Identification of an ATP-sensitive potassium channel in mitochondria

Abstract

Mitochondria provide chemical energy for endoergonic reactions in the form of ATP, and their activity must meet cellular energy requirements, but the mechanisms that link organelle performance to ATP levels are poorly understood. Here we confirm the existence of a protein complex localized in mitochondria that mediates ATP-dependent potassium currents (that is, mitoK_ATP). We show that-similar to their plasma membrane counterparts-mitoK_ATP channels are composed of pore-forming and ATP-binding subunits, which we term MITOK and MITOSUR, respectively. In vitro reconstitution of MITOK together with MITOSUR recapitulates the main properties of mitoK_ATP. Overexpression of MITOK triggers marked organelle swelling, whereas the genetic ablation of this subunit causes instability in the mitochondrial membrane potential, widening of the intracristal space and decreased oxidative phosphorylation. In a mouse model, the loss of MITOK suppresses the cardioprotection that is elicited by pharmacological preconditioning induced by diazoxide. Our results indicate that mitoK_ATP channels respond to the cellular energetic status by regulating organelle volume and function, and thereby have a key role in mitochondrial physiology and potential effects on several pathological processes.

Extended data figure 1. Mitok overexpression causes mitochondrial dysfunction.

(a) Membrane (Pellet) and soluble (Surnatant) proteins were separated from isolated liver mitochondria using ice-cold 0.1M Na₂CO₃ (pH 11.5). Western blot is representative of three independent experiments. (b) Proteinase K protection assay in isolated liver mitochondria. Similar results were obtained in three independent reactions. (c) Mitochondrial morphology of control and Mitok-overexpressing HeLa cells (representative of five independent experiments; scale bar is 10 μm). (d) Representative images and average ± s.d. traces of control and Mitok-GFP expressing HeLa cells loaded with TMRM. n = 9 biologically replicates from three independent experiments. (e) [Ca²⁺]_mt measurements (mean ± s.d.) in intact HeLa expressing the indicated constructs; n = 4 biological replicates (representative of three independent experiments), * p < 0.001 using one-way ANOVA with Holm-Sidak correction. (f) qPCR analyses of transcripts from HeLa cells or the indicated human tissues using specific (isoform 1 and isoform 2) or non-specific (pan) primer pairs for MITOK. Data were normalized to ACTIN and expressed as mean ± s.d. For HeLa cells, n=3 biologically independent samples. For human tissues, n=3 technical replicates. (g) Protein expression of MITOK isoforms in HeLa cells transfected with the indicated constructs. Asterisk indicates a non-specific band. Image is representative of two independent experiments.

Extended data figure 2. Localization and function of human MITOK isoforms.

(a) Immunolocalization of MITOK (green) and the mitochondrial marker HSP60 (cyan) in HeLa cells transfected with the indicated constructs (images are representative of two independent experiments; scale bar is 10 μm). (b) [Ca²⁺]_mt measurements (mean ± s.d.) in intact HeLa expressing the indicated constructs; n ≥ 6 independent samples, * p < 0.001 using one-way ANOVA with Holm-Sidak correction.

Extended data figure 3. Mitok is a cation channel.

(a-b) (a) Western blots and Coomassie of Mitok expressed and purified from E.coli (a) or WGL (b), representative of three independent experiments. (c) Current traces showing two channels gating together resulting in flickering activity (upper panel) or normal single channel activity (lower panel). The two traces were recorded in the same experiment, performed in 100 mM Kgluconate medium and representative of five independent experiments. (d) Current trace showing burst like activity and cooperative transition between dual- or multi-states of two channels. Recording was performed in 100 mM Kgluconate medium. Similar activity was present in more than six recordings. (e) Representative traces of Mitok channel activity at the indicated voltages. Similar results were obtained in three independent experiments. (f) Voltage ramp (from -120 mV to 0 mV) of Mitok recorded in 100 mM Kgluconate symmetrical medium (n=3 independent experiments). (g) Single-channel i-V curves under symmetric (black) and asymmetric (grey) ionic conditions (mean value ± s.d., n = 50 from 3-4 different experiments for each point). Fitting revealed an E_rev = -21.6 ± 1 mV (n=4 independent experiments). (h) Representative traces (upper panels) and amplitude histograms (lower panels) before and after the addition of 2 mM Ba²⁺ (n=4 independent experiments) in 100 mM KCl medium. (i) Representative traces (upper panels) and amplitude histograms (lower panels) of Mitok activity before (Control) and after paxilline (40 μM) addition (n=4 independent experiments).

Extended data figure 4. Mitok alone is insensitive to ATP.

(a) Activity of Mitok in 100 mM Na-gluconate (n=5 independent experiments). (b) Current traces of Mitok channel activity obtained from 60 s recordings (in 100 mM K-gluconate) before (upper panel) and after (lower panel) addition of 2 mM Mg/ATP (representative of 8 independent experiments). Voltages of cis side are reported. (c) Representative traces (left panels) and amplitude histograms (right panels) of Mitok activity before (Control) and after 5-HD (100 μM) addition (n=5 independent experiments).

Extended data figure 5. Biophysical characterization of recombinant Mitok and MITOSUR.

(a-c) Thermal shift assay analysis of Mitok and MITOSUR: average curves (a), graphs of Taggr (b, expressed as mean ± s.d.) and Western blot (c), representative of 4 independent experiments. (d) Membrane extraction and Western blot (representative of two independent experiments) of in vitro co-expressed Mitok and MITOSUR incorporated into liposomes. (e-f) Membrane topology assessed by proteinase K protection assay in reconstituted liposomes and probed for Mitok (e) and MITOSUR (f). (g) The same experiment shown in Figure 2a is here represented with a different time scale. (h) Amplitude histograms of channel activity before (Control) and after first addition of 500 μM Mg/ATP, second addition of 500 μM Mg/ATP, and third addition of 30 μM diazoxide. Similar results were obtained with 4 independent preparations. (i) Single channel current (i)-voltage (V) relationship of Mitok and MITOSUR. Linear fitting revealed a chord conductance of 63±3 pS (n=4 independent experiments). (j) Activity in the absence (Control, left panel) and presence of 1 mM Mg²⁺ (right panel) in 100 mM Kgluconate medium (n=3 independent experiments). (k) Activity of Mitok + MITOSUR in 100 mM K-gluconate, 5 mM EDTA, 10 mM Hepes, pH 7.4 (n=3 independent experiments). (l) Mitok + MITOSUR channel activity in 100 mM Na-gluconate medium (n=4 independent experiments).

Extended data figure 6. MITOK and MITOSUR interact in situ.

(a) Proteinase K protection assay in isolated HeLa mitochondria. Similar results were obtained in two independent reactions. (b) Co-immunoprecipitation of endogenous Mitok using mitochondria isolated from HeLa cells. FT: flow-through fraction; W3: third (last) CoIP wash. Representative of two independent experiments (c) Co-immunoprecipitation between overexpressed Mitok and mutant MITOSUR^K513A. Representative of two independent experiments.

Extended data figure 7. Genetic ablation of MITOK in HeLa cells.

(a) Schematic representation of the human MITOK gene. The expanded regions were used to design Cas9 guides (highlighted in red). (b) Western blot of wild type and MITOK-KO HeLa cell lines (representative of three independent experiments). (c) Mitochondrial morphology in wild type and MITOK-KO HeLa cells. (scale bar is 10 μm). Asterisks are located near doughnut-shaped mitochondria. Similar results were obtained in five independent experiments. (d-e) Δψ_m measurements in control and MITOK-KO cells. Cells were loaded with TMRM and normalized fluorescence in different regions was monitored through time. (d) Representative traces of single mitochondria. (e) Pseudo-colored representative images of a HeLa^MITOK-KO#1 cell loaded with TMRM at the indicated time points. Similar results were obtained in four independent experiments. (f) Western blot in HeLa cells of the indicated genotype. Representative of two independent experiments.

Extended data figure 8. Loss of MITOK causes mitochondrial dysfunction.

(a) OCR measurements in wt and MITOK-KO HeLa cells treated with either vehicle or 1pM valinomycin for 1 hour. Representative of three independent experiments (b) OCR measurements in wt and MITOK-KO HeLa cells transfected with control or mitoK_ATP expressing (MITOSUR-P2A-Mitok) plasmids. Representative of three independent experiments. (c) Maximal cristae width in the indicated genotype. n ≥ 12 individual cells (approximately 20 cristae per cell were measured) from two independent preparations, * p ≤ 0.013 using two-way ANOVA with Holm-Sidak correction. (d) OPA1 crosslinking (using 1mM BMH) in wt and MITOK-KO cells. Similar results were obtained in three independent experiments. (e-f) ECAR (e) and OCR (f) measurements in intact cells of the indicated genotype. n = 5 biological replicates, representative of two independent experiments. (g) ROS production during energy stress. Cells were incubated in 5.5 mM of either glucose or 2-deoxyglucose in the presence or absence of 30 μM diazoxide and fluorescence were monitored for 16 hours. Box plots indicate the rate of ROS production over this time frame. n ≥ 10, * p < 0.05 using three-way ANOVA with Holm-Sidak correction. (h) Cell death analysis in HeLa cells treated with 0, 100 or 500 μM H₂O₂. Data were normalized to the untreated condition and expressed as mean ± s.d. n = 3 independent experiments, * p < 0.003 using two-way ANOVA with Holm-Sidak correction.

Figure 1. Biochemical and functional characterization of MITOK.

(a) Representation of human and murine MITOK proteins. Transmembrane (TM) and coiled-coil (CC) domains are indicated. (b) Immunolocalization of MITOK (green) and the mitochondrial marker HSP60 (red) (representative of four independent experiments; scale bar is 10 μm). (c) Subcellular fractionation of mouse liver (replicated twice). (d) Representation of MITOK membrane topology. (e) TEM images of control and Mitok-overexpressing HeLa cells (replicated three times). (f-g) Representative current traces with Mitok purified from E.coli (f, n=5 biological replicates from two independent preparations) or expressed in vitro (g, n=23 biological replicates from 10 independent preparations).

Figure 2. Electrophysiological characterization of recombinant Mitok co-expressed with MITOSUR.

(a) Current traces before (Control) and after first addition of 500 μM Mg/ATP, second addition of 500 μM Mg/ATP, and third addition of 30 μM diazoxide. All traces were obtained from the same experiment, representative of 4 independent experiments. (b) Current recordings before and after addition of 30 μM glibenclamide. The channel was re-activated by subsequent addition of 100 μM diazoxide (n= 4 for inhibition by glibenclamide, n=2 for reactivation by diazoxide). (c) Representative histograms before (upper panel) and after (lower panel) addition of 5-HD (100 μM, n = 5 independent experiments).

Figure 3. MITOK and MITOSUR form the mitoK_ATP in situ.

(a) Co-immunoprecipitation between overexpressed Mitok and MITOSUR (representative of three independent experiments). (b) Blue-native PAGE of digitonin-permeabilized mitochondria (representative of two independent experiments). (c) Co-immunoprecipitation of endogenous Mitok using liver mitochondria (representative of two independent experiments). (d) Δψ_m measurements in HeLa cells transfected with the indicated constructs. n ≥ 9 biological replicates from three independent experiments, *p ≤ 0.01 using two-way ANOVA with Holm-Sidak correction. (e) [Ca²⁺]_mt measurements (mean ± s.d.) in HeLa expressing the indicated constructs; n = 8 biological replicates (representative of three independent experiments), *p < 0.001 using two-way ANOVA with Holm-Sidak correction. (f) Current traces (left panels) and histograms (right panels) of Mitok together with MITOSUR^K513A before and after the addition of 2 mM Mg/ATP (representative of three independent preparations).

Figure 4. Loss of MITOK impairs mitochondrial structure and function.

(a) Swelling traces of wild type mitochondria (three independent experiments with similar results). (b) Mitochondrial swelling rates in K⁺-based media. n=2 independent experiments, *p ≤ 0.009 using two-way ANOVA with Holm-Sidak correction. (c) Quantification of Δψ_m flashes. n > 10 independent experiments, *p ≤ 0.001 using two-way ANOVA with Holm-Sidak correction. (d) OCR measurements. n = 5 biological replicates, representative of three independent experiments. (e) TEM images of mitochondrial ultrastructure, representative of two independent preparations. (f) Analysis of mitochondrial morphology during energy stress. Box plots indicated the % of organelle area occupied by elongated (cyan), intermediate (grey) or fragmented (magenta) mitochondria. Scale bars indicate 10μm. n ≥ 22 individual cells from 3 independent experiments, *p < 0.01 using one-way ANOVA with Holm-Sidak correction. (g) Schematic representation of mitoK_ATP channels.

Figure 5. Mitok is required for diazoxide-induced cardioprotection.

(a) ⁸⁶Rb⁺ flux in isolated mitochondria. n = 3 independent experiments. (b) ⁸⁶Rb⁺ uptake rate. n = 3 independent experiments, * p ≤ 0.029 using two-tailed Student's t-test. (c) Western blot of wt and Mitok-KO liver mitochondria. (d-e) Heart injury after ischemia/reperfusion, evaluated as % of LDH release (d, mean ± s.d., n ≥ 5 independent animals, *p < 0.001 using two-way ANOVA with Holm-Sidak correction), or % of infarct area after TTC staining (e, mean ± s.d., n ≥ 7 independent animals *p = 0.008). Single measurements are provided in Source Data Tables.