Disruption of protein quality control of the human ether-à-go-go related gene K⁺ channel results in profound long QT syndrome

Abstract

BACKGROUND

Long QT syndrome (LQTS) is a hereditary disease that predisposes patients to life-threatening cardiac arrhythmias and sudden cardiac death. Our previous study of the human ether-à-go-go related gene (hERG)–encoded K⁺ channel (K_v11.1) supports an association between hERG and RING finger protein 207 (RNF207) variants in aggravating the onset and severity of LQTS, specifically T613M hERG (hERG_T613M) and RNF207 frameshift (RNF207_G603fs) mutations. However, the underlying mechanistic underpinning remains unknown.

OBJECTIVE

The purpose of the present study was to test the role of RNF207 in the function of hERG-encoded K⁺ channel subunits.

METHODS

Whole-cell patch-clamp experiments were performed in human embryonic kidney (HEK 293) cells and human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) together with immunofluorescent confocal and high resolution microscopy, auto-ubiquitinylation assays, and co-immunoprecipitation experiments to test the functional interactions between hERG and RNF207.

RESULTS

Here, we demonstrated that RNF207 serves as an E3 ubiquitin ligase and targets misfolded hERG_T613M proteins for degradation. RNF207_G603fs exhibits decreased activity and hinders the normal degradation pathway; this increases the levels of hERG_T613M subunits and their dominant-negative effect on the wild-type subunits, ultimately resulting in decreased current density. Similar findings are shown for hERG_A614V, a known dominant-negative mutant subunit. Finally, the presence of RNF207_G603fs with hERG_T613M results in significantly prolonged action potential durations and reduced hERG current in human-induced pluripotent stem cell–derived cardiomyocytes.

CONCLUSION

Our study establishes RNF207 as an interacting protein serving as a ubiquitin ligase for hERG-encoded K⁺ channel subunits. Normal function of RNF207 is critical for the quality control of hERG subunits and consequently cardiac repolarization. Moreover, our study provides evidence for protein quality control as a new paradigm in life-threatening cardiac arrhythmias in patients with LQTS.

Introduction

The pore-forming α subunit of the K_v11.1 channel (human ether-à-go-go related gene [hERG], encoded by the KCNH2 gene) underlies the rapid component of the delayed rectifier K⁺ current (I_Kr).^1,2 The hERG-encoded K⁺ channel is known to play a critical role in ventricular repolarization. Loss-of-function mutations in hERG-encoded K⁺ channels result in decreased I_Kr, delayed repolarization, and the prolonged QT interval characteristic of long QT syndrome (LQTS) type 2.^3,4

Genetic mutations in cardiac ion channels, collectively known as cardiac ion channelopathies, have been shown to cause a significant percentage of LQTS cases.^5–8 Currently, it is estimated that 1 in 5000 people carry an LQTS mutation. Defects in hERG-encoded K⁺ channels are the second leading cause of LQTS. Moreover, hERG-encoded K⁺ channels represent the most common target for drug-induced LQTS.^1,3,4,9

Previous studies have provided new evidence that LQTS manifests not only from mutations in cardiac ion channels (as in LQT1, LQT2, and LQT3) but also from mutations in ion channel interacting proteins (such as calmodulin in LQT14–LQT16).^10,11 One potential ion channel interacting protein is RING finger protein 207 (RNF207), encoded by its gene located on chromosome 1. RNF207 is a specialized type of zinc finger protein, shown to be associated with prolongation of the QT interval.^12–14

Our previous study revealed RNF207 as a potential modifier of the hERG channel.¹⁵ Specifically, a term female infant presented with perinatal LQTS with recurrent life-threatening ventricular arrhythmias. Using whole-genome sequencing (WGS), a paternally inherited, known pathogenic variant in KCNH2 leading to a missense mutation (p.T613M, hERG_T613M) was identified. In addition, WGS revealed a maternally inherited variant of unknown significance in the RNF207 gene (a frameshift mutation, RNF207_G603fs). The patient was heterozygous for both variants; however, neither parent presented with LQTS. The findings suggest that the patient’s presentation of LQTS may result from the simultaneous inheritance of the heterozygous RNF207 variant with the hERG_T613M channel (Figure 1A).

Figure 1.

A Dominant-negative effect and altered channel gating kinetics by hERG_T613M^. A: Amino acid sequence alignments for hERG_WT and hERG_T613M (residues 596–630) and RNF207_WT and RNF207_G603fs (residue 596 to the end of the polypeptide). The conserved selectivity filter of the hERG subunit GFG is underlined. The asterisk indicates mutated residues (613 in the hERG sequence and 603 in the RNF207 sequence). B: Representative recordings from HEK 293 cells expressing hERGWT alone, hERGWT:hERGT613M, hERGT613M alone, or nontransfected cells. C: Summary data of current densities for hERGWT alone (black traces) compared with hERG_WT:hERG_T613M (blue traces). n = 20–30 cells; *P < .05. D: Summary data for voltage-dependent activation using the peak tail current density fitted using the Boltzmann function (see Online Supplemental Table 1). E–H: Time constants of inactivation, recovery from inactivation, activation, and deactivation, respectively, from hERG_WT alone (black traces) compared with hERG_WT:hERG_T613M (blue traces). The insets show voltage-clamp protocols and representative current traces. n = 6–11 for inactivation; n = 9–11 for recovery from inactivation; n = 3 for activation; n = 5–10 for deactivation. *P < .05, **P < .01. Data shown are mean ± SEM. Analyses were performed using the Student t test. Expanded current traces in insets in panels E–H are shown in Online Supplemental Figure 3. ANOVA = analysis of variance; HEK 293 = human embryonic kidney 293; hERG = human ether-à-go-go related gene; I = current; ms = millisecond; mV = millivolt; pA/pF = picoampere per picofarad; RNF207 = ring finger protein 207; s = second; SEM = standard error of the mean; V = voltage; WT = wild-type.

Previous literature has shown that the T613M hERG mutation (hERG_T613M, located in the CpG sequence of the pore helix, Figure 1A) produces no detectable current and very low surface expression, suggesting a trafficking defect.¹⁶ Coexpression of hERG_T613M with wild-type hERG (hERG_WT) subunits presented very low current density, indicating that hERG_T613M may exert a dominant-negative effect when forming heterotetramers with the hERG_WT subunit.¹⁶ There is evidence for an interaction between hERG_WT and RNF207 proteins via RNF207’s RING domain.¹² This RING domain, as well as RNF207’s structural similarity to tripartite motif-containing proteins, suggests a potential function of RNF207 as an E3 ubiquitin ligase and a possible role in facilitating protein degradation.¹⁷ However, the exact molecular mechanisms remain unexplored.

Misfolded proteins are frequently degraded by endoplasmic reticulum (ER)–associated degradation (ERAD).^17–20 After polyubiquitination, proteins are dislocated from the ER and degraded by proteasome within the cytosol. Several quality control proteins targeting hERG-encoded K⁺ channel subunits have been identified.^20,21 RNF207 has been shown to interact with chaperones heat shock protein, involved in the regulation of hERG subunits.^12,22,23 Therefore, RNF207 may play a critical role in ER-associated degradation.

Here, we demonstrate that RNF207 serves as one of the ubiquitin ligases and targets misfolded hERG_T613M proteins for degradation. Mutant RNF207 (RNF207_G603fs) exhibits decreased activity and hinders the normal degradation pathway; this increases the levels of hERG_T613M subunits and their dominant-negative effect on the WT subunits, ultimately resulting in decreased current density. Our study provides novel mechanisms whereby dysfunction in degradation-dependent quality control plays a key role in aggravating the effects of existing LQTS mutations.

Methods

Please see detailed Materials and Methods in the Online Supplement.

Results

hERG_T613M exerts a dominant-negative effect when coexpressed with hERG_WT

Human embryonic kidney 293 cells were transfected with hERG and an accessory subunit, KCNE2 (MiRP1).^24,25 The hERG_T613M subunit failed to conduct I_Kr because of membrane trafficking defects (Figures 1B and 1D; Online Supplemental Figures 1 and 2), consistent with previous studies.¹⁶ Coexpression of hERG_WT with hERG_T613M only partially rescued I_Kr with significantly reduced current density compared with expression of hERG_WT subunits alone (Figures 1B–1D). Moreover, the hERG_WT:hERG_T613M current exhibited significantly faster time constants for activation and inactivation and significantly slower time constants for deactivation and recovery from inactivation (Figures 1E–1H and Online Supplemental Figure 3), supporting the formation of heterotetramers between WT and mutant subunits.

hERG-encoded K⁺ channel subunits colocalize with RNF207_WT

Immunofluorescence experiments in guinea pig and rabbit ventricular cardiomyocytes showed strong colocalization of hERG and RNF207 proteins as well as colocalization of hERG and RNF207 with α-actinin2, used to mark the z lines (Figure 2A and Online Supplemental Figures 4 and 5). Both hERG and RNF207 proteins localized near the z lines (Figure 2A and Online Supplemental Figure 4C), consistent with the previous literature.²⁶ We further used proximity ligation assay (PLA), which detects whether 2 proteins of interest are ≤40 nm apart (Figures 2B and 2C).²⁷ Robust PLA signals were observed in guinea pig ventricular myocytes co-labeled with anti-hERG, anti-RNF207, and anti–α-actinin2 antibodies (Figure 2B and Online Supplemental Movie 1). In contrast, PLA signals were completely absent when one of the primary antibodies was omitted (Figure 2C). The results further support a close association (≤40 nm) between hERG-encoded K⁺ channel subunits and RNF207.

Figure 2.

Colocalization of hERG and RNF207 in guinea pig ventricular cardiomyocytes. A: Confocal images showing colocalization among α-actinin2, hERG K⁺ channel subunits, and RNF207. Scale bar = 10 μm. The right panels show the corresponding fluorescence intensity profiles perpendicular to the z lines. B: Proximity ligation assay (PLA) for α-actinin2, hERG K⁺ channel subunits, and RNF207. C: Quantification of PLA signals per cell area (puncta/μm²). n = 15, 11, 15, 10, 9, and 9 cells from left to right bars; *P < .05. D: Auto-ubiquitinylation assay for RNF207_WT (lane 3, right) vs negative control (lane 1, left) and MDM2, a known E3 ubiquitin ligase (positive control, lane 2). Transfected HEK 293 cells were immunoprecipitated (IP) for RNF207-FLAG. An auto-ubiquitinylation assay was conducted on isolated protein, followed by SDS-PAGE and Western blot analysis (IB). Proteins were incubated in the presence of E1 and E2 ubiquitin enzymes, ubiquitin, and ATP. E: Ubiquitination assays. Ubiquitinated proteins were absent in the negative control (lane 1). The E3-ubiquitin band appeared for hERG_T613M incubated with RNF207_WT (lane 3), but not in the presence of RNF207_G603fs (lane 5) or with hERG_WT subunits (either with RNF207_WT [lane 2] or with RNF207_G603fs [lane 4]). F: Quantification of the data from panel E. n = 5 independent experiments for each group; *P < .05. Data shown are mean ± SEM. Analyses were performed using 1-way analysis of variance (ANOVA) with Brown-Forsythe post hoc analyses. AU = arbitraty units; HEK 293 = human embryonic kidney 293; hERG = human ether-à-go-go related gene; RNF207 = ring finger protein 207; SDS-PAGE = sodium dodecyl sulfate-polyacrylamide gel electrophoresis; SEM = standard error of the mean; Ub = ubiquitin; WT = wild-type.

RNF207_G603fs exhibits decreased E3 ubiquitin ligase activity

Using an auto-ubiquitinylation assay, we demonstrated that RNF207 is, indeed, an E3 ubiquitin ligase (Figure 2D). Compared with the known auto-ubiquitin ligase MDM2, RNF207_WT displayed strong auto-ubiquitinylation (Figure 2D). In the presence of hERG_WT, we did not observe significant ubiquitination by RNF207_WT (Figure 2E, lane 2). In contrast, incubation of RNF207_WT with hERG_T613M resulted in significant ubiquitination of hERG_T613M subunits (Figure 2E, lane 3). Importantly, RNF207_G603fs failed to tag the hERG_T613M mutant subunit with ubiquitin (Figure 2E, lane 5, and Figure 2F).

RNF207_WT decreases the hERG_T613M membrane bound population

Hemagglutinin (HA) and c-Myc tags were inserted into the S1-S2 linker of hERG_WT and hERG_T613M, respectively, in order to quantify the ratio of hERG subunits on the membrane vs those in the cytosol (nonpermeabilized to permeabilized fluorescence ratios, Figures 3A and 3B).²⁸ hERG_T613M subunits failed to show fluorescence signals under nonpermeabilized conditions (Figure 3A). The hERG_WT:hERG_T613M group, however, showed normal fluorescent levels.

Figure 3.

Surface membrane and cytosolic expression of hERG K⁺ channel subunits after coexpression with RNF207_WT vs RNF207_G603fs. A: Immunofluorescence confocal microscopic imaging of HEK 293 cells transfected with different combinations of hERG_WT, hERG_T613M, RNF207_WT, and RNF207_G603fs. NP = nonpermeabilized; P = permeabilized. Scale bar = 10 μm). The last panel in each row represents a higher magnification image from the merged image as outlined in a red box. B: Schematic diagrams of hERG_WT-HA and hERG_T613M-Myc fusion constructs. C: Summary data of the fluorescence ratios (555 nm/633 nm). n = 6–10 cells; *P < .05, **P < .01. Data shown are mean ± SEM. Analyses were performed using ANOVA with Tukey’s post hoc analyses. ANOVA = analysis of variance; DAPI = 4’6-diamidino-2-phenylindole; GFP = green fluorescent protein; HA = hemagglutinin; HEK 293 = human embryonic kidney 293; hERG = human ether-à-go-go related gene; RNF207 = ring finger protein 207; SEM = standard error of the mean; WT = wild-type.

When hERG_WT:hERG_T613M:RNF207_WT were coexpressed (Figures 3A and 3C), hERG_WT-HA showed normal fluorescence on the membrane, whereas hERG_T613M-Myc was primarily seen in the cytosol (Figures 3A and 3C). The addition of RNF207_G603fs, however, resulted in an increase in hERG_T613M subunits on the membrane, suggesting decreased degradation of hERGT613M by RNF207G603fs.

RNF207_WT increases the degradation of hERG_T613M subunits

Brefeldin A, an inhibitor of ER-to-Golgi transport, decreased the fully glycosylated 155 kDa form, while the cytosolic core glycosylated 135 kDa form of hERG subunits increased over time (Figure 4A). There was a significant accumulation of hERG subunits when hERG_WT:hERG_T613M subunits were coexpressed with RNF207_WT:RNF207_G603fs (closed Δ) compared with coexpression with RNF207_WT (closed O) (Figures 4A and 4B). The differences between these 2 groups were reduced by MG 132, a proteasome inhibitor, suggesting that RNF207_WT may function through the ubiquitin proteasome system (Figures 4A and 4C). Similar effects were also seen with bafilomycin A1, a lysosome inhibitor, suggesting an additional role of lysosomal degradation (Figures 4A and 4D).

Figure 4.

Distinct effects of RNF207_WT vs RNF207_G603fs on the degradation of hERG K⁺ channel subunits. A: Degradation assay for HEK 293 cells collected before (time = 0) vs 3, 6, 12, and 24 hours after treatment. Cells were treated with brefeldin A (left panel), brefeldin A + MG 132 (middle panel), or brefeldin A + bafilomycin A1 (Baf A1; right panel). B–D: Summary data from panel A. *P < .05 in panel B. Data shown are mean ± SEM. n = 5 different independent experiments; *P < .05. Analyses were performed using ANOVA with Tukey’s post hoc analyses. ANOVA = analysis of variance; Baf A1 = bafilomycin A1; BFA = brefeldin A; HEK 293 = human embryonic kidney 293; hERG = human ether-à-go-go related gene; IB = immunoblot; kDa = kilodaltons; RNF207 = ring finger protein 207; SEM = standard error of the mean; WT = wild-type.

RNF207 interacts with hERG K⁺ channel subunits via the C terminus

Co-immunoprecipitation experiments revealed that hERG_WT or hERG_T613M successfully immunoprecipitated RNF207_WT, suggesting that the 2 proteins form multiprotein complexes (Figure 5A). RNF207_G603fs showed a weaker interaction with hERG_WT, but also interacted with hERG_T613M.

Figure 5.

Multiprotein complexes formed by hERG K⁺ channel subunits and RNF207. A: Transfected (T_x) HEK 293 cells were immunoprecipitated (IP) using anti-hERG antibody, and Western blot analysis (IB) was performed using anti-hERG and anti-FLAG antibodies to target the RNF207-FLAG fusion protein. Western blot analyses from lysates and IP are shown in lanes 1–5 and lanes 6–10, respectively. The negative control using IgG for IP is shown in lane 11. B: Schematic of hERG_WT ΔN, hERG_WT ΔC, hERG N-terminus, and hERG C-terminus constructs (tagged with HA). C: Coimmunoprecipitation of hERG-HA fragments that were cotransfected (Tx) in HEK 293 cells with RNF207_WT-FLAG. Immunoprecipitation (IP) followed by immunoblotting (IB) was performed using anti-HA (IP:HA, upper panel) and anti-FLAG (IB:FLAG, upper panel) antibodies, respectively. The reverse experiments were conducted as shown in the lower panel. Lanes 1–4 are lysate samples (20 μg each), and lanes 5–8 show immunoprecipitated samples transfected with hERG_WT ΔN, hERG_WT ΔC, hERG N-terminus, hERG C-terminus, and IgG negative control. Some nonspecific bands are seen in lanes 5–8 in the lower panel. The blue arrow in lane 5 in the lower panel shows a band of the expected size for hERG_WT ΔN. D: Rosetta model of hERG (from top to bottom): (1) view from the extracellular side of the membrane, (2) view from the transmembrane side, and (3) view from the intracellular side of the membrane. The backbone is shown in ribbon representation and colored by the rainbow color scheme from the N terminus (blue) to the C terminus (red). The side chain of the T613 residue in each subunit is shown using a space filling representation. ANOVA = analysis of variance; cyt = cytosolic; HA = hemagglutinin; HEK 293 = human embryonic kidney 293; hERG = human ether-à-go-go related gene; IB = immunoblot; IP = immunoprecipitation; kDa = kilodaltons; RNF207 = ring finger protein 207; SEM = standard error of the mean; Tx = transfection; WT = wild-type.

We designed 4 additional hERG-HA constructs: hERGΔN, hERG ΔC, hERG N terminus, and hERG C terminus (Figure 5B). We coexpressed RNF207_WT-FLAG with the hERG fragments to determine the interaction domain of the hERG channel with RNF207 (Figure 5C). The lysate samples are shown in lanes 1–4. hERG ΔN and hERG C terminus were able to successfully immunoprecipitate RNF207_WT (Figure 5C, upper panel, lanes 5 and 8), while hERG ΔC and hERG N terminus failed to pull down RNF207 (Figure 5C, upper panel, lanes 6 and 7). The reverse experiments were performed in the lower panel. There are some nonspecific bands from the immunoprecipitated samples (Figure 5C, lower panel, lanes 5–8); however, a distinct band of the expected size could be discerned in lane 5, corresponding to hERG_WT ΔN. The data suggest that the C terminus of hERG_WT plays an important role in its interaction with RNF207_WT.

To determine the mechanistic underpinning of hERG_T613M on the alterations in the time-dependent kinetics of the channel (Figures 1E–1H), we took advantage of the published cryogenic-electron microscopy (cryo-EM) of the hERG structure²⁹ using Rosetta modeling software (Figure 5D). The hERG backbone is shown with the rainbow color scheme from the N terminus (blue) to the C terminus (red). The T613 residue is seen to position near the extracellular side of the pore, which likely plays critical roles in channel activation and inactivation kinetics as demonstrated in Figures 1E–1H.

RNF207_G603fs decreases the I_Kr density

Expression of hERG_WT with RNF207_WT exhibited the normal hERG current density (Figures 6A–6C, black traces). Although hERG_T613M and hERG_WT coexpression produces very low current density with altered kinetics (Figures 1B–1F), the addition of RNF207_WT rescued the current density to near hERG_WT levels (Figures 6A–6C, red traces), suggesting the enhanced degradation and subsequent decrease in dominant-negative effects from hERG_T613M subunits. Additionally, the time- and voltage-dependent kinetics were restored to near hERG_WT levels (Figures 6D–6G), consistent with the notion that the functional channels consist of mostly hERG_WT subunits in the form of homotetramers instead of heterotetramers as in Figures 1E–1H. In contrast, RNF207_G603fs failed to restore the current density, despite the coexpression with RNF207_WT (Figures 6A–6C, blue traces). The rectification also decreased (Figure 6B). The time constants of recovery from inactivation (closed Δ, Figure 6E) and the slow component of deactivation (open Δ, Figure 6G) were significantly slower in the presence of RNF207_G603fs, while the activation time constant was significantly faster (Figure 6F).

Figure 6.

Differential effects of RNF207_WT vs RNF207_G603fs on the hERG current. A: Representative whole-cell voltage-clamp recordings. B: Summary data of current density for hERG_WT:RNF207_WT (black traces) compared with hERG_WT:hERG_T613M:RNF207_WT (red traces) and hERG_WT:hERG_T613M:RNF207_WT:RNF207_G603fs (blue traces). n = 14–19.C: Summary data for voltage-dependent activation using the peak tail current density fitted using the Boltzmann function (see Online Supplemental Table 1). D and E: Time constants of inactivation and recovery from inactivation for hERG_WT:RNF207_WT (black traces) compared with hERG_WT:hERG_T613M:RNF207_WT (red traces) and hERG_WT:hERG_T613M:RNF207_WT:RNF207_G603fs (blue traces). F and G: Time constants for activation and deactivation, where closed Δ, □, and O represent the fast component and open Δ, □, and O represent the slow component of deactivation. n = 5–10 for inactivation; n = 6–12 for recovery from inactivation; n = 3–5 for activation; n = 5–10 for deactivation. *P < .05, **P < .01. In panels B and C, **P < .01 for hERG_WT:hERG_T613M:RNF207_WT:RNF207_G603fs compared with hERG_WT:RNF207_WT and hERG_WT:hERG_T613M:RNF207_WT throughout positive voltages and was shown only at the end of the curves for clarity. Data shown are mean ± SEM. Analyses were performed using ANOVA with Tukey’s post hoc analyses. ANOVA = analysis of variance; HEK 293 = human embryonic kidney 293; hERG = human ether-à-go-go related gene; RNF207 = ring finger protein 207; SEM = standard error of the mean; WT = wild-type.

To further evaluate whether the RNF207_G603fs variant may affect the other known dominant-negative hERG mutant subunit, we generated the hERG_A614V mutant subunit, a known dominant-negative subunit³⁰ (Online Supplemental Figure 6). Expression of hERG_WT with RNF207_WT exhibited normal hERG current density (Online Supplemental Figures 6A–6C, black traces). In contrast, cells expressing hERG_WT:hERG_T613M:RNF207_WT:RNF207_G603fs showed a significantly reduced current density (Online Supplemental Figures 6A–6C, blue traces). Coexpression of hERG_WT and hERG_A614V with RNF207_WT rescued the abnormal current (Online Supplemental Figures 6A–6C, red traces), suggesting that RNF207 serves as a quality control mechanism for LQT2 that is applicable to other KCNH2 variants.

Coexpression of hERG_T613M and RNF207_G603fs in human-induced pluripotent stem cell–derived cardiomyocytes results in prolonged action potential durations and a significant decrease in E-4031–sensitive currents

We took advantage of human-induced pluripotent stem cell–derived cardiomyocytes (hiPSC-CMs, iCell, FUJIFILM Cellular Dynamics, Inc., Madison, WI) as a platform (Figures 7A–7F). Cells expressing hERG_WT:hERG_T613M:RNF207_WT:RNF207_G603fs (denoted as the “mutant” group) showed significantly longer action potential durations at 50% and 90% repolarization compared with cells expressing hERG_WT:RNF207_WT (WT group) or nontransfected cells (Figures 7A–7C). More importantly, expression of RNF207_WT resulted in the “rescue” of delayed repolarization in the mutant group (hERG_WT:hERG_T613M:RNF207_WT, the “rescue” group), consistent with the findings from Figure 6.

Figure 7.

Regulation of APDs of hiPSC-CMs by RNF207. A: Representative action potential recordings (iCell, Cellular Dynamics) in cells expressing hERG_WT:RNF207_WT (black trace), hERG_WT:hERG_T613M:RNF207_WT (red trace), and hERG_WT:hERG_T613M:RNF207_WT:RNF207_G603fs (blue trace) as well as a nontransfected cell (gray trace). B–F: Summary data for action potential recordings in nontransfected cells (labeled “Non-TF”; gray bar) compared with hERG_WT:RNF207_WT (labeled “WT”; black bar), hERG_WT:hERG_T613M:RNF207_WT (labeled “Rescue”; red bar), and hERG_WT:hERG_T613M:RNF207_WT:RNF207_G603fs (labeled “Mutant”; blue bar) at baseline (solid bars) vs 1 μM E-4031 (striped bars). Data are shown for average diastolic potential (panel B), peak action potential (panel C), action potential amplitude (panel D), action potential duration at 50% repolarization or APD₅₀ (panel E), and action potential duration at 90% repolarization or APD₉₀ (panel F). Data shown represents the average of 5 action potentials per cell, with n = 6–9 cells for baseline recordings and n = 3–5 cells for E-4031 recordings. *P<.05, **P<.01, §P<.001. Data shown are mean ± SEM. Analyses were performed using ANOVA with Tukey’s post hoc analyses. ANOVA = analysis of variance; APD = action potential duration; APD50 and APD90 = APD at 50% and 90% repolarization; hERG = human ether-à-go-go related gene; hiPSC-CM = human induced pluripotent stem cell-derived cardiomyocytes; non-TF = nontransfected cells; RNF207 = ring finger protein 207; SEM = standard error of the mean; WT = wild-type.

Finally, we recorded the E-4031–sensitive currents from the same 3 groups of hiPSC-CMs as in Figure 8. There was a significant decrease in E-4031–sensitive current in hiPSC-CMs expressing hERG_WT:hERG_T613M:RNF207_WT:RNF207_G603fs compared with hERG_WT:RNF207_WT. Importantly, expression of RNF207_WT resulted in the “rescue” of the E-4031–sensitive hERG current (hERG_WT:hERG_T613M:RNF207_WT).

Figure 8.

Regulation of hERG currents by RNF207 in hiPSC-CMs and a schematic diagram of RNF207 interaction with hERG-encoded K⁺ channels in adult ventricular myocytes. A: Representative E-4031–sensitive currents recorded from hiPSC-CMs expressing hERG_WT:RNF207_WT (black traces), hERG_WT:hERG_T613M:RNF207_WT (red traces), and hERG_WT:hERG_T613M:RNF207_WT:RNF207_G603fs (blue traces). B: Summary data of current density for the 3 groups of cells. n = 5–6. C: Summary data for voltage-dependent activation using the peak tail current density fitted using the Boltzmann function (see Online Supplemental Table 1). n = 5–6. In panel B, *P < .05 for hERG_WT:RNF207_WT compared with hERG_WT:hERG_T613M:RNF207_WT:RNF207_G603fs throughout positive voltages and was shown only at the end of the curve for clarity. In panel C, *P < .05 for hERG_WT:RNF207_WT and hERG_WT:hERG_T613M:RNF207_WT compared with hERG_WT:hERG_T613M:RNF207_WT:RNF207_G603fs throughout positive voltages and was shown only at the end of the curves for clarity. Analyses were performed using ANOVA with Tukey’s post hoc analyses. D: Schematic diagram of RNF207 interaction with hERG-encoded K⁺ channels (K_v11.1) with trafficking and degradation pathways (generated using BioRender, Toronto, Canada). ANOVA 5 analysis of variance; hERG = human ether-à-go-go related gene; hiPSC-CM 5 human induced pluripotent stem cell-derived cardiomyocytes; I = current; RNF207 = ring finger protein 207; SEM = standard error of the mean; V = voltage; WT = wild-type.

Discussion

While the majority of LQTS cases are attributed to defects in the ion channels, previous studies have provided new evidence where mutations in ion channel interacting proteins act as potential culprits in the development of LQTS. Calmodulin’s role in LQT14, LQT15, and LQT16 is 1 such example.¹⁰ Here, we propose that quality control mechanisms play a key role in either aggravating or appeasing the existing ion channel mutations.

Protein quality control as a new paradigm for cardiac arrhythmias in LQTS

In our previously published patient,¹⁵ we observed severe LQTS with early onset. WGS identified that the patient was heterozygous for the hERG_T613M and RNF207_G603fs variants. The absence of LQTS symptoms in the parents suggests that the combination of these 2 mutations may be responsible for aggravating the LQTS phenotype. Our present study suggests that quality control mechanisms, including degradation by ubiquitin ligases such as RNF207, are critically important in regulating the impacts of these ion channel mutations. Indeed, we demonstrated that RNF207_WT rescued the detrimental effects of hERG_T613M coexpression with hERG_WT while the addition of RNF207_G603fs failed to rescue I_Kr and aggravate the LQTS phenotype. Our results support the underlying hypothesis that quality control mechanisms are critical regulators of ion channel function and consequently disorders such as LQTS and sudden cardiac death.

Identification of RNF207 as an E3 ubiquitin ligase

We demonstrated that RNF207 colocalizes with the hERG channel, near the z lines, in ventricular cardiomyocytes. We confirmed that RNF207_WT functions as an E3 ubiquitin ligase for hERG_T613M, primarily through the ubiquitin proteasome system. Co-immunoprecipitation experiments confirm that RNF207 and hERG reside within the same multiprotein complex. RNF207_WT successfully rescues the hERG current density via ubiquitination when WT and mutant subunits are coexpressed, whereas RNF207_G603fs fails to ubiquitinate and rescue I_Kr.

Functional roles of T613 residues in the time-dependent kinetics of hERG channels

Previous literature has shown that hERG_T613M fails to traffic and produce currents when expressed alone but exhibits decreased current density when coexpressed with hERG_WT, consistent with our findings.¹⁷ Additionally, hERG_T613M, when coexpressed with hERG_WT, alters the time- and voltage-dependent kinetics. Specifically, the activation and inactivation time constants are significantly faster while the deactivation and recovery from inactivation are significantly slower when hERG_T613M and hERG_WT are coexpressed. The changes in inactivation are in accordance with our molecular model of hERG_T613M, as the rectification from the hERG channel’s inactivation process has been shown to be similar to C-type inactivation, which takes place near the outer region of the pore helix.³¹

Findings from hiPSC-CMs show the unexpected effects of mutant subunits on the diastolic potentials. Future studies are required to explore other K⁺ currents including inwardly rectifying K⁺ current (I_K1) that may contribute to the observed findings.

Findings from previous studies

Previous genome-wide association studies have identified single nucleotide polymorphisms within the gene encoding RNF207 and QT interval prolongation.^12–14 Our findings are consistent with a previous study showing that RNF207 and the hERG-encoded K⁺ channel interact and colocalize.¹² The previous study suggests that RNF207 overexpression significantly increases hERG protein trafficking, membrane expression, and the current density. Our present study provides additional evidence for a novel role of RNF207 as a channel interacting partner that serves critical roles in the quality control of the proteins.

Conclusion

New knowledge of the intricate and precise regulatory mechanisms of ion channels will provide an enormous opportunity to uncover new therapeutic targets that may serve to “fine-tune” ion channel function, instead of “blocking” ion channels directly as in our current forms of antiarrhythmic drugs, thus serving as a paradigm shift in our new rationale for the development of antiarrhythmic therapy.