⁺]_i (C). The τ_rec (D) is shorter and τ_entry (E) is longer in presence of Ca²⁺ compared to the Ca²⁺-free condition. Na_V1.5 in 10 μmol/L Ca²⁺ (Figure 1C) is shown for comparison. The symbols are the same in B through E.

Figure 3.

IQ/AA retains Ca²⁺ sensitivity of inactivation gating. A, Families of representative Na_V1.5_IQ/AA currents with and without Ca²⁺. The IV relationship (B) and the kinetics of activation gating of Na_V1.5_IQ/AA (dotted lines C) of Na_V1.5_IQ/AA are not altered by changes in [Ca²⁺]_i. The voltage dependence of steady-state inactivation of Na_V1.5_IQ/AA remains sensitive to [Ca²⁺]_i. τ_rec (D) and τ_entry (E) respond similarly to Na_V1.5 and Na_V1.5_Δ1885 with changes in [Ca²⁺]_i. The symbols are the same in B through E.

Unmasking of Modulation of Inactivation by CaM

The IQ motif is not required for Ca²⁺ modulation of Na_V1.5; however, the existence of direct IQ-CaM regulation of gating of other Na channel isoforms^6,10 with homologous carboxyl termini has led to the speculation that there could be latent tuning of normal channel function by CaM binding to the IQ motif in Na_V1.5. To test the idea of latent CaM-mediated modulation of Na_V1.5 gating, we examined the effects of CaM overexpression in the absence of Ca²⁺ on wild-type channel function. Coexpression of CaM with wild-type Na_V1.5 significantly shifts channel availability in the depolarizing direction compared to Na_V1.5 without CaM coexpression in the Ca²⁺-free condition (Figure 4A). Interestingly mutant CaM₁₂₃₄ also shifts the steady-state V_1/2 of the inactivation curve in the Ca²⁺-free condition (Figure 4A). Neither CaM nor CaM₁₂₃₄ affects the activation curve, recovery from or entry into inactivation of Na_V1.5 (Table). However, in physiological [Ca²⁺]_i (0.5 μmol/L), the effect of CaM overexpression on Na_V1.5 inactivation gating is not significantly different from that observed in the absence of CaM overexpression. Notably, overexpression of CaM₁₂₃₄ produces a significant depolarizing shift in inactivation (Table and Figure 4A and 4B), consistent with retained binding to the CT and modulation of gating. To verify the involvement of the IQ motif in latent CaM-mediated regulation of the cardiac Na channel, we studied the effect of CaM overexpression on inactivation of the IQ-deficient truncation mutant Na_V1.5_Δ1885 in Ca²⁺-free conditions. The steady-state inactivation curve of Na_V1.5_Δ1885 is not affected by overexpression of CaM or CaM₁₂₃₄ compared to the channel alone and neither CaM nor CaM₁₂₃₄ effects activation of Na_V1.5_Δ1885 (Table). The data suggest that CaM/CaM₁₂₃₄-mediated modulation of Na_V1.5 in Ca²⁺-free conditions is through binding to the IQ motif.

Figure 4.

CaM modulation of gating. A, The steady-state inactivation curves of Na_V1.5 in Ca²⁺-free conditions are shifted in the depolarizing direction with overexpression of both CaM or CaM₁₂₃₄ compared with Na_V1.5 expressed alone. B, In 0.5 μmol/L Ca²⁺, only CaM₁₂₃₄ overexpression significantly shifts the steady-state inactivation curve. Neither CaM nor CaM₁₂₃₄ overexpression affects the current through Na_V1.5_4X (C) or IV relationships (D) in 0.5 μmol/L Ca²⁺. E, CaM and CaM₁₂₃₄ shifts the steady-state inactivation curves in the hyperpolarizing and depolarizing direction, respectively, compared to Na_V1.5_4X expressed alone. In contrast, CaM or CaM₁₂₃₄ overexpression does not alter recovery from inactivation τ_rec (F) or entry into inactivation τ_entry (G). The symbols are the same in D through G.

We then tested whether CaM modulation of Na_V1.5 is unmasked in the EFL mutant channel in physiological [Ca²⁺]_i. Neither CaM nor CaM₁₂₃₄ overexpression alters the Na current or IV relationship (Figure 4C and 4D). In contrast to wild-type Na_V1.5, coexpression of CaM with Na_V1.5_4X significantly shifts channel availability in the hyperpolarizing direction (≈7 mV with CaM) compared to Na_V1.5_4X expressed alone in 0.5 μmol/L [Ca²⁺]_i (Figure 4E). Conversely, coexpression of apo-CaM (CaM₁₂₃₄) significantly shifts the V_1/2 in the depolarizing direction (≈5 mV with CaM₁₂₃₄; Figure 4E). Compared to Na_V1.5_4X alone, recovery from (Figure 4F), or entry into, inactivation (Figure 4G) of Na_V1.5_4X was not significantly affected by CaM/CaM₁₂₃₄ overexpression. Thus, elimination of Ca²⁺ sensing by the EFL mutation unmasks a CaM-mediated effect on the voltage dependence of steady-state inactivation.

There are 2 lines of evidence that suggest that the CaM or apo-CaM effects on Na_V1.5_4X occur through an IQ-CaM interaction. First, substitution of the IQ motif (IQ/AA) in Na_V1.5_4X (Na_V1.5_4X-IQ/AA) abolishes the CaM/CaM₁₂₃₄ shift of the inactivation curve without altering the Na current or the IV relationship (Figure 5A through 5C). The kinetics of recovery from or entry into inactivation of Na_V1.5_4x-IQ/AA remains unaffected by CaM or CaM₁₂₃₄ (Table). Thus, Na_V1.5_4X-IQ/AA loses its sensitivity to coexpressed CaM and CaM₁₂₃₄.

Figure 5.

IQ motif mediates the CaM regulation of Na_V1.5_4X. A, Representative current families through Na_V1.5_4X-IQ/AA in the presence of CaM and CaM₁₂₃₄ overexpression. Neither CaM nor CaM₁₂₃₄ overexpression alters the IV relationship (B) or steady-state inactivation curves (C) of Na_V1.5_4X-IQ/AA. D, Representative raw current records of Na_V1.5_4X with overexpression of CaM or CaM₁₂₃₄ and intracellular application of AIP290-309. The IV relationship (E) and steady-state inactivation (F) are not affected by CaM or CaM₁₂₃₄ in the presence of AIP290-309. The symbols are the same in B and C and in E and F.

The second line of evidence that the IQ motif is involved in CaM/CaM₁₂₃₄ regulation of Na_V1.5_4x is the effect of the CaM anti-peptide AIP290–309 on Na_V1.5_4x gating. In the presence of AIP290-309, overexpression of CaM or CaM₁₂₃₄ does not affect the Na current, IV relationship (Figure 5D and 5E), or V_1/2 of steady-state inactivation of Na_V1.5_4x (Figure 5F). Similarly, recovery from and entry into inactivation of Na_V1.5_4x are unaffected by CaM or CaM₁₂₃₄ in the presence of AIP290-309 (Table). These data suggest that AIP290-309 competes for CaM binding; displaces CaM from the channel, preventing CaM-induced modulation of the voltage dependence of steady-state inactivation; or binds to an unknown alternate binding site(s) to counteract the effect of CaM on inactivation. The AIP290-309 data and the effect of the IQ/AA mutation in Na_V1.5_4x are consistent with the IQ motif mediating latent CaM/CaM₁₂₃₄ modulation of the cardiac Na channel.

Physical Association of CaM With Na_V1.5 and Na_V1.5_4X

The electrophysiological data suggest that CaM/CaM₁₂₃₄ interacts with the EFL mutant channel Na_V1.5_4X and produces a shift in the voltage dependence of gating. Wild-type Na_V1.5 channels remain insensitive to CaM in physiological (0.5 μmol/L) [Ca²⁺]_i (Table and Figure 4B), but steady-state inactivation is regulated by CaM and CaM₁₂₃₄ in the Ca²⁺-free condition (Figure 4A). In the skeletal muscle channel, Na_V1.4, CaM is tethered to the channel CT and modulates inactivation gating through the IQ motif.¹⁰ The CT of Na_V1.4 and Na_V1.5 are highly homologous, both containing EFL and IQ motifs. We hypothesize that CaM interacts with Na_V1.5, but the functional effects of the interaction on inactivation gating are concealed at physiological [Ca²⁺]_i and that the EFL mutant channel, Na_V1.5_4X, reveals latent regulation of inactivation by CaM binding to the CT of Na_V1.5 (Figure 4A and 4E). To test this hypothesis, we looked for direct interaction of CaM with the CT of Na_V1.5 and Na_V1.5_4X by coexpressing Na_V1.5–enhanced yellow fluorescent protein (EYFP) or Na_V1.5_4X-EYFP with ECFP-CaM, ECFP-CaM₁₂₃₄, or ECFP alone. We used donor dequenching FRET in live cells to study the proximity of ECFP-CaM to the CT of Na_V1.5 or Na_V1.5_4X. The channel fusions were designed with the fluorescent proteins linked to the carboxyl terminus, which is close in the linear amino acid sequence to the IQ motif, optimizing the possibility of FRET detection on CaM binding. Both the Na_V1.5 and Na_V1.5_4X fusion constructs expressed in HEK293 cells display a distinct fluorescent enrichment at the cell perimeter consistent with surface membrane expression in contrast to the uniform cytosolic expression of ECFP-CaM and ECFP-CaM₁₂₃₄ (Figure 6A).

Figure 6.

Physical association of CaM and CaM₁₂₃₄ with Na_V1.5 variants. A, Epifluorescence image of Na_V1.5_4X-EYFP showing expression of the tagged channel protein (top left panel) and the bright field image of the same cell (bottom right panel). The line scan intensity plot shows channel expression in the cell interior as well as at the perimeter indicating membrane targeting (bottom left panel). To avoid contamination of the FRET signal from cytoplasmic EYFP, we measured FRET by focusing on a region of interest (ROI) only at the membrane (see the online data supplement). In contrast to channel expression, ECFP-CaM is expressed more homogenously in the cytosol (top right). Only cells expressing both fluorophores were chosen for donor dequenching FRET. B, Donor dequenching FRET in HEK293 cells expressing the indicated constructs. The Na_V1.5_4X-EYFP coexpressed with ECFP-CaM or ECFP-CaM₁₂₃₄ yielded significantly more FRET compared to Na_V1.5.

The FRET efficiency (percentage) using donor dequenching was computed as previously described.¹⁰ A mean FRET efficiency of 7.2±0.4% (Figure 6B) was observed when ECFP-CaM was coexpressed with EYFP-tagged wild-type Na_V1.5 channels, indicating that the fluorophores in Na_V1.5-EYFP and ECFP-CaM were separated by less than 100 Å. Mutant CaM₁₂₃₄ also associated with the resting channel and exhibited a mean FRET efficiency 5.6±0.5% (Figure 6B) when ECFP-CaM₁₂₃₄ was coexpressed with EYFP-tagged Na_V1.5 channels. These data support a physical association between CaM/CaM₁₂₃₄ and Na_V1.5, with no apparent effects of CaM overexpression on channel gating in physiological [Ca²⁺]_i.

We then assessed the direct CaM interaction with Na_V1.5_4X channels to study the role of the IQ motif in CaM-mediated regulation of gating of the EFL mutant channel (Figure 4E). The Na_V1.5_4X-EYFP interaction with ECFP-CaM and ECFP-CaM₁₂₃₄ assessed by donor dequenching FRET showed enhanced mean FRET efficiency between Na_V1.5_4X-EYFP and both donors, ECFP-CaM (18.1±1.5%), and ECFP-CaM₁₂₃₄ (18.0±1.3%; Figure 6B). Coexpression of free cytosolic ECFP with Na_V1.5-EYFP or Na_V1.5_4X-EYFP and, co expression of ECFP, ECFP-CaM or ECFP-CaM₁₂₃₄ with mutant Na_V1.5_4XIQ/AA-EYFP does not yield FRET (Figure 6B). The absence of FRET between the Na_V1.5or Na_V1.5_4X and free cytosolic ECFP, excludes possibility of FRET because of generalized enrichment of CaM/CaM₁₂₃₄ at the cell membrane, and nonspecific binding of CaM to the channel. In addition, the absence of FRET between the EFL and IQ compound mutant Na_V1.5_4X-IQ/AA indicates a specific IQ-CaM/CaM₁₂₃₄ interaction in the CT of Na_V1.5. Thus, the enhanced interaction between Na_V1.5_4X channels and CaM/CaM₁₂₃₄ compared to Na_V1.5 channels is consistent with unmasking of the latent regulation of Na_V1.5_4X by CaM/CaM₁₂₃₄.

Gating Defects in an LQT3 EFL Mutant Channel

A clinically relevant defect in Ca²⁺-mediated regulation of Na_V1.5 was explored by studying an LQT3 mutation in the EFL, D1790G.¹⁵ Na_V1.5_D1790G expresses robust Na-selective current; the IV relationship (Figure 7A and 7B) and activation voltage dependence are insensitive to changes in [Ca²⁺]_i (Figure 7C). Similar to Na_V1.5 _4x, Na_V1.5_D1790G significantly shifts V_1/2 of inactivation by ≈15 mV in the hyperpolarizing direction compared to Na_V1.5, an effect that was unperturbed by changes in [Ca²⁺]_i (Figure 7C). Changes in [Ca²⁺]_i do not alter recovery from (Figure 7D) or entry into inactivation (Figure 7E and Table). However, D1790G significantly enhances entry into inactivation compared to wild-type channel in high [Ca²⁺]_i (Figure 7E and Table).

Figure 7.

Ca²⁺/CaM regulation of the LQT3 mutant Na_V1.5_D1790G. A, Representative families of Na_V1.5_D1790G currents in presence and absence of intracellular Ca²⁺. Changes in [Ca²⁺]_i do not alter the normalized IV relationships (B), the activation curves (dotted lines in panel C) or the steady-state inactivation (solid lines in C) of Na_V1.5_D1790G. Similar to Na_V1.5_4X, the V_1/2 of steady-state inactivation of Na_V1.5_D1790G is significantly shifted in the hyperpolarizing direction compared to Na_V1.5. Increasing [Ca²⁺]_i does not alter τ_rec (D) or τ_entry (E) into inactivated states compared with the Ca²⁺-free condition. In 0.5 μmol/L [Ca²⁺]_i, neither CaM nor CaM₁₂₃₄ overexpression alters currents (F) or the IV relationship (G) of Na_V1.5_D1790G. H, CaM significantly shifts V_1/2 of inactivation in hyperpolarizing direction, whereas CaM₁₂₃₄ shifts V_1/2 in depolarizing direction compared with Na_V1.5_D1790G expressed alone. The symbols are the same in panels B–E and, panels G and H.

We tested Na_V1.5_D1790G for CaM/CaM₁₂₃₄ regulation. Similar to Na_V1.5_4X, the current kinetics and the IV relationship were not affected by overexpression of CaM or CaM₁₂₃₄ (Figure 7F and 7G), but CaM significantly shifts the channel availability in the hyperpolarizing direction in physiological 0.5 μmol/L [Ca²⁺]_i (Figure 7H). Conversely, coexpression of CaM₁₂₃₄ significantly shifts the V_1/2 in the depolarizing direction (Figure 7H), and neither CaM nor CaM₁₂₃₄ alters the recovery from or entry into inactivation (Table) of Na_V1.5_D1790G. Changes in [Ca²⁺]_i have no significant effect on the persistent current of Na_V1.5_D1790G channel (supplemental Figure II). This single point mutation in the EFL eliminates Ca²⁺ sensitivity of the mutant channel and unmasks a CaM-mediated effect on the voltage dependence of steady-state inactivation similar to Na_V1.5_4X.

Discussion

Our study demonstrates the capacity for regulation of cardiac Na channels by the Ca²⁺ regulatory protein CaM, through an IQ motif-CaM interaction at physiological [Ca²⁺]_i. Surprisingly, low [Ca²⁺]_i or mutations in a predicted EFL, upstream of the IQ motif, unmask latent regulation of the Na_V1.5 channel by CaM. Furthermore, Na_V1.5 is directly and manifestly influenced by [Ca²⁺]_i through the EFL independent of the CaM-IQ interaction. Thus, 2 processes of Ca²⁺-mediated regulation of Na_V1.5 gating exist⁸ and could be interdependent, particularly when EFL-mediated Ca²⁺ sensing is compromised by mutations (eg, Na_V1.5_D1790G or Na_V1.5_4X) or extremely low local [Ca²⁺]_i.

Latent CaM Regulation of Na_v1.5 Is Through the IQ Motif

The possibility of latent regulation of Na_V1.5 is based on the physical association of CaM/CaM₁₂₃₄ with the wild-type channel CT in vitro⁶ and in the intact channel in live cells (Figure 6B). Mutations in the EFL shift inactivation in the hyperpolarizing direction (Figure 1D and 7C) simulating low [Ca²⁺]_i, which is in part counteracted by CaM and CaM₁₂₃₄ in the Ca²⁺-free solution (Figure 4A) and CaM₁₂₃₄ in 0.5 μmol/L Ca²⁺ (Figure 4B). These data are consistent with an interaction between the EFL and IQ motifs in the CT of Na_V1.5. The EFL and IQ motifs are at the N- and C-terminal ends of the proximal structured portion of the CT of Na_V1.5, respectively. In the presence of CaM and Ca²⁺, it is easy to imagine either competitive or synergistic regulation of the channel by these 2 motifs. The existence of manifest regulation of other isoforms of voltage-gated Na channels¹⁰ by CaM binding to the IQ suggests the possibility of latent tuning of Na_V1.5 channel function by CaM. Our FRET data conclusively demonstrate tethering of CaM to the CT of intact Na_V1.5 channels in live cells (Figure 6B) without any apparent effect on channel function at physiological [Ca²⁺]_i, which we interpret as a latent CaM regulation of the wild-type channel. The FRET signal between CaM/CaM₁₂₃₄ and Na_V1.5 is more pronounced in the EFL mutant channel (Figure 6B) and is eliminated by the IQ/AA mutation, Na_V1.5_4X-IQ/AA, further supporting the role of the IQ motif in CaM binding and latent and manifest modulation of Na_V1.5 inactivation gating. The CaM anti-peptide AIP290–304 abolishes CaM/CaM₁₂₃₄ regulation of inactivation, consistent with CaM or apo-CaM acting on Na_V1.5_4X by binding to the IQ motif.

Remarkably increasing [Ca²⁺]_i shifts the inactivation V_1/2 of Na_V1.5 in the depolarizing direction, opposite to the effect of CaM on the voltage dependence of Na_V1.5_4X inactivation gating. In addition, mutant CaM₁₂₃₄, which does not bind Ca²⁺, shifts the Na_V1.5_4X inactivation V_1/2 in depolarizing direction in contrast to hyperpolarizing shift of the V_1/2 of Na_V1.5 observed in the Ca²⁺-free condition. It is possible that CaM₁₂₃₄ is displacing endogenous CaM from the IQ in Na_V1.5_4X or binding to an alternate binding site(s) accounting for the depolarizing shift in inactivation. The CaM₁₂₃₄-mediated depolarizing shift of Na_V1.5_4X is consistent with the depolarizing shift in inactivation mediated by CaM in Na_V1.5 channels in the absence of Ca²⁺, where CaM may be predominantly in its apo form, similar to CaM₁₂₃₄. In any case, these data confirm the importance of the IQ motif for CaM binding, which could be an evolutionarily conserved process, with concealed modulation of some isoforms and manifest modulation of others (eg, Na_V1.4).

Previous studies have suggested that a conformational change in the CT is associated with an interaction involving the III–IV linker that stabilizes the inactivated state of the channel and promotes CaM unbinding from the IQ.^4,8 Our current data on Na_V1.5 and previous work with Na_V1.4¹⁰ suggest that an interaction between Na_V1 channels and CaM is independent of voltage and possibly channel conformation. This suggests that stabilization of inactivation by the interaction of the CT and III–IV is not associated with complete dissociation of CaM from the CT, although we cannot rule out unbinding of a lobe of CaM from the IQ during inactivation.

IQ-Independent Ca²⁺ Modulation of Na_V1.5 Channel Kinetics

The amino acid sequences of the CT of all Na_V isoforms, especially the IQ motif regions, are highly homologous, yet regulation of the voltage dependence and kinetics of channel gating mediated through IQ-CaM interactions differ among isoforms.^6,7,10,16,17 Na_V1.5 channels may be regulated by Ca²⁺ in 2 ways, acting through the CT of the channel: by direct interaction with the EFL domain and indirectly by binding to CaM and then the IQ domain. Our data show that mutation (Na_V1.5_IQ/AA) or deletion of the IQ motif of Na_V1.5 (Na_V1.5_Δ1885) does not eliminate the regulation of inactivation gating by intracellular Ca²⁺. Na_V1.5_Δ1885 channels exhibited a greater rightward shift of the steady-state inactivation curve in the presence of Ca²⁺ than Na_V1.5 channels; thus, the deletion mutant without the IQ domain remains sensitive to [Ca²⁺]_i. The IQ motif does influence Ca²⁺-mediated channel modulation, but its absence does not render the channel insensitive to [Ca²⁺]_i, indicating EFL-mediated regulation of Na_V1.5 by Ca²⁺ may be independent of IQ-mediated regulation of the channel (Figures 2C and 3C).

Structural Implications of EFL-IQ–Mediated Inactivation

Inactivation of cardiac Na channels is thought to involve an interaction of the III–IV linker loop and CT.⁴ Our data suggest removal of Ca²⁺ exposes negatively charged acidic residues of EFL, leading to conformational modification and disruption of allosteric coupling with other components of the inactivation gating mechanism (eg, III–IV linker loop or inactivation gate receptors formed by IIIS4-S5, IVS4-S5, and IVS6),⁴ leading to a stabilization of inactivation reflected in the negative shift of fast inactivation of Na_V1.5 at low [Ca²⁺]_i (Figure 1D). However, stabilization of channel closure during inactivation is compromised, resulting in an increase in persistent current and may be the result of a change in the interaction with the III–IV linker induced by mutations in EFL. Similar gating alterations are observed with the LQT3 point mutation D1790G, consistent with a congruous disruption of the EFL structure altering the Ca²⁺-sensing capacity of the EFL loop.¹⁸

The EFL and IQ motifs are not immediately contiguous in the linear amino acid sequence but after folding of the CT are predicted to be adjacent.¹⁴ In our study, although Na_V1.5_4X is known to disable Ca²⁺ sensing by the EFL, it is not known whether or not this mutation alters the interaction between the IQ and the EFL. It is possible that CaM access to the IQ domain is enhanced in Na_V1.5_4X compared with wild-type channels, consistent with the enhanced FRET in the mutant (Figure 6B). However, this is speculation and our experiments do not address the structural details of the IQ-EFL interaction.

Conclusion

Ca²⁺ interaction with the EFL motif plays a critical role in controlling Na_V1.5 availability and Ca²⁺ regulation of gating is independent of the CaM-IQ interaction. Disease-associated mutations in the EFL region of Na_V1.5 may be associated with an exaggerated, possibly arrhythmogenic, CaM-induced shift in gating; however, in the intact cardiac Na channel, the concerted action of the 2 motifs finely tune channel inactivation.