Fig 2: The binding domain of TMEM43 for TASK-1 interaction. (A) TMEM43 protein structure prediction based on AlphaFold. (B) Left: amino acids of TMEM43 protein and its predicted topology based on AlphaFold. TM domains are colored in pink. The truncated domains are dotted-lined with different colors. Right: amino acids of TMEM43-?Loop1 with three GGS linkers. (C) Co-IP results performed with TASK-1 antibody using lysates of HEK293T cells expressing TASK-1 and TMEM43-FLAG full length and Loop1 truncation mutant (?Loop1). Expected band size for TMEM43-?Loop1 is 20 kDa. (D) Co-IP data showing that both TMEM43-WT and TMEM43-?IM are immuno-pulled down with anti-TASK-1 antibody. The expected band size for TMEM43-?IM is 41 KDa.

Fig 3: Overexpression of CG4928 induces apoptosis and alters the membrane potential. The CG4928:eGFP-transfected cells had a decline in cell number at 24 h post-transfection, and therefore apoptosis, cytotoxicity, and viability were monitored using the ApoTox-GloTM Triplex kit. Fluorescence and luminescence were measured, mean fluorescence and luminescence signals (±SEM) (n = 2 with 12 measured points in each) were plotted, and differences were studied using unpaired t-test (*p < 0.05, **p < 0.01, ***p < 0.001). CG4928:eGFP-transfected cells had (A) higher apoptosis, (B) cytotoxicity, and (C) lower viability. The CG4928:eGFP-transfected cells were maintained in different media [normal Dulbecco’s modified Eagle’s medium (DMEM), high sodium, high potassium, low sodium, and low potassium media] to investigate changes in cell death. The mean fluorescent and luminescence signals (±standard error of the difference) (n = 6) were plotted and CG4928:eGFP-transfected cells on normal DMEM media was used as control. Differences were calculated using appropriate ANOVA [one-way ANOVA (apoptosis) or Kruskal–Wallis (cytotoxicity)] with unpaired t-test or Mann–Whitney as post hoc test (adjusted p-values: *p < 0.0489, **p < 0.0099, ***p < 0.0001). (D) Changes in apoptosis were observed in transfected cells maintained on high sodium, high potassium, and low sodium, (E) while only high sodium and high potassium altered the cytotoxicity. The speculated interaction between CG4928 and TASK1 channels suggests that the results could be linked to the membrane potential. Therefore, the membrane potential was measured using the FluoVoltTM membrane potential kit. The average fluorescent intensity per area was calculated from each frame at each timepoint per well (n = 4, images taken for two sites with an interval of 10 s for each of the 10 timepoints). Graphs with 95% confidence interval were generated and unpaired t-tests were performed (*p < 0.05, **p < 0.01, ***p < 0.001). (F) The basal membrane potential was lower in CG4928:eGFP-transfected cells compared with the control. (G) Representative image of the difference in fluorescent intensity between CG4928:eGFP-transfected cells and the controls. (H) Taken together, it can be speculated that CG4928 is important for TASK1 channel function and/or regulation. Overexpression of CG4928 probably caused potassium leakage, leading to disturbed membrane potential and cell death. Cell death could be hindered by culturing the CG4928-overexpressing cells in high sodium concentrations, but the mechanisms behind the rescue still remain uncertain.

Fig 4: TMEM43 interacts with TASK-1 in the heterologous overexpression system. (A) Schematic diagram illustrating the heterologous DNA expression in HEK293T cell line and procedures of the co-IP assay. Tmem43, Cx26, and Cx30 coding sequences tagged with FLAG and Task-1 coding sequences were cloned into IRES2 vectors. Anti-TASK-1 antibody was used for immunoprecipitation, and the pulled-down protein was blotted with an anti-FLAG antibody. (B) Co-IP result, performed with normal IgG or TASK-1 antibody using lysates of HEK293T cells co-expressing TMEM43-FLAG and TASK-1. (C) Co-IP results of TASK-1 with Cx26-FLAG and Cx30-FLAG. Anti-FLAG antibody was used to detect the connexin channels. The first lane is a negative control. (D) A cartoon depicting the principle of the Duolink PLA. If the proteins of interest are in close proximity, the DNA probes hybridize to make circular DNA. This DNA can be amplified and visualized by fluorescently labeled complementary oligonucleotide probes. (E) Duolink PLA with anti-TMEM43 and anti-TASK-1 antibodies. Duolink PLA signal was recognized as far-red fluorescent (?ex 644 nm, ?em 669 nm), indicative of proximity of TMEM43 and TASK-1 (<40 nm). The lower panel is a negative control, using only an anti-TMEM43 antibody.

Fig 5: Endogenous TMEM43 interacts with TASK-1 in mouse cochlea tissue. (A) An illustration of materials used for ex vivo study. C57BL/6 pups of p5~p8 were used, and the cochlea tissue was acutely dissected from the inner ear. (B) Co-IP data performed with cochlea tissue. Tissue lysates were pulled down with anti-TASK-1 antibody and blotted with anti-TMEM43 antibody. (C) Immunostaining of cochlea tissue with anti-TMEM43 and anti-TASK-1 antibodies. Inset images show magnified GLSs with TMEM43 and TASK-1 signals. (D) Duolink PLA with anti-TMEM43 and -TASK-1 antibodies, performed with acutely dissected mouse cochlea tissue. Duolink PLA signal was positive at GLSs (green). The amplified PLA red fluorescent signal was pseudo-colored as green for better data display. Inset: magnified images. The lower panel is a negative control, using only an anti-TMEM43 antibody.