Short SCN5A Transcript Yields a NaV1.5 Fragment Influencing Cardiac Metabolism

Abstract

Background:

SCN5A encodes the cardiac voltage-gated Na⁺ channel, NaV1.5, classically known for initiating action potentials and recently implicated in cardiomyocyte metabolism via mitochondrial Na⁺/Ca²⁺ exchange. SCN5A variants are linked to arrhythmias and heart failure, but mechanisms controlling SCN5A/NaV1.5 expression and its metabolic interface remain understudied.

Methods:

We used bioinformatic approaches to identify novel SCN5A regulatory features and discovered an alternative polyadenylation (APA) signal downstream of exon 2, which is conserved in humans and several other species but not mice. To test its function, we generated knock-in mice harboring the human APA signal. Western blotting, cell fractionation and fluorescence microscopy were used to characterize the resulting truncated protein isoform that localizes to mitochondria. Mitochondrial functions and/or metabolites were assessed in neonatal rat cardiomyocytes, human IPSC-derived cardiomyocytes, and mouse hearts overexpressing the novel isoform.

Results:

We identified a well-conserved APA signal downstream of SCN5A exon 2, yielding a truncated transcript isoform (SCN5A-short). Reanalysis of cardiac APA-seq and mRNA-seq data reveals reduced SCN5A-short expression in failing human hearts. Knock-in of the human APA signal into mice enables expression of SCN5A-short, while decreasing full-length SCN5A mRNA. SCN5A-short encodes a novel N-terminal fragment of NaV1.5 (NaV1.5-NT) that localizes to the mitochondrial matrix in cardiomyocytes and mouse hearts. Exogenous expression of NaV1.5-NT in cultured cardiomyocytes enhances mitochondrial respiration, ATP production, and mitochondrial ROS, while depleting NADH. Native-PAGE analyses indicate that this coincides with enhanced complex I activities, as well as context-dependent alterations of CV assembly. Moreover, moderate cardiomyocyte-targeted NaV1.5-NT expression in mice was sufficient to rewire the cardiac metabolome, with suggestive evidence of increased fatty acid oxidation.

Conclusion:

APA-mediated regulation of SCN5A produces a short transcript encoding NaV1.5-NT, a novel mitochondrial-targeted peptide that supports cardiomyocyte metabolism. While the precise molecular mechanisms remain unresolved, these findings highlight an unforeseen alternative pathway for expanding SCN5A-mitochondrial crosstalk, with potential implications for metabolic changes in heart failure and arrhythmias.

INTRODUCTION

Heart failure (HF) and arrhythmias are major contributors to sudden cardiac death and frequently coexist, with one often exacerbating the other¹. Despite advances in treatment and prevention strategies, HF remains the world’s leading cause of death and affects ~2% of people globally². A host of underlying molecular mechanisms coincide with deterioration of cardiac function and arrhythmic death risk, including shifts in gene expression and ion channel dysregulation³. While the primary drivers (e.g. specific ion channels) mediating the cardiac action potential are known, there remains a need to continue characterizing the expansive regulatory mechanisms controlling these genes/proteins, as well as the many others that are central to HF onset and progression.

SCN5A encodes the pore-forming subunit of the voltage-gated Na⁺ channel, NaV1.5, that is responsible for inward Na⁺ current during the fast depolarization phase of the cardiac action potential⁴. Fine-tuned NaV1.5 activity is required for normal heart function⁵, and several genetic variants in SCN5A cause inherited arrhythmias (e.g. Brugada syndrome and long-QT syndrome) and dilated cardiomyopathy (DCM)^6,7. Prior research has focused on defining NaV1.5 functional properties, effects of disease-associated variants, and pharmacological targeting^4,8–10. Beyond that, multiple studies have also found decreased SCN5A/NaV1.5 expression and dysregulated function in HF, spurring investigations of its roles beyond conduction^11,12. Along these lines, we previously reported that lower SCN5A expression associates with increased HF patient mortality, which was surprisingly linked to worsening ejection fraction as opposed to increased risk for arrhythmic sudden cardiac death (inferred from implantable cardioverter-defibrillator shock rates)¹³.

NaV1.5 is subject to multilayered regulation at several points during its lifecycle (e.g. intracellular trafficking, membrane anchoring, and degradation¹⁴, post-translational modification¹⁵, and interactions with metabolites^16–19). Further influence is imparted by transcriptional and post-transcriptional controls of SCN5A mRNA expression, with prior studies characterizing several alternatively spliced transcripts^20,21, transcriptional regulators^22–24, and influence of RNA-binding proteins^25,26 and microRNAs^13,25,27. However, continued characterizations of SCN5A/NaV1.5 regulatory controls and non-canonical functions are needed to expand understanding of its diverse roles in cardiac biology and disease, particularly in light of recent evidence implicating NaV1.5 in mitochondrial crosstalk via Na⁺/Ca²⁺ exchange¹².

One notable route of SCN5A gene regulation that remains unexplored is alternative polyadenylation (APA), which is an RNA control mechanism that generates distinct 3’ transcript ends through differential usage of polyadenylation sites. APA involves ~85 distinct regulatory proteins that generate unique APA signatures in ~70% of human genes^28,29. Notably, APA can occur in upstream introns to produce truncated protein isoforms³⁰, and differential APA has been described in several conditions such as proliferation, differentiation, and membrane excitability^31–36. Previously generated transcriptome-wide maps of APA events in human heart tissues suggest that APA is dynamically altered in HF. Global shortening of 3’-UTR lengths attributed to APA has also been observed in hypertrophied mouse hearts^37,38. Despite these findings, the overall regulation and biological consequences of APA in the heart remain largely underexplored, both globally and at the single gene level.

Herein, we report our serendipitous discovery of a novel APA signal in the SCN5A locus that influences the expression of full-length SCN5A mRNA and produce a short transcript isoform encoding an N-terminal fragment of NaV1.5 (NaV1.5-NT). Our initial characterizations demonstrate that NaV1.5-NT localizes to the mitochondrial matrix and enhances cardiomyocyte respiration, providing a direct and unexpected mode of SCN5A-mitochondrial crosstalk that expands the interface beyond mechanisms of Na⁺/Ca²⁺ exchange.

METHODS

Data Availability Statement

All data that support the findings of this study are available in the primary figures, the Supplemental Material, or from the corresponding author upon reasonable request. Extended methods and resources used are provided in the Supplemental Material and the Major Resources Table.

Mice:

All animal studies were approved by the Institutional Animal Care and Use Committees (IACUC) at the University of Iowa, and all experiments conform to the appropriate regulatory standards.

Seahorse Assays:

NRCM Seahorse assays were performed using XF media (pH 7.4) supplemented with 5 mM HEPES, 10 mM glucose, 2 mM L-glutamine, and 1 mM sodium pyruvate. For mitochondrial stress test, cells were sequentially exposed to 2 μM oligomycin, 0.3 μM FCCP (additional FCCP doses shown in the Supplemental Material), and 1 μM each rotenone plus antimycin A. For ATP production rate assay, cells were sequentially exposed to 2 μM oligomycin and 1 μM each rotenone plus antimycin A, and ATP rates were calculated using the Agilent Seahorse ATP rate calculator. hiPSC-derived cardiomyocyte assays were performed using Seahorse DMEM media supplemented with 5.6 mM glucose, 2 mM glutamine, and 1 mM sodium pyruvate and cells were sequentially exposed to 1 μM oligomycin, 1 μM FCCP, and 0.5 μM each rotenone plus antimycin A. All measurements were normalized to total protein content.

NAD+/NADH Quantification:

48-hours post-transduction, NRCMs were washed with ice-cold PBS and NAD+/NADH content was quantified using the Promega NAD+/NADH-Glo Assay (G9071), according to the manufacturer’s instructions. Signal was normalized to total protein content.

Statistical methods:

Statistical analyses were performed using Graphpad Prism 10 and data are expressed as mean ± SEM, unless otherwise indicated. Raw p-values are shown with at least two significant digits and values less than 0.050 were considered significant. P-values were determined by an unpaired t-test for experiments with two groups or one-way ANOVA with Dunnett’s multiple comparisons correction for experiments with more than two groups. Normality tests were performed on experiments with n>6, and non-parametric tests were used where normality could not be established. In these cases, Mann-Whitney U tests were used for experiments with two groups or Kruskal-Wallis with Dunn’s multiple correction tests for experiments with more than two groups. Representative immunofluorescence images, western blots, and BN-PAGE gel images were chosen to closely match the mean for the described experiment. Uncropped and unedited blot and gel images are provided as a supplemental file.

RESULTS

An alternative polyadenylation signal downstream of SCN5A first coding exon generates a truncated transcript that is downregulated in failing human hearts.

While examining SCN5A gene structure and expression using the UCSC Genome Browser³⁹, we noted the presence of intronic RNA-seq reads (GTEx human heart RNA-seq tracks) extending downstream of the first coding exon (exon 2), with apparent dissipation near a canonical polyadenylation signal (AATAAA; Fig. 1A,B). This pattern is also found in several myocyte-like cell lines (AC16, Rh30, and iPS-CMs) and in cardiac RNA-seq data from other species (e.g. cat, pig, and monkey; data not shown). Notably, this upstream polyA signal is reasonably well-conserved across mammals (present in 27/62 species) but is not present in mice or rats (Fig. 1B). Targeted 3’-end polyA-seq data³⁷ from nonfailing and failing human heart tissues shows a convincing transcript terminus mapping ~30 bases downstream of the polyA signal in all samples analyzed, consistent with canonical 3’-end mRNA processing⁴⁰ (Fig. 1A, S1). These data indicate that this previously overlooked polyA signal generates a truncated SCN5A transcript that we refer to as SCN5A-short. To assess if SCN5A APA might be altered in HF, we performed targeted re-analysis of the Creemers et al. human cardiac 3’-end polyA-seq data³⁷, revealing a trending reduction in the ratio of SCN5A-short/full-length polyA usage in patients with DCM (versus nonfailing controls, p=0.222 for nonparametric and p<0.09 for parametric tests, n=5/group; Fig. 1C). Moreover, re-analysis of human cardiac RNA-seq data from two larger cohorts^41,42 provides further support that SCN5A-short transcript expression (relative to full-length SCN5A mRNA) decreases in failing hearts (DCM and ischemic and hypertrophic cardiomyopathy [ICM and HCM respectively], versus nonfailing controls, p<0.05, Fig. 1D,E). Based on these overall data, we estimate the upstream polyA signal is used ~10–20% of the time in human hearts, with a diverse range across individuals. Notably, this usage rate may also vary largely across cell-types and/or conditions; for example, RNA-seq read coverage profiles support predominant utilization (>50%) of the upstream polyA signal in MDA-MB-231 breast cancer cells, which endogenously express full-length SCN5A⁴³ (Fig. S2).

Figure 1. An alternative polyadenylation signal downstream of SCN5A first coding exon generates a truncated transcript that is downregulated in failing human hearts.

A, Screenshots of the SCN5A exon 2 locus from the UCSC Genome Browser (hg38) with NCBI RefSeq gene annotations, GTEx RNA-seq (representative human atrial and left ventricular), and GWIPS-viz Ribo-seq tracks are shown. Representative RNA-seq from AC16, Rh30, and iPS-cardiomyocyte cell lines and 3’-end polyA (pA)-seq from human heart tissue are also shown. Red arrow and box indicate position and alignment of the polyadenylation signal sequence (AATAAA), which leads to the production of a truncated transcript isoform, SCN5A-short. B, Clustal-Omega alignment of the alternative polyadenylation signal in selected mammals, showing its conservation in several species but absence in mice and rats. C, Targeted analysis of 3’-end polyA-seq in healthy versus dilated cardiomyopathy (DCM) human heart samples. The relative ratio of reads at the upstream polyA site to the full-length polyA site are shown (p-value was determined by Mann-Whitney U test; n=5/group; plotted as mean ± SEM) D, Re-analysis of human heart mRNA-seq data (Colorado study) comparing the relative TPM ratio of short to full-length SCN5A transcript expression in nonfailing (NF; n=14) versus ischemic cardiomyopathy (ICM; n=13) and dilated cardiomyopathy (DCM; n=37) samples (p-values were determined by a one-way ANOVA with Dunnett’s post-hoc). E, Re-analysis of human heart mRNA-seq data from the MAGNet (UPenn) study comparing the relative TPM ratio of short to full-length SCN5A transcript expression in nonfailing (NF; n=166) versus dilated cardiomyopathy (DCM; n=195) and hypertrophic cardiomyopathy (HCM; n=28) samples (p-values were determined by Kruskal-Wallis test with Dunn’s post-hoc). Unless otherwise stated, data are shown as violin plots.

The SCN5A alternative polyadenylation signal influences full-length SCN5A mRNA expression.

To test if the upstream polyA site influences full-length SCN5A mRNA and NaV1.5 protein expression, we generated knock-in mice (SCN5A-APA KI), replacing mouse intronic sequence with human SCN5A sequence harboring the polyA signal and additional surrounding sequence context (Fig. 2A). RT-qPCR done on wildtype (WT) and SCN5A-APA KI mouse heart RNA samples indicates that the APA site is clearly used in KI mice to produce SCN5A-short (Fig. 2B). RT-qPCR analyses also revealed that full-length SCN5A mRNA expression is trending lower in APA-KI mouse hearts (versus WT controls, n≥4/group, p=0.0635; Fig. 2C), consistent with a competitive usage model for the two polyA sites. Interestingly, full-length NaV1.5 protein expression appears unchanged (Fig. 2D). We also assessed cardiac conduction in these mice by electrocardiography and found no changes in PR or QRS intervals, which are influenced by NaV1.5 activity. In addition, electrocardiograms recorded from WT or APA-KI mice showed no significant differences and patch-clamp electrophysiological measures showed no significant differences in Na⁺ currents in isolated APA-KI mouse cardiomyocytes, relative to WT controls (Fig. S3A–C). Overall, these findings support that the SCN5A APA signal is functional and can influence full-length SCN5A mRNA expression, while possible compensatory regulation may contribute to maintaining normal levels of NaV1.5. Taken together, this may suggest that SCN5A APA evolved to serve an alternative purpose.

Figure 2. The SCN5A alternative polyadenylation signal influences full-length SCN5A mRNA expression in mouse hearts.

A, Schematic of how the SCN5A-APA knock-in (KI) mouse line was generated. CRISPR/Cas9 technology was employed to replace the mouse intronic region with the human version, which carries the alternative polyA signal sequence. Red region indicates the edited sequence. B, RT-qPCR for SCN5A-short transcripts in wildtype (WT; n=5) versus SCN5A-APA KI (n=4) mouse heart RNA samples (corrected for minimal gDNA background signal), supporting that this polyA signal can produce SCN5A-short transcripts. C, Relative full-length SCN5A mRNA expression, determined by RT-qPCR, in WT (n=5) versus KI (n=4) mouse hearts, indicating alternative polyadenylation regulates its expression. D, Western blot and relative quantitation of NaV1.5 protein expression in WT (n=6) and KI (n=5) mouse heart lysates. All data are plotted as mean ± SEM and p-values were determined by Mann-Whitney U tests.

SCN5A-short transcript generates a novel protein corresponding to the N-terminus of NaV1.5.

In addition to APA-mediated regulation of full-length SCN5A expression, the truncated transcript isoform could generate a novel 134 amino acid (aa) protein composed of an N-terminus identical to NaV1.5 (residues 1–91) and a unique C-terminus derived from “intronic” sequence (residues 92–134), termed NaV1.5-NT (Fig. 3A). To begin characterizing this protein, we first assessed its stability using an adenovirus to express NaV1.5-NT (Ad-NaV1.5-NT) in neonatal rat cardiomyocytes (NRCMs) and subjected cell lysates to western blotting using a commercial NaV1.5 N-terminal antibody (recognizes aa 42–70) and a custom antibody targeted to the unique C-terminal region (aa 109–123). This revealed three unique bands (migrating at ~17–20 kDa) that derive from the NaV1.5-NT transgene (Fig. 3B). We next generated a cardiomyocyte-specific AAV (AAV9-NaV1.5-NT, driven by the human cardiac troponin promoter) to assess NaV1.5-NT expression in mouse hearts three weeks after AAV administration. Western blotting with commercial NaV1.5-NT antibody revealed unique NaV1.5-NT-derived bands running at ~17 kDa and ~34 kDa (possible SDS-resistant dimers; not observed in NRCM lysates, data not shown), the former of which appears to migrate at the same size as a background band in vehicle-injected controls. Notably, 17 kDa NaV1.5-NT is co-immunoreactive with our custom antibody, which does not appear to detect 34 kDa NaV1.5-NT (Fig. 3C), suggesting that the antibody epitope is not accessible in the dimeric form. We also assessed NaV1.5-NT expression in SCN5A-APA KI mouse hearts; however, western blotting with the available antibodies failed to detect any convincing bands (data not shown). This was somewhat anticipated given that these antibodies only modestly detect overexpressed NaV1.5-NT in mouse hearts (Fig. 3C). Overall, these data support that NaV1.5-NT can be stably expressed in various forms in cardiomyocytes both in vitro and in vivo.

Figure 3. SCN5A-short transcript generates a novel protein corresponding to the N-terminus of NaV1.5.

A, Schematic of NaV1.5-NT protein generated from SCN5A-short, corresponding to the N-terminus of NaV1.5 (blue region; amino acids 1–91) and a unique C-terminal region corresponding to intronic sequence (orange region; amino acids 92–134). Two antibody epitopes are shown: a commercial NaV1.5 antibody (Abcepta; recognizing amino acids 42–70) and a custom antibody that we generated (recognizing amino acids 109–123). B, Western blots of NRCMs transduced with Ad-GFP or Ad-NaV1.5-NT using commercial or custom NaV1.5-NT antibodies show unique NaV1.5-NT-derived bands running near the expected size. C, Western blots of mouse hearts transduced with cardiomyocyte-specific AAV-GFP or AAV-NaV1.5-NT using commercial or custom NaV1.5-NT antibodies. Specific bands are detected at 17 kDa and 34 kDa (red arrows). D, Western blot of human hearts alongside vehicle or AAV treated mouse hearts using commercial NaV1.5-NT antibody shows bands co-migrating with 34 kDa and 17 kDa NaV1.5-NT that are present in human samples.

To assess if NaV1.5-NT is endogenously expressed in humans, we performed western blot (commercial N-terminal antibody) on human heart tissue lysates, revealing bands running at the same sizes (~17 and 34 kDa) as a positive control (heart lysate from AAV-NaV1.5-NT injected mouse; Fig. 3D). Note: we find that freeze-thawing these control lysates appears to alter distribution of NaV1.5-NT (17 versus 34 kDa) and/or increases prominence of the 17 kDa background band, now seen strongly in both vehicle and AAV treated samples. To further interrogate whether the 17 kDa band could be NaV1.5-NT, we blotted a separate cohort of nonfailing human heart samples to assess if this band is co-reactive with both commercial and custom antibodies, targeting the N- and C-terminal portions of NaV1.5-NT respectively. Notably, the 17 kDa band was detected by both antibodies, with band intensities for commercial versus custom significantly correlating across samples (Fig. S4; p=0.015, with no correlation of either to loading controls), providing further support that this band is NaV1.5-NT. To gain additional supporting evidence of the presence of NaV1.5-NT in humans, we also searched publicly available mass spectrometry data using PepQuery, an online peptide identification algorithm and database⁴⁴. This search yielded impressive coverage, with five high-confidence peptides identified across several independent experiments in human cells/tissues (each with >100 matches) that correspond to and span much of the unique NaV1.5-NT C-terminus, providing further indication of its natural existence in humans (Table S2).

NaV1.5-NT localizes primarily to mitochondria

We next examined the molecular function of NaV1.5-NT, initially hypothesizing that it may influence NaV1.5 channel activity, based on previously published reports showing that transgene-mediated expression of a similar but artificially-engineered NaV1.5 N-terminal fragment (Nter; aa 1–132) was sufficient to enhance NaV1.5 current densities and surface localization^45–47. Along these lines, we tested if Ad-NaV1.5-NT in NRCMs altered NaV1.5 activity but found no significant differences in Na⁺ current densities under low extracellular Na⁺ concentration nor at physiologic Na⁺ concentration, the latter of which drives higher Na⁺ influx (Fig. S5).

While examining NaV1.5-NT aa sequence using various in silico prediction algorithms, we found that it contains a probable mitochondrial targeting sequence (MTS, predicted by both MitoProtII⁴⁸ and MitoFates⁴⁹; aa 1–34). To investigate this, we transfected NRCMs with NaV1.5-NT expression plasmids and performed fluorescence confocal microscopy; NaV1.5-NT immunostaining (not present in adjacent non-transfected cells) shows clear co-localization with mitochondrial staining (MitoTracker), which is abolished in deletion constructs lacking the MTS (ΔMTS; Fig. 4A). These findings were reproduced in experiments using the A549 immortalized cell line (Fig. S6). Interestingly, in parallel, we found that the previously reported artificial Nter fragment also predominantly co-localizes with mitochondria in both A549 cells and NRCMs, raising interesting questions regarding how Nter may promote NaV1.5 activity (Fig. S6–7). We next further assessed NaV1.5-NT localization by western blotting of mitochondrial isolates collected from Ad-infected (NaV1.5-NT or GFP) NRCMs, revealing enrichment of two out of three NaV1.5-NT derived bands in mitochondrial lysates, evidenced by decreased signal in the mitochondria-depleted supernatant (Sup) samples and appearance in mitochondria pellets (Fig. 4B). Notably, the smallest band is not enriched in mitochondrial pellets and may derive from an alternative in-frame ATG start codon located downstream of the MTS.

Figure 4. NaV1.5-NT localizes primarily to mitochondria.

A, Representative confocal images of immunostained (commercial NaV1.5-NT antibody) NRCMs transfected with plasmids expressing NaV1.5-NT or NaV1.5-NT-ΔMTS (mitochondrial targeting sequence deleted). Co-localization with mitochondria (labeled with MitoTracker) was evaluated (scale bar = 20μm). B, Mitochondria were isolated from adenovirus-transduced NRCMs (Ad-GFP or Ad-NaV1.5-NT). Fractions corresponding to the total homogenate (Hom), mitochondria-depleted supernatant (Sup), or mitochondrial pellets were blotted for NaV1.5-NT (commercial antibody) and TOM20, a classic marker for mitochondria. C, NRCM mitochondria isolated from Ad-GFP or Ad-NaV1.5-NT transduced cells were treated with isotonic (Iso), hypotonic (Hypo; to rupture the outer membrane), or isotonic plus triton (Iso+TX) in the presence or absence of proteinase K (PK) and subjected to western blot analysis. Proteins with known localization to the mitochondrial matrix (Ogdh) or outer membrane (Tom20) were blotted for as controls.

To further define the intra-mitochondrial localization of NaV1.5-NT, we performed classical proteinase K extraction experiments with western blotting, indicating that NaV1.5-NT resides in the mitochondrial matrix (similar to Ogdh control), consistent with its lack of a transmembrane domain (per Phobius⁵⁰ prediction algorithm; Fig. 4C). Beyond NRCM studies, blotting of mitochondrial pellet lysates collected from AAV-treated mouse hearts also supports that NaV1.5-NT localizes to mitochondria in vivo (Fig. S8). Together, these data indicate that NaV1.5-NT is most likely a mitochondrial matrix protein and may exist in dimeric form.

NaV1.5-NT expression influences mitochondrial respiration

We next explored whether NaV1.5-NT functionally influences mitochondrial physiology. Given prior work suggesting that SCN5A/NaV1.5 may influence mitochondrial Ca²⁺ through crosstalk with the mitochondrial Na⁺/Ca²⁺ exchanger (Nclx)¹², we assessed whether NaV1.5-NT expression might influence mitochondrial Ca²⁺ dynamics in NRCMs; however, we found no significant differences in mitochondrial Ca²⁺ levels nor uptake/efflux rates (Fig. S9). We subsequently measured respiration in Ad-transduced NRCMs by Seahorse assay (mitochondrial stress test). Baseline oxygen consumption rates are clearly and consistently elevated in NaV1.5-NT transduced cells compared to multiple controls (untreated, Ad-GFP, or Ad-βGal; Fig. 5A). Notably, this increased basal respiration is inhibited by oligomycin (a classic complex V uncoupler). NaV1.5-NT also increases maximal respiration, albeit to a lesser degree with some inconsistency across experiments (see extended data in Fig. S10). Similar findings were obtained in a complementary experiment (Seahorse ATP rate assay), with NaV1.5-NT expression inducing a much higher rate of inferred mitochondrial ATP production (i.e. derived from oxidative phosphorylation, OXPHOS; Fig. 5B, S11). We further examined this using a ratiometric ATP-sensitive GFP sensor⁵¹ and found that NaV1.5-NT expression increases ATP levels in NRCMs (Fig. 5C).

Figure 5. NaV1.5-NT expression influences mitochondrial respiration.

A, Seahorse assay (mitochondrial stress test; 0.3um FCCP dose) of Ad-NaV1.5-NT transduced NRCMs (untreated, Ad-GFP, and Ad-βGal controls are also shown). Basal, ATP-linked, and maximal respiration are shown to the right (n=4/group). Refer to Fig. S10 for additional measures. B, Mitochondria-derived ATP production rates calculated from Seahorse ATP rate assay (n=3–4/group). Refer to Fig. S11 for the OCR trace from which the data are calculated. C, Relative ATP levels in Ad-transduced NRCMs, quantified by fluorescence intensity of a ratiometric ATP-sensitive GFP reporter (n=120 [UT], 125 [Ad-BGal], 120 [Ad-NaV1.5-NT]; p-values determined by one-way ANOVA with Dunnett’s multiple comparisons test) cells across 3 wells, shown as violin plot). D, Mitochondrial membrane potential was quantified by TMRM staining intensity in intact NRCMs treated with Ad-NaV1.5-NT or Ad-GFP (control). Background staining (determined by staining intensity after FCCP treatment) was subtracted (n=12 wells/group across 2 independent experiments; p-values determined by Mann-Whitney U test). E, Relative mitochondrial ROS levels in Ad-transduced NRCMs, quantified by mitochondrial-localized ratiometric redox sensitive GFP sensor (Ad-roGFP; n=124 [UT], 120 [Ad-BGal], 117 [Ad-NaV1.5-NT]; p-values determined by one-way ANOVA with Dunnett’s multiple comparisons test) cells across 3 wells; shown as violin plot). F, Relative NAD+, NADH, and NAD+/NADH ratio quantification in adenovirus transduced NRCMs (n=4/group). All data are plotted as mean ± SEM and p-values were determined by a Kruskal-Wallis with Dunn’s multiple comparisons tests, unless otherwise noted.

We next evaluated if these differences were linked to altered mitochondrial integrity and/or respiratory by-products and/or co-factors. We measured mitochondrial membrane potential using TMRM in intact Ad-treated NRCMs (NaV1.5-NT versus GFP control) and found no significant differences relative to control cells (Fig. 5D). Given that NaV1.5-NT increases mitochondrial respiration through OXPHOS, we examined if this is linked to elevated mitochondrial ROS and depleted NADH pools. We measured mitochondrial and cytosolic ROS using genetically-encoded redox sensors⁵² and found that NaV1.5-NT expression significantly increases mitochondrial-derived ROS, with no statistically significant effects on cytosolic ROS levels (Fig. 5E, S12). In addition, we found that Ad-NaV1.5-NT decreases the amount of total NADH levels but had no apparent effects on NAD+, resulting in overall increased NAD+/NADH ratio (Fig. 5F).

NaV1.5-NT modulates complex V subunits and increases complex I-containing supercomplexes

Given that NaV1.5-NT expression strikingly increases baseline mitochondrial oligomycin-sensitive oxygen consumption rate and decreases NADH supply, we speculate that NaV1.5-NT may influence OXPHOS complexes, specifically complexes V and I respectively. To begin assessing this, we performed blue native-PAGE (BN-PAGE) and western blot on mitochondrial isolates collected from Ad-transduced (NaV1.5-NT or GFP control) NRCMs to measure levels of OXPHOS (super)complexes (antibody cocktail; Fig. 6A). We found no obvious differences in the amounts of primary OXPHOS complexes but note the clear presence of several bands (migrating at ~440 kDa) that increase with NaV1.5-NT expression.

Figure 6. NaV1.5-NT modulates complex V subunits and increases complex I-containing supercomplexes.

A, Representative BN-PAGE blotting of Ad-GFP and Ad-Nav1.5-NT treated NRCM mitochondrial isolates using an OXPHOS cocktail antibody. Red arrows indicate subcomplexes affected by NaV1.5-NT. B, Representative BN-PAGE blots of Ad-GFP and Ad-Nav1.5-NT treated NRCM mitochondrial isolates using an Atp5b (F₀-subunit), Atp5g1/2/3 (F₁-subunit), or NaV1.5-NT (commercial) antibody. Bands corresponding to F₁ or F₀ subcomplexes, intact complex V (CV) or complex V dimer (CV₂) are annotated. Relative intensities of bands corresponding to CV and F₁-subunits are quantified to the right (n=4/group; see Fig. S13 for additional blots). C, BN-PAGE in-gel complex V activity assay performed on Ad-fected NRCM (Ad-GFP vs. Ad-NaV1.5-NT) mitochondria isolates. White/light blue staining indicates CV activity and relative staining intensity was quantified to the right for bands corresponding to CV and F₁-subunits (n=4/group). D, In Ad-GFP and Ad-NaV1.5-NT treated NRCM mitochondrial isolates, representative data for BN-PAGE blot using a complex I antibody (Ndufb8; left) or in-gel complex I activity (violet staining; right). Bands corresponding to complex I (CI), complexes I+III₂ (CI+III₂), and supercomplexes (SCs) are annotated. CI-containing supercomplex abundance (blot) or activity were quantified to the right and the ratio of supercomplex band intensity to CI intensity (n=3/group). All data are plotted as mean ± SEM and p-values were determined by Mann-Whitney U tests.

We next examined complex V more closely given that several papers show that bands appearing in the 440–500 kDa range can correspond to complex V subunits. Complex V (also known as mitochondrial- or F₁/F₀-ATP synthase) has two major subunits, the F₁ that comprises the enzymatic ATPase activity and the F₀ subunit that forms the proton (i.e. H⁺) porous c-ring^53,54. BN-PAGE and western blot for Atp5b F₁ subunit clearly indicates that these bands of interest (i.e. those increased with NaV1.5-NT expression) correspond to F₁ subcomplexes, suggesting that NaV1.5-NT influences complex V biogenesis and/or stability (Fig. 6B, S13). BN-PAGE and western blot for subunit c of the ATP synthase (Atp5g1/2/3), shows increased levels of free F₀-ATPase (c-ring) after NaV1.5-NT expression, suggesting that NaV1.5-NT might impede assembly or promote disassembly of F₁-F₀ (Fig. 6B). Moreover, in-gel complex V activity assay, which measures “reverse-mode” ATP to ADP conversion, also showed free F₁ bands in controls that were increased in Ad-NaV1.5-NT treated samples (Fig. 6C). Of note, we found no evidence of overall decreased levels of intact complex V (or cV dimers) by BN-PAGE western blotting nor in-gel complex V activity assay (Fig. 6B–C, S13).

The pore forming F₀ subunit can act as a leaky proton pore⁵⁵, but given that NaV-1.5-NT expression did not disrupt mitochondrial membrane potential (Fig. 5D), we presumed that possible proton leak via F₀ could be synchronously offset by enhanced complex I activities moving protons back to the intermembrane space. We therefore assessed levels and activities of complex I. In our initial BN-PAGE assessment of OXPHOS complexes (Fig. 6A), signal for intact complex I was weak and unclear given the high abundance of complex V; thus, we ran additional gels and blotted specifically for complex I subunit Ndufb8 (Fig. 6D). This revealed that NaV1.5-NT expression increases the abundance of high molecular weight supercomplexes containing complex I, with no noticeable effect on the levels of primary complex I or total Ndufb8 expression (SDS-PAGE, Fig. S14). Notably, this increase in supercomplexes was also reflected in in-gel complex I activity assays (Fig. 6D), which could account for the observed NADH depletion in NaV1.5-NT expressing NRCMs.

Overall, these data combined with our mitochondrial phenotyping data lead us to speculate that NaV1.5-NT may play a multifaceted role in influencing mitochondrial physiology. Our data support that 1) NaV1.5-NT increases baseline oxygen consumption rate possibly by promoting complex V subcomplex conformations that enhance proton leak, and 2) NaV1.5-NT increases overall respiratory efficiency by enhancing formation and/or stability of complex I-containing respiratory supercomplexes, though the precise underlying mechanisms remain to be fully resolved.

NaV1.5-NT enhances mitochondrial respiration and complex I activity in hIPSC derived cardiomyocytes

To assess for similar roles of NaV1.5-NT in human cardiomyocyte model systems, we generated human IPSC-derived cardiomyocyte (hIPSC-CM) cultures to measure respiration rates and complex activities after NaV1.5-NT overexpression. Ad-NaV1.5-NT transduced cells showed higher baseline, ATP-linked, and maximal OCR relative to multiple controls (untreated, Ad-BGal, and Ad-GFP; p<0.0001; Fig. 7A). BN-PAGE and in-gel activity assays showed no obvious differences in complex I-containing supercomplexes nor complex V subcomplexes in Ad-NaV1.5-NT treated cells, albeit with some inconsistency across sample replicates (Fig. 7B and data not shown). However, NaV1.5-NT samples showed clear increases in free complex I abundances and activities (Fig. 7B). Combined with our NRCM study results, the data support that NaV1.5-NT overexpression consistently enhances mitochondrial respiration, likely in part through modulation of complex I, whereas possible influences on complex V may be context-dependent.

Figure 7: NaV1.5-NT enhances mitochondrial respiration and complex I activity in hIPSC derived cardiomyocytes.

A, Seahorse assay OCR trace (mitochondrial stress test) of Ad-NaV1.5-NT (n=22) transduced hIPSC-CMs (untreated [UT, n=21], Ad-BGal [n=19], and Ad-GFP [n=18] controls are also shown). Basal, ATP-linked, and maximal OCR are quantified to the right (data are representative of 2 independent experiments; p-values were determined by one-way ANOVA with Dunnett’s multiple comparisons correction). B, BN-PAGE complex I activity (left) or blot (Ndufb8; right) of Ad-treated hIPSC-CM whole-cell lysate (Ad-NaV1.5-NT versus Ad-GFP or Ad-BGal controls). Relative complex I activities or abundances are quantified to the right (n=4 wells/group; p-values were determined by Mann-Whitney U tests). All data are plotted as mean ± SEM.

Moderate cardiomyocyte-targeted NaV1.5-NT overexpression is sufficient to rewire the cardiac metabolome in mice.

Cultured NRCMs and hIPSC-CMs can be considered metabolically immature model systems, and adenoviral-mediated overexpression can result in supraphysiologic protein levels. To begin evaluating the in vivo effects of moderate NaV1.5-NT overexpression on overall heart function and metabolism, we injected mice with cardiomyocyte-targeted AAVs to express NaV1.5-NT, or vehicle and GFP controls (myoAAV3A capsid⁵⁶ driven by the human cardiac troponin promoter). Baseline echocardiography and electrocardiogram recordings showed no effects on overall heart pumping or electrical function at 4 weeks post-injection (Fig. S15).

As a first effort to address whether modest expression of NaV1.5-NT is sufficient to influence cardiac metabolism in vivo, we performed broad LC-MS metabolomic profiling on whole heart tissues from myoAAV-injected mice, along with controls (myoAAV-GFP or vehicle; harvested at 4 weeks post-injection, after a 4 hour fast). Western blot analyses on cardiac tissue lysates collected in parallel show that NaV1.5-NT overexpression in vivo is considerably more modest relative to that achieved with Ad vectors in cultured CMs (Fig. S16). Metabolomic profiling identified 258 quantifiable metabolites, with 19 metabolites showing trending changes after NaV1.5-NT expression (based on unadjusted p<0.05, Fig. 8A and Table S3). Pathway enrichment analyses (MetaboAnalyst⁵⁷) performed on metabolites showing trending changes (unadjusted p<0.2, NaV1.5-NT versus controls) revealed enrichments for decreased pentose phosphate pathway (PPP) and glycolytic intermediates (Fig. 8B), as well as reductions in several saturated medium/long-chain fatty acids (Fig. 8B–C) in AAV-NaV1.5-NT treated mouse hearts, possibly signaling increased fatty acid oxidation (FAO) with concomitant decreased glucose utilization. In addition, elevated metabolites were broadly enriched for amino acids, including branched-chain amino acids (BCAAs, Fig. 8B–C). Together, these overall findings could be considered consistent with NaV1.5-NT causing increased mitochondrial respiration favoring complex I-linked FAO and decreased reliance on reciprocally controlled metabolic pathways (e.g. lowering glycolytic and PPP flux and demand for BCAA catabolism)^58–62. It is important to note that steady-state pools do not directly measure pathway flux and that these signature changes could be explained by other possible metabolic alterations. Nevertheless, these data provide initial support that even modest cardiomyocyte-targeted NaV1.5-NT overexpression is sufficient to drive metabolic reprogramming in vivo.

Figure 8: Moderate cardiomyocyte-targeted NaV1.5-NT overexpression in mice is sufficient to rewire the cardiac metabolome.

LC-MS metabolomics measurements were taken from AAV-NaV1.5-NT (NT, n=8) or control (AAV-GFP and vehicle, n=4 each) heart tissues 4 weeks post-AAV injection. A, Volcano plot depicting measured metabolites, highlighting those showing trending changes (decreased metabolites colored in blue and increased colored in red; based on unpaired t-test p<0.05, unadjusted). B, Pathway enrichment analysis (SMPDB or KEGG) performed on trending decreased (left) and increased (right) metabolites (p<0.2, unadjusted). Pathway enrichment FDR<10% for the top three decreased pathways and top four increased pathways. C, Violin plot depicting the relative levels of selected metabolites (plotted relative to control). BCAAs refers to combined summation of valine, leucine and isoleucine for each sample. P-values were determined by unpaired t-test.

DISCUSSION

Through careful examination of genome annotation tracks and various RNA-sequencing datasets, we discovered that the SCN5A gene contains a well-conserved and functional polyadenylation signal downstream of the second exon. This leads to the biosynthesis of a truncated transcript isoform that is decreased in failing human hearts, perhaps hinting at a relevant role in heart disease risk and/or pathogenesis. Notably, this short transcript isoform generates a protein homologous to the N-terminus of NaV1.5 but with a unique C-terminus, which we call NaV1.5-NT. This protein appears to be stable when expressed in isolated cardiomyocytes and in mouse hearts, the latter of which shows two bands running at 17kDa (expected singlet size) and 34kDa (possible SDS-resistant dimer). It is unclear whether NaV1.5-NT functions as a monomer and/or in dimeric form or influences potential interacting partners, as we are unable to detect dimeric 34kDa NaV1.5-NT in cultured NRCMs. Nevertheless, the presence of endogenous NaV1.5-NT in humans is supported by western blot on human cardiac tissue samples and bolstered by the detection of several peptides identified in publicly available mass spectrometry data collected from human cells and tissues. While this evidence is highly supportive of the natural existence of NaV1.5-NT in human tissue, future studies will need to implement high-resolution mass spectrometry of human heart tissues with proper control peptides.

NaV1.5-NT localizes to the mitochondrial matrix where it influences mitochondrial respiration. Baseline mitochondrial oxygen consumption is strikingly elevated upon NaV1.5-NT expression in NRCMs and hIPSC-CMs, with increases in ATP-linked and maximal respiration. BN-PAGE analyses of OXPHOS complexes shows increased levels of disassembled complex V in NRCMs, which could lead to elevated proton leak via the low-conductance pore forming F₀ subunit, resulting in enhanced oxygen consumption⁵⁵ that can be blunted by oligomycin (binds to the F₀ pore to inhibit proton transport⁶³). While we did not find altered mitochondrial membrane potential after NaV1.5-NT expression, subtle differences can often be missed due to the highly dynamic nature of mitochondrial membrane potential. Future experiments assessing real-time mitochondrial membrane potential kinetics and/or electrophysiological recordings of isolated mitoplasts will be needed to further corroborate whether NaV1.5-NT overexpression promotes proton leak via disassembled complex V.

Concurrently, we also found that NaV1.5-NT expression in NRCMs increases levels of complex I-containing supercomplexes, while hIPSC-CMs showed elevations in free complex I. While this could be secondary to enhanced proton leak, whereby complex I compensates to maintain membrane potential, another possible explanation is that NaV1.5-NT plays a more direct and multifaceted context-dependent role in coordinating complex V and complex I supercomplex (dis)assembly and/or stability. Our initial attempts to co-immunoprecipitate NaV1.5-NT interacting proteins were unsuccessful due to poor antibody specificity and very low capture efficiency (<30%). Notably, our BN-PAGE analyses show that NaV1.5-NT appears to smear with high molecular weight CI supercomplexes and complex V subunits in NRCMs (Fig. 6B). Interestingly, interactome data from BioGRID⁶⁴ identifies ATP5H (a complex V peripheral stalk protein) and ATP5A1 (F₁ subcomplex alpha subunit; interacts with the peripheral stalk oligomycin sensitivity conferral protein, OSCP) as a potential interactors with NaV1.2 and/or NaV1.3 N-terminal regions, which share high sequence similarity with NaV1.5. Based on this, NaV1.5-NT might interact with the complex V peripheral stalk to influence F₁-F₀ assembly / stability, however follow-up studies using panels of epitope-tagged constructs suitable for co-immunoprecipitation and/or BioID proximity labeling will be needed to rigorously define NaV1.5-NT protein interactors.

Our unbiased metabolomics data provide initial support for NaV1.5-NT action in vivo. Modest NaV1.5-NT overexpression caused decreases in several saturated fatty acid species and glycolytic / PPP intermediates, as well as increased BCAAs, hinting at enhanced fatty acid utilization relative to other fuel sources. While this could reflect a boost in mitochondrial respiration (complex I activity), NaV1.5-NT did not alter NAD+/NADH levels in mouse hearts, as was observed in NRCMs. Nevertheless, several of these observations would be considered consistent with the Randle cycle (glucose-fatty acid usage switch), as well as other prior literature describing similar effects (in models with increased FAO) or opposing effects (with FAO inhibition using etomoxir)^58–62. Ultimately, follow-up investigations will be needed to more rigorously characterize metabolic alterations in AAV-NaV1.5-NT treated mouse hearts using metabolite tracing methods and cardiac myofiber respiration studies.

Our findings may also have notable relevance to other reported studies examining the possible therapeutic effects of an artificially engineered NaV1.5 N-terminal fragment (termed Nter). Interestingly, these studies support that Nter expression increases NaV1.5 levels and current densities in heterologous cell culture systems, perhaps through acting as a chaperone^45,46. Most recently, Gizon et al. report data supporting that AAV-mediated expression of Nter is sufficient to rescue NaV1.5 surface localization in SCN5A haploinsufficient mice and iPS-cardiomyocytes, although the specific molecular mechanisms remain unknown⁴⁷. Intriguingly, we find that Nter, like NaV1.5-NT, co-localizes with mitochondria in NRCMs, raising interesting questions regarding proposed actions of Nter on NaV1.5 levels and activities. In addition, some discrepancy remains given that NaV1.5-NT expression does not appear to influence NRCM Na⁺ currents in our studies. Future experiments will need to carefully resolve this, with the exciting possibility of revealing an expanded NaV1.5-mitochondrial crosstalk beyond Na⁺/Ca²⁺ balance.

In summary, we serendipitously discovered a novel APA signal in SCN5A, which regulates full-length SCN5A mRNA expression and generates a truncated transcript isoform encoding a mitochondrial-targeted N-terminal fragment of NaV1.5 that influences respiration via complex V and/or complex I. Future studies in cardiomyocytes (human iPS-derived) and SCN5A APA-KI mice will need to examine how SCN5A APA is influenced in models of cardiac stress, and whether there is regulatory interplay with known modulators of NaV1.5 activity, e.g. NAD+^16,19 and ROS¹⁷. Moreover, our work raises further questions regarding how many other classic cardiac disease-related genes might be regulated by APA and/or give rise to alternative short protein fragments. The uniqueness of functional APA signals in far upstream regions remains largely unclear in cardiomyocytes, since these have not been thoroughly examined in human heart 3’-end-seq datasets³⁷. In addition, extensive follow-up experiments are needed to expand our studies to rigorously assess translationally relevant outcomes by examining the long-term effects of AAV-NaV1.5-NT expression in mouse hearts at baseline and under stress (e.g. ischemic or pressure-overload), performing echocardiography, cardiac myofiber respiration assays, metabolite tracing, deep mitochondrial phenotyping, and unbiased metabolomics to gain additional insights into NaV1.5-NT physiologic and molecular actions. Finally, our studies highlight a novel direct route for SCN5A gene products to impact mitochondria, complementing possible indirect influences linked to NaV1.5 activity, and future work should assess the potential bimodal interplay of these mechanisms by testing if NaV1.5-NT actions are coordinately influenced by mitochondrial Na⁺/Ca²⁺.

Supplementary Material

Document S1:

Supplemental Materials and Methods

Figures S1–16

Table S1–2

Major Resources Table

References 66–80.

Document S2: Uncropped Western Blots

Document S3: LC-MS metabolomics data

NOVELTY AND SIGNIFICANCE.

What Is Known?

• Heart failure and arrhythmias are major contributors to sudden cardiac death, which remains a leading mortality cause worldwide, yet many related underlying biological mechanisms remain to be elucidated.

• The SCN5A gene encodes the primary cardiac voltage-gated sodium channel, NaV1.5, that is responsible for inward sodium current triggering the cardiac action potential with each heartbeat; mutations in SCN5A and genetic variants altering SCN5A expression are linked to arrhythmias and heart failure mortality risk.

• Recent evidence suggests that inward sodium passing through NaV1.5 modulates cardiomyocyte metabolism through influence on mitochondrial sodium-calcium exchange, which could influence cardiac disease risk and outcomes.

What New Information Does This Article Contribute?

• The SCN5A gene contains an overlooked conserved alternative polyadenylation signal capable of producing a short transcript isoform encoding a N-terminal fragment of NaV1.5 (NaV1.5-NT).

• NaV1.5-NT localizes to mitochondria in cardiomyocytes and increases mitochondrial respiration, energy production, and reactive oxygen species generation, potentially through modulation on electron transport chain and ATP synthase complexes.

• NaV1.5-NT overexpression in mouse heart is sufficient to rewire cardiac metabolism, possibly by promoting fatty acid oxidation relative to utilization of other fuel sources.

• This work unveils an unexpected direct route for SCN5A gene products to impact mitochondria, perhaps providing additional means to modulate cardiomyocyte metabolism to complement or balance the effects mediated by inward Na+ through NaV1.5.

This manuscript describes our serendipitous discovery of a novel alternative polyadenylation signal in the SCN5A locus that influences the expression of full-length SCN5A mRNA and produces a short transcript isoform encoding an N-terminal fragment of NaV1.5 (NaV1.5-NT). Our initial characterizations demonstrate that NaV1.5-NT localizes to the mitochondrial matrix and enhances cardiomyocyte respiration, providing a direct and unexpected mode of SCN5A-mitochondrial crosstalk that expands the interface beyond mechanisms of sodium-calcium exchange.

NON-STANDARD ABBREVIATIOS AND ACRONYMS

MTS

mitochondrial targeting sequence

APA

alternative polyadenylation

polyA

polyadenylation

OXPHOS

oxidative phosphorylation

PPP

pentose phosphate pathway

FAO

fatty acid oxidation

BCAAs

branched-chain amino acids

HF

heart failure

DCM

dilated cardiomyopathy

ICM

ischemic cardiomyopathy

HCM

hypertrophic cardiomyopathy

NRCM

neonatal rat cardiomyocytes

hiPSC-CMs

human induced pluripotent stem cell-derived cardiomyocytes

AAV

adeno-associated virus

Ad

adenovirus

BN-PAGE

blue native polyacrylamide gel electrophoresis