The mpkCCDc₁₄ line is an immortalized mouse collecting duct principal cell line, which was cultured as described previously (Bens et al., 1999). These cells were cultured in a 1:1 mixture of DMEM and Ham’s F-12 medium (GIBCO) supplemented with 20 mM HEPES, 2 mM L-glutamine, 50 nM dexamethasone, 1 nm triiodothyronine, 2% heat-inactivated FBS, and 0.1% penicillin-streptomycin. The mpkCCD_c14 cells were plated at a density of 75,000 cells·cm⁻¹ and grown on permeable supports to maintain cell polarization (Costar Transwells; 0.4 µm pore, 24 mm diameter) and cultured for at least 7 days prior to the experiments.

Biotinylation of the plasma membrane from mpkCCDc₁₄ cells was performed as described previously (Wu et al., 2016). Briefly, after each treatment, the cells were incubated with a freshly prepared solution of 1.0 mg/ml EZ-Link sulfo-N-hydroxysuccinimide disulfide-biotin (Pierce, 21331) in borate buffer for 30 min at 4°C. The biotin reaction was quenched for 5 min with 0.1 mM lysine. An equal amount of lysate protein (1 mg) from each sample was respectively incubated with 50 µl of immobilized streptavidin-agarose beads (Pierce, 20349) at 4°C for overnight with gentle shaking. The beads were washed four times with RIPA buffer. Equal amounts of samples from either whole-cell or biotinylated plasma membranes were loaded and separated by a 10% SDS-polyacrylamide gel and transferred to polyvinylidene fluoride membranes. The membranes were then blocked in 5% non-fat dry milk for 1 h, followed by incubation with rabbit polyclonal anti-TRPM4 antibody (1:200 dilution; Alomone Labs; ACC-044) at 4°C for overnight. Rabbit polyclonal anti-GAPDH (1:1000 dilution; Santa cruz; sc-25778) was used as internal controls. Bands were visualized with enhanced chemiluminescence (Bio-Rad, Cat. No., 170-5061) and quantified via densitometry using the ImageJ software (NIH ImageJ software).

Lipid raft fractionation was isolated as described previously (Liu et al., 2015). Briefly, mpkCCDc14 cells suspension were homogenized in 0.5% Brij 96V (Sigma)/TNEV buffer (10 mM Tris-HCl, pH 7.5; 150 mM NaCl; 5 mM EDTA; 2 mM Na vanadate; and protease inhibitor cocktail) on ice for 30 min. The supernatant (500 μl) was mixed with an equal volume of 80% sucrose in TNEV and transferred into a centrifuge tube (13 × 51 mm; Beckman Coulter, Palatine, IL, United States). Three milliliters of 35% sucrose in TNEV was carefully layered on top of the mixture, followed by another 1 ml layer of 5% sucrose. The sucrose gradient was then centrifuged in a SW 50.1 rotor (Beckman Coulter) at 34,000 rpm (∼ 110,000 g) for 20 h at 4°C. After centrifugation, fractions were collected starting from the top to bottom of the tube. Thirteen fractions (∼ 400 μl) were collected, and equal volumes of each fraction were analyzed by 10% gradient SDS-PAGE and immunoblotted with TRPM4 antibodies and caveolin-1 (1:1000 dilution; Cell Signaling Tech cat# 3267S).

Single-channel currents were recorded from mpkCCD_c14 cells under the voltage-clamp mode with an Axopatch-200B amplifier (Molecular Devices, Sunnyvale, CA), using inside-out patch-clamp configurations, as described previously (Wu et al., 2016). Data were acquired and sampled with a low-pass, 1 kHz, eight-pole Bessel filter using a Digidata 1440A analog-digital interface (Axon Instruments, Inc.). The mpkCCDc₁₄ cells were thoroughly washed with NaCl solution containing (in mM) 145 NaCl, 1 MgCl₂, 1 CaCl₂, 10 glucose, and 10 HEPES, adjusted pH to 7.4 with NaOH. This NaCl solution was used for filling the bath in the patch chamber and filling the patch pipette. The patch pipette was pulled with Borosilicate glass, giving a tip resistance of 5–8 MΩ when filled with NaCl solution. Single-channel currents were obtained at a holding potential of 80 mV for inside-out recordings, and only the patches with the seal resistance >2 GΩ were used. Experiments were conducted at room temperature (22–25°C). Prior to analysis, the single-channel traces were further filtered at 100 Hz. The single-channel amplitude was constructed by all-point amplitude histogram and the histograms were fit using multiple Gaussians and optimized using a simplex algorithm. P _O was calculated as P _O = NP _O /N, where N (N was estimated by the current amplitude histogram during at least 5 min recording period when the channel was maximally activated by Ca²⁺) is the number of active channels in the patch. According to a total of 20 inside-out patches, we found that the channel activity can be maximally activated after we excised the patch membrane and exposed the inner leaflet of the membrane to 1 mM Ca²⁺ in the bath. Under the condition, we recorded more than 5 min and the channel activity remained very high. This maximal activation of the channel after we excised the patch membrane into a bath solution containing 1 mM Ca²⁺ was also used for the estimation of N. We also switched to a solution containing 10 mM EGTA without any calcium to mark the zero current levels. More importantly, since lovastatin significantly inhibited the TRPM4 open probability by reducing its sensitivity to Ca²⁺. In order to obtain the number of active channels in the patches, single channel current was recorded from inside-out parches exposed the patch membrane to the bath containing 5 mM CaCl₂, followed by a bath solution with 10 mM EGTA without any calcium. The free Ca²⁺ concentration after chelating CaCl₂ with EGTA was determined using free Web software Winamac (Stanford University, Stanford, CA, United States), as previously described (Wu et al., 2014).

Confocal microscopy experiments were performed as previously reported (Wu et al., 2016). Briefly, after fixation with 4% paraformaldehyde at room temperature for 10 min, the cells were permeabilized with 0.2% Triton X-100 in PBS for 10 min and blocked with 5% BSA/PBS-T for 30 min. Rabbit polyclonal anti-TRPM4 antibody (alomone ACC-044; 1:100 dilution) was added in 1% BSA/TBS-T for overnight at 4°C. The sections were washed in TBS-T and incubated with Alexa Fluor 488 conjugated donkey anti-rabbit IgG (Invitrogen A21206, 1:1000 dilution) and Alexa-594-conjugated cholera toxin B (CTB) (Invitrogen C34777) for 1 h. All slides were imaged using a confocal microscope (Olympus, Fluoview1000, Japan). To detect cholesterol levels in the plasma membrane of mpkCCDc₁₄ cells, the cells were incubated with 5 μg/ml filipin (Sigma, Cat#: F9765) for 30 min. Filipin staining was viewed by confocal microscope using DAPI filter. The control fluorescent intensity is used as a calibrator, and relative fluorescent intensity is calculated against this calibrator. All slides were imaged using a confocal microscope (Olympus, Fluoview1000, Japan) and analyzed using Olympus Fluoview FV1000 version 3.1 software. Identical acquisition settings were used for all images. To quantify colocalizations, the image analysis program ImageJ was used. Both Pearson and Manders coefficients were calculated.

To test the lipid-binding properties of TRPM4, Protein lipid overlay assays were performed using PIP Strips from Invitrogen (Chicago, IL, United States) as previously described (Zhang et al., 2010). The strip is a piece of nitrocellulose membrane on which 15 phospholipids at 100 pM and a blank sample were loaded by the manufacturer. Briefly, strips were blocked in TBS-Tween (0.1%, TBS-T) and 3% BSA for 1 h. The mpkCCDc₁₄ cells lysate was then diluted in the blocking buffer to 0.5 μg·ml⁻¹ and incubated overnight at 4°C. To detect the possible binding of TRPM4 to spotted phospholipids, the Strips were incubated with rabbit polyclonal antibodies directed against TRPM4 (1:200; ACC-044; Alomone Labs, Jerusalem, Israel) similar to the western blot method.

All chemicals for electrophysiological recordings were purchased from Sigma-Aldrich (St Louis, MO, United States) except when specified. DiC8-PI(4,5)P₂ was purchased from Echelon Biosciences. Both wortmannin and prostaglandin E2 (PGE2) were used to pre-treat mpkCCDc₁₄ cells for 30 min for patch-clamp experiments.

Data are reported as mean values ±SEM. Statistical analysis was performed with GraphPad Prism 5 software (GraphPad; La Jolla, CA) was used for all statistical calculations. Student t test was used between two groups. Analysis of variance was used for multiple comparisons. Results were considered significant if P < 0.05.

To manipulate the plasma membrane cholesterol content, lovastatin and exogenous cholesterol were used as we previously described (Song et al., 2014). The mpkCCD_c14 cells were treated with 5 μM lovastatin, 30 μg/ml exogenous cholesterol, or 5 μM lovastatin plus 30 μg/ml exogenous cholesterol for 48 hrs. Then, plasma membrane cholesterol levels in mpkCCD_c14 cells were evaluated by filipin staining. The data show that exogenous cholesterol significantly increased, whereas lovastatin significantly decreased, the cholesterol levels in the plasma membrane of mpkCCDc₁₄ cells, and that co-treatment of the cells with both lovastatin and exogenous cholesterol did not alter cholesterol levels (Figures 1A,B). To further determine whether these treatments affect TRPM4 channel activity by regulating its sensitivity to Ca²⁺, the channel activity was recorded by exposing the patch membrane to the bath containing different concentrations of free Ca²⁺ using the inside-out patch-clamp technique. We have previously demonstrated that the concentration of Ca²⁺ required for 50% of maximal activation of TRPM4 (EC50) was ∼32.6 μM under basal conditions without manipulation of membrane cholesterol levels (Wu et al., 2016). Here, we found that the EC50 after reducing membrane cholesterol with lovastatin was only ∼5.76 mM and that this effect was significantly reversed to ∼10.2 μM by enrichment of membrane cholesterol with exogenous cholesterol (Figure 2). These results indicate that elevation of membrane cholesterol increases the TRPM4 channel activity by enhancing its sensitivity to Ca²⁺ in mpkCCD_c14 cells.

To further determine whether acute depletion of cholesterol with MβCD affects TRPM4 channel activity, we performed excised inside-out patch-clamp experiments in mpkCCD_c14 cells. Here, we found that acute extraction of cholesterol out of the inner leaflet of the patch membrane with MβCD significantly reduced TRPM4 Po and that application of exogenous cholesterol reversed the reduction of TRPM4 channel activity induced by MβCD (Figures 3A,B). These data together suggest that cholesterol in the inner leaflet of the plasma membrane is required for maintaining TRPM4 activity.

We have previously shown that cholesterol regulates ROMK channels by altering PI(4,5)P₂ localization (Liu et al., 2015). It is known that PI(4,5)P₂ stimulates TRPM4 channels (Nilius et al., 2006). Therefore, we would ask whether PI(4,5)P₂ is required for cholesterol to stimulate TRPM4. Wortmannin, at high concentrations, is a PI4K inhibitor, therefore preventing the synthesis of PI(4,5)P₂ and depleting membrane PI(4,5)P₂ (Saleh et al., 2009). Our data show that depletion of PI(4,5)P₂ by wortmannin (20 μM) abrogated exogenous cholesterol-induced TRPM4 channel activity (Figures 4A,B). In contrast, application of 20 nM wortmannin, which is unable to alter PI(4,5)P₂ levels, had no effects on cholesterol-induced TRPM4 channel activity (Figures 4C,D). Since PGE2 depletes PI(4,5)P₂ via activation of Gq-coupled EP1 receptors (Harraz et al., 2018), PGE2 (2 μM) was also used to examine whether cholesterol can stimulate TRPM4 without PI(4,5)P₂. Our data showed that PGE2 abolished the stimulatory effects of exogenous cholesterol on TRPM4 activity, whereas exogenous PI(4,5)P₂ (diC8-PI(4,5)P₂, a water-soluble analog) increased the effects (Figures 4E,F). These data suggest that elevation of plasma membrane cholesterol stimulates TRPM4 via a PI(4,5)P₂-dependent mechanism.

Previous studies have demonstrated that PI(4,5)P₂ is a strong positive modulator of TRPM4 (Nilius et al., 2006). Consistently, our inside-out data show that diC8-PI(4,5)P₂ (20 μM) significantly increased the TRPM4 activity (Figures 5A,B). To further determine whether TRPM4 can bind to PI(4,5)P₂, lipid-protein overlay experiments were performed by using PIP Strips. Our data show that TRPM4 physically binds to almost all phosphatidylinositols (PI) including PI(4,5)P₂ (Figure 5). These data suggest that PI(4,5)P2 stimulates TRPM4 channels probably via a physical interaction.

Our previous studies have shown that cholesterol in lipid rafts maintains PI(4,5)P₂ in lipid rafts (Liu et al., 2015). To determine whether TRPM4 is also located in lipid rafts to physically interact with PI(4,5)P₂, we labeled lipid rafts with fluorescence-tagged cholera toxin (CTX) and TRPM4 with its specific antibody. Quantitative analysis with ImageJ showed that TRPM4 channel was colocalized with lipid rafts (Pearson coefficient was 0.726 ± 0.049 and Manders coefficients were 0.902 ± 0.042 [M₁], red color and 0.932 ± 0.037 [M₂], green color). Confocal microscopy data showed that TRPM4 channel was co-localized with lipid rafts (Figure 6A). Consistently, the data from sucrose density gradient assays also show that fractions 2–5 are denoted as the lipid raft fractions as indicated by caveolin-1, a marker of membrane lipid rafts and TRPM4 was mainly enriched in lipid raft membranes in mpkCCD_c14 cells (Figure 6B). These data indicate that plasma membrane cholesterol stimulates TRPM4 by holding PI(4,5)P₂ in lipid rafts.

Since plasma membrane cholesterol could regulate TRPM4 activity, we examined whether manipulation of cholesterol levels will affect the expression of TRPM4 in mpkCCD_c14 cells. Confocal microscopy experiments were performed using control cells and cells treated for 48 hrs with 30 μg/ml exogenous cholesterol, 5 μM lovastatin, or 5 μM lovastatin plus 30 μg/ml exogenous cholesterol. These data showed that enrichment of cholesterol or inhibition of cholesterol biosynthesis has no effect on TRPM4 expression in mpkCCDc14 cells (Figures 7A,B). To confirm the results from confocal microscopy experiments, Western blot and cell-surface biotinylation assay data also showed that the total and membrane levels of TRPM4 in mpkCCD_c14 cells were unaltered by cholesterol enrichment or inhibition of cholesterol biosynthesis (Figures 7C,D). To further examine whether manipulating the membrane cholesterol of mpkCCD_c14 cells can affect TRPM4 cell membrane abundance, the number of active channels was recorded by exposing the patch membrane to the bath containing 5 mM CaCl_2, followed by a bath solution with 10 mM EGTA and no calcium. Our data showed that neither cholesterol enrichment nor inhibition of cholesterol biosynthesis affects the number of active channels in the patches from mpkCCD_c14 cells (Figures 7E,F). Thus, the potentiating effect of cholesterol on TRPM4 activity cannot be attributed to an enhanced its surface expression.