Landing Pad Stable Cell Line Generation

Stable cell lines were introduced into the landing pad HEK293 cell line as described previously (Matreyek et al., 2017) with minor adjustments (Fig. 1A). Landing pad HEK293 cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% FBS, 100 U/ml penicillin, and 0.1 mg/ml streptomycin, and plated into 6-well plates 1 day before transfection. For cells in one well of the plate, 1.5 μg pCAG-NLS-Bxb1 and 1.5 μg attB recombination plasmid (containing the K_ATP channel subunits) were cotransfected with 6 μl Fugene 6. Two days after transfection, the medium was replaced by the culture media supplemented with doxycycline (4 μg/ml) and puromycin (10 mM). Death of cells without DNA recombination was observed within 2 days. The cells were then cultured for 2–3 weeks with occasional replacement of the culture medium. The established stable cell lines lost blue fluorescence and uniformly acquired red fluorescence under a fluorescent microscope (Fig. 1B, Supplemental Fig. 2).

Open in a separate window

Fig. 1.

Landing pad system for generation of stable cell lines. (A) Insertion of cDNAs containing Kir6.1 and SUR2B into the genome of Landing pad HEK293 cells. (B) Representative micrographs of untransfected Landing pad cells (top) and cells transfected with hKir6.1/hSUR2B channel constructs (bottom). 4′,6-Diamidino-2-phenylindole (DAPI)–filtered images show blue fluorescence in untransfected cells disappears in transfected cells, and red fluorescent protein filter shows appearance of red fluorescence uniformly in stably modified cells.

Thallium Flux Assay

This procedure is adapted from a previously published paper (Raphemot et al., 2013) with minor adjustment. One day before the assay, Landing pad HEK cells, WT, and mutant hKir6.1+hSUR2B stable cell lines were separately suspended in medium with doxycycline at a density of ~800–1000 k/ml, and 100 μl cell suspensions were plated into each well of a 96-well plate precoated with 0.01% poly-L-lysine. Approximately 2 hours before the assay, 5 μl Fluozin-2 (0.5 μg/μl in DMSO) was mixed with 2.5 μl Pluronic F-127/DMSO solution (20% w/v) and then dissolved in 5 ml Hanks’ balanced salt solution (HBSS) buffer forming the dye solution. Tl⁺ solution (1 mM) was prepared and added to two columns (300 μl/well) of another 96-well plate as compound source.

The medium in the 96-well plate was discarded and the cells were washed three times (“flick and slam”) with 80 μl HBSS for each well. Fluozin-2 dye solution (45 μl) was added to each well for 1 hour, after which the cells were washed another three times with HBSS. HBSS solution (160 μl) was then added to cover the cells, or only 80 μl if there were additional components needed for the assay. In this case, another 80 μl, with these components, was added immediately (4 minutes) prior to assay.

A Flexstation 3 fluorescent plate reader was used to conduct fluorescence measurements. Fluorescence mode was used for the whole assay with excitation wavelength of 485 nm and emission wavelength of 528 nm. The signal gain level was set at medium, any required additions for this assay were at t = –4 minutes, and then fluorescence was measured every 2 seconds for 2 minutes. At the 20 second timepoint, Tl⁺ solution was added (40 μl/well) and at a height equal to 150 μl liquid. Signals from replicate wells were averaged and plotted versus time for each experiment. Data were saved as Excel files for later calculation.

The recipe for HBSS is: NaCl 137 mM, KCl 5.4 mM, CaCl₂ 1.3 mM, MgSO₄ 0.4 mM, MgCl₂ 0.5 mM, Na₂HPO₄ 0.4 mM, KH₂PO₄ 0.4 mM, Glucose 5.6 mM, NaHCO₃ 4.2 mM, HEPES 20 mM. The recipe for 4× Tl⁺ solution is: CaSO₄ 1.8 mM, MgSO₄ 1 mM, glucose 5 mM, HEPES 10 mM, TlSO₄ 4 mM. 4× Tl⁺ was diluted with HBSS to get the final 1 mM thallium solution.

Thallium Influx Curve Fitting

For analysis (Fig. 2), signal intensity of wells without cells and dye was subtracted, and the data were then normalized (relative fluorescence units [RFU]) to the dye fluorescence prior to Tl⁺ addition. RFU versus time traces were fit with single exponential eq. 1 using Origin 8:

where y = RFU, A = maximum amplitude, t = time, τ = time constant, and y₀ = offset.

Open in a separate window

Fig. 2.

Principle of cell-based, fluorescence high-throughput screening assay. (A) Potassium channel-facilitated thallium influx is measured from the increase of Fluozin-2 fluorescence on binding to thallium. (B) Representative raw data traces of Landing pad only (L-p) cells (or blank wells, no cell) in the presence or absence of 0.2 mM Tl⁺ added at 20s. In (B)–(E), experiments were all carried out in the presence of 30 μM pinacidil. (C) Fluorescence changes with increasing [Tl⁺] (0, 0.1 or 0.2 mM as indicated) in Landing pad only cells, normalized to initial fluorescence after subtraction of the fixed fluorescence of empty wells. (D) Fluorescence changes with increasing [Tl⁺] (0, 0.1 or 0.2 mM as indicated) in cells transfected with WT hKir6.1 and hSUR2B, normalized to initial fluorescence after subtraction of the fixed fluorescence of empty wells. (E) Time-dependent fluorescence change (WT transfected cells with 0.2 mM Tl⁺, normalized as in D), was fit with a single exponential function, used to calculate initial rate of uptake.

The derivative function:

was calculated at t = 0 as the initial rate of uptake after obtaining value y₀ and τ.

Pinacidil and glibenclamide dose-response relationships were fitted with modified Hill equations:

where Grel = relative conductance, G₀ = conductance in zero drug, G_sat = conductance in saturating [drug], [Pin] and [Glib] are pinacidil or glibenclamide concentration, EC₅₀ and IC₅₀ are the half-maximally effective drug concentrations, and n = Hill coefficient using Origin 8. Because pinacidil activation curves resulted in both shifts of apparent EC₅₀ to lower values and higher Gsat, we calculated an activity index:

Cell Imaging

Landing pad HEK293 cells, routinely cultured in 35-mm dishes with doxycycline, were imaged using an inverted EVOS M5000 microscope (Thermo Fisher Scientific, Carlsbad, CA) with a 20× objective lens. Images in Fig. 1 were made with transmitted light and 4′,6-diamidino-2-phenylindole (DAPI) or red fluorescent protein filter sets.

Membrane Potential Assessments by Voltage-Sensitive Fluorescent Dye

Transparent cell-culture 96-well plates were coated with poly-L-lysine (0.1 mg/ml) 1 day before stable cells were plated at an appropriate density to ensure an even distribution without layering.

Cells were cultured with 100 μl Dulbecco’s modified Eagle’s medium containing 10% FBS, 100 U/ml penicillin, 0.1 mg/ml streptomycin, and 4 μg/ml doxycycline. After 1 day in culture, the medium was discarded and the cells were washed twice with low [K] (1 mM) buffer (80 μl/well). Cells were then incubated with 3 μM DiBAC4(3) in low K buffer (or mixtures of low K and high K buffer), with or without additional channel modulators, for at least 30 minutes. Separate images of green and red fluorescence were generated with EVOS M5000 imaging system (10× lens). In the built-in software, the intensity was 0.005 for green fluorescence and 0.379 for red fluorescence.

The recipe for low K buffer is: NaCl 139 mM, KCl 1 mM, CaCl₂ 2 mM, MgCl₂ 1 mM, HEPES 10 mM, glucose 10 mM. The pH was adjusted to 7.4 with NaOH.

The recipe for high K buffer is: KCl 140 mM, CaCl₂ 2 mM, MgCl₂ 1 mM, HEPES 10 mM, glucose 10 mM. The pH was adjusted to 7.4 with potassium hydroxide (KOH).

Exported tiff images were analyzed by CellProfiler with custom designed programs for Landing pad cells and generated stable cell lines. Briefly, cells were detected using red fluorescence (or green for depolarized nontransfected Landing pad cells) and defined as objects by the software, then regions of noncellular background fluorescence were identified and subtracted. Multiple measurements of each cell were exported as a csv format file and the mean intensities of all cells in each image were extracted with a custom script written in Python. These mean intensities were used to describe the membrane potential for each specific cell line with a certain treatment in one image. The average intensity value of duplicate images was calculated as one data point for the statistical analysis and data presentation.

Estimation of Potassium Channel Conductance with DiBAC4(3) Assay Results

To estimate conductance from DiBAC4(3) measurements, we used a simplified Goldman-Hodgkin-Katz equation that assumed that permeability is conductance, and that all non-K conductances can be lumped into a single “sodium” conductance:

where V_m = membrane potential, R = gas constant, T = temperature (°K), z = 1, F = Faraday’s constant, G_K = K conductance, G_Na = sodium conductance, and [K]_in, [K]_out, [Na]_in, [Na]_out are the ion concentrations on either side of the membrane.

Whole-Cell Patch-Clamp Experiments

Micropipettes were pulled from soda-lime glass microhematocrit tubes (Kimble-Chase 2502) using a P-97 puller (Sutter Instruments) and pipette tips with resistance of 2–4 MΩ were used for recordings. Whole-cell currents were recorded at a V_m –60 mV, using an Axopatch 200B amplifier, Digidata 1322A (Molecular Devices) and Axon pCLAMP software (Molecular Devices, LLC). High-Na⁺ bath solution was (mM): NaCl 136, KCl 4, CaCl₂ 2, MgCl₂ 1, HEPES 10, and glucose 10, pH 7.4 adjusted with NaOH. High-K⁺ bath solution was (mM): KCl 140, CaCl₂ 2, MgCl₂ 1, HEPES 10, and glucose 10, pH 7.4 adjusted with KOH. The pipette solution was: potassium aspartate 110, KCl 30, NaCl 10, MgCl₂ 1, HEPES 10, CaCl₂ 0.5, K₂HPO₄ 4, ATP 0.1, and EGTA 5, pH 7.2 adjusted with KOH.

Noise Analysis

Recorded traces were offline sampled at 100Hz and divided into small fragments, each of 6.4 second Mean current and variance (Var²) were estimated in each fragment. In high Na⁺, theoretical K_ATP-specific current ≈0 pA, and mean currents and variance in high Na were subtracted from each high K fragment. Var² was plotted versus number of open channels (n = mean current/single channel current 2.1 pA) in each fragment and fitted with the equation:

where i (single channel current) = 35 pS * (–60 mV) = –2.1 pA and n (number of open channel) = average current value/(–2.1). N denotes the total number of channels on the cell membrane.

Generation of Stable hKir6.1/hSUR2B Cell Lines with Candidate CS Mutations in hKir6.1 or hSUR2B

Multiple hKir6.1 and hSUR2B mutations were introduced into the attB plasmids and used to generate stable cell lines expressing mutant channel complexes (Fig. 3A). These included previously characterized [C176S] and [V65M] (Brownstein et al., 2013; Cooper et al., 2014, 2017) CS mutations, as well as CS-associated, but uncharacterized [E189K] (Chihara et al., 2020), [A46V], and [D338E]mutations in hKir6.1(the latter two mutations were identified in patients clinically diagnosed with CS using Clinical Laboratory Improvement Amendments-approved sequencing). Mutations in hSUR2B include previously characterized [A478V] (Cooper et al., 2015), [R1154Q], [R1154W] (van Bon et al., 2012; McClenaghan et al., 2018), and indel1055 (Gao et al., 2023) CS mutations. After puromycin treatment, all generated cell lines lost blue fluorescence and showed mCherry red fluorescence in almost every single cell (Fig. 1, Supplemental Fig. 2).

Open in a separate window

Fig. 3.

WT and Cantú mutant Kir6.1/SUR2B channel properties characterized by Tl influx. (A) Location of introduced mutations in hKir6.1/hSUR2B complex (PDB: 7MJO) color-coded as in subsequent figures. (B) Representative Tl influx traces of untransfected, WT, Kir6.1[A46V], and SUR2B[R1154Q] cells under basal condition, MI (oligomycin, 2.5 mM and 2-deoxy-D-glucose, 1 mM), forskolin (10 μM), and pinacidil (30 μM). (C) Summary of initial rate from each of the four treatment conditions in (A) (n = 3–6). Data from individual experiments were plotted together with mean ± S.D. Statistical significance compared with the WT group is denoted by ***P < 0.001, **P < 0.01, and *P < 0.01. “No significance” is not marked in the figure and is defined as P ≥ 0.05 according to one-way ANOVA test with Dunnett’s multiple comparisons.

Differential Responses of Kir6.1 and SUR2B Mutants to KCOs, Metabolism Inhibition, and Phosphorylation

Kir6.1/SUR2B is known to be activated by multiple mechanisms (Isomoto et al., 1996), including inhibition of cell metabolism [Metabolic inhibition (MI)] (Li et al., 2016) and by protein kinase A and protein kinase G (PKA/PKG)-dependent pathways (Haynes and Cook, 2006; Yang et al., 2008). We first examined the effect of these activating conditions on WT and mutant hKir6.1/hSUR2B channels (Fig. 3A) using the Tl-flux assay. Cell lines were treated with the potassium channel opener (KCO) pinacidil (100 μM), MI (induced with 2.5 mg/ml oligomycin and 1 mM 2-deoxy-D-glucose), or 10 μM forskolin. In untransfected cells, the Tl-influx rate was not increased by any of these conditions, and 10 μM forskolin actually decreased the influx signal (Fig. 3B). Although basal Tl-influx rate was not markedly different between any transfected cell lines (Fig. 3C), WT and Kir6.1 mutant channels all showed much higher Tl-influx rates in the presence of pinacidil, reflecting increased channel activation (Fig. 3C). Although there was no obvious increase for WT, Tl-influx rates increased for hKir6.1[V65M], [C176S], [E189K], and [D338E] mutations with MI or forskolin (Fig. 3C).

Very surprisingly, all cell lines with hSUR2B mutations were essentially insensitive to pinacidil activation, as well to MI and 10 μM forskolin (Fig. 3C), even though A478V, indel1055, R1154Q, and R1154W have all been shown to be GOF mutations, with enhanced sensitivity to activation by Mg-nucleotides and by pinacidil, in previous studies with Kir6.2/SUR2A mutant channels (Harakalova et al., 2012; Cooper et al., 2015; McClenaghan et al., 2018; Gao et al., 2023). To examine the generality of this markedly different response of hKir6.1 and hSUR2B mutants to the KCO pinacidil, we further tested the response of hKir6.1[V65M] and hSUR2B[R1154Q] mutations to other KCOs (diazoxide, cromakalim, and nicorandil) at presumably maximal concentrations. WT channels were only activated by pinacidil and cromakalim (Fig. 4, A and B), whereas hKir6.1[V65M] showed enhanced activation by these drugs and measurable activation by nicorandil and diazoxide (Fig. 4, A and B). Note, however, that hSUR2B[R1154Q] was not activated by any of the KCOs (Fig. 4, A and B).

Open in a separate window

Fig. 4.

Response of WT and Cantú mutant Kir6.1/SUR2B channels to different KCOs. (A) Representative thallium influx traces of WT, Kir6.1[V65M], and SUR2B[R1154Q] cells treated with pinacidil at indicated concentrations (μM). (B) Summary of the mean initial rate values of data as in (A), for treatment with pinacidil, diazoxide, cromakalim, and nicorandil treatments (n = 3–4, concentrations indicated in μM). Data from individual experiments are plotted together with mean ± S.D. Statistical significance compared with the WT group of the same drug concentration group is denoted by ***P < 0.001, **P < 0.01, or *P < 0.01, according to one-way ANOVA test with Tukey’s multiple comparisons.

We carried out whole-cell patch-clamp experiments on Kir6.1/SUR2B channels and the apparently pinacidil-insensitive SUR2B[R1154Q] mutant channels (Fig. 5). Nonstationary noise analysis (Stevens, 1972) was used to measure channel density and open probability (Fig. 5B). Although mutant channel density is slightly reduced (to ~60% that of WT) (Fig. 5C), this is not sufficient to explain the lack of pinacidil response in mutant Tl fluxes. These data show that first, although the basal open probability of WT channels is extremely low (~0.02), it is much higher for the SUR2[R1154Q] mutant channels (Fig. 5D). They further show that, although WT current is increased ~150-fold by pinacidil, it is only increased ~10-fold for R1154Q mutant channels (Fig. 5E), and they additionally confirm a markedly reduced glibenclamide sensitivity of R1154Q mutant channels (Fig. 5F), as seen in previous experiments with recombinant expression (McClenaghan et al., 2018).

Open in a separate window

Fig. 5.

Loss of pinacidil sensitivity in Kir6.1/SUR2B[R1154Q Cantú mutant channels. (A) Representative whole-cell patch-clamp recordings from stably transfected WT and R1154Q mutant channels exposed to nonconducting (high Na), basal (high K), pinacidil-activated (Pin), and glibenclamide-inhibited (Glib) at concentrations indicated. (B) Nonstationary noise analysis (see Materials and Methods) for the two traces in (A), during exposure to basal, Pin, or Pin+Glib conditions as indicated by colored circles. (C) Number of channels per cell from experiments as in (A) and (B). (D) Open probability from experiments as in (A) and (B) during the final ~30 seconds in each condition. (E) Ratio of the current in pinacidil to the basal current (I_Pin/I_Basal), and (F) ratio of the current in pinacidil+glibenclamide to the current in pinacidil (I_Pin+Glib/I_Pin) from experiments as in (A) and (B). In each of graphs in (C)–(F), data from individual experiments are plotted together with mean ± S.D. Statistical significance compared with the WT under each condition is denoted by ***P < 0.001 or *P < 0.05, according to one-way ANOVA test with Tukey’s multiple comparisons.

Dose-dependence of the response of Kir6.1/SUR2B mutants to pinacidil activation was measured to further quantitatively assess drug sensitivities. WT channels only started to show measurable activation of Tl⁺ influx at 10 μM pinacidil and reached maximum at 100 μM pinacidil. Human Kir6.1 mutations A46V, V65M, E189K, and D338E all showed higher sensitivity to pinacidil, with markedly lower EC₅₀, and C176S additionally showed a much higher maximum channel activity (Fig. 6, A and B). Unexpectedly, at very high concentration of pinacidil (300 μM), the activity of V65M channels was slightly decreased.

Open in a separate window

Fig. 6.

Dose-dependent activation of WT and Cantú Kir6.1 mutant Kir6.1 channels by pinacidil. (A) Representative thallium influx traces of WT and Kir6.1 mutant cells treated with 0, 0.3, 1, 3, 10, 30, 100, and 300 μM pinacidil. (B) Summary of the mean initial rate values of data as in (A). Each data point represents mean ± S.D. of three individual experiments. The mean of average initial rate values were fitted with eq. 3. EC₅₀ and Hill coefficient values are WT: 44.0 ± 9.5 μM, 2.6 ± 1.0; Kir6.1[A46V]: 38.1 ± 3.0 μM, 1.9 ± 0.2; Kir6.1[V65M]: 13.7 ± 0.9 μM, 1.2 ± 0.0; Kir6.1[C176S]: 71.9 ± 18.0 μM, 1.2 ± 0.1; Kir6.1[E189K]: 37.3 ± 17.9 μM, 1.3 ± 0.4; Kir6.1[D338E]: 8.0 ± 0.3 μM, 1.7 ± 0.1.

Variable Sensitivity of Kir6.1 Mutants to Glibenclamide Inhibition

WT and hKir6.1 mutant channels were next evaluated for their response to glibenclamide, in the presence of 100 μM pinacidil (Fig. 7, A and B). All channels were inhibited to a similar maximal extent, but all GOF mutations showed reduced sensitivity to glibenclamide (Fig. 7B), with a similar rank order to that seen for enhancement of pinacidil sensitivity (Fig. 6B). Plotting the IC₅₀ values for glibenclamide inhibition against EC₅₀ values for pinacidil activation (Fig. 7C) reveals a close fit of the data to the dashed line that represents the hypothesis that the action of glibenclamide is purely on decreasing the binding affinity for pinacidil (i.e., that they are reciprocally related).

Open in a separate window

Fig. 7.

Dose-dependent inhibition of pinacidil activated WT and Cantú mutant Kir6.1/SUR2B channel current by glibenclamide. (A) Representative thallium influx traces of WT and Kir6.1[D338E] cells treated with 100 μM pinacidil with 0, 1 nM, 10 nM, 100 nM, 1 μM, and 10 μM glibenclamide. (B) Summary of the mean initial rate values of data as in (A). Each data point represents mean ± S.D. of four to seven experiments. The data were fitted with eq. 4. IC₅₀ and Hill coefficient values are WT: 102.0 ± 6.7 nM, 2.9 ± 6.0; Kir6.1[A46V]: 213.5 ± 31.9 nM, 1.9 ± 0.3; Kir6.1[V65M]: 565.1 ± 132.4 nM, 1.8 ± 0.8; Kir6.1[C176S]: 261.8 ± 139.0 nM, 1.1 ± 0.5; Kir6.1[E189K]: 642.0 ± 32.6 nM, 1.0 ± 0.2; Kir6.1[D338E]: 591.3 ± 172.4 nM, 1.5 ± 0.4. (C) IC₅₀ for glibenclamide inhibition versus EC₅₀ for pinacidil activation. The dashed line models the hypothesis that glibenclamide acts solely by displacing pinacidil, with the equation: EC₅₀(WT) × IC₅₀(WT) = EC₅₀(mut) × IC₅₀(mut).

Measurements of K_ATP Channel Activities by Voltage Sensing Dye DiBAC4(3)

The sensitivity of the thallium flux assay being such that we can only distinguish channel-specific signals in the presence of KCOs is a major limitation to assessing physiologic activities. We therefore developed a complementary method, using DiBAC4(3) to detect channel activity by effects on the membrane potential that allows us to characterize mutational effects under conditions without KCOs (Fig. 8A). Intracellular DiBAC4(3) fluorescence is reported to be linearly positively correlated with cell membrane potential (Yamada et al., 2001). To confirm this, pilot experiments were carried out in which untransfected Landing pad cells, or WT, hKir6.1[A46V], and hKir6.1[V65M] transfected cells, loaded with 3 μM DiBAC4(3) dissolved in buffer, were incubated in extracellular K⁺ concentrations between 1–140 mM (Fig. 8B). Consistent with a relatively high K⁺ conductance that clamped the membrane potential close to E_K under basal conditions, the fluorescence from V65M cells was linear with log[K]_o, to below 3 mM (Fig. 8B). Both WT and Landing pad-only cells showed a much shallower dependence on [K]_o, consistent with much lower overall K⁺ conductances, and essentially no K_ATP-specific conductance, whereas A46V behavior was intermediate (Fig. 8B).

Open in a separate window

Fig. 8.

Characterization of Kir6.1/SUR2B channel activities with DiBAC4(3) assay. (A) Representative images of untransfected Landing pad, WT, Kir6.1[A46V], and Kir6.1[V65M] cells, 30 minutes after incubation with 3 μM DiBAC4(3) in [K⁺]_o 1 mM or 140 mM. (B) Fluorescence intensities of DiBAC4(3)-treated untransfected Landing pad, WT, Kir6.1[A46V], and Kir6.1[V65M] cells as a function of [K⁺]_o (1–140 mM). The data are expressed as mean ± S.D. (n = 3 experiments in each case). (C) DiBAC4(3) fluorescence of untransfected Landing Pad and WT cells to different concentrations of pinacidil (PIN) and diazoxide (DZX), normalized to untreated cells in each group. The data are expressed as mean ± S.D. (n = 5–6 for PIN, n = 4 for DZX). ***P < 0.001 for pinacidil group, ^##P < 0.01, and ^###P < 0.001 for diazoxide group compared untreated control in each group. One-way ANOVA with Dunnett’s multiple comparisons test. (D) DiBAC4(3) fluorescence (relative to untreated group) for untransfected Landing pad and WT cells in response to MI (oligomycin, 2.5 mM) and 2-deoxy-D-glucose, 1 mM) or forskolin (10 μM). Data are expressed as mean ± S.D. (n = 4–6). **P < 0.01, ***P < 0.001 compared with data points of control cells. One-way ANOVA with Dunnett’s multiple comparisons test.

At 1 mM [K⁺]_o, fluorescence signals of the four cell lines were well detected and well separated (hKir6.1[V65M] < hKir6.1[A46V] < WT ≈ untransfected cells) (Fig. 8B), and we thus used this extracellular [K⁺] for the following measurements. We first conducted DiBAC4(3) assays on WT hKir6.1/hSUR2B channels under the same experimental conditions used in the thallium influx assay. Although pinacidil and diazoxide had apparently qualitatively different effects in the thallium flux assay (i.e., no effect of 100 μM diazoxide) (Fig. 4B), both drugs dose-dependently hyperpolarized WT cells (Fig. 8C), with pinacidil being more potent than diazoxide. Notably, diazoxide generated clear K_ATP-dependent hyperpolarization signals in the DiBAC4(3) assay at all concentrations above 1 μM (Fig. 8C). Similarly, although MI and 10 μM forskolin failed to generate measurable Tl⁺ influx signals, both caused significant hyperpolarization of WT cells (Fig. 8D), confirming that intracellular ATP depletion and phosphorylation can cause channel activation.

Overlapping the dose response to pinacidil in the two assays, we can more clearly compare the sensitivity of the two methods. With any given concentration of pinacidil, the potassium channel activity in the two assays should theoretically be the same. However, the pinacidil-sensitivity of signal amplitudes was very different between the two assays. Although thallium assay signal changes are only detected above 10 μM, DiBAC4(3) signal changes are almost saturated at this concentration (Fig. 9A). We used a simple model of the system (see Materials and Methods), assuming that the membrane potential is controlled only by K conductance and Na conductance to predict the K conductance reported by the DiBAC4(3) signal for WT channels (Fig. 9B). In the range of pinacidil from ~0.3–10 μM, the model provides reasonable agreement with the thallium flux-reported conductance, suggesting that relative conductance could be directly inferred with the DiBAC4(3) method, but also that ~80% of the DiBAC4(3) signal change (that occurs at <3 μM pinacidil) results from activation of <20% of the maximal available WT conductance. With this recognition that very low levels of WT channel activity are detected by the DiBAC4(3) assay, we measured the DiBAC4(3) fluorescence of each of the mutant cell lines under basal conditions (Fig. 9C). WT cells showed similar fluorescence amplitude to untransfected cells, but all of the mutant cells exhibited markedly lower fluorescence. Notably, CS hSUR2B mutations, all of which showed no activation by pinacidil in the thallium flux assay, clearly showed comparably significant cell hyperpolarization effects (Fig. 9C), reflecting some level of basal activation. For the hKir6.1 mutations, we plotted the calculated relative conductance from these DiBAC4(3) measurements versus the activity index assessed from Tl⁺ flux assays (see Materials and Methods). Consistent with both methods assessing the same underlying behaviors, there is a reasonable correlation between the two (Fig. 9D, see Discussion).

Open in a separate window

Fig. 9.

Sensitivity of DiBAC4(3) assays in response to Kir6.1/SUR2B activities and cellular diagnosis of CS mutations. (A) Dose–response of pinacidil-induced signal changes in thallium influx (orange, from Fig. 5B) and DiBAC4(3) (blue, n = 4–5, [K⁺]_o 1 mM) assays. Data are expressed as mean ± S.D. in each case. (B) Calculated GK_ATP from Tl flux and DiBAC4(3) assays. Data shown for mean values. For Tl flux assay, the conductance is directly proportional to initial rate and scaled to 1. For DiBAC4(3) assay, eq. 6, with values of background gNa = 0.001, gK = 0.003. (C) Basal DiBAC4(3) fluorescence from untransfected Landing pad, WT, and indicated Kir6.1 or SUR2B mutant cells, measured without drug treatment in 1 mM [K⁺]_o. Individual values, mean, and S.D. are plotted (n = 9–16). One-way ANOVA with Dunnett’s multiple comparisons test. ***P< 0 .001. (D) Calculated conductance in the DiBAC4(3) assay plotted versus activity index in the Tl⁺ flux assay for Kir6.1 mutants.

High Throughput Approaches to Ion Channel Mutation Screening

We have therefore adapted the Landing pad system (Matreyek et al., 2020) to provide a rapid approach to generating stable cell lines expressing hKir6.1/hSUR2B channels and variants at similar expression levels. Second, we have developed thallium flux (Raphemot et al., 2014) and DiBAC3(4) assays (Gopalakrishnan et al., 1999; Yamada et al., 2001; Baczko et al., 2004) for high throughput and semi-quantitative assessment of channel activity. In both assays, cells are intact with normal cellular composition—critical for measuring physiologically relevant K_ATP channel activities. One potential shortcoming of our diagnosis pipeline is that mutations are expressed homomerically, whereas all patients to date are heterozygous. Heterozygous expression might be achieved using an additional IRES to drive a normal gene copy, but this will increase the complexity of the attB plasmid and require further optimization. Otherwise, the present assays should provide accurate reflection of CS mutational effects in relevant physiologic conditions. They can now be used for comprehensive mutational analysis and may be further scaled up for random mutagenic library-based approaches and combined with fluorescence-activated cell sorting and next-generation sequencing for large-scale high throughput (Coyote-Maestas et al., 2022) and generation of clinical “look up” tables.

In considering the thallium flux versus DiBAC4(3) assays, it is clear that the dynamic range of K conductance assayed by the two is quite distinct and overlaps over only a small range. The thallium flux data indicate that basal K_ATP conductance is very low in CS Kir6.1 mutations but is activated by pinacidil with higher sensitivity and to a greater extent than in WT. However, thallium flux data also suggest the erroneous conclusion that none of the SUR2B mutations result in any meaningful GOF. That this is not correct is made clear by the DiBAC4(3) data, which reveal a hyperpolarizing effect of all SUR2B mutations and provide an estimate of the relative effect of different mutations. In the thallium flux assay, there was also no observable effect of MI or forskolin on WT Kir6.1/SUR2B channels, but again there was significant decrease of DiBAC4(3) fluorescence, confirming the activating effects of both on Kir6.1/SUR2B channels (Shi et al., 2007; Yang et al., 2008; Li et al., 2016).

Our data (Fig. 9A) show that the DiBAC4(3) assay is extremely sensitive to even very small changes of conductance, such that only 1%–10% of the maximal pinacidil-activated conductance measured in the thallium flux being sufficient to maximally hyperpolarizes the cells. The simplified analysis in Fig. 9B suggests that both assays do indeed assess the same conductances but over a very different absolute range. Although neither assay is as precise as patch-clamp, both are relatively high throughput and technically easy and do not require the advanced operator skills of electrophysiology. By predicting basal conductance of Kir6.1 mutants from DiBAC4(3) measurements, and the predicted basal activity given the pinacidil sensitivity in the Tl flux assay, we see that there is reasonable agreement between the two methods in the rank effect of mutations (Fig. 9D). The much higher sensitivity of the DiBAC4(3) assay, avoiding the need for pinacidil activation, may make it particularly advantageous over the thallium assay for clinical diagnosis.

CS: Mechanistic Implications

One potential explanation for the remarkable, unexpected, loss of pinacidil sensitivity for each SUR2B mutation is that these mutations dramatically decrease trafficking of the channel complex to the surface, with a small number of channels that do make it to the surface still being sufficient to elevate basal K_ATP activities to cause measurable hyperpolarization. Whole-cell patch-clamp experiments confirm dramatic loss of pinacidil activation for SUR2B[R1154Q], but with only a slight reduction in channel density that cannot explain lack of conductance in Tl flux experiments. A more probable explanation is that these mutations decrease KCO binding to the SUR2 subunit or decrease the coupling between KCO binding and channel activation. MgADP decreases the binding of the related KCO P1075 to SUR2B (Hambrock et al., 1999), and it is possible that the SUR2B mutations favor this MgADP induced dissociation of KCOs.

Additionally, hKir6.1 CS mutations all caused variable increases in maximal conductance in saturating [pinacidil] and leftward shifts of the [pinacidil]-dependence. Assuming that pinacidil binding affinities are unaffected, the left-shifted dose–response curves reflect increased coupling of binding to channel activation. In general, analogous mutations in Kir6.2 increase the stability of the open channel, and hence the maximum open probability (Trapp et al., 1998; Loussouarn et al., 2000; Enkvetchakul et al., 2001; Cooper et al., 2017), which may give rise to the higher plateau fluxes at saturating pinacidil concentrations.

The synthesized human SUR2A cDNA was a kind gift of Rod MacKinnon (Rockefeller University). The authors declare that all the data supporting the findings of this study are available within the paper.

All original data will be made available to interested parties on reasonable request.