The crystal structure of TRPM2 MHR1/2 domain reveals a conserved Zn2+ -binding domain essential for structural integrity and channel activity.

Transient receptor potential melastatin 2 (TRPM2) is a Ca2+ -permeable, nonselective cation channel involved in diverse physiological processes such as immune response, apoptosis, and body temperature sensing. TRPM2 is activated by ADP-ribose (ADPR) and 2'-deoxy-ADPR in a Ca2+ -dependent manner. While two distinct binding sites exist for ADPR that exert different functions dependent on the species, the involvement of either binding site regarding the superagonistic effect of 2'-deoxy-ADPR is not clear yet. Here, we report the crystal structure of the MHR1/2 domain of TRPM2 from zebrafish (Danio rerio), and show that both ligands bind to this domain and activate the channel. We identified a so far unrecognized Zn2+ -binding domain that was not resolved in previous cryo-EM structures and that is conserved in most TRPM channels. In combination with patch clamp experiments we comprehensively characterize the effect of the Zn2+ -binding domain on TRPM2 activation. Our results provide insight into a conserved motif essential for structural integrity and channel activity.

2′‐deoxy‐adenosine diphosphate ribose

adenosine diphosphate ribose

Dulbecco's Modified Eagle Medium

Danio rerio (zebrafish)

ethylenediaminetetraacetic acid

electron microscopy

Homo sapiens (human)

immobilized metal affinity chromatography

isopropyl β‐d‐1‐thiogalactopyranoside

isothermal titration calorimetry

TRPM homology region

nicotinamide adenine dinucleotide

Ni‐nitrilotriacetic acid

N‐methyl‐d‐glucamine

poly‐ADPR glycohydrolase

poly‐ADPR polymerase

reactive oxygen species

single‐wavelength anomalous diffraction

sodium dodecyl sulfate‐polyacrylamide gel electrophoresis

standard deviation

tobacco etch virus

transient receptor potential melastatin

INTRODUCTION

Members of the melastatin subfamily of transient receptor potential channels (TRPM channels) are widely expressed and contribute to cellular Ca2+ signaling either directly or indirectly. They play an important role in physiological processes such as temperature sensing and regulation (TRPM8, TRPM2 , ), the immune response, Mg2+ homeostasis (TRPM6/TRPM7 ), taste transduction (TRPM5 ), the response to oxidative stress (TRPM2 ) and apoptosis (TRPM2 ). Many diseases are linked to TRPM proteins explaining the increased attention they receive as potential drug targets. , , The activity of TRPM channels is regulated by various influences such as voltage and temperature as well as changes in concentrations of small molecules, ions, or lipids (reviewed in the literature ). Structurally, all TRPM family members share a common core architecture: N‐terminal TRPM homology regions (MHR1‐4), six transmembrane helices, a TRP helix and a C‐terminal coiled‐coil domain.

The Ca2+‐permeable, nonselective cation channel TRPM2, which is involved in cell death, diverse immune cell functions, temperature sensing and the control of body temperature harbors a unique C‐terminal domain. The name of this domain (NUDT9‐H) arises from the homology to the Nudix box enzyme NUDT9, a soluble pyrophosphatase, that hydrolyzes adenosine 5′‐diphosphoribose (ADPR). Due to this homology, the NUDT9‐H domain was predicted to contain a binding pocket for ADPR, the endogenous ligand of TRPM2. Although human TRPM2 does not possess ADPR hydrolase activity, the recently determined structure of human TRPM2 illustrated that the NUDT9‐H domain in fact binds ADPR. However, the cryo‐EM structure of TRPM2 from zebra fish revealed a distinct ADPR‐binding site in the N‐terminal MHR1/2 domain. Since this site is conserved across TRPM2 from different species including humans, human TRPM2 harbors two distinct ADPR‐binding pockets. , , The current view of gating of TRPM2 by ADPR is, that upon ADPR binding in the N‐terminal domain, the clamshell‐like shape of the MHR1/2 domain closes, inducing a rotation. This rotation and binding of Ca2+ at the membrane‐cytosol interface as well as binding of a second ADPR molecule within the NUDT9‐H domain lead to further conformational changes in the tetrameric channel provoking an activated state. , ,

The cellular nucleotide ADPR is a metabolite of nicotinamide adenine dinucleotide (NAD) and can arise from the hydrolysis of NAD by the multifunctional enzyme CD38. , Another source is poly‐ADPR, synthesized by sequential attachment of ADPR derived from NAD to proteins by the poly‐ADPR polymerase (PARP). The hydrolysis of poly‐ADPR and the formation of free monomeric ADPR are catalyzed by the poly‐ADPR glycohydrolase (PARG) and the terminal ADPR protein glycohydrolase (TARG). The PARP/PARG pathway is triggered by oxidative stress explaining the connection between the oxidative stress response and TRPM2. Another TRPM2 activator is 2′‐deoxy‐ADPR, which proved to be a more effective TRPM2 agonist than ADPR. 2′‐Deoxy‐APDR could also be detected endogenously and thus may act as second messenger.

Like a number of other Ca2+ channels, the human TRPM channels TRPM3, TRPM6, and TRPM7 as well as the homologous single TRPM channels from Drosophila melanogaster have been shown to be permeable to Zn2+ and to mediate Zn2+ entry upon activation. While TRPM2 is inhibited by high concentration of extracellular Zn2+, it has been shown that the cytosolic Zn2+ concentration in TRPM2‐expressing cells increases upon activation of a photoactivatable ADPR analogue. This indicates that TRPM channels contribute to the cellular Zn2+ homeostasis. Zn2+ is an essential trace metal that is required as a cofactor for some enzymes and as a structural component for a large number of proteins (reviewed in the literature ). It has also been considered to exert signaling function and act as second messenger. Similar to Ca2+, high concentrations of Zn2+ are cytotoxic and contribute to excitotoxicity and other pathological processes. A recent study of the role of TRPM2 in cell death of microglial cells showed that extracellular Zn2+ could also activate TRPM2 through multiple steps, involving reactive oxygen species (ROS) that induce PARP/PARG. While the mechanisms remain unknown, there is evidence that also a change in the intracellular Zn2+ concentration influences the oxidative stress response and TRPM2. ,

Another member of the TRP family, TRPC5, comprises an intracellular Zn2+‐binding motif that is conserved within TRPC channels and suggests a direct modulatory role for Zn2+ ions on these related proteins. ,

In order to improve our understanding of the regulatory role of ADPR and 2′‐deoxy‐ADPR on the TRPM2 channel, we biophysically characterized ligand binding to the N‐terminal MHR1/2 domain of zebrafish TRPM2 and electrophysiologically investigated possible functional consequences. The crystal structure of this domain revealed a novel Zn2+‐binding motif that is conserved within the TRPM family and is essential for structural integrity and activity of the channel.

RESULTS AND DISCUSSION

Biophysical and electrophysiological characterization of ADPR and 2′‐deoxy‐ADPR binding to drMHR1/2

The N‐terminal MHR1/2 domain of TRPM2 contains a binding site for the channel‐activating ligand ADPR, which has been identified in TRPM2 cryo‐EM structures from human and zebrafish. , Whether the TRPM2 superagonist 2′‐deoxy‐ADPR binds to the same site is currently unknown. In order to investigate the thermodynamic parameters of ligand binding to MHR1/2 and to confirm the structural as well as functional integrity of our sample, we determined and compared binding of ADPR and 2′‐deoxy‐ADPR to the isolated zebrafish (Danio rerio) MHR1/2 domain (drMHR1/2) using several biophysical techniques.

Isothermal titration calorimetry (ITC) revealed endothermic binding in the low micromolar range for both ligands with 2′‐deoxy‐ADPR showing a slightly tighter binding than ADPR (Figure 1a). To investigate whether ligand binding has an effect on the stability of the protein, we used nano differential scanning fluorimetry (nDSF) titrations. We found that ADPR as well as 2′‐deoxy‐ADPR both concentration‐dependently stabilize the drMHR1/2 domain (Figure 1b; Figure S1).

FIGURE 1

Biophysical and electrophysiological characterization of ADPR and 2′‐deoxy‐ADPR binding to drMHR1/2. (a) Binding of zebrafish MHR1/2 (drMHR1/2) and ADPR/2′‐deoxy‐ADPR (2dADPR) as measured by isothermal titration calorimetry (ITC). KD values are averages of triplicate experiments. (b) Binding of drMHR1/2 and ADPR/2dADPR shown by shift of melting temperature observed by differential scanning fluorimetry (nDSF). (c) Whole‐cell patch clamp experiments with HEK293 cells transiently transfected with zebrafish TRPM2 (drTRPM2). Intracellular solution contained 100 nM free Ca2+ with 100 μM ADPR/2dADPR or without ADPR/2dADPR (buffer control). Extracellular solution contained 100 μM Ca2+. Mean maximum currents do not differ significantly between ADPR (130 pA) and 2dADPR (146 pA).

2′‐deoxy‐ADPR acts as a superagonist of human TRPM2 that induces significantly higher currents than ADPR. The mechanisms behind the longer open times and slower inactivation caused by 2′‐deoxy‐ADPR are currently unknown, neither is known whether this behavior is unique to the human channel or holds true for other orthologues as well. To address whether 2′‐deoxy‐ADPR also acts as a superagonist of zebrafish TRPM2 (drTRPM2), we transiently transfected HEK293 cells with an expression vector for drTRPM2 and performed whole‐cell patch clamp experiments (Figure 1c, Figure S2). Under the conditions we normally use for the human channel, we observed much higher currents than with HEK293 cells transfected with human TRPM2 when ADPR was infused, in agreement with a recent report by Kühn and coworkers. Under these conditions, most cells ruptured during the recording. We thus decided to reduce the extracellular Ca2+ concentration to 100 μM to reduce the driving force for Ca2+ and thus the positive feedback by Ca2+. Under these conditions, we could follow channel activation by both ADPR and 2′‐deoxy‐ADPR. The maximum current during the recording does not differ significantly upon activation with ADPR or 2′‐deoxy‐ADPR, respectively, indicating that 2′‐deoxy‐ADPR does not act as a superagonist of drTRPM2. This lack of superagonist effect of 2′‐deoxy‐ADPR on drTRPM2 could reflect differences in the relative affinity of the MHR1/2 domain from hsTRPM2 and drTRPM2 for ADPR and 2′‐deoxy‐ADPR. Unfortunately, we were not able to express and purify the isolated human MHR1/2 domain in E. coli.

Alternatively, it could indicate a role of the second nucleotide binding site (NUDT9‐H domain) in the superagonistic effect of 2′‐deoxy‐ADPR. In this regard, it is of interest that it has recently been shown that nvTRPM2 and drTRPM2 on one hand and hsTRPM2 on the other hand differ in their response to some ADPR analogues: Inosine 5′‐diphosphoribose (IDPR), a partial agonist of human TRPM2 that is assumed to bind to the NUDT9‐H domain (it is a substrate of NUDT9), does not activate drTRPM2, whereas 8‐(thiophen‐3‐yl)‐ADPR and 8‐(3‐acetyl‐phenyl)‐ADPR, analogues that act as ADPR antagonists on human TRPM2, act as agonists of drTRPM2, implying a fundamentally different role of the two nucleotide binding sites in the human and zebrafish channel.

Crystal structure of TRPM2 MHR1/2 domain

The TRPM2 MHR1/2 domain plays a crucial role in the regulation of the full‐length channel. , , We determined the crystal structure of the zebrafish TRPM2 MHR1/2 domain (drMHR1/2). Two molecules were present in the asymmetric unit, with crystal contacts generated through the exposed loops. The structure was determined using Se‐SAD phasing to 2.0 Å resolution (Table S1). We could resolve almost an entire MHR1/2 domain with unambiguous side chain information (residues 33–418, with nine residues missing in the loop comprising residues 201–209). The overall structure (Figure 2a) shows a bi‐lobed clamshell‐like shape and superimposes well with the previously published cryo‐EM structure of the full‐length zebrafish TRPM2 channel (PDB: 6DRK) (Figure 2b,c).

FIGURE 2

Crystal structure of drMHR1/2. (a) The overall fold of zebrafish MHR1/2 (drMHR1/2) shows a bi‐lobed clamshell‐like structure. The structure is colour‐coded from blue (N‐terminus) to red (C‐terminus). Zn2+ ion is shown as magenta sphere. (b) Overlay of the drMHR1/2 crystal structure (wheat) with the MHR1/2 domain from a zebrafish TRPM2 (drTRPM2) cryo‐EM structure (green) (PDB: 6DRK). (c) Superposition of the drMHR1/2 crystal structure (wheat) on the tetrameric drTRPM2 cryo‐EM structure (white with one monomer in green), indication of the location of the novel Zn2+‐binding domain (PDB: 6DRK).

While we failed to crystallize drMHR1/2 in complex with ADPR or 2′‐deoxy‐ADPR, comparison with the horseshoe‐like binding mode of ADPR as observed in the cryo‐EM structure of ADPR‐bound drTRPM2 (PDB: 6DRJ) indicates that the 2′ hydroxyl group, which distinguishes ADPR and 2′‐deoxy‐ADPR, is solvent‐exposed in the complex and thus does not contribute to the binding. This is in agreement with the similar binding parameters obtained for these two ligands by ITC and nDSF.

Identification and characterization of a conserved Zn2+‐binding domain

The high resolution of the drMHR1/2 structure allowed unambiguous model building and identified a domain (residues 53–95), located between β1 and β2, which was not resolved in the lower resolution cryo‐EM structure of drTRPM2. Surprisingly, this novel domain revealed clear electron density for an ion that is coordinated by three cysteines and one histidine (Figure 3a). The interacting residues C53, C65, C67, and H74 coordinate the ion tetrahedrally (Figure 3b) with geometry and bond length typical for Zn2+ ion coordination (according to CheckMyMetal server, ). An x‐ray fluorescence energy scan of the crystal near the zinc absorption K edge (9.6586 keV) unambiguously confirmed the presence of a Zn2+ ion (Figure 3c).

FIGURE 3

Novel Zn2+‐binding domain in drMHR1/2 that is conserved in TRPM channels. (a) Electron density of the novel Zn2+‐binding site (grey mesh) and PHENIX POLDER electron density after omission of the Zn2+ ion (green mesh) with the atomic model (wheat) and the Zn2+ ion (grey). (b) Detailed structure view of the Zn2+‐binding domain with a tetrahedral coordination of the Zn2+ ion by C53, C65, C67, H77. (c) X‐ray energy scan of the fluorescence emitted by the sample near the zinc absorption K edge (9.6586 keV) confirming ion identity. (d) Multiple sequence alignment of novel Zn2+‐binding domain in TRPM channels. Conservation of the four Zn2+‐coordinating residues (marked) in the N‐terminus of most TRPM channels. Alignment of all human (Homo sapiens) TRPM channels and TRPM2 from sea starlet anemone (Nematostella vectensis), chicken (Gallus gallus), mouse (Mus musculus), rat (Rattus norvegicus), chimpanzee (Pan pansicus), zebrafish (Danio rerio).

A multiple sequence alignment revealed that the four residues forming the Zn2+‐binding domain (C53, C65, C67, H74) are conserved between TRPM2 orthologues from different species and most human TRPM members (Figure 3d). This strict evolutionary conservation from invertebrates to mammals strongly indicates that the novel Zn2+‐binding domain is physiologically relevant. As the Zn2+‐domain with its adjacent ß‐stem makes extensive interactions with the remaining MHR1/2 domain (cyan/grey interface in Figure 5), it is likely that the presence of the Zn2+‐binding motif causes a stabilization of the MHR1/2 domain (Figure 5). Indeed, we could show its importance for protein integrity/stability by mutating two of the cysteine residues to alanine (C65A and C67A). The mutant drMHR1/2 sequence was recombinantly expressed in E. coli, but the protein seemed to be insoluble and thus probably incorrectly folded (Figure S3).

FIGURE 5

Importance of Zn‐binding domain for MHR1/2 stability and ADPR binding. (a) Zn2+‐binding domain (cyan) makes substantial interactions with the remaining part of the MHR1/2 domain (grey). (b) The presence of the Zn2+‐binding domain in drMHR1/2 leads to correct positioning of loop 263–273 (magenta) containing the conserved tyrosine residue Y271 which stacks with the adenine moiety of ADPR (see PDB:6DRJ, ). According to this model, the stabilization and loop positioning caused by the Zn2+‐domain primes the MHR1/2 domain for ligand binding.

The Zn2+‐binding domain is required for TRPM2 function

We next set out to investigate the role of the Zn2+‐binding domain in the context of full‐length TRPM2. Full‐length human TRPM2 (hsTRPM2) containing equivalent mutations of the Zn2+‐coordinating residues (in human: C89A and C91A) could be successfully expressed in transfected HEK293 cells and localized to the cell surface, albeit with reduced expression levels compared to the wild‐type protein (Figure 4a). Whole‐cell patch clamp measurements revealed that the Zn2+‐binding site is required for channel activity as the hsTRPM2 mutants on the cell surface with mutations of Zn2+‐coordinating residues did not evoke a current upon infusion of ADPR like the wild‐type protein (Figure 4b, Figure S4).

FIGURE 4

The novel Zn2+‐binding site in the MHR1/2 domain is important for protein integrity and TRPM2 channel activity. (a) Cell surface biotinylation assay with Zn2+‐binding site mutants of human TRPM2 (hsTRPM2) proving the importance of the motif. The expression level of the mutants is lower on the cell surface and in the total membrane fraction. (b) Whole‐cell patch clamp measurements showing that the Zn2+‐binding site is also important for channel activity. The remaining fraction of mutant protein on the cell surface of HEK293 cells transfected with hsTRPM2 variants does not invoke a current upon infusion of 200 nM free Ca2+ with 100 μM ADPR via the patch pipette. The extracellular solution contained 1 mM Ca2+.

Taken together, we could show that the novel motif is crucial for correct protein folding, which subsequently affects TRPM2 channel activity. Since the channel with mutations within the motif could not be activated by ADPR, an endogenous ligand of TRPM2, we postulate that the presence of an intact Zn2+‐domain leads to correct positioning of loop 263–273 (287–297 in hsTRPM2). This loop contains the conserved tyrosine residue Y271 (Y295 in hsTRPM2), which stacks with the adenine moiety of ADPR (see PDB: 6DRJ, ). In this model, the stabilization and loop positioning caused by the Zn2+‐domain primes the MHR1/2 domain for ligand binding (Figure 5).

Interestingly, a structurally similar Cys3‐His1 Zn2+‐binding motif has recently been discovered in the N‐terminal region of human TRPC5. While being nonrelated in sequence to our motif identified in TRPM channels, this motif is also conserved across all TRPC channels, indicating that both TRPM and TRPC channels contain a conserved intracellular Zn2+‐binding site.

MATERIALS AND METHODS

Key resources table

Materials

All chemicals were of analytical grade and obtained from Roth (Karlsruhe, Germany) or Sigma Aldrich/Merck (Darmstadt, Germany).

Protein expression and purification

The sequence coding for the zebrafish TRPM2 MHR1/2 domain (drMHR1/2, residues 1–419 of drTRPM2) was cloned into a pnEK‐vH vector bearing a TEV‐cleavable N‐terminal His6‐tag. Protein production was carried out in E. coli BL21 Gold (DE3) cells in Terrific Broth (TB) media supplemented with 25 μg/ml kanamycin. Cells were grown at 37°C until a cell density of 1 (measured at 600 nm) was reached. After induction with 0.1 mM isopropyl β‐d‐1‐thiogalactopyranoside (IPTG), the cells were grown for further 16 hr at 20°C. The production of selenomethionine‐labeled protein was carried out in the methionine auxotroph E. coli strain B834 in M9 minimal medium supplemented with 70 mg/L l‐selenomethionine (Serva) and 25 μg/ml kanamycin. Cells were grown at 37°C to an OD600 of 0.6 and after induction with 0.1 mM IPTG, the target protein was expressed for further 16 hr at 20°C.

The cells were harvested by centrifugation (5,000g for 25 min) and lysed in 25 mM Tris pH 7.5, 300 mM NaCl, 5% (v/v) glycerol, 5 mM 2‐mercaptoethanol using a high‐pressure homogenizer (EmulsiFlex‐C3, Avestin). Cell debris was removed by centrifugation (39,000g for 45 min) and the His6‐tagged protein was purified from the supernatant using immobilized metal affinity chromatography (IMAC) with Ni‐NTA resin (Roth). After His‐tag removal by TEV protease (1/10 w:w) and reverse Ni‐NTA purification, the target protein was further purified by gel filtration on a Superdex S200 increase 10/300 column using buffer M (for wild‐type protein, 25 mM HEPES pH 7.5, 150 mM NaCl, 5 mM CaCl2, 5 mM MgCl2) or buffer MS (for Se‐Met‐labeled protein, 25 mM HEPES pH 7.5, 150 mM NaCl, 5 mM CaCl2, 5 mM MgCl2, 5 mM 2‐mercaptoethanol). Peak fractions were pooled and protein identity confirmed by SDS‐PAGE and mass spectrometry.

Isothermal titration calorimetry

ITC measurements were carried out at 25°C using a MicroCal ITC‐200 isothermal titration calorimeter (Malvern Panalytcal) and thermodynamic parameters were analyzed using the MicroCal ORIGIN™ software. The ligands ADPR and 2′‐deoxy ADPR were each dissolved in buffer M (see purification) to a concentration of 500 μM and placed in the syringe. After an initial injection of 0.5 μl, 18 regular injections of 2 μl were added to 10 μM drMHR1/2 in the sample cell. The individual injections were interspaced by 150 s and stirring speed was set to 750 rpm. Heat of dilution was obtained by titrating ADPR into buffer M and baseline corrections were carried out accordingly. All ITC experiments were performed as triplicates and errors are reported as standard deviations of the mean KD value.

Differential scanning fluorimetry

nDSF measurements were carried out on a Prometheus NT.48 system (Nanotemper). The 10 μM drMHR1/2 protein was mixed with varying amounts (600 nM to 2 mM) of ADPR or 2′‐deoxy ADPR in buffer M (see purification). Thermal unfolding was measured by following intrinsic tryptophan fluorescence during a thermal ramp (1°C/min). The PR.ThermControl software (Nanotemper) was used to determine melting temperatures. Binding parameter analysis was performed by simple Hill fit in the ORIGIN™ software.

Crystallization

Crystals of selenomethionine‐labeled drMHR1/2 protein were grown by sitting drop vapor diffusion technique. The 1 μl of purified protein (4 mg/ml) was mixed with 1 μl of the precipitant mix (Tris, BICINE, diethylene glycol, triethylene glycol, tetraethylene glycol, pentaethylene glycol, glycerol; PEG4000, Jeffamine M‐600). Crystals of triangular shape appeared after 1–3 days, reaching sizes of approximately 50–120 μm.

Structure determination

X‐ray diffraction data were collected at 100 K at the PETRA III/EMBL P14 beamline. All datasets were processed with XDS and merged with AIMLESS. Heavy atom site identification and phasing was performed with SHELXCD. A combination of ARP/wARP and COOT was used for automatic and manual model building, respectively. Refinement was carried out in PHENIX. The final model corresponds to residues 38–423. All data collection and refinement statistics are summarized in Table S1 (supplementary information).

Multiple sequence alignment

Multiple sequence alignment of TRPM2 and homologues was performed using the Clustal Omega server. Results were visualized with Jalview (https://www.jalview.org/).

Cell culture and transfection vectors

HEK293 wild type cells were cultivated in Dulbecco's modified eagle medium (DMEM) (with GlutaMAX‐I, 4.5 g/L d‐Glucose and Phenol Red; Gibco) supplemented with 10% FBS (Sigma), 100 U/ml Penicillin and 100 μg/ml Streptomycin (Gibco) and kept at 37°C and 5% CO2.

Cells were transiently transfected with pIRES2‐EGFP expression vectors containing either full‐length human TRPM2 variants (wild type, C89A, C91A, C89A C91A double mutant) or full‐length zebrafish TRPM2 (wild type). Transfection was verified by fluorescence microscopy.

Cell surface biotinylation assay

Transfection of hsTRPM2 variants or empty vector was performed with Lipofectamine LTX (Thermo Fisher) according to the manufacturer's instructions. Cells were grown for 48 hr after transfection, washed with D‐PBS (Gibco) and incubated with 1 mg/ml EZ‐Link Sulfo‐NHS‐LC‐Biotin (Thermo Fisher) in order to biotinylate cell surface proteins. After detachment with 2 mM EDTA in D‐PBS, the cells were collected, centrifuged (500 g for 5 min) and washed with D‐PBS. Cell lysis and membrane protein extraction was performed using the ProteoExtract Native Membrane kit (Merck Millipore) according to the manufacturer's protocol. Solubilized membrane protein samples were quantified by Bradford assay.

NeutraVidin Agarose beads (70 μl, Thermo Fisher) were used to isolate the biotinylated cell surface proteins. The beads were incubated with total membrane protein samples (600 μg) for 18 hr while rotating at 4°C. After washing with extraction buffer II from the kit mentioned above, the beads and the total membrane protein samples were mixed with SDS sample buffer (with 5% 2‐mercapto ethanol) and heated to 75°C for 5 min.

The samples were analyzed by western blot. After SDS‐PAGE (4%–20% Protean precast gel, BioRad) and transfer to a polyvinylidene difluoride membrane (Merck Millipore), the membrane was cut at 140 kDa in order to simultaneously detect TRPM2 (171 kDa) and the reference Na+/K+‐ATPase (100 kDa). The membrane parts were probed for 18 hr with anti‐hsTRPM2 antibody from rabbit (Novus #nb500‐241 at 1:50,000 dilution) and anti‐Na+/K+‐ATPase antibody from rabbit (Cell Signaling #3010 at 1:1000 dilution), respectively. Both primary antibodies were detected with an HRP‐conjugated anti‐rabbit secondary antibody (Dianova #111‐035‐045 at 1:10,000 dilution) for 1 h. Chemiluminescent detection was performed using the SuperSignal West Pico substrate (Thermo Fisher).

Electrophysiology (patch clamp)

hsTRPM2

HEK293 cells were transfected 24 hr prior to experiments using jetPEI reagent (Polyplus transfection). The transfection complex (5 μg DNA and 10 μl jetPEI reagent in 150 mM NaCl solution at a total volume of 250 μl) was incubated for 30 min before 2.5 × 105 cells were added. The cell suspension was subsequently seeded to 35 mm dishes at low density. Patch clamp experiments were performed at room temperature. Before the start of the experiments the medium was replaced by extracellular solution (in mM: 140 N‐methyl‐d‐glucamine [NMDG], 5 KCl, 3.3 MgCl2, 1 CaCl2, 5 d‐Glucose, 10 HEPES, adjusted to pH 7.4 with HCl). Patch pipettes were pulled from 1.05 × 1.50 × 80 mm glass capillaries and filled with intracellular solution (in mM: 0.1 ADPR, 120 KCl, 8 NaCl, 1 MgCl2, 10 HEPES, adjusted to pH 7.2 with KOH; The intracellular Ca2+ concentration was set to 200 nM with 10 mM EGTA and 5.6 mM CaCl2, as calculated with Maxchelator.). After pipette resistances were determined (1.5–2.9 MΩ), currents were compensated for fast and slow capacity transients and recorded in whole‐cell configuration. The series resistance compensation was set to 70%. Using voltage clamp, repetitive voltage ramps of 140 ms spanning the range from −85 mV to +20 mV were applied from a holding potential of −50 mV every 5 s over a measuring period of 445 s. For data analysis, the maximum outward current at +15 mV of each measuring period was extracted from all experiments. Cells with a series resistance >10 MΩ during the experiment were excluded from data analysis.

drTRPM2

Cells were transfected with jetPEI reagent as described above (transfection complex: 1 μg DNA and 5 μl jetPEI reagent in 150 mM NaCl solution at a total volume of 250 μl). Due to large whole‐cell currents and subsequent rupturing of the cells, the patch clamp protocol was adjusted for drTRPM2. Compared to human TRPM2, zebrafish TRPM2 has a higher permeability for Ca2+. To reduce Ca2+ inward current and thus positive feedback by Ca2+, the Ca2+ concentration in the extracellular solution was reduced to 100 μM. Additionally, the Ca2+ concentration in the intracellular solution was reduced to 100 nM. The holding potential was set to −40 mV. Since under these Ca2+‐reduced conditions, whole‐cell currents were still too high, only cells with relatively low drTRPM2 expression levels were measured. The drTRPM2 expression level was quantified by measuring the EGFP fluorescence intensity. Using a fluorescence filter, a picture of each cell was taken prior to the experiment (Leica DMi8 microscope). The exposure time was set to 1,000 ms. EGFP Fluorescence intensities are given as gray‐level values and are the same in each measuring group (image analysis with ImageJ) (Figure S2C).

Statistical analysis

Statistical analysis was performed with GraphPad Prism (v9, GraphPad Software). After normal distribution was confirmed by Kolmogorov–Smirnov test, one‐way ANOVA test was applied followed by Dunnett (for hsTRPM2) and Tukey (for drTRPM2) test for multiple comparisons.

AUTHOR CONTRIBUTIONS

Simon Sander: Conceptualization (equal); data curation (lead); investigation (lead); methodology (equal); visualization (equal); writing – original draft (equal); writing – review and editing (equal). Jelena Pick: Data curation (supporting); investigation (equal); methodology (supporting); visualization (supporting); writing – original draft (supporting); writing – review and editing (supporting). Ellen Gattkowski: Investigation (supporting). Ralf Fliegert: Conceptualization (equal); funding acquisition (equal); investigation (equal); project administration (equal); supervision (equal); writing – original draft (equal); writing – review and editing (equal). Henning Tidow: Conceptualization (equal); funding acquisition (equal); investigation (equal); project administration (lead); resources (equal); supervision (lead); writing – original draft (equal); writing – review and editing (equal).

CONFLICT OF INTEREST

The authors declare no competing financial interest.

Appendix S1. Supplementary Information

Click here for additional data file.

Reagent type (species) or resource	Designation	Source or reference	Identifiers
Gene	TRPM2	GenScript	ODa47363D
Strain, strain background (E. coli)	BL21 Gold (DE3)	Agilent	230132
Strain, strain background (E. coli)	B834 (DE3)	Merck	69041
Cell line, human	HEK293	ATCC	CRL‐1573
Chemical compound	L selenomethionine	Serva	77765.01
Chemical compound	ADPR	Sigma‐Aldrich	A0752
Chemical compound	2′‐deoxy‐ADPR	BioLog	D 227‐01
Software, algorithm	XDS	42	http://xds.mpimf‐heidelberg.mpg.de/
Software, algorithm	AIMLESS	43	http://www.ccp4.ac.uk/html/aimless.html
Software, algorithm	SHELXCD	44	https://shelx.uni‐goettingen.de/
Software, algorithm	ARP/wARP	45	https://www.embl‐hamburg.de/ARP/
Software, algorithm	Coot 0.8.9.1	46	https://www2.mrc‐lmb.cam.ac.uk/personal/pemsley/coot/
Software, algorithm	PHENIX 1.16	47	https://www.phenix‐online.org/
Software, algorithm	Clustal Omega	48	https://www.ebi.ac.uk/Tools/msa/clustalo/