Differential regulation of cardiac sodium channels by intracellular fibroblast growth factors

In the heart, voltage-gated sodium (Na_V) channels are responsible for the rapid upstroke of action potentials (APs) in atrial and ventricular myocytes, and in Purkinje fibers (Nerbonne and Kass, 2005). The predominant Na_V channel pore-forming (α) subunit responsible for generating these channels in the mammalian myocardium is Na_V1.5 (Abriel and Kass, 2005). Precise regulation of Na_V1.5 gating is essential for the proper electromechanical functioning of the heart and the disruption of gating in acquired and inherited cardiac diseases predisposes individuals to potentially life-threatening arrhythmias (Noble and Noble, 2006; Mangold et al, 2017; Ton et al, 2021). Mutations in SCN5A, the gene which encodes Na_V1.5, for example, are linked to congenital cardiac arrhythmia syndromes, including the long QT type 3 (LQT3) and Brugada syndromes (Ruan et al., 2009). Similar to other Na_V α-subunits, Na_V1.5 has four homologous repeats (I–IV), each of which contains six α-helical transmembrane segments (S1–S6; Yu and Catterall, 2003). The S1–S4 segments in each repeat form the voltage sensor domains VSD-I through VSD-IV, and the S5–S6 segments in the four repeats combine to constitute the Na⁺-selective pore. The activation of each VSD on membrane depolarization is coupled to conformational changes in the channel, leading to a rapid opening of the activation gate, followed closely by channel inactivation (Varga et al., 2015).

Native myocardial Na_V channels function in macromolecular complexes, comprising the Na_V1.5 α-subunit and various auxiliary subunits, including Na_V β-subunits (Calhoun and Isom, 2014), calmodulin (CaM; Gardill et al, 2019), and intracellular fibroblast growth factors (iFGFs; Yang et al., 2016; Pitt and Lee, 2016) that influence channel stability, trafficking, and biophysical properties (Abriel and Kass, 2005; Meadows and Isom, 2005; Abriel, 2010). It has been previously demonstrated that the auxiliary β-subunits differentially regulate the expression, gating, and pharmacological sensitivity of Na_V channels (Dhar Malhotra et al, 2001; Zhu et al, 2017; Angsutararux et al, 2021b). Although less well-studied, iFGFs have been shown to bind to the C-terminal domain of Na_V α-subunits, including Na_V1.5, and to modulate the time- and voltage-dependent properties of heterologously expressed (Liu et al., 2001; Liu et al., 2003) and native (Pablo and Pitt, 2017; Goldfarb et al., 2007) Na_V1.5-encoded currents.

The iFGFs, also known as fibroblast growth factor homologous factors (FHFs), are a subfamily of FGFs, consisting of iFGF11–iFGF14 that have similar “core” sequences and distinct N-termini (Olsen et al., 2003; Goldfarb, 2005). Alternative exon usage and differential N-terminal splicing generate further iFGF protein diversity (Munoz-Sanjuan et al., 2000; Pablo and Pitt, 2017). The crystal structure of the Na_V1.5 C-terminus in complex with iFGF13 and CaM revealed that the binding of iFGF13 is mediated by amino acids in the core domain (Wang et al., 2012; Musa et al., 2015; Hennessey et al., 2013). The earlier structure of iFGF12 alone (Goetz et al., 2009) and the later structure of iFGF12B in complex with Na_V1.5 C-terminus and CaM (Wang et al., 2014) reveal iFGF12 structures nearly identical to iFGF13, especially in the core region. Although these findings suggest that all iFGFs might modulate the properties of Na_V channels encoded by a given Na_V α-subunit (such as Na_V1.5) similarly, this is not the case. Indeed, various iFGF proteins have been shown to exert dramatically different effects on the gating of Na_V channels that depend on both the subtype of Na_V α-subunit and the specific isoform of iFGF (Munoz-Sanjuan et al., 2000; Liu et al., 2003; Goetz et al., 2009; Wang et al., 2011b; Chakouri et al., 2022). For example, three iFGF13 splice variants, iFGF13S (iFGF13A), iFGF13U (iFGF13B), and iFGF13VY, were shown to produce distinct effects on the gating of Na_V1.5 channels (Yang et al., 2016). Two iFGF14 isoforms, iFGF14A and iFGF14B, have also been shown to differentially modulate Na_V1.5-encoded currents and to produce functional effects markedly different from iFGF13-mediated effects (Lou et al., 2005).

The most prominent iFGF expressed in the human heart, iFGF12B, has been linked to inherited cardiac arrhythmias (Hennessey et al., 2013; Li et al., 2017). In addition, LQT3 and Brugada syndrome-linked mutations in Na_V1.5 have been shown to disrupt iFGF binding and/or to modify iFGF-mediated effects on the properties of Na_V1.5-encoded currents (Musa et al., 2015; Liu et al., 2003). The mechanisms underlying iFGF12B-mediated regulation of cardiac Na_V1.5 channels, however, have not been defined. The experiments here were designed to address this knowledge gap and, in addition, to directly compare the effects of iFGF12B on cardiac Na_V1.5 channels with those of the extensively studied iFGF13VY, the predominant iFGF expressed in rodent ventricles (Wang et al., 2011a; Park et al., 2016).

Adult (8–20-wk-old) male and female wild-type (WT), Fgf12KO, Fgf13 floxed, and cFgf13KO C57BL/6J mice were used in the experiments here. The Fgf12KO line (Goldfarb et al., 2007) was obtained from Drs. Mitchell Goldfarb (City University of New York, New York, NY) and David Ornitz (Washington University School of Medicine, St. Louis, MO). The Fgf13 floxed and cFgf13KO lines were generated as described below. The transgenic C57BL/6J mouse line (Tg(Myh6-cre)2182Mds), expressing Cre-recombinase driven by the cardiac-specific α-myosin heavy chain (α-MHC) promoter, was purchased from The Jackson Laboratory. Adult female Xenopus laevis of sizes >9 cm were obtained from Xenopus 1 Corp, and oocytes were harvested using previously described methods (Varga et al., 2015). All animals were handled in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals, and all experimental protocols were approved by the Washington University Institutional Animal Care and Use Committee (IACUC).

To enable conditional deletion of Fgf13, mice in which the Fgf13 locus (on the X chromosome) was floxed were generated by the Genome Editing and Stem Cell (GESC) Center in the McDonnell Genome Institute and the Mouse Genetics Core at Washington University using CRISPR-Cas9 gene editing technology (Doudna and Charpentier, 2014), modified to provide robust and reliable generation of floxed alleles (Sentmanat et al., 2022). Briefly, guide RNAs (gRNAs) and single-stranded oligodeoxynucleotides (ssODNs) were designed to introduce a 5′-side LoxP (5′ LoxP) site into the intron upstream of exon 2 and a 3′-side LoxP (3′ LoxP) site into the intron downstream of exon three of the Fgf13 gene (Fig. S1 A). These gRNAs and ssODNS, together with Cas9, were electroporated into single-cell C57BL/6J embryos, and these embryos were transferred into the oviducts of pseudopregnant females, ∼20 embryos per recipient.

Following preliminary screening of the founder (F0 generation) pups, proper insertion and localization of the loxP sites were confirmed by sequencing. To ensure germline transmission and that both loxP sites were on the same allele, F0 animals positive for both loxP sites were crossed to WT C57BL/6J animals and the resulting (F1 generation) offspring were also screened by sequencing. Heterozygous Fgf13 floxed (Fgf13^fl/+) female and hemizygous Fgf13 floxed (Fgf13^fl/y) male (F1) offspring were mated to produce (F2) females homozygous for the floxed Fgf13 locus, Fgf13^fl/fl (Fig. S1 A). The Fgf13^fl/fl and Fgf13^fl/y animals were then crossed with transgenic animals (Tg(Myh6-cre)2182Mds) expressing Cre-recombinase driven by the (cardiac specific) α-MHC promoter (Fig. S1 A). Crossing Cre-recombinase-expressing hemizygous male (Fgf13^fl/y) offspring with Fgf13^fl/fl females (or Cre-recombinase-expressing heterozygous Fgf13^fl/+ females with Fgf13^fl/y males) provided cardiac-specific Fgf13 targeted deletion hemizygous male (cFgf1^−/y) and homozygous female (cFgf13^−l−) animals, referred to here collectively as cardiac-specific Fgf13 knockouts, cFgf13KO. Offspring (from this and subsequent crosses) were screened by PCR using the primers given in Table S1 A, and representative results are illustrated in Fig. S1 B.

To confirm the elimination of the iFGF13 protein in the myocardium of cFgf13KO animals, protein lysates were prepared from WT and cFgf13KO left ventricles (LVs). Following protein fractionation and transfer, membranes were probed with a polyclonal anti-FGF13 antibody (generously provided by Dr. Geoffrey Pitt, Weill Cornell Medical School, New York, NY). As illustrated in Fig. S1 C (left panel), two bands, at ∼32 and ∼24 kD, were detected with the anti-iFGF13 antibody in the WT LV samples, whereas these bands were not detected in the cFgf13KO LV samples, consistent with the elimination of the iFGF13 protein. Similar analyses, conducted on LV protein lysates prepared from Fgf12KO animals, revealed iFGF13 protein expression indistinguishable from that observed in WT LV (Fig. S1 C, right panel).

AAV9 vectors were prepared by the Hope Center Viral Vectors Core at Washington University, as described previously (Zolotukhin et al, 2002). Briefly, HEK-293 cells, maintained in Dulbecco’s modified Eagle’s medium (DMEM), supplemented with 5% fetal bovine serum (FBS), 100 U/ml penicillin, and 100 µg/ml streptomycin at 37°C in a 95% air/5% CO₂ incubator, were plated at 30–40% confluence in CellSTACS (Corning). Approximately 24 h later, cells were cotransfected with 0.6 mg of the AAV transfer plasmid containing the construct of interest (e.g., human iFGF12B, hFGF12B) and 1.8 mg of a helper plasmid (pXYZ1) using the calcium phosphate precipitation method (Zolotukhin et al, 2002). The transfection medium was removed after 6 h, and the cells were incubated at 37°C for 3 d. For harvesting, cells were lysed by three freeze/thaw cycles and the cell lysates were collected and treated with 50 U/ml of Benzonaze, followed by iodixanol gradient centrifugation. The iodixanol gradient fraction was further purified on a HiTrap Q column and concentrated (Zolotukhin et al, 2002). The vector titer of hFGF12B, determined by Dot blot assay, was 3.8 × 10¹³ vg/ml (viral genomes/ml).

Adult (8–12 wk) male and female WT and cFgf13KO C57BL/6J mice were anesthetized by intraperitoneal (ip) injection of 1 ml/kg of a ketamine/xylene (30 mg/ml/4 mg/ml) cocktail; the body temperature was maintained at 37°C with a feedback-controlled heating pad. Each animal was placed on its left side with the head facing to the right, and virus injections were made into the right retro-orbital sinus. Gentle downward pressure was applied to the skin dorsal and ventral to the eye, making the eye protrude slightly. Retro-orbital injection of a 1:1 mixture of the hFGF12B-expressing (20 μl) and the eGFP-expressing AAV9 (20 μl) viruses was then made. For injections, the virus solution (20–40 μl) was drawn up into an (0.5 ml) insulin syringe and the (12.7 mm) syringe needle was then placed at the medial canthus of the eye at an angle of ∼30° and with the bevel facing down. After inserting the needle until it reached the back of the orbit, the virus solution was injected (10 μl/min) slowly. The needle was then gently and slowly removed to prevent damage to the eye. The grips around the eye were then released, the eyelid was closed, and mild pressure was applied to the injection site for a few seconds. Animals were monitored for ∼2 h to ensure that there were no unexpected deleterious effects of the anesthesia or the injections and were then returned to their home cages.

To confirm the expression of iFGF12 and determine the time course of eGFP/iFGF12 expression in the ventricles of virus-injected animals, protein lysates were prepared from cFgf13KO LV at 2, 3, and 4 wk following the injections of the hFGF12B-expressing and eGFP-expressing AAV9 viruses. Following protein fractionation and transfer, membranes were probed with a polyclonal anti-eGFP antibody (Fig. S4 A) or with a polyclonal anti-iFGF12 antibody (Fig. S4 B). The anti-iFGF12 antibody was validated using protein lysates prepared from WT and Fgf12KO adult mouse left and right atria. A prominent band at ∼20 kD was detected with the anti-iFGF12 antibody in the WT, but not in the Fgf12KO, atrial protein samples (Fig. S4 C), consistent with the elimination of iFGF12 proteins and validating the anti-FGF12 antibody for Western blot analyses of iFGF12 protein expression.

cDNAs encoding human SCN5A, Fgf12B, and Fgf13VY were produced from pMAX and pBSTA vectors accordingly. Point mutations were made in the SCN5A gene using the QuikChange II site-directed mutagenesis kit (Agilent) with primers from Sigma-Aldrich containing the mutation. The DNA constructs of the iFGF chimeras were synthesized by GeneArt Gene Synthesis (Thermo Fisher Scientific). All mutations and chimeras were confirmed by sequencing. Complementary RNAs (cRNAs) were synthesized with the mMessage mMACHINE T7 Transcription kit (Life Technologies) after linearizing the corresponding cDNAs with the appropriate restriction enzyme and purifying with the NucleoSpin Gel and PCR Clean-up kit (Macherey-Nagel). Each cRNA was reconstituted with double-distilled H₂O at a concentration of ∼1 µg/μl.

Myocytes were isolated from the LV of adult (10–20-wk-old) male and female WT, Fgf13 floxed (female Fgf13^fl/fl and male Fgf13^fl/y), cFgf13KO (female cFgf13^−/− and male cFgf13^−/y), and Fgf12KO mice by enzymatic and mechanical dissociation using previously described methods (Xu et al. 1999; Brunet et al., 2004). In addition, LV myocytes were isolated from cFgf13KO mice (4–6 wk) after retro-orbital (eGFP- + FGF12B-expressing) AAV9 injections. Briefly, hearts were quickly removed from avertin-anesthetized mice and perfused retrogradely through the aorta with collagenase-containing solution (Type II, Worthington) at 37°C. After 15–20 min perfusion, the LV was separated, minced, and dispersed by gentle trituration. The resulting cell suspension was filtered and resuspended in serum-free medium-199 (M-199; Sigma-Aldrich). Isolated myocytes were plated on laminin-coated coverslips and maintained in a 95% air-5% CO₂ incubator at 37°C until used (within 24–48 h) in electrophysiological experiments.

Whole-cell voltage-clamp recordings were obtained from adult mouse LV myocytes within 24–36 h of isolation at room temperature (22–24°C) using an Axopatch 1D (Axon Instruments) amplifier interfaced with a Digidata 1332 data acquisition system (Axon Instruments) and the pClamp 10 (Axon Instruments) software to a Dell computer. Recording pipettes contained (in mM) 90 CsCH₃O₃S, 20 CsCl, 5 NaCl, 5 MgATP, 0.4 TrisGTP, 10 EGTA, and 10 HEPES (pH 7.3; 300 mOsm). Pipette resistances were routinely 1.5–3.0 MΩ when filled with the recording solution. The bath solution contained (in mM) 20 NaCl, 65 CsCl, 50 TEACl; 2 MgCl2, 1 CaCl₂, 0.5 CdCl2, 10 HEPES, and 10 glucose (pH 7.3; 310 mOsm).

Electrophysiological data were acquired at 100 KHz and signals were low-pass filtered at 5 kHz prior to digitization and storage. After the formation of a gigaohm-seal (>1 GΩ) and establishment of the whole-cell configuration, brief (10 ms) ± 10 mV voltage steps from the holding potential (HP) of −90 mV were presented to allow measurements of whole-cell membrane capacitances (C_m), input resistances (R_in), and series resistances (R_s). In each cell, C_m and R_s were compensated electronically by ∼85%; voltage errors resulting from uncompensated series resistances were always <2 mV and were not corrected. Leak currents were always <50 pA and were not corrected.

To determine the voltage-dependences of steady-state inactivation of I_Na in WT, Fgf13 floxed, cFgf13KO, Fgf12KO, and eGFP- + hFGF12B-expressing cFgf13KO LV myocytes, whole-cell Na_V currents, evoked at −30 mV following brief (25 ms) conditioning voltage steps (ranging from −130 to −15 mV in 5 mV increments), presented from a HP of −90 mV, were recorded. Peak I_Na, evoked at −30 mV from each conditioning potential in each cell, was measured and normalized to the maximal peak current (I_Na,max) evoked (in the same cell) from the most hyperpolarized conditioning potential (of −130 mV). Mean ± SEM normalized current amplitudes (I_Na/I_Na,max), determined in WT, Fgf13 floxed, cFgf13KO, Fgf12KO, and hFGF12B- + eGFP-expressing cFgf13KO LV myocytes, were then plotted as a function of the conditioning membrane potential (V_m) and fitted using the Boltzmann equation (Eq. 1).

Harvested oocytes were then digested into single cells with collagenase (Sigma-Aldrich). Oocytes were injected with cRNAs and incubated in ND93 solution (93 mM NaCl, 5 mM KCl, 1.8 mM CaCl₂, 1 mM MgCl₂, 5 mM HEPES, 2.5 mM Na pyruvate, and 1% penicillin/streptomycin, pH 7.4) at 18°C for 3–6 d before electrophysiological recordings. Each oocyte was injected with 50–56 ng of either SCN5A cRNA alone or a combination of SCN5A and Fgf12B/Fgf13VY cRNAs at a 4:1 M ratio for coexpression.

Cut-open voltage-clamp recordings were performed at 19°C, maintained with a temperature controller (HCC-100A; Dagan Corporation), using a cut-open amplifier (CA-1B; Dagan Corporation), interfaced to a computer using an A/D converter (Digidata 1440; Molecular Devices) and the pClamp version 10 software (Molecular Devices). Data were collected using Clampex (Molecular Devices) and analyzed using Clampfit 10.7 (Molecular Devices). The internal recording solution was composed of 105 mM N-methyl-D-glucamine (NMDG), 10 mM 2-(N-Morpholino)ethanesulfonic acid (MES) sodium salt (Na-MES), 20 mM HEPES, and 2 mM EGTA, at a pH level of 7.4, and the external recording solution contained 25 mM NMDG, 90 mM Na-MES, 20 mM HEPES, and 2 mM Ca-MES₂, at a pH level of 7.4.

For VCF, oocytes were labeled for 30 min on ice with 10 µmol/liter of methanethiosulfonate–carboxytetramethylrhodamine (MTS-TAMRA; Santa Cruz Biotechnology) in a depolarizing solution composed of 110 mM KCl, 1.5 mM MgCl₂, 0.8 mM CaCl₂, 0.2 mM EDTA, and 10 mM HEPES at a pH of 7.1. Simultaneous recordings of ionic current and fluorescence emissions were collected on a custom rig as described previously (Rudokas et al., 2014; Varga et al., 2015; Zhu et al., 2017). After teal LED light illumination by the SPECTRA X (Lumencor), fluorescence emission was measured by a photodiode (PIN-040A; United Detector Technology) and a patch-clamp amplifier (Axopatch-200A; Molecular Devices).

Voltage-clamp records were analyzed with Clampfit (Molecular Devices) and Excel (Microsoft). Steady-state activation curves were generated from currents recorded during 100-ms voltage steps to various test potentials (−120 to 60 mV) from a HP of −120 mV. The calculated conductance (G) was normalized to the maximum conductance at 20 mV. For steady-state inactivation curves, cells were preconditioned at voltages ranging from −150 to 20 mV for 200 ms and the channel availability at −20 mV was then determined. The normalized conductance-voltage (G-V) curves were fitted with a Boltzmann equation (Eq. 1). The time constants describing the fast inactivation kinetics of the currents were calculated by fitting the peak current decay to a sum of two exponentials function (Eq. 2).

Recovery from inactivation was determined by a double-pulse protocol with varying recovery durations between two pulses. The first depolarizing pulse was applied at −20 mV for 200 ms to induce fast inactivation. The second pulse, applied at −20 mV for 20 ms, was used to test channel recovery. The recovery duration was measured from 1 to 1,000 ms, and the normalized recovery curve was fitted to a double exponential equation (Eq. 3).

Electrophysiological data were analyzed using Clampfit (Molecular Devices) and Prism 9 (Graph Pad). The statistical significance of apparent differences between/among data sets was evaluated with Welch ANOVA, followed by Dunnett’s post-hoc test. Comparisons between the two groups were done using the unpaired Welch’s t test. Pearson correlation analysis was used to determine correlation coefficients (r), together with the coefficients of determination (R²). Multiple linear regression analysis was applied to test the relationship between parameters of different measurements.

Protein lysates were prepared from WT, Fgf12KO, cFgf13KO, and hFGF12B- + eGFP-expressing cFgf13KO ventricles and from Fgf12KO atria in 20 mM HEPES + 150 mM NaCl buffer with 0.5% CHAPS and a protease inhibitor tablet (Roche) using previously described methods (Marionneau et al., 2008). Lysates were fractionated by SDS-PAGE (4–15% gradient), transferred to polyvinylidene fluoride (PVDF) membranes (Biorad), and probed for: iFGF13 expression using a rabbit polyclonal anti-iFGF13 antibody (generous gift of Dr. Geoffrey Pitt, Weill Cornell Medical School); eGFP expression using a rabbit polyclonal anti-eGFP antibody (catalog number AB3080; Sigma-Aldrich); or iFGF12 expression using a rabbit polyclonal anti-iFGF12 antibody (catalog number SAB2700759; Sigma-Aldrich). Blots were also probed with a mouse monoclonal anti-tubulin (1:10,000; Abcam) antibody to verify equal protein loading of the gel lanes.

Quantitative RT-PCR analyses were completed using previously described methods (Marionneau et al., 2008). Briefly, for the preparation and analyses of RNA, adult (9–10 wk) WT (n = 6), cFgf13^−/− (n = 6), and Fgf12^−/− (n = 6) animals were sacrificed by cervical dislocation under isoflurane anesthesia and the hearts were rapidly removed. LVs were dissected and snap-frozen in liquid nitrogen. After homogenization, total RNA was isolated using the TRIzol Reagent (Invitrogen), and DNase was treated with the RNeasy Fibrous Tissue Mini Kit (Qiagen). The concentration of total RNA in each sample was measured spectrophotometrically using a NanoDrop ND-1000 (NanoDrop Technologies). RNA quality was examined using gel electrophoresis, and genomic DNA contamination was assessed by PCR amplification of total RNA without prior cDNA synthesis; no genomic DNA was detected.

First-strand cDNA was synthesized from 2 μg of total RNA using the High-Capacity cDNA Archive Kit (Applied Biosystems). The expression levels of the Fgf13VY, Fgf13S(A), Fgf13U(B), Fgf12B, Fgf12A, Fgf11, Fgf14A, and Fgf14B transcripts were determined by quantitative real-time RT-PCR using SYBR Green PCR Master Mix (Applied Biosystems). PCR reactions were performed on 10 ng of cDNA using sequence-specific primer pairs (Table S1 B) and the ABI PRISM 7900HT Sequence Detection System. The cycling conditions included a hot start at 95°C for 10 min, followed by 40 cycles at 95°C for 15 s, and 60°C for 1 min. All primer pairs were tested using mouse cDNA as the template, and templates giving 90–100% efficiency were chosen. In all cases, a single amplicon of the appropriate melting temperature or size was detected using the dissociation curve or gel electrophoresis. Data were collected with instrument spectral compensations using the Applied Biosystems SDS 2.2.2 software and analyzed using the comparative threshold cycle (C_T) relative quantification method (Schmittgen and Livak, 2008). The hypoxanthine-guanine phosphoribosyl-transferase gene (HPRT) was used as an endogenous control to normalize the data (deKok et al., 2005). Individual sample measurements (n = 6) were averaged and 2Δ^CT values for each gene, corresponding to the relative expression level of that gene compared with HPRT, were calculated and are reported here; all C_T values were <30. Negative control experiments using RNA samples incubated without reverse transcriptase during cDNA synthesis showed no amplification.

Fig. S1 illustrates the generation and validation of cardiac-specific deletion of Fgf13 in mice. Fig. S2 shows the measurement of Fgf transcripts expression in WT, cFgf13KO, and Fgf12KO adult mouse LVs. Fig. S3 shows the amplitudes/densities of I_Na in cFgf13KO and Fgf12KO mouse LV myocytes. Fig. S4 shows the time dependence of AAV9-mediated expression of iFGF12B and eGFP in adult mouse LV. Fig. S5 portrays results from the iFGF chimeras experiment switching the N-terminal domains between iFGF12B and iFGF13VY. Fig. S6 demonstrates the effects of multiple iFGF chimeras on the voltage dependence of VSD-IV activation. Fig. S7 illustrates additional correlation and linear regression analysis between the VSD-IV activation and I_Na gating kinetics. Table S1 lists primers used in screening Fgf13 floxed (Fgf13^fl/+ and Fgf13^fl/fl) and cFgf13KO mice. Table S2 lists biexponential fits to the decay phases of I_Na recorded at various test potentials from LV myocytes isolated from WT, cFgf13KO, and cFgf13KO + iFGF12B mice. Table S3 lists biexponential fits to the decay phases of I_Na evoked at various test potentials from oocytes expressing Na_V1.5 alone or Na_V1.5 combined with iFGF12B or iFGF13VY. Table S4 lists properties of I_Na recorded from Xenopus oocytes expressing Na_V1.5 in combination with iFGF chimeras.

To generate mice (cFgf13KO) with cardiac-specific deletion of Fgf13, we floxed the Fgf13 locus, using CRISPR-Cas 9 gene editing technology, and crossed these mice with transgenic mice expressing Cre-recombinase driven by the cardiac-specific α-MHC promoter as described in Materials and methods (Fig. S1, A and B). The loss of the Fgf13 transcript and the iFGF13 protein in the heart was confirmed by quantitative RT-PCR (Fig. S2) and Western blot analyses (Fig. S1 C, left; and Fig. S2). Whole-cell voltage-clamp recordings revealed that the waveforms and the densities of the voltage-gated I_Na in LV myocytes isolated from WT and cardiac-specific Fgf13 deletion (cFgf13KO) mice are remarkably similar (Fig. 1 A). Quantitative analyses of peak I_Na densities revealed no significant differences in Fgf13 floxed, compared with WT (P values >0.10 at all voltages), or in cFgf13KO, compared with Fgf13 floxed or WT (P values >0.15 at all voltages; Fig. S3 A). Analysis of the voltage-dependences of steady-state activation and inactivation of I_Na in WT and cFgf13KO LV myocytes revealed a hyperpolarizing shift in the voltage at which half the channels are inactivated (V_1/2, _inact), whereas the voltages at which half the channels are activated (V_1/2, _act) in WT and cFgf13KO LV myocytes are indistinguishable (Fig. 1 B and Table 1). The finding that steady-state inactivation of I_Na is selectively affected in cFgf13KO LV myocytes is consistent with previous studies with acute, shRNA-mediated, knockdown of Fgf13 or the targeted deletion of Fgf13 (Wang et al., 2011a; Park et al., 2016; Wang et al., 2017; Park et al., 2020).

Additional experiments revealed that the kinetics of I_Na recovery from inactivation, determined from measurements of the normalized peak I_Na amplitudes as a function of the recovery interval as described in Materials and methods, are similar in WT and cFgf13KO LV myocytes (Fig. 1 C and Table 1). The records in Fig. 1 A and the overlay of I_Na recordings evoked at −20 mV (Fig. 1 D, inset), however, reveal that the conditional knockout of Fgf13 accelerated peak I_Na decay. In WT LV myocytes, the decay phases of the currents were best fitted with a sum of two exponentials, providing the fast (Ƭ_f) and slow (Ƭ_s) time constants of peak I_Na decay, corresponding to the fast and intermediate modes of open-state inactivation (Silva, 2014; Silva and Goldstein, 2013a; Silva and Goldstein, 2013b). Analyses of I_Na in cFgf13KO LV myocytes, however, revealed that the decay phases of the currents in many (13 of 26) cFgf13KO cells were not well-fitted by the sum of two exponentials. Rather, the decay phases of I_Na in these cells were well-described by single exponential with Ƭ values similar to the fast component (Ƭ_f) determined in WT cells. For all cells in which two exponentials were required, Ƭ_f and Ƭ_s were similar to the values determined in WT LV myocytes (Fig. 1 D and Table S2), although the relative amplitude of the slow component of I_Na decay (%Aslow) in the (13) cFgf13KO LV myocytes with two decay components was much lower than in WT LV myocytes (Fig. 1 E and Table S2). The apparent acceleration of I_Na decay in cFgf13KO LV myocytes (Fig. 1, A and D, inset), therefore, reflects a reduction (or, in many cells, elimination) of the slow component of current decay. Expression of iFGF13VY, therefore, reduces the fast component of I_Na decay.

Additional experiments revealed that the amplitudes and properties of I_Na recorded from LV myocytes isolated from Fgf12KO (Fig. S3, B and C) mice were indistinguishable from the currents in WT cells (Table 1), consistent with the negligible expression of Fgf12 in adult mouse ventricles (Fig. S2).

To examine the effects of iFGF12B on native I_Na, we generated an iFGF12B-expressing AAV9 and delivered this virus, mixed at a 1:1 ratio with eGFP-expressing AAV9, into cFgf13KO mice by retro-orbital sinus injection (see Materials and methods and Fig. S4). Representative recordings of I_Na from control cFgf13KO LV myocytes and an iFGF12B- + eGFP-expressing cFgf13KO LV myocytes are presented in Fig. 2 A. As illustrated in Fig. 2 B, the expression of iFGF12B shifted the voltage-dependence of I_Na inactivation in the depolarizing direction relative to the results in cFgf13KO myocytes, i.e., back toward the WT G-V plot (Table 1). In addition, however, iFGF12B expression shifted the activation curve for I_Na, also in the depolarizing direction relative to the G-V curve for cFgf13KO and WT (Fig. 2 B) LV myocytes. The addition of iFGF12B did not alter the amplitudes of I_Na when compared to I_Na recorded from cFgf13KO (or WT) myocytes (Fig. 2 A). The addition of iFGF12B also accelerated the kinetics of I_Na recovery from inactivation, relative to the data in cFgf13KO myocytes (Fig. 2 C and Table 1). Analyses of the waveforms of the currents in cFgf13KO LV myocytes expressing iFGF12B, however, revealed that, similar to cFgf13KO LV myocytes, the decay phases of the currents in many (13 of 19) cells were best described by single exponentials characterized by Ƭ values similar to Ƭ_f measured in WT and cFgf13KO LV myocytes (Table S2). For the remaining (6 of 19) cFgf13KO LV myocytes expressing iFGF12B, I_Na was well-described by the sum of two exponentials with Ƭ_f and Ƭ_s values and %A_slow similar to those measured in cFgf13KO LV myocytes (Fig. 2, D and E; and Table S2). In contrast with iFGF13VY, therefore, iFGF12B does not measurably affect the fast or the slow component of decay of mouse ventricular I_Na.

The results presented above demonstrate that iFGF13VY and iFGF12B have distinct effects on the time- and voltage-dependent properties of native (mouse) myocardial I_Na. To explore the molecular mechanisms underlying these differences, we turned to a heterologous expression system (Xenopus oocytes) and obtained voltage-clamp current and VCF recordings from oocytes expressing Na_V1.5 alone or in combination with iFGF12B or iFGF13VY.

In subsequent experiments, Na_V1.5-encoded currents were recorded from Xenopus oocytes expressing Na_V1.5 alone or Na_V1.5 combined with iFGF12B (Fig. 3 A). In comparison to Na_V1.5 alone, iFGF12B coexpression shifted the steady-state activation and inactivation curves in the depolarizing direction (Fig. 3 B and Table 2). In the presence of iFGF12B, therefore, Na_V1.5 channels begin to open at more depolarized membrane potentials. The rate of recovery of I_Na from inactivation was also faster in the presence of iFGF12B (Fig. 3 C and Table 2).

Superimposing recordings of I_Na evoked by a −20-mV voltage step from oocytes expressing Na_V1.5 alone or Na_V1.5 with iFGF12B revealed that current decay was accelerated with iFGF12B coexpression (Fig. 3 D, inset). Analyses of the kinetics of peak I_Na decay revealed that the data were best fit to the sum of two exponentials, and the measured time constants (Ƭ_f and Ƭ_s) were similar for the currents produced by Na_V1.5 alone and Na_V1.5 combined with iFGF12B (Fig. 3 D and Table S3). The fractional amplitudes of the slower component of inactivation (%A_slow), however, were markedly reduced in the presence of iFGF12B (Fig. 3 E and Table S3). iFGF12B, therefore, enhanced the faster component of inactivation of heterologously expressed (human) Na_V1.5.

Next, we employed VCF to probe the molecular mechanism(s) contributing to iFGF12B modulation of Na_V1.5 channel gating (Mannuzzu et al., 1996; Stefani and Bezanilla, 1998; Gandhi and Olcese. 2008; Rudokas et al., 2014). A cysteine mutation was introduced into the extracellular linker S3–S4 in each VSD for fluorophore (MTS-TAMRA) tagging (Varga et al., 2015; Zhu et al., 2016). The movement of S4 upon membrane depolarization can then be detected by a change in fluorescence emission, reflecting an altered surrounding environment (Varga et al., 2015; Zhu et al., 2016). To improve sensitivity, a cysteine mutation was created in the background of WT-LFS (large fluorescence signal) bearing Y1977A-C373Y mutations that ablate a ubiquitination site and prevent non-specific labeling (Varga et al., 2015). Previous work showed that the LFS mutations do not substantially affect Na_V channel activation or inactivation kinetics (Varga et al., 2015). Plotting changes in the fluorescence signals as a function of voltage (i.e., F versus V curves) for each VSD allows estimation of the conformational changes that occur during VSD activation.

As illustrated in Fig. 4, coexpression of iFGF12B caused no changes in the half-maximal voltages (V_1/2) of VSD-I, VSD-II, or VSD-III F-V curves compared with Na_V1.5 expressed alone. For VSD-III, however, we observed a change at potentials greater than −50 mV. Multiple lines of evidence suggest that VSD-III can activate in two discrete steps (Chanda and Bezanilla, 2002; Hsu et al., 2017; Zhu et al., 2017). iFGF12B did not affect the initial activation at very hyperpolarized potentials. However, the second translation of VSD-III was modulated by iFGF12B, as seen by the deflection at higher potentials (Fig. 4). This second VSD-III activation has been proposed to regulate the open-state inactivation (Angsutararux et al., 2021a).

iFGF12B coexpression induced marked changes in the F-V curve of VSD-IV F-V, specifically causing a hyperpolarizing shift in V_1/2 and a decrease in the slope factor (k; Fig. 4 and Table 3). The alterations in the F-V curve in the presence of iFGF12B reflected the facilitated activation of VSD-IV and its accelerated outward translation, resulting in its fully activated conformation at less depolarized potentials. Together, these results suggest that the facilitation of VSD-IV activation underlies the observed effects of iFGF12B on the gating properties and kinetics of I_Na inactivation.

Next, we investigated the effect of iFGF13VY on Na_V1.5. Analysis of the ionic currents (Fig. 5 A) produced on the expression of Na_V1.5 alone and Na_V1.5 coexpressed with iFGF13VY revealed a shift in the steady-state inactivation curve toward depolarizing potentials in the presence of iFGF13VY, with no effect on the voltage-dependence of current activation (Fig. 5 B and Table 2). The depolarizing shift in the V_1/2 for inactivation seen with iFGF13VY (Fig. 5 B and Table 2) was substantially larger than the shift seen with coexpression of iFGF12B (Fig. 3 B and Table 2). Recovery from inactivation was also facilitated by iFGF13VY (Fig. 5 C and Table 2). Compared to Na_V1.5 alone, we also found that iFGF13VY coexpression accelerates the apparent rate of current decay (Fig. 5 D, inset) by reducing %A_slow (Fig. 5 E), as Ƭ_f and Ƭ_s are unchanged in the presence of iFGF13VY (Fig. 5 D and Table S3).

Examination of the activation of the individual VSDs revealed marked effects of iFGF13VY on the F-V curves for VSD-I and VSD-IV (Fig. 6 and Table 3). The VSD-I F-V curve was shifted toward positive potentials with iFGF13VY coexpression (Fig. 6), an effect that is not observed with iFGF12B (Fig. 4). iFGF13VY coexpression also increases the steepness (slope) of the F-V curve for VSD-IV and results in a hyperpolarizing shift in the V_1/2. The modulatory effects of iFGF13VY on VSD-IV activation (Fig. 6), therefore, are distinct from the effect of iFGF12B (Fig. 4). iFGF13VY markedly increased the slope of the F-V curve (reduced slope factor k), whereas iFGF12B shifted the V_1/2 in the hyperpolarizing direction.

Like other FGFs, the iFGFs consist of three main domains: the N-terminus, the β-trefoil core, and the C-terminus. The core domains of iFGF11–iFGF14 are highly homologous (Olsen et al., 2003). According to resolved crystal structures, the amino acid residues that mediate iFGF interaction with the C-termini of Na_V α subunits are in the core domain (Goetz et al., 2009; Wang et al., 2012; Musa et al., 2015; Hennessey et al., 2013). If the same residues are involved in the binding of iFGF12B and iFGF13VY to Na_V1.5, the distinct functional effects on Na_V channel gating must reflect interactions with other parts of these proteins which are distinct. Alignment of the iFGF12B and iFGF13VY proteins reveals the prominent difference in the N-termini of these two proteins (Fig. 7 A). Previous studies have shown that the unique N-terminal sequences of iFGF14 splice variant N-termini result in distinct effects on Na_V channel gating (Lou et al, 2005; Laezza et al, 2009; Barbosa et al, 2017; Ali et al, 2018; Singh et al, 2020, 2021). Similarly, the differing N-terminal sequences of the five splice variants of iFGF13 (iFGF13S(A), iFGF13U(B), iFGF13V, iFGF13Y, and iFGF13VY) result in distinct regulatory effects on Na_V1.5 channel gating (Wang et al., 2017). In addition, it has been reported that the specific sequence of iFGF A variants, including iFGF12A, can inhibit the late sodium current, which is not observed with other splice variants, including iFGF12B (Chakouri et al., 2022). Together, these studies point to the significance of the iFGF N-termini in determining the specific modulatory effects of the different variants.

To determine whether the differences we observed in the effects of iFGF12B and iFGF13VY on Na_V1.5-encoded currents are the result of the distinct N-termini of these two proteins, we created two iFGF chimeras by swapping the N-terminal domains (NTD) of iFGF12B (aa1–4) and iFGF13VY (aa1–72), and iFGF12B/13 (12NTD-13Core-13CTD) and iFGF13VY/12 (13NTD-12Core-12CTD; Fig. S5 A). If the NTD was sufficient to produce an iFGF-specific effect, we would expect the shifts in the G-V and VSD-IV F-V curves produced by the iFGF12B/13 and iFGF13VY/12 chimeras to resemble those produced by iFGF12B and iFGF13VY, respectively. These iFGF chimeras, however, produced modulatory effects on Na_V channels G-V and VSD-IV F-V curves that are different from the effects of iFGF12B and iFGF13VY (Fig. S5 B and Table S4), implying that the N-terminal sequence alone is not enough to confer the specificity of iFGF-mediated effects. Therefore, we generated additional chimeras switching (1) the C-termini or (2) both the N- and C-termini of iFGF12B and iFGF13VY. Remarkably, neither of these chimera pairs replicated the modulation of the VSD-IV F-V curves produced by iFGF12B and iFGF13VY (Fig. S6 and Table S4), suggesting that all iFGF domains contribute to VSD-IV regulation.

Activation of VSD-IV was previously shown to facilitate Na_V channel fast inactivation (Chahine et al., 1994; Horn et al., 2000; Chanda and Bezanilla, 2002; Hsu et al., 2017). In addition, charge neutralization mutations in the S4 segment of VSD-IV caused a shift in the V_1/2 of inactivation of Na_V1.5 currents (Capes et al., 2013). To explore the relationships between VSD-IV regulation and Na_V1.5 gating kinetics, we investigated possible correlations between/among the measured voltage-dependences of I_Na activation and inactivation, I_Na decay kinetics, and VSD-IV F-V curves for Na_V1.5 coexpressed with each of the iFGF chimeras (Table S4). Surprisingly, the iFGF and iFGF chimera data showed no correlation between the V_1/2 values determined for steady-state inactivation of I_Na (V_{1/2, inact}) and the VSD-IV F-V curves, yielding a Pearson’s correlation coefficient (r) equal to 0.23 and an R² of 0.054 (Fig. S7 A), and only modest correlation with the slope (k) values (r = 0.62, R² = 0.4; Fig. S7 C). Similarly, no correlation was found between the voltage-dependences of I_Na activation (V_{1/2, act}) and the VSD-IV F-V curves (for V_1/2: r = 0.005, R² = 3 × 10⁻⁵, for k: r = −0.34, R² = 0.12; Fig. S7, B and D).

These analyses, however, did reveal that the slope (k values) of the VSD-IV F-V curves correlated with the parameters describing I_Na decay, specifically with the slow inactivation time constants (Ƭ_s) and the fractional amplitudes of the slow component of inactivation (%A_slow). As illustrated for I_Na evoked at 0 mV, the r and R² values between VSD-IV k values and the Ƭ_s were 0.79 and 0.63 (Fig. 7 B). Similarly, the calculated r and R² values for the correlation between the measured %A_slow and VSD-IV k values were 0.76 and 0.59, respectively (Fig. 7 C). No correlations, however, were established between the V_1/2 values of the VSD-IV F-V curves and neither the Ƭ_slow nor the %A_slow of I_Na (for Ƭ_slow: r = 0.53, R² = 0.29, for %A_slow: r = 0.06, R² = 0.004; Fig. S7, E and F).

Combining the two parameters derived from the VSD-IV F-V curves (i.e., V_1/2 and k values) predicts the %A_slow with higher accuracy than achieved with the V_{1/2, inact}. The R² values between the predicted and actual values for %A_slow and the V_{1/2, inact} were 0.72 and 0.58, respectively (Fig. 7 D and Fig. S7 G). Together, these results suggest that the slope of VSD-IV F-V curve is a strong determinant of the slower component of Na_V channel inactivation. We also found that a steeper VSD-IV F-V curve, corresponding to the earlier completion of VSD-IV activation at more hyperpolarized potentials, contributes to a reduction in the relative amplitude of the slower component of inactivation. The absence of correlation between the V_{1/2, inact} and the V_1/2 of VSD-IV F-V curve hints that at least one additional element is involved in iFGF modulation of the voltage-dependence of steady-state inactivation of Na_V1.5 currents.

In this study, we demonstrate that the differential regulation of the time- and voltage-dependent properties of Na_V1.5-encoded currents by iFGF12B and iFGF13VY reflects distinct effects on VSD-IV activation. Although previous studies have reported the importance of the iFGF N-terminal domain in the isoform-specific modulation of I_Na gating (Lou et al, 2005; Laezza et al, 2009; Yang et al, 2016; Barbosa et al, 2017; Ali et al, 2018; Singh et al, 2020, 2021; Chakouri et al, 2022), our results show that switching the N-termini of iFGF12B and iFGF13VY is not sufficient to confer the iFGF-specific modulation of Na_V1.5 channel gating and VSD-IV activation. Rather, our results indicate that all regions of the iFGF proteins (i.e., the N-terminal, core, and C-terminal domains) contribute to determining iFGF-specific effects on I_Na gating and VSD-IV activation.

In native mouse cardiomyocytes, we observed a change in the voltage dependence of steady-state inactivation of I_Na with the deletion of iFGF13. This alteration results in decreased numbers of channels available to activate and conduct I_Na. The expression of iFGF12B in cFgf13KO ventricular myocytes rescues Na_V channel steady-state availability, revealing roles for the iFGF proteins in regulating cardiac myocyte excitability. Patients carrying mutations on the binding interface of iFGF12B and Na_V1.5 exhibit phenotypes associated with Brugada (Hennessey et al., 2013) and long-QT type 3 syndrome (Wang et al., 2012). In addition, a missense mutation in Fgf12 was shown to increase arrhythmia susceptibility in mice (Veliskova et al., 2021). The expression of iFGF12B also exerts an additional effect on Na_V channel steady-state activation. The shift in voltage dependence of I_Na activation toward more positive potentials with iFGF12B means that larger depolarizations are required to activate the channels and, therefore, to reach the threshold for triggering APs. This effect will lead to a longer refractory period, where myocytes cannot be activated and could suppress arrhythmic activity.

Previous studies of Fgf13 knockdown or knockout in mice myocytes reported a reduction in the peak I_Na density (Wang et al., 2011a; Park et al., 2016; Wang et al., 2017). Here, however, we did not observe any significant reduction in the I_Na density with cardiac-specific iFGF13 knockout or with the addition of iFGF12B in cFgf13KO ventricles. These differences may reflect the mouse strain or knockout techniques employed or the I_Na recording conditions used in the previous versus the present studies. Interestingly, Park and colleagues reported a reduction in peak I_Na density in mouse ventricular myocytes lacking iFGF13 at 30°C but not at 25°C (Park et al., 2016). Whether iFGF12B and/or iFGF13VY directly affect cardiac Na_V1.5 trafficking and membrane localization, as demonstrated for iFGFs in some neuronal cells (Laezza et al., 2009), requires further study.

The regulation of VSD-IV activation is evident not only with iFGFs but also with other Na_V channel accessory proteins including the β subunits. Past studies showed that both the Na_Vβ1 and Na_Vβ3 subunits induce depolarizing shifts in VSD-IV activation, i.e., effects opposite iFGFs-mediated effects (Zhu et al, 2017, 2021). The presence of Na_Vβ subunits and other Na_V channel accessory subunits in mouse myocytes, therefore, could affect iFGF-mediated changes in I_Na decay kinetics, accounting for observed differences in inactivation kinetics compared with the results observed in the Xenopus oocyte expression systems. In addition, the binding of Ca²⁺/calmodulin (CaM) on the Na_V C-terminus is in the same vicinity as iFGF binding and might influence the function of iFGF, similar to how iFGF enhances the CaM function in Na_V1.5 variants of reduced CaM binding affinity (Gade et al., 2020; Abrams et al., 2020). Different levels of Ca²⁺/calmodulin could thus contribute to system-specific effects of iFGF. Finally, posttranslational modification, such as phosphorylation on the Na_V1.5 C-terminus, that occurs in mammalian cells, but not in Xenopus oocytes, can also affect the binding of iFGFs (Burel et al., 2017; Iqbal and Lemmens-Gruber, 2019; Lorenzini et al., 2021). The synergistic effects of multiple regulatory proteins and posttranslation channel modifications suggest that VSD-IV may act as a signaling hub where the regulation by many different Na_V channel complex components is integrated to determine inactivation kinetics. In addition, other Na_V α subunits, such as Na_V1.6, might be present in native mouse cardiomyocytes (Maier et al, 2004; Lopez-Santiago et al, 2007), producing I_Na with different gating properties and differentially affected by iFGF coexpression.

The structures of the iFGF13/iFGF12 proteins in complex with the Na_V1.5 C-terminus reveal that the common iFGF core domain is critical for binding (Wang et a., 2012; Wang et al., 2014). In addition, the folding of the core domain is nearly identical in iFGF13 (PDB accession nos. 3HBW [Goetz et al., 2009] and 4DCK [Wang et a., 2012]) and iFGF12 (PDB accession nos. 1Q1U [Goetz et al., 2009] and 4JQ0 [Wang et al., 2014]). It should also be noted that the iFGF N- and C-terminal domains are omitted from all of these structures. Both ends of the core domain structures, where the N- and C-terminal domains are located, however, are in the vicinity of the iFGF–Na_V1.5 interaction sites. The N- and C-terminal domains of the iFGFs, therefore, might be expected to also interact with other parts of the Na_V α subunit or assembled Na_V1.5 channel and influence channel/current properties. Consistent with this hypothesis, it has previously been reported that mutation of a single proline residue in iFGF12 outside the iFGF–Na_V C-terminus binding interface altered the (iFGF–Na_V C-terminus) binding affinity (Wang et al., 2011b). Similar functional effects were observed when the corresponding proline residue in iFGF13 was modified. These observations led the authors of this study to suggest that the iFGFs may contain multiple Na_V α subunit interaction sites, each contributing to the specific modulatory effects of the iFGF on different Na_V α subunits and, therefore, on different Na_V channels (Wang et al., 2011b). It has also been demonstrated that the perturbation of iFGF binding sites at two distinct interfaces, for example, using (two) peptides mimicking different sequences of iFGF14, led to differential modulation of Na_V1.6-encoded channel gating (Ali et al, 2018; Singh et al, 2021). The differential effects of the various iFGF proteins on Na_V channels/currents encoded by different Na_V α subunits is thus likely the result of multiple and variable interactions, involving the diverse iFGF N- and C-termini, as well as the canonical binding interface, and resulting in distinct functional effects on VSD-IV dynamics and Na_V channel gating.

Analyses of electrophysiological data obtained from iFGF12B/13VY chimeras coexpressed with Na_V1.5 revealed a strong correlation between the slope(s) of the VSD-IV activation curve(s) and the kinetics of I_Na decay. This finding is consistent with previous studies demonstrating that the activation of VSD-IV is necessary for initiating fast inactivation (Chahine et al., 1994; Horn et al., 2000; Chanda and Bezanilla, 2002). Additional studies using site-3 toxins, which bind to the extracellular loop near repeat IV S4, showed that intermediate activation of VSD-IV prevents Na_V channel transition into a fast inactivated state (Hanck and Sheets, 2007; Campos et al., 2004). These results imply that the full activation of VSD-IV is required for Na_V fast inactivation (Campos and Beirão, 2006; Campos et al., 2008; Jiang et al., 2021a; Jiang et al., 2021b). According to the cryo-EM structure of mouse Na_V1.5 with α-scorpion toxin, the intermediate-activated VSD-IV made the binding of the IFMT motif less stable and the activation gate wider (Jiang et al., 2021a). The presence of iFGFs facilitates the movement of VSD-IV toward the fully activated state and thus promotes the binding of the IFMT motif or the fast component of inactivation, which also recovers at a faster rate. The competition between two kinetically different pathways of inactivation implies that fast inactivation may limit the contribution of the slow component of inactivation (Richmond et al, 1998; Ransdell et al, 2022). The larger impact on the slope of VSD-IV F-V thus correlates with a larger fraction of the fast component of inactivation (equivalent to reduced %A_slow) that can recover faster and partially contribute to increasing channel availability.

Modulation of VSD-IV is likely the direct effect of iFGF being bound to the Na_V C-terminal domain. A two-switch model for fast inactivation has been proposed based on the chimeric structures of Na_VPaS-hNa_V1.7 in complex with the α-scorpion toxin to link VSD-IV activation and the fast inactivation (Clairfeuille et al., 2019). It is hypothesized that the interaction between resting VSD-IV and Na_V C-terminus, termed switch 1, causes the sequestration of the III–IV linker by the CTD (switch 2), and the activation of VSD-IV alters the CTD conformation to release III–IV linker. We proposed that the binding of iFGF could potentially change the Na_V CTD interaction, mediating earlier activation of VSD-IV (switch 1) and the preceding channel inactivation, as illustrated in the schematic diagram of the iFGF mechanism in Fig. 8.

Additional modulation of the Na_V channel inactivation gate by the iFGF protein(s) is also possible. It has, for example, previously been reported that the E1784K mutation in Na_V1.5 disrupts the switch 2 interaction and results in iFGF12B being ineffective in reducing the persistent component of Na_V1.5-encoded currents (Gade et al., 2020), implying that the interaction between the Na_V III–IV linker and the C-terminal domain is important for the iFGF-mediated modulation of the persistent Na_V current. Another auxiliary protein bound to the Na_V1.5 C-terminal domain, calmodulin, has also been suggested to modulate intramolecular interactions in the Na_V α subunit (Gade et al., 2020), providing an additional mechanism to regulate the persistent Na_V current (Abrams et al., 2020; Chakouri et al., 2022). The binding of iFGF to the Na_V1.5 C-terminal domain, however, might additionally impact the calmodulin-binding conformation (Gabelli et al., 2014) and negatively affect the calmodulin function, as has been demonstrated for Ca²⁺/calmodulin-dependent inactivation of skeletal muscle Na_V1.4-encoded channels (Gardill et al, 2019). Further studies are needed to fully understand the complex interactions, including antagonistic and/or synergistic interactions, and how these impact the regulation/modulation of Na_V channel function by its various auxiliary subunits or combinations of auxiliary subunits.

The modulation of VSD-I activation by iFGF13VY, but not iFGF12B, may be due to the longer N-terminus of iFGF13VY disrupting the coupling between VSDs and resulting in an indirect modulation of VSD-I. Because of the domain-swapped arrangement in Na_V α subunit, the VSD of one repeat is placed near the S5–S6 linker of the neighboring repeat, allowing two VSDs to be coupled (Chanda et al., 2004). Previous work showed that the binding of local anesthetic drugs inside Na_V channel pores can alter VSD coupling, especially between VSD-I and VSD-IV, resulting in their F-V curves shifting in opposite directions (Muroi and Chanda, 2009). The longer N-terminus of the A-type iFGF isoform, iFGF14A, was also found to localize near the channel intracellular gate where it could regulate the pore (White et al., 2019; Yan et al., 2014). It is therefore possible that the N-terminus of iFGF13VY may remodel VSD coupling, causing altered VSD-I activation as an indirect effect that is connected to VSD-IV modulation.

In this study, we demonstrated distinct effects of iFGF12B and iFGF13VY on the regulation of Na_V1.5 gating kinetics. Our results demonstrate that iFGFs facilitate the outward translation of VSD-IV from resting to activated conformation, leading to a reduction in the slow component of inactivation. The iFGF-specific effects are partially derived from the differential regulation of VSD-IV activation. The ability to modulate the activation of VSD-IV, however, is not unique to the iFGFs. Other regulatory proteins, toxins, and antiarrhythmic drugs can modulate VSD-IV activation and affect Na_V channel inactivation. Taken together, we present a novel mechanism of iFGF modulation of cardiac I_Na that accounts for distinct iFGF effects.

Olaf S. Andersen served as editor.

The authors thank Dr. A. Burkhalter for assistance with and training in retro-orbital virus injections, Drs. M. Goldfarb and D. Ornitz for providing the Fgf12KO mouse line, Dr. G. Pitt for the polyclonal anti-iFGF13 antibody, and R. Wilson for expert technical assistance.

We gratefully acknowledge the financial support provided by the National Heart Lung and Blood Institute of the NIH (R01 142520 to J.M. Nerbonne and R01 HL150637 to J.M. Nerbonne and J.R. Silva), the NIH National Center for Research Resources (UL1 RR024992 to the Institute for Clinical and Translational Sciences at Washington University), and the Children’s Discovery Institute Pediatric Disease Mouse Models Core at Washington University. Viruses were generated in the Hope Center Viral Vectors Core, supported by a Neuroscience Blueprint Core grant (P30 NS057105) at Washington University Medical School. None of these funding sources were involved in study design, data collection, data analyses, data interpretation, manuscript preparation, or the decision to submit this article for consideration for publication.

The authors declare no competing financial interests.

Author contributions: P. Angsutararux, J.M. Nerbonne, and J.R. Silva intiated the project; P. Angsutararux, M. Marras, J.M. Nerbonne, and J.R. Silva supervised the work; P. Angsutararux, A.K. Dutta, M. Marras, C. Abella, R.L. Mellor, and J. Shi conducted the research; P. Angsutararux, J.R. Silva, and J.M. Nerbonne wrote the manuscript. All authors provided edits and feedback.