Biophysical Investigation of Sodium Channel Interaction with β-Subunit Variants Associated with Arrhythmias

Abstract

Background:

Voltage-gated sodium (Na_V) channels help regulate electrical activity of the plasma membrane. Mutations in associated subunits can result in pathological outcomes. Here we examined the interaction of Na_V channels with cardiac arrhythmia-linked mutations in SCN2B and SCN4B, two genes that encode auxiliary β-subunits.

Materials and Methods:

To investigate changes in SCN2B^R137H and SCN4B^I80T function, we combined three-dimensional X-ray crystallography with electrophysiological measurements on Na_V1.5, the dominant subtype in the heart.

Results:

SCN4B^I80T alters channel activity, whereas SCN2B^R137H does not have an apparent effect. Structurally, the SCN4B^I80T perturbation alters hydrophobic packing of the subunit with major structural changes and causes a thermal destabilization of the folding. In contrast, SCN2B^R137H leads to structural changes but overall protein stability is unaffected.

Conclusion:

SCN4B^I80T data suggest a functionally important region in the interaction between Na_V1.5 and β4 that, when disrupted, could lead to channel dysfunction. A lack of apparent functional effects of SCN2B^R137H on Na_V1.5 suggests an alternative working mechanism, possibly through other Na_V channel subtypes present in heart tissue. Indeed, mapping the structural variations of SCN2B^R137H onto neuronal Na_V channel structures suggests altered interaction patterns.

Introduction

Voltage-gated sodium (Na_V) channel mutations can cause dramatic alterations in cellular excitability leading to potentially life-threatening disorders.¹ In cardiac myocytes, Na_V1.5 is the main Na_V channel subtype responsible for heart muscle function.^2,3 Consequently, mutations in this channel have indeed been associated with long QT syndrome (LQTS)⁴ and Brugada syndrome (BrS).⁵ However, it is now well established that Na_V channels do not function in isolation but form membrane-embedded macromolecular complexes that include auxiliary proteins such as β-subunits,^1,6–8 fibroblast growth factor proteins, calmodulin, and others.^9–16 β-subunits are known to modify gating properties, regulate trafficking, and influence plasma membrane recruitment of Na_V channels.⁶ Four main β-subunit isoforms (β1–4) have been identified and despite a shared, single transmembrane-segment protein topology, each interacts distinctly with and exerts disparate effects on specific Na_V channel subtypes (Na_V1.1–1.9).^6,17–19 All four β-subunits have been identified in atrial and ventricular cardiomyocytes,² and mutations have been implicated in cardiac disorders.²⁰

Potential disease-associated mutations have also been found in SCN4B (p.Ile80Thr) and SCN2B (p.Arg137His), encoding the Na_V channel β4- and β2-subunit, respectively.^1,21 These mutations were identified during genetic screening in two patients affected by familial atrial fibrillation and LQTS, respectively (www.ncbi.nlm.nih.gov/clinvar/). In this study, we investigated the functional and structural consequences of these mutations on Na_V channel complex function, providing initial evidence of altered interactions.

Methods

Electrophysiological recordings in Xenopus laevis oocytes

DNA sequencing of human (h)Na_V1.2, hNa_V1.5 C373Y, hβ2, and hβ4 (Origene, USA) was confirmed by automated Sanger sequencing.²² β-subunit mutants were generated using the Gibson assembly method. cRNA was synthesized using T7 polymerase (the mMessage mMachine Kit; Life Technologies, USA) after linearizing the DNA by restriction enzyme digest. hNa_V channel constructs were expressed in Xenopus laevis oocytes (sourced from Xenopus one^®) with or without a β-subunit (1:5 molar ratio) and electrophysiological recordings were taken 2–4 days post cRNA injection. Oocytes were maintained at 17°C in Barth's medium (96 mM NaCl, 2 mM KCl, 5 mM HEPES, 1 mM MgCl₂, and 1.8 mM CaCl₂, 50 μg/mL gentamicin, pH 7.6) and studied using the two-electrode voltage clamp recording technique (OC-725C; Warner Instruments, USA) with a 150 μL recording chamber. All data were filtered at 4 kHz and digitized at 20 kHz using pClamp 10 software (Molecular Devices, USA). The external recording solution used was ND100 (100 mM NaCl, 5 mM HEPES, 1 mM MgCl₂, and 1.8 mM CaCl₂, pH 7.6) and microelectrode resistances were 0.5–1.0 MΩ when filled with 3 M KCl. Recordings were performed at room temperature (∼22°C) and leak and background conductance, identified by blocking Na_V1.2 and the tetrodotoxin (TTX)-sensitive Na_V1.5 C373Y channel with TTX (Alomone Labs, Israel), was subtracted for all channel currents. Off-line data analysis was performed using Clampfit 10 (Molecular Devices), Excel (Microsoft, USA), and Prism 7 (GraphPad, USA).

Qualitative surface biochemistry

X. laevis oocytes expressing hNa_V1.5, hβ4, hβ4-I80T, hβ2, hβ2-R137H were washed 2 × with ND100 and incubated with 0.5 mg/mL Sulfo-NHS-LC-biotin (Pierce, USA) for 1 h at 4°C. Oocytes were then washed 6 × in ND100 before lysis by sonication in 20 μL/oocyte Nonidet P-40 (NP-40, 1%) with protease inhibitor cocktail (Pierce). All subsequent steps were performed at 4°C. Lysates were gently shaken for 15 min and then centrifuged for 15 min at 14,000 rpm. The pellet was discarded and 40 μL of the supernatant was reserved as the total protein fraction. The surface fraction was generated by adding the remaining supernatant to 200 μL of MyOne™ Streptavidin C1 magnetic beads (Invitrogen, USA), raising the volume to 1 mL with NP-40, and then rotating overnight at 4°C. Beads were then washed 6 × with NP-40 and resuspended in 30 μL NP-40, after which the biotinylated protein was dissociated by addition of 1 × Bolt™ Sample Buffer plus Reducing Agent (Thermo Fisher Scientific, USA) and boiling at 95°C for 5 min. Protein concentrations were measured by bicinchoninic acid assay (Pierce) and samples were run on a Bolt 4–12% Bis-Tris Plus gel (Thermo Fisher Scientific), then subsequently analyzed by western immunoblotting. Membrane images were cropped to show relevant bands only.

Statistical analyses

Mean data points for fast inactivation time constants (τ), persistent current, and recovery from fast inactivation (RFI) were analyzed using a one-way analysis of variance. For data on normalized conductance-voltage and steady-state inactivation relationships presented in Table 1, unpaired Student's t-test against wild-type channel or wild-type β-subunit was used. Values in all cases reflect the mean and error bars reflect standard error of the mean. All analyses were carried out using Excel (Microsoft).

Table 1.

Gating Parameters of Na_V Channels Without and in the Presence of β-Subunits and Mutants

	Activation					Channel availability
	N	V_1/2	SEM	d(x)	SEM	N	V_1/2	SEM	d(x)	SEM
Na_V1.5	7	−39.8	0.1	4.6	0.1	5	−73.8	0.5	8.0	0.3
β4	6	−35.6^a	0.2	5.7^a	0.1	6	−74.4	0.5	9.6^a	0.3
I80T	5	−40.1^b	0.2	4.8^b	0.2	5	−78.5^a,b	0.5	8.9	0.2
β2	5	−34.7^a	0.3	6.5^a	0.3	7	−67.2^a	0.4	7.9	0.2
R137H	5	−35.3^a	0.5	7.0^a	0.5	5	−68.7^a	0.5	7.8	0.3
Na_V1.1	6	−20.5	0.2	6.3	0.2	5	−32.3	0.4	7.6	0.3
β4	5	−22.1	0.1	5.7	0.1	6	−31.4	0.2	7.3	0.3
I80T	4	−20.6	0.2	6.3	0.1	6	−31.8	0.2	7.7	0.1
Na_V1.2	5	−22.5	0.3	5.4	0.3	5	−37.7	0.4	9.00	0.4
β4	5	−21.9	0.4	5.1	0.4	5	−34.9	0.3	7.1^a	0.3
I80T	5	−21.9	0.3	6.0	0.2	4	−35.6	0.3	9.1^b	0.2
β2	5	−18.8^a	0.4	4.5	0.4	5	−34.9^a	0.4	9.3	0.4
R137H	5	−22.0^b	0.6	4.9	0.5	5	−36.0	0.6	9.4	0.6

N = number of experiments, d(x) = slope value.

p < 0.01 compared with WT channel by unpaired t-test.

p < 0.01 compared with WT β-subunit by unpaired t-test.

SEM, standard error of the mean; WT, Wild-Type.

Recombinant protein production and crystallographic analysis

Human β4 C58A (residues 32–157) and β2 C55A (residues 30–153) were cloned into a modified pET28 vector.^17,18 The β4-C58A/I80T and β2 C55A/R137H mutations were introduced using the QuikChange Kit from Agilent Technologies (USA). Both mutants were expressed at 18°C in Escherichia coli Rosetta (DE3) pLacI strains (Novagen, USA), induced at an OD600 of ∼0.6 with 0.4 mM IPTG, and grown overnight before harvesting. Cells were lysed through sonication in buffer A (250 mM KCl and 10 mM HEPES at pH 7.4), supplemented with 25 mg/mL DNaseI and 25 mg/mL lysozyme. After centrifugation, the supernatants were applied to a PorosMC column (Tosoh Biosep, USA), washed with buffer A plus 10 mM imidazole, and eluted with buffer B (250 mM KCl plus 500 mM imidazole pH 7.4).

The mutants were dialyzed overnight against buffer A and cleaved simultaneously with recombinant tobacco etch virus (TEV) protease. Next, the samples were run on another PorosMC column in buffer A, and the flowthrough was collected and dialyzed against buffer C (10 mM KCl plus 10 HEPES at pH 7.4), applied to a HiloadQ column (GE Healthcare, USA), and eluted with gradients from 0% to 30% buffer D (2 M KCl plus 10 mM HEPES at pH 7.4). Finally, samples were run on a Superdex75 (GE Healthcare) gel filtration column in buffer A. β4-C58A/I80T and β2-C55A/R137H were exchanged to 50 mM KCl plus 10 mM HEPES (pH 7.4), concentrated to 9.5 and 2.8 mg/mL, respectively using Amicon concentrator (3K MWCO; Millipore, USA), and stored at −80°C.

Crystallization, data collection, and structure solution

Crystals were grown using the hanging-drop method at 4°C. β4-C58A/I80T was crystallized in 0.1 M MES (pH 6.5), 15–17% (w/v) PEG 2000-MME and β2-C55A/R137H crystals were grown in 0.1 M HEPES (pH 7.0), 14% PEG 6000 using seeds from β2-C55A crystals at various dilutions. Crystals were flash frozen after transfer to the mother liquor supplemented with 30% glycerol. The datasets were collected at the beam line BL9-2 of the Stanford Synchrotron Radiation Lightsource and beam line 23-ID-D at the Advanced Photon Source (Chicago). Data were processed using XDS²³ and HKL3000.²⁴ Phases were obtained through molecular replacement through Phaser²⁵ using the available crystal structures of β4-C58A (4MZ2.pdb) and β2-C55A (5FEB.pdb). The structures were completed by manual model building in COOT²⁶ and refinement using Phenix.²⁷ Simulated annealing composite omit maps were calculated with CNS²⁸ software to verify the absence of residual model bias. All structure figures were prepared using PYMOL (DeLano Scientific, San Carlos, USA). Protein Data Bank (PDB) IDs for mutant β-subunits are 6VRR and 6VSV.

Circular dichroism spectroscopy

All proteins were diluted to 16 μM in 50 mM KCl, 10 mM K-phosphate, pH 7.6 supplemented with 5 mM β-mercaptoethanol. Spectra were measured from 200–260 nm in a Jacso J-810. The initial thermal melt of β4-C58A from 25°C to 95°C in a Peltier controlled sample chamber connected to a water bath showed highest secondary structural changes at 213 nm. For all subsequent melts, the signal at 213 nm was monitored as a function of temperature.

Results

We examined the effects of SCN4B, SCN4B^I80T (β4^I80T), SCN2B, and SCN2B^R137H (β2^R137H) on Na_V channel gating. These mutations were chosen from the ClinVar database (www.ncbi.nlm.nih.gov/clinvar/), which contains many gene variants of unknown or unsubstantiated clinical significance. In this database, both variants are linked to LQTS 10, a cardiac electrophysiological disorder characterized by QT prolongation and T-wave abnormalities associated with tachyarrhythmias and syncope events that typically occur during exercise or emotional stress.^29,30 Original data were submitted to ClinVar by Invitae, a genetic testing company based in California (USA). In this study, we exploit this information to investigate possible changes in SCN2B and SCN4B interaction with Na_V channels, which may lay the foundation for future experiments in which a link with the clinical phenotype is investigated.

Electrophysiological analysis

Since the clinical phenotype revolves around heart function, we used human (h)Na_V1.5 as our model channel.³ Similar to wild-type hβ4 and hβ2, both hβ4^I80T and hβ2^R137H express in Xenopus oocytes and translocate to the membrane, indicating that protein trafficking is not affected, at least in a heterologous Xenopus oocyte expression system (Figs. 1 and 2). To discern potential subtle effects of hβ4 and hβ2 on Na_V channel function, we recorded and compared multiple gating parameters, such as the conductance-voltage (G-V) relationship, channel availability (I-V), RFI, and persistent current between hNa_V1.5 alone or in the presence of hβ4 or hβ4^I80T. Channel opening kinetics were not studied due to inherent technical limitations associated with electrophysiological recordings in Xenopus oocytes.

FIG. 1.

(A–C) Representative traces of hNa_V1.5 alone or coexpressed with hβ4 (red) or hβ4^I80T (magenta). Oocytes were held at −120 mV and pulsed from −90 to 15 mV for 50 ms. (D) Western blot probing for the C-terminal myc-tag of hβ4. Signal is seen for wild-type hβ4 as well as hβ4^I80T, in both whole cell (closed circles) and surface (open circle) fractions. (E) Normalized conductance voltage (G-V, filled circles) and channel availability (I-V, open circles) relationships for hNa_V1.5 either without hβ4 (black), with hβ4 (red), or with hβ4 I80T (magenta). Summary data of G-V and I-V V_1/2 values, slope values, and p-values determined by unpaired t-test are reported in Table 1. (F) Rate (τ) of channel fast inactivation using the same color scheme described above. Time constant was obtained using a single-exponential fit of the current decay phase over a 45 mV range. Error bars in all cases reflect SEM with n = 5–7; *p < 0.05 or **0.01 using one-way ANOVA. ANOVA, analysis of variance; hNa_V, human voltage-gated sodium; SEM, standard error of the mean.

FIG. 2.

(A, B) Representative traces of hNa_V1.5 coexpressed with hβ2 (red) or hβ2^R137H (cyan). Oocytes were held at −120 mV and pulsed from −90 to 15 mV for 50 ms. (C) Normalized conductance voltage (G-V, filled circles) and channel availability (I-V, open circles) relationships for hNa_V1.5 without a β-subunit (black) or with either hβ2 (red) or with hβ2 R137H (cyan). Summary data of G-V and I-V V_1/2 values, slope values, and p-values determined by unpaired t-test are reported in Table 1. (D) Rate (τ) of channel fast inactivation using the same color scheme described above. Time constant was obtained using a single-exponential fit of the current decay phase over a 45 mV range. (E) Western blot probing for the C-terminal myc-tag of hβ2. Signal is seen for wild-type hβ2 as well as hβ2^R137H, in both whole cell (closed circles) and surface (open circle) fractions. Error bars in all cases reflect SEM. with n = 5–7; *p < 0.05 or **0.01 using one-way ANOVA.

We found that the G-V relationship of hNa_V1.5/hβ4^I80T is shifted to more hyperpolarized potentials (±5 mV; p < 0.01) when compared with hNa_V1.5/hβ4. A similar shift was seen in channel availability (±4 mV; p < 0.01). Another apparent effect of hβ4^I80T emerged when fitting the current decay phase with a single-exponential fit. Over a 45-mV range, hβ4^I80T caused Na_V1.5 currents to inactivate significantly faster when compared with wild-type hβ4 (p < 0.05). Persistent current and RFI of hNa_V1.5 were unaffected (Fig. 3). The G-V and I-V relationships of two neuronal Na_V channel subtypes, hNa_V1.1 and hNa_V1.2, were not disrupted by the hβ4^I80T mutation (Fig. 3). Markedly, hβ2^R137H did not have any apparent effect on the tested hNa_V1.5 gating parameters as compared with wild-type hβ2 (Fig. 2).

FIG. 3.

(A) Normalized persistent current of hNa_V1.5 (black), hNa_V1.5 +hβ4 (red), or hNa_V1.5 +hβ4^I80T (magenta) measured 20 ms after depolarization over a 30 mV range. (B) Normalized recovery from fast inactivation measured over a 50 ms timeframe using a double-pulse protocol to the maximum current of the I-V in Figure 1. Recovery voltage used for this protocol was −10 mV. (C, D) Normalized conductance voltage (G-V, filled circles) and channel availability (I-V, open circles) relationships of hβ4 (red) or hβ4^I80T (magenta) coexpressed with either hNa_V1.1 (C) or hNa_V1.2 (D). Both wild-type channels without β-subunit coexpression are shown in black. Summary data of G-V and I-V V_1/2 values, slope values, and p-values determined by unpaired t-test are reported in Table 1.

Crystallographic analysis

Since hβ2 and hβ4 can be crystallized,^17,18 we prepared crystals for the hβ2^R137H and hβ4^I80T variants (PDB IDs are 6VRR and 6VSV). Both resulted in high-resolution structures of 1.45 and 1.65 Å, respectively (Table 2). Figure 4 shows a comparison of wild-type and hβ2^R137H. The two structures can be superposed with an root-mean-square deviation (RMSD) of 0.79 Å for 90 Cα atoms (with Cα atoms of residues with double conformations not included). The R137H mutation results in multiple structural changes at distances >12 Å away from the mutation site. In wild-type hβ2, the Arg¹³⁷ side chain is stacked against the side chain of Tyr¹²⁸, making cation–π interactions. It also forms two hydrogen bonds with a water molecule involved in a hydrogen-bond network, including Ser⁶⁴ and the Tyr¹²⁸ main chain. In the hβ2^R137H, these interactions are lost, resulting in a reorientation of the Ser⁶⁴ side chain. Instead, His¹³⁷ adopts a dual conformation and a 1.6 Å displacement of its main chain, which in turn propagates to several neighboring residues. This includes a near 10 Å shift in the side chain of Asn²⁸, which makes a hydrogen bond with one of the His¹³⁷ conformers, and large shifts of the Met¹³⁰ side chain, His¹³⁶, and Arg¹³⁵. Overall, the R137H mutation results in significant changes in the local area around the mutation site. Although this seems at odds with the lack of effect of this mutation on Na_V1.5 function (Fig. 2), it suggests that the structural changes do not disrupt the interface between hβ2 and Na_V1.5. However, this mutation may lead to altered interactions with other Na_V channel subtypes.

FIG. 4.

Crystallographic analysis of hβ2^R137H. Wild-type hβ2 (containing the C55A mutation to facilitate crystallization¹⁷) is shown in blue, and the R137H mutant in green. R137 is shown in black sticks and H137 in pink. (A) Overall superposition of both structures. (B) Details around the site of the mutation. Selected residues are shown in sticks and labeled. Hydrogen bonds are shown with dotted lines. In wild-type hβ2, the Arg137 side chain is involved in a hydrogen bond network as well as a cation–π interaction with Y128. Substitution by H137 results in multiple changes in side chain orientations, as well as the main chain conformation of the N-terminal region around N28.

Table 2.

Data Collection and Refinement Statistics

Data collection	β4-C58A/I80T	β2-C55A/R137H
PDB code
Space group	P3₂21	P2₁2₁2₁
Cell dimensions
a, b, c (Å)	43.1, 43.1, 107.5	40.3, 53.6, 59.3
α, β, γ (°)	90, 90, 120	90, 90, 90
Resolution (Å)	50–1.62 (1.68–1.65)^a,b	40.3–1.45 (1.53–1.45)^a,b
R_meas	7.9 (65.4)	4.3 (62.5)
I/σI	33.5 (1.9)	24.4 (2.75)
Completeness (%)	99.8 (99.4)	99.9 (99.6)
Redundancy	5.1 (4.7)	6.9 (6.54)
CC (1/2)^c	76.0	92.9
Refinement
Resolution (Å)	37.3–1.62	40.3–1.45
No. of reflections	15397	43929
Molecules	1	1
R_work/R_free	18.47/20.81	14.8/18.35
No. of atoms
Protein	990	2436
Ligand	—	28
Water	147	163
B-factors
Protein	33.2	23.9
Ligand	—	44.2
Water	48.2	34.6
R.m.s. deviations
Bond lengths (Å)	0.012	0.013
Bond angles (°)	1.229	1.623
Ramachandran plot^d
Favored (%)	96.3	97.6
Allowed (%)	3.7	2.4
Disallowed (%)	—	—

One crystal for each structure was used for data collection and structure determination.

Values in parentheses are for highest-resolution shell.

Value is for highest-resolution shell.

Using Procheck of CCP4 package.

The situation is very different for the hβ4^I80T mutation. Wild-type and mutant subunit superpose with an RMSD of 0.48 Å for 105 Cα atoms (Fig. 5). In wild-type hβ4, Ile⁸⁰ is buried within the hydrophobic core of the Ig-like domain, which consists of Trp⁶⁸, Leu⁷⁹, Val⁹³, Leu¹⁰³, and Ile¹⁰¹. Substitution by a hydrophilic residue would thus be predicted to be destabilizing. In addition, as Thr is smaller than Ile, the packing is less optimal; resulting in a change in the side chain of Leu¹⁰³, which now adopts a dual side chain conformation. Importantly however, no structural changes are visible on the surface. As the surface of hβ4 would be interacting with Na_V1.5, there is an apparent contradiction with the functional data, which clearly show an effect of the I80T mutation on the G-V relationship and steady-state inactivation (Fig. 1). However, as this protein crystallized at low temperature (4°C), and crystals were flash cooled with liquid N₂, a possible temperature-dependent effect on the structure may not be observed. We thus postulate that the suboptimal packing in the hβ4 core can result in changes on the surface at elevated temperature, such as used in electrophysiological experiments or present in vivo.

FIG. 5.

Crystallographic analysis of hβ4^I80T. Wild-type hβ4 (containing the C58A mutation to facilitate crystallization) is shown in blue, and the I80T mutant in orange. I80 is shown in black sticks, and T80 in pink. (A) Overall superposition of both structures. (B) Details around the site of the mutation. Select residues are shown in sticks and labeled. I80 is involved in packing of the hydrophobic core, making interactions with other hydrophobic residues. Substitution by Thr results in minor structural changes, including a rearrangement of the I101 side chain conformation.

Protein stability

Amino acid substitutions may perturb protein folding. Although both β-subunit mutants still express to levels similar to wild type in Xenopus oocytes (Figs. 1 and 2), it is possible that these substitutions cause a thermal destabilization at higher temperatures. Such effect could explain how a mutant, which shows minimal structural changes in a crystallized form at low temperature, can still have a functional effect at physiologically relevant temperatures. We, therefore, prepared purified, recombinantly expressed extracellular subunits of hβ4, hβ2, and the two sequence variants, and utilized Circular Dichroism (CD) to measure thermal melting curves (Fig. 6). Monitoring the CD signal at 213 nm, we increased the temperature from 25°C to 95°C in increments of 0.5°C. Wild-type hβ4 displays a melting temperature (T_M) of 61.1°C ± 0.3°C, whereas the I80T mutation results in a significant destabilization with a T_M of 51.4°C ± 0.1°C (Fig. 6).

FIG. 6.

Thermal melting curves monitored through CD recorded at 213 nm.(A) Wild-type hβ4 (containing the C58 mutation—black) and the I80T mutant (red). (B) Wild-type hβ2 (containing the C58A mutation—black) and the R137H mutant (green). The T_M values were obtained from the midpoints of the transitions. CD, circular dichroism.

Although the actual T_M in the context of a native cell, under oxidizing conditions and in the presence of a Na_V channel is unknown, our observation that the mutation has an effect on stability in our assay hints at possible structural changes at elevated temperatures. Although the T_M is well above 37°C, a subtle degree of unfolding is already present at this temperature. Of note, the CD signal measures secondary structure, and it is thus conceivable that significant local structural perturbations, which do not affect secondary structure content, already occur at physiologically relevant temperatures, which would affect the interface between hβ4 and Na_V1.5. Wild-type hβ2 is significantly less stable than hβ4, with a T_M of 47.4°C ± 0.1°C. However, in this case, no significant impact on protein stability was observed for the R137H mutation (T_M = 45.8°C ± 0.1°C).

Mapping β-subunit variants on cryoelectron microscopy structures

Next, we wondered whether the β-subunit mutations studied here are part of an interface with Na_V1.5. A cryoelectron microscopy (EM) structure is available for Na_V1.5 as well as crystal structures of cytosolic portions in complex with auxiliary proteins^{9,10,12–15,31} and also cryo-EM structures of the related Na_V1.2, Na_V1.4, and Na_V1.7 channels, in complex with one or more auxiliary β-subunits.^7,8,32 This includes structures of hNa_V1.2 in complex with β2 (PDB 6J8E), hNa_V1.4 in complex with β1 (PDB 6AGF), and hNa_V1.7 in complex with β1 and β2 (PDB 6J8G). In each of the β2 complexes, a disulfide bond was observed with the pore-forming subunit. This corresponds to Cys⁹¹⁰ in Na_V1.2, a residue we predicted to form a disulfide bond with Cys⁵⁵ in hβ2.¹⁷ Although no cryo-EM structures have captured β4 in complex with a Na_V channel, this subunit is also predicted to make a similar disulfide bond.^17,18 However, as Na_V1.5 does not have a Cys residue at the corresponding position, the interaction between β2/β4 with Na_V1.5 cannot rely on this disulfide bridge.^17,18,31

Figure 7 shows the structure of the Na_V1.2-β2 complex (PDB 6J8E), illustrating that the interaction also involves several other residues that make extensive interactions with the Domain II pore-forming region of the channel. From these data, it is clear that the R137H variant is in a position to perturb this interaction. Although R137 is not directly located at the interface, the structural changes we observed for this mutation affect residues located at the β2-Na_V1.2 interface, including Arg,²⁸ Arg¹³⁵, and His¹³⁶. Thus, the hβ2 R137H variant could perturb the interface with Na_V1.2, and likely also closely related Na_V channel subtypes expressed in the heart (i.e., Na_V1.1–1.3).³³ Since hβ2^R137H has no functional effect on Na_V1.5 compared with hβ2, whereas hβ4^I80T does, it seems unlikely that β2 and β4 interact with Na_V1.5 similar to the Na_V1.2-β2 complex. Indeed, Na_V1.5 lacks the Cys equivalent to position 910 in Na_V1.2, suggesting that the manner in which these two β-subunits interact with Na_V1.5 is fundamentally different.^17,31,34

FIG. 7.

Mapping the mutation sites on the Na_V1.2-β2 cryo-EM structure. (A) Overall cryo-EM structure of Na_V1.2 bound to hβ2 (PDB 6J8E). The pore-forming domain and VSDs are shown in gray and red, respectively, and hβ2 in blue. (B). Details of the interaction. Selected residues in hβ2 are labeled. Notice that several of the hβ2 residues, shown to be affected by the R137H mutation, are directly at the interface with Na_V1.2. EM, electron microscopy; VSDs, voltage-sensing domains.

Na_V1.2 gating and β2^R137H

Since neuronal Na_V channel subtypes expressed in the heart² contain the Cys equivalent to position 910 in Na_V1.2,^7,17 the structural changes induced by the β2 R137H mutation may disrupt channel function. To test this hypothesis with links to available structural data, we used hNa_V1.2 as our model channel, coexpressed with wild-type β2 or β2^R137H. While we observe no changes in channel availability of hNa_V1.2, the G-V relationship and fast inactivation in the presence of β2^R137H differs substantially from that observed with wild-type β2 (p < 0.001) and is virtually identical to the condition when no subunit is present (Fig. 8).

FIG. 8.

(A) Normalized conductance voltage (G-V, filled circles) and channel availability (I-V, open circles) relationships for hNa_V1.2 without a β-subunit (black) or with either hβ2 (red) or with hβ2 R137H (cyan). Summary data of G-V and I-V V_1/2 values, slope values, and p-values determined by unpaired t-test are reported in Table 1. (B) Rate (τ) of channel fast inactivation using the same color scheme described above. Time constant was obtained using a single-exponential fit of the current decay phase over a 45 mV range. Error bars in all cases reflect SEM. with n = 5; *p < 0.05 or **0.01 using one-way ANOVA.

Discussion

We investigated the functional and structural effects of two β-subunit variants, SCN4B (p.Ile80Thr) and SCN2B (p.Arg137His), found in patients with familial atrial fibrillation and LQTS 10. Our functional investigation showed a direct effect of the hβ4^I80T, but not the hβ2^R137H variant, on the function of hNa_V1.5 (Figs. 1 and 2). However, structural investigation using X-ray crystallography revealed opposite results, with hβ2^R137H causing significant structural changes and hβ4^I80T resembling the wild-type protein under the crystallization conditions used (Figs. 4 and 5). A structural change does not necessarily have to cause a functional impact, especially if the perturbation is local and does not affect a protein–protein interface. The results presented here suggest that the region around Arg¹³⁷ in hβ2 may not be involved in the interface with Na_V1.5. However, the hβ4^I80T results seem contradictory given that the absence of significant structural changes suggests that the interface with Na_V1.5 should be unaffected.

To better understand this conundrum, it is important to consider a possible effect of temperature. Proteins will unfold, both locally and globally, with rising temperatures, and a mutation could affect this thermal stability.^35,36 If any unfolding of the mutant occurs at physiological temperatures, and this affects a relevant interface, then the mutant is likely to have a functional effect. In this regard, the hβ4^I80T caused a significant reduction in thermal stability (Fig. 6). This is not surprising, as our structural analysis showed that Ile⁸⁰ is buried in a hydrophobic core, and substitution by a hydrophilic residue should be destabilizing. One could thus postulate that the mutation causes local changes in structure at or near the interface between hβ4 and Na_V1.5 at physiologically relevant temperatures, leading to alterations in channel gating. Importantly, thermal melting curves obtained via CD only represent changes in secondary structure, so small structural perturbations may not be measured. Thus, although the melting temperature is above a physiologically relevant temperature, small local changes that do not affect secondary structure are still possible at physiological temperatures. In addition, the altered stability can affect the dynamics of hβ4, thus affecting its interactions with Na_V1.5 indirectly.

Temperature-dependent effects have previously been reported as a result of mutations in Na_V1.5. This includes amino acid substitutions in the EF-hand domain leading to thermal destabilization of this locus.^37,38 Thermal destabilization induced by disease-causing mutations has also been described in other ion channel families.^39,40 Since the hβ4^I80T mutation was found in a LQTS patient, an increase in body temperature (e.g., fever or exercise) could elicit structural and functional changes in the cardiac Na_V channel complex, thereby causing an increase in QT interval time and subsequently an increased risk to develop torsade de pointes and polymorphic ventricular tachycardia. Indeed, fever-related ventricular arrhythmia and sudden death have been associated extensively with BrS⁴¹; however, few reports also describe fever-induced QT prolongation and ventricular arrhythmias in individuals with type 2 congenital LQTS.⁴² As such, it is possible that the SCN4B p.Ile80Thr mutation could favor the appearance of an arrhythmic phenotype in concomitance with fever. We note that this remains a hypothesis, and it needs to be shown whether the mutation elicits small structural or dynamic changes at physiological temperatures.

Although the hβ2^R137H variant did not appear to affect any of the electrophysiological properties of Na_V1.5, mapping of the mutation on the Na_V1.2-hβ2 cryo-EM structure showed that this residue can affect channel–subunit interactions. Although not directly involved in the interface, we found that the structural changes we observed affect residues directly participating in their interaction, including R135 and H136 (Fig. 7). Subsequently, hβ2^R137H indeed restored hNa_V1.2 function as if no hβ2-subunit was present when considering channel closure (i.e., the fast inactivation process) (Fig. 8). Since several neuronal Na_V channel isoforms are expressed in the heart at functionally important locations,^2,33 our result hints at a contribution of this hβ2 variant to arrhythmia,^43,44 despite the absence of functional effects on the main cardiac isoform, Na_V1.5. From a clinical perspective, hβ2^R137H was found in a patient with familial atrial fibrillation and since intrinsic and extrinsic cardiac innervations are pivotal in determining physiological modulations in cardiomyocytes,⁴⁵ it may be feasible that this mutation can modify neural activity in the heart.

Finally, it is worth noting that the extracellular Ig domain of β-subunits can participate in homo- and heterophilic cell adhesion processes, which may have implications for cardiomyocyte function.³⁴ Combined with previous work,^{17,18,30,31,34,46,47} it is unlikely that the Ig domains of hβ2 and hβ4 interact with Na_V1.5 in the same way that they do with other Na_V channel subtypes, since Na_V1.5 does not have a free Cys residue on the domain II extracellular loop.^17,31 As such, the exact interface of hβ2 or hβ4 with Na_V1.5 remains to be described, but likely does not involve hβ2 residue R137 or the neighboring residues whose positions are affected by the R137H mutation. Ig domains of hβ4 (and presumably hβ2) can form both cis and trans cell adhesion contacts and some of the same amino acids involved in binding to the Na_V channel also contribute to these cis interactions.³⁴ Therefore, hβ2 and hβ4 Ig domains that associate with Na_V1.5 may be involved in crosslinking of associated channels suggesting the involvement of higher-order assemblies in potentially causing cardiac defects and illustrating the complexities in linking genetic data to patient phenotype.

Footnotes

Author Contributions

F.B and F.V.P. conceived the experiments, analyzed data, and wrote the article. J.D.W., L.C., and A.S. helped write the article and interpret (clinical) data. S.D. performed all crystallographic and thermal melt experiments. J.P.L. performed all electrophysiology and biochemical experiments. All coauthors have reviewed and approved of the article before submission.

Acknowledgments

We acknowledge the staff at the Stanford Synchrotron Radiation Lightsource beamline BL9-2, and the Advanced Photon Source beamline 23-ID-D.

Disclaimer

The article has been submitted solely to this journal and is not published, in press, or submitted elsewhere.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

J.P.L. was supported by the James H. Gilliam Fellowships for Advanced Study through the Howard Hughes Medical Institute. F.B. received support from the National Institutes of Health (1R01NS091352). F.V.P. received support for this study by a grant from the Canadian Institute for Health Research (CIHR, PJT-148632). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the article.

References

Ahern

, Payandeh

, Bosmans

, et al. The hitchhiker's guide to the voltage-gated sodium channel galaxy. J Gen Physiol, 2016; 147:1–24.

Kaufmann

, Westenbroek

, Maass

, et al. Distribution and function of sodium channel subtypes in human atrial myocardium. J Mol Cell Cardiol, 2013; 61:133–141.

Rook

, Evers

, Vos

, et al. Biology of cardiac sodium channel Nav1.5 expression. Cardiovasc Res, 2012; 93:12–23.

George

Jr.

Inherited disorders of voltage-gated sodium channels. J Clin Invest, 2005; 115:1990–1999.

Brugada

, Brugada

. Right bundle branch block, persistent ST segment elevation and sudden cardiac death: A distinct clinical and electrocardiographic syndrome. A multicenter report. J Am Coll Cardiol, 1992; 20:1391–1396.

O'Malley

, Isom

. Sodium channel beta subunits: Emerging targets in channelopathies. Annu Rev Physiol, 2015; 77:481–504.

Pan

, Li

, Huang

, et al. Molecular basis for pore blockade of human Na+ channel Nav1.2 by the mu-conotoxin KIIIA. Science, 2019; 363:1309–1313.

Shen

, Liu

, Wu

, et al. Structures of human Nav1.7 channel in complex with auxiliary subunits and animal toxins. Science, 2019; 363:1303–1308.

Wang

, Chung

, Yan

, et al. Crystal structure of the ternary complex of a Nav C-terminal domain, a fibroblast growth factor homologous factor, and calmodulin. Structure, 2012; 20:1167–1176.

10.

Wang

, Chung

, Yan

, et al. Structural analyses of Ca(2+)/CaM interaction with Nav channel C-termini reveal mechanisms of calcium-dependent regulation. Nat Commun, 2014; 5:4896.

11.

Yan

, Wang

, Marx

, et al. Calmodulin limits pathogenic Na+ channel persistent current. J Gen Physiol, 2017; 149:277–293.

12.

Gabelli

, Boto

, Kuhns

, et al. Regulation of the Nav1.5 cytoplasmic domain by calmodulin. Nat Commun, 2014; 5:5126.

13.

Yoder

, Ben-Johny

, Farinelli

, et al. Ca(2+)-dependent regulation of sodium channels Nav1.4 and Nav1.5 is controlled by the post-IQ motif. Nat Commun, 2019; 10:1514.

14.

Sarhan

, Tung

, Van Petegem

, et al. Crystallographic basis for calcium regulation of sodium channels. Proc Natl Acad Sci U S A, 2012; 109:3558–3563.

15.

Johnson

, Potet

, Thompson

, et al. A mechanism of calmodulin modulation of the human cardiac sodium channel. Structure, 2018; 26:683–694 e683.

16.

Gardill

, Rivera-Acevedo

, Tung

, et al. Crystal structures of Ca(2+)-calmodulin bound to Nav C-terminal regions suggest role for EF-hand domain in binding and inactivation. Proc Natl Acad Sci U S A, 2019; 116:10763–10772.

17.

Das

, Gilchrist

, Bosmans

, et al. Binary architecture of the Nav1.2-beta2 signaling complex. eLife, 2016; 5:e10960.

18.

Gilchrist

, Das

, Van Petegem

, et al. Crystallographic insights into sodium-channel modulation by the beta4 subunit. Proc Natl Acad Sci U S A, 2013; 110:E5016–E5024.

19.

Namadurai

, Balasuriya

, Rajappa

, et al. Crystal structure and molecular imaging of the Nav channel beta3 subunit indicates a trimeric assembly. J Biol Chem, 2014; 289:10797–10811.

20.

Watanabe

, Koopmann

, Le Scouarnec

, et al. Sodium channel beta1 subunit mutations associated with Brugada syndrome and cardiac conduction disease in humans. J Clin Invest, 2008; 118:2260–2268.

21.

Pancaroglu

, Van Petegem

. Calcium channelopathies: Structural insights into disorders of the muscle excitation-contraction complex. Annu Rev Genet, 2018; 52:373–396.

22.

Gilchrist

, Dutton

, Diaz-Bustamante

, et al. Nav1.1 modulation by a novel triazole compound attenuates epileptic seizures in rodents. ACS Chem Biol, 2014; 9:1204–1212.

23.

Kabsch

. XDS. Acta Crystallogr D Biol Crystallogr, 2010; 66:125–132.

24.

Minor

, Cymborowski

, Otwinowski

, et al. HKL-3000: The integration of data reduction and structure solution—From diffraction images to an initial model in minutes. Acta Crystallogr D Biol Crystallogr, 2006; 62:859–866.

25.

McCoy

, Grosse-Kunstleve

, Adams

, et al. Phaser crystallographic software. J Appl Crystallogr, 2007; 40:658–674.

26.

Emsley

, Lohkamp

, Scott

, et al. Features and development of Coot. Acta Crystallogr D Biol Crystallogr, 2010; 66:486–501.

27.

Adams

, Afonine

, Bunkoczi

, et al. PHENIX: A comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr D Biol Crystallogr, 2010; 66:213–221.

28.

Brunger

, Adams

, Clore

, et al. Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr D Biol Crystallogr, 1998; 54:905–921.

29.

, Wang

, Xu

, et al. Mutations of the SCN4B-encoded sodium channel beta4 subunit in familial atrial fibrillation. Int J Mol Med, 2013; 32:144–150.

30.

Medeiros-Domingo

, Kaku

, Tester

, et al. SCN4B-encoded sodium channel beta4 subunit in congenital long-QT syndrome. Circulation, 2007; 116:134–142.

31.

Jiang

, Shi

, Tonggu

, et al. Structure of the cardiac sodium channel. Cell, 2020; 180:122–134.e10.

32.

Yin

, Hassan

, Haroun

, et al. Arrhythmogenic calmodulin mutations disrupt intracellular cardiomyocyte Ca2+ regulation by distinct mechanisms. J Am Heart Assoc, 2014; 3:e000996.

33.

Veeraraghavan

, Gyorke

, Radwanski

. Neuronal sodium channels: Emerging components of the nano-machinery of cardiac calcium cycling. J Physiol, 2017; 595:3823–3834.

34.

Salvage

, Huang

, Jackson

. Cell-adhesion properties of beta-subunits in the regulation of cardiomyocyte sodium channels. Biomolecules, 2020; 10:989.

35.

Clapham

, Miller

. A thermodynamic framework for understanding temperature sensing by transient receptor potential (TRP) channels. Proc Natl Acad Sci U S A, 2011; 108:19492–19497.

36.

Chowdhury

, Jarecki

, Chanda

. A molecular framework for temperature-dependent gating of ion channels. Cell, 2014; 158:1148–1158.

37.

Gardill

, Rivera-Acevedo

, Tung

, et al. The voltage-gated sodium channel EF-hands form an interaction with the III-IV linker that is disturbed by disease-causing mutations. Sci Rep, 2018; 8:4483.

38.

Abdelsayed

, Peters

, Ruben

. Differential thermosensitivity in mixed syndrome cardiac sodium channel mutants. J Physiol, 2015; 593:4201–4223.

39.

Lau

, Van Petegem

. Crystal structures of wild type and disease mutant forms of the ryanodine receptor SPRY2 domain. Nat Commun, 2014; 5:5397.

40.

Yuchi

, Lau

, Van Petegem

. Disease mutations in the ryanodine receptor central region: Crystal structures of a phosphorylation hot spot domain. Structure, 2012; 20:1201–1211.

41.

Antzelevitch

, Patocskai

. Brugada syndrome: Clinical, genetic, molecular, cellular, and ionic aspects. Curr Probl Cardiol, 2016; 41:7–57.

42.

Amin

, Meregalli

, Bardai

, et al. Fever increases the risk for cardiac arrest in the Brugada syndrome. Ann Intern Med, 2008; 149:216–218.

43.

Bennett

, Yazawa

, Makita

, et al. Molecular mechanism for an inherited cardiac arrhythmia. Nature, 1995; 376:683–685.

44.

Wan

, Chen

, Sadeghpour

, et al. Accelerated inactivation in a mutant Na+ channel associated with idiopathic ventricular fibrillation. Am J Physiol Heart Circ Physiol, 2001; 280:H354–H360.

45.

Shivkumar

, Ajijola

, Anand

, et al. Clinical neurocardiology defining the value of neuroscience-based cardiovascular therapeutics. J Physiol, 2016; 594:3911–3954.

46.

Cortada

, Brugada

, Verges

. Trafficking and function of the voltage-gated sodium channel beta2 subunit. Biomolecules, 2019; 9:604.

47.

Dhar Malhotra

, Chen

, Rivolta

, et al. Characterization of sodium channel alpha- and beta-subunits in rat and mouse cardiac myocytes. Circulation, 2001; 103:1303–1310.

Biophysical Investigation of Sodium Channel Interaction with β-Subunit Variants Associated with Arrhythmias

Abstract

Background:

Materials and Methods:

Results:

Conclusion:

Introduction

Methods

Electrophysiological recordings in Xenopus laevis oocytes

Qualitative surface biochemistry

Statistical analyses

Recombinant protein production and crystallographic analysis

Crystallization, data collection, and structure solution

Circular dichroism spectroscopy

Results

Electrophysiological analysis

Crystallographic analysis

Protein stability

Mapping β-subunit variants on cryoelectron microscopy structures

NaV1.2 gating and β2R137H

Discussion

Footnotes

Author Contributions

Acknowledgments

Disclaimer

Author Disclosure Statement

Funding Information

References

Na_V1.2 gating and β2^R137H