Mechanistic insights into robust cardiac IKs potassium channel activation by aromatic polyunsaturated fatty acid analogues

Voltage-gated potassium (KV) channels are important regulators of cellular excitability and control action potential repolarization in the heart and brain. KV channel mutations lead to disordered cellular excitability. Loss-of-function mutations, for example, result in membrane hyperexcitability, a characteristic of epilepsy and cardiac arrhythmias. Interventions intended to restore KV channel function have strong therapeutic potential in such disorders. Polyunsaturated fatty acids (PUFAs) and PUFA analogues comprise a class of KV channel activators with potential applications in the treatment of arrhythmogenic disorders such as long QT syndrome (LQTS). LQTS is caused by a loss-of-function of the cardiac IKs channel – a tetrameric potassium channel complex formed by KV7.1 and associated KCNE1 protein subunits. We have discovered a set of aromatic PUFA analogues that produce robust activation of the cardiac IKs channel, and a unique feature of these PUFA analogues is an aromatic, tyrosine head group. We determine the mechanisms through which tyrosine PUFA analogues exert strong activating effects on the IKs channel by generating modified aromatic head groups designed to probe cation–pi interactions, hydrogen bonding, and ionic interactions. We found that tyrosine PUFA analogues do not activate the IKs channel through cation–pi interactions, but instead do so through a combination of hydrogen bonding and ionic interactions.


Introduction
The delayed rectifier potassium channel (I Ks ) underlies a critical repolarizing current that determines the timing of the ventricular action potential (Nerbonne and Kass, 2005). The cardiac I Ks current is mediated by the association of the voltage-gated K + channel K V 7.1 α-subunit with the KCNE1 β-subunit (Abramochkin et al., 2018;Wang et al., 1998;Barhanin et al., 1996). The K V 7.1 α-subunit consists of six transmembrane spanning segments, denoted S1-S6 where S1-S4 form the voltagesensing domain (VSD) and S5-S6 form the pore domain (PD) (Sun and MacKinnon, 2017). The S4 segment contains several positively charged arginine residues that allow S4 to move outward, toward the extracellular side of the membrane, when the membrane becomes depolarized (Broomand et al., 2003). This outward movement of the S4 is transformed into pore opening as a result of conformational changes in the S4-S5 linker of K V 7.1 (Kalstrup and Blunck, 2018). Co-expression of KCNE1 with K V 7.1 imparts a more depolarized voltage dependence of activation, slower activation kinetics, and increased single-channel conductance compared to K V 7.1 alone (Sun et al., 2012;Yang and Sigworth, 1998). Loss-of-function mutations in the cardiac I Ks channel can lead to an arrhythmogenic disorder known as long QT syndrome (LQTS), which predisposes individuals to ventricular fibrillation and sudden cardiac death (Fernández-Falgueras et al., 2017;Roden, 2008;Watanabe et al., 2005). Current treatments for LQTS include pharmacological intervention with β-blockers or surgical implantation of a cardioverter defibrillator (Waddell-Smith and Skinner, 2016). However, limitations of these treatments generate a need for novel therapeutic interventions to treat LQTS.
Polyunsaturated fatty acids (PUFAs) are amphipathic molecules composed of a charged hydrophilic head group and a long, polyunsaturated hydrophobic tail group (Jump, 2002). It is well documented that PUFAs form a group of I Ks channel activators that interact with the channel VSD, thus influencing the voltage dependence of I Ks channel activation (Moreno et al., 2015;Moreno et al., 2012;Elinder and Liin, 2017). The ability of PUFAs to activate the I Ks channel makes them great candidates for potential LQTS therapeutics. PUFAs promote I Ks channel activation through an electrostatic interaction between the negative charge of the hydrophilic PUFA head and positively charged arginine residues in the S4 segment of the I Ks channel (Elinder and Liin, 2017;Börjesson and Elinder, 2011;Larsson et al., 2020;Börjesson et al., 2008). This electrostatic activation of the I Ks channel is seen as a leftward shift in the voltage dependence of I Ks channel activation that leads to increases in I Ks current. Recently, it has been reported that PUFAs increase I Ks current through two independent effects: one on S4 (as described above) and one on the PD through an electrostatic interaction with a positively charged lysine residue located in S6 (K326) (Yazdi et al., 2021;Liin et al., 2018). This electrostatic interaction with the K326 mediates an increase in the maximal conductance (G max ) of the I Ks channel (Yazdi et al., 2021;Liin et al., 2018). The mechanism through which the negatively charged PUFA head group interacts with positive charges of S4 and S6 is called the lipoelectric hypothesis where the polyunsaturated tail of PUFAs and PUFA analogues incorporates into the cell membrane via hydrophobic interactions and electrostatically attracts the outermost gating charges of S4 as well as positively charged K326 in the S6 segment (Börjesson et al., 2008;Yazdi et al., 2021;Bohannon et al., 2019;Bohannon et al., 2020).
PUFA analogues that have the most robust effects on increasing I Ks current are those that have a low pKa value and thus possess a negatively charged head group at physiological pH (Bohannon et al., 2020). Examples include PUFAs with glycine or taurine head groups that possess either a carboxyl or sulfonyl head group, respectively (Bohannon et al., 2020;Liin et al., 2016). We have observed that another PUFA analogue, N-(α-linolenoyl) tyrosine (NALT), has robust effects on I Ks current. NALT is unique in that it possesses a large aromatic tyrosine head group rather than a carboxyl or sulfonyl group present in most of the PUFAs and PUFA analogues that we have characterized. NALT induces a potent leftward shift in the voltage dependence of I Ks channel activation and an increase in the maximal channel conductance, thus increasing overall I Ks current. Here, we aim to determine the mechanism behind I Ks activation by NALT using PUFA analogues with aromatic and modified aromatic head groups. Better understanding of the mechanism of how these aromatic PUFA analogues have more potent I Ks channel activation effects could help to aid future drug development for LQTS.

Results
Diverse PUFA analogues with a tyrosine head group activate the I Ks channel To measure the effects of the aromatic PUFA analogues on the cardiac I Ks channel, we expressed the I Ks channel complex in Xenopus laevis oocytes ( Figure 1A). We co-injected mRNA for the K V 7.1 α-subunit and the KCNE1 β-subunit to achieve expression of tetrameric I Ks channels. Using two-electrode Structures of NALT, Lin-tyrosine, DHA-tyrosine, and Pin-tyrosine (with Lin-glycine for comparison). (D-F) I/I 0 , (E) ΔV 0.5 , and (F) G max dose-response curves for NALT (black circles) (n = 4), Lin-tyrosine (pink circles) (n = 4), DHA-tyrosine (teal circles) (n = 3), Pin-tyrosine (purple circles) (n = 5), and Lin-glycine (gray dotted line) (n = 3). Values for all compounds and concentrations available in Figure 1-source data 1.

Distal -OH group is necessary for robust activation of the I Ks channel
Amino acids with aromatic groups (like tryptophan, tyrosine, and phenylalanine) can participate in cation-pi interactions (Dougherty, 2007). Cation-pi interactions take place between the pi-electrons of an aromatic ring and positively charged (cationic) groups (such as arginine and lysine) (Infield et al., 2021). If tyrosine PUFAs activate the I Ks channel via cation-pi interactions, we would expect that other aromatic groups (such as phenylalanine) would similarly affect I Ks activation. We tested two different PUFA analogues that both contain a phenylalanine head group -linoleoyl phenylalanine (Lin-phe) and N-(α-linolenoyl) phenylalanine (NAL-phe) ( Figure 2A). Lin-phe and NAL-phe both increase I/I 0 (Lin phe: 2.6 ± 0.3; and NAL-phe: 2.4 ± 0.5 at 20 μM; Figure 2B-D), causing a modest leftward shift in the V 0.5 (Lin-phe: ΔV 0.5 = -13.1 ± 2.9 mV; and NAL-phe: -12.5 ± 3.8 mV at 20 μM; Figure 2E and F). However, Lin-phe and NAL-phe have minimal effects on the G max (Lin phe: 1.2 ± 0.1; and NALphe: 1.2 ± 0.2 at 20 μM; Figure 2G and H). All of these effects (I/I 0 , ΔV 0.5 , and G max ) are reduced in comparison with tyrosine PUFAs, with Lin-phe and NAL-phe having an attenuated effect compared to NALT (p=0.51 and p=0.44, respectively; Figure 2D) and resulting in a significantly smaller increases in I/I 0 compared to Lin-tyrosine (***p=0.0004 and ***p=0.0004, respectively; Figure 2D). In addition, both NALT and Lin-tyrosine cause a significantly greater ΔV 0.5 compared to NAL-phe and Lin-phe (****p<0.0001; Figure 2F). Together, these differences suggest that cation-pi interactions are not the primary mechanism through which tyrosine PUFAs activate the I Ks channel. Rather, our data suggest that it is actually the presence of the distal -OH group on the aromatic head group that is critical for the potent activation of the I Ks channel because the loss of this -OH group (Lin-phe and NAL-phe) results in pronounced reductions in PUFA efficacy.

Hydrogen bonding is important for pronounced leftward shifts in I Ks channel voltage dependence
The presence of the -OH group on tyrosine PUFA analogues or the addition of electronegative groups to the phenylalanine head group improves I Ks activation. However, a persistent and striking difference between tyrosine PUFAs and modified phenylalanine PUFAs in the magnitude of their voltage-shifting effects with the tyrosine PUFAs having an almost twice as big voltage shift effect than the modified phenylalanine PUFAs ( Figure 3E and F). One explanation for this discrepancy is that the -OH group can also behave as a hydrogen bond donor. To determine whether hydrogen bonding contributes to the activating effects of tyrosine PUFA analogues, we applied the modified aromatic PUFA analogue N-(αlinolenoyl)-3-fluoro-L-tyrosine (3F-NALT), which has a fluorine atom adjacent to the tyrosine hydroxyl group ( Figure 4A). The addition of the fluorine atom reduces the pK a of the distal hydroxyl group and increases the hydrogen bonding ability of said group in 3F-NALT as compared to NALT. Overall, the maximum effects on I/I 0 are similar for 3F-NALT and NALT (3F-NALT: 5.0 ± 1.0; NALT: 5.14 ± 1.2 at 20 μM; p=0.7257, ns; Figure 4B-D). Notably, 3F-NALT induces a significantly greater maximum shift in the V 0.5 (ΔV 0.5 = -69.3 ± 1.4 at 20 μM) compared to NALT (-56.1 ± 3.6 AT 20 μM) (p=0.0298*; Figure 4E and F), while the effects on G max are not significantly different between 3F-NALT and NALT (3F-NALT: 1.3 ± 0.3; NALT: 1.4 ± 0.3 at 20 μM; p=0.7324, ns; Figure 4G and H). These data demonstrate that increasing the hydrogen bonding capacity of the -OH group increases the maximum shift in I Ks channel voltage dependence. This implicates hydrogen bonding as an important mechanism for I Ks activation and preferentially influences the effects on the voltage dependence of I Ks activation.

Aromatic PUFAs appear to activate the I Ks channel in similar mechanisms as non-aromatic PUFAs
To better understand the mechanism of these more effective aromatic PUFAs, we mutated residues previously shown to be important for non-aromatic PUFA activating effects on I Ks channels. The residue R231, located in the voltage sensor (S4) ( Figure 5A), has been previously shown to be important for the V 0.5 shifting effect of non-aromatic PUFAs (Liin et al., 2018). We tested Lin-tyr, the largest V 0.5 shifting aromatic PUFA, on the I Ks channel with the mutation R231Q+Q234R to assess whether R231 is also important for the aromatic PUFA V 0.5 shifting mechanism. The additional mutation Q234R is necessary to preserve the voltage dependence of activation in I Ks channels with the R231Q mutation (Liin et al., 2018;Panaghie and Abbott, 2007;Wu et al., 2010). The V 0.5 shifting effect of Lin-tyr was significantly decreased from ΔV 0.5 = -74.4 mV ± 4.1 at 20 μM in the wild-type (WT) I Ks channel toΔV 0.5 = -36.5 mV ± 7.3 at 20 μM with the R231Q+Q234R mutation (****p < 0.0001; Figure 5B and C). This reduction indicates that R231 contributes to more than half of the voltage dependence shifting effect of Lin-tyr. The remaining shift is most likely due to PUFA head group interactions with other nearby S4 charges such as R228 and Q234R.
The online version of this article includes the following source data for figure 2: Source data 1. Source Data for Panels C-H. G max increasing aromatic PUFA, on the I Ks channel with the mutation K326C to assess whether K326 is also important for the aromatic PUFA G max increasing mechanism ( Figure 5D). The G max increasing effect of 3,4,5F NAL-phe was significantly decreased from 2.4 ± 0.4 at 20 μM in the WT I Ks channel to 1.22 ± 0.2 at 20 μM (*p=0.0287; Figure 5E and F). This reduction indicates that K326 is necessary for 3,4,5F NAL-phe's G max increasing effect.

Figure 2 continued
To test whether the V 0.5 shift and G max increasing effects of aromatic PUFAs are independent of each other, we compare the Lin-tyr effects on both of the above mutations. The mutation R231Q+Q234R significantly decreases the V 0.5 shift effect from ΔV 0.5 = -74.4 mV ± 4.1 at 20 μM in the WT I Ks channel to ΔV 0.5 = -36.5 mV ± 7.3 at 20 μM in the mutation (****p < 0.0001), as mentioned above, but does not change G max increasing effect of Lin-tyr ( Figure 5G and H). Similarly, the mutation K326C significantly decreases the G max effect of Lin-tyr from G max = 2.04 ± 0.59 at 20 μM in the WT I Ks channel to G max = 1.08 ± 0.15 at 20 μM in the mutation (***p=0.0015) but does not change the V 0.5 shift effect of Lin-tyr ( Figure 5G and H). Overall, Figure 5 demonstrates that aromatic PUFAs are shifting the voltage dependence and increasing the current maximum by two independent mechanisms: one via interactions with S4 charges (R231) and one via interactions with K326.

Residue T224 in the S3-S4 loop is a novel locus for hydrogen bond formation between the I KS channel and tyrosine PUFAs
Our experiments using fluorinated NALT (NAL-3F-tyr) to improve the hydrogen bonding capacity of the tyrosine head group demonstrated that hydrogen bonding by the tyrosine's para-hydroxyl group is the reason for the large effect of PUFAs with tyrosine head groups on the I Ks channel voltagedependent activation. To identify the residue with which the tyrosine head group hydrogen bonds, we mutated residues in the S3-S4 loop capable of hydrogen bond formation. We chose residues in the S3-S4 loop because it is located near the top of the PUFA head group in the proposed binding site for PUFAs to have the voltage-dependence effect. We individually mutated serine 217 (S217A), glutamine 220 (Q220L), threonine 224 (T224V), and serine 225 (S225A) and compared the effects of NALT on mutated channels compared to the WT I Ks channel ( Figure 6A and B). We found that S217A, Q220L, and S225A showed similar maximum shifts in voltage-dependent activation compared to the wild-type channel (WT + NALT: ΔV 0.5 = -56.1 ± 3.6 mV; S217A + NALT: -65.9 ± 3.7 mV; Q220L + NALT: -59.5 ± 11.1 mV; S225A + NALT: -52.4 ± 3.7 mV at 20 μM, ns; Figure 6C and D). However, the T224V mutation significantly attenuated the leftward shift in the voltage dependence of activation in response to NALT application from ΔV 0.5 = -56.1 ± 3.6 mV in WT channels to ΔV 0.5 = -32.1 ± 7.0 at 20 μM (*p=0.03; Figure 6D). To determine whether this effect was specific to compounds with the ability to form hydrogen bonds, we compared the effects of hydrogen-bonding NALT and nonhydrogen-bonding NAL-phe on T224V mutant channels ( Figure 6E). In contrast to the attenuation of the overall voltage shift observed when NALT was applied to the T224V, there was no difference in the voltage-shifting effects of NAL-phe between the T224V mutant and WT channels (WT + NALphe:ΔV 0.5 = -12.5 ± 3.8 mV; T224V + NAL-phe: -13.2 ± 2.6 mV at 20 μM, ns; Figure 6F and G). These data demonstrate that the T224V mutation only reduces the efficacy of aromatic PUFAs that contain a hydrogen-bonding group like tyrosine. As a result, we have identified a novel interaction between the S3-S4 loop residue T224 and hydrogen bonding moieties of aromatic PUFA head groups. Figure 7A shows the distance between the residues R231 and T224V in KCNQ1 is 9.1 Å and the length of a tyrosine head group is 8.5 Å. The tyrosine head group fits nicely in between these two residues in silico, demonstrating the possibility of this interaction to occur ( Figure 7A).
The online version of this article includes the following source data for figure 3: Source data 1. Source Data for Panels C-H.  the distal -OH is hydrogen bonding with T224 (in the S3-S4 loop) ( Figure 7B). In the other site, the carboxyl group is interacting with K326 (in S6) and the distal electronegative atom stabilizes the PUFA via an electrostatic interaction ( Figure 7B).

Discussion
We have found that PUFA analogues with tyrosine head groups are strong activators of the cardiac I Ks channel. Tyrosine PUFAs shift the voltage dependence of activation to negative potentials and increase the maximal conductance, which together contribute to increases in overall I Ks current. The tyrosine head group is an aromatic ring with a distal -OH group in the para-position. Tyrosine PUFA analogues have the potential to interact with the I Ks channel through several candidate mechanisms involving either the aromatic ring or the -OH group (or both). The aromatic ring could modulate I Ks channel function through cation-pi interactions with positively charged groups on the I Ks channel. In addition, the -OH group could participate in electrostatic interactions and/or act as a hydrogen bond donor. In this work, we elucidate the mechanisms of this PUFA-induced activation of the I Ks channel by applying PUFA analogues with modified aromatic head groups designed to test specific chemical interactions between the PUFA head group and the I Ks channel.
If cation-pi interactions were the primary mechanism through which tyrosine PUFAs activate the I Ks channel, we would expect similar activating effects of PUFA analogues with aromatic rings that lack the -OH group, such as phenylalanine. However, PUFA analogues with phenylalanine head groups (Lin-phe and NAL-phe) do not activate the I Ks channel to the same degree as PUFA analogues with a tyrosine head group (Lin-tyr and NALT) and display significant reductions in efficacy for increases in I/I 0 and shifts in the V 0.5 . Further evidence that cation-pi interactions are not a predominant mechanism for I Ks channel activation by tyrosine PUFA analogues comes from experiments applying fluorinated phenylalanine PUFAs (4F-NAL-phe and 3,4,5F-NAL-phe), which can be used as a tool to probe cation-pi interactions in ion channel function (Pless et al., 2014). Pless et al., 2014 demonstrated that tri-fluorination of phenylalanine disperses the electrostatic surface potential that is necessary for cation-pi interactions (Pless et al., 2014). Disruption of the electrostatic surface potential through addition of fluorine atoms to the NAL-phe head group (3,4,5F-NAL-phe), therefore, is expected to reduce the efficacy of 3,4,5F-NAL-phe in comparison to NAL-phe alone. However, we find the opposite when we apply 3,4,5F-NAL-phe to the cardiac I Ks channel, and see that 3,4,5F-NAL-phe is a more potent activator of the I Ks channel compared to NAL-phe alone. Together, these data suggest that cation-pi interactions are not the primary mechanism through which these aromatic PUFA analogues activate the cardiac I Ks channel.
When we look at several fluorinated and brominated phenylalanine PUFA analogues, we find specifically that 3,4,5F-NAL-phe has significantly greater effects on I/I 0 and ΔV 0.5 compared to NALphe alone. While not statistically significant, 4Br-, 4F-, and 3,4,5F-NAL-phe also lead to some of the most consistent increases in G max among the PUFA analogues tested in this work, with each of these compounds leading to a twofold increase in G max . These data suggest that aromatic PUFA analogues with highly electronegative atoms on the distal end of the aromatic head group have the most pronounced effects on the maximal conductance of the I Ks channel. Although brominated and fluorinated phenylalanine analogues increase the maximal conductance of the I Ks channel, these modified PUFAs still fail to recapitulate the leftward ΔV 0.5 observed with tyrosine PUFA analogues. While the -OH group of tyrosine PUFA analogues is indeed strongly electronegative, it can also act as a hydrogen bond donor. When we applied a fluorinated tyrosine PUFA (3F-NALT) to increase hydrogen bonding abilities, we found that this leads to a stronger leftward shift in the voltage dependence of I Ks activation. This suggests that hydrogen bonding via the -OH group contributes to the leftshifting effects of voltage-dependent activation through effects on the I Ks channel voltage sensor. Most notably, these results suggest that specific modifications to the aromatic PUFA head group can effects on (D) I/I 0 , (F) ΔV 0.5 , and (H) G max (at 20 μM) for 3F-NALT (n = 3) and NALT (n = 4). Values for all compounds and concentrations available in Figure 4-source data 1.

Figure 4 continued
The online version of this article includes the following source data for figure 5: Source data 1. Source Data for Panels B, C, E, F, G, and H. preferentially improve either the voltage-shifting or maximal conductance effects of PUFA analogues. Our data suggests that adding highly electronegative groups to an aromatic ring, such as bromine and fluorine, most consistently improve the maximal conductance-increasing effects and reduce voltage dependence-shifting effects relative to PUFA analogues with a tyrosine or phenylalanine head group. On the other hand, we found that reducing the pK a of the -OH group (and increasing the potential for hydrogen bonding), while leaving the effect on G max intact, preferentially improves the voltage-shifting effects on the I Ks channel.
Previous work has demonstrated that PUFA analogues have two independent effects on I Ks channel activation. PUFA analogues are known to shift the voltage dependence of activation in the I Ks channel through electrostatic effects on the channel voltage sensor (Börjesson et al., 2008;Liin et al., 2015). This is mediated by interactions of the negative PUFA head group with the outermost positively charged arginine residues located in the S4 segment (Liin et al., 2016;Liin et al., 2015). Recently, though, a second effect on the I Ks channel pore has been reported to influence the maximal conductance of the I Ks channel (Liin et al., 2018). This is mediated through electrostatic interactions between the PUFA head groups and a positively charged lysine residue in the S6 segment -K326 (Liin et al., 2018). In addition, molecular dynamics (MD) simulations with the K v 7.1 (KCNQ1) channel (the poreforming domain for the I Ks channel) (Yazdi et al., 2021) identified two separate high-occupancy sites for linoleic acid: site 1 at R231 in the S4 segment, and site 2 at K326 in the S6 segment (Yazdi et al., 2021). We here show that the more effective aromatic PUFAs also act on these sites in S4 and S6. To do this, we selected the best V 0.5 shifting aromatic PUFA (Lin-tyr) to test on the I Ks channel with the S4 mutation R231Q. Additionally, we selected the best G max increasing aromatic PUFA (3,4,5F-NAL-phe) to test on the I Ks channel with S6 mutation K326C. The mutation R231Q decreases the V 0.5 shifting effect of Lin-tyr by half, indicating that Lin-tyr is shifting the voltage dependence by creating an electrostatic interaction with the positive charges on the voltage sensor. Conversely, the mutation K326C almost completely removed the G max increasing effect of 3,4,5F-NAL-phe. To demonstrate that these effects are also independent of each other in aromatic PUFAs, we evaluate the V 0.5 shifting effect and the G max increasing effect of Lin-tyr on both mutations. The R231Q mutation decreases V 0.5 shifting effect but does not change the G max effect of Lin-tyr. While the K326 mutation decreases the G max effect but does not change the V 0.5 shifting effect of Lin-tyr. We therefore propose that the increased effects of the aromatic PUFAs, compared to non-aromatic PUFAs, are due to the additional hydrogen bonding in site 1 and electrostatic interactions in site 2 to better anchor them in these binding sites to increase their effects ( Figure 7B). As mentioned above, we also show that the aromatic rings have the potential to be modified to give preferential effects on either the I Ks channel voltage sensor or channel pore.
The online version of this article includes the following source data for figure 6: Source data 1. Source Data for Panel C, D, F, and G. dependence of mutation T224V was shifted significantly less than the WT I Ks channels. This shows that the -OH group on the tyrosine of NALT hydrogen bonds with T224V, thereby improving the PUFA's ability to shift the voltage dependence. This hydrogen bond interaction between PUFAs and the 3-4 loop of the I Ks channel is a novel mechanism to increase the effect of PUFAs to activate the I Ks channel. These data suggest that the drugs designed to target this interaction would be more effective at shifting I Ks channel voltage dependence. Overall, our findings suggest that different aromatic PUFA analogues not only increase PUFA efficacy on activating the I Ks channel, but their specific effects on I Ks function can be modulated independently, either increasing the maximal conductance or voltage-shifting effect. Patients with LQTS can have mutations in the I Ks channel that causes either a decrease in G max or a rightward shift of voltage dependence. By independently modulating PUFAs to either increase G max or cause a leftward shift in voltage dependence, we may be able to design therapies that are more personalized to each patient's specific LQTS mutation. This novel mechanistic understanding of how aromatic PUFAs have these effects on the I Ks channel may help to aid drug development for LQTS.

Two-electrode voltage-clamp (TEVC)
X. laevis oocytes, co-expressing wild-type K V 7.1 and KCNE1, were recorded in the two-electrode voltage-clamp (TEVC) configuration. Recording pipettes were filled with 3 M KCl. The recording chamber was filled with ND96 (96 mM NaCl, 2 mM KCl, 1 mM MgCl 2 , 1.8 mM CaCl 2 , 5 mM Tricine; pH 7.5). Dilutions of PUFAs and PUFA analogues were perfused into the recording chamber using the Rainin Dynamax Peristaltic Pump (Model RP-1) (Rainin Instrument Co., Oakland, CA). Electrophysiological recordings were obtained using Clampex 10.3 software (Axon, pClamp, Molecular Devices). During the application of PUFAs, the membrane potential was stepped every 30 s from -80 mV to 0 mV for 5 s before stepping to -40 mV and back to -80 mV to ensure that the PUFA effects on the current at 0 mV reached steady state ( Figure 1D). A voltage-step protocol was used to measure the current vs. voltage (I-V) relationship before PUFA application and after the PUFA effects had reached steady state for each concentration of PUFA. Cells were held at -80 mV followed by a hyperpolarizing prepulse to -140 mV to make sure all channels are fully closed. The voltage was then stepped from -100 to 60 mV (in 20 mV steps) followed by a subsequent voltage step to -20 mV to measure tail currents before returning to the -80 mV holding potential.

Data analysis
Tail currents were analyzed using Clampfit 10.3 software in order to obtain conductance vs. voltage (G-V) curves to determine the voltage dependence of channel activation. V 0.5 , the voltage at which half the maximal current occurs, was obtained by fitting the G-V curves from each concentration of PUFA with a Boltzmann equation: where G max is the maximal conductance at positive voltages and s is the slope factor in mV. The current values for each concentration at 0 mV (I/I 0 ) were used to plot the dose-response curves for each PUFA. These dose-response curves were fit using the Hill equation to obtain the K m value for each PUFA: where A is the fold increase in current caused by the PUFA at saturating concentrations, K m is the apparent affinity of the PUFA, x is the concentration, and n is the Hill coefficient. Fitted maximum values derived from the dose-response curves are reported for each of the effects (I/I 0 , ΔV 0.5 , and G max ) from the different PUFAs tested. In some cases, there is variability in the V 0.5 between batches of oocytes. In order to correct for variability due to oocytes, when the V 0.5 was greatly different than 20 mV in control solution, we applied a correction in order to more accurately measure PUFA-induced I Ks current increases. We subtracted the V 0.5 (given by fitting the G-V with a Boltzmann equation) by 20 mV and used the current measured at the resulting voltage. The maximum conductance (G max ) was calculated by taking the difference between the maximum and minimum current values (using the G-V curve for each concentration) and then normalizing to the value G max 0 in control solution (0 μM). Graphs plotting mean and standard error of the mean (SEM) for I/I 0 , ΔV 0.5 , G max , and K m were generated using GraphPad Prism (GraphPad Software, La Jolla, CA).

Statistics
Unpaired t-tests and one-way ANOVA with multiple-comparisons statistics were computed using GraphPad Prism (GraphPad Software). Results were considered significant if p<0.05.

Materials availability statement
Mutations and newly synthesized PUFAs are available from the corresponding author upon reasonable request.
HPL and SIL identified as inventors. H Peter Larsson: A patent application (#62/032,739) including a description of the interaction of charged lipophilic compounds with the KCNQ1 channel has been submitted by the University of Miamiwith HPL and SIL identified as inventors.Dr Hans Peter Larsson is the equity owner of VentricPharm, a company that operates in the same field of research as the study. The other authors declare that no competing interests exist. The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.