The Natural Flavone Acacetin Blocks Small Conductance Ca2+-Activated K+ Channels Stably Expressed in HEK 293 Cells

The natural flavone acacetin inhibits several voltage-gated potassium currents in atrial myocytes, and has anti-atrial fibrillation (AF) effect in experimental AF models. The present study investigates whether acacetin inhibits the Ca2+-activated potassium (KCa) currents, including small conductance (SKCa1, SKCa2, and SKCa3), intermediate conductance (IKCa), and large-conductance (BKCa) channels stably expressed in HEK 293 cells. The effects of acacetin on these KCa channels were determined with a whole-cell patch voltage-clamp technique. The results showed that acacetin inhibited the three subtype SKCa channel currents in concentration-dependent manner with IC50 of 12.4 μM for SKCa1, 10.8 μM for SKCa2, and 11.6 μM for SKCa3. Site-directed mutagenesis of SKCa3 channels generated the mutants H490N, S512T, H521N, and A537V. Acacetin inhibited the mutants with IC50 of 118.5 μM for H490N, 275.2 μM for S512T, 15.3 μM for H521N, and 10.6 μM for A537V, suggesting that acacetin interacts with the P-loop helix of SKCa3 channel. However, acacetin at 3–10 μM did not decrease, but induced a slight increase of BKCa (+70 mV) by 8% at 30 μM. These results demonstrate the novel information that acacetin remarkably inhibits SKCa channels, but not IKCa or BKCa channels, which suggests that blockade of SKCa by acacetin likely contributes to its anti-AF property previously observed in experimental AF.


INTRODUCTION
Potassium channels are the largest and the most diverse super-family of ion channels in living organisms from bacteria and insects to animals including humans. Among them, Ca 2+ -activated potassium channels (K Ca ) comprise many members. They are divided into three subfamilies: big (or large) conductance (BK Ca , Slo, or K Ca 1.1, encoded by KCNMA1), intermediate conductance (IK Ca or K Ca 3.1, encoded by KCNN4), and small conductance channels (SK Ca 1, SK Ca 2, and SK Ca 3 or K Ca 2.1, K Ca 2.2, and K Ca 2.3, encoded by KCNN1, KCNN2, and KCNN3, respectively) (Girault et al., 2012;Gueguinou et al., 2014). The three SK Ca channels are expressed in excitable tissues (e.g., neurons, skeletal muscle, adrenal gland, and heart) and also in some non-excitable tissues (e.g., liver, vascular endothelium, cancers, etc.) (Wei et al., 2005). In neurons, apamin-sensitive SK Ca current is responsible for afterhyperpolarization (Weatherall et al., 2010) and regulates firing frequency as well as learning and memory (Adelman et al., 2012). In the cardiovascular system, SK Ca channels contribute to cardiac repolarization (Xu et al., 2003;Li et al., 2009;Zhang et al., 2014), endotheliumderived hyperpolarization-type arterial dilation in response to increased hemodynamics (Wulff and Kohler, 2013), and also provide negative feedback on sympathetic tone (Taylor et al., 2003). Results from recent studies suggest that SK Ca channels play a role in atrial fibrillation (AF) (Diness et al., 2010;Ellinor et al., 2010;Qi et al., 2014;Haugaard et al., 2015), tumor cell migration and metastasis (Chantome et al., 2013), and overactive bladder (Soder et al., 2013). A recent report demonstrated that the SK Ca inhibitor apamin may cause ventricular arrhythmias in failing rabbit hearts ; however, blockade of SK Ca channels is very effective in anti-AF (Diness et al., 2010;Qi et al., 2014;Haugaard et al., 2015). The development of SK channel blockers has been considered as a new therapeutic strategy in the treatment of AF (Zhang et al., 2015).
We have previously reported that the natural flavone acacetin from the traditional Chinese medicinal herb Xuelianhua (Saussurea involucrata) prolongs the atrial effective refractory period and prevents or terminates the experimentally induced AF in anesthetized dogs without increasing the QT interval (Li et al., 2008;Liu et al., 2016) by inhibiting atrial I Kur (ultra-rapidly activating delayed rectifier potassium current) or Kv1.5, I K.ACh (acetylcholine-activated potassium current), and I to (transient outward potassium current) (Wu et al., 2011(Wu et al., , 2013a. The present study investigated the effects of acacetin on SK Ca 1, SK Ca 2, SK Ca 3, IK Ca , and BK Ca currents in HEK 293 cells stably expressing corresponding channel genes with a conventional whole-cell patch voltage-clamp technique. Our results showed that acacetin inhibited the three subtypes of SK Ca channels, but not IK Ca and BK Ca channels, suggesting that the blockade of SK Ca channels may also participate in the anti-AF previously observed in experimental canine models.

Cell Line Culture and Gene Transfection
The pCDNA3/rSK Ca 2 (KCNN2), pCDNA3/hSK Ca 3 (KCNN3), and pCDNA3/hIK Ca (KCNN4) plasmids obtained as generous gifts from Dr. Nicole Schmitt (Department of Biomedical Sciences, University of Copenhagen, Copenhagen, Denmark) were transfected into HEK 293 cells (ATCC, Manassas, VA, United States) using Lipofectamine 2000 TM . The HEK 293 cell lines stably expressing the SK Ca 1, SK Ca 2, and SK Ca 3 channels were established as described previously (Wu et al., 2012). The cell lines were maintained in Dulbecco's modified Eagle's medium (Invitrogen, Hong Kong, China) supplemented with 10% fetal bovine serum and G418 (400 µg/ml). HEK 293 cell line (Wu et al., 2013c) stably expressing human BK Ca (KCNMA1) was also maintained in the same culture conditions. Cells were seeded on glass cover slips for electrophysiological recording.
The primers of SK Ca 3 mutants were synthesized by the Genome Research Center, the University of Hong Kong (Hong Kong), and the mutants were generated using a QuikChange kit (Stratagene, La Jolla, CA, United States). After confirmed by DNA sequencing, the mutants were transiently expressed in HEK 293 cells in a 35 mm culture dish using Lipofectamine 2000 TM (10 µl) with SK Ca 3 mutant cDNA plasmid (4 µg).

Electrophysiology
The HEK 293 cells on a coverslip were placed into a cell chamber mounted on the stage of an inverted microscope (Olympus, IX70, Japan), and superfused with Tyrode's solution (2 ml/min). Whole-cell current was recorded with a patch clamp amplifier (EPC-10, HEKA Elektronik, Lambrecht, Germany) as described previously (Wu et al., 2011(Wu et al., , 2012(Wu et al., , 2013cSun et al., 2014). Briefly, glass electrodes (1.2 mm OD) were pulled with a Brown-Flaming puller (Model P-97, Sutter Instrument Co., Novato, CA, United States). Resistance of the glass pipettes was 2-3 M when filled with the pipette solution. Whole-cell configuration was established by a gentle suction after a gigaohm-seal was obtained. Electrical signal was stored on the hard disk of a PC computer. All experiments were performed at room temperature (22-23 • C).

Statistical Analysis
The data were expressed as means ± SEM. Unpaired Student's t-tests were used as appropriate to evaluate the differences between two group means, and ANOVA was used for multiple groups. A value of P < 0.05 was considered to indicate statistical significance.

Effect of Acacetin on SK Ca 1 Current
The effect of acacetin on SK Ca 1 current was determined in HEK 293 cells stably expressing human KCNN1. Figure 1A displays the voltage-dependent SK Ca 1 current recorded with 200ms voltage steps between −70 and +80 mV from a holding potential of −80 mV in a representative cell. The current was inhibited by 10 µM acacetin (10 min exposure), and the inhibition was partially reversed by washout. Figure 1B displays the current-voltage (I-V) relationships of SK Ca 1 determined in another typical experiment with a voltage ramp in the absence and presence of acacetin. The I-V relationships of SK Ca 1 current showed a reversal potential around −70 mV and an inward rectification property, typical SK Ca current as described previously (Girault et al., 2011;Wu et al., 2013b). The current was significantly descreased by 10 µM acacetin in bath solution, and the inhibition was partially reversed on washout. Figure 1C illustrates the concentration-dependent inhibition of SK Ca 1 current (at +80 mV) by acacetin. The concentrationresponse curve was fitted to a Hill equation to obtain IC 50 (the concentration of inhibiting 50% current) value. The IC 50 of acacetin for inhibiting SK Ca 1 at +80 mV was 12.4 µM (Hill co-efficient, 0.8).

Effect of Acacetin on SK Ca 2 Current
The effect of acacetin on SK Ca 2 current was determined in HEK 293 cell line expressing rat KCNN2. Figure 2A shows the voltagedependent SK Ca 2 current in a typical experiment with the voltage protocol as shown in the inset. The current was significantly decreased by 10 µM acacetin (10 min exposure) at all testing potentials, and the inhibition was partially reversed by washout. I-V relationships of SK Ca 2 current determined by a ramp voltage protocol also showed an inward rectification. Inward and outward components of the current were decreased by 10 µM acacetin, and the inhibition was partially recovered on drug washout ( Figure 2B). Figure 2C illustrates the concentrationresponse relationship of acacetin for inhibiting SK Ca 2 current (at +80 mV). The curve was fitted to a Hill equation. The IC 50 of acacetin for inhibiting SK Ca 2 current was 10.8 µM (Hill coefficient, 0.8).

Inhibition of SK Ca 3 Current by Acacetin
The inhibitory effect of acacetin on SK Ca 3 was determined in HEK 293 cell line expressing human KCNN3 gene. The voltage-dependent step SK Ca 3 current was determined with the voltage protocol as shown in the inset in a typical experiment ( Figure 3A). SK Ca 3 current at all test potentials was inhibited by 10 µM acacetin with 10 min incubation, and the inhibition partially recovered on washout for 10 min. I-V relationships of SK Ca 3 current was determined with a ramp voltage protocol in another cell ( Figure 3B) before and after application of acacetin. The current also displays an inward rectification and was reversibly decreased by 10 µM acacetin. The concentrationresponse curve of acacetin for inhibiting SK Ca 3 current was fitted to a Hill equation (Figure 3C). The IC 50 of acacetin for inhibiting SK Ca 3 was 11.6 µM (with a Hill coefficient of 0.8).

Molecular Determinant of Acacetin for Inhibiting SK Ca Channels
The potential molecular determinant of acacetin for inhibiting SK Ca channels was investigated using SK Ca 3 mutants, H490N, S512T, H521N, and A537V in P-loop helix and S6, generated by site-directed mutagenesis as described previously (Wu et al., 2013a,b). Figure 4A illustrates the I-V relationships of wild type (WT) SK Ca current and mutant currents recorded in representative cells expressing WT SK Ca 3 or the mutant H490N, S512T, or H521N with a voltage ramp protocol before (control) and after 10 µM acacetin. The inhibitory effect of acacetin for the mutant H490N and S512T currents was clearly reduced, compared with WT SK Ca 3 current. Figure 4B illustrates the percent values of current inhibition by 10 µM acacetin for WT SK Ca 3, and the mutants H490N, S512T, H521N, and A537V currents at +80 mV. Acacetin at 10 µM decreased the current by 45.7 ± 4.1% for WT SK Ca 1 current (n = 11), 21.9 ± 4.5% for H490N current (n = 7, P < 0.01 vs. WT), 17.9 ± 3.9% for S512T current (n = 7, P < 0.01 vs. WT), 48.8 ± 3.5% for H521N current (n = 7, P > 0.05 vs. WT), and 40.8 ± 7.6% for A537V current (n = 6, P > 0.05 vs. WT), respectively. Figure 5A displays the concentration-response relationships of acacetin for inhibiting WT SK Ca 3 current, H490N current, S512T current, H521N current, and A537V current at +80 mV. The concentration-dependent inhibition curves were fitted to a Hill equation. The IC 50 of acacetin was 11.6 µM for WT SK Ca 3 current, 118.5 µM for H490N current, 275.2 µM for S512T current, 15.3 µM for H521N current, and 10.6 µM for A537V current, respectively. The efficacy of acacetin for inhibiting H490N current and S512T current was dramatically reduced, which suggests that acacetin blocks SK Ca 3 channel by interacting with H490 and S512 in the P-loop helix of the channel ( Figure 5B).
voltages as shown in the inset in a typical experiment before and after application of acacetin. Acacetin (10 and 30 µM) slightly decreased the current, and the inhibition was partially recovered on washout. Similar results were observed for the I-V relationships of the current recorded with a voltage ramp in another representative cell ( Figure 6B). IK Ca shows a linear I-V relationship, similar to those previously recorded in HEK 293 cell line expressing IK Ca (Girault et al., 2011). Figure 6C shows that acacetin (10 and 30 µM) decreased IK Ca (+70 mV) to 95.0 ± 4.5% (n = 7, P > 0.05) and 89.3 ± 5.5% of control (n = 7, P < 0.05 vs. control, 0 µM), respectively. These results suggest that acacetin has a slight inhibition of IK Ca current.

Effect of Acacetin on BK Ca Current
The effect of acacetin on BK Ca current was examined in HEK 293 cell line expressing human KCNMA1 gene. Voltage-dependent BK Ca current was recorded with the step voltage protocol as shown in the inset in a representative cell before and after application of acacetin ( Figure 7A). Acacetin had no effect on BK Ca at 3 and 10 µM, whereas it slightly increased the current at 30 µM. I-V relationships ( Figure 7B) of BK Ca current determined with a voltage ramp showed a similar response to acacetin. Acacetin did not affect the current at 10 µM, but slightly increased the outward component of BK Ca current at 30 µM. The BK Ca inhibitor paxilline (1 µM) almost fully suppressed the current. The percent values of BK Ca current at +70 mV illustrated in Figure 7C show that no significant effect of acacetin was observed at 3 and 10 µM, whereas 30 µM acacetin increased the current to 108.1 ± 5.7% of control (n = 7, P < 0.05 vs. control). These results suggest that acacetin may stimulate BK Ca channel at high concentration of 30 µM.

DISCUSSION
The present study provides the novel information that the natural flavone acacetin blocks the three SK Ca channel subtypes: SK Ca 1, SK Ca 2, and SK Ca 3, stably expressed in HEK 293 cells with similar efficacy. The IC 50 values of acacetin for inhibiting SK Ca 1, SK Ca 2, and SK Ca 3 are 12.4, 10.8, and 11.6 µM, respectively. Point mutagenesis of SK Ca 3 channel reveals that acacetin mainly interacts with H490 and S512 in the P-loop helix of the channel. However, acacetin at a high concentration of 30 µM induces only a small decrease in IK Ca channel and a small increase in BK Ca channel stably expressed in HEK 293 cells. The very limited effect of acacetin on IK Ca channel is similar to that reported previously for other SK Ca channel blockers (Girault et al., 2011).
An earlier study demonstrated that SK Ca channels were expressed in rat skeletal muscles, and sensitive to blocking by apamin (Blatz and Magleby, 1986). Then, the sequence of the transmembrane segments of SK Ca 1, SK Ca 2, and SK Ca 3 are found 80-90% identical (Kohler et al., 1996). However, the three subunits have different sensitivity to blocking by apamin (SK Ca 2 > SK Ca 1 > SK Ca 3), and are highly conserved among mammalian species, and are identified in many organisms from Drosophila to humans (Adelman et al., 2012). SK Ca subunits assemble to form homomeric (Kohler et al., 1996) or heteromeric (Tuteja et al., 2010) tetramers. SK Ca channels are identified in human and mouse atrial myocytes (Tuteja et al., 2005;Skibsbye et al., 2014), neurons (Church et al., 2015), and tumor cells (Girault et al., 2011(Girault et al., , 2012Gueguinou et al., 2014).
In the heart, activation of SK Ca channel may be antiarrhythmic or proarrhythmic, depending on the myocardial pathophysiological conditions (Chang and Chen, 2015). Chua et al. (2011) reported that the SK Ca channel current was heterogeneously upregulated in failing rabbit ventricles and SK Ca blocker apamin suppressed post-shock shortening of action potential duration in the failing hearts with ventricular fibrillation. On the other hand, apamin induced ventricular arrhythmias in slowly paced failing rabbit ventricles . The proarrhythmic effect was also observed with apamin in isolated normal canine left atrium .
However, the results from other groups demonstrated that blockade of SK Ca channels prolongs atrial effective refractory period, and SK Ca channels are therefore considered as a promising therapeutic target in the treatment of AF (Diness et al., 2010Qi et al., 2014;Haugaard et al., 2015). Several SK Ca channel blockers, e.g., NS8593, UCL1684, N-(pyridin-2-yl)-4-(pyridin-2-yl)thiazol-2-amine (ICA) and apamin, have been used for anti-AF studies (Diness et al., 2010(Diness et al., , 2015. In perfused guinea pig hearts, NS8593, UCL1684, and ICA effectively terminated AF induced with a combination of acetylcholine with electric stimulation (Diness et al., 2010). Injection of NS8593, UCL1684 or apamin reduced the duration of pacing-induced AF in vivo rat model . UCL1684 and NS8593 had significant anti-AF effect in a rat paroxysmal AF with hypertension-induced atrial remodeling . Interestingly, in large animals such as dogs (Qi et al., 2014) and horses (Haugaard et al., 2015), intravenous administration of NS8593 terminated all induced AF episodes, increased atrial effective refractory period, and decreased AF duration and vulnerability without QTc interval prolongation, suggesting that SK Ca channel blockers can be considered as promising anti-AF drugs. Moreover, recent studies showed that acute myocardial infarction might activate SK Ca channels, and apamin, UCL-1684 or ICA reduced ventricular burden arrhythmia by prolonging ventricular action potential duration and effective refractory period in rats with acute myocardial infarction (Gui et al., 2013;Hundahl et al., 2017).
In this study, we demonstrated that acacetin inhibited SK Ca 1, SK Ca 2, and SK Ca 3 channels in HEK 293 cell line expressing the corresponding genes. The blockade of SK Ca channels by acacetin likely also contributes to the anti-AF effect observed in canine models in addition to blocking I Kur /Kv1.5, I to /Kv4.3, and I K.ACh (Li et al., 2008;Wu et al., 2011Wu et al., , 2013a. These studies suggest that acacetin blocks multiple atrial-selective channels, and would be more effective in anti-AF than the blocker that specifically inhibits one type of atrial channel. However, whether acacetin, as apamin and other SK Ca blockers, is effective in improving learning and memory (Adelman et al., 2012) remains to be studied in the future. Moreover, additional studies are required for clarifying whether the SK Ca blocking effect of acacetin is related to its anti-cancer effect (Salimi et al., 2016;Zeng et al., 2017).
In our previous reports, we demonstrated that acacetin blocked Kv1.5 channel by binding to both its resting and open states by interacting with V505, I508, and V512 within the S6 domain (Wu et al., 2011), and inhibited the closed channel and blocked the open state of Kv4.3 by binding to both P-loop selectivity filter and S6 domain (Wu et al., 2013a). In the present study, we found that H490 and S512, but not H521, of P-loop helix are the binding sites of acacetin for blocking SK Ca 3 channel. The pore blocking of SK Ca 3 by acacetin is applicable to SK Ca 1 and SK Ca 2, because SK Ca 1, SK Ca 2, and SK Ca 3 share the same sequence in the range of pore helix 1 . This differs from the molecular determinants of acacetin for blocking Kv1.5 or Kv4.3 channel. On the other hand, the acacetin blockade of SK Ca 3 channel is different from the organic SK Ca blocker NS8593 and the archetypical peptide SK Ca blocker apamin. NS8593 interacts with S507 of P-loop helix and A532 of S6 domain (Jenkins et al., 2011), while apamin binds to a residue of S3-S4 extracellular loop of outside pore of the channel to produce a high-sensitivity block without selectivity filter contact (Weatherall et al., 2011). While the binding sites of various SK Ca blockers differ, it is important to develop these potential blockers into feasible drug candidates for future clinical application. A water soluble prodrug of acacetin has been developed, which can be intravenously administered for future clinical application (Liu et al., 2016).
Acacetin showed increased BK Ca current at concentration of 30 µM. Although the concentration for activating BK Ca channel is greater than those of blocking I Kur /Kv1.5, I K.ACh , I to /Kv4.3, and also SK Ca channels; this effect may account in part for the vascular dilation reported in a previous study (Calderone et al., 2004).
A limitation of the present study was that all the experiments were conducted only in HEK 293 line expressing SK Ca 1, SK Ca 2, or SK Ca 3 channels and lack of data from native cardiomyocytes. However, this does not affect the conclusion that acacetin blocks SK Ca channels. Future effort is required to obtain the data for the effect of acacetin on SK Ca current in native cardiomyocytes from an animal species whose heart has no or less expression of I Kur /Kv1.5 and I to /Kv4.3, because acacetin also inhibits these currents in native human atrial myocytes (Li et al., 2008).
Collectively, the present study demonstrates for the first time that acacetin is a SK Ca channel blocker and inhibits three subtypes of the SK Ca channels stably expressed in HEK 293 cells. The SK Ca channel blocking effect may be involved in its anti-AF property previously observed in experimentally induced AF in dogs.

AUTHOR CONTRIBUTIONS
K-HC, M-WJ, G-SX, YW, and G-RL conceived and designed the project. K-HC, HL, and H-YS conducted the experiments. K-HC, HL, H-YS, and G-RL analyzed the data. K-HC and G-RL prepared the manuscript. All authors read and approved the manuscript.

ACKNOWLEDGMENTS
This work was supported in part by a grant (ITS/339/09) from Innovation and Technology Commission of the Hong Kong SAR Government, China, a Seeding Fund from the University of Hong Kong, and a Key Cardiovascular Laboratory Fund (3502Z20150050) from Department of Xiamen Science and Technology, Xiamen, China