Molecular mechanisms of function deficiencies in KCNQ1 variants associated with Jervell and Lange–Nielsen syndrome

ABSTRACT

Jervell and Lange-Nielsen syndrome (JLNS) is characterized by congenital bilateral sensorineural hearing loss, a prolonged QT interval (QTc) on an electrocardiogram (ECG), and a high incidence of sudden death in childhood. More than 90% of JLNS cases are associated with variants in the potassium voltage-gated channel subfamily Q member 1 gene, KCNQ1 (Kv7.1). Herein, eighteen identified JLNS-related KCNQ1 variants were examined, including I145S, Y148S, G168R, Y171X, S182R, G186D, R190Q, G269D, G272D, A302V, G306V, V307V, S333F, A344A, F351L, K422S, T587M, and R594Q. Using an integrative method, we systematically characterized the biophysical properties, functional, and membrane trafficking of KCNQ1 variants distributed in different structural domains of the channel. The results demonstrated that all the variants resulted in functional deficiencies, with impaired localization in the plasma membrane being the most common cause. Although many variants exhibited normal cell surface expression consistent with protein stability, structural simulation analysis revealed that these KCNQ1 variants disrupt either KCNQ1-KCNE1 or KCNQ1-calmodulin (CaM) interaction, leading to channel dysfunction. These finding provide significant implications for the future treatment and prevention of JLNS.

Introduction

Jervell and Lange-Nielsen syndrome (JLNS, OMIM No.220400) is an autosomal recessively inherited disease clinically characterized by congenital bilateral sensorineural hearing loss, a markedly prolonged correct QT interval (QTc) on a documented ECG, and a high incidence of sudden death secondary to polymorphic ventricular arrhythmias in childhood [1–9]. Approximately 95% of the arrhythmic events are triggered by emotions or exercise [4,10–12]. JLNS occurs worldwide, affecting families of diverse backgrounds ranging from Asian, European, and American origins [11,12].

JLNS is associated with biallelic or compound heterozygous mutations in KCNQ1 or KCNE1, which encode the α subunits and β subunits of the voltage-gated potassium channel, respectively [10,13–24]. Mutations in the KCNQ1 gene account for more than 90% of JLNS cases [4,23,25]. In the heart, KCNQ1 and KCNE1 encode the slowly activated delayed rectifier current (IKs). IKs, together with the rapidly activating delayed rectifier potassium current (IKr), mediates the late phase of repolarization of the cardiac action potential (AP) [26–32]. In the ear, the KCNQ1 channel is involved in potassium-rich endolymph production by inner ear hair cells [33–36]. The KCNQ1 subunit contains six transmembrane segments (S1–S6), a pore loop between S5 and S6, and two intracellular domains (the N-terminus and C-terminus) [36–38]. The C-terminus includes a region (~100 amino acids) called A-domain, which directs KCNQ1 α-subunits to specifically assemble with KCNE β-subunits but not with other KCNQ α-subunits. The A-domain is also involved in proper channel trafficking and normal cell surface expression [39,40].

Several KCNQ1 variants (Table S1) have been reported to induce channel function deficiencies, which underlies the cellular mechanisms of JLNS. While linkage analysis and genotyping have established the role of KCNQ1 variants in deafness and cardiac phenotypes, and although molecular mechanisms have been investigated, functional analysis of genotype-phenotype correlations remain incomplete. Herein, the molecular mechanisms of eighteen identified JLNS-related KCNQ1 variants (I145S, Y148S, G168R, Y171X, S182R, G186D, R190Q, G269D, G272D, A302V, G306V, V307V, S333F, A344A, F351L, K422S, T587M and R594Q) were investigated. We performed comprehensive functional characterization of eighteen KCNQ1 variants, systematically evaluating channel function, membrane trafficking, and biophysical properties. These analyses demonstrated that impaired trafficking to the plasma membrane represents the predominant, though not exclusive, mechanism responsible for KCNQ1 channel dysfunction. In addition, KCNQ1 variants disrupt either KCNQ1-KCNE1 interactions or KCNQ1- CaM binding, leading to channel dysfunction. In summary, the integrative study of JLNS-related KCNQ1 variants revealed the molecular mechanisms of channel function deficiencies, providing significant implications for the future treatment and prevention of JLNS.

Results

To date, fifty-five KCNQ1 variants have been identified in JLNS pedigrees worldwide. These variants, reported in families from America, Europe, and Asia, are distributed across diverse regions of the KCNQ1 protein (Table S1) [8,10,11,13,15,18–21,41–80]. In this study, eighteen KCNQ1 variants from previous studies were functionally summarized. We systematically characterized compound heterozygous KCNQ1 variants, including missense mutations, splice site alterations, nonsense mutations, and frameshift variants. The eighteen KCNQ1 variants investigated reside in distinct domains of the α-subunit (Figure 1(A,B)), including the N-terminal domain, the voltage-sensing domain (VSD), the pore domain (PD), the cytosolic domains, and the unstructured distal C-terminus³⁶. The selected KCNQ1 variants were analyzed based on annotations in ClinVar [81] and/or the Human Genome Mutation Database [82].

Figure 1.

Structural mapping of the eighteen analyzed KCNQ1 variants.(A) Topology of the KCNQ1 channel α-subunit with the localization of the eighteen variants highlighted by blue circles in this study. (B) Left: the sites of the variants mapped onto the structure of KCNQ1. The structure of the KCNQ1 channel was determined by cryo-EM [protein databank (PDB) id: 6uzz]. The tetrameric helical bundle hd domain structure, unstructured N- and C-termini, and loops between the HA and HB helices and between the HC and HD helices are predicted by alphafold3. The Cα atoms of the backbone of the analysed variant sites are shown as green stick models. Right: close-up views of the VSD, pd, and hd domains of KCNQ1 structures.

JLNS-related KCNQ1 variants induced channel function deficiency

To examine the effects of KCNQ1 variants on channel function (including missense and truncating mutations), wild-type (WT) KCNQ1 and eighteen variants were heterologously expressed in CHO-K1 cells through pIRES2-EGFP vector (with nonfusion EGFP as a transfection marker), and the whole-cell currents was recorded. The functional consequences of the variants were assessed by measuring the peak current density, the V_1/2 of the activation voltage dependence, and the time constants for activation and deactivation. Cells transfected with WT KCNQ1 presented robust outwards currents under voltage steps ranging from −100 mV to +60 mV, with a holding potential of −80 mV (Figure 2(A)). The current density of WT KCNQ1 channels at +40 mV was 41.90 ± 5.99 pA/pF (n = 8) (Figure 2(B)). In addition, no measurable current was detected in untransfected CHO-K1 cells under the same stimulation conditions (Figure 2(A)). Compared with the WT KCNQ1 channel, sixteen pathogenic variants (I145S, Y148S, G168R, Y171X, G186D, R190Q, G269D, G272D, A302V, G306V, V307V, A344A, S333F, F351L, K422S, T587M, and R594Q) significantly reduced the current density (Figure 2(B), Figure S1A). Additionally, the S182R variant exhibited a statistically rightward shift in V_1/2 activation without differences in current density (Figures 2(B,C) and Figure S1B). The shifts in activation V_1/2 are associated with channel dysfunction [83,84].

Figure 2.

Electrophysiological characterization of KCNQ1 variants or KCNQ1 variants co-expressed with KCNE1 in this study. (a) Representative current traces of wild-type KCNQ1 and the S182R, V307V, and A344A variants expressed in cho cells. The cells were held at −80 mV. The step voltages ranged from −100 to +60 mV in 20-mV increments, with steps of 3 s, followed by a pulse of 50 mV. (b) The mean current densities of wt KCNQ1 and KCNQ1 variant channels at +60 mV. N ≥ 4, all error bars mean ± sem. **, p < 0.01; ***, p < 0.001; ns, p ≥ 0.05. (c) Voltage of half-maximal activation (V_1/2) of wt KCNQ1, S182R, V307V and A344A variants. N ≥ 4, all error bars mean ± sem. *, p < 0.05; ns, p ≥ 0.05. (d) Representative current traces of wt KCNQ1 and the S182R, V307V, and A344A variants co-expressed with KCNE1 in cho cells. The cells were held at −80 mV. The step voltages ranged from −100 to +60 mV in 20-mV increments, with steps of 3 s, followed by a pulse of 50 mV. (e) The mean current densities of wt KCNQ1 and KCNQ1 variant channels co-expressed with KCNE1 at +60 mV. N ≥ 4, all error bars mean ± sem. ***, p < 0.001; ns, p ≥ 0.05. (f) The voltage of half-maximal activation (V_1/2) of wt KCNQ1, the S182R, V307V and A344A variants co-expressed with KCNE1. N ≥ 4, all error bars mean ± sem. **, p < 0.01; ns, p ≥ 0.05.

Since KCNQ1 forms the cardiac IKs channel by co-assembling with its β-subunit KCNE1, the WT KCNQ1 or KCNQ1 variants with KCNE1 in CHO-K1 cells were co-expressed for whole-cell recording (Figure 2(D)). The current density of WT KCNQ1 with KCNE1 channels at +40 mV was 308.91 ± 23.10 pA/pF (n = 17) (Figure 2(E)). Compared with WT KCNQ1 co-expressed with the KCNE1 channel, sixteen pathogenic variants (I145S, Y148S, G168R, Y171X, G186D, R190Q, G269D, G272D, A302V, G306V, V307V, A344A, S333F, F351L, K422S, T587M, and R594Q) significantly reduced the IKs current (Figure 2(E), Figure S2A). Strikingly, only the S182R variant co-expressed with KCNE1 exhibited a significant rightward shift in V_1/2 compared to WT KCNQ1 without affecting current density (Figures 2(E,F) and Figure S2B), suggesting the voltage sensitivity in this variant was altered.

Electrophysiological characterization of KCNQ1 variants co-expressed with wt KCNQ1 and KCNE1

Furthermore, to mimic the heterozygous state of carriers in the JLNS pedigrees of the cardiac phenotype, the electrophysiological characteristics of KCNQ1 variants co-expressed with WT KCNQ1 (1:1 ratio) in the presence of KCNE1 were examined. The results revealed that I145S, G186D, R190Q, A344A, and T587M variants co-expressed with equal amounts of WT KCNQ1 significantly reduced current density compared with those of WT KCNQ1 (Figure 3(A), Figure S3A and Table S2). Y148L, G168R, Y171X, S182R, G269D, A302V, G306V, S333F, F351L, K422S, T587M, and R594Q presented the greatest IKs current reduction (Figure 3(A), Figure S3 A and Table S2). In contrast, the V307V variant showed no significant change compared to WT KCNQ1 (Figure 3(A), Figure S3A and Table S2). Next, the effect of the heterozygous KCNQ1 variants when co-expressed with KCNE1 on the voltage dependence of IKs channel activation were analyzed. Only the S182R variant co-expressed with KCNE1 demonstrated a significant rightward shift in the V_1/2 of activation (Figure 3(B), Figure S3B and Table S2). Our results demonstrate that heterozygous KCNQ1 variants act a dominant-negative effect on channel function.

Figure 3.

Electrophysiological characterization of KCNQ1 variants co-expressed with KCNQ1 and KCNE1 in this study.(A) The mean current densities of wt KCNQ1 and KCNQ1 variant channels co-expressed with KCNQ1 and KCNE1 at +60 mV. The cells were held at −80 mV. The step voltages ranged from −100 to +60 mV in 20-mV increments, with steps of 3 s, followed by a pulse of 50 mV. N ≥ 4, all error bars mean ± sem. ***, p < 0.001; **, p < 0.01; ns, p ≥ 0.05. (b) The voltage of half-maximal activation (V_1/2) of WT and KCNQ1 variants co-expressed with KCNQ1 and KCNE1. N ≥ 4, all error bars mean ± sem. *, p < 0.05; ns, p ≥ 0.05.

Mistrafficking is a common cause of mutation-induced channel lof

Previous studies have shown that several KCNQ1 variants generate nonfunctional channels due to defective membrane trafficking [85]. To further investigate this, non-permeabilized immunostaining was performed to assess the cell surface expression of WT KCNQ1 and its variants when co-expressed with KCNE1. A Myc epitope tag inserted into the extracellular loop between TM1 and TM2 of KCNQ1, enabled specific labeling of plasma membrane under non-permeabilized conditions [85–88] (Figure 4(A)). Twelve KCNQ1 variant channels (G168R, G186D, R190Q, G269D, G272D, G306V, S333F, F351L, K422S, and R594Q) did not expressed on the cell surface (Figure 4(B,C)). In contrast, the remaining KCNQ1 variants (S182R, A302V, V307V, A344A, and T587M) showed plasma membrane localization similar to WT KCNQ1 (Figure S4). These findings demonstrate that impaired membrane trafficking represents a major mechanism of LOF in JLNS-related KCNQ1 variants.

Figure 4.

The thermal destabilization and cell surface expression of KCNQ1 variants. topology schematic of KCNQ1. Purple squares represent the six transmembrane domains of KCNQ1. The pink dot represents myc tag. A Myc tag was inserted into the extracellular loop between the TM1 and TM2 transmembrane domains of KCNQ1. (b) Non-permeabilized staining showed that the wt KCNQ1 protein expressed on the cell surface co-expressed with KCNE1. Green (KCNQ1), red (KCNE1), pink (Myc-tag). (c) Non-permeabilized staining showed JLNS-related KCNQ1 variants were not expressed on the cell surface. Green (KCNQ1), red (KCNE1), pink (Myc-tag). (d) Representative thermal unfolding curves of wt KCNQ1 and KCNQ1 missense mutants. The band signal intensity was normalized to the signal intensity of the respective sample at 37°C. Curves were fitted to a four-parameter symmetric sigmoidal curve to determine T_agg. (e) T_agg values for the missense mutations studied in this study. All error bars represent the mean ± SEMs **, p < 0.01; *, p < 0.05; ns, p ≥ 0.05. N = 6 for wt KCNQ1 WT and N = 3 for each KCNQ1 variant. (f) Comparison of Tagg and surface expression values. For surface expression (up), horizontal dashed lines indicate the cell surface expression efficiency of wt KCNQ1 and the level of reduced cell surface expression relative to wt, respectively. All error bars represent the mean ± sem, ***, p < 0.001; N = 3 for each wt KCNQ1 and KCNQ1 variant. For the Tagg (down), horizontal dashed lines indicate the mean cell surface expression efficiency for wt KCNQ1 and the variants threshold for thermal destabilization, respectively. All error bars represent the mean ± SEMs **, p < 0.01; *, p < 0.05; ns, p ≥ 0.05. N = 6 for wt KCNQ1 WT and N = 3 for each KCNQ1 variant. (data from Figure 4(E)).

Destabilizing KCNQ1 variants represents a major cause of JLNS

Previous studies have demonstrated that LQT1-associated KCNQ1 variants are destabilizing, frequently resulting in impaired channel trafficking and LOF [84,89]. To investigate the mutations associated with JLNS in our study, we first predicted the stability of KCNQ1 variants using MuPro, a web server predicting protein stability changes from single amino acid mutations [90,91]. The results revealed that all missense mutations led to protein destabilization (Table S3). Next, the cellular thermal shift assay (CETSA) was adapted to determine the thermal stability of full-length KCNQ1 expressed in cultured HEK293T cells, as previously described [89,92–94].

For CETSA analysis, HEK293 cells transiently expressing KCNQ1 variants were cultured for 48 hours and then harvested and heated to incremental temperature gradients to induce protein thermal unfolding and irreversible aggregation. The thermal aggregation temperature (T_agg) was determined by quantifying monomeric KCNQ1 signal intensities as a function of incubation temperature. The CETSA analysis was performed on eleven KCNQ1 missense mutations distributed across all functional domains (Figure S5). Among these variants, eight predominantly pathogenic variants exhibited thermal destabilization compared with the WT KCNQ1 protein (Figure 4(D,E)).

Furthermore, destabilizing KCNQ1 variants were also exhibited impaired cell surface trafficking, likely due to intracellular retention or targeted degradation. To confirm the correlation between the stability of the KCNQ1 variants and their surface expression, we first counted quantitative cell surface expression using immunofluorescence staining. Myc-tagged KCNQ1 was transiently expressed using a pEGFP-C5 vector. The Myc-tagged KCNQ1 and eGFP represent the surface and total expression levels, respectively. The cell surface expression efficiency was defined as the ratio of surface to total expression (surface/total × 100). The results showed that the G168R, G186D, R190Q, G269D, G272D, G306V, F351L, and R594Q variants failed to express on the cell surface, while the S182R, A302V, and T587M variants showed reduced cell surface expression compared with WT KCNQ1 (S182R 87.67 ± 5.61, A302V 84.67 ± 4.47, T587M 84.33 ± 4.91 WT KCNQ1 97.00 ± 1.53) (Figure 4(F)). In addition, we observed a correlation between the thermal stability and the surface expression of KCNQ1 variant proteins, except for A302V, nearly all thermally destabilized variants showing impaired membrane trafficking (Figure 4(F)). These findings establish protein misfolding-induced trafficking defects as a common LOF mechanism for pathogenic KCNQ1 variants.

Structural basis of KCNQ1 variants causing channel dysfunction

Collectively, our data indicated that many pathogenic KCNQ1 variants frequently display thermodynamic instability, which coincides with impaired membrane trafficking. While defective membrane localization represents the most common pathogenic mechanism for KCNQ1-releted JLNS, some variants exhibited normal membrane localization but retained functional deficiency, suggesting mechanisms beyond trafficking deficiency. The pore domain variant A302V correctly localized in the plasma membrane but lacked voltage-dependent currents. To investigate the structural basis of this functional impairment, we performed molecular modeling studies using AlphaFold3 based WT KCNQ1. Based on AlphaFold3 simulations, the three-dimensional structure revealed that A302 directly interacted with residues S277, S298, and G306 in KCNQ1. However, V302 cannot directly interact with residue S277, which may lead to an unstable channel structure (Figure 5(A)). Simulation results revealed alterations in the hydrogen bonds due to the substitution of the residues, which possibly perturbed the interactions among the amino acid residues and led to the instability of the channel.

Figure 5.

Structural analysis and key interactions involving mutation sites in KCNQ1.(A) Comparison of residue interactions before and after mutation. Structural model of KCNQ1-A302V based on homology modelling by Alphafold3 (PDB accession code 6uzz). Amino acid side chains are drawn as sticks. The A302 residues and V302 residues are colored purple and green, respectively. Amino acid residues with interactions are shown in orange, whereas those with altered interactions are highlighted in blue. The predicted hydrogen bond interactions are indicated by yellow dotted lines. (b) Comparison of residue interactions before and after mutation. Structural model of KCNQ1-S182R-CaM based on homology modelling by Alphafold3 (PDB accession code 6uzz). The S182 and R182 residues are colored purple and green, respectively. Amino acid residues with interactions are shown in orange, whereas those with altered interactions are highlighted in blue. The predicted hydrogen bond interactions are indicated by yellow dotted lines. (c) Comparison of residue interactions before and after mutation. Structural model of KCNQ1-G306V-KCNE1 based on homology modeling by Alphafold3 (PDB accession code 6uzz). The G306 and V306 residues are colored purple and green, respectively.

In addition, the S2-S3 linker of the KCNQ1 protein interacts with CaM to regulate channel function. The S182, G186, and R190 residues reside within this linker. Based on AlphaFold3 simulations, the three-dimensional structure demonstrated that, unlike the WT KCNQ1-S182-CaM complex, the R182 mutant residue fails to maintain the interaction with K183 in KCNQ1 and increases its interaction with E141 in CaM, affecting the CaM-mediated channel regulation (Figure 5(B)). KCNQ1 co-assembled with KCNE1 to form the slowly activated delayed rectifier potassium current (IKs). mutations in KCNQ1 can disrupt this assembly. In addition, for the variant G306V, structural analysis revealed that the KCNQ1-V306 variant exhibited altered position of its S1 domain relative to KCNE1 compared to the wild-type complex, suggesting impaired KCNE1-mediated channel regulation and membrane localization (Figure 5(C)). Taken together, altered interactions between amino acid residues may resulted in channel dysfunction.

Discussion

Our study provides a comprehensive analysis of the molecular mechanisms for the functional deficiencies of KCNQ1 variants in JLNS. These findings have significant implications for the future treatment and prevention of JLNS, as well as the development of more precise therapeutic approaches for this life-threatening disorder.

The clinical characteristics of patients with KCNQ1 variants causing JLNS

Among all pathogenic KCNQ1 variants associated with JLNS, missense mutations serve as the most prevalent variant class [7,95]. Notably, many of these mutations were initially identified as pathogenic variants for long QT syndrome (LQTS). However, when these variants are combined with other KCNQ1 mutations – typically frameshift or protein-truncating mutations – they manifest as JLNS. For example, the c.605-2A > G variant and c.1032 G > A (p.A344A) variant were initially characterized as LQTS-causing mutations [96]. However, the compound heterozygosity of these two variants (c.605-2A > G +c.1032 G > A) results in a more severe JLNS phenotype [64]. Interestingly, some homozygous KCNQ1 variants do not result in deafness, demonstrating phenotypic heterogeneity in JLNS. The homozygous c.3875T > A variant [97] causes severe cardiac manifestations due to reduced IKs currents, yet homozygous mutation shows no deafness phenotype. These findings suggest that inner ear function can be maintained by the residual current induced by KCNQ1. Kanovsky et al [61] reported a family in which two affected individuals carried the same homozygous KCNQ1 mutation (p.R190L). One patient exhibited incomplete JLNS with subclinical hearing impairment, while the other presented normal auditory function. This observation highlights the clinical heterogeneity in cardiac and auditory phenotypes associated with KCNQ1 mutations.

The functional deficiency of KCNQ1 variants causing JLNS

Our study revealed that most KCNQ1 mutations associated with JLNS result in complete LOF. However, three variants-S182R, A344A, and V307V- exhibit partial LOF. The c.546C > A (p. S182R) variant showed normal whole-cell current density but induced a significant rightward shift in the voltage dependence of activation. c.546C > A (p. S182R) combined with c.1831 G > A (p.D611N) caused JLNS in a previous study [71]. The S182R mutation, located in the S2-S3 linker of KCNQ1, computational predictions supported its pathogenic potential. Compared with WT KCNQ1, the S182R variant mainly slows channel activation kinetics while maintaining normal current density, indicating its pathogenic potential. In addition, although A344A (c.1032 G > A) and V307V (c.921A > G) are synonymous mutations, functional studies confirmed they induce LOF when paired with other pathogenic alleles.

In families with JLNS, carriers typically exhibit prolonged QT intervals on electrocardiograms without severe hearing impairment. To maintain normal cardiac function, the precisely coordinated rhythmic activity of specialized cardiomyocytes must be regulated by regulated by voltage- and time-dependent currents [98]. In the inner ear, a unidirectional K⁺ flux from the spiral ligament to marginal cells (MCs) is maintained by KCNQ1 channels. K⁺ cycling through the cochlear duct maintains a high-capacity K⁺ flux across the lateral wall, generating the endocochlear potential (EP) [33–36]. Thus, the essential requirement for the generation and maintenance of the EP is the continuous flow of K⁺. Moreover, K⁺ channel tetramerization occurs stochastically, resulting in all possible combinations of WT and mutant subunits in heterozygous carriers. This random assembly preserves sufficient channel function for EP maintenance in the inner ear.

Mechanistic landscape of KCNQ1 variants revealed by integrative analysis

Previous studies established that JLNS pathogenesis involves defective plasma membrane localization of KCNQ1 variants [85]. Consistent with these findings, our study demonstrated that trafficking deficiencies – caused by mutations distributed across folded domains (VSD, PD, and cytosolic HA/HB helices) – represent the primary pathogenic basis of JLNS. Moreover, low cell surface expression always coincides with protein instability. Interestingly, the G186D, located in the S2-S3 linker exhibited impaired cell surface expression without significant destabilization, consistent with the class IV classification in Kathryn R. Brewer’s study [89]. This suggests G186D avoids thermally induced irreversible aggregation of the full-length protein but causes localized domain destabilization. Such structural defects trigger membrane protein quality control mechanisms, leading to intracellular retention and subsequent degradation.

Several LOF variants exhibit intact cell surface expression but functional deficiency through multiple mechanisms, including deleterious effects on ion conduction; altered voltage sensitivity; disrupted KCNQ1-KCNE1, KCNQ1- CaM, or KCNQ1-PIP2 interactions; or impaired kinetics of the conformational changes required for channel opening [99]. Structural studies revealed that the S2-S3 loop region interacts with CaM to regulate the gating function of the KCNQ1 channel [100]. In our study, S182, G186, and R190 reside located in S2-S3 extracellular loop. The G186D and R190Q mutations exhibited defective membrane trafficking as their primary molecular mechanism. However, the S182R showed normal cell surface expression but exhibited altered voltage dependence, likely resulting from impaired CaM-mediated regulation due to altered KCNQ1-CaM interaction, leading to channel dysfunction. Additionally, as KCNE1 interacts with the S1-S2 linker of KCNQ1 to modulate channel function [101], the V306A variant may impair this KCNQ1-KCNE1 interaction, leading to channel dysfunction.

In summary, we systematically evaluated eighteen JLNS-related KCNQ1 variants for channel function, membrane trafficking, and biophysical properties were evaluating. While impaired plasma membrane trafficking represents the main molecular mechanism for KCNQ1 channel dysfunction, we found that the mutations-induced conformational changes significantly disrupt the interactions either in KCNQ1-KCNE1 or KCNQ1-CaM, representing another important molecular mechanism.

Method

Plasmid construction

Full-length human KCNQ1 (NP_004691.2, isoform a) was purchased from GeneCopoeia (China). The WT KCNQ1 and KCNQ1 variants used in the patch-clamp recordings were subcloned and inserted into the mammalian expression vector pIRES2-EGFP. For the subcellular localization of WT KCNQ1 and its variants, a Myc tag was inserted in the extracellular loop between TM1 and TM2 of the KCNQ1 protein as described in the previous study [85–88]. For the KCNE1 co-transfection experiment, EGFP was substituted with mCherry by a homologous recombination method. For the CETSA experiments, N flag-tagged human WT KCNQ1 and its variants were subcloned and inserted into a pN3 vector. All the plasmids were verified by Sanger sequencing.

Cell culture and transfection

The human embryonic kidney (HEK) 293T and Chinese hamster ovary (CHO-K1) cell lines were obtained from the American Type Culture Collection (ATCC). The CHO-K1 cell line and HEK293T cell line was cultured in F-12/DMEM and DMEM, respectively, supplemented with 10% Gibico, cat#A5256701 and a 1% antibiotic‒antimycotic mixture (Invitrogen, cat#15640055) at 37°C in 5% CO₂. For patch-clamp recording, CHO-K1 cells plated in 12-well plates were transiently co-transfected with plasmids encoding either WT or variant KCNQ1 channels (0.5 μg) plus KCNE1 (1.0 μg) via Lipofectamine 3000 reagent (Thermon, cat# L3000015) according to the manufacturer’s protocol. For CETSA experiments, HEK293T cells were plated in 60 cm cell culture dishes and transfected with 6 μg of either WT KCNQ1 or KCNQ1 variants using jetPRIME-DNA reagent (Polyplus, cat# 101,000,001) at 80 to 90% confluent. For non-permeabilized immunostaining, glass coverslips were coated with poly-L-lysine prior to cell seeding. HEK293T cells were seeded into 12-well plates and co-transfected with either WT KCNQ1 or variant KCNQ1 variants (0.5 μg) plus KCNE1 (1.0 μg) via the Lipofectamine 3000 reagent (Thermon, cat# L3000015) at 50–60% confluence. Electrophysiological and immunostaining experiments were conducted 24 h after transfecion, whereas CETSA experiments were performed at 48 h.

Patch-clamp recordings

The WT KCNQ1 and mutated KCNQ1 plasmids were heterologously expressed in CHO-K1 cells. Whole-cell patch-clamp recordings were performed using an Axopatch 700B amplifier (Axon Instruments, USA) and Digidata 1550B (Axon Instruments) at room temperature (20–24 °C). Borosilicate glass pipettes (BF150-86–10, Sutter Instrument) were pulled using a horizontal p-97 Flaming – Brown micropipette puller (Sutter Instruments, Novato, CA, USA) with a tip resistance of 3–6 MΩ. The bath solution for the whole-cell patch-clamp contained (in mM) 145 NaCl, 4 KCl, 1.8 CaCl₂, 0.5 MgCl₂, 10 HEPES, and 5 D-glucose (pH 7.4) with NaOH. The pipette mixture contained (in mM) 140 KCl, 1 MgCl₂, 10 HEPES, 10 EGTA, 1 CaCl₂, and 4 K₂ATP (pH 7.2) with KOH. The signals were filtered at 1 kHz and sampled at 10 kHz. Before current acquisition, the cell membrane capacitance (Cm) and series resistance (Rs) were compensated using a circuit of the patch-clamp amplifier. The series resistances were compensated70–90% and no less than 15 MΩ. The KCNQ1 current traces were generated by voltage steps ranging from −100 to +60 mV with 20-mV increments at a holding potential of −80 mV. The currents at different voltage potentials were measured at peak levels. The currents were then divided by the cell capacitance (pF) to generate the current density‒voltage relationship. All current trace data were analyzed using Clampfit 11.1.

Non-permeabilized immunostaining and confocal imaging

HEK293T cells were transfected approximately 24 h before non-permeabilized staining. Briefly, the cells were fixed with 4% paraformaldehyde (PFA) solution for 10 min at room temperature, followed by blocking in 3% BSA in PBS at room temperature for 1 h. After blocking, the sections were incubated with a primary antibody (anti-MYC rabbit monoclonal antibody, 1:250, ABclonal, cat# AP0082) overnight at 4°C and with a secondary antibody (goat anti-rabbit IgG, 1:1000, Sigma, cat# A32733) for 1 h at room temperature. Images were acquired with a Zeiss LSM980 confocal microscope.

Cellular thermal shift assay (CETSA)

HEK293T cells transfected with either WT or variant N-Flag-KCNQ1 constructs were harvested 48 hours after transfection. The cells were washed with calcium and magnesium-free phosphate-buffered saline (PBS, Vazyme), detached with 0.25% trypsin-EDTA, and quenched with complete DMEM. After centrifugation at 1000 rpm for 3 minutes, the cells were washed with PBS and resuspended in PBS to a final concentration of 1 × 10 [7] cells/mL. The cell suspensions were supplemented with a protease inhibitor cocktail at a final volume of 1:100 (P8340, Sigma). 50 μL samples were placed in PCR tubes and heat shocked at their respective temperatures for 3 minutes via a PCR system (Bio-Rad, T100 Thermal Cycler). The temperature of heat shock ranged between 40°C and 80°C at 4°C increments. Untransfected and KCNQ1-transfected control samples maintained at 37°C were excluded from heat treatment, as these samples represented the cell culture temperature. Following heat shock, the cells were lysed with 1% digitonin, subjected to freezing in liquid nitrogen for approximately 15 s and allowed to thaw at room temperature for a total of three freeze‒thaw cycles. The lysates were then centrifuged at 15,000 rpm for 15 minutes. The supernatant was collected and supplemented with 10 mM TCEP prior to SDS‒PAGE and Western blot analysis to separate KCNQ1 tetramers into a monomeric band on the gel. CETSA experiments were performed 3 times for each variant, for a total of 6 times for WT KCNQ1.

Western blot analysis

For the CETSA samples, the supernatants were prepared for gel electrophoresis addition of protein loading buffer (Yamei). SDS‒PAGE was conducted via the addition of SDS running buffer to 12.5% Ga‒Tris gels at 110 V for 90 minutes. Proteins were transferred from SDS‒PAGE gels to nitrocellulose membranes using the Trans-blot Turbo Transfer System (Bio-Rad). Following transfer, the membranes were blocked in TBST buffer (TBS and 0.1% Tween 20) supplemented with 5% nonfat powdered milk for 60–90 minutes at room temperature. The membranes were then incubated at 4°C in blocking buffer with a 1:2500 dilution of anti-flag antibody (ABclonal, cat# AE-061) overnight. The next day, after three TBST washes, the membranes were incubated with 1:10,000 HRP-conjugated anti-mouse IgG secondary antibody (ABclonal, cat# AS080) in TBST for 1 h at room temperature. Following additional TBST washes, protein signals were detected using a Bio-Rad ChemiDoc MP imaging system (Bio-Rad). The band intensity of the monomeric flag-tagged KCNQ1 band (78 kDa) was calculated using the band peak quantification plug-in in FIJI [102]. The signal intensity was normalized to that of the KCNQ1-transfected sample at 37°C for each respective CETSA experiment. CETSA curves were fit to a four-parameter symmetric sigmoidal curve in GraphPad Prism 10.4.0 (“Sigmoidal, 4PL, X is concentration” in the software). The IC₅₀ calculated for each curve was used as T_agg.

Molecular modeling

The structure of the KCNQ1 [UniProtKB: P51787, Protein Databank (PDB) ID: 6uzz] tetramer from the Protein Data Bank (https://www.rcsb.org/) was used as a template for molecular modeling analysis. Alterations in the KCNQ1 variants structure were predicted in silico using Alphafold3 (https://alphafoldserver.com/) and visualized through PyMOL. The effect of variation on protein stability was determined through Mupro (https://mupro.proteomics.ics.uci.edu/) by calculating the change in thermodynamic free energy (ΔΔG) and the direction of change after variation, where ΔΔG > 0 indicated stabilization, whereas ΔΔG < 0 indicated destabilization.

Statistical analysis

Statistical analysis and graph generation were performed using GraphPad Prism 10.4.0 (GraphPad Software). The Shapiro-Wilk and Kolmogorov-Smirnov tests were applied to check for normal distribution in all samples prior to running statistical analysis. The F test in GraphPad Prism 10.4.0 to was used to assess variance homogeneity. When assumptions of normality and equal variances were met, unpaired two-tailed Student’s t tests were used. Conversely, an unpaired two-tailed Student’s t test with Welch’s correction was applied. For data non-normally distributed, Mann-Whitney tests were conducted. A p value less than 0.05 was considered significant. The data are presented as mean ± SEM. The statistical details of the experiments are shown in the figure legends.

Funding Statement

Hebei Provincial Graduate Student Innovation Program (CXZZBS2021086), Postdoctoral Fellowship Program of CPSF (GZB20250190), Postdoctoral Foundation of Hebei Medical University (cp00218) and Hebei Provincial Graduate Student Innovation Program (XCXZZB202308). Postdocoral Foundation of Hebei Medical university

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available from the corresponding author, Mingshun Lu, upon reasonable request.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/19336950.2025.2580177