Introduction

The voltage-gated K⁺ channel KCNQ1, also referred to as Kv7.1, is expressed in the heart. Here the channel associates with the accessory KCNE1 subunit, generating the so-called slow delayed-rectifier (IKs) current, an important contributor to the repolarizing phase of the ventricular action potential (AP) ^1,2,3. Loss-of-function mutations of the KCNQ1/KCNE1 complex are associated with Long QT Syndrome (LQTS) ^{4, 5}. These LQTS mutations cause a reduction of the channel current, leading to a significant prolongation of the ventricular AP waveform that can be seen on the electrocardiogram as a prolonged QT interval ³ (Fig. 1A). These channel mutations, or dysfunction, increase the risk of developing cardiac arrythmias, which can lead to sudden cardiac death ⁶. At present, treatment of LQTS is mainly based on the usage of β-blockers (most used are long-lasting preparations such as nadolol and atenolol) and implantable cardioverter-defibrillators, especially for patients with high risk of sudden death and frequent syncope ⁷. However, both approaches do not shorten the QT interval duration. Restoring the physiological duration of the QT interval could be achieved by increasing the activity of the KCNQ1/KCNE1 channel complex ⁸ (Fig. 1A).

Figure 1.

Like other Kv channels, KCNQ1 has a typical tetrameric structure of four α-subunits. Each α-subunit is composed of six transmembrane segments, with S1-S4 forming the voltage sensor domain (VSD) and S5-S6 forming the pore domain (PD) (Figure 1D). However, KCNQ1/KCNE1 channel complex has a very small single channel conductance and a very low open probability compared to other Kv channels ⁹. The mechanisms behind the small conductance and low open probability are not understood.

Polyunsaturated fatty acids (PUFA), in particular omega 3, are known to exert a protective effect on sudden cardiac death and are recommended in the diet at least twice a week ¹⁰. PUFAs have been shown to increase IKs currents by a dual mechanism of action, characteristically described as a “Lipoelectric mechanism” ¹¹. According to this mechanism, PUFAs can increase IKs currents by shifting the voltage-dependence of activation (ΔV0.5) toward negative voltages and increase the maximum conductance of the channel (Gmax) ^{12, 13, 14, 15, 16} (Figure 1B-C).

The two PUFA effects (ΔV0.5 and Gmax) on KCNQ1 are independent of each other and originate from the binding of PUFA to two different sites on KCNQ1, conventionally indicated as Site I and Site II ^{14, 18}. Site I is found in the VSD, where the PUFA head group interacts with the positive charges in the S4 segment and causes the channel to open at more negative potentials ¹⁹. Site II is found in the pore domain, where electrostatic interactions between the PUFA head group and the positively charged residue, K326, facilitate an increase in the maximal conductance of the channel¹⁸ (Figure 1E-F).

Although previous studies have shown the importance of K326 for PUFA in increasing Gmax through binding at site II ¹⁸, the mechanism by which PUFAs increase Gmax is still not understood. In this study we use a combination of two-electrode voltage clamp, single channel recordings, site-directed mutagenesis, and recent cryoEM structures to gain insight into the molecular mechanism underlying the Gmax effect following PUFA binding at site II. We propose a novel mechanism where the selectivity filter of KCNQ1 is normally unstable and the binding of PUFA to site II stabilizes a network of interactions at the selectivity filter, which in turn, leads to a more open and stable conformation of the KCNQ1 pore and thus an increase in channel open probability.

Material and methods

Molecular biology

KCNQ1 (UniProt: P51787), KCNE1 (Uniprot: P15382) cRNA were transcribed using the mMessage mMachine T7 kit (Ambion). KCNQ1 was co-expressed with KCNE1 subunit, following a 3:1, weight: weight (Q1:E1) cRNA ratio to make up the KCNQ1/KCNE1 currents. Site-directed mutagenesis was performed using the Quickchange II XL Mutagenesis Kit (QIAGEN Sciences) for mutations in KCNQ1. 50 ng of complementary RNA was injected into defolliculated Xenopus laevis oocytes (Ecocyte Bioscience, Austin; Xenopus 1 Corp, Dexter) for channel expression. After RNA injection, cells were incubated for 24-96 h, in standard ND96 solution (96 mM NaCl, 2 mM KCl, 1 mM MgCl₂, 1.8 mM CaCl₂, 5 mM HEPES, 1M sodium pyruvate and penicillin (10,000 units)-streptomycin (10 mg/ml in 0.9% NaCl; pH 7.5) at 16°C before experiments.

Two-electrode voltage clamp

KCNQ1/KCNE1 currents were recorded using the two-electrode voltage clamp (TEVC) technique. The recording chamber was filled with ND96 solution (96 mM NaCl, 2 mM KCl, 1 mM MgCl₂, 1.8 mM CaCl₂, and 5 mM HEPES; pH 7.5). Pipettes were filled with 3M KCl. The polyunsaturated fatty acid (PUFA) Lin-Glycine was purchased from Cayman Chemicals (Ann Arbor, MI), kept in stock of 100 mM with ethanol at - 20°C and diluted in ND96 solution the day of the experiments. Electrophysiological recordings were acquired using Clampex 10.7 software (Axon, pClamp, Molecular Devices). Lin-Glycine was perfused into the recording chamber in a pre-application step until reaching steady state, followed by an (I-V) protocol to measure the current-voltage relationship before and after perfusion. During PUFA application, cells were held at −80 mV and depolarization from −80 mV to 0 mV (5-s) was applied every 30 s, before stepping to −40 mV and then back to – 80 mV. In the I-V protocol, a hyperpolarized step from −80 mV to - 140 mV is applied before 20-mV voltage steps from −100 mV to +60 mV. Tail currents are recorded at −20 mV before returning to holding potential.

Single channel recordings

Single channel currents were recorded from transiently transfected mouse ltk- fibroblast cells (LM cells) using Lipofectamine 2000 (Thermo Fisher Scientific) as per the manufacturers protocol. Cells were transfected with pcDNA3 containing a linked KCNE1-KCNQ1 construct ²⁰, to ensure fully KCNE1-saturated complexes, in addition to a plasmid containing green fluorescent protein (GFP) to identify transfected cells. Cells were recorded from 24-48 hours after transfection using an Axopatch 200B amplifier, a Digidata 1440A and pClamp10 software (Molecular Devices, San Jose CA, USA). The bath solution contained (in mM): 135 KCl, 1 MgCl₂, 1 CaCl₂, 10 HEPES, 10 Dextrose (pH 7.4 with KOH). The pipette solution contained (in mM): 6 NaCl, 129 MES, 1 MgCl₂, 5 KCl, 1 CaCl₂, 10 HEPES (pH 7.4 with NaOH). Pipettes were pulled from thick-walled borosilicate glass (Sutter Instruments, Novato, CA, USA), using a linear multistage electrode puller (Sutter Instruments), fire polished and coated with Sylgard (Dow Corning, Midland, MI, USA). Electrode resistance was between 40 and 80 MΩ after polishing. Currents were sampled at 10 kHz, low-pass filtered at 2 kHz at acquisition and subsequently digitally filtered at 200 Hz for presentation and analysis.

Data analysis

To measure the conductance vs voltage (G-V) curve, KCNQ1/KCNE1 tail currents were measured at −20 mV and obtained values were plotted against the activation voltages and fitted to a Boltzmann function:

where G_min is the minimal conductance, G_max is the maximal conductance, V₅₀ the midpoint (which describes the voltage at which the conductance is half the maximal conductance established from the fit), and s is the slope of the curve in mV. The difference in Gmax effect, before and after application of Lin-Glycine in each oocyte is used as a measure of the change in maximal conductance. To understand the concentration dependence of LIN-Glycine effect on G_max, the following concentration-response curve was fitted to the data:

where B is the maximum relative increase in G_max {(G_max - G_max0)/G_max0}, [PUFA]₅₀ the PUFA concentration needed to cause 50% of the maximal effect, and H the Hill coefficient. Average values are expressed as mean ± SEM and n represents the number of experiments.

Statistical analysis was conducted using GraphPad Prism 8 (GraphPad Software, Boston, Massachusetts, USA) and statistical tests used were One-way ANOVA with Dunnett’s post-hoc multiple comparisons test and student t test.

Data were analyzed using Clampfit 10.7 (pCLAMP), Origin Pro (OriginLab Corporation) and GraphPad Prism 8 software (GraphPad Software, Boston, Massachusetts, USA).

Results

Lin-Glycine drastically increases the chance of KCNQ1/KCNE1 channel opening

Lin-Glycine has been shown to increase the Gmax in whole-oocyte recordings of KCNQ1/KCNE1 channels 2.5-fold ¹³. To better understand how Lin-Glycine increases Gmax, we here extended our analysis to the single-channel level to study the behavior of KCNQ1/KCNE1 in the absence and presence of Lin-Glycine (20µM) (Fig. 2). Representative traces (Fig. 2A-B) and all points histograms (Fig. 2C-D) suggest that there is no change in single channel conductance between KCNQ1/KCNE1 in control and with Lin-Glycine. This shows that the Gmax effect is not due to an increase in the single channel conductance. However, Lin-Glycine caused a decrease in the latency to first opening (In control, 1.78 ± 0.36 s, n = 41 sweeps and after 20 µM Lin-Glycine, 1.03 ± 0.05 s, n = 99 sweeps; p = 0.0025) and an increase in the number of non-empty sweeps. In control, there are a high number of empty sweeps as shown before for KCNQ1/KCNE1 channels ¹⁷. The average current during the single channel sweeps was increased by 2.3 ± 0.33 times (n = 4 patches, p =0.0006) by the application of Lin-Glycine (Fig. 2E). The number of non-empty sweeps increased 2.85-fold (2.85 ± 0.85, n = 200; p = 0.001), going from a total of 41/200 non-empty sweeps in control to 99/200 when Lin-Glycine was applied. In contrast, once the channel opens, it seems to behave very similarly in control and Lin-Glycine, because, in non-empty sweeps, the open probability (Po) in the last second of the traces was almost identical between control and Lin-Glycine conditions (Po = 0.75 ± 0.12 (n = 8) in control vs 0.87 ± 0.04 (n = 3) in Lin-Glycine) (Fig. 2F). This high open probability was maintained in control conditions even for longer depolarizing voltage pulses, as if once the channels had opened they stayed open for the remaining of the voltage steps (Suppl. Fig. S1). The 2.85-fold increase in the number of non-empty sweeps is very similar to the Gmax increase seen in macroscopic currents.

Figure 2.

The decrease in first latency is most likely due to an effect of Lin-Glycine on Site I in the VSD and related to the shift in voltage dependence caused by Lin-Glycine. In contrast, the increase in the number of non-empty sweeps is most likely an effect on Site II in the pore and related to the Gmax effect. We conclude that the Gmax effect of Lin-Glycine on KCNQ1/KCNE1 is mainly due to an increase in the Po by increasing the number of non-empty sweeps.

Supplementary Figure S1.

PUFAs increase Gmax by binding to a crevice in the pore domain

It was previously shown that PUFAs binding at Site II electrostatically interact with the positively charged residue K326, located just outside the selectivity filter ¹⁸. We tested whether another residue very close to K326, the aspartic acid at position 301, is important for the PUFA interaction at Site II. Electrophysiological analysis revealed that when KCNQ1_WT/KCNE1 was mutated to KCNQ1_D301E/KCNE1, the ΔV0.5 was shifted similarly to the WT channel (Supplementary figure S2) but a dramatic decrease of the Gmax effect of Lin-Glycine was observed, showing the importance of this residue for PUFAs interaction at Site II (Figure 3A-C). However, it is not clear how PUFA binding at Site II increases Gmax.

Figure 3.

In our previous MD simulations ¹⁴ based on the cryoEM structure of KCNQ1 with S4 in the activated state, PUFAs bind at Site II in a narrow crevice between K326 and D301 (Figure 3D-E). PUFA fits into this space with both the head and the tail (Figure 3D-E). However, in the recent cryoEM structure of KCNQ1 with S4 in the resting state ²¹ the crevice between K326 and D301 is now so narrow that PUFAs cannot fit into Site II (Figure 3F). In addition, there are large rearrangements in the selectivity filter when comparing the cryoEM structures with S4 activated (PDB: 8SIK) and S4 resting (PDB: 8SIN) (Figure 4). We have made a video showing the changes and the reorganization of the pore between these two structures (Supplement Video S1). The conformation of the pore significantly varies between what seems like a conductive (PDB: 8SIK) and a non-conductive (PDB: 8SIN) state of the channel. For example, in the structure with S4 in the resting state, Y315 and D317 have swung outwards and left their positions in which they made hydrogen bonds with two tryptophans, W304 and W305, respectively. Y315 belongs to the K⁺ channels signature sequence, a stretch of 8 amino acids including TXXTXGYG, highly conserved among K⁺ channels of different families and D317 is just next to this sequence ^{22, 23}. W304 and W305 belong to the aromatic ring cuff (Figure 4C-D) that has been suggested to be important for the stability of the open state of the selectivity filter in Kv channels ^24,25.

Figure 4.

In addition, P320, which in the structure with an activated S4 sits in between W304 and W305 from two subunits (Figure 4C), has also swung outwards in the structure with S4 in the resting state and exposed a gap between W304 and W305 in the aromatic cuff (Figure 4D, see also supplementary Video S2). The homologous proline in other Kv channels sits between the two tryptophans and stabilizes the aromatic ring cuff.

Supplementary Figure S2.

Based on our single channel data with many empty sweeps, we propose the hypothesis that the KCNQ1 selectivity filter is inherently unstable and transitions between the two conformations - one conductive state that generates the non-empty sweeps and one non-conductive state that generates the empty sweeps in the single channel data. We further propose that PUFA binding to Site II stabilizes and promotes the conductive state of the pore. If PUFA binding into the narrow crevice between K326 and D301 promotes a series of interactions that stabilize the pore, then residues at the upper part of the selectivity filter, such as Y315 and D317, might be involved in those interactions that promote an open-conductive conformation of the channel.

Residues that stabilize the selectivity filter are necessary for Gmax effect

If our pore stability hypothesis is correct, we should see altered effects of Lin-Glycine on Gmax when residues important for the PUFA-promoted open-conductive conformation are mutated. We therefore mutated KCNQ1_Y315F and KCNQ1_D317E. We tested Lin-Glycine on KCNQ1_Y315F/KCNE1 and KCNQ1_D317E/KCNE1 channels and compared the Gmax effect to that of the WT channel. The effect of Lin-Glycine was reduced on both mutant channels, with a particularly dramatic decrease for Y315F (Figure 5A). For KCNQ1_Y315F/KCNE1 we compared the effect on the voltage-dependence of activation (ΔV0.5) and we observed a similar effect to the WT channel (Supplementary Figure S3), with both showing an ∼-25 mV shift. We also tested the Y315F mutation at the single channel level. As expected, Lin-Glycine did not increase the number of non-empty sweeps in KCNQ1_Y315F/KCNE1 channels (52/478 (10.9 %) of traces were non-empty in control and 44/533 (8.3%) of traces were non-empty in LIN-Gly) (Fig. 6). The mutation Y315F reduced the single channel current slightly from I = 0.4 pA in wildtype to I = 0.3 pA in the Y315F mutation (Fig. 6 C-D). The current average of all traces showed no increase in average current by Lin-Glycine (Fig. 6 E). This data shows that Y315 and D317 are necessary for the ability of Lin-Glycine to increase Gmax.

Figure 5.

Figure 6.

We also found that another residue in the pore helix, T309, was important for the Gmax effect of Lin-Glycine. The homologous residue in the Shaker channel was earlier suggested to be important for the stabilization of the selectivity filter by hydrogen bonding to one of the tryptophans in the aromatic ring ²⁶. We generated the mutant channel KCNQ1_T309S/KCNE1 and measured the effect of Lin-Glycine. As shown in Figure 5, the effect of Lin-Glycine on Gmax of the KCNQ1/KCNE1 mutant channel was noticeably reduced compared to the WT channel showing that this residue contributes to the Gmax effect (Figure 5A). We also tested the involvement of proline, which makes up a part of the aromatic cuff in the activated state of the channel (Fig. 4C) by creating the mutant channel KCNQ1_P320L/KCNE1 and found a significant reduction of the Gmax effect for this mutant (Figure 5A). All these data suggest that mutations of residues at the outer portion of the selectivity filter do affect the Gmax increase by Lin-Glycine. These data are consistent with our hypothesis that PUFAs increase the Gmax by affecting interactions that stabilize the selectivity filter.

The residues involved in the Gmax effect are found near the external region of the selectivity filter, suggesting that the network of interactions that are altered during the switch from non-conductive to conductive state are confined to the outer region of the selectivity filter and pore helix of KCNQ1/KCNE1 channels. To test the specificity of the network localization, we mutated two residues, T312 and I313, in the internal portion of the selectivity filter. As a confirmation of our hypothesis, we found that for both KCNQ1/KCNE1 mutant channels, T312S and I313S, the effect of Lin-Glycine in increasing the maximum conductance resembled values obtained in the WT channels (2.67 ± 0.98 and 2.18 ± 0.05, respectively) (Figure 5B). This indicates that mutations of residues in the more intracellular region of the selectivity filter do not affect the Gmax increase and that the interactions that stabilize the channel involve only residues located near the external region part of the selectivity filter.

Supplementary Figure S3.

Taken together, our results suggest that the binding of PUFA to Site II increases Gmax by promoting a series of interactions that stabilize the channel pore in the conductive state. For instance, we speculate that in the conductive state, hydrogen bonds between W304-D317 and W305-Y315, which are likely absent in the non-conductive conformation of KCNQ1, are created and that PUFA binding to Site II favors the transition towards the conductive state of the channel.

Figure 7.

Discussion

We have previously shown that the effects of PUFAs on IKs channels involve the binding of PUFAs to two independent sites: one at the voltage sensor (Site I) and one at the pore domain (Site II). The result is a potent activation of IKs channels, with an increase of the maximum conductance by binding to Site II and a shift of the voltage dependence of activation of the channel by binding to Site I. The mechanism of the voltage shift effects at Site I, where the PUFA head group electrostatically interacts with positively charged S4 residues and thereby facilitates channel activation, has been investigated in previous studies^{14, 27, 18, 28}. However, less is known about the molecular mechanism by which PUFA increases the maximum conductance (Gmax) of the channel following binding to Site II. A positively charged lysine at position 326 was suggested to be critical for PUFA-channel interaction, as mutations of K326 completely abolished the Gmax effect ¹⁸. We have here shown that the interaction at Site II also involves another residue, the aspartic acid 301, as mutagenesis analysis revealed that the effect of Lin-Glycine on Gmax was abolished when the residue was substituted with a glutamic acid.

Our single channel recordings revealed that Lin-Glycine decreases the latency to first opening and increases the Po. The decrease in the latency to first opening is most likely due to the binding of Lin-Glycine to Site I at the VSD and related to the shift in voltage dependence caused by Lin-Glycine. In addition, our single channel recordings revealed that Lin-Glycine did not change the single channel conductance, but instead increased Gmax by increasing the Po of the channel. This increase in Po was mainly due to an increase in non-empty sweeps when Lin-Glycine was applied in comparison to control solutions.

We tested the hypothesis that the effect of Lin-Glycine involved conformational changes in the selectivity filter following PUFA binding to the small crevice between residues K326 and D301 by making several mutations of residues known to stabilize the selectivity filter in potassium channels (Y315F, D317E, T309S and P320L; Fig. 5A). The Lin-Glycine effect on the Gmax was much reduced in these mutants suggesting that these residues are necessary for the Gmax effect. To gain further insight into the molecular interactions that could underline the Gmax effect by PUFA binding to site II, we used the latest KCNQ1 structure with the S4 segment in the resting state ²¹. The selectivity filter in this structure showed a very different conformation of the pore region and selectivity filter of KCNQ1 compared to the structures with activated VSDs. Many of the interactions that have been shown to be important for the stability of the selectivity filter are missing in the KCNQ1 structure with a resting VSD. These changes in the selectivity filter can best be seen in our interpolation video between the states with the S4 segments moving from the resting state to the activated state (Supplement video 1). This gave us the idea that the effect of PUFAs in increasing the maximum conductance of the KCNQ1 channel is linked to their ability to stabilize the pore of the channel in a conductive state. It was previously shown that several interactions at the pore region of K⁺ channels are important for ensuring channel conductivity. For example, a feature conserved among K⁺ channels is the aromatic ring cuff that stabilizes the conducting state and plays a role in C-type inactivation ^{24, 25, 26, 29}. This structure is made up of two tryptophan residues (W) and a proline residue (P) that sits in between the two tryptophan residues (Supplement video 2). In Shaker and KcsA K⁺ channels, the two tryptophans are involved in C-type inactivation and modification of those interactions manipulate the extent of inactivation ^{30, 31}. In Shaker, breaking the hydrogen bonds between the two tryptophans of the aromatic ring cuff and the aspartic acid and tyrosine at the selectivity filter causes an acceleration of the rate C-type inactivation ²⁶. In KCNQ1, the two tryptophans of the aromatic ring cuff correspond to W304 and W305. Another conserved residue in K⁺ channels important for the stability of the selectivity filter is P320 (equivalent to P450 in Shaker) that makes up part of the aromatic ring cuff. In the transition between the non-conductive and conductive state of KCNQ1, the proline pulls away from its position close to W304 and W305 and flips outwards (Supplement Video 2). We hypothesize that those hydrogen bonds between W304-D317 and W305-Y315 and the proline interactions are likely absent in the non-conductive conformation of KCNQ1, thereby generating the high number of empty sweeps in control conditions. Occasionally the channel will reform these bonds and interactions and transition into the conductive conformation, thereby generating the non-empty sweeps. We hypothesize that the binding of Lin-Glycine to K326 and D301 at site II biases the transition towards the conductive conformation, thereby increasing the number of non-empty sweeps and increasing Gmax. We here show that mutations of residues involved in these hydrogen bonds in the selectivity filter greatly reduce or abolish the Gmax effect of Lin-Glycine, as if Lin-Glycine increases the Gmax by stabilizing a conductive state of the selectivity filter through these hydrogen bonds.

We noticed that the arrangement of the aromatic ring cuff is slightly different between Shaker K⁺ channels and KCNQ1 channel. For instance, in Shaker, the proline residue sits in between the two tryptophan residues and stabilizes the aromatic ring cuff (Supplementary Figure S4 A). In contrast, in KCNQ1 the proline is positioned a little further outward and away from the two tryptophan residues, thus generating a looser arrangement of the aromatic ring cuff (Supplementary Figure S2 B). This difference might contribute to rendering the pore of KCNQ1 more unstable, resulting in a large number of empty sweeps and the characteristic flickering nature of channel openings ³². In Shaker channels the aromatic cuff seems more stable since it displays few empty sweeps and fewer flickering during bursts. This difference in the stability of the aromatic cuff between Shaker and KCNQ1 might explain why the effect of PUFA on Gmax is large for KCNQ1 but not seen in Shaker ^{33, 34} even if the residues involved in the Gmax effect are conserved between these two channels.

Supplementary figure S4.

Our studies suggest that the mechanism of PUFAs to increase KCNQ1/KCNE1 maximum conductance relies on the ability of PUFAs to favor a conductive conformation of the selectivity filter: PUFAs promote interactions between residues in the selectivity filter that stabilize the channel pore. Furthermore, we elucidated that Lin-Glycine binds to a crevice between K326 and D301 that is present in structures of KCNQ1 with activated VSD but not in structures where the VSD is in the resting state. We propose that PUFA binding in this crevice stabilizes the network of interactions that form the conductive form of the selectivity filter. The result is a more stable and conductive pore, which explains the increase in Gmax by PUFAs.