Human venous blood from anonymized healthy donors was obtained from a blood bank. The peripheral blood lymphocytes (PBLs) were isolated through Histopaque1077 (Sigma-Aldrich Hungary, Budapest, Hungary) density gradient centrifugation. Cells obtained were resuspended in RPMI 1640 medium containing 10% fetal calf serum (Sigma-Aldrich), 100 μg/mL penicillin, 100 μg/mL streptomycin, and 2 mM L-glutamine, seeded in a 24-well culture plate at a density of 5 × 10⁵ cells per mL, and grown in a 5% CO₂ incubator at 37°C for 2–5 days. Phytohemagglutinin A (PHA, Sigma-Aldrich) was added in 5, 7 or 10 μg/mL concentrations to the medium to boost the K⁺ channel expression.

Chinese hamster ovary (CHO) cells (gift from Yosef Yarden, Weizmann Institute of Science, Rehovot, Israel) were maintained by culturing in Dulbecco’s modified Eagle medium (DMEM, Gibco) supplemented with 2 mM L-glutamine, 10% FBS, 100 μg/mL streptomycin and 100 U/mL penicillin-G (Sigma-Aldrich) at a density of 0.5–1 × 10⁶ cells per mL in a humidified incubator at 37°C and 5% CO₂. Cells were passaged 3 times in a week following a 2–5 min incubation in 0.05% trypsin-EDTA solution at 37°C. Cultures were used up to passage number 20. PCR-based tests were routinely used to detect mycoplasma infection, only mycoplasma-free cultures were used for experiments.

CHO cells that do not express endogenous voltage-gated ion currents (Voros et al., 2018) were transiently transfected with the following plasmids encoding hKCa3.1 and turboGFP in a pCMV6-AC-GFP (OriGene Technologies) vector; H192A-hKCa3.1 in a pEGFP-C1 vector (a kind gift from Bernard Attali, Tel Aviv University, Israel); hKCa2.2 in a pCDN3 plasmid (a kind gift from Bernard Attali, Tel Aviv University, Israel) and T79D-rCaM in a pcDNA3 plasmid (a kind gift from Bernard Attali, Tel Aviv University, Israel). TurboGFP is a modified version of ppluGFP2, derived from Pontellina plumate, and is characterized by a fluorescence up to three times higher than EGFP (Evdokimov et al., 2006). Transfections were performed using the Lipofectamine 2000 kit (Invitrogen, Carlsbad, CA) following the manufacturer’s protocol. The cells were grown under standard conditions (see CHO cell culturing). GFP-positive transfectants were identified using Nikon TMS fluorescence microscope (Nikon, Tokyo, Japan), and currents were recorded 24–48 h post transfection.

Electrophysiology measurements were carried out using the patch-clamp technique in voltage-clamp mode. Whole-cell currents were recorded from peripheral blood lymphocytes and transfected CHO cells using a Multiclamp 700B amplifier connected to a DigiData 1440A digitizer (Molecular Devices, Sunnyvale, CA, United States). Micropipettes were pulled from GC 150 F-15 borosilicate capillaries (Harvard Apparatus Kent, UK) resulting in 3–5 MΩ resistance in the bath solution. Current traces were lowpass-filtered through the built-in analog 4-pole Bessel filters of the amplifiers and sampling frequency was set at least twice the filter cutoff frequency. Recordings were carried out at room temperature (20°C–25°C), The patch-clamped cell was perfused with control and test solutions using a gravity-driven custom-built perfusion system and excess bath solution was removed constantly by vacuum suction. The cells were held at −85 mV holding potential to minimize the holding current and allow the calculation of the leak conductance. 150-ms-long voltage ramps (from −120 to +50 mV, in 150 ms) were applied every 10 s to evoke the KCa3.1 currents. The pClamp 10.5, 10.7, and 11.2 software packages were used to acquire the data.

For offline leak correction and the calculation of the K⁺ conductance a Python based custom written program was used (available upon request). The algorithm averaged the holding current (I_hold) for n = 600 data points at −85 mV, which is close to the K⁺ equilibrium potential calculated from the Nernst equation (E_K = −89 mV). I_hold was considered as leak current and used to calculate G_leak as I_hold/–85 mV. Every data point was corrected for linear leak using I_K = I_m – (G_leak × E_m), where I_K is the leak subtracted K⁺ current, I_m is the measured membrane current at E_m membrane potential. I_K was then displayed as a function of E_m during the ramp and the region for linear current-voltage relationship was selected and G_K was calculated as the slope of the straight line fitted to the I_K data points (G_K = ΔI_K/ΔV). The linear region was typically between −120 mV and −60 mV for lymphocytes, at more depolarized membrane potentials the activation of the Kv1.3 current caused a significant deflection from the linear I-V relationship.

The following parameters were derived from the leak subtracted G_K data:

G_K,200/G_K,0: ratio of the K⁺ conductances determined at a time point t >200 s (G_K,200) in S-ECS over G_K at the beginning of the experiment (0 s) in S-ECS (G_K,0) in the absence of KCa3.1 activators; G_K,x/G_K,7.4: normalized conductance, G_K values determined in various pH_e solutions (G_K,x, where x is the pH_e) in a given cell normalized to G_K,7.4 recorded at pH_e = 7.4 in the same cell; G_K,act/G_K: “fold increase in conductance,” K⁺ conductance measured in the presence of the activator at a given concentration (G_K,act) divided by the K⁺ conductance in the drug-free solution (G_K) “fold increase normalized to 7.4”: fold increase in conductance” caused by the activator at a given pH_i and pH_e combination normalized to the “fold increase in conductance” measured with the activators in S- ECS at pH_e = 7.4.

G_K,end/G_K,start: G_K,end and G_K,start are the activator-enhanced K⁺ conductance at the end of the experiment (≥800 s) and the activator-enhanced K⁺ conductance at the beginning of the experiment, respectively.

The activator concentration-response curves were fit using: fold increase in conductance = Bottom + T o p − B o t t o m × X n H X n H + E C 50 n H where “fold increase in conductance” is defined above, [X] is the activator concentration, EC₅₀ is the concentration of agonist that gives a response halfway between Bottom (min value of “fold increase in the conductance”) and Top (maximum value of the “fold increase in conductance”), and n_H is the Hill slope. The Clampfit 10.7 and 11.1 software packages (Molecular Devices Inc., Sunnyvale, CA) were used to analyze the data and subtract the leak currents in the pilot experiments. Statistical analyses were performed with GraphPad Prism 8.4.3 (GraphPad Software, Inc., San Diego, CA).

The standard extracellular solution (S-ECS) was a Na⁺-aspartate (Na⁺Asp⁻)-based solution with 2.5 mM CaCl₂, and 10 mM HEPES titrated to pH_e = 7.4 with NaOH. The extracellular solution having pH_e = 8.0 (8.0-ECS) was buffered with 10 mM HEPES whereas those having pH_e = 6.0 (6.0-ECS); pH_e = 6.5 (6.5-ECS) or pH_e = 6.9 (6.9-ECS) were buffered using 10 mM MES, pH was titrated to the desired value with NaOH (see Table 1 for the composition of the solutions). The standard intracellular solution (S-ICS) was a K⁺-aspartate (K⁺Asp⁻)-based solution with 8.5 mM CaCl₂, 10 mM EGTA and 10 mM HEPES (titrated to pH_i = 7.2 with Tris). This solution has an estimated free Ca²⁺ concentration of ∼1-2 μM based on the MaxChelator program WEBMAX-C software (C. Patton, Stanford University, retrieved here: https://somapp.ucdmc.ucdavis.edu/pharmacology/bers/maxchelator/webmaxc/webmaxcE.htm). Since pH has a strong effect on the affinity of EGTA for Ca²⁺ we substituted EGTA with BAPTA, a less pH-sensitive Ca²⁺ chelator (Bers et al., 2010), when the pH of the intracellular solution was titrated to various levels. To keep the free Ca²⁺ concentration around ∼1-2 μM the pipette-filing solution titrated to pH_i = 6.5 (6.5-ICS) contained 11 mM BAPTA whereas the one titrated to pH_i = 8.0 (8.0-ICS) contained 10 mM BAPTA. For the experiments in which we used the KCa3.1 activators at pH = 7.2 we set the free Ca²⁺ concentration in the pipette-filing solution to ∼250 nM (7.2-ICS-250), as suggested by Jenkins et al. (2013) and Ca²⁺ was buffered using EGTA (10 mM EGTA and 5.7 mM CaCl₂). When the pH of the pipette-filling solution, with 250 nM Ca²⁺ concentration, was set to pH_i = 6.5 (6.5-ICS-250) or pH_i = 8.0 (8.0-ICS-250) the amount of BAPTA was set to 10 mM, but the concentration of CaCl₂ was adjusted to reflect the pH-dependence of the buffer capacity of BAPTA. More precisely, in case of pH_i = 6.5 we used 3.25 mM CaCl₂ and for pH_i = 8.0 we used 4.3 mM. All pipette-filling solutions were titrated with Tris. All titrations were done at 25°C and the pH of solutions was checked before every experiment.

All salts and components of the solutions were purchased from Sigma-Aldrich Budapest, Hungary. The KCa3.1 activators Riluzole (6-(trifluoromethoxy) benzothiazol-2-amine) and SKA-31 (Naphtho [1,2-d]thiazol-2-ylamine) and the blocker TRAM-34 (1-[(2-chlorophenyl)-di (phenyl) methyl] pyrazole), a gentle gift by prof. Heike Wulff, were kept in DMSO at a stock concentration of 5–10 mM and suitably diluted in the extracellular solutions (see Table 1) as needed.

The number of KCa3.1 channels in resting T cells is ∼10 (Cahalan and Chandy, 2009) and considering the ∼11–40 pS conductance of the channel (Aldrich et al., 2023), the magnitude of the whole-cell current is too small for precise pharmacological experiments. The expression of KCa3.1 is transcriptionally upregulated when T cells are stimulated with mitogens (Ghanshani et al., 2000; Wulf and Schwab, 2002). We took advantage of this and stimulated T cell proliferation using phytohemagglutinin A (PHA) to increase KCa3.1 expression. In addition to PHA activation of the T cells we also used 1 µM Ca²⁺ concentration in the pipette filling solution to fully activate the KCa3.1 current (Sforna et al., 2018). PHA-treatment also upregulates the expression of the voltage-gated Kv1.3 K⁺ channel. One can record isolated KCa3.1 currents in activated T cells using either the pharmacological separation, i.e., blockage of the Kv1.3 current using peptide toxins (Varga et al., 2021), or analyze the whole-cell currents below the activation threshold of Kv1.3. We chose the second scenario to avoid the use of multiple pharmacological tools when KCa3.1 activators were studied. The records in Figure 1A were obtained in a whole-cell patch-clamped T cell at several extracellular pH values during voltage ramps ranging from −120 mV to +50 mV. The records can be split on the voltage axis into two regions: a linear part between −120 mV and −50 mV and a very strongly outwardly rectifying part corresponding to the voltage-dependent activation of Kv1.3 at membrane potentials more positive than −50 mV (Panyi et al., 2006). Consequently, we have assigned the currents to KCa3.1 in the region between −120 mV and −60 mV (see inset to Figure 1A). The inset shows clearly that the current-voltage relationship is linear in this range and thus, the slope of the straight lines fitted to the highlighted part of the traces (inset) were used to determine the KCa3.1-specific K⁺ conductance and were denoted as G_K thereafter (calculated as G_K = ΔI/ΔV). The reversal potential (E_rev) of the leak-corrected currents is between −75 mV and −100 mV which indicates the K⁺ selectivity of the current (theoretical E_rev of a K⁺ selective conductance calculated from the Nernst equation is E_K = −89 mV). The reversal potential of the current being close to the theoretical equilibrium potential for K⁺ is a strong indicator for the quality of the records (i.e., small contamination from leak). Thus, we used G_K to assess quantitatively the magnitude of the KCa3.1 current as the K⁺ gradient remained constant throughout the experiments. Moreover, we used TRAM-34 to confirm that the current between −120 mV and −60 mV is KCa3.1 (Supplementary Figure S1) The slope of the current is reduced substantially upon perfusion the recording chamber with S-ECS containing 20 nM TRAM-34, which is a high affinity and specific inhibitor of KCa3.1 (IC₅₀ = 20 nM, (Wulff et al., 2000), Supplementary Figure S1A, inset). The G_K determined from the slopes decreased gradually following the start of the perfusion (Supplementary Figure S1B) with TRAM-34. We used TRAM-34 (20 nM) routinely at the end of the experiments to confirm that the currents between −120 mV and −60 mV were KCa3.1-specific.

Figure 1A also shows that the activation threshold of Kv1.3 becomes progressively more positive as the extracellular pH decreases from pH_e = 8.0 to pH_e = 6.0. This is consistent with the pH_e-dependence of the activation threshold for Kv1.3 reported earlier (Deutsch and Lee, 1989). Figure 1A and the inset shows qualitatively that the KCa3.1 currents recorded at pH_e values ranging from pH_e = 8.0 to pH_e = 6.0 are superimposable when the S-ICS (pH = 7.2) was used as the pipette filling solution. Figure 1B shows the G_K obtained at different pH_e values with S-ICS in the pipette (in the same experiment as in A). The G_K values were consecutively determined when the recording chamber was perfused with extracellular solutions having five different pH_e values. The perfusion started with the pH_e = 7.4 solution (S-ECS) as the reference solution and then sequentially switched to a different pH_e solutions, as indicated, and back to the pH_e = 7.4 solution. Figure 1B indicates that changing of pH_e alters minimally but fully and readily reversibly G_K. Data in Figure 1A was obtained using the standard pipette-filling solution of pH_i = 7.2 (S-ICS). The same set of experiments as in Figures 1A, B were repeated using pipette-filling solution having pH_i = 8.0 (8.0-ICS) and pH_i = 6.5 (6.5-ICS) (Figures 1C–F). The results were essentially the same as at pH_i = 7.2: the raw currents below the activation threshold of Kv1.3 and recorded in different pH_e solutions are virtually superimposable (Figures 1C, E, and insets), the currents reverse negative to −75 mV and the G_K values are minimally sensitive to changing pH_e of the extracellular solution (Figures 1D, F). Moreover, G_K values at pH_e = 7.4 are relatively constant throughout a given experiment, regardless of the pH_i. Taking the ratio of the K⁺ conductances determined at a time point t > 200 s (G_K,200) in S-ECS over G_K at the beginning of the experiment (0 s) in S-ECS (G_K,0) resulted in G_K,200/G_K,0 ∼1 thereby indicating the stability of the whole-cell hKCa3.1 currents (Figure 1G). The statistical analysis of the G_K values at all pH_e and pH_i combinations used in this study is in Figure 1H. G_K values determined in various pH_e solutions (G_K,x) in a given cell were normalized to the ones recorded at pH_e = 7.4 in the same cell (normalized conductance = G_K,x/G_K,7.4) to allow the comparison of the data obtained in different cells. The bar graph in Figure 1H shows that at pH_i = 7.2 and pH_i = 8.0 there is a small, but significant decrease in the normalized conductance when the extracellular pH was changed to acidic ones, whereas the conductances were essentially the same when pH_i was 6.5 regardless of the pH_e.

We repeated the same set of experiments as in Figure 1. Using cells transfected with the hKCa3.1 gene with the following motivations: 1) transfection can be optimized to increase KCa3.1 conductance therefore minimize the effect of leak on the whole-cell K⁺ conductance and 2) CHO cells do not express Kv1.3 (Yu and Kerchner, 1998), therefore one can minimize the contamination of the data by other K⁺ conductances. These become critical when the effect of the activators was studied (see below). Moreover, CHO can be transfected with mutant KCa3.1 and Calmodulin constructs (see below). Figure 2A shows the K⁺ current recorded in a CHO cell transfected with the hKCa3.1 gene that contains a turboGFP tag for the identification of the transfectants. Upon obtaining whole-cell configuration voltage ramps were applied and robust KCa3.1 currents (several nA at +50 mV) were recorded regardless of the extracellular pH. The current-voltage relationships are linear for voltages up to 0 mV and the reversal potential of the currents is more negative than −75 mV. Similarly to human lymphocytes, KCa3.1 currents recorded at pH_e values ranging from pH_e = 8.0 to pH_e = 6.0 in CHO cells are superimposable when the S-ICS (pH = 7.2) was used as the pipette filling solution. Figure 2B shows G_K obtained at different pH_e values (same experiment as in Figure 2A). The G_K values were consecutively determined when the recording chamber was perfused with extracellular solutions having five different pH_e values. Figure 2B indicates that changing of pH_e alters minimally but fully and readily reversibly G_K. Moreover, G_K determined at pH_e = 7.4 is constant during the >200 s duration of the experiment. The same set of experiments as in Figures 2A, B were repeated using pipette-filling solution having pH_i = 8.0 (8.0-ICS, Figures 2C, D) and pH_i = 6.5 (6.5-ICS, Figures 2E, F). The current density measured at +50 mV was insensitive to the pH_i (Supplementary Figure S2). Figure 2G allows us to draw a similar conclusion to what was obtained for human peripheral blood lymphocytes, i.e., G_K values at pH_e = 7.4 are constant throughout a given experiment, regardless of the pH_i. The statistical analysis of the G_K values at all pH_e and pH_i combinations were performed the same way as for human peripheral blood lymphocytes (see details above). The normalized conductance (G_K,x/G_K,7.4) values in the bar graph in Figure 2H show that at some pH_i and pH_e combinations there was a small, but significant decrease in the normalized conductance.

Riluzole and its more potent derivative SKA-31 (Sankaranarayanan et al., 2009) are positive modulators of the KCa3.1 currents in micro- and nanomolar-range concentrations, respectively. However, the pH-dependence of the action of these compounds has not been elucidated. To ensure that the applied concentration of the activators is not saturating (i.e., either loss or increase in their potency can be measured at different pH_i-pH_e combinations) we have experimentally determined the EC₅₀ values for Riluzole and SKA-31 in CHO cells. As the potentiation of the KCa3.1 current by the activators is more pronounced at low cytosolic Ca²⁺ concentrations as compared to ∼1 µM Ca²⁺ concentration used above (Figures 1, 2), we used pipette filling solutions having 250 nM Ca²⁺ concentration (7.2-ICS-250) in this set of experiments (Jenkins et al., 2013). The KCa3.1 currents in lymphocytes in 7.2-ICS-250 are very small and this means that the K⁺ conductance determined from the slopes of the voltage-ramps can be contaminated by leak. This becomes critical since the potency of the activators is expressed as fold change in the KCa3.1 conductance (Jenkins et al., 2013), where G_K in the absence of the activator is in the denominator. To overcome this, especially at low activator concentrations, we chose to obtain the EC₅₀ values in CHO overexpressing KCa3.1 where G_K in the absence of the activators can be determined more precisely than in human peripheral blood lymphocytes.

Figure 3 shows representative activator concentration-response experiments for SKA-31 (Figure 3A) and for Riluzole (Figure 3C). The G_K values were determined from voltage-ramp experiments in control extracellular solution (S-ECS) and S-ECS solutions containing the activators in the indicated concentrations. Figures 3A, C show, especially at higher activator concentrations, that the K⁺ conductance increases drastically and rapidly upon perfusing the recording chamber with the activators and declines with rapid kinetics when the chamber is perfused with S-ECS. The activator concentration-response curves (Figures 3B, D) were constructed by calculating the “fold increase in conductance” by dividing the K⁺ conductance measured in the presence of the activator at a given concentration (G_K,act) with the K⁺ conductance in the drug-free solution (G_K), i.e., G_K,act/G_K. The best-fit Hill equations to the data points obtained for individual plots (Figures 3B, D) resulted in EC₅₀ = 570 ± 101 nM (n = 6) for SKA-31 and 6.0 ± 1.3 μM (n = 6) for Riluzole (Figure 3E). Based on the EC₅₀ values we chose 1 µM SKA-31 and 5 µM Riluzole concentrations where the fold increase in the G_K was 52.3 ± 10.4-fold (n = 8) for SKA-31 and between 4.5 (2 µM) and 23-fold (10 µM) for Riluzole. The highest concentration of SKA-31 tested in our experiments (4 µM) caused 79.3 ± 17.5-fold increase in G_K (n = 6), the same parameter for 30 µM Riluzole was 37.4 ± 17.9-fold increase (n = 5).

To study the pH_e-dependence of the effect of the KCa3.1 activators, Riluzole (5 µM) or SKA-31 (1 µM) were added to the extracellular solutions with pH_e ranging from 6.0 to 8.0. The intracellular pH in whole-cell patch-clamped PBLs was buffered at pH_i = 7.4, pH_i = 8.0 or pH_i = 6.5 using appropriate ICS solution (Table 1) and the intracellular free Ca²⁺ was maintained at 250 nM. This Ca²⁺ concentration allows robust activator effects to be measured (Jenkins et al., 2013). The raw current traces evoked by voltage ramps in Figure 4A; Supplementary Figure S3A show that the baseline KCa3.1 currents below the activation threshold of Kv1.3 are almost negligible using the 7.2-ICS-250 pipette filling solution containing 250 nM free Ca²⁺. Upon perfusing the recording chamber with the activators in S-ECS the slope of the current-voltage relationship increased drastically. Moreover, the slopes of the current traces recorded in the presence of the activators at various pH_e values are superimposable. The right-hand side panels show thea KCa3.1-specific K⁺ conductance values determined by the slope of the straight lines fitted to the current traces below the activation threshold of Kv1.3. The KCa3.1 conductance is instantaneously increased upon perfusion of either 1 µM SKA-31 (Figures 4B, D) or 5 µM Riluzole (Supplementary Figure S3B, D), regardless of the pH_e of the extracellular solution and the washout of the activators is completed in 2–4 episodes corresponding to 20–40 s. The reversible effect of the activators can be repeated over extended periods of time without significant loss in the KCa3.1 conductance (Figure 4E, left side). Qualitatively similar results were obtained when the 8.0-ICS-250 pipette filling solution was used (Figures 4C, D and Supplementary Figure S3C, D): the KCa3.1-specific K⁺ conductance increased remarkably in the presence of both Riluzole and SKA-31, the magnitude of the conductance was similar in all pH_e solutions and the reversible effect of the activators could be repeated over extended periods of time (Figure 4E, right side). The quantitative analysis in Figures 4F, G supported the above-mentioned observations. To compare the effect of the activators we first calculated the “fold increase in conductance” (see e.g., Figure 3B) caused by the activator at a given pH_i and pH_e combination and normalized it to the “fold increase in conductance” measured with the activators in S-ECS at pH_e = 7.4. The “fold increase normalized to 7.4” values obtained this way scatter around 1 for both Riluzole (Figure 4F) and SKA-31 (Figure 4G), regardless of the pH_i (7.2 or 8.0) or the pH_e (8.0, 6.9, 6.5 or 6.0). In summary, these data indicate that the potency of Riluzole and SKA-31 in activating the KCa3.1 current is independent of the extracellular pH, measured either using 7.2-ICS-250 or 8.0-ICS-250 solutions, i.e., when the pH_i was 7.2 or 8.0.

The response of the KCa3.1 current to Riluzole and SKA-31 is completely different from the above when the pipette-filling solution was 6.5-ICS-250, i.e., the pH_i was 6.5. Figures 5A, B show representative K⁺ conductance values recorded upon activation of the KCa3.1 current by Riluzole and SKA-31, respectively. The potency of both Riluzole and SKA-31 become progressively lost during continuous application of either of the drugs. To characterize the loss of the potency of the modulators over time we used the G_K,end/G_K,start ratio where G_K,end and G_K,start are the modulator-enhanced K⁺ conductance at the end of the experiment (≥800 s) and the modulator-enhanced K⁺ conductance at the beginning of the experiment, respectively. The loss of the conductance was similar when 5 µM Riluzole or 1 µM SKA-31 were administered repeatedly and the phenomenon was independent of the pH_e, i.e., it progressed continuously at all pH_e values until saturating at ∼20% of the activator-evoked conductance at the beginning of the experiment (Figure 5C). The cells, however, did not respond uniformly to the activators: 2 out of 12 cells displayed constant activator potency over time.

We have demonstrated above that the characteristics of the potentiation of the KCa3.1 current by Riluzole and SKA-31 are similarly under the same experimental conditions (i.e., pH_e-pH_i combinations). Therefore, we restricted the subsequent experiments to the more potent and selective SKA-31 and repeated the same set of experiments in CHO cells expressing KCa3.1 as in PBLs (see Figures 4, 5). Figure 6A shows whole-cell currents recorded in a CHO cell using 7.2-ICS-250 pipette filling solution. The current in the absence of SKA-31 (dark-toned colors) increases robustly upon perfusing the cells with solutions of different pH_e and supplemented with 1 µM SKA-31 (lighter-toned colors). The currents recorded in various pH_e values are superimposable, i.e., the potency of SKA-31 was the same regardless of the pH_e. This is demonstrated more clearly in Figure 6B where the time-course of the experiment is displayed. The K⁺ conductance, determined from the slope of the current-voltage relationships, increases rapidly upon starting the perfusion with SKA-31-containing solutions. The effect of SKA-31 is quickly reversible upon perfusing the cell with activator-free solutions. The wash-in and wash-out of SKA-31could be repeated over an extended period without loss in the K⁺ conductance. Moreover, the K⁺ conductances measured in the presence of 1 µM SKA-31 were similar regardless of the pH_e. The properties of the potentiation of the KCa3.1 current by SKA-31 were qualitatively the same when the activator was applied to a cell patch-clamped using 8-ICS-250 pipette filling solution (Figures 6C, D). The bar graph in Figure 6E addresses if the K⁺ conductance could be increased by SKA-31 over an extended period without decline in the potency of the activator. The data shown in Figure 6E were obtained by calculating G_K,end/G_K,start ratio induced by 1 µM SKA-31 in S-ECS as described for Figure 5C. This ratio is ∼1 in Figure 6E which indicates that the potency of the activators remains constant throughout the experiment for both pH_i = 7.2 and pH_i = 8.0 pipette filling solutions. The observation that the effect of SKA-31 is independent of the pH_e of the recording solution is demonstrated quantitatively in Figure 6F. The “fold increase normalized to 7.4” variable (see above, description of Figure 4) values scatter around 1, like what was observed for PBLs (see Figure 4G). There was no statistical difference between this variable measured at various pH_e values when the pH_i of the pipette filling solution was either 7.2 or 8.0 (Figure 6F).

Similar to the results obtained in PBLs the KCa3.1 conductance current progressively decreased when SKA-31 was repeatedly administered to the same cell at pH_i = 6.5 (6.5-ICS-250) (Figure 7). Figure 7A shows that the progressive loss of the SKA-31-induced conductance continued at an apparently similar rate regardless of the pH_e of the extracellular solution. Moreover, the K⁺ conductance could not be recovered upon re-application of SKA-31 in the control S-ECS solution with pH_e = 7.4. This loss-of-potency phenotype was also observed when the extracellular solution was kept constant, i.e., S-ECS, and pH_i was slightly more alkaline (pH_i = 6.7, Figure 7B) or more acidic (pH_i = 6.2 Figure 7C). Similar to PBLs, the cells did not respond uniformly to SKA-31, 5 out of 30 cells displayed relatively constant K⁺ currents upon repeated administration of the activator. The frequency of the cells showing stable K⁺ conductances vs. decline upon repeated SKA-31 application was statistically the same in PBLs and CHO cells (p > 0.80, Chi-square test with Yates’ correction). The loss of the K⁺ conductance over the time-course of the experiment was characterized quantitatively using the G_K,end/G_K,start ratio as described for PBLs (see details in Figure 4E). The bar chart in Figure 7D shows that the cells showing loss-of-potency phenotype upon repeated SKA-31 application displayed a large scatter in G_K,end/G_K,start ratio with a median of 0.17.

To get more insight into the nature of loss-of-potency phenotype we studied the response of the currents to SKA-31 at a constant pH_e of 7.4 (S-ECS) and pH_i of 6.2 and varied the pattern of SKA-31 application. The first special protocol started with SKA-31 administration at the beginning of the experiment followed by a washout (Figure 8A). Thereafter the cell was repeatedly depolarized using voltage ramps every 10 s for 400 s in the absence of SKA-31. SKA-31 was reapplied after this period but failed to potentiate the current to the same extent as in its first application to this cell. The quantitative analysis of the drop in the potentiation of the K⁺ conductance is in Figure 8B. The G_K,end/G_K,start ratio (see above) is significantly smaller than 1 thereby indicating the decline in the potentiation of the K⁺ conductance by the end of the experiments. Thus, the loss-of-potency phenotype cannot be prevented by inserting a drug-free period into the protocol where voltage ramps are repeatedly applied. In the next special protocol (Figure 8C) the KCa3.1 conductance was first potentiated by SKA-31 application. Thereafter, while the SKA-31-containng solution was perfused constantly on the cell, the delivery of the voltage ramps was suspended for 360 s and the cell was kept at the holding potential of −85 mV during this period. The repeatedly applied voltage ramps restarted after the gap. As Figure 8C shows that the SKA-31-induced KCa3.1 conductance was much smaller after the gap in the recording than at the beginning of the experiment (Figure 8D). The G_K,end/G_K,start ratio (see above) is significantly smaller than 1 thereby confirming the loss-of-potency phenotype by the end of the experiments. Thus, the decline in the potency of SKA-31 can neither be prevented by inserting a voltage-ramp-free period into the protocol, nor by inserting a drug-free period into the protocol, indicating that the loss-of-potency phenotype must be attributed to the acidic pH_i.

KCa3.1 and the members of the KCa2.x family share several structural features: they are made up of four alpha subunits, each containing six transmembrane domains and a Calmodulin-binding domain in the C-terminal portion. KCa2.x channels are activated by intracellular Ca²⁺ and the same modulators as KCa3.1 (Guéguinou et al., 2014), albeit KCa2.x channels are ∼10 times less sensitive to molecules like Riluzole and SKA-31 than KCa3.1. We chose KCa2.2 as the representative channel of this family and assessed the ability of SKA-31, at a higher concentration of 5 μM, to activate the current at neutral and acidic intracellular pH conditions. At pH_i of 7.2 (Figure 9A) the robust activation of the KCa2.2 by SKA-31 can be repeatedly evoked with negligible loss in the modulator’s efficacy over time. On the contrary, when the intracellular solution was set at pH_i = 6.5 a rapid and irreversible loss of the potency of SKA-31 to activate KCa2.2 was observed (Figure 9C), similar to the findings for KCa3.1. The quantitative analysis of the data in Figures 9B, D show very similar results to the ones obtained in CHO cells transfected with hKCa3.1, i.e., there is a significant loss in the potency of SKA-31 to activate KCa2.2 at acidic intracellular pH (cfr. Figures 6E, 7D).

A recent cryo-EM derived structure of hKCa3.1 revealed a role of the S4-S5 linker in the formation of the functional and structural connection between Calmodulin and the C-terminal portion of the KCa channels (Lee and MacKinnon, 2018). Residue His192 is in this linker and faces the pocket where the activators are thought to exert their effect. Mutating this histidine to a non-charged alanin (H192A) caused the disruption of the interaction between BA6b9, a KCa3.1 blocker devised to be structurally similar to Riluzole/1-EBIO and KCa3.1 (Burg et al., 2022). Moreover, it is known that protonation of histidine residues in acidic environments can disrupt the ability of toxins to bind to ion channels, making the toxins less or non-functional (Aiyar et al., 1995). Considering the structural similarity between BA6b9 and the modulators used in this study and that they fit in overlapping binding pockets, we checked whether the H192A mutation would influence the activation of hKCa3.1 by SKA-31 in neutral and in acidic conditions. Figure 10A shows that SKA-31 activates the H192A-KCa3.1 current, in a reversible manner, similar to the wild type KCa3.1. The wash-in-wash-out cycles could be repeated several times when the intracellular solution was set to 7.2. The potency of SKA-31 in activating the current was relatively constant over extended periods of time (Figure 10B). However, at pH_i = 6.5 SKA-31 gradually lost its potency in activating the H192A-KCa3.1 current during repeated administration (Figure 10C, D). This suggests that the loss of SKA-31 potentiation of the current at acidic intracellular pH is oblivious to whether the protonated His or the neutral Ala is in position 192.

Calmodulin (CaM) is constitutively bound to KCa channels (Lee and MacKinnon, 2018) which, besides Ca²⁺ ions, also requires membrane-bound PIP₂ as co-agonist (Burg et al., 2022).Thr79 in CaM is the target of the Casein Kinase-2 (CK2) (Bildl et al., 2004). When Thr79 is phosphorylated, it results in the loss of the sensitivity of the KCa2.2 channel to PIP₂ (Zhang et al., 2014) and Ca²⁺ (Allen et al., 2007). The phosphorylation of Thr79 can be mimicked by the phosphomimetic mutation T79D, this mutation decreases the K⁺ current both in KCa2.2 (Zhang et al., 2014) and KCa3.1 (Burg et al., 2022). The disturbed network of activators and co-activators in the presence of T79D-CaM may be reflected in the modulation of KCa3.1 activation by SKA-31 at acidic pH_i. To test this, we co-transfected hKCa3.1 and T79D-CaM into CHO cells and studied the potentiation of the whole-cell current at neutral and acidic pH_i. Figures 11A, B show that using neutral pH_i condition the current could be activated by SKA-31 similar to what was obtained in cells transfected with hKCa3.1 only (see Figure 6). The activation cycles by SKA-31 resulted in consistently increased K⁺ conductance over extended time periods (Figure 11B). On the contrary, when the pH_i = 6.5 was used SKA-31 gradually lost its potency over time to activate G_K (Figures 11C, D). The average loss of the G_K by the end of the experiment (>800 s, Figure 11D) is slightly reduced as compared to when CHO cells were transfected with wild-type (Figure 7D) or H192A KCa3.1 constructs (Figure 10D), but the G_K,end/G_K,start ratio was non significantly different among the three groups (One-way ANOVA, p > 0.05). Moreover, some cells displayed a very slow restoration of the potency of SKA-31 over time (Figure 11C). This phenomenon was not investigated any further due to inherent limitations of the whole-cell patch-clamp over extended durations beyond 15–20 min.

KCa3.1 is extremely sensitive to the intracellular concentration of Ca²⁺, presenting an EC₅₀ ranging from 100 to 400 nM (Brown et al., 2019) and a characteristic sigmoid activation curve (Bailey et al., 2010). When SKA-31 was applied to hKCa3.1-expressing CHO and 1 μM Ca²⁺ concentration was used in the pipette at pH_i of 7.2 we found the expected: 1) an elevated base-line KCa3.1 conductance due to the higher intracellular Ca²⁺ (Figures 12A, B) and 2) a reduced potentiation of G_K by SKA-31 (∼2 fold vs. ∼50-fold at 250 nM cytosolic Ca²⁺ concentration, see Figure 3) due to near-saturation levels in the Ca²⁺ sensitivity of the channel. Interestingly, at 1 μM cytosolic Ca²⁺ concentration the potency of SKA-31 to upregulate KCa3.1 conductance remained constant even at acidic pH_i = 6.5 (Figure 12B). The G_K,end/G_K,start parameter obtained at pH_i = 6.5 and 1 μM Ca²⁺ did not differ statistically from the data obtained at pH_i = 7.2 at either 1 µM or 250 nM cytosolic Ca²⁺ concentration (Figure 12C). This means that at saturating concentration of intracellular Ca²⁺, the potency of SKA-31 in activating KCa3.1 remains constant regardless of the pH_i.