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_V7.1 α-subunit and the KCNE1 β-subunit to achieve expression of tetrameric I_Ks channels. Using two-electrode voltage-clamp recordings, we applied depolarizing voltage steps to activate the I_Ks channel (Figure 1B) before and after applying four different tyrosine PUFA analogues: N(α-linolenoyl)-tyrosine (NALT), linoleoyl tyrosine (Lin-tyrosine), docosahexanoyl tyrosine (DHA-tyrosine), and pinoleoyl tyrosine (Pin-tyrosine) (Figure 1C). From these voltage-clamp experiments, we are also able to acquire dose–response curves for different aspects of I_Ks channel activation, including changes in overall I_Ks current (I/I₀, Figure 1D), changes in the voltage dependence of activation (ΔV_0.5, Figure 1E), and changes in the maximal channel conductance (Gmax/Gmax₀, Figure 1F). NALT, Lin-tyr, DHA-tyr, and Pin-tyr all activate the cardiac I_Ks channel by shifting the voltage dependence of I_Ks channel activation to more negative voltages (NALT: ΔV_0.5 = –56.2 ± 3.6 mV; Lin-tyr: –74.4 ± 4.1 mV; DHA-tyr: –72.0 ± 4.9 mV; and Pin-tyr: –60.5 ± 5.8 mV at 20 μM; Figure 1E) and increasing the maximal conductance (NALT: 1.43 ± 0.3; Lin-tyr: 2.0 ± 0.6; DHA-tyr: 1.5 ± 0.2; and Pin-tyr: 1.8 ± 0.4 at 20 μM; Figure 1F). Together, the left shift in V_0.5 and the increase in G_max increase the overall I_Ks current measured in response to a voltage step close to 0 mV (NALT: 5.14 ± 1.2; Lin-tyr: 12.8 ± 2.1; DHA-tyr: 5.0 ± 0.9; and Pin-tyr: 5.8 ± 0.9 at 20 μM; Figure 1D: see ‘Materials and methods’ for calculation of I/I₀). In comparison, Lin-glycine (a known I_Ks channel activator; Figure 1C) causes only modest leftward voltage shifts (ΔV_0.5 = –30.8 ± 5.4 mV at 20 μM; Figure 1E), but similar increases in maximal conductance (2.6 ± 0.5 at 20 μM; Figure 1F) and I/I₀ current (6.7 ± 1.1 at 20 μM; Figure 1D) as for most tyrosine PUFAs.

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/I₀ (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 NAL-phe: 1.2 ± 0.2 at 20 μM; Figure 2G and H). All of these effects (I/I₀, 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/I₀ 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.

Our data thus far indicates that it is the presence of the –OH group, not cation–pi interactions, that is critical for pronounced I_Ks channel activation by tyrosine PUFAs. The –OH group found in tyrosine PUFAs is highly electronegative. To test how electronegativity influences I_Ks channel activation, we compared three modified phenylalanine PUFAs, which all include a highly electronegative group(s) attached to the aromatic ring. We compared N-(α-linolenoyl)–4-bromo-L-phenylalanine (4Br-NAL-phe), N-(α-linolenoyl)–4-fluoro-L-phenylalanine (4F-NAL-phe), and N-(α-linolenoyl) 3,4,5-trifluorophenylalanine (3,4,5F-NAL-phe) (Figure 3A). 4Br-NAL-phe, 4F-NAL-phe, and 3,4,5F-NAL-phe application increases I/I₀ (4Br-NAL-phe: 4.6 ± 0.1 at 20 μM; 4F-NAL-phe: 4.6 ± 1.1 at 20 μM; 3,4,5F-NAL-phe: 7.1 ± 1.0 at 20 μM; Figure 3B–D), causes a leftward shift in the V_0.5 (4Br-NAL-phe: ΔV_0.5 = –22.8 ± 2.0 mV; 4F-NAL-phe: –23.9 ± 0.8 mV; 3,4,5F-NAL-phe: –32.4 ± 4.9 mV at 20 μM; Figure 3E and F) and increases the G_max (4Br-NAL-phe: 1.9 ± 0.1 at 20 μM; 4F-NAL-phe: 2.0 ± 0.5; 3,4,5F-NAL-phe: 2.4 ± 0.4 at 20 μM; Figure 3G and H). Increasing the number of highly electronegative groups significantly improves the effects of phenylalanine PUFAs on increasing I/I₀ and shifting the V_0.5, evidenced by significant increases in I/I₀ (*p=0.0186; Figure 3D) and a significantly greater leftward shift in the V_0.5 (**p=0.0096; Figure 3F) from 3,4,5F-NAL-phe compared to NAL-phe alone. Interestingly, though, NALT still causes the most prominent left-shift in the V_0.5 compared to 4Br-, 4F-, and 3,4,5F-NAL-phe (***p=0.0003; ***P=0.00028; and **p=0.0021, respectively). These data suggest that the presence of highly electronegative groups improve the activating effects of phenylalanine PUFAs on the I_Ks channel. However, they do not completely recapitulate the effects of tyrosine PUFAs on the shift in V_0.5 of the I_Ks channel.

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-(α-linolenoyl)–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/I₀ 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.

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 residue K326, located near the pore, has been previously shown to be important for the G_max increasing effect of non-aromatic PUFAs (Liin et al., 2018). We tested 3,4,5F NAL-phe, the largest 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.

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.

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 voltage-dependent 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 non-hydrogen-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 + NAL-phe: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).

In our proposed mechanism, aromatic PUFA analogues modulate the I_Ks channel via two independent interactions (Figure 5). In one site, the carboxyl group is interacting with R231 (in S4) and 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).

K_V7.1 and KCNE1 channel cRNA were transcribed using the mMessage mMachine T7 kit (Ambion). 50 ng of cRNA was injected at a 3:1, weight:weight (K_V7.1:KCNE1) ratio into defolliculated X. laevis oocytes (Ecocyte, Austin, TX) for I_Ks channel expression. Injected cells were incubated for 72–96 hr in standard ND96 solution (96 mM NaCl, 2 mM KCl, 1 mM MgCl₂, 1.8 mM CaCl₂, 5 mM HEPES; pH = 7.5) containing 1 mM pyruvate at 16°C prior to electrophysiological recordings.

Primers for K_V7.1 mutations:

Commercially available PUFAs N-(α-linolenoyl) tyrosine (NALT) item number 10032 and linoleoyl phenalanine (Lin-phe) item number 20063 were obtained from Cayman Chemical (Ann Arbor, MI) or Larodan (Solna, Sweden). Linoleoyl tyrosine (Lin-tyr), docosahexaenoyl tyrosine (DHA-tyr), and pinolenoyl tyrosine (Pin-tyr) were synthesized as described previously (Bohannon et al., 2020). NAL-phe, 4Br-NAL-phe, 4F-NAL-phe, 3,4,5F-NAL-phe, and 3F-NALT were synthesized similarly, with detailed descriptions of the synthesis procedures for each compound provided in the supplemental methods. PUFA analogues were kept at –20°C as 100 mM stock solutions in ethanol except 4Br-NAL-phe, 4F-NAL-phe, 3,4,5F-NAL-phe, and 3F-NALT where stock solutions were prepared as needed on the day of recording. Serial dilutions of the different PUFAs were prepared from stocks to make 0.2 μM, 0.7 μM, 2.0 μM, 7.0 μM, and 20 μM concentrations in ND96 solutions (pH = 7.5). NALT is a commercially available compound. The other compounds used in this study were synthesized in small amounts. Thus, NALT is used as the focus of this study.

X. laevis oocytes, co-expressing wild-type K_V7.1 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 MgCl₂, 1.8 mM CaCl₂, 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.

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:G(V)=Gmax1+e(V0.5−V)/s

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/I₀) 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:II0=1+ A1+Kmnxn

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/I₀, 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_max0 in control solution (0 μM). Graphs plotting mean and standard error of the mean (SEM) for I/I₀, ΔV_0.5, G_max, and K_m were generated using GraphPad Prism (GraphPad Software, La Jolla, CA).

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.

Mutations and newly synthesized PUFAs are available from the corresponding author upon reasonable request.