Cell culture and transfection protocols described in our previous work were adopted here with minimal modifications [36]. Briefly, human embryonic kidney 293 cells were grown in Dulbecco's Modified Eagle Medium (GIBCO; Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 1× antibiotic-antimycotic reagent (Life Technologies, Carlsbad, CA, USA) and 10% fetal bovine serum. Cells grown to a confluency of 80%–90% were transiently transfected using Lipofectamine 2000 (Thermo Fisher Scientific) transfection regent according to the manufacturer's specifications. 1 µg of cDNA and 3 µl of Lipofectamine each were separately mixed with 50 µl OptiMEM (Thermo Fisher Scientific) in two 1.5 ml tubes and incubated at RT for 5 min. Following the incubation, the two mixtures were mixed together by rapid trituration. The resulting DNA-Lipofectamine mixture was incubated for 20 min at RT and put into a selected well in a 12-well cell culture plate. Before adding the DNA-Lipofectamine mixture, the culture media in the 12-well cell culture plate was replaced with OptiMEM. Cells were seeded on to glass coverslips 24 h after the transfection and were used in patch clamping experiments 36–48 h after transfection. 0.45 µg pCDNA3-hSlo3 (NP_114016.1) was co-transfected with 0.45 µg hLRRC52 (NP_114016.1) and 0.1 µg of pCDNA3.1-GFP for Slo3 experiments.

Electrophysiological protocols and saline solutions described in our previous work are adopted in this work with slight modifications [36]. GC 150TF-7.5 Glass capillaries (Harvard Apparatus, Holliston, MA, USA) were pulled using a vertical pipette puller (PC-10; Narishige Group Products, Tokyo, Japan) with heat levels adjusted to 60.4 Ω in the first step and 50.0 Ω in the second step. These pipettes possessed a 3–4 MΩ resistance when filled with K-aspartate based pipette solutions described as follows. Responses of hSlo3 to internal pH was studied using the following pipette solutions: 130 mM K-aspartate, 10 mM NaCl, 1 mM ethylene glycol bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid (EGTA), 5 mM N-(2-hydroxyethyl)piperazine-N′-2-ethansulfonic acid (HEPES), 15 mM D-glucose (pH adjusted using KOH/HCl). The effect of internal free calcium on Slo3 was studied using the following pipette solutions: divalent-free solution; 130 mM K-aspartate, 10 mM NaCl, 1 mM EGTA, and 20 mM HEPES, (pH 7.3 using KOH); intracellular solutions with different free Ca²⁺ concentrations; 130 mM K-aspartate, 10 mM NaCl, 5 mM EGTA, XCaCl₂ and 20 mM HEPES; (pH 7.3 using KOH). Free calcium concentrations were calculated using the MaxChelator tool [37]. External bath saline solutions were as follows. Standard extracellular solution; 140 mM NaCl, 5 mM KCl, 10 mM HEPES 2 mM CaCl₂ and 1 mM MgCl₂ (pH adjusted to 7.4 using NaOH). Whole-cell voltage-clamp experiments were performed at RT using an Axopatch 200B capacitor-feedback patch-clamp amplifier (Molecular Devices, Sunnyvale, CA, USA) connected to a Digidata-1440A Digitizer (Molecular Devices). Voltage protocol used for step pulses was as follows. 400 ms long step pulses with +25 mV voltage increments were given from −125 mV with a one second interval at −100 mV holding potential (Fig. 1A). Continuous voltage ramps of +140 mV (400 ms long) every 10 sec were given with a holding potential of −100 mV to monitor continuous gradual changes in channel activity. Membrane potential was corrected for the liquid junction potential (LJP) by deducting the LJP from the measured pipette potential (V_P). LJP was calculated using the inbuilt LJP calculator in Pclamp 10.7.0.3 software (Molecular Devices).

Statistical analysis was performed according to the guidelines provided on experimental design and analysis in pharmacology [38]. Data acquired from Pclamp 10.7.0.3 software (Molecular Devices) were directly analyzed with Clampfit 10.7.0.3 software (Molecular Devices) or imported to Origin Pro 8.1 (OriginLab Corporation, Northampton, MA, USA) for further analysis. Generation of current voltage (IV) curves from step pulse data was done from IV plot tool in Clampfit. Other calculations and analysis were done using the integrated statistical analysis tools in Origin Pro 8.1 as described below. Conductance at different command voltages was measured by calculating the slope of the IV curves using a standard linear fit (y = mx + c). Calculated conductance values for each cell was normalized relative to the maximum conductance recorded in the same recording. The relative conductance values were plotted against voltage to calculate half activation voltage values using a Boltzman fit (y = A2 + (A1 − A2) / (1 + exp((x − x0) / dx)).

Drug responses shown as percentage relative current remaining were calculated by dividing the current at +125 mV after treatment (I_post) by the current before treatment of drug (I_pre) in each cell, to avoid unnecessary variations arising from cell to cell differences. Current at +125 mv was calculated using the inbuilt IV tool in clampfit software by averaging the current between 350–400 ms at the +125 mV command step. These calculated relative percentage current values were compared to control values using the Wilcoxon signed-rank test. Non-normalized paired data were compared using a paired sample t-test. Other non-paired data were compared with a one-way ANOVA followed by a post hoc Tukey's mean comparison test. Dose response data were fitted using the dose response fit model (y = A1 + (A2 − A1) / (1 + 10^{(LOGx0 − x)*p)}). All data are shown as mean ± standard error of the mean represented using error bars, followed by the sample size in brackets. A level of probability below 0.05 was used as the threshold probability for statistical significance.

The brown outer layers of the onions were extracted three times with ethanol under reflux for 3 h. The combined filtrate was concentrated using a rotary evaporator, freeze dried, and stored at −20℃ until further analysis. The HPLC profile of the onion extract and the compounds identified from it was standardized using a validated HPLC assay [21]. Stock solutions of OPE were made by dissolving them in the appropriate amount in DMSO to a final concentration of 100 mg/ml, which were then stored at −20℃. GF109203X (GFX), H-89 dihydrochloride, and wortmannin were purchased from Tocris bioscience (Bristol, UK). Quercetin, luteolin and morin were purchased from Cayman Chemical (Ann Arbor, MI, USA). Stock solutions of chemicals were made by dissolving the appropriate powder mass of each chemical in either DMSO or distilled water, which were then stored at −80℃ until further use. Working solutions of chemicals were freshly made each day prior to experiments by dissolving the stock chemical solutions in electrophysiological saline solution.

We first re-established a successful Slo3 expression system in HEK 293 cells for electrophysiological recording as previously reported [36] (Fig. 1A, B). The measured Slo3+LRRC52 current responded to classical Slo3 modulators such as progesterone and Tetraethylammonium (Fig. 1C). Furthermore, the measured Slo3 currents were sensitive to internal pH (Fig. 2F) but lacked the sensitivity to internal calcium (Fig. 3B) which is a previously been reported as a key feature of Slo3 channels that set them apart from calcium sensitive Slo1 channels [39]. After confirming the standard phenotype for Slo3 currents, we tested the effect of OPE on Slo3 currents under the same conditions. A concentration of 100 µg/ml OPE inhibited Slo3 and LRRC52 current from 300.45 ± 12.69 pA/pF to 156.75 ± 8.96 pA/PF (p < 0.001). Furthermore, the inhibition of Slo3+LRRC52 currents by OPE was dose dependent with an EC₅₀ of 95.75 µg/ml (Fig. 1F). This dose response could be fitted to a double-dose response curve as shown by the dotted lines (Fig. 1F). The reason for double-dose response fitting in Slo3 could be due to the use of crude OPE: while quercetin was the main active ingredient, the mixture contains several other chemicals [21]. Furthermore, the inhibition of Slo3 by OPE was reflected by the reversible depolarization in membrane potential upon exposure to OPE (Fig. 1G, H). Membrane potential depolarized from −23.03 ± 1.69 mV to −17.22 ± 1.17 mV upon OPE administration (p < 0.001). This inhibition of Slo3 by OPE decreased with intracellular alkalization (Fig. 1I). The voltage-dependent activation of Slo3 currents was shifted to the left with intracellular alkalization (Fig. 1J, K), and OPE administration shifted the voltage-dependent activation of Slo3 towards the right, meaning that the voltage required to activate Slo3 moved towards more depolarized potentials (Fig. 1I, J). The half activation voltage value shifted from 62.35 ± 2.18 mV to 69.97 ± 15.77 mV at pH 6.0, 45.04 ± 3.00 mV to 100.42 ± 10.87 mV at pH 7.3, 32.76 ± 3.17 mV to 77.41 ± 6.59 mV at pH 8.0 and 29.92 ± 5.49 mV to 95.62 ± 7.59 mV at pH 9.0.

Current measured in non-transfected cells and those transfected with Slo3 alone (without LRRC52) was negligible (p = 0.19 and p = 0.06 respectively) (Fig. 1D), displaying a minimal response to OPE. Similarly, the membrane potential of non-transfected cells slightly depolarized from −18.00 ± 0.51 mV to −17.00 ± 1.67 mV (p = 0.50) upon exposure to OPE (Fig. 1H). The combination of Slo3+LRRC52 yielded characteristic Slo3 currents, whereas the Slo3 alone cells and non-transfected HEK cells did not. Therefore, from here onwards we refer to current measured from Slo3+LRRC52 co-transfected cells as ‘Slo3 current’. Taken together these results indicated that OPE inhibits Slo3 currents in a dose-dependent and internal pH dependent manner which depolarizes the cell membrane.

Quercetin is the major active component of OPE, which is a mix of several other active ingredients [21]. In order to investigate the contribution of quercetin, we tested its effects on Slo3 currents in isolation, without the other components of OPE. Quercetin, similarly to OPE, inhibits Slo3 current amplitude in a dose dependent and pH dependent manner with an EC₅₀ value of 11.27 µM at internal pH 7.3 (Fig. 2A, B). As an example, quercetin inhibited Slo3 currents down to 33.17 ± 1.71% compared to current prior to exposure to queen at internal pH 6.0, whereas further pH increases to 7.3, 8.0 and finally to 9.0 (by to internal alkalization) gradually decreased the inhibitory effects of quercetin by increasing the level of reaming current to 40.78 ± 5.60%, 70.23 ± 4.04%, and 67.74 ± 2.69% respectively compared to the control (Fig. 2E). Unlike OPE however, quercetin did not alter the voltage dependence of Slo3 activation (Fig. 2E) indicating some involvement of other minor constituents of OPE in Slo3 regulation. The calculated half activation voltages from Boltzmann fitting of relative conductivity data in Fig. 2F at internal pH 6, 7.3, 8 and 9 were 59.30 ± 5.65 mV, 38.21 ± 2.58 mV, 28.40 ± 1.53 mV and 22.36 ± 2.23 mV before exposure to quercetin and 59.82 ± 4.04 mV, 28.62 ± 2.55 mV, 25.82 ± 2.66 mV and 11.52 ± 3.63 mV after exposure to quercetin, respectively (Fig. 2F, G). Furthermore, similarly to OPE, quercetin depolarized the cell membrane of Slo3+LRRC52 expressing cells (Fig. 2C, D). These results demonstrate that quercetin is the main component of OPE that inhibits Slo3 in a dose and pH dependent manner.

In sperm, internal pH change regulates calcium influx via the pH sensitive CatSper channel. Therefore, we wanted to test whether internal calcium variation has an effect on the inhibitory action of quercetin on Slo3. Interestingly, we found that inhibition of Slo3 by quercetin is dependent on the internal free calcium concentration (Fig. 3A). Increasing the internal free calcium concentration hinders the inhibitory effect of quercetin on Slo3: when internal free calcium was buffered by EGTA at levels of 0 µM, 20 µM and 200 µM, relative current remaining after quercetin treatment was 40.78 ± 5.60%, 60.45 ± 5.78% and 65.51 ± 4.79% respectively (Fig. 3A). Even though, inhibitory effects of quercetin was lowered due to increasing calcium, quercetin, however, did not change the voltage dependency of Slo3 at internal free calcium concentrations of 20 µM and 200 µM (Fig. 3B, C). Furthermore, increasing the internal free calcium had a minimal effect on the half activation voltage of Slo3 probably due to the fact that Slo3 lacks the calcium binding calcium bowl that is present in closely related calcium sensitive, Slo1 (BK) channels. Therefore, it is possible that the hindering of the inhibition of Slo3 by quercetin is mediated through a calcium dependent signaling pathway rather than by calcium directly acting on Slo3.

As the inhibitory efficacy of quercetin on Slo3 was strongly dependent on internal pH, we wanted to test whether the external pH would have any influence on inhibitory effect of quercetin on Slo3 particularly because the external pH in the female reproductive tract is acidic, whereas it is alkaline in the male reproductive tract. Similarly, to the internal pH, raising the external pH reduced the inhibitory effect of quercetin on Slo3 (Fig. 4A). However, unlike the internal pH, external pH changes did not affect Slo3 currents in quercetin untreated cells. Therefore, it's likely that pH has a direct effect on quercetin besides its direct effect on channel gating. Furthermore, we noticed that the color of 50 µM quercetin containing saline solutions changed when the pH of the saline was changed (data not shown) indicating alteration to moieties association to the quercetin flavonoid ring [40]. In order to investigate this, we tested the effects of luteolin and morin on Slo3. Luteolin and morin are analogues of quercetin which share a common flavonoid ring with slight variations in functional moieties (Fig. 4D). 50 µM luteolin, which has the highest pKa value of 7.05, along with four hydrogen bond donors, displayed a 53.47 ± 3.15% inhibition. On the other hand 50 µM morin, which has a pKa value of 6.88 (closer to quercetin's pKa value of 6.90) and a total of 5 hydrogen bond donors (similar to quercetin), inhibited Slo3 currents up to 42.15 ± 2.16% , which was not statistically different from that elicited by quercetin (Fig. 4B) [41]. The current-voltage relationship elicited by step pulses remained the same characteristic shape after exposure to the quercetin analogues (Fig. 4B). These results raised the question as to whether variations in the polyphenolic ring affect the function of quercetin on Slo3. As an example, quercetin's activity as a PI3 kinase inhibitor is strongly dependent on the moeities at the 2- and 3-positions of the polyphenolic ring [42]. Furthermore, allosteric regulation of Slo3 by phenolic rings contacting quinidine occurs via a hydrophobic binding site [43]. The hydrophobicity of quercetin can also be affected by alterations to patterns of substitution on the flavonoid ring. However, when we tested mutant Slo3 that lack the quinidine binding site, quercetin was still able to inhibit both Slo3 mutants and Slo3 wildtype to a similar degree (Supplementary Fig. 1).

Quercetin is a biologically active chemical that can inhibit both PKC and phosphoinositide 3-kinase (PI3K) activity [333442]. The stimulatory effects of quercetin on sperm Hv1 occur via the inhibition of PKC [21]. Furthermore, we found that the inhibitory effects of quercetin decrease with increasing calcium concentration (Fig. 3) indicating a possibility that the promotion of calcium dependent protein kinase activity C can negate the inhibitory effects of quercetin on Slo3. Therefore, we investigated whether Slo3 currents respond to protein kinase modulators. To do this, we tested the effects of protein kinase A inhibitor H-89 and protein kinase C inhibitor GFX on Slo3. Both protein kinase inhibitors were applied with external perfusion at 5 µM concentration, and both attenuated Slo3 currents (Fig. 5A, B). The relative current remaining was 48.27 ± 2.66% and 28.30 ± 4.40 after H-89 and GFX treatment respectively (Fig. 5C). However, introducing quercetin in the presence of H-89 or GFX further inhibited the Slo3 current to 47.72 ± 14.33% and 36.40 ± 12.57% respectively (Fig. 5D) indicating that PKA inhibition or PKA inhibition does not explain the inhibitory effect of quercetin on Slo3. Other than PKC inhibitory function, quercetin is known also as a phosphoinositide kinase inhibitor [3442]. The PI3K inhibitory action of quercetin is strongly dependent on moieties at the 2- and 3-positions of the polyphenolic ring which can be altered due to changes in pH [42]. Therefore, we tested whether Slo3 currents are inhibited by PI3K inhibitor wortmannin (25 µM) (Fig. 5F). The current remaining after wortmannin treatment was 95.14 ± 6.87%. Quercetin treatment after wortmannin treatment inhibited the Slo3 current to a level similar to non-wortmannin treated cells, indicating a minimal effect of PI3K inhibition on Slo3 function (Fig. 5H, I). Therefore, there is not enough evidence to conclude that the inhibitory effect of quercetin on Slo3 is due to PKC or PI3K inhibition.