The cDNA of the human TRPM3 channel (accession number AJ505026) with C-terminal GFP tag was provided by Christian Harteneck (University of Tübingen, Tübingen, Germany) (Grimm et al., 2003). Alternative splicing patterns of TRPM3 is highly conserved across human and rodents (Oberwinkler et al., 2005). To date, 25 isoforms of mTRPM3 protein have been identified, including a recently discovered variant - TRPM3γ3. Splicing events affect exons 8, 13, 15, 17, 20, 24, and 28 of TRPM3. α variants lack exon 2; β variants lack exon 1; and γ variants lack a large part of exon 28 (Oberwinkler and Philipp, 2014; Uchida et al., 2019). Our hTRPM3 cDNA contains all 30 exons, where 389 amino acids in exon 28 have been replaced with the alternative carboxy terminus of seven residues.This truncation does not affect functional activity of the ion channel (Grimm et al., 2003; Oberwinkler and Philipp, 2014). Mutations of hTRPM3-GFP were generated by site-directed mutagenesis (performed by GENEWIZ Inc.). The predicted mutations were verified by sequencing analysis.

Human embryonic kidney (HEK)-293 cells were used to transiently overexpress wild-type hTRPM3-GFP and its mutants. The cells were grown in DMEM/F12 medium (Fisher Scientific, Waltham, MA, United States. Catalog no. MT10090CV) supplemented with 10% bovine growth serum (HyClone, Logan, UT, United States. Catalog no. SH30541.03), 100 U/ml penicillin/100 mg/ml streptomycin (Fisher Scientific, catalog no. SV30010) at 37°C in a 5% CO₂, humidity-controlled incubator. Lipofectamine 2000 (Thermo Fisher Scientific, Waltham, MA, United States. Catalog no. 18324012) was used for the transfection of TRPM3 into the cells in a 35 mm culture dish according to the manufacturer’s instructions. Successfully transfected cells were identified by their fused GFP when illuminated at 480 nm excitation wavelength. Electrophysiological recordings were conducted between 36- and 48-h post-transfection.

All patch-clamp experiments were performed at room temperature (20–22°C). TRPM3 whole-cell currents were recorded using an Axopatch 200B amplifier. Data were digitized at 10 kHz and digitally filtered offline at 5 kHz. Patch electrodes were pulled by Sutter P-97 micropipette puller and fire-polished to the resistance of 3–5 MΩ when filled with internal solutions. Series resistance (Rs) was compensated up to 90% to reduce series resistance errors to <5 mV. Cells in which Rs was >8 MΩ were discarded (Du et al., 2009).

For whole-cell current recording, ramp voltage stimuli (250 ms duration) were delivered at 1 s intervals and ranging from −100 to +100 mV. The internal pipette solution for whole-cell current recordings contained (in mM): 115 Cs-methanesulfonate (CsSO₃CH₃), 8 NaCl, 10 Cs-EGTA, 5 Na₂-ATP and 10 HEPES, with pH adjusted to 7.2 with CsOH. In high intracellular Ca²⁺ experiments, 0.93 mM CaCl₂ was added to the above-mentioned intracellular solution and EGTA was reduced to 1 mM, resulting in 1 µM free intracellular Ca²⁺. MaxChelator (https://somapp.ucdmc.ucdavis.edu/pharmacology/bers/maxchelator/downloads.htm) software from the University of California, Davis was used to calculate free [Ca²⁺]_i. To avoid proton activated Cl⁻ currents conducted by endogenous anion channels of HEK-293 cells (Lambert and Oberwinkler, 2005), NaCl in standard Tyrode solution was replaced with Na-glutamate for all whole-cell current recordings. This external solution contained (in mM): 145 Na-glutamate, 5 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, and 10 glucose, with pH adjusted to 7.4 with glutamic acid. Internal and external acidic pH solutions were prepared as described previously with slight modifications (Du et al., 2009). In brief, 10 mM HEPES was replaced by 10 mM MES for the solutions at pH ≤ 6.0. Bath solutions containing 1 mM–60 mM NH₄Cl were prepared by decreasing Na⁺ concentrations to 85 mM to keep the osmolarity constant Osmolarity was adjusted to 300 ± 10 mOsm with mannitol. In experiments designed to test protons permeability of TRPM3, pipette solutions contained (in mM): 120 NMDG, 108 glutamic acid, 10 HEPES, 10 EGTA, with pH adjusted to 7.2 with NMDG. External solutions for testing proton permeability contained (in mM): 145 NMDG, 10 HEPES, and 10 Glucose; and pH was adjusted with glutamic acid. To prepare the pH 5.5 external solution for proton permeability, 10 mM HEPES was replaced with 10 mM MES. PS was dissolved in DMSO to prepare 100 mM stock solution, and an adequate volume of stock PS solution was added to the external solution to achieve the required concentration. All the chemicals used in electrophysiological experiments were from Sigma-Aldrich, St. Louis, MO, United States.

Statistical data were analyzed using GraphPad Prism 8. Pooled data are presented as mean ± SEM. Concentration-response curves were fitted by an equation of the form: E = E _max {1/[1+(EC₅₀/C)ⁿ]} where E is the effect at concentration C, E _max is the maximal effect, IC₅₀ is the concentration for half-maximal effect, and n is the Hill coefficient (Du et al., 2009). The concentration of proton required for half-maximal inhibition is denoted by IC₅₀ (when proton concentration is expressed as molar concentration) and pH₅₀ (when proton concentration is expressed by pH value). Statistical comparison of two groups was performed by unpaired Student’s t-test, with p < 0.05 considered statistically significant. Statistical comparison of three or more groups was performed by one-way ANOVA with Tukey’s post hoc multiple comparisons.

We studied the effects of low pH on TRPM3 by overexpressing hTRPM3-GFP in HEK-293 cells and recording whole-cell currents in response to PS. We found that low extracellular pH inhibited TRPM3 in a reversible manner (Figures 1A–C). To avoid proton activated endogenous anion channel conducted Cl⁻ currents, external Cl⁻ was replaced by glutamate (See Methods). The inhibitory effect of low extracellular pH (pH_o) was only observed below pH 6.0. Specifically, at pH_o 7.0 and 6.0, recorded TRPM3 currents were equivalent to pH_o 7.4 (p > 0.05 in both groups). At a pH_o below 6.0, acidic conditions exhibited significant inhibition of TRPM3. pH_o 5.5, caused ∼60% reduction in TRPM3 whole-cell current induced by PS (Figure 1). Although the concentration-dependent effect of extracellular pH on TRPM3 activity was weak, fitting these data in a non-linear regression curve resulted in a concentration-dependent relationship with a pH₅₀ value of 5.5 ± 0.13, and, Hill coefficient of n_H = 2.5 (Figure 1D). When PS was applied continuously while reducing extracellular pH (Figures 1E–G), a concentration-dependent relationship was observed with a pH₅₀ value of 5.5 ± 0.17, and a Hill coefficient of n_H = 2.6 (Figure 1H). To further measure the inhibitory effect of low extracellular pH on TRPM3, we held the membrane potential at +60 mV continuesly during the recordings. We found that low pH significantly blocked PS-induced TRPM3 currents (Supplementary Figure S1). Interstingly, as shown in Figure 1A, application of low extracellular pH (pH ≤ 5.5) with PS produced an initial activation of TRPM3 before blocking it. This suggested that the onset of low pH_o inhibition is slower than the PS activation. Because acidic pH has been shown to block TRPM channels intracellularly (Du et al., 2009; Starkus et al., 2010; Yang et al., 2010), we next asked if intracellular low pH could inhibit TRPM3. To answer this question, we first investigated the effects of intracellular low pH (pH_i) on TRPM3. Whole-cell TRPM3 currents were recorded using the low-pH pipette solutions (see experimental procedures) while keeping extracellular pH constant at 7.4 (Figure 2A). To investigate the extent of modulation of TRPM3 by protons, we also introduced higher pH_i than the physiological pH_i of 7.2. we found that the TRPM3 current plateaued at about pH_i 7.6. Acidic pH had similar inhibitory effects on both the outward and inward TRPM3 currents (Figure 2A). There was no significant difference in the steady-state inhibition between inward and outward currents, suggesting that there were no voltage-dependent effects of acidic pH_i on inward and outward TRPM3 currents (Figure 2B). Low pH_i markedly reduced TRPM3 current in a concentration-dependent manner with a pH₅₀ value of 6.90 ± 0.11 (outward current at +100 mV) (Figure 2C) and pH₅₀ value of 6.90 ± 0.15 (inward current at −100 mV) (Figure 2D). The Hill coefficient was n_H = 1.6. Combined, our findings suggesting that extracellular acidic pH 6.0 did not affect TRPM3 current but low intracellular low pH had an inhibitory pH₅₀ value of 6.9. Thus, these data suggest that TRPM3 sensitivity to pH was stronger on the cytoplasmic side.

PS binds directly to the extracellular side of the TRPM3 channel to activate it. TRPM3 channel stays both functional and unaffected by the presence of intracellular PS (Wagner et al., 2008). Previous studies have shown that PS activates TRPM3 in a concentration-dependent manner with an EC₅₀ value of 12 and 23 µM for outward and inward current respectively (Wagner et al., 2008). To investigate the effect of protons on PS concentration-dependent activation of TRPM3, we perfused the cells with a wide range of PS concentrations (1–500 µM). pH_i was held constant at 7.2 or 6.0. TRPM3 whole-cell outward and inward currents showed very similar PS concentration-dependence in both pH_i conditions (Figures 3A,B). The EC₅₀ values for the outward currents were 16 µM (pH_i 7.2) and 15 µM (pH_i 6.0) (Figure 3C). This result suggested that higher proton concentrations inside the cell decreased TRPM3 maximal activation at any given PS concentration. However, protons did not affect the PS concentration-dependent activation of TRPM3 as EC₅₀ values for the inward currents were 21 µM (pH_i 7.2) and 26 µM (pH_i 6.0) (Figure 3D). This result indicated that the PS concentration-dependent curve of TRPM3 inward currents did not show any change in response to low pH_i. In summary, these data suggested that intracellular protons do not compete with PS for binding sites.

Extracellular application of NH₄Cl produces a rise in pH_i resulting from an influx of NH₃. NH₃ binds intracellular protons and causes alkalization inside the cells (Jacobs, 1922; Warburg, 1922). We observed that protons blocked TRPM3 from the cytoplasmic side. Thus, we tested whether raising pH_i by applying NH₄Cl could rescue the TRPM3 currents. We perfused the cells with solutions where part of NaCl was replaced by an equal amount of NH₄Cl (see experimental procedures), to test whether increasing pH_i while keeping extracellular pH the same can reverse the blocking effects of protons on TRPM3. As shown in Figure 4A,B, 30 and 60 mM NH₄Cl potentiated the PS-induced TRPM3 currents significantly, while potentiation was only slightly different on the outward and inward currents (Figure 4C) We attribute this increase to the influx of NH₃ into the cell and removing bound protons from the TRPM3 cytoplasmic side. Overall, this result supported our finding that TRPM3 could be blocked by an acidic intracellularpH Indeed, different concentrations of NH₄Cl were applied to the same cell, and all NH₄Cl applications changed intracellular pH. To minimize the residual effects of NH₃, we allowed adequate washing time between two consecutive NH₄Cl + PS applications (Figure 4D–F).

Subsequently, we tested whether external NH₄Cl exhibited a concentration-dependent rescuing effect on TRPM3 activity or not. We perfused the cells with modified Tyrode solutions containing different concentrations of NH₄Cl (1–60 mM) (Figure 4D–F). Na⁺ concentrations in all solutions were kept the same, and osmolarities were adjusted to 300 ± 10 mOSM with mannitol. We observed a five-fold increase in TRPM3 activity at only 3 mM (NH₄Cl) concentration (Figure 4D). Even 1 mM NH₄Cl rescued some TRPM3 activity (p < 0.05) (Figure 4D). Although, when compared with 3 mM NH₄Cl, some of the higher concentrations showed a significant increase in TRPM3 activity, there was no substantial increase in TRPM3 activity by further increasing the NH₄Cl concentration. The PS concentration-dependent manner with a EC₅₀ value of 2.7 ± 0.3 µM, the Hill coefficient is nH = 1.1 for the outward currents (Figure 1G) and inward currents (Figure 1H). We also observed a higher increase in inward current than corresponding outward current (Figure 4I), with the ratio of inward current to the outward current increasing from 0.13 at 1 mM NH₄Cl to 0.54 at 60 mM NH₄Cl. When plotting the correlation between the outward and inward currents, we observed a linear realationship between the outward and inward currents (Figure 1I). It is important to note that because we did not measure the exact pH_i changes following external NH₄Cl applications, the pH_i differences before and after NH₄Cl applications, are unknown. However, despite this limitation, these experiments support the claim that TRPM3 activity is highly sensitive to changes in intracellular pH changes in the physiological pH_i range (pH_i 6.0–8.0). ITo ensure that application of NH₄Cl and PS for an extended period of time did not affect TRPM3 activity and thus confound our results, we also examined individual cells following only one exposure to different concentrations of NH₄Cl. The results of these experiments are summarized in Supplementary Figure S2A. Perfusion of individual cells with different concentrations of NH₄Cl (Supplementary Figure S2B) showed similar response when compared to TRPM3 currents from our earlier experiments (Figure 4D). In summary, these data suggest that the increase of intracellular pH boosts TRPM3 activities.

The lipid bilayer (eg., cell membrane) presents a strong barrier for the transport of charged ions through eukaryotic cell membranes. Although the permeability of protons is higher than other monovalent cations (Tepper and Voth, 2005), it is unlikely that these mechanisms can transport sufficient protons across the membrane to have a direct impact unless protons are passing through the overexpressed ion-channel itself. We thus examined whether protons can permeate through TRPM3. To test this, we recorded TRPM3 inward current in HEK-293 cells, overexpressing hTRPM3-GFP, by holding the membrane at −100 mV and applying PS and low pH solutions externally. We maintained pH_i at 7.6 to provide a higher concentration gradient for protons. All cations except protons were removed by using NMDG in the external and internal solutions (Figure 5A). PS, along with NMDG pH 5.5 extracellular solution, produced a small transient inward current in TRPM3 transfected HEK-293 cells (Figure 5B). Indeed, the amplitude of this current was significantly lower than the TRPM3 currents observed using theTyrode solution (Figures 5C,D). This result was likely due to removal of all cations besides a limited amount of protons, with the extracellular and intracellular proton concentrations calculated to be 3.2 × 10^–3 mM (pH 5.5) and 2.5 ×10^–5 mM (pH 7.6), respectively. These cation concentrations are several orders of magnitude lower than the total cation concentrations of the Tyrode solution. However, since there were no other cations in our solution and although low in amplitude, the current we observed suggests the passage of proton through the TRPM3 channel. We spectulate that the transient nature of the current is because the intracellular protons inhibit TRPM3 after their permeation. Mock transfected cells did not produce any inward current in response to PS (Data not shown). This evidence indicates that TRPM3 is potentially permeable to protons.

TRPM3 displays higher permeability for divalent cations than monovalent cations. For the splice variant TRPM3α2, 24% of total TRPM3 current is expected to result from Ca²⁺ (Przibilla et al., 2018), suggesting a large increase in [Ca²⁺]_i following TRPM3 activation. Multiple studies verified this effect showing an increase in [Ca²⁺]_i in a micromolar range following TRPM3 activation (Vriens et al., 2011; Straub et al., 2013; Przibilla et al., 2018). A recent study demonstrated that an increase in [Ca²⁺]_i, independent of TRPM3 activity, inhibits TRPM3 in a calmodulin-dependent manner (Przibilla et al., 2018). This suggests a potential negative feedback mechanism that regulates a high increase in [Ca²⁺]_i resulting from TRPM3 activation. Studies conducted by others enlist [Ca²⁺]_i as a regulator of TRPM3 activity. In our study, we found that intracellular protons block TRPM3. Hence, we asked the question, how do these two regulatory mechanisms interact with each other? To find the effect of [Ca²⁺]_i on the concentration-dependent inhibition of TRPM3 by protons, we tested TRPM3 activity in two different [Ca²⁺]_i conditions, while providing a wide range of pH_i. In addition to the inhibition of TRPM3 current by [Ca²⁺]_i observed by Przibilla (Przibilla et al., 2018), we observed an additionally delayed inhibition of TRPM3 current by Ca²⁺. For example, low [Ca²⁺]_i (<10 nM) did not show a delayed inhibition of TRPM3 current (red point, Figure 6A). Under 1 µM [Ca²⁺]_i, the recorded current showed a further inhibition after initial activation and the residual current amplitudes were less than 50% of the initial activation (red point, Figures 6B,C). The inhibition was pH dependent (Figure 6D). However, the pH_i concentration-dependence of TRPM3 was not affected by [Ca²⁺]_i, as they showed similar pH₅₀ values (7.0 ± 0.1, n_H = 2.1; 7.1 ± 0.1, n_H = 1.4; 7.2 ± 0.5, n_H = 1.6) for low [Ca²⁺]_i and high Ca²⁺, (Figure 6E). It is noteworthy that despite having similar pH₅₀ values, TRPM3 outward current densities were significantly decreased in high [Ca²⁺]_i conditions (Figure 6E). We also analyzed the ratio of delayed current intensities compared to the initial current intensities. We did not observe any significant difference between the ratios resulting from different pH_i (Figure 6F). Overall, these results indicate that, although [Ca²⁺]_i inhibits TRPM3, it does not affect the regulation of TRPM3 by intracellular protons.

Glutamate, aspartate, histidine, and lysine residues are potential proton acceptors, especially glutamate and aspartate, which present negative charges (Zhou and Pang, 2018). To identify the amino acid residues of TRPM3 accountable for its sensitivity to pH_i, we prepared hTRPM3-GFP mutant plasmids with mutations in the pore region. We selected all eight glutamate and aspartate amino acids, in the vestibule of the loop between S5 and S6 transmembrane domains (Figure 7A). These eight residues were either glutamate and aspartate amino acid and mutated to glutamine. We created two double mutants (E1034Q–E1035Q, E1072Q–D1073Q) and six single mutants (E1055Q, D1059Q, D1062Q, E1069Q, E1072Q, and D1073Q). We expressed these mutants in HEK-293 cells and recorded elicited currents in response to PS, while perfusing with physiological and low pH external solutions (Figure 7B). Except for E1055Q, which resulted in a non-functional ion-channel, all other mutants exhibited identical I-V relation to WT-hTRPM3, although most of the mutants showed markedly reduced current amplitudes (Figure 7B). We perfused cells expressing these mutants with external solutions of pH 7.4 and pH 5.5. The pH of 5.5 was selected as representative of low pH external solutions because at this pH, WT-TRPM3 currents were blocked significantly, yet sufficient activity remained for analysis. We compared the percent decrease in TRPM3 current from pH_o 7.4 to pH_o 5.5 for all the mutants and WT-TRPM3 (Figure 7C). The double mutant (E1034Q–E1035Q), E1069Q, and E1072Q showed similar sensitivity to protons when compared with the WT-TRPM3. Mutants D1059Q and D1062Q were found to be more sensitive to protons, as the reduction of whole-cell current due to low pH_o was increased to 93.9 and 84.1% in D1059Q and D1062Q, respectively, compared to 56.5% observed in WT-TRPM3. Mutant D1073Q showed significantly less sensitivity toward protons, as pH_o 5.5 reduced its current amplitude only by 25%. As summarized in Figure 7C, these results establish the amino acid residues D1059Q, D1062Q, and D1073Q as key determinants of proton sensitivity in TRPM3. Although it is unclear at this moment why mutating these residues produce variable effects (increased or decreased proton sensitivity), these data suggested that the pore vestibule of TRPM3 is critical for pH sensitivity. Further studies are required to delineate the underlying mechanism responsible for these variable effects.