Vasodilation of rat skeletal muscle arteries by the novel BK channel opener GoSlo is mediated by the simultaneous activation of BK and Kv7 channels

Background and Purpose BK channels play important roles in various physiological and pathophysiological processes and thus have been the target of several drug development programmes focused on creating new efficacious BK channel openers, such as the GoSlo‐SR compounds. However, the effect of GoSlo‐SR compounds on vascular smooth muscle has not been studied. Therefore, we tested the hypothesis that GoSlo‐SR compounds dilate arteries exclusively by activating BK channels. Experimental Approach Experiments were performed on rat Gracilis muscle, saphenous, mesenteric and tail arteries using isobaric and isometric myography, sharp microelectrodes, digital droplet PCR and the patch‐clamp technique. Key Results GoSlo‐SR compounds dilated isobaric and relaxed and hyperpolarised isometric vessel preparations and their effects were abolished after (a) functionally eliminating K+ channels by pre‐constriction with 50 mM KCl or (b) blocking all K+ channels known to be expressed in vascular smooth muscle. However, these effects were not blocked when BK channels were inhibited. Surprisingly, the Kv7 channel inhibitor XE991 reduced their effects considerably, but neither Kv1 nor Kv2 channel blockers altered the inhibitory effects of GoSlo‐SR. However, the combined blockade of BK and Kv7 channels abolished the GoSlo‐SR‐induced relaxation. GoSlo‐SR compounds also activated Kv7.4 and Kv7.5 channels expressed in HEK 293 cells. Conclusion and Implications This study shows that GoSlo‐SR compounds are effective relaxants in vascular smooth muscle and mediate their effects by a combined activation of BK and Kv7.4/Kv7.5 channels. Activation of Kv1, Kv2 or Kv7.1 channels or other vasodilator pathways seems not to be involved.

Key Results: GoSlo-SR compounds dilated isobaric and relaxed and hyperpolarised isometric vessel preparations and their effects were abolished after (a) functionally eliminating K + channels by pre-constriction with 50 mM KCl or (b) blocking all K + channels known to be expressed in vascular smooth muscle. However, these effects were not blocked when BK channels were inhibited. Surprisingly, the K v 7 channel inhibitor XE991 reduced their effects considerably, but neither K v 1 nor K v 2 channel blockers altered the inhibitory effects of GoSlo-SR. However, the combined blockade of BK and K v 7 channels abolished the GoSlo-SR-induced relaxation. GoSlo-SR compounds also activated K v 7.4 and K v 7.5 channels expressed in HEK 293 cells.
Conclusion and Implications: This study shows that GoSlo-SR compounds are effective relaxants in vascular smooth muscle and mediate their effects by a combined activation of BK and K v 7.4/K v 7.5 channels. Activation of K v 1, K v 2 or K v 7.1 channels or other vasodilator pathways seems not to be involved.

| INTRODUCTION
Large conductance, calcium-activated potassium channels (BK channels or K Ca 1.1 channels) are expressed in all tissues and organs. They contribute to a wide array of physiological functions in the kidney and neurons (Latorre et al., 2017), as well as in the heart (Balderas, Zhang, Stefani, & Toro, 2015) and both vascular (Brayden & Nelson, 1992) and visceral smooth muscle (Burdyga & Wray, 2005). Altered BK channel function has been suggested to contribute to a variety of disease states including hypertension (see discussion in Kyle & Braun, 2014), diabetes (Lu et al., 2005;McGahon et al., 2007) and detrusor overactivity (Chang et al., 2010). Thus, BK channels play important roles in various physiological processes and changes in their function may contribute to pathophysiological states.
Recently, a more efficacious family of BK channel openers called the GoSlo-SR compounds have been developed (Roy et al., 2012;Roy et al., 2014). The efficacy of some of them has been reported to depend on the presence of BK channel regulatory β and γ-subunits (Kshatri et al., 2017;Large et al., 2015;Webb et al., 2015). These compounds, in particular GoSlo-SR-5-130 and GoSlo-SR-5-6,activated BK channels in freshly isolated smooth muscle cells from rabbit bladder , rabbit corpus cavernosum (Hannigan et al., 2016), and bronchial smooth muscle (Bradley et al., 2018). Furthermore, Webb et al. (2015) demonstrated that the effects of GoSlo-SR-5-6 were reduced by >80%, when a triplet of mutations were introduced on the S4/S5 linker and S6 helix.
Although GoSlo-SR compounds reliably activate BK channels in electrophysiological experiments, their effects on the contractility of intact smooth muscle tissues appear variable. Thus, Large et al. (2015) showed that GoSlo-SR-5-130 decreased rabbit bladder spontaneous contractility but did not alter contractions in response to electrical field stimulation or carbachol application. In contrast, the closely related compound, GoSlo-SR-5-6 failed to alter bladder contractility . In rabbit corpus cavernosum, GoSlo-SR-5-130 decreased spontaneous contractility and its effects (like those on the rabbit bladder) were reversed by iberiotoxin (Hannigan et al., 2016), suggesting that these compounds mediate their effects exclusively by activating BK channels.

What is already known
• BK channels play important roles in various physiological and pathophysiological processes.
• Several drug programmes are focused on creating new efficacious BK channel openers (e.g. GoSlo-SR compounds).

What this study adds
• GoSlo-SR compounds are effective relaxants in vascular smooth muscle.
• They mediate their effects by a combined activation of BK and K v 7.4/K v 7.5 channels.

What is the clinical significance
• GoSlo-SR compounds may be beneficial against combined BK and K v 7 channel dysfunction (e.g, in hypertension).
Even though the effects of GoSlo-SR compounds have been established on urogenital and airways smooth muscle, little is known about their effect on vascular smooth muscle, or if these compounds open other K channels. Given that the contractility of vascular smooth muscle is modulated by a variety of K channels including BK channels (Tykocki, Boerman, & Jackson, 2017), we tested the hypothesis that GoSlo-SR compounds dilate rat arteries exclusively by activating BK channels.

| Animals
The investigation conforms with the U.S. Guide for the Care and Use of Laboratory Animals (Eighth Edition, National Academy of Sciences, 2011). Animal studies are reported in compliance with the ARRIVE guidelines (McGrath & Lilley, 2015) and with the recommendations made by the British Journal of Pharmacology. Approval for the use of laboratory animals in this study was granted by a governmental committee on animal welfare (I-17/17). Adult,male Wistar rats were obtained from Janvier (France; RRID: RGD_13508588). Rats have been used for studies on K + channel function for many years (Tykocki, Boerman, & Jackson, 2017). The animals were provided with food and water ad libitum and housed in a room with a controlled temperature and a 12-hr light-dark cycle in IVC cages.

| Vessel preparation
The rats were killed under CO 2 narcosis by decapitation. The lower extremity (limb), the tail and the mesentery were quickly removed and placed in an ice-cold physiological saline solution composed of (in mM) 145 NaCl, 4.5 KCl, 1.2 NaH 2 PO 4 , 0.1 CaCl 2 , 1.0 MgSO 4 , 0.025 EDTA, 5 HEPES at pH 7.4. All arteries were isolated by removing all surrounding skeletal muscle and connective tissue. Small rings 2 mm in length were used for further experiments.
The microscope image of the vessel was viewed with a CCD camera and digitised by a frame-grabber card (Hasotec, Gemany). Based on the vessel image, diameter changes were measured continuously at a sampling rate of 0.5 Hz using a custom-made programme (Fischer, Mewes, Hopp, & Schubert, 1996). Vessels were exposed to a pressure of 80 mmHg without any luminal flow at a temperature of 37 C. To ensure complete non-flow conditions, leaking vessels were discarded at any stage of the experiment. After development of a spontaneous myogenic tone, vessel viability was tested with noradrenaline at 10 −5 M to test smooth muscle cell function and acetylcholine (ACh) at 10 −6 M to test endothelial cell function. At the end of the experiments, all vessels were exposed to calcium-free solution to determine the fully relaxed diameter at 80 mmHg. The fully relaxed diameter of the vessels in this study was in the range from 240 to 368 μm. All diameter values were normalised to the diameter of the fully relaxed vessel at 80 mmHg in a calcium-free solution. Normalisation was done in order to eliminate variability due to differences in the size of different vessels.

| Isometric mounting of arteries
Isolated vessels were mounted in a wire myograph (model 410A or 610M, Danish Myotechnology, Denmark) for recording of isometric tension on two wires with a diameter of 40 μm. Data acquisition and analysis was performed using Labchart (ADInstruments, USA).
The arteries were stretched to their optimal lumen diameter (90% of the diameter they would have at a transmural pressure of 100 mmHg; Mulvany & Halpern, 1977; wall tension under these conditions corresponds to a pressure of about 45 mmHg according to the law of Laplace) and studied in PSS consisting of (in mM) 120 NaCl, 4.5 KCl, 1.2 NaH 2 PO 4 , 1.0 MgSO 4 , 1.6 CaCl 2 , 0.025 EDTA, 5.5 glucose, 26 NaHCO 3 , and 5 HEPES at pH 7.4 oxygenated with carbogen (95% O 2 and 5% CO 2 ) at 37 C. Viability of the vessels was tested with methoxamine (MX) at 10 −5 M to test smooth muscle cell function and ACh at 10 −5 M after preconstriction with 10 −7 M methoxamine to test endothelial cell function. The solution containing 50 mM KCl was prepared based on PSS by equimolar replacement of NaCl. Vessel tension was normalised to the peak tension developed in response to 10 −5 M methoxamine applied directly after the viability test in order to eliminate variability due to differences in the contractility of different vessels. To be able to compare vessel responses to different interventions, special care was taken to carefully match vessel tension before the intervention. For example, pre-constrictions obtained (a) before application of the GoSlo-SR compound by application of methoxamine alone in the control group of vessels and (b) by application of methoxamine together with IBTX in the treatment group of vessels (different vessels compared to the control group) were the same (see Figure 5b, time point "0").

| Functional removal of the endothelium
In the experiments of this study, the endothelium of the vessels was removed. In isobaric experiments, this was done by passing an air bubble through the lumen of the vessel. In isometric experiments, mechanical disruption of endothelium using a rat whisker was performed. Functional removal of the endothelium was considered successful when ACh-induced vasodilation was absent during the viability test.

| Membrane potential recordings
Intracellular recordings of membrane potential in smooth muscle cells of intact mesenteric arteries were made using microelectrodes pulled from aluminosilicate glass and filled with 3 M KCl. An amplifier (DUO 773, World Precision Instruments) was used to record the membrane potential. A micromanipulator (UMP, Sensapex) was employed to make impalements from the adventitial side of the vessel. The following criteria for acceptance of membrane potential recordings were used: (a) an abrupt change in membrane potential upon cell penetration; (b) a constant electrode resistance when compared before, during, and after the measurement; (c) a stable reading of the membrane potential lasting longer than 1 min; (d) no change in the baseline when the electrode was removed.

| Digital droplet PCR
Vessels, isolated as described above, were cut into small pieces and homogenised for 3 min at 30 Hz in the TissueLyser (Qiagen). Total RNA was isolated using the "miRNeasy Mini-Kit" (Qiagen) according to the manufacturer instructions. Optional On-Column DNase Digestion using the RNase-Free DNase Set (Qiagen) was performed as described. In the final step, RNA was collected from the affinity column using 30 μl H 2 O, which was passed twice over the column. RNA concentration was determined on the Tecan infinite 200.
Samples were quantified by two-step digital droplet PCR. Reverse transcription to cDNA was done using the iScript™ cDNA Synthesis Kit (Bio-Rad, Hercules, CA, Cat#170-8890) according to the manufacturer's standard protocol. All samples were diluted to a starting concentration of 12.5 ng RNA per microliter of reaction.
Primers and probes were either purchased from Bio-Rad or selfdesigned and ordered from Sigma-Aldrich (St. Louis, MO). All probes were FAM-labelled at the 5 0 -end, except GAPDH which was HEXlabelled, and BHQ1-labelled at the 3 0 -end. Amplicon context sequence and amplicon length can be found on the Bio-Rad Homepage (www. CCCGCGGCTGGGCGTTCGTCT), and GAPDH (dRnoCPE5188005).
The annealing temperature was set to 58.0 C based on a temperature gradient run. The limit of detection, the linearity of amplification, and the possibility to do duplex measurements were checked by dilution of synthetic oligonucleotides corresponding to the specific probe sequences. All primers and probes were used at a concentration of 900 and 250 nM, respectively.
Samples were quantified with digital droplet PCR (Hindson et al., 2011)

| Cell isolation
A piece of a tail or mesenteric artery was placed into a microtube containing 1 ml of an isolation solution consisting of (in mM) 55 NaCl, 6 KCl, 88 Na glutamate, 2 MgCl 2 , 10 HEPES, 10 glucose, pH 7.4, as well as 0.6 mgÁml −1 papain, and 1.2 mgÁml −1 DL-DTT for 20 min at 37 C. Thereafter, the artery was moved into a microtube containing 1 ml of the isolation solution as well as 1.2 mgÁml −1 collagenase F, 1.0 mgÁml −1 trypsin inhibitor, and 0.5 mgÁml −1 elastase for 12 min at 37 C for cells from tail arteries or in isolation solution containing 1 mgÁml −1 collagenase (types F and H; ratio, 30% and 70%, respectively) and 0.1 mM CaCl 2 for 16 min at 37 C for cells from mesenteric arteries. Single cells were released by trituration with a polyethylene pipette into the experimental bath solution consisting of (in mM) 126 NaCl, 4. All experiments were carried out in the whole cell configuration of the patch-clamp technique (Hamill, Marty, Neher, Sakmann, & Sigworth, 1981). Cells were held at −80 mV and stepped from −100 to +50 mV for 1 s in 10 mV increments with a 10 s interval between steps. Activation curves were constructed from the peak tail current evoked by a repolarisation back to −120 mV following depolarising voltage steps. Data were fitted with the Boltzmann equation of the form: where V 1/2 was the membrane potential at which there was half maximal activation, K the slope factor, and Vm the membrane potential (mV). The change in activation V 1/2 (ΔV 1/2 ) caused by drugs was obtained by subtracting the V 1/2 in control from that in the presence of the drugs. Leak current was estimated from the current at the end of the −120 mV repolarisation step in the absence of any drugs and was digitally subtracted.

| Statistics
The However, data analysis was performed semi-blinded by an independent analyst. Outliers were included in data analysis and presentation.
Statistical analysis was performed using GraphPadPrism 6.0 (RRID: SCR_002798; GraphPad Software, Inc.) employing ANOVA (parametric test as there was no significant variance inhomogeneity; post hoc tests were conducted only if F in ANOVA achieved P < .05), unpaired or paired Student's t-tests, as appropriate and only on groups with at least n = 5. For methodological reasons, a few groups did not reach n = 5, these data have not been subjected to statistical analysis. A value of P < .05 was considered statistically significant.

| Nomenclature of targets and ligands
Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2019/20 (Alexander et al., 2019).

| Effect of GoSlo-SR compounds on myogenic tone
Isobaric preparations of rat Gracilis muscle arteries possessing spontaneous myogenic tone at 80 mmHg were dilated in a concentration- nimodipine also produced dilation initiated, however, at lower concentrations ( Figure 1d).
Recent publications (Hannigan et al., 2016;Large et al., 2015) suggest that the inhibitory effects of the GoSlo-SR family of compounds are mediated via activation of BK channels. Therefore, we were surprised to find that inhibition of BK channels with IBTX (10 −7 M; Galvez et al., 1990)    Having established the transcriptional expression of KCNQ, we next examined the effects of a K v 7 channel blocker, XE991 (3 × 10 −6 M; Greenwood & Ohya, 2009) on isobaric preparations pressurised to 80 mmHg. As shown in Figure 4a,b, GoSlo-SR-5-6 produced a concentration-dependent dilation of these arteries and this effect was reduced but not abolished when K v 7 channels were blocked with XE991. Of note, XE991 significantly contracted these vessels by 6.4 ± 1.1% which was not different from the effect of IBTX as reported above. Interestingly, this effect of GoSlo-SR-5-6 was unaltered by the K v 7.1 channel blocker HMR1556 (10 −5 M, n = 5), suggesting that either GoSlo-SR-5-6 did not mediate its effects by activating K v 7.1 channels or K v 7.1 channels are not functionally expressed in these vessels. We

| Effect of GoSlo-SR compounds on mesenteric, saphenous and tail arteries
To understand whether the effect of GoSlo-SR compounds is unique to the Gracilis artery, the effect of GoSlo-SR-5-6 was studied on mesenteric, saphenous, and tail arteries. These arteries have been selected, because they represent different vascular beds and Of note, the functional availability of BK and K v 7 channels during contraction induced by MX, the agent used to produce preconstriction when the effect of GoSlo-SR compounds was tested, was observed to be different in these vessels. Thus, in mesenteric ( Figure 12a) and saphenous arteries (Figure 12b), both 10 −7 M IBTX and 3 × 10 −6 M XE991 increased methoxamine-induced contractile responses to a similar degree. In contrast, in tail arteries, this effect was observed only for IBTX (Figure 12c).

| Effect of GoSlo-SR compounds on membrane potential and BK currents
To get more direct evidence for the involvement of ion channels in the relaxations induced by GoSlo-SR compounds, electrophysiological experiments were performed. We first measured the membrane potential in smooth muscle cells of intact mesenteric arteries using microelectrodes. The mesenteric artery was selected for these experiments because it was possible to get reliable membrane potential measurements (for criteria, see Section 2.6). We observed that the GoSlo-SR-5-6-induced relaxation was associated with a hyperpolarisation from −29 ± 4 to −45 ± 3 mV (n = 7; Figure 13). We were able to reverse this hyperpolarisation and relaxation with the K v 7 channel blocker XE991 (Figure 13). In a few of these vessels, we were also able to measure the membrane potential after the subsequent addition of IBTX, in the continued presence of XE991. In these preliminary experiments (n = 3), membrane potential was −33 ± 2 mV in suggesting that BK channels did contribute to the hyperpolarisation.
However, these experiments were difficult to perform, and impalements were often lost in XE991 and IBTX, due to the induction of spontaneous activity by these ion channel blockers. To explore the role of BK channels further, we also measured BK currents in freshly isolated tail artery smooth muscle cells, which have been well documented to possess BK currents (Schubert, Noack, & Serebryakov, 1999;Schubert, Serebryakov, Engel, & Hopp, 1996). We observed that GoSlo-SR-5-6 induced a considerable increase in BK currents ( Figure 14). A similar increase in BK currents was obtained in freshly isolated rat mesenteric artery smooth muscle cells studied. However, a pronounced current run-down precluded more detailed studies with these cells.

| Effect of GoSlo-SR compounds on K v 7.4 and K v 7.5 channels
To test if GoSlo-SR compounds activated K v 7 channels, we examined the effects of extracellularly applied GoSlo-SR-5-6 (10 −5 M) on whole cell currents recorded from HEK cells transiently transfected with human K v 7.4 cDNA. Preliminary experiments established that the EC 50 for GoSlo-SR-5-6 on K v 7.4 applied at a potential of −40 mV was 6.4 × 10 −6 ± 0.5 μM (n = 5), suggesting that it was slightly less potent on K v 7.4 channels, compared to BK channels (2.3 × 10 −6 μM) reported previously (Roy et al., 2012). Figure 15a shows a family of currents recorded from a cell expressing K v 7.4 channels. In the absence of any drugs, these currents activated slowly at potentials positive to −60 mV (Figure 15c). Application of GoSlo-SR-5-6 (10 −5 M) increased current amplitude at all voltages tested and dramatically slowed tail current deactivation (Figure 15b). These effects were reversible on washout, and the currents were blocked in the presence of XE991 (10 −5 M, exploratory data, n = 4, data not shown).
The activation curve in the presence of GoSlo-SR-5-6 (10 −5 M) was characterised by a shift of V 1/2 by approximately −40 mV (n = 6; P < .05; paired t test); the slope factor was increased from 18 ± 1 mV to 28 ± 2 mV (n = 6; P < .05; paired t test; Figure 15c). It is important to note that the application of GoSlo-SR-5-6 increased the amplitude of the peak tail current at all potentials recorded. It is clear from these data that GoSlo-SR-5-6 activates K v 7.4 channels and shifts their voltage-dependent activation to more negative potentials. Another K v 7 channel activator, ML213 (10 −5 M), also significantly shifted the activation V 1/2 by −35 ± 2 mV (from −20 ± 2 mV to −54 ± 2 mV, ) and increased G max to 2.2 ± 0.2 and these effects were abolished in the W242L mutant (n = 5, data not shown). In contrast, the effects of GoSlo-SR-5-6 were not altered in K v 7.4 channels with a W242L mutant (n = 5, data not shown), suggesting that, in contrast to retigabine, residue W242 was not essential for GoSlo-SR compounds to mediate their effects.
GoSlo-SR-5-130 (10 −5 M) also activated K v 7.4 channels ( Figure   15,e,f), but it was clearly less efficacious than GoSlo-SR-5-6 ( Figure 15a,b,c). We were unable to determine an EC 50 for this compound on K v 7.4 because maximal effects were not reached at 30 μM and the compound came out of solution at higher concentrations. Nevertheless, we only observed a small increase in the steady state current amplitude with 10 −5 M GoSlo-SR-5-130. However, the tail currents were clearly slowed compared to the control currents. GoSlo-SR-5-130 significantly shifted the activation V 1/2 by approximately −20 mV (n = 6) , and this was significantly less effective than GoSlo-SR-5-6 (ΔV 1/2 approximately −40 mV). The slope factor K was unaffected (19 ± 1 mV under control conditions and 21 ± 2 mV in GoSlo-SR-5-130).
We next examined the effects of the two GoSlo-SR compounds on HEK cells expressing K v 7.5 channels. We were unable to determine the EC 50 of either GoSlo-SR-5-6 or 5-130 on K v 7.5 due to limited solubility of these compounds in Hank's solution at concentrations above 30 μM. A brief inspection of the current amplitude in the first 50 ms demonstrates that GoSlo-SR-5-6 (10 −5 M) increased the amplitude of the K v 7.5 current at all voltages, but this effect was particularly apparent at negative potentials (Figure 16a,b).
At potentials positive to −60 mV, it is clear that although GoSlo-SR-5-6 increased the initial current amplitude, the currents decreased during the depolarising pulse, presumably as a result of open channel block. When the cell was repolarised from positive potentials back to −120 mV, the apparent block was relieved, and the tail current amplitude consequently increased over time. The activation curve of the control current was constructed from tail currents measured 100 ms after the repolarisation step began, to minimise any distortion of the relationship caused by the block at positive potentials and was characterised by a V 1/2 of −29 ± 7 mV and the slope factor was 19 ± 4 mV. It was not possible to measure either V 1/2 or slope factor in the presence of GoSlo-SR-5-6, as the activation curve was approximately linear over the entire voltage range recorded. What is clear, however, is that the current amplitude in GoSlo-SR-5-6 was much greater at every voltage recorded, compared to control.
Further, it significantly shifted V 1/2 by approximately −50 mV (slope factors were unaffected 20 ± 3 mV in control compared to 18 ± 4 mV in the presence of the drug (Figure 16f)).

| DISCUSSION
We utilised isometric tension and isobaric diameter recordings of rat arteries to examine the effects of two recently disclosed BK channel openers on vascular smooth muscle. Our focus was to ascertain if these compounds relaxed vascular smooth muscle and determine if these effects were mediated exclusively through activation of BK channels.

| Effect of GoSlo-SR compounds on rat arteries
Two different GoSlo-SR compounds, GoSlo-SR-5-6 and GoSlo-SR-5-130, produced a strong relaxation of isobaric as well as isometric preparations of rat Gracilis muscle arteries. To the best of our knowledge, the effects of GoSlo-SR compounds on blood vessels have not been reported before. These findings are supported by the reports demonstrating that GoSlo-SR-5-130 reduced spontaneous contractility in rabbit visceral smooth muscle. However, both GoSlo-SR compounds did not affect induced contractility in these preparations (Hannigan et al., 2016;Large et al., 2015). Taken together, these findings show that GoSlo-SR compounds are much more effective relaxants in vascular compared to visceral smooth muscle.
Previous studies showed that some effects of GoSlo-SR compounds on visceral smooth muscle contractility are abolished by pretreatment with specific BK channel blockers (Hannigan et al., 2016;Large et al., 2015). Consequently, in the present study on vascular smooth muscle, we tested whether K + channels in general mediate the effect of GoSlo-SR compounds. The influence of K + channels on vessel tension was functionally eliminated by pre-constricting the vessels with 50 mM KCl. At this extracellular KCl concentration, the equilibrium potential for K + is close to the actual membrane potential of smooth muscle cells. Thus, the driving force for potassium ions is negligible. Even if K + channels were open, there would be no K + efflux, hence no alteration of the membrane potential and no change in contractility. Importantly, vasodilators acting on other mechanisms except K + channels would retain their ability to affect vessel tone; only K + channel openers would lose this capability. Indeed, this was observed; the GoSlo-SR compounds were without any effect after pre-constriction of the vessels with 50 mM KCl. Moreover, the same effect was seen after blocking functionally important K + channels in vascular smooth muscle (Nelson & Quayle, 1995;Tykocki, Boerman, & Jackson, 2017) using the BK channel inhibitor iberiotoxin (Galvez et al., 1990), the K v 2 channel blocker stromatoxin (Escoubas, Diochot, Celerier, Nakajima, & Lazdunski, 2002), the K v 1 channel F I G U R E 1 2 IBTX and XE991 affect methoxamine-induced contraction. (a) Effect of 10 −7 M IBTX and 3 × 10 −6 M XE991 on methoxamineinduced contraction in mesenteric arteries (repeated measures ANOVA: con vs. IBTX: n = 8; P < .05; con vs. XE991: n = 8; P < .05; IBTX vs. XE: n = 8; P = .20); (b) effect of 10 −7 M IBTX and 3 × 10 −6 M XE991 on methoxamine-induced contraction in saphenous arteries (repeated measures ANOVA: con vs. IBTX: n = 9; P < .05; con vs. XE991: n = 9; P < .05; IBTX vs. XE: n = 9; P = .10); (c) effect of 10 −7 M IBTX and 3 × 10 −6 M XE991 on methoxamine-induced contraction in tail arteries (repeated measures ANOVA: con vs. IBTX: n = 7; P < .05; con vs. XE991: n = 7; P = .39) inhibitor DPO-1 (Lagrutta, Wang, Fermini, & Salata, 2006;Tsvetkov et al., 2016) and the K v 7 channel blocker XE991 (Greenwood & Ohya, 2009). The fact that the GoSlo-SR compounds were unable to relax blood vessels after either treatment supports the idea that GoSlo-SR compound-induced vasodilation was due to activation of K + channels. This is further supported by previous findings showing that GoSlo-SR-5-6 and 5-130 had no significant effect on smooth muscle L-type calcium currents , that GoSlo-SR-5-130 was not able to affect contractile activity in rabbit bladder in the presence of IBTX  and that GoSlo-SR-5-130 had no effect on KCl-induced contractions in rabbit corpus cavernosum (Hannigan et al., 2016).
In conclusion, the data presented in this study show that GoSlo-SR compounds mediate their vasodilator effects exclusively by

| Role of BK channels in mediating the effect of GoSlo-SR compounds on rat arteries
Recent publications (Hannigan et al., 2016;Large et al., 2015) suggest that the inhibitory effects of the GoSlo-SR family of compounds on urogenital smooth muscles are mediated by activation of BK channels.
Thus, we hypothesised that BK channels may play a leading role in the effect of GoSlo-SR compounds on vascular smooth muscle. Unexpectedly, we were, initially, unable to get any direct support for an involvement of BK channels in the effect of GoSlo-SR compounds on vascular smooth muscle, despite evidence that BK channels were functional in these preparations. Thus, pretreatment of Gracilis, mesenteric, or saphenous arteries with the most specific BK channel inhibitor, iberiotoxin (Galvez et al., 1990) did not alter the vasodilation induced by the GoSlo-SR compounds, even when IBTX was applied at a high concentration (3 × 10 −7 M). However, it is important to note that IBTX alone abolished the GoSlo-SR-induced relaxation in rat tail artery, which, as discussed later, is because K v 7 channels are functionally unavailable in these vessels.
In the Gracilis artery, two other widely used BK channel inhibitors, TEA at 10 −3 M, a concentration affecting primarily BK channels (Nelson & Quayle, 1995), and the specific BK channel inhibitor penitrem A at 10 −7 M (Knaus et al., 1994), were also unable to modify the effect of the GoSlo-SR compounds. Finally, we F I G U R E 1 5 GoSlo-SR compounds activate K v 7.4 channels. (a) Typical family of currents obtained from a HEK cell during a series of voltage steps from −100 to +60 mV in 10-mV increments lasting 1 s. Cells were held at −80 mV and repolarised back to −120 mV to obtain tail currents; (b) currents from the same cell during incubation with 10 −5 M GoSlo-SR-5-6. Tail current deactivation (τ) recorded at −120 mV following a step to +40 mV increased from 15 ± 1 ms to 47 ± 5 ms (n = 6; P < .05; paired Student's t-test); (c) summary activation curves obtained by measuring tail currents in six cells before (open circles) and during (blue circles) application of GoSlo-SR-5-6; (d) and (e) currents obtained from a different cell, held at −80 mV and stepped from −100 mV to +50 mV in 10 mV increments, in the absence and presence of GoSlo-SR-5-130 (10 −5 M), respectively. Tail currents recorded at −120 mV following a step to +40 mV decayed with a τ of 15 ± 2 ms in control conditions compared to 26 ± 1 ms in GoSlo-SR-5-130 (n = 6; P < .05; paired t test); (f) summary activation curves obtained from six cells in the absence (open circles) and presence (pink circles) of GoSlo-SR-5-130 considered the possibility that the BK channel blocker may affect the efficiency of the GoSlo-SR compounds to dilate vessels but not their maximal effect at saturating concentrations. However, neither IBTX nor penitrem A altered the effect of the GoSlo-SR compound at a lower concentration not producing full vasodilation. These data suggested that GoSlo-SR compounds either interfered with the binding of BK channel blockers or that they activated other K + channels. The former explanation appears unlikely, given that the effects of GoSlo-SR compounds have been shown to be blocked with IBTX and penitrem A in tissue strips (Hannigan et al., 2016;Large et al., 2015) and in single cells  and that in our study IBTX inhibited the effect of GoSlo-SR compounds in the tail artery and regained a blocking effect against GoSlo-SR compounds in the presence of the K v 7 channel inhibitor XE991 (for more details, see below).
In conclusion, the data presented in this study show that GoSlo-SR compounds mediated their vasodilator effects exclusively by activating K + channels but are not consistent with the idea that activation of BK channels is the predominant mechanism mediating their effect in either rat Gracilis, mesenteric or saphenous arteries.

| Role of K v 7 channels in the effect of GoSlo-SR compounds on rat arteries
In view of our conclusion that other K + channels, in addition to BK channels, mediate the vasodilator effect of GoSlo-SR compounds, we hypothesised that K v 7 channels may play a leading role in this effect.
Furthermore, we obtained novel data proposing an involvement of K v 7 channels in the effect of GoSlo-SR compounds on vascular smooth muscle. Thus, the widely used K v 7 channel inhibitor XE991 (Greenwood & Ohya, 2009;Zavaritskaya et al., 2013) reduced the vasodilating, as well as the hyperpolarising effects of the GoSlo-SR compounds considerably. XE991 has been employed with great success to identify specific roles of KCNQ-encoded channels in the circulatory system (Greenwood & Ohya, 2009;Mackie & Byron, 2008). Of note, we observed that the partial inhibition of the GoSlo-SR-induced dilation by XE991 was not altered further after elevating the concentration of XE991 from 3 × 10 −6 to 10 −5 M, suggesting that activation of K v 7 channels was not the only mechanism mediating GoSlo-SR-induced vasodilation.
The GoSlo-SR-induced vasodilation that remained in the presence of XE991 was not affected when XE991 was co-applied with either the specific K v 1 channel inhibitor DPO-1 (Lagrutta, Wang, Fermini, & Salata, 2006;Tsvetkov et al., 2016) or the specific K v 2 channel inhibitor stromatoxin (Escoubas, Diochot, Celerier, Nakajima, & Lazdunski, 2002). K v 1 and K v 2 channels are the other major K v channel subtypes expressed in arterial smooth muscle (Albarwani et al., 2003;Amberg & Santana, 2006). Thus, either K v 1 or K v 2 channels are not activated by the GoSlo-SR compounds or these K v channels are not functionally available in rat Gracilis arteries. The latter explanation seems unlikely, because we have observed in an ongoing study that DPO-1 and stromatoxin are able to modify myogenic constriction of this vessel (data not published).
Thus, K v 1 and K v 2 channels are functionally available in rat Gracilis arteries but are not involved in the vasorelaxant effects of GoSlo-SR compounds.
Based on the transcriptional expression data, our results suggest that the GoSlo-SR-induced vasodilation in Gracilis muscle, mesenteric and saphenous arteries is mediated mainly by K v 7.4 and K v 7.5 channels. We excluded any contribution from K v 7.1 channels in this response since the specific K v 7.1 channel inhibitor HMR1556 (Chadha et al., 2012;Gogelein, Bruggemann, Gerlach, Brendel, & Busch, 2000) failed to affect the vasodilatory activity of GoSlo-SR-5-6.
Importantly, we observed that GoSlo-SR compounds activated K v 7.4 and K v 7.5 currents, an effect associated with a shift of the activation properties of these channels to more negative potentials. In this respect, GoSlo-SR-5-130, the GoSlo-SR compound with weaker vasodilating capacity, appeared to be much less efficacious than GoSlo-SR-5-6. Taken together, our data strongly suggest that activation of K v 7.4 and/or K v 7.5 channels or of K v 7.4/7.5 heteromeric channels (Brueggemann et al., 2014;Chadha et al., 2014) contribute to the vasorelaxant effects of GoSlo-SR compounds.

| Role of BK and K v 7 channels in the effect of GoSlo-SR compounds on rat arteries
As discussed so far, blockade of K v 7 channels only partially reduced the effect of the GoSlo-SR compounds in Gracilis, mesenteric and saphenous arteries. However, the combined application of IBTX and XE991 to either isobaric or isometric vessel preparations abolished the vasodilating effect of the GoSlo-SR compounds completely. This supports the idea that they relaxed Gracilis muscle arteries by activating both BK and K v 7 channels. This dual action of GoSlo-SR compounds on BK and K v 7.4/K v 7.5 channels is further supported by the findings made on single cells expressing these channels, that is, (a) our findings reported in the present study demonstrating that GoSlo-SR compounds produce a large stimulation of native BK currents and K v 7.4 and K v 7.5 channels and (b) previously published findings showing that GoSlo-SR compounds activate expressed as well as native BK channels (Hannigan et al., 2016;Kshatri et al., 2017;Large et al., 2015;Roy et al., 2012;Roy et al., 2014;. Incidentally, the BK channel opener NS11021 has been shown to stimulate expressed K v 7.4 channels (Bentzen et al., 2007), and BMS204352 activates both BK and K v 7 channels (Schroder, Strobaek, Olesen, & Christophersen, 2003). Thus, the joint activation of BK and K v 7.4/K v 7.5 channels by BK channel opener compounds is not without precedent and perhaps suggests that they interact with a common site on both BK and K v channels. Future studies will be focused at determining the precise location of this site in K v 7 channels.
Importantly, the degree of contribution of BK and K v 7.4/K v 7.5 channels to the GoSlo-SR compound-induced vasodilation varied depending on the experimental conditions. Thus, GoSlo-SR compound-induced vasodilation (a) was not affected by inhibition of BK channels when K v 7.4/7.5 channels were not blocked (b) but was completely abolished by inhibition of BK channels when K v 7.4/7.5 channels were blocked. Of note, the latter finding is emphasised by our data on tail arteries. In this artery, in contrast to all other vessels studied, K v 7 channels appeared to be functionally unavailable during MX-induced contraction, as evidenced by the absence of an effect of XE991 on this contraction (see Figure 12c). Here, again in contrast to the Gracilis, mesenteric, and saphenous arteries, the GoSlo-SR compound-induced vasodilation was completely abolished when BK channels alone were blocked. Together, this suggests that when K v 7.4/7.5 channels are not functionally available, the effect of GoSlo-SR compounds on BK channels was sufficient to relax the blood vessels. However, when K v 7.4/7.5 channels were functionally available, blockade of BK channels failed to reduce the response to GoSlo-SR compounds. In addition, except in the tail artery, GoSlo-SR-induced vasodilation was reduced by inhibition of K v 7 channels either partly, when BK channels were not blocked, or fully when BK channels were blocked.
A possible explanation for our observations has been suggested recently (Coleman, Tare, & Parkington, 2017). If the GoSlo compounds produce a considerable hyperpolarisation, the membrane potential will be much closer to the potassium equilibrium potential resulting in a small driving force for potassium ions. Due to the small driving force, blockade of potassium channels under these conditions will result in only a small change in membrane potential and vessel tension. Thus, our data are consistent with the idea that when IBTX has blocked functional BK channels at small driving force, membrane potential was almost not affected, and the effect of GoSlo was unchanged. When XE991 has blocked functional K v 7 channels (with a somewhat larger impact, compared to blocking BK channels) under conditions where the driving force is small, membrane potential was presumably affected and the effect of GoSlo was reduced. However, when IBTX together with XE991 has blocked functional BK and K v 7 channels, the effects on GoSlo on membrane potential were presumably blocked, and as a result, the GoSlo-induced relaxation was attenuated.
In conclusion, the data presented in this study show that GoSlo compounds are much more effective relaxants in vascular compared to visceral smooth muscle. Like other small molecule BK channel openers, GoSlo-SR compounds mediate their vasodilator effects by a combined activation of BK and K v 7.4/K v 7.5 channels. Activation of K v 1, K v 2, or K v 7.1 channels or other vasodilator pathways, for example, voltage-gated calcium channels, seems not to be involved in this effect. Whereas the joint activation of BK and K v 7.4/K v 7.5 channels by the GoSlo-SR compounds is not without precedent, the GoSlo-SR compound-induced vasodilation was characterised by a special feature. This vasodilation was mediated by K v 7.4/7.5 channels only when BK and K v 7.4/7.5 channels were available but was mediated by BK channels when K v 7.4/7.5 channels were not available. This special mechanism of action of GoSlo-SR compounds may be beneficial for their clinical use as K + channels openers, for example, against combined BK and K v 7 channel dysfunction like in hypertension. This idea has to be confirmed in future studies.