1. Structural Biology and Molecular Biophysics
Download icon

Fatty acid analogue N-arachidonoyl taurine restores function of IKs channels with diverse long QT mutations

  1. Sara I Liin  Is a corresponding author
  2. Johan E Larsson
  3. Rene Barro-Soria
  4. Bo Hjorth Bentzen
  5. H Peter Larsson  Is a corresponding author
  1. University of Miami, United States
  2. Linköping University, Sweden
  3. University of Copenhagen, Denmark
Research Article
  • Cited 8
  • Views 879
  • Annotations
Cite this article as: eLife 2016;5:e20272 doi: 10.7554/eLife.20272

Abstract

About 300 loss-of-function mutations in the IKs channel have been identified in patients with Long QT syndrome and cardiac arrhythmia. How specific mutations cause arrhythmia is largely unknown and there are no approved IKs channel activators for treatment of these arrhythmias. We find that several Long QT syndrome-associated IKs channel mutations shift channel voltage dependence and accelerate channel closing. Voltage-clamp fluorometry experiments and kinetic modeling suggest that similar mutation-induced alterations in IKs channel currents may be caused by different molecular mechanisms. Finally, we find that the fatty acid analogue N-arachidonoyl taurine restores channel gating of many different mutant channels, even though the mutations are in different domains of the IKs channel and affect the channel by different molecular mechanisms. N-arachidonoyl taurine is therefore an interesting prototype compound that may inspire development of future IKs channel activators to treat Long QT syndrome caused by diverse IKs channel mutations.

https://doi.org/10.7554/eLife.20272.001

eLife digest

Every heartbeat relies on an electric wave that travels through the heart. This wave must reach different parts of the heart in a specific sequence to ensure that the heart muscle cells contract in a coordinated manner. Such coordinated contractions enable the heart to pump enough blood around the body. By allowing specific ions to flow into or out of the heart muscle cell, proteins called ion channels in the cell membrane generate the electric wave, keep it going and stop it. One such protein called the IKs channel controls the flow of potassium ions, and in doing so stops the electric wave in heart muscle cells.

About 300 different mutations in the IKs channel have been shown to cause abnormal heart rhythms in individuals with a disorder called long QT syndrome. People with this condition may suddenly black out because their heart develops prolonged electric waves that prevent blood from being pumped properly.

To investigate how mutations in the IKs channel produce heart rhythm abnormalities, Liin et al. genetically engineered the egg cells of African clawed frogs to have one of eight mutant forms of the human IKs channel. Studying these channels revealed that the mutations reduce how well the channels work in a wide variety of ways. However, treating the cells with a particular fatty acid helped to normalize how each of the mutant channels worked. Therefore, variants of the fatty acid could potentially form a useful treatment for people with heart rhythm problems caused by mutations in the IKs channel.

More studies are needed to confirm whether the fatty acid is as effective at combating the effects of the mutations in whole hearts and animals. As ion channels related to the IKs channel are found in many types of cells, it is also important to investigate whether treatment with the fatty acid could cause any side effects that affect other organs.

https://doi.org/10.7554/eLife.20272.002

Introduction

Long QT syndrome (LQTS) is a condition of the heart which in most cases is caused by a mutation in cardiac ion channels (Hedley et al., 2009; Morita et al., 2008). In LQTS, the action potential of the heart is prolonged, which is observed as a prolonged QT interval in the electrocardiogram. LQTS patients have an increased risk of developing ventricular tachyarrhythmias called torsades de pointes when exposed to triggers such as adrenergic stress (Morita et al., 2008; Cerrone et al., 2012). These arrhythmias can cause palpitation, syncope or sudden death due to ventricular fibrillation. To improve the clinical outcome of LQTS patients, it is therefore critical to prevent these LQTS-induced life-threatening arrhythmias.

Most mutations causing LQTS are located in the KCNQ1 gene (Hedley et al., 2009). KCNQ1 codes for the potassium channel KV7.1, which in the heart co-assembles with the beta-subunit KCNE1 to form the slowly-activating, voltage-dependent potassium channel IKs (Barhanin et al., 1996; Sanguinetti et al., 1996). The IKs channel provides one of the important delayed rectifier outward potassium currents that repolarizes the cardiomyocyte and terminates the cardiac action potential (Nerbonne and Kass, 2005). Reduced IKs function therefore tends to delay cardiomyocyte repolarization, thereby causing prolonged cardiac action potential durations and a prolonged QT interval. The cardiac IKs channel consists of four KV7.1 subunits and two to four KCNE1 subunits (Nakajo et al., 2010; Plant et al., 2014; Murray et al., 2016). Throughout this work, we will refer to the IKs channel as KV7.1+KCNE1. KV7.1 has six transmembrane segments named S1-S6 (Liin et al., 2015) (Figure 1a). S1-S4 of each KV7.1 subunit forms a voltage-sensing domain where S4 is the voltage sensor with three positive gating charges. S5 and S6 from all four KV7.1 subunits form the pore domain with a putative gate in S6 that needs to move to open the ion-conducting pore of the channel. KCNE1 has a single-transmembrane segment (Figure 1a) and is proposed to be localized in the otherwise lipid-filled space between two voltage-sensing domains of neighbouring KV7.1 subunits (Nakajo and Kubo, 2015). Upon cardiomyocyte depolarization, the voltage sensor of KV7.1 moves outward in relation to the membrane. It has been proposed that this movement of the voltage sensor is transferred to the pore domain via the S4-S5 linker and induces channel opening by moving the S6 gate (Liin et al., 2015).

Figure 1 with 4 supplements see all
Biophysical properties of LQTS and LQTS-like KV7.1+KCNE1 channel mutants expressed in Xenopus oocytes.

(a) Topology of KV7.1 and KCNE1, and position of tested LQTS and LQTS-like mutants. (b) G(V) midpoints (V50) from the Boltzmann fits for mutants co-expressed with KCNE1. n = 5–11. Data as mean ± SEM. The statistics represent one-way ANOVA with Dunnett’s Multiple Comparison Test to compare the mutants to wild-type KV7.1+KCNE1; **p<0.01; ns is p≥0.05. # denotes lowest estimate. Dashed line denotes wild-type V50. (c) Representative example of KV7.1/S225L+KCNE1 G(V) (black line and symbols) compared to wild-type KV7.1+KCNE1 (blue line and symbols, mean ± SEM, n = 5). (de) Representative example of KV7.1/S225L+KCNE1 opening kinetics and KV7.1+KCNE1/K70N closing kinetics (black lines) compared to wild-type KV7.1+KCNE1 (blue lines).

https://doi.org/10.7554/eLife.20272.003

Altogether, about 300 mutations in KCNQ1 and KCNE1 have been identified in patients suffering from LQTS (Hedley et al., 2009) (http://www.fsm.it/cardmoc/). These mutations are distributed throughout the channel sequence and are therefore likely to cause channel dysfunction by different mechanisms, which are, however, largely unknown. Potential mechanisms for KV7.1+KCNE1 channel loss of function by a mutation could, for example, be interference with voltage sensor movement, gate opening, or membrane expression. LQTS is today treated with drugs that prevent the triggering of arrhythmic activity, such as beta-blockers, or with arrhythmia-terminating implantable cardioverter defibrillator (Hedley et al., 2009). A different treatment strategy for LQTS caused by loss-of-function mutations in the KV7.1+KCNE1 channel would be to pharmacologically augment the KV7.1+KCNE1 channel function of these LQTS mutants, thereby shortening the prolonged QT interval and lower the risk of arrhythmia development. However, there is currently no clinically approved KV7.1+KCNE1 channel activator.

In this study, we investigate the biophysical properties and potential mechanism of action of LQTS-associated KV7.1+KCNE1 channel mutations and test the ability of the fatty acid analogue N-arachidonoyl taurine (N-AT) to restore the function of these mutants.

We selected eight mutations of residues mutated in patients with LQTS located in different segments of the KV7.1+KCNE1 channel and that were previously shown to form active channels (Bianchi et al., 2000; Yamaguchi et al., 2003; Eldstrom et al., 2010; Henrion et al., 2009; Yang et al., 2013; Yang et al., 2002; Harmer et al., 2010; Splawski et al., 1997). We measure the movement of the S4 voltage sensor in selected mutants using voltage clamp fluorometry to further our understanding of the molecular mechanisms underlying the defects caused by the diverse mutations. We find that the eight LQTS-associated mutations affect the voltage dependence and/or closing kinetics, in some cases by different molecular mechanisms. Moreover, we find that N-AT restores much of the channel activity in these eight LQTS-associated KV7.1+KCNE1 mutants. This suggests that N-AT may function as a general activator of KV7.1+KCNE1 channels with diverse mutational defects.

Results

LQTS mutants show altered biophysical properties

We first study the biophysical properties of six point mutations in KV7.1 (F193L, V215M, S225L, L251P, F351S, R583C), and two in KCNE1 (K70N, S74L) identified in patients with LQTS (Yamaguchi et al., 2003; Yang et al., 2002; Splawski et al., 1997; Priori et al., 1999; Napolitano et al., 2005; Lai et al., 2005) (Figure 1a). As L251P and F351S did not produce functional channels (Napolitano et al., 2005; Deschenes et al., 2003) (Figure 1—figure supplement 1), we engineered the milder L251A and F351A mutants instead. L251A and F351A will be referred to as 'LQTS-like mutants'. When expressed alone in Xenopus oocytes, all investigated KV7.1 mutants, except F193L and V215M, display a shifted conductance versus voltage curve (G(V)) compared to the wild-type KV7.1 channel (Figure 1—figure supplement 2; Supplementary file 1). S225L, L251A and F351A shift the G(V) towards positive voltages compared to wild-type KV7.1. In contrast, R583C shifts the half-maximal activation, V50, ~10 mV towards negative voltages compared to wild-type KV7.1. This apparent negative shift is likely caused by the pronounced inactivation of the R583C mutant (Figure 1—figure supplement 3a), which is seen to a considerable smaller extent in the other KV7.1 mutants and wild-type KV7.1 (inset in Figure 1—figure supplement 3a). When a fraction of the channels are released from inactivation, by introducing a brief hyperpolarizing pulse between the test pulse and the tail pulse, R583C has a V50 fairly comparable to wild-type KV7.1 (Figure 1—figure supplement 3b).

When the KV7.1 mutants are co-expressed with KCNE1, all KV7.1 and KCNE1 mutants except KV7.1/F193L+KCNE1 have a G(V) that is shifted towards positive voltages compared to the wild-type KV7.1+KCNE1 channel (Figure 1b). KV7.1/F351A causes the most dramatic change by shifting V50 more than +30 mV. We are therefore only able to record the foot of the G(V) curve of KV7.1/F351A+KCNE1, and a shift in V50 of +30 mV is a lower estimate of the change in V50 (ΔV50). One of the other mutants with dramatically shifted G(V) is KV7.1/S225L+KCNE1. V50 for KV7.1/S225L+KCNE1 is shifted almost +30 mV compared to wild-type KV7.1+KCNE1 (Figure 1c; Supplementary file 1). S225L also slows down KV7.1+KCNE1 channel opening kinetics (p<0.01; Figure 1d; Supplementary file 1). All mutations, except for L251A, accelerate channel closing kinetics compared to wild-type KV7.1+KCNE1 (Supplementary file 1). K70N has the most dramatic effect on KV7.1+KCNE1 channel closing by accelerating the closing kinetics by approximately a factor of 5 (Figure 1e; Supplementary file 1). When comparing the amplitude of K+ currents generated by these mutants with the current amplitude of the wild-type KV7.1+KCNE1 channel in the same batch of oocytes, we note that all mutants generate smaller currents than wild-type over a large voltage range (Figure 1—figure supplement 4). Although defective trafficking may contribute to these reduced currents in Xenopus oocytes, the current amplitudes for most mutants matches fairly well with the predicted current amplitude from channels with G(V) curves shifted towards positive voltages as observed for these mutants (Figure 1—figure supplement 4a), suggesting that the reduced current amplitudes in Xenopus oocytes are mainly a result of gating defects (and not trafficking defects).

To summarize, all mutations change channel function by altering the voltage dependence of opening and/or the kinetics of opening and/or closing. Reduced function of the KV7.1+KCNE1 channel induced by these LQTS and LQTS-like mutations may largely be explained by the right-shifted G(V) and the faster closing kinetics caused by these mutations. F193L does not alter the G(V), but speeds up KV7.1+KCNE1 channel closing by a factor of 2 (Supplementary file 1). These results are consistent with previous reported findings for some of these mutants (Bianchi et al., 2000; Yamaguchi et al., 2003; Eldstrom et al., 2010; Henrion et al., 2009; Yang et al., 2013; Yang et al., 2002; Harmer et al., 2010).

Heterozygous expression reduces LQTS mutant severity

Patients with LQTS mutations can be either homozygous or heterozygous for the mutation. To mimic heterozygous expression, we co-inject the mutated KV7.1 subunit and KCNE1 subunit together with the wild-type KV7.1 subunit (or wild-type KCNE1 subunit for KCNE1 mutants) (cartoon in Figure 2). We refer to this as heterozygous expression. Figure 2a–b compares the homozygous expression (KV7.1wt+KCNE1mut or KV7.1mut+KCNE1wt) with heterozygous expression (KV7.1wt+KV7.1mut+KCNE1wt or KV7.1wt+KCNE1wt+KCNE1mut) for KV7.1/S225L (Figure 2a) and KCNE1/K70N (Figure 2b). Both of these examples show that heterozygous expression generates channels with more wild-type like opening or closing kinetics and G(V) compared to homozygous expression of the mutant subunit. A milder biophysical phenotype upon heterozygous expression is generally seen for the LQTS and LQTS-like mutants in terms of G(V), current amplitude, and/or closing kinetics (Figure 2c–d, Figure 1—figure supplement 4, Supplementary file 2). This milder phenotype indicates that the wild-type subunit can partly restore KV7.1+KCNE1 function. Alternatively, for mutants with a G(V) that is very shifted to positive voltages (e.g. F351A), it may be that channel complexes that contain the mutated subunits are largely out of the physiological voltage range and therefore do not contribute substantially to the recorded current. Also, for mutants with low membrane expression (e.g. possibly F193L [Yamaguchi et al., 2003]), it may be that channels containing the wild-type subunit are favoured so that in most KV7.1+KCNE1 channel complexes the majority (or all) of the subunits will be wild-type subunits.

Comparison of homozygous and heterozygous expression of LQTS and LQTS-like mutants.

(ab) Representative example of kinetics (middle panel) and G(V) (right panel) for homozygous expression and heterozygous expression of S225L (a) and K70N (b). Currents in response to steps from –80 mV to +40 mV (a, middle pane) and from +40 mV to –20 mV (b, middle panel). Homozygous expression (black), heterozygous expression (gray), and KV7.1+KCNE1 wild-type (blue). n = 7–13. (cd) Summary of V50 (c) and T50 for closing (d) for homozygous and heterozygous expression. Data as mean ± SEM. n = 5–13. The statistics represent one-way ANOVA with pair-wise Bonferroni’s Test to compare homozygous and heterozygous expression; **p<0.01; ***p<0.001; ns is p≥0.05. # denotes lowest estimate. Not determined (nd). The statistics was not calculated for F351A. Dashed lines denote corresponding values for wild-type KV7.1+KCNE1.

https://doi.org/10.7554/eLife.20272.008

Different mutants display different fluorescence versus voltage profiles

Although most of the mutations shift channel voltage dependence and affect channel closing kinetics, the underlying mechanism of mutation-induced changes in KV7.1+KCNE1 channel function is most likely different for different mutations. For instance, mutations located in S5 and S6 (e.g. F351A) may mainly affect gate movement, while mutations in S1–S4 (e.g. S225L) are more likely to affect voltage sensor movement. To explore whether different mutations interfere with different gating transitions, we use voltage clamp fluorometry, in which the movement of the voltage sensor in KV7.1 can be tracked by the fluorescence change from the fluorescent probe Alexa-488-maleimide attached to G219C in the S3-S4 loop (referred to as G219C*) (Barro-Soria et al., 2014; Osteen et al., 2010; Osteen et al., 2012). Voltage sensor movement (measured by fluorescence) and gate movement (measured by ionic currents) are then monitored under two-electrode voltage clamp. The KV7.1/G219C* construct by itself or co-expressed with KCNE1 gives voltage-dependent fluorescence changes (Figure 3a). As previously reported, the fluorescence versus voltage (F(V)) curve of KV7.1/G219C* correlates well with the G(V) curve (Figure 3a, left panel), while the F(V) curve of KV7.1/G219C*+KCNE1 is divided into two components (Figure 3a, right panel) (Barro-Soria et al., 2014; Osteen et al., 2010; Osteen et al., 2012). For KV7.1/G219C*+KCNE1, the first fluorescence component (F1) has been suggested to represent the main voltage sensor movement and the second fluorescence component (F2) to be correlated with gate opening (Barro-Soria et al., 2014). We introduce G219C into KV7.1/S225L and KV7.1/F351A. The G(V) curves of both KV7.1/G219C*/S225L and KV7.1/G219C*/F351A are shifted towards more positive voltages compared to the wild-type channel, but the F(V) curves are differentially affected by the two mutations (Figure 3b–c, left panels). For KV7.1/G219C*/S225L, the F(V) curve is shifted to a similar extent as the G(V) curve, while for KV7.1/G219C*/F351A, the F(V) curve is shifted to a considerably smaller extent (Osteen et al., 2010). When these mutants are co-expressed with KCNE1, we observe different effects on the voltage dependence of the two fluorescent components F1 and F2 induced by the mutations. The S225L mutation primarily shifts F1 towards positive voltages so that F1 and F2 of KV7.1/G219C*/S225L+KCNE1 are hardly distinguishable in the F(V) curve (Figure 3b, right panel). In contrast, the F351A mutation primarily shifts F2 towards positive voltages so that F1 and F2 are clearly separated (Figure 3c, right panel). Thus, S225L and F351A seem to shift the G(V) curve of KV7.1+KCNE1 towards positive voltages by interfering with different gating transitions.

Figure 3 with 3 supplements see all
Voltage-clamp fluorometry recordings of wild-type and mutated KV7.1+KCNE1 channels.

(a-c) Representative fluorescence traces and mean F(V)/G(V) curves for KV7.1/G219C* (a), S225L (b), and F351A (c). Left panels without KCNE1 and right panels with KCNE1. The holding voltage is –80 mV, the pre-pulse –120 mV for 2 s (left panels) and –160 mV for 5 s (right panels), and test voltages between –140 and +80 mV for 3 s (left panels) and between –160 and +80 mV for 5 s (right panels) in 20 mV increments. The tail voltage is –80 mV (left panels) and −40 mV (right panels). For KV7.1/G219C*/F351A+KCNE1, the pre-pulse is –120 mV for 3 s, and test voltages ranging between –160 and +100 mV. The bottom of the fit of the KV7.1/G219C*/S225L+KCNE1 F(V) curve (which saturates fairly well at negative voltages) is set to 0 in the normalized F(V) curves in the right panels. The F1 amplitude of KV7.1/G219C*/F351A+KCNE1 is normalized to the F1 amplitude of wild-type. Data as mean ± SEM. n = 4–14. The dashed lines in (b) and (c) denote F(V) (red) and G(V) (black) for wild-type (from a).

https://doi.org/10.7554/eLife.20272.009

Kinetic modeling recapitulates experimental findings

To further explore the different effects of S225L and F351A in the voltage-clamp fluorometry experiments, we use two kinetic models previously developed to reproduce the currents and fluorescence from KV7.1/G219C* (Osteen et al., 2012) and KV7.1/G219C*+KCNE1 channels (Barro-Soria et al., 2014), respectively. The KV7.1/G219C* model is an allosteric model with 10 states (Figure 3—figure supplement 1a), where the horizontal transition is the main S4 movement (which generates the main fluorescence component F1) and the vertical transition is channel opening accompanied by an additional smaller S4 movement (that generates a smaller additional fluorescence component F2) (Osteen et al., 2012; Zaydman et al., 2014). The KV7.1/G219C* model allows for channel opening after only a subset of four S4s are activated, which thereby generates F(V) and G(V) that are close in the voltage dependence (reference (Osteen et al., 2012); and Figure 3—figure supplement 2a). The KV7.1/G219C*+KCNE1 model has 6 states (Figure 3—figure supplement 1b), where the horizontal transition is the main S4 movement (which generates the main fluorescence component F1) and the vertical transition is channel opening accompanied by an additional smaller S4 movement (that generates a smaller additional fluorescence component F2) (Osteen et al., 2012; Zaydman et al., 2014). The KV7.1/G219C*+KCNE1 model only allows for channel opening after all four S4s are activated, which thereby generates F(V) and G(V) that are separated in voltage dependence (reference [Barro-Soria et al., 2014]; and Figure 3—figure supplement 2a).

Using these models, we can reproduce the main features of the fluorescence and currents from KV7.1/G219C*/S225L and KV7.1/G219C*/S225L+KCNE1 by only shifting the main voltage sensor movement by +50 mV in both models (Figure 3—figure supplement 2b), as if the S225L mutation mainly affects the main S4 movement. In the KV7.1 model, shifting the main voltage sensor movement by +50 mV shifts both the G(V) and F(V) curves by +35–40 mV, similar to the effect induced by the S225L mutation in the experimental data. In the KV7.1+KCNE1 model, shifting the main voltage sensor movement by +50 mV results in that the F1 and F2 components overlap in voltage, such that it is hard to distinguish the two components, and that the G(V) is shifted by +10 mV. Both effects are similar to the effects induced by the S225L mutation in the experimental data (cf. Figure 3b).

We can reproduce the main features of the fluorescence and currents from KV7.1/G219C*/F351A and KV7.1/G219C*/F351A+KCNE1 by only shifting the voltage dependence of the opening transition by +140 mV in both models (Figure 3—figure supplement 2c), as if the F351A mutation mainly affects the opening transition. In the KV7.1 model, shifting the opening transition by +140 mV shifts the G(V) by +100 mV whereas the F(V) is shifted less and has a shallower slope, similar to the effects induced by the F351A mutation in the experimental data. In the KV7.1+KCNE1 model, shifting the opening transition by +140 mV results in that the F1 and F2 components are further separated in voltage and that the G(V) is shifted by +100 mV. Both effects are similar to the effects induced by the F351A mutation in the experimental data (cf. Figure 3c).

In summary, our voltage-clamp fluorometry experiments together with kinetic modeling are compatible with a model in which the S225L mutation primarily interferes with the main S4 movement, whereas the F351A mutation interferes with later gating transitions associated with pore opening. One note of caution is that the interpretation of the mutational effects is dependent on the models used for the wild-type channels. Other models for KV7.1 and KV7.1+KCNE1 channels have been proposed (Zaydman et al., 2014; Ruscic et al., 2013), but these have not been as extensively tested or developed as our models. Although other alternative mechanisms for the effects of these mutations are possible, the different impacts of S225L and F351A on the fluorescence versus voltage relationships suggest that these mutations introduce distinct molecular defects.

N-AT enhances the activity of all tested LQTS and LQTS-like mutants

We previously observed that the effect of regular polyunsaturated fatty acids, such as docosahexaenoic acid, on KV7.1 is impaired by co-expression with the KCNE1 subunit (Liin et al., 2015). In contrast, we found that the PUFA analogue N-arachidonoyl taurine (N-AT, structure in Figure 4) retained its ability to activate the KV7.1 channel also in the presence of KCNE1. N-AT activated the wild-type KV7.1+KCNE1 by shifting the G(V) roughly –30 mV (Liin et al., 2015) (Figure 4—figure supplement 1). The magnitude of this N-AT-induced shift is comparable to, but in the opposite direction, to the G(V) shifts observed for several of the LQTS and LQTS-like mutants. We therefore here test the ability of N-AT to enhance the function of the eight KV7.1+KCNE1 mutant channels. Figure 4a–b shows representative effects of 7–70 µM N-AT on KV7.1/S225L+KCNE1. 70 µM N-AT increases current amplitude by a factor of 16 at +20 mV (Figure 4a) and shifts the G(V) curve by about –50 mV (Figure 4b, Supplementary file 3). Steady state of N-AT effects is reached within a few minutes (Figure 4—figure supplement 2). We note a small instantaneous ‘leak’ component in the 70 µM N-AT trace of KV7.1/S225L+KCNE1 (Figure 4a). This leak component in KV7.1/S225L+KCNE1 is observed also in the absence of N-AT, but at more positive voltages (Figure 4—figure supplement 3). We do not observe this leak component in wild-type KV7.1+KCNE1 upon application of N-AT (Figure 4—figure supplement 1a), which suggests that this phenomenon is associated with the S225L mutation. The human ventricular action potential has a duration of about 300–400 ms and a systolic voltage range of about 0 to +40 mV (O'Hara et al., 2011; Piacentino et al., 2003). To test the behaviour of the S225L mutation during shorter stimulating pulses, we apply repetitive 300 ms pulses to +40 mV at a frequency of 1 Hz and at 28°C (37°C was not tolerated by the oocytes). In response to this protocol, the KV7.1/S225L+KCNE1 channel barely opens and thus generates only minor currents (Figure 4c). In contrast, we observe large KV7.1/S225L+KCNE1 currents upon application of 70 µM N-AT (Figure 4c). N-AT also restores the gradual increase in current amplitude during repetitive pulsing seen experimentally (inset in Figure 4c) and in computer simulations (Silva and Rudy, 2005) for the wild-type KV7.1+KCNE1 channel.

Figure 4 with 4 supplements see all
Effect of N-AT on LQTS and LQTS-like mutants.

All these experiments are done in the presence of KCNE1. Structure of N-AT is shown. (ab) Representative effect of 7–70 µM N-AT on current amplitude (a) and G(V) (b) of KV7.1/S225L+KCNE1. Dashed line in (a) denotes 0 µA. (c) Representative currents generated by KV7.1/S225L+KCNE1 during pulsing at 1 Hz and +28°C in control solution (black) and after the cell had been bathed continuously in 70 µM N-AT (light to dark green, # denotes sweep order). Inset: corresponding currents from wild-type KV7.1+KCNE1 scaled similarly as KV7.1/S225L+KCNE1. Light grey trace denotes sweep #1, grey trace denotes sweep #2, and dark grey trace denotes sweep #20. (d) Summary of V50 for LQTS and LQTS-like mutants before and after 70 µM N-AT application. Dashed line denotes V50 for wild-type KV7.1+KCNE1. (e–f) Summary of ΔV50 (e) and ΔΔGo (f) for LQTS and LQTS-like mutants induced by 70 µM N-AT. # denotes an approximation. Dashed lines denote corresponding ΔV50 and ΔΔGo induced by 70 µM N-AT for wild-type KV7.1+KCNE1. The statistics in (f) represent one-way ANOVA with Dunnett’s Multiple Comparison Test to compare the N-AT-induced change in ΔΔGo of mutants to N-AT-induced change in ΔΔGo of wild-type KV7.1+KCNE1; *p≤0.05. Only significant differences shown in (f), other comparisons have p>0.05. (g) Estimate of the ability of 70 µM N-AT to restore LQTS and LQTS-like mutant current amplitude at +40 mV. The mean N-AT induced increase in current amplitude for each mutant (from Figure 4—figure supplement 4b) is multiplied with the control amplitude for each mutant (from Figure 1—figure supplement 4d). Not determined (nd). Data as mean ± SEM. n = 5–12. Dashed line denotes relative wild-type KV7.1+KCNE1 current amplitude in control solution (i.e. without N-AT).

https://doi.org/10.7554/eLife.20272.013

Further testing of N-AT show that 70 µM N-AT shifts the G(V) curve of all tested mutants by 30–50 mV towards more negative voltages (Figure 4d–e, Supplementary file 3). The G(V) curve of wild-type KV7.1+KCNE1 is shifted by –27.0 ± 2.5 mV (Liin et al., 2015). Thus, 70 µM N-AT completely corrects the positive G(V) shifts induced by the mutations so that in the presence of N-AT the G(V) is similar to or shifted negative compared to the G(V) of the wild-type KV7.1+KCNE1 channel (Figure 4d, F351A homozygous expression was not included in this analysis because of the very shifted G(V) curve of this mutant). The G(V) of mutants is shifted about equally by N-AT for homozygous and heterozygous expression (Figure 4e). The slope of the G(V) curve varies slightly (10.4 to 16.3) among the mutants (Supplementary file 3). To correct for this difference in slope and to better compare the functional effect of N-AT-induced G(V) shifts on the different mutants, we also calculate the change in Gibbs free energy for channel opening (ΔΔGo) that 70 µM N-AT induces. 70 µM N-AT reduces the energy required to open the channel by 5.3–9.0 kJ/mol depending on mutant (4.9 ± 0.7 kJ/mol (n = 5) for wild-type) (Figure 4f). To estimate the functional effect of N-AT on the KV7.1+KCNE1 current amplitude of each mutant, we calculate the ratio of the current amplitude at the end of the 5 s test pulse before and after application of N-AT at +20 and +40 mV. The 5 s voltage pulse to +20 mV (or + 40 mV) at room temperature was chosen to make the KV7.1+KCNE1 channel activate to a similar extent as during a ventricular action potential (300–400 ms) at body temperature (note that KV7.1+KCNE1 channels have a relatively high Q10 of around 5–7.5 [Busch and Lang, 1993; Seebohm et al., 2001]). 70 µM N-AT increases the current amplitude of all mutants at these voltages (Figure 4—figure supplement 4a–b, Supplementary file 3). As expected, current amplitude is most increased for those mutants that have the most shifted G(V) curve towards more positive voltages (e.g. V215M and S225L). This is because these mutants are still at the foot of their G(V) curve at +20 and +40 mV and a N-AT-induced shift towards more negative voltages results in a relatively larger increase in the current amplitude. By multiplying these relative N-AT-induced increases in current amplitude with the relative current amplitude of each mutant (compared to wild-type KV7.1+KCNE1 channels, from Figure 1—figure supplement 4c–d), we observe that 70 µM N-AT compensates fairly well (or overcompensates) for the mutation-induced reduction in current amplitude (Figure 4g, Figure 4—figure supplement 4c). Moreover, for all mutant and wild-type KV7.1+KCNE1 channels, 70 µM N-AT speeds up the opening kinetics at +40 mV by a factor of 1.3–2.5 (Supplementary file 3). 70 µM N-AT also slows down the closing kinetics for most mutants and wild-type KV7.1+KCNE1 (Supplementary file 3). For F351A heterozygous expression and R583C homozygous expression, 70 µM N-AT restores the closing kinetics so that the closing kinetics is not statistically different (p>0.05) from wild-type KV7.1+KCNE1 closing kinetics (737 ± 62 ms and 833 ± 74 ms, respectively compared to 967 ± 47 ms for wild-type). In the presence of KCNE1, channels made with F193L heterozygous expression, L251A homozygous expression, and R583C heterozygous expression have wild-type like closing kinetics already before application of N-AT.

N-AT affects both S4 movement and gate opening in mutants

We next use voltage clamp fluorometry on KV7.1/G219C*/S225L+KCNE1 and KV7.1/G219C*/F351A+KCNE1 to explore the mechanism by which N-AT enhances the activity of two mechanistically different mutants. Surprisingly, N-AT caused a dramatic decrease in the fluorescence from Alexa488-labeled KV7.1/G219C*+KCNE1 channels (Figure 5—figure supplement 1a). In contrast, N-AT did not decrease the fluorescence from Alexa488-labeled KV7.1/G219C* channels nor did high concentrations of taurine decrease the fluorescence from unbound Alexa488 (even up to concentrations of 0.5 M taurine; Figure 5—figure supplement 1b), suggesting that N-AT is not a collisional quencher of Alexa488. The mechanism of the N-AT-induced decrease of fluorescence from Alexa488-labeled KV7.1/G219C*+KCNE1 channels is not clear, but could be due to N-AT inducing a conformational change in KCNE1 or KV7.1 that brings a quenching residue close to Alexa488.

Due to the dramatic decrease in the fluorescence signal from Alexa488-labeled KV7.1/G219C*+KCNE1 channels, we have to normalize the F(V) curves obtained in N-AT to the amplitude of the F(V) in control solutions. With this normalization, voltage clamp fluorometry experiments on KV7.1/G219C*/S225L+KCNE1 indicate that N-AT shifts both the voltage dependence of the first part (which represents F1) and the second part (which represents F2) of the F(V) curve towards more negative voltages (Figure 5—figure supplement 1c). However, due to the not completely saturating F(V) for KV7.1/G219C*/F351A+KCNE1, we are unable to reliably normalize the F(V) curves in the presence of N-AT to the control F(V) curves. We instead explore the effect of N-AT on the kinetics of the two fluorescence components: F1, which is seen as a fast fluorescence change at negative voltages, and F2, which is seen as a slow fluorescence change on top of the F1 component at positive voltages (Barro-Soria et al., 2014). F1 correlates with the measured gating currents in KV7.1+KCNE1 channels (and the initial delay in the KV7.1+KCNE1 ionic currents), whereas F2 correlates with the opening of KV7.1+KCNE1 channels (Barro-Soria et al., 2014). For both mutants, 70 µM N-AT speeds up F1 kinetics (Figure 5a,d, measured at –40 mV where virtually no channels open and the fluorescence is mainly composed of F1). Numeric values for N-AT effects on channel kinetics are summarized in Figure 5f. Moreover, N-AT accelerates the channel opening kinetics (Figure 5b,e) and both the F1 and F2 fluorescence components at +80 mV for KV7.1/G219C*/S225L+KCNE1 (Figure 5f). The change in the F2 component is probably larger than what the fits of a double-exponential function suggest, because the slow part of the fluorescence, mainly F2, overlay nicely on the currents in both the presence and absence of 70 µM N-AT (Figure 5c, upper panel). As a control, we show that the fluorescence in N-AT does not, however, overlay the currents in control solutions and vice versa (Figure 5c, middle and lower panel). For KV7.1/G219C*/F351A+KCNE1, the G(V) curve and the F2 component are so shifted towards depolarizing voltages that we cannot reliably quantify the F2 component in our fluorescence traces. 70 µM N-AT does, however, speed up KV7.1/G219C*/F351A+KCNE1 current kinetics (Figure 5e), which suggests that N-AT also speeds up F2 in KV7.1/G219C*/F351A+KCNE1. Altogether, these results suggest that N-AT accelerates both conformational changes during the main gating charge movement and channel opening.

Figure 5 with 1 supplement see all
Effect of 70 µM N-AT on S4 movement and gate opening in S225L and F351A mutants.

(a–c) Representative example of the effect of 70 µM N-AT on F1 kinetics (a), current opening kinetics (b), and F2 kinetics (c) in KV7.1/G219C*/S225L+KCNE1. Control fluorescence (red) and current (black). N-AT fluorescence (magenta) and current (green). Top in (c) shows an overlay of the later part of the fluorescence (after most of F1 has occurred) and the later part of the currents (after the initial delay) before and after application of N-AT. Middle and lower (c) show that there is not a great overlap of the fluorescence in the presence of N-AT and the current in control solution (middle) or the fluorescence in control solution and the current in the presence of N-AT (lower). (d–e) Representative example of effect of 70 µM N-AT on F1 kinetics (d) and current opening kinetics (e) in KV7.1/G219C*/F351A+KCNE1. Same colouring as in (a–b). Dashed line in (b) and (e) denotes 0 µA. Fluorescence traces and all traces in (c) have been normalized to better allow temporal comparison. (f) Summary of the effect of 70 µM N-AT on the kinetic parameters of KV7.1/G219C*/S225L+KCNE1 and KV7.1/G219C*/F351A+KCNE1. Kinetics of the fast (F1) and slow (F2) fluorescence components were deduced from a double-exponential function fitted to the fluorescence traces. The kinetics of currents were deduced from a single-exponential function fitted to current traces. Ratios of time constants (τN-ATCtrl) were calculated pair-wise (control compared to N-AT) in each oocyte and analysed using two-tailed one sample t-test where ratios were compared with a hypothetical value of 1. Data as mean ± SEM. n = 4 (3for fluorescence kinetics for KV7.1/G219C*/F351A+KCNE1). *p<0.05; **p<0.01. nd = not determined.

https://doi.org/10.7554/eLife.20272.018

Discussion

We show that all studied LQTS and LQTS-like mutations i) shift the G(V) of KV7.1+KCNE1 towards more positive voltages, and/or ii) accelerate KV7.1+KCNE1 closing. This suggests that at least part of the mechanism underlying the reduced ability of these mutants to generate K+ currents is by altering these biophysical properties of the KV7.1+KCNE1 channel. Using voltage clamp fluorometry in combination with kinetic modeling, we further suggest that these altered biophysical properties in mutants may be caused by interference with different gating transitions. Our experimental data and kinetic modeling are consistent with a model in which KV7.1/S225L primarily causes the reduced channel function by altering the main voltage sensor movement, while KV7.1/F351A alters later gating transitions associated with pore opening. The different effects of S225L and F351A on the fluorescence versus voltage relationships in KV7.1/G219C* and KV7.1/G219C*+KCNE1 suggest that these mutations cause channel dysfunction via different molecular mechanisms. Note that we used the LQTS-like F351A mutant, because the LQTS mutant F351S did not generate any currents (Figure 1—figure supplement 1). However, during the review process of this manuscript a new LQTS mutation, F351L, was found (Vyas et al., 2016). The current and fluorescence of this LQTS mutant is very similar to the current and fluorescence of F351A (Figure 3—figure supplement 3), suggesting that our conclusions on the LQTS-like F351A is also relevant for the LQTS mutant F351L.

One of the mutations, F193L, has only minor effects on the biophysical properties of KV7.1+KCNE1. This mutant was previously reported to have reduced current amplitude compared to the wild-type KV7.1+KCNE1 channel and a mild clinical phenotype (Yamaguchi et al., 2003). The F193L mutation may therefore cause loss of function by faster deactivation kinetics and lower current density. Heterozygous expression of mutated subunits and wild-type subunits in equal molar ratios results in general in a milder biophysical phenotype (more close to the wild-type phenotype). This is in line with a milder clinical phenotype generally reported for heterozygous carriers of LQTS mutations compared to individuals with homozygous genotypes (Priori et al., 1998; Jackson et al., 2014; Zhang et al., 2008). Moreover, for different mutations different biophysical effects of the mutations could be dominant or recessive: For S225L and L251A, heterozygous expression in the presence of KCNE1 partially or completely restores wild-type like V50, whereas heterozygous expression does not improve closing kinetics compared to homozygous expression. For KCNE1/K70N and KCNE1/S74L, co-expression with wild-type KCNE1 subunits also restores wild-type like V50, whereas wild-type like closing kinetics is only partially restored. In contrast, for KV7.1/R583C, heterozygous expression restores wild-type like closing kinetics, but not wild-type like V50. However, because of uncertainties regarding the stoichiometry of mutant to wild-type subunits in assembled KV7.1+KCNE1 channels (as mentioned in the Results section), further studies will be required to understand the mechanisms underlying these apparent dominant or recessive effects and to evaluate possible physiological impact of these effects.

Our results show that all tested mutants respond to N-AT. This is in contrast to previously reported KV7 channel activators on disease-causing KV7 mutants, for which mutants show markedly different sensitivity (Seebohm et al., 2003; Xiong et al., 2007; Leitner et al., 2012). 70 µM N-AT shifts the G(V) curve of the wild-type KV7.1+KCNE1 channel and of all LQTS and LQTS-like mutants by approximately (–50)–(–30) mV, accelerates channel opening and slows down channel closing. In the presence of 70 µM N-AT, the V50 of all LQTS and LQTS-like mutants are similar to or more negative than V50 for the wild-type KV7.1+KCNE1 channel. For most mutants, 70 µM N-AT overcompensates for the shift in G(V) and reduction in current amplitude caused by the mutations, indicating that a lower N-AT concentration or a less potent N-AT analogue could be used to restore wild-type like G(V) and current amplitudes. Moreover, KV7.1+KCNE1 opening and closing kinetics are partially or completely restored by N-AT. Also, although the disease aetiology of the F193L mutant is likely mainly reduced channel expression, the N-AT induced augmentation caused by a shift in G(V) and increased currents may at least in part overcome the reduction in currents caused by the reduced channel expression. This general ability of N-AT to, at least partly, compensate for the reduced function of mutants with mutations in different parts of the KV7.1+KCNE1 channel complex and with seemingly different molecular defects, as long as a population of these mutant channels reaches the plasma membrane, suggests that N-AT is an interesting model compound for development of future anti-arrhythmics to treat LQTS caused by diverse KV7.1+KCNE1 mutations.

Defective trafficking of mutant KV11.1 ion channels is a common cause of LQTS type 2. About 80-90% of LQTS type 2-associated hERG mutants are estimated to suffer from defective trafficking (Anderson et al., 2014; Sanguinetti, 2010). The corresponding number for LQTS-associated KV7.1 and KCNE1 mutants is not known. Previous studies identify both trafficking defective and trafficking competent KV7.1 and KCNE1 mutants, e.g. (Anderson et al., 2014; Sanguinetti, 2010). We are mainly interested in understanding the mechanism that underlies abnormal gating of KV7.1 and KCNE1 mutants. To avoid mutants with severe trafficking defects, we therefore selected mutants that have previously been shown to localize abundantly enough to the cell membrane to generate detectable K+ currents. Several of the selected mutants have been shown to traffic well in mammalian systems (KV7.1/V215M and KCNE1/S74L [Eldstrom et al., 2010; Harmer et al., 2010]) or generate clearly detectable currents in mammalian cells (KV7.1/R583C [Yang et al., 2002]). Our Xenopus oocyte experiments that compare mutant current amplitudes with wild-type current amplitudes (Figure 1—figure supplement 4) suggest that the reduced ability of the selected mutants to generate currents in Xenopus oocytes may largely be explained by the shifted G(V) of mutants. Trafficking defects could be disguised in Xenopus oocytes that are cultured at low temperatures that may rescue some trafficking defects (Anderson et al., 2014; Delisle et al., 2004). These current amplitude experiments should therefore be interpreted with caution until trafficking of specific KV7.1 and KCNE1 LQTS mutants in mammalian systems has been explored. Previous studies show that membrane expression of trafficking-defect channel mutants (e.g. for KV11.1 and CFTR) can be pharmacologically rescued using compounds that are suggested to stabilize channel conformation during folding and trafficking (Anderson et al., 2014; Delisle et al., 2004; Sato et al., 1996). However, rescue of membrane expression may only partially compensate for mutation-induced loss of function, if these mutants also suffer from defective gating (Perry et al., 2016). Our proposed N-AT model for pharmacological correction of ‘G(V)’ LQTS mutants could therefore potentially complement pharmacological correction of trafficking-defect LQTS mutants to improve the outcome of patients suffering from LQTS.

We previously suggested that polyunsaturated fatty acids and their analogues (such as N-AT) attract the voltage sensor S4 in KV7.1 by an electrostatic mechanism and thereby shift the G(V) towards more negative voltages and speed up channel opening (Liin et al., 2015). We therefore initially hypothesized that N-AT only would restore the function of those LQTS mutations with altered S4 movement. We were pleasantly surprised when N-AT seems to be able to restore the function of many LQTS and LQTS-like mutants, with diverse mutational defects (such as S225L and F351A). Using voltage clamp fluorometry, we have previously shown that both the main gating charge movement and the gate opening of KV7.1+KCNE1 channels are accompanied by fluorescence signals from fluorophores attached to S4 (Barro-Soria et al., 2014). This suggests that S4 moves both during the main gating charge movement and during the subsequent channel opening in KV7.1+KCNE1 channels (Barro-Soria et al., 2014), which is similar to observations in Shaker KV channels (Börjesson and Elinder, 2011; Pathak et al., 2005; Phillips and Swartz, 2010). Therefore, N-AT could affect both the main gating charge movement and gate opening by acting on the S4 voltage sensor, as has been shown for hanatoxin which targets the voltage-sensing domain in the Shaker KV channel (Milescu et al., 2013). This hypothesis is supported by our voltage-clamp fluorometry experiments using KV7.1/S225L and KV7.1/F351A in which N-AT accelerates the fluorescence components associated with both the main S4 movement (F1) and gate opening (F2), as well as accelerates the kinetics of channel opening. This proposed mechanism would explain why N-AT can restore the function of mutations that mainly target either the main S4 movement or gate opening. However, the dramatic decrease in the fluorescence signal caused by N-AT makes it hard for us to completely determine the effect of N-AT on the F(V) of mutants. Therefore, the complete mechanism of N-AT in the different mutations is not clear.

Future studies are required to assess the clinical utility of PUFA analogues in cardiomyocytes and animal models. We see channel specificity of PUFA analogues as one major challenge and recognize the need to improve PUFA analogue affinity to KV7.1+KCNE1 to reduce required therapeutic concentrations and minimize potential adverse effects. Despite these challenges, our data show that the magnitude of the N-AT-induced voltage shifts are in a similar range as the shifts induced by several LQTS mutations, thereby serving as proof of concept that this PUFA analogue, at least partly, restores channel function in diverse LQTS and LQTS-like mutants.

Materials and methods

Experiments were approved by The Linköping Animal Ethics Committee at Linköping University and The Animal Experiments Inspectorate under the Danish Ministry of Food, Agriculture and Fisheries (University of Copenhagen).

Experiments on Xenopus laevis oocytes

Molecular biology

Request a detailed protocol

Expression plasmids human KV7.1 (GenBank Acc.No. NM_000218) in pXOOM and KCNE1 (NM_000219) in pGEM have been previously described (Jespersen et al., 2002; Schmitt et al., 2007). LQTS and LQTS-like point mutations and G219C were introduced into KV7.1 or KCNE1 using site-directed mutagenesis (QuikChange Stratagene, CA). All newly generated constructs were sequenced to ensure integrity (Genewiz, NJ). cRNA was prepared from linearized DNA using the T7 mMessage mMachine transcription kit (Ambion, TX). RNA quality was checked by gel electrophoresis, and RNA concentrations were quantified by UV spectroscopy.

Two-electrode voltage-clamp electrophysiology

Request a detailed protocol

Xenopus laevis oocytes (from EcoCyte Bioscience, TX, or prepared in house) were isolated and maintained as previously described (Börjesson et al., 2010). 50 nl cRNA (~50 ng KV7.1 for KV7.1-only expression, 25 ng KV7.1 + 8 ng KCNE1 for homozygous expression, or 12.5 ng KV7.1wt + 12.5 ng KV7.1mut + 8 ng KCNE1wt alternatively 25 ng KV7.1wt + 4 ng KCNE1wt + 4 ng KCNE1mut for heterozygous expression) was injected into each oocyte. Currents were measured at room temperature 2–5 days after injection with the two-electrode voltage-clamp technique (CA-1B amplifier, Dagan, MN). For the current amplitude experiments presented in Figure 1—figure supplement 4, the current amplitude of mutants were normalized to the current amplitude of wild-type KV7.1+KCNE1 expressed in the same batch of oocytes and incubated under identical conditions for the same time period. Currents were sampled at 1–3.3 kHz, filtered at 500 Hz, and not leakage corrected. The control solution contained (in mM): 88 NaCl, 1 KCl, 15 HEPES, 0.4 CaCl2, and 0.8 MgCl2 (pH adjusted to 7.4 using NaOH). The holding voltage was generally set to –80 mV. Activation curves were generally elicited by stepping to test voltages between –110 and +60 mV (3–5 s durations and 10 mV increments) followed by a tail voltage of –20 mV. Voltage clamp fluorometry experiments were performed as previously described on oocytes labeled for 30 min with 100 µM Alexa-488-maleimide (Molecular Probes) at 4°C (Barro-Soria et al., 2014; Osteen et al., 2010; Osteen et al., 2012). For voltage clamp fluorometry experiments on KV7.1/G219C*, the holding voltage was –80 mV, the pre-pulse –120 mV for 2 s, and test voltages ranging between –140 and +80 mV for 3 s in 20 mV increments. The tail voltage was –80 mV. For KV7.1/G219C*/KCNE1, the holding voltage was –80 mV, the pre-pulse –160 mV for 5 s, and test voltages ranging between –160 and +80 mV for 5 s in 20 mV increments. The tail voltage was –40 mV. N-arachidonoyl taurine was purchased from Cayman Chemical (MI, USA) and stored, diluted and applied to the oocyte chamber as previously described (Liin et al., 2015). Control solution was added to the bath using a gravity-driven perfusion system.

Electrophysiological analysis

Request a detailed protocol

To quantify effects on the G(V), tail currents (measured shortly after initiation of tail voltage) were plotted against the pre-pulse (test) voltage. The following Boltzmann relation was fitted to the data

(1) GK(V)=Gmax/(1+exp((V50V)/s)),

where V50 is the midpoint (i.e. the voltage at which the conductance is half the maximal conductance estimated from the fit) and s the slope factor (shared slope for control and N-AT curves within the same cell). In figures showing Itail vs voltage, the curves are normalized to the fitted Gmax. The same single Boltzmann relation was used to fit the F(V) from voltage clamp fluorometry recordings of KV7.1 without KCNE1 co-expression, where fluorescence at the end of the test pulse was plotted versus the test voltage (Barro-Soria et al., 2014). For voltage-clamp fluorometry recordings of KV7.1 with KCNE1 co-expression (and F351A without KCNE1), a double Boltzmann relation was used (Barro-Soria et al., 2014). For experiments where conductance or fluorescence did not clearly show signs of saturation in the experimental voltage range, these fits should be considered as an approximation. To estimate the effect of N-AT on Gibbs free energy, the following relation was used:

(2) ΔΔGo=zΔV50F,

Where z is the gating charge of each channel deduced from the slope of the Boltzmann fits according to z=25/s, ΔV50is the N-AT induced shift in the V50 values from the Boltzmann fits, and F is Faraday’s constant (Li-Smerin and Swartz, 2001; Monks et al., 1999; DeCaen et al., 2008). This analysis assumes a two-state model and tends to underestimate the z (Chowdhury and Chanda, 2012). The calculated ΔΔGo should therefore be seen as an approximation. For opening and closing kinetics, T50,open was defined as the time it takes to reach 50% of the current in the end of a 3 s (5 s for KCNE1 co-expression) long test pulse to +40 mV. T50,close was defined as the time it takes to reduce the amplitude (= instantaneous tail current – steady state tail current) of the tail current by 50% when stepping to a tail pulse to –20 for 5 s. To analyze the effect of N-AT on fluorescence and current kinetics, single or double exponentials were fitted to the fluorescence or current traces. The ratios of time constants before and after application of N-AT were then calculated.

Modeling

Request a detailed protocol

Fluorescence and currents from the KV7.1+KCNE1 models were simulated using Berkeley Madonna (Berkeley, CA).

Statistics

Average values are expressed as mean ± SEM. Mutant parameters (e.g. V50 and ΔΔGo) were compared to wild-type parameters using one-way ANOVA with Dunnett’s Multiple Comparison Test. Comparison of homozygous and heterozygous expression was done using one-way ANOVA with pair-wise Bonferroni’s Test. The effects of N-AT on fluorescence and current kinetics were analysed using two-tailed one sample t-test where ratios were compared with a hypothetical value of 1. p<0.05 is considered as statistically significant.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
    Mechanisms of I(Ks) suppression in LQT1 mutants
    1. L Bianchi
    2. SG Priori
    3. C Napolitano
    4. KA Surewicz
    5. AT Dennis
    6. M Memmi
    7. PJ Schwartz
    8. AM Brown
    (2000)
    American Journal of Physiology Heart and Circulatory Physiology 279:H3003–3011.
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
    Dual-function vector for protein expression in both mammalian cells and Xenopus laevis oocytes
    1. T Jespersen
    2. M Grunnet
    3. K Angelo
    4. DA Klaerke
    5. SP Olesen
    (2002)
    BioTechniques 32:536–538.
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
    HERG1 channelopathies
    1. MC Sanguinetti
    (2010)
    Pflügers Archiv - European Journal of Physiology 460:265–276.
    https://doi.org/10.1007/s00424-009-0758-8
  46. 46
  47. 47
  48. 48
  49. 49
  50. 50
  51. 51
  52. 52
  53. 53
  54. 54
  55. 55
  56. 56
  57. 57
  58. 58
  59. 59

Decision letter

  1. Kenton J Swartz
    Reviewing Editor; National Institutes of Health, United States

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "Fatty Acid Analogue N-Arachidonoyl Taurine Restores Function of IKs Channels with Diverse Long QT Mutations" for consideration by eLife. Your article has been favorably evaluated by Richard Aldrich (Senior Editor) and three reviewers, one of whom, Kenton J Swartz (Reviewer #1), is a member of our Board of Reviewing Editors. The following individual involved in review of your submission has agreed to reveal his identity: Christopher J Lingle (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

This paper shows that LQTS deficits resulting from diverse loss-of-function mutations in Kv7.1+KCNE1 channels can potentially be rescued by the polyunsaturated fatty acid analogue, N-arachidonyl taurine (N-AT). To make this point, the authors focus on two mutations, S225L and F351A, drawing the conclusion that the S225L mutation produces a rightward shift in voltage-sensor activation while the F351A mutation largely results in less effective channel opening once voltage-sensors are activated. The results are nicely supported by a comparison of fluorescence-voltage and conductance-voltage curves for each construct with for Kv7.1 expressed alone or when coexpressed with KCNE1. To validate the idea that each mutation does reduce activation in mechanistically distinct ways, the authors use previously developed models of Kv7.1 activation +/- KCNE1 to show that a 50 mV rightward shift in VSD activation nicely accounts for both the FV and GV shifts for the S225L mutation, while the F351A mutations can be approximated by a +140 mV shift in the voltage-dependence of channel opening. Although this approach to validate the idea that the two mutations inhibit Kv7.1 activation by distinct mechanism might be viewed as somewhat dependent on the validity of the activation models (which remains a bit uncertain), the fact that changes in largely a single parameter accounts for the FV and GV relationships both in the absence and presence of KCNE1 is satisfying.

The authors make the argument that, with appropriate development of compounds of greater specificity and affinity, this may be a plausible strategy for treatment of various LQTS diseases. A number of arguments can be made against the idea that this approach could ever work, such as:

1) Deficits in expression may underlie the effects of many LQTS mutations;

2) Fatty acid analogues are likely to have effects on many targets;

3) Concentrations of any potentially effective compound may have to be adjusted for any given LQTS mutation.

However, overall, we felt that the authors did a good job of discussing many of these challenges and providing perspective about what would have to be done before this approach might be of benefit. As a proof of principle that an activator of Kv7.1/KCNE channels can rescue function, irrespective of the nature of the functional deficit, this paper accomplishes that goal. The idea that an activator may qualitatively restore a channel's function irrespective of the origins of the original gating deficit is intriguing.

Essential revisions:

More data needs to be included to show the effects of the PUFA on fluorescence signals. The text states that the compound greatly diminishes fluorescence but otherwise little is shown or explained. Do the authors understand the mechanism? Is N-AT a collisional quencher? Is there any voltage-dependence to quenching? If not, why not show F-V relations unscaled and then scaled so the reader can appreciate what is going on? Given the large effects of N-AT on fluorescence, should we be concerned about interpreting small changes in kinetics? Were the traces in Figure 5 normalized to allow comparison of the temporal properties? Although the authors would like to conclude that N-AT has generalized effects that affect both the VSD activation and channel opening, it seems like it could still be the case that effects on channel opening are primary. As such, the FV and GV data might help put this issue on firmer foundation.

[Editors’ note: a previous version of this study was rejected after peer review, but the authors submitted for reconsideration. The first decision letter after peer review is shown below.]

Thank you for submitting your work entitled "Fatty Acid Analogue N-Arachidonoyl Taurine Restores Function of IKs Channels with Diverse Long QT Mutations" for consideration by eLife. Your article has been favorably evaluated by Richard Aldrich (Senior Editor) and three reviewers, one of whom, Kenton J Swartz (Reviewer #1), is a member of our Board of Reviewing Editors. The following individual involved in review of your submission has agreed to reveal his identity: Christopher J Lingle (Reviewer #3).

Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered further for publication in eLife.

All three reviewers found your work to be interesting because it attempts to understand disease causing mutations at a mechanistic level and provides proof of concept evidence that PFU could potentially be used to correct the deficit. Although the reviewers were somewhat divergent in their initial enthusiasm for the manuscript, after discussion there was consensus that the study would need substantial work to be appropriate for eLife. All three reviewers offer specific suggestions, which you should consider carefully in going forward with publication. These include 1) clarification of the mechanism by which the two studied mutants affect gating, 2) examination of the impact of mutations on surface levels of the KCNQ channel in order to assess whether restoring the gating properties with PUFs would be sufficient to correct the deficits produced by the mutants, and 3) clarification of the impact of PUFs on other ion channels, cardiac action potentials and EKGs. The overarching concern was that insufficient evidence was presented to support the potential therapeutic utility of PUFs. Our policy is to reject manuscripts requiring the amount of effort that we feel yours would require, but we hope that the reviews will be helpful in revising your manuscript for publication elsewhere. If you feel that you can address the concerns of the reviewers, we would be willing to reconsider the manuscript as a new submission.

Reviewer #1:

The authors investigate a series of mutations in the KCNQ channel that cause long QT syndrome in humans and use voltage-clamp fluorimetry with a fluorescent probes attached to the voltage sensors to track conformational changes while measuring channel activation. They see that the mutants alter the voltage-activation relations to varying degrees, but typically shifting activation to positive voltages, slowing activation and speeding deactivation. From comparisons of the F-V and G-V relations in the absence and presence of KCNE, the authors propose that the one mutant located in S4 alters voltage-sensor activation while another located in S6 alters pore opening. They then show that a polyunsaturated fatty (PFU) acid can correct the effects of the mutants, and do proof of concept experiments in iPSC-derived cardiomyocytes and guinea pig hears where a KCNQ inhibitor is used to induce LQT and show that the PUF can restore normal LQT. This is an interesting study because it attempts to understand disease causing mutations at a mechanistic level and provides proof of concept evidence that PFU could potentially be used to correct the deficit. Although the authors' interpretations are plausible, the data are not always so clean and I can imagine more complex mechanisms may be involved.

1) The discussion of the combined F-V and G-V data in Figure 3 in the subsection “KV7.1+KCNE1 channel dysfunction is caused by different underlying mechanisms” needs some work and it is not clear to me that S225L affects voltage-sensor activation and F351A affects pore opening. The data to me seem much more complex than the authors acknowledge. I would like to see a clearer presentation of the evidence supporting the key conclusions and a more nuanced discussion of the underlying complexity. The authors may be correct, but I am not sure we understand why the F-V and G-V overlap for KCNQ alone and what KCNE is doing to those relations, which makes interpretation of what is seen with the mutants less than straightforward.

2) What is the logic of using N-AT instead of other PUFs the authors have previously studied? Given those studies, I would imagine N-AT might have substantial effects on other Kv channels, and possibly even Nav and Cav channels. What makes the authors think that N-AT will have any selectivity for KCNQ? This feels like a stretch to imagine using these in humans at high μM concentrations.

Reviewer #2:

This paper investigates a selection of Long QT syndrome Type 1 and Type 5 mutations that shift the voltage-dependence of channel opening. Like others, they have found that such mutants have altered voltage-dependence of channel opening, and changed closing kinetics, that can be partially normalized by "heterozygous" expression. N-AT moves activation back to more negative potentials in the LQT mutants, as the authors have previously shown for IKs, and N-AT accelerates all components of fluorescence movement from the mutant channels incorporating G219C. In addition, it is shown that N-AT can affect action potential duration in iPSC-derived cardiomyocytes and intact guinea-pig hearts, although these data are less convincing. The significance of this work is that N-AT is an activator of both wild type Kv7.1 +KCNE1 and LQTS mutants. It does not seem so surprising, or indeed particularly significant that N-AT affects LQT mutants in the same way as WT channels. The authors have already reported that PUFA analogues shift the activation gating of Kv7.1+KCNE1 channels, so it is to be expected that the LQT mutants have their activation affected as well, as simple missense mutants. Given that, most of the biophysics presented is logical and seems to suggest that N-AT acts non-specifically on the channels (unlike their previous suggestion that WT channels are affected by an "electrostatic mechanism" affecting the VSD). For most in the field, interest in this work is lessened as it is generally accepted that about 2/3 of LQT phenotypes come about as a result of lowered expression at the cell surface from misfolding and other events during synthesis and trafficking.

1) Expression in oocytes bypasses these problems due to the low incubation temperatures, so it is unclear whether these gating effects are relevant in the greater scheme of increasing overall current density at the mammalian cell surface. Do the authors know what the expression and biophysical characteristics are for these LQT mutants at physiological temperature compared with wild type to understand how N-AT may alter the abnormal functioning of these mutants?

2) There is a little bit of mixture of models in this study. Most experiments are carried out at room temperature in oocytes, but sometimes CHO cells are used for hKv11.1, derived cardiomyocytes and intact hearts are used for action potential and QT studies. Some data that could tie together studies in the different models would support the potential translational importance of the idea.

3) The data from iPSC-derived cardiomyocytes and intact guinea-pig hearts are unconvincing. In the intact hearts it looks like other changes affect the early action potential to shorten it in the presence of N-AT. Changes in the isoelectric region of the EKG suggest a dispersion of action potential effects by N-AT. In intact hearts, the slope of phase 3 repolarization in the HMR+N-AT trace is shallower than in control or with HMR1556 alone. This suggests less net outward current when N-AT is added, the opposite of what is contended. Perhaps this is because N-AT blocks many other cardiac currents.

4) The authors suggest that PUFAs could provide a treatment option for LQTS, but there is a large literature on the effects of PUFAs on humans, in animal models and also on individual ion currents in heart showing that almost all ion currents, inward and outward, including IKr, are decreased by PUFAs. The well-established exception is IKs, but the overall effect on the QT interval duration could be pro- or antiarrhythmic depending on the physiology of the tissue and the pathology affecting the heart.

Reviewer #3:

The main point of this paper is that LQTS deficits regulating from diverse loss-of-function mutations in Kv7.1+KCNME1 channels can potentially be rescued by the polyunsaturated fatty acid analogue, N-arachidonyl taurine (N-AT). This is an exciting and interesting conclusion. The paper is clearly written and the results are straight forwards and support the conclusions. A particularly compelling aspect of the paper is that the authors show that N-AT can rescue function in two mechanistically distinct types of LQTS mutants, those likely to affect voltage sensor function and those affecting channel opening equilibria. This provides some assurance that this sort of pharmacological strategy might have general applicability for a number of different categories of LQTS mutations.

1) One topic that probably requires some additional clarification concerns the presentation of GVs and how the V50's and z values were generated. For many of the mutant GV curves, a saturating level of activation is never obtained, but the plotted GV curves are simply normalized to the maximum level of conductance. I think for some (many) readers this may give a misleading impression of what the relative conductances as a function of voltage may be among different constructs (Figures 1C, 2A, and others) or in response to N-AT application (e.g., Figure 4B), since the assumption may be made that the voltage of the half maximal conductance after normalization corresponds to the true V50. Although sometimes it is natural to not trust the extrapolated gmax from a Boltzmann fit, plotting the GVs normalized to the fitted Gmax may be a better proxy for the true channel behavior. In such cases, the observed 0.5 level of conductance directly informs a reader of the approximate Vh, and it also informs the reader regarding changes in the slope, which may otherwise be obscured. In Figure 4B, the normalization used by the authors probably tends to minimize the effects at the lower concentrations, while normalizing to the fitted gmax might avoid this. One might also ask in regard to Figure 4B, since the control and N-AT tails are all measured in the same patches, why not use the maximal conductance with 70 μM N-AT as the basis for normalization? My concern regarding the GV displays and the estimates of Vh and z become most critical in regards to the DDG estimates (I think, as the process is described in the methods, this was done appropriately, but some clarifications about this might help).

https://doi.org/10.7554/eLife.20272.024

Author response

Essential revisions:

More data needs to be included to show the effects of the PUFA on fluorescence signals. The text states that the compound greatly diminishes fluorescence but otherwise little is shown or explained. Do the authors understand the mechanism? Is N-AT a collisional quencher? Is there any voltage-dependence to quenching? If not, why not show F-V relations unscaled and then scaled so the reader can appreciate what is going on? Given the large effects of N-AT on fluorescence, should we be concerned about interpreting small changes in kinetics? Were the traces in Figure 5 normalized to allow comparison of the temporal properties? Although the authors would like to conclude that N-AT has generalized effects that affect both the VSD activation and channel opening, it seems like it could still be the case that effects on channel opening are primary. As such, the FV and GV data might help put this issue on firmer foundation.

We agree that the previous description of N-AT-induced reduction in overall fluorescence intensity was minimal. To test whether N-AT is a collisional quencher, we monitored the fluorescence from unbound Alexa488 in control solution and in N-AT-supplemented control solution (up to 0.5 M N-AT). In these experiments where no oocytes or channels were present, we did not see any quenching effect of N-AT on Alexa488 fluorescence, suggesting that N-AT is not a collisional quencher of Alexa488. N-AT actually slightly increases the Alexa488 fluorescence by 10-20% (in 0.5 M N-AT, Figure 5—figure supplement 1B). Although further experiments will be required to fully understand how N-AT decreases the fluorescence from Alexa-488 bound to Kv7.1+KCNE1 channels, these experiments suggest that the reduced fluorescence observed in our voltage clamp fluorometry experiments are not caused by direct N-AT quenching of the fluorophore. Instead, it is possible that N-AT induces some rearrangement in the channel that brings some unknown Kv7.1 or KCNE1 residue, such as a trp, close to Alexa488 and thereby N-AT indirectly quenches the fluorescence.

Due to the slow kinetics of the Kv7.1+KCNE1 channel and the wide voltage range over which it activates, it takes about 7 minutes to record a full F(V) curve on Kv7.1+KCNE1. During this time course, roughly 50% of the fluorescence signal disappears (now shown in Figure 5—figure supplement 1A). As a consequence, the voltage-dependent changes in fluorescence intensity caused by the S4 movement will be underestimated at more positive voltages compared to negative voltages (we run the protocol from negative to positive test voltages). In addition, the F(V) curves for WT Kv7.1+KCNE1 and Kv7.1/F351A+KCNE1 channels are hard to normalize (especially when N-AT has reduced the fluorescence signal to noise), due to not complete saturation at either the negative or positive end of the F(V) curve. We therefore feel that we presently are unable to reliably quantify N-AT-induced shifts in the F(V) curve for WT Kv7.1+KCNE1 and Kv7.1/F351A+KCNE1. However, we now include a normalized F(V) curve of Kv7.1/G219C*/S225L+KCNE1 (which shows better saturation at both ends of the F(V) curve) in the presence of N-AT, to compare to the corresponding F(V) in control solution (Figure 5—figure supplement 1C). This normalized F(V) curve indicates that N-AT shifts both the first part of the F(V) curve (which represents the main outward S4 movement, i.e. VSD activation) and the second part of the F(V) curve (which represents S4 rearrangements associated with channel opening) towards more negative voltages.

We now clarify in the figure legend that fluorescence traces in Figure 5A,D are normalized to allow kinetic comparison.

To fully understand the mechanism by which N-AT and other PUFA analogues activate the Kv7.1+KCNE1 channel, we need to develop means to reliably measure complete F(V) curves in future studies.

[Editors’ note: the author responses to the first round of peer review follow.]

Reviewer #1:

1) The discussion of the combined F-V and G-V data in Figure 3 in the subsection “KV7.1+KCNE1 channel dysfunction is caused by different underlying mechanisms” needs some work and it is not clear to me that S225L affects voltage-sensor activation and F351A affects pore opening. The data to me seem much more complex than the authors acknowledge. I would like to see a clearer presentation of the evidence supporting the key conclusions and a more nuanced discussion of the underlying complexity. The authors may be correct, but I am not sure we understand why the F-V and G-V overlap for KCNQ alone and what KCNE is doing to those relations, which makes interpretation of what is seen with the mutants less than straightforward.

We agree that the previous discussion of the mechanism of the two mutations were minimal. We have now included some more discussion and added kinetic modeling to show that our proposed mechanisms for these two mutations are plausible. In this kinetic modeling, we have used the two models that our group has previously developed for wild type Kv7.1 and IKs channels (Osteen et al., 2012; Barro-Soria et al., 2014). These models are based on extensive data from many types of experiments (ionic currents, gating currents, fluorescence, linked constructs, complex voltage protocols). We feel that we have earlier extensively shown that the main reason for why the GV and FV overlap are that Kv7.1 channels open after only a subset of S4s are activated (Osteen et al., 2012). In contrast, IKs channels need all four S4s to be activated before they open (Osteen et al., 2012; Barro-Soria et al., 2014). The main result supporting these two activation mechanisms is our data from linked Kv7.1 subunits with different numbers of permanently activated S4s (Osteen et al. 2012). Linked Kv7.1 channels with two permanently activated S4s are open even at hyperpolarized potentials where the two wt S4s have clearly deactivated, showing that Kv7.1 channels open even with a subset of S4s activated. In contrast, the same linked construct co-expressed with KCNE1 is closed at hyperpolarized voltages, showing that IKs channels don’t open after only a subset of S4s have activated. Using these two models, we can show that S225L can be explained in both models by just shifting the voltage dependence of the main S4 movement, whereas F351A can be explained in both models by just shifting the voltage dependence of opening. That the effects of each mutation can be explained in both the Kv7.1 and IKs models by changing a single parameter (different in each mutation) by the same amount is a strong argument for the robustness of the models and the proposed hypothesized mechanisms of the two mutations. Of course, more complex mechanisms of action of these mutations could also generate the same effects on the currents and fluorescence, but we here present the mechanisms we feel are the most simple that can explain most of the phenotypes.

Other models for Kv7.1 and IKs have been proposed, but none of these have been as extensively tested as ours and, in reality, these models fail some of our previous experimental tests. The model that is most similar to ours is the model from Jianmin Cui’s group (Zaydman et al., 2014). In principle, Jianmin’s and our models are very similar: the main difference is that their models have more states and ours are simplified versions of their models. Unfortunately, they use parameters from the model by Rudy and Silva (e.g. a slow S4 and a fast gate), which we have shown is not compatible with more complex triple pulse voltage protocols (Barro-Soria et al., 2014). So the actual parameters in their model would have to be refitted to more complex voltage protocols before their models can be used. However, we now acknowledge that there are other models for Kv7.1 and IKs and that our conclusions about the mechanism of the mutations are model dependent. However, independent of what models are used to describe the channel, it is clear from the differences in their effects on the fluorescence that these two mutations affect the channels by different mechanisms and that N-AT restores most of the function in both mutations.

2) What is the logic of using N-AT instead of other PUFs the authors have previously studied? Given those studies, I would imagine N-AT might have substantial effects on other Kv channels, and possibly even Nav and Cav channels. What makes the authors think that N-AT will have any selectivity for KCNQ? This feels like a stretch to imagine using these in humans at high μM concentrations.

We have toned down the clinical utility of N-AT and instead termed it as a good starting compound for future drug development. We use N-AT because it gives large robust effects. However, its affinity is not ideal. We have recently been able to combine, in novel PUFA compounds, the robust effects of N-AT with the higher affinity effects of other PUFAs. These compounds need further testing and will be presented in future work.

Reviewer #2:

1) Expression in oocytes bypasses these problems due to the low incubation temperatures, so it is unclear whether these gating effects are relevant in the greater scheme of increasing overall current density at the mammalian cell surface. Do the authors know what the expression and biophysical characteristics are for these LQT mutants at physiological temperature compared with wild type to understand how N-AT may alter the abnormal functioning of these mutants?

We agree that trafficking could be a problem for these mutations. However, we do show that for all of the mutations there are gating defects that would be in addition to any putative trafficking effects. Our goal with N-AT is to restore the gating defects, not to restore any trafficking defects per se. We show in Xenopus oocytes that N-AT can restore the function of all mutants, as measured by the current amplitude of wt channels in control solutions compared to mutant channels in N-AT solutions (Figure 4G, Figure 4—figure supplement 4C). In addition, in Xenopus oocytes, most reductions in current amplitude (if not all) can be explained by the gating defects caused by the mutations (Figure 1—figure supplement 4). So, in Xenopus oocytes, we don’t see much of any trafficking defect in these mutants.

However, as the reviewer points out, there might be trafficking defects in mammalian cells that are not present in Xenopus oocytes. Other studies on trafficking and mammalian cells are now referred to (Eldstrom et al., 2010; Harmer et al., 2010). For any mutation that displays trafficking defects, one has to find a cure for this trafficking defect. However, this is outside of the scope of our study. But even if the trafficking is corrected, these mutants would also need to be corrected for their gating defect. Otherwise, one would end up with surface-expressed, defective IKs channels, which could still cause arrhythmia. We now clearly state all these caveats and what N-AT could do to these mutants in these possible cases.

2) There is a little bit of mixture of models in this study. Most experiments are carried out at room temperature in oocytes, but sometimes CHO cells are used for hKv11.1, derived cardiomyocytes and intact hearts are used for action potential and QT studies. Some data that could tie together studies in the different models would support the potential translational importance of the idea.

The data from CHO, cardiomyocytes, and intact hearts have been removed.

3) The data from iPSC-derived cardiomyocytes and intact guinea-pig hearts are unconvincing. In the intact hearts it looks like other changes affect the early action potential to shorten it in the presence of N-AT. Changes in the isoelectric region of the EKG suggest a dispersion of action potential effects by N-AT. In intact hearts, the slope of phase 3 repolarization in the HMR+N-AT trace is shallower than in control or with HMR1556 alone. This suggests less net outward current when N-AT is added, the opposite of what is contended. Perhaps this is because N-AT blocks many other cardiac currents.

This data has been removed.

4) The authors suggest that PUFAs could provide a treatment option for LQTS, but there is a large literature on the effects of PUFAs on humans, in animal models and also on individual ion currents in heart showing that almost all ion currents, inward and outward, including IKr, are decreased by PUFAs. The well-established exception is IKs, but the overall effect on the QT interval duration could be pro- or antiarrhythmic depending on the physiology of the tissue and the pathology affecting the heart.

We have removed the QT interval data on cardiomyocytes and intact heart. We have toned down the therapeutic effect of N-AT and instead focused on the biophysical effect of PUFA analogues on the IKs channel.

Reviewer #3:

1) One topic that probably requires some additional clarification concerns the presentation of GVs and how the V50's and z values were generated. For many of the mutant GV curves, a saturating level of activation is never obtained, but the plotted GV curves are simply normalized to the maximum level of conductance. I think for some (many) readers this may give a misleading impression of what the relative conductances as a function of voltage may be among different constructs (Figures 1C,2A, and others) or in response to N-AT application (e.g., Figure 4B), since the assumption may be made that the voltage of the half maximal conductance after normalization corresponds to the true V50. Although sometimes it is natural to not trust the extrapolated gmax from a Boltzmann fit, plotting the GVs normalized to the fitted Gmax may be a better proxy for the true channel behavior. In such cases, the observed 0.5 level of conductance directly informs a reader of the approximate Vh, and it also informs the reader regarding changes in the slope, which may otherwise be obscured. In Figure 4B, the normalization used by the authors probably tends to minimize the effects at the lower concentrations, while normalizing to the fitted gmax might avoid this. One might also ask in regard to Figure 4B, since the control and N-AT tails are all measured in the same patches, why not use the maximal conductance with 70 μM N-AT as the basis for normalization? My concern regarding the GV displays and the estimates of Vh and z become most critical in regards to the DDG estimates (I think, as the process is described in the methods, this was done appropriately, but some clarifications about this might help).

We have changed the normalization in all graphs. We also better explain in text how we estimate the V50. The V50 is always determined from the fit, as suggested by the reviewer.

https://doi.org/10.7554/eLife.20272.025

Article and author information

Author details

  1. Sara I Liin

    1. Department of Physiology and Biophysics, University of Miami, Miami, United States
    2. Department of Clinical and Experimental Medicine, Linköping University, Linköping, Sweden
    Contribution
    SIL, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article
    For correspondence
    sara.liin@liu.se
    Competing interests
    SIL: A patent application (62/032,739) based on these results has been submitted by the University of Miami with SIL and HPL identified as inventors
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-8493-0114
  2. Johan E Larsson

    Department of Clinical and Experimental Medicine, Linköping University, Linköping, Sweden
    Contribution
    JEL, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article
    Competing interests
    No competing interests declared.
  3. Rene Barro-Soria

    Department of Physiology and Biophysics, University of Miami, Miami, United States
    Contribution
    RB-S, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article
    Competing interests
    No competing interests declared.
  4. Bo Hjorth Bentzen

    1. The Danish Arrhythmia Research Centre, University of Copenhagen, Copenhagen, Denmark
    2. Department of Biomedical Sciences, University of Copenhagen, Copenhagen, Denmark
    Contribution
    BHB, Conception and design, Drafting or revising the article
    Competing interests
    No competing interests declared.
  5. H Peter Larsson

    Department of Physiology and Biophysics, University of Miami, Miami, United States
    Contribution
    HPL, Conception and design, Analysis and interpretation of data, Drafting or revising the article
    For correspondence
    PLarsson@med.miami.edu
    Competing interests
    HPL: A patent application (62/032,739) based on these results has been submitted by the University of Miami with SIL and HPL identified as inventors

Funding

National Institutes of Health (R01GM109762)

  • H Peter Larson

American Heart Association (16GRNT30990060)

  • H Peter Larson

Svenska Sällskapet för Medicinsk Forskning

  • Sara I Liin

Vetenskapsrådet (524-2011-6806)

  • Sara I Liin

Northwest Lions Foundation

  • Sara I Liin

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Frida Starck Härlin (Linköping University) and Briana Watkins (University of Miami) for help with some experiments and Drs. Fredrik Elinder (Linköping University), Laura Bianchi and Feng Qiu (University of Miami), Nicole Schmitt, Mark Skarsfeldt and Federico Denti (University of Copenhagen) for valuable comments.

Ethics

Animal experimentation: Experiments were performed in strict accordance with the recommendations of The Linköping Animal Ethics Committee at Linköping University and The Animal Experiments Inspectorate under the Danish Ministry of Food, Agriculture and Fisheries. Protocols were approved by The Linköping Animal Ethics Committee at Linköping University (#53-13 ) and The Animal Experiments Inspectorate under the Danish Ministry of Food, Agriculture and Fisheries (University of Copenhagen; #2014-15-2934-01061).

Reviewing Editor

  1. Kenton J Swartz, National Institutes of Health, United States

Publication history

  1. Received: August 3, 2016
  2. Accepted: September 28, 2016
  3. Accepted Manuscript published: September 30, 2016 (version 1)
  4. Version of Record published: October 26, 2016 (version 2)
  5. Version of Record updated: November 1, 2016 (version 3)

Copyright

© 2016, Liin et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 879
    Page views
  • 257
    Downloads
  • 8
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, Scopus, PubMed Central.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Structural Biology and Molecular Biophysics
    Keenan C Taylor et al.
    Research Article
    1. Biochemistry and Chemical Biology
    2. Structural Biology and Molecular Biophysics
    Esteban Suárez-Delgado, León D Islas
    Insight