This study was approved by the Gunma University Ethical Review Board for Medical Research Involving Human Subjects (approval number: HS 2017–15). Written informed consents of the subjects (the proband and her mother) were obtained from the proband’s mother. The Genomic DNA was extracted from peripheral blood lymphocytes. Target panel sequencing of 72 genes, including inherited arrhythmia syndrome-related genes, was performed using the proband’s sample as previously described (Nakajima et al., 2020a). Confirmation of nucleotide substitutions in the proband and her mother was performed using Sanger sequencing. Exon 5 of KCNQ1 (NM_000218.3) was analyzed as previously described (Imai et al., 2014).

Full-length cDNA encoding human wild-type (WT)-KCNQ1 subcloned into a pIRES2-EGFP expression vector (WT-KCNQ1 cDNA) was kindly provided by Prof. Jacques Barhanin (Laboratories of Excellence, Ion Channel Science and Therapeutics, Nice, France). Q234K-KCNQ1 cDNA was constructed using a Quick Change II XL site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA, United States). Full-length cDNA encoding human KCNE1 subcloned into the pCDNA3.1 expression vector (KCNE1 cDNA) was obtained as described previously (Wu et al., 2014).

For electrophysiological characterization of WT-KCNQ1 (WT alone) channels and Q234K-KCNQ1 (Q234K alone) channels, 0.5 μg of WT-KCNQ1 cDNA alone or 0.5 μg of Q234K-KCNQ1 cDNA alone was transiently transfected into tsA201-cells using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, United States). For electrophysiological characterization of WT-KCNQ1+KCNE1 (WT + E1) channels and Q234K-KCNQ1+KCNE1 (Q234K + E1) channels, 0.5 μg of WT-KCNQ1 cDNA or 0.5 μg of Q234K-KCNQ1 cDNA in combination with 0.5 μg KCNE1 cDNA was transiently transfected. For electrophysiological characterization of WT-KCNQ1+Q234K-KCNQ1+KCNE1 (WT + Q234K + E1) channels, 0.25 μg of WT-KCNQ1 cDNA plus 0.25 μg of Q234K-KCNQ1 cDNA in combination with 0.5 μg KCNE1 cDNA was transiently transfected. Transfected cells were maintained in DMEM at 37°C for 42–54 h before current recordings as described previously (Nakajima et al., 2022). Cells that exhibited green fluorescence were chosen for the current recordings.

Membrane currents were recorded using whole-cell patch-clamp techniques at room temperature (23°C–25°C). The bath solution contained (in mmol/L) 140 NaCl, 4 KCl, 2 CaCl₂, 1 MgCl₂, 5 glucose and 10 HEPES (pH 7.4 with NaOH), and the pipette solution contained (in mmol/L) 140 KAsp, 1 MgCl₂, 4 MgATP, 10 EGTA and 10 HEPES (pH 7.3 with KOH). The electrode resistance ranged from 2.0 to 3.0 MΩ. Data acquisition was carried out using an Axopatch 200B amplifier and pCLAMP10.3 software (Molecular Devices, Sunnyvale, CA, United States). Currents were acquired at 20–50 kHz, and low pass-filtered at 5 kHz using an analog-to-digital interface (Digidata 1440A acquisition system, Molecular Devices). Current densities at each test potential were obtained by dividing the expressed currents by cell capacitance.

All pulse protocols are shown in each of Figures or Figure legends. The voltage dependence of activation was fitted with a Boltzmann function of the following form: y = 1–1/{1+exp[(V_m-V_1/2)/K]}, where y is the relative current, V_m is the membrane potential, V_1/2 is the voltage at which half of the channels are available to open, and K is the slope factor. The time course of activation was fitted with a single exponential function of the following form: I(t)/I_max = A₀+A₁exp(-t/τ), and the time course of deactivation was fitted with a single exponential function of the following form: I(t)/I_max = A₀+A₁[1-exp(-t/τ)], where A and τ refer to the amplitude and time constant, respectively, I refers to the current, and t refers to the time.

To avoid potential endogenous current contamination, recordings from the cells exhibiting peak outward current amplitudes of <0.8 nA were excluded from the analyses.

In order to understand how the changes of the channel kinetics or current density in I_Ks might affect APD or QT interval, we mimicked the experimental data for the WT or the Q234K model and simulated action potentials across the transmural myocardium or ECG (Itoh et al., 2009; Kojima et al., 2020). O’Hara and Rudy model was available for this simulation (O'Hara et al., 2011), and we searched for appropriate parameters according to the experimental data under the Q234K heterozygous condition. For the analysis of I_Ks, we used the experimental data in which KCNQ1 would be co-expressed with KCNE1. We modified a parameter associated with Hodgkin-Huxley model in I_Ks and simulated the Q234K + WT + E1 model as shown below. W T + E 1 ;   τ x , S 2 = 1 0.01 × ⁡ exp V − 50 20 + 0.0193 × ⁡ exp   − V + 66.54 31 Q 234 K + W T + E 1 ;   τ x , S 2 = 5 0.01 × ⁡ exp V − 50 20 + 0.0193 × ⁡ exp   − V + 66.54 31

To simulate the transmural myocardium and pseudo-ECG, we linearly connected 165 myocardial cells through epicardial (1–60), mid-myocardial (61–105), and endocardial cells (106–165). The cells were stimulated in the endocardial cell for 3 min at the cycle length of 1,000 ms. The pseudo-ECG was simulated using Gima and Rudy model (Gima and Rudy, 2002).

All data are expressed as mean ± standard error, and statistical comparisons were tested using the unpaired Student’s t-test with p < 0.05 considered to be statistically significant. In some figures, the standard error bars are smaller than the data symbols.

A 13-year-old female (II-1-1) (Figure 1B) was referred to our institution because QT prolongation on her ECG had been pointed out at medical check-up in junior high school. Her 12-lead ECG showed QTc prolongation (QTc: 486 ms) (Figure 1C). She experienced an episode of syncope several months ago, but the cause was unknown. Transthoracic echocardiography revealed no structural heart diseases. An exercise stress test using a treadmill did not induce any ventricular arrhythmias, but induced QTc prolongation at peak exercise and at recovery phase of 4-min (QTc: 489 ms) (Figure 1C). Her EEG showed frequent spike-and-wave complexes (Figure 1D), although there was no epileptic seizure while recording EEG.

12-lead ECG of her mother (I-2) (Figure 1B) also showed QTc prolongation (QTc: 475 ms) (Figure 1E). She had experienced several episodes of syncope since she was 1 year of age. Prodromal symptoms, such as blinking her eyes, were present before syncope, and her EEGs recorded repeatedly showed spike-and-wave complexes, which led to a diagnosis of epilepsy at 3 years of age. Sodium valproate had been prescribed for many years, but she discontinued treatment on her own judgment several years ago. She had been asymptomatic since she was 4 years of age. Her EEG recorded at age 39 also showed spike-and-wave complexes (Figure 1F).

Target panel sequencing of 72 genes, including inherited arrhythmia syndrome-related genes, was performed in the proband (II-1) (Figure 1B). A novel KCNQ1 Q234K variant was identified and validated using Sanger sequencing (Figure 1B). Sanger sequencing of the proband’s mother (I-2) (Figure 1B) revealed that she harbored the same variant. This variant was not found in gnomAD (https://gnomad.broadinstitute.org/) and ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/), and was deleterious in SIFT (https://sift.bii.a-star.edu.sg/) and probably damaging in PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2/).

We first compared the electrophysiological properties of the WT alone channels and Q234K alone channels. The current density of the Q234K alone channels, measured at the end of a 2-s depolarizing potential of 100 mV, was not different from that of the WT alone channels (Table 1) (Figures 2A,B), indicating that trafficking of the Q234K alone channels might not be impaired. However, the voltage dependence of activation (VDA) of the Q234K alone channels was significantly shifted to positive potentials (≈5.8 mV) in comparison to the WT alone channels (Table 1) (Figure 2B).

The activating currents were fitted with a single exponential function. The time constants of the Q234K alone channels were significantly larger than those of the WT alone channels, indicating that the Q234K alone channels exhibited a delay of the activation rate in comparison to the WT alone channels (Table 1) (Figures 2A,C). Notably, the WT alone channels exhibited inactivation at higher depolarizing potentials while the Q234K alone channels did not exhibit apparent inactivation (Figure 2A). The absence of inactivation in the Q234K alone channels might, at least in part, contribute to the delayed activation rate in comparison to the WT alone channels.

Tail currents of the WT alone channels, recorded at −50 mV after depolarizing pulses, exhibited hooks (Figure 2A), the initial component of which indicated that the current increase may have resulted from recovery from inactivation, as described previously (Tristani-Firouzi and Sanguinetti, 1998; Franqueza et al., 1999), whereas those of the Q234K alone channels did not exhibit apparent hooks (Figure 2A). The hooks were also present at repolarizing potentials after depolarization in the WT alone channels, but not in the Q234K alone channels (Figure 3A). The presence of the hook in the tail currents of the WT alone channels suggested that these channels inactivated during depolarization and recovered from inactivation during repolarization, which is compatible with the notion that the WT alone channels inactivate during depolarization (Tristani-Firouzi and Sanguinetti, 1998). On the other hand, the lack of a hook in the tail currents of the Q234K alone channels suggested that these channels did not inactivate during depolarization or that they recovered from inactivation very quickly. The disappearance of the hook has also been observed in other variants of KCNQ1 (Franqueza et al., 1999; Moreno et al., 2017).

The deactivation rate was assessed by fitting deactivating currents with a single exponential function. In the WT alone channels, an initial increase in the hook was not included in the fitting procedure (Figure 3A). Intriguingly, the time constants of the Q234K alone channels were significantly larger at hyperpolarizing potentials (below the reversal potential) but were significantly smaller at depolarizing potentials (over the reversal potential) in comparison to the WT alone channels (Table 2) (Figure 3B), indicating that the deactivation rate of the Q234K alone channels was decelerated at hyperpolarizing potentials but accelerated at depolarizing potentials.

The interaction of KCNE1 with KCNQ1 forms I_Ks and dramatically affects WT alone channels. Therefore, we compared the electrophysiological properties of the WT-KCNQ1+KCNE1 (WT + E1) channels and Q234K-KCNQ1+KCNE1 (Q234K + E1) channels. The current density, measured at the end of the 2-s depolarizing potential of 100 mV, of the Q234K + E1 channels was not different from that of the WT + E1 channels (Table 3) (Figures 4A,B). The VDA of the Q234K + E1 channels was significantly shifted to positive potentials (≈8.5 mV) in comparison to the WT + E1 channels (Table 3) (Figure 4B).

Because activating currents during depolarizing pulses for both channels could not be fitted with a single or double exponential function, we assessed the half activation time (HAT) during the 2-s depolarizing pulses. The HAT of the Q234K + E1 channels was significantly larger than that of the WT + E1 channels (Table 3) (Figures 4A,D), indicating that the activation rate of the Q234K + E1 channels was significantly decelerated in comparison to the WT + E1 channels.

We also assessed the current densities at 2-s, 1-s, and 400-ms depolarizing pulses of various depolarizing potentials (20 mV, 50 mV, and 100 mV). Although the current density of the Q234K + E1 channels, measured at the 2-s depolarizing potential of 100 mV, was comparable to that of the WT + E1 channels as mentioned above, those of the Q234K + E1 channels, measured at the 2-s depolarizing potentials of 20 mV and 50 mV, were smaller than those of the WT + E1 channels, mainly due to the depolarizing shift of the VDA. Notably, the current densities, measured at 400-ms and 1-s depolarizing potentials of 100 mV, of the Q234K + E1 channels were significantly smaller than those of the WT + E1 channels (Table 3) (Figures 4C,E), mainly due to the delayed activation rate in the Q234K + E1 channels. On the other hand, the current densities of the Q234K + E1 channels, measured at 400-ms and 1-s depolarizing potentials of 50 mV and 20 mV, were smaller than those of the WT + E1 channels (Table 3) (Figures 4C,E), due to both the depolarizing shift of the VDA and delayed activation rate in the Q234K + E1 channels.

Because 2-s depolarizing pulses appeared to be insufficient to fully activate both channels, 8-s depolarizing pulses were employed. The current density, measured at the end of the 8-s depolarizing potential of 100 mV, of the Q234K + E1 channels was significantly larger than that of the WT + E1 channels (Table 4) (Figures 5A,B). The VDA of the Q234K + E1 channels was still slightly but significantly shifted to positive potentials (≈6.2 mV) in comparison to the WT + E1 channels (Table 4) (Figure 5C).

In contrast, the current density, measured at a 400-ms depolarizing potential of 100 mV, of the Q234K + E1 channels was significantly smaller than that of the WT + E1 channels (Table 4) (Figures 5A,D), as obtained by the 2-s depolarizing pulse protocol. The HAT of the Q234K + E1 channels was significantly larger than that of the WT + E1 channels (Table 4) (Figure 4E), indicating that the activation rate of the Q234K + E1 channels was significantly slower than that of the WT + E1 channels, as obtained by the 2-s depolarizing pulse protocol. Like these, using the 8-s depolarizing pulse protocol, we obtained almost the same results as the 2-s depolarizing pulse protocol, except for an increased current density in the Q234K + E1 channels.

The deactivation rates were assessed by fitting deactivating currents during repolarizing pulses with a single exponential function (Figure 6A). The time constants of the Q234K + E1 channels were significantly larger than those of the WT + E1 channels (Table 5) (Figure 6B), indicating that the Q234K + E1 channels caused a significant delay of the deactivation rate.

Since the patients in this study harbored the KCNQ1 Q234K variant in a heterozygous manner, we characterized the electrophysiological properties of the WT-KCNQ1+Q234K-KCNQ1+KCNE1 (WT + Q234K + E1) channels.

At the end of the 8-s depolarizing potential of 100 mV, the current density of the WT + Q234K + E1 channels was significantly larger than that of the WT + E1 channels (Table 4) (Figures 5A,B). The VDA of the WT + Q234K + E1 channels was slightly but significantly shifted to positive potentials (≈6.7 mV) in comparison to the WT + E1 channels (Table 4) (Figure 5C). The HATs, obtained by both the 2-s and 8-s depolarizing pulse protocols, of the WT + Q234K + E1 channels were significantly larger than those of the WT + E1 channels (Tables 3 and 4) (Figures 4A,D, 5A,E), indicating that the activation rate of the WT + Q234K + E1 channels was significantly slower than the WT + E1 channels. Using both the 2-s and 8-s depolarizing pulse protocols, the current densities, measured at 400-ms depolarizing potential of 100 mV, of the WT + Q234K + E1 channels were significantly smaller than those of the WT + E1 channels (Tables 3 and 4) (Figures 4A,C, 5A,D).

The time constants of deactivation of the WT + Q234K + E1 channels were significantly larger than those of the WT + E1 channels (Table 5) (Figure 6B), indicating that the deactivation rate of the WT + Q234K + E1 channels was slower than the WT + E1 channels.

The I_Ks (WT + E1 channels) have been reported to display frequency-dependent current accumulation (Romey et al., 1997). Therefore, we examined whether the Q234K + E1 channels and WT + Q234K + E1 channels display frequency-dependent current accumulation during the physiological range of ventricular APD. For this purpose, repeated 400-ms depolarizing pulses at 1 Hz were applied. The WT + E1 channels displayed frequency-dependent current accumulation, while the Q234K + E1 channels and the WT + Q234K + E1 channels displayed weak frequency-dependent current accumulation in comparison to the WT + E1 channels (Table 6) (Figures 7A,B).

In order to elucidate how I_Ks with the Q234K would affect the action potential or QT interval, we focused on three characteristics of the Q234K channel following the experimental data; the shift of the voltage dependence of activation toward the depolarizing side, the slower deactivation time curse, and the larger current density. The modification of the time constant, τ x , S 2 , alone was able to mimic the characteristics of activation and deactivation in the Q234K channel (Figures 8A–C). The V_1/2 for activation was 18.3 mV for WT and 28.7 mV for the Q234K heterozygous model. As shown in Figures 8D,E, the Q234K heterozygous model without the modification of current density had prolonged the APD and QT interval which were intermediate between the WT and the model without I_Ks current like type-1 LQTS. The APD or QT interval in the pseudo-ECG failed to be shortened even though the current density would be increased to 1.2 times of WT corresponding to the experimental data, and the simulated action potential and ECG were almost similar to the Q234K heterozygous model without the modification of current density.

The S4 of α-subunits of voltage-gated potassium channels typically consists of several positively-charged AAs aligned periodically in every three AAs, and is postulated to function as a voltage-sensor. It is widely accepted that the S4 in the VSD senses depolarizing voltage changes across the membrane and moves outwards within the membrane. This outward movement leads to conformational changes in the VSD that trigger structural rearrangements in the PD (VSD-PD coupling or electromechanical coupling), which induces channel opening (Kuenze et al., 2019; Nakajo, 2019; Wu and Larsson, 2020). Unlike Shaker-like voltage-gated potassium channels (Kv1.2), the alignment of positively-charged AAs in the S4 of KCNQ1 is unique in terms of having a neutral-charge AA, glutamine (Q) at the third position (Q3) (Wu et al., 2010; Cui, 2016; Sun and MacKinnon, 2017).

We identified a novel KCNQ1 Q234K (or Q3R) variant in the S4, a substitution for positively-charged lysine, in patients with LQTS and epileptiform activity on EEG. It was expected that an addition of a positive charge on Q3, a critical position of S4, might provide unique gating abnormalities of I_Ks thus resulting in unique phenotypic manifestations.

The interaction of KCNE1 with KCNQ1 slows the channel activation rate, increases voltage-dependent current amplitude, shifts the VDA to depolarizing potentials, and eliminates channel inactivation (Barhanin et al., 1996; Sanguinetti et al., 1996; Wu and Larsson, 2020). However, how KCNE1 associates with and functionally interacts with KCNQ1 remains unknown.

KCNE1 is suggested to directly bind to the pore, particularly S6, of the KCNQ1 channels to control KCNQ1 gating, and to have contact with S4 and the extracellular end of S1 in the KCNQ1 channel (Wu and Larsson, 2020). An interaction between KCNE1 and KCNQ1 may separate the voltage-sensor movements of the I_Ks (WT + E1 channels) into two steps. The first step involves S4 gating charge movement, and the second step involves a slow voltage-sensor movement opening the channel. It is assumed that KCNQ1 alone channels can open after the first step, whereas WT + E1 channels only open after both steps (Barro-Soria et al., 2014; Wu and Larsson, 2020). On the other hand, Nakajo et al. assumed that there might be three states in I_Ks channel gating: closed state (S4 in the down state and PD closed), intermediate stage (S4 in the up state and PD closed), and open state (S4 in the full up state and PD open) (Nakajo, 2019).

When we focused on the activation rate of I_Ks, the KCNQ1 Q234K variant induced a marked delay in the activation rate of I_Ks. Although the mechanism underlying the marked delay in the activation rate is unclear, there may be several possible explanations: the Q234K might affect the first step or the second step of the S4 movements, or the conformational change of the gate via VSD-PD coupling, or a combination thereof. It is also undetermined which state the Q234K affects: transition from a closed state to an intermediate state, from an intermediate state to a closed state, or a combination of the two. It is noteworthy that the electrophysiological properties of the WT + Q234K + E1 channels are generally more similar to those of the Q234K + E1 channels than the WT + E1 channels. However, the deactivation kinetics of the WT + Q234K + E1 channels resemble those of the WT + E1 channels (Figure 6), which may weaken the degree of the gain-of-function of I_Ks. These findings suggest that the Q234K subunits in heterotetrameric channels may not equally affect the transition from the closed state to the intermediate state and the transition from the intermediate state to the open state, and vice versa, in the VSD-PD coupling. To clarify these matters, voltage-clamp fluorometry experiments in combination with ionic current recordings may be required.

Functional changes in the KCNQ1 channels induced by amino acid substitutions in the S4 of KCNQ1 have been reported (Table 7) (Nakajo and Kubo, 2007; Panaghie and Abbott, 2007; Shamgar et al., 2008; Wu et al., 2010). In particular, substitutions of positively-charged arginines for negatively-charged glutamic acids (R1E and R2E) in the N-terminal half (before Q3) of S4 are susceptible to becoming constitutively open in the absence and presence of KCNE1, indicating that basic residues in the N-terminal half of S4 stabilize the resting state (Wu et al., 2010). In contrast, substitutions of basic residues for glutamic acids (R4E and H5E) in the C-terminal half (after Q3) of S4 are susceptible to shifting the VDA to depolarizing potentials in the absence and presence of KCNE1, indicating that basic residues in the C-terminal half of S4 stabilize the activated state (Wu et al., 2010).

Regarding neutral-charge glutamine (Q3), the effects of amino acid substitutions on the channel function appear to differ according to the substituted amino acids and the absence or presence of KCNE1 (Table 7). A substitution for cysteine (Q3C or Q234C) shifted the VDA to hyperpolarizing potentials and increased the current density in the absence and presence of KCNE1 (Nakajo and Kubo, 2007). A substitution for tryptophan (Q3W or Q234W) shifted the VDA to depolarizing potentials and decreased the current density in the absence of KCNE1, but did not affect the VDA and current density in the presence of KCNE1 (Shamgar et al., 2008). A substitution for positively-charged arginine (Q3R or Q234R) shifted the VDA to depolarizing potentials in the absence and presence of KCNE1 (Wu et al., 2010). In contrast, a substitution for negatively-charged glutamic acid (Q3E or Q234E) shifted the VDA to depolarizing potentials in the absence of KCNE1, while it shifted the VDA to hyperpolarizing potentials in the presence of KCNE1, suggesting that a positive charge at Q3 would favor the resting state (Wu et al., 2010).

Several pathogenic variants in the S4 of KCNQ1 have been reported (Table 7). KCNQ1 G229D and R231C variants, located in the N-terminal half of S4, caused constitutively open I_Ks and were associated with atrial fibrillation and/or LQTS (Bartos et al., 2011; Hasegawa et al., 2014; Zhou et al., 2019). In contrast, KCNQ1 D242N, R243C and R243H variants, located in the C-terminal half of S4, strongly shifted the VDA to depolarizing potentials or produced low I_Ks and are thus associated with LQTS (Franqueza et al., 1999; Chouabe et al., 2000; Park et al., 2005; Moreno et al., 2017). In the case of the R243H variant, reduced phosphaditylinositol-4,5-bisphosphate (PIP₂) affinity of the mutant may also be involved in reduced I_Ks (Park et al., 2005). Like these, variants in the N-terminal half (before Q3) of S4 generally cause constitutively open I_Ks and are associated with atrial fibrillation and/or LQTS, whereas variants in the C-terminal half (after Q3) of S4 generally shift the VDA to depolarizing potentials thus producing low I_Ks and are associated with LQTS. Huang et al. reported several other variants, I227L, Q234P, L236P, and L236R in S4 associated with LQTS (Huang et al., 2018). Cell surface expression levels were not reduced, but current densities were remarkably reduced in the I227L and Q234P channels, while both the cell surface expression levels and current densities were reduced in the L236P and L236R channels.

We found that the Q3K (or Q234K) shifted the VDA to depolarizing potentials in the absence and presence of KCNE1, which resembles the effects of Q3R (or Q234R) (Wu et al., 2010), although the effects were weaker than those of the Q234R. Thus, our data support the notion that a positive charge at position 234 in KCNQ1 favors the resting state (Wu et al., 2010). Notably, we found, for the first time, that a single missense variant in the S4, Q3K (or Q234K), markedly delayed both activation and deactivation rates in the presence of KCNE1, although the activation and deactivation rates were rarely assessed in previous experiments of AA substitutions in the S4. Moreover, in the presence of KCNE1, the Q3K (or Q234K) paradoxically induced a gain-of-function during long-depolarization, even though the Q234K + E1 channels favor the resting state, whereas it induced a loss-of-function during short-depolarization (the physiological range of ventricular APD), which is due to the delayed activation rate and depolarizing shift of the VDA. Our data emphasize the importance of assessing not only the VDA but also the activation and deactivation rates of I_Ks to reveal the association between altered channel function and phenotypic manifestation.

In patients with the KCNQ1 Q234K variant in this study, one allele contains KCNQ1 Q234K and the other contains KCNQ1 WT, we therefore assessed the WT + Q234K + E1 channels. The WT + Q234K + E1 channels displayed an increased current density at the end of 8-s depolarizing potentials (gain-of-function of I_Ks), depolarizing shift of the VDA, and decelerated activation and deactivation rates. However, at 400-ms depolarizing potentials, the physiological range of the ventricular APD, the WT + Q234K + E1 channels displayed a decreased current density (loss-of-function of I_Ks) in comparison to the WT + E1 channels. Furthermore, contrary to our expectation, the WT + Q234K + E1 channels displayed a weak frequency-dependent current accumulation, despite their delayed deactivation rates. These findings indicate that the WT + Q234K + E1 channels induce a loss-of-function of I_Ks during the physiological range of ventricular APD, which underlies QT prolongation. The simulation study could prove that electrophysiological properties of the WT + Q234K + E1 obtained by patch-clamp experiments can prolong ventricular APD and QT interval.

Both KCNQ1 and KCNE1 are reported to be expressed in the brain, where KCNQ1 interacts with KCNE1 (Goldman et al., 2009; Rannals et al., 2016), and there have been several reports that KCNQ1 variants are associated with both LQTS and epilepsy (Goldman et al., 2009; Tiron et al., 2015). In addition to the LQTS phenotype, both the proband and her mother showed epileptiform activity on EEG, and her mother had been diagnosed with epilepsy, although the causes of their syncopal events are unknown. In any case, their EEGs showed apparent abnormalities compatible with epilepsy.

In the brain cells, potassium channels function as setting the resting membrane potential, reducing excitability as well as controlling the duration, shape and firing frequency of action potentials (Greene and Hoshi, 2017). Therefore, I_Ks loss-of-function by the KCNQ1 Q234K variant in the brain cells may induce hyperexcitability of the brain cells. However, in case of long-lasting firings (that mimic a long-depolarization), I_Ks gain-of-function may act in the direction of preventing firings. It may be conceivable that the instability of membrane potential of the brain cells may be associated with epileptiform activity (or epilepsy). Moreover, the fact that KCND3 variants which cause unique electrophysiological abnormalities (both gain- and loss-of-function) of I_to are associated with a novel cardiocerebral channelopathy may also suggest the association between unique electrophysiological properties of I_Ks and epileptiform activity (or epilepsy) (Takayama et al., 2019; Nakajima et al., 2020b).