Fig 1.
Summary of NUDT9H mutations of hTRPM2 and nvTRPM2 and their effects on channel function.
Characteristic TRPM2 regions like the N-terminal TRPM homology region, the six transmembrane segments, as well as the C-terminal coiled-coiled region and the NUDT9H domain are indicated. The NUDT9H domain is divided into the putative N-terminal ADPR binding region and the C-terminal catalytic domain. (A) Human TRPM2 (green) with the mutation N1326D in the ADPR binding region which abolishes both the sensitivity to ADPR and to H2O2 (symbolized with ↓↓). In contrast, the corresponding mutation N1365D in nvTRPM2 (blue) does not change the sensitivity to ADPR (+/-) but produces strong sensitivity to H2O2 (↑↑). (B) Human TRPM2 in which the double mutation IL to EF within the catalytic site abolishes both the sensitivity to ADPR and to H2O2. The deletion of 15 amino acid residues immediately downstream of the catalytic site which is characteristic for hTRPM2 is also indicated. In contrast, the reciprocal mutations in the catalytic site of nvTRPM2 (EF to IL or the deletion Δ15) again do not change the sensitivity to ADPR but produce sensitivity to H2O2.
Fig 2.
Functional characterisation of nvTRPM2 variants with mutations in the NUDT9H region.
(A-D) Representative whole cell patch-clamp experiments. The variants were nvTRPM2-N1326D (A and B) and nvTRPM2-RILRQE (C and D). Stimulation of currents was performed either with ADPR (50–100 μM) in the pipette (A and C) or with H2O2 (10 mM) applied to the bath at the time point indicated by an arrow (B and D). As negative control for the stimulation with H2O2, a recording of wild-type nvTRPM2 is included in panel B (inset). The intracellular Ca2+ concentration was adjusted to 1 μM. Current flow was inhibited by superfusion of the cells with a solution containing NMDG as main cation (horizontal bars). For each variant similar results were obtained from a least 3 independent experiments. (E) Summary of calcium imaging experiments. Maximal increases in (Ca2+)i, as indicated by an increased F340/F380 ratio, were evoked by extracellular H2O2 (10 mM). The variants (see above) were compared with mock-transfected cells as well as with cells transfected with wild-type nvTRPM2. *** indicates a significant difference (P < 0.001) evaluated with a one-way ANOVA and the Bonferroni correction. (n = 9–13). Error bars are s.e.
Fig 3.
Western blot of surface expressed variants of hTRPM2 and nvTRPM2 and endogenous NUDT9 expression.
(A) Cell surface expression, assessed with biotinylation assays, of full-length and truncated channel variants (as indicated), each containing a C-terminally attached 3xHA tag. Western blots on the NeutrAvidin-bound fractions (top) and on total HEK-293 cell lysates (bottom) were probed with anti-HA antibody. Reduced β-actin staining in the Avidin-bound faction rules out biotinylation of cytosolic proteins in damaged cells. (B) Western blots on enriched plasma membrane (left) and intracellular membrane fractions (right) of HEK-293 cells expressing the indicated TRPM2 variants. Membrane fractions were prepared with the differential centrifugation method [18] and probed with anti-HA antibody. Mock-transfected cells were used as negative control. Two independent experiments gave similar results. (C) Western blot on total HEK-293 cell lysates of mock-transfected cells and cells transfected with cDNA of the human NUDT9 enzyme (as indicated) probed with monoclonal mouse anti-hNUDT9 antibody. For each of the Western blots at least two independent experiments were performed to give similar results.
Fig 4.
Functional characterisation of TRPM2 variants with the NUDT9H domain removed.
(A) Control patch-clamp experiment performed on hTRPM2 stimulated with ADPR (0.3 mM) and Ca2+ (< 10 nM) in the pipette solution. Note the typical slow onset of the current and block by NMDG. (B) Absent response with ADPR (0.6 mM) and Ca2+ (1 μM) in the pipette solution in hTRPM2 lacking the NUDT9H domain. For both hTRPM2- Δ NUD variants (with or without 3xHA-tag) similar results were obtained from a least 10 independent experiments. (C) Characteristic current development induced by ADPR (0.6 mM) in nvTRPM2-3xHA. (D) Current response to ADPR (0.1 mM) in the nvTRPM2 variant in which the NUDT9 domain is lacking (in both nvTRPM2- Δ NUD variants the same results were obtained with or without 3xHA-tag). The intracellular Ca2+ concentration was adjusted to 1 μM. (E) Relation of current densities to ADPR concentration in cells transfected either with nvTRPM2-3xHA or with nvTRPM2- Δ NUD-3xHA. Already the smallest concentration of ADPR evoked significant currents (*** P < 0.001; Student's t-test, n = 8) in comparison to the absence of a stimulus. Error bars are s.e. (F) Direct activation of inside-out patches by ADPR. The three traces were from TRPM2 channel variants as indicated and obtained rapidly after establishing the inside-out configuration, with ADPR (0.4 mM) already present in the bath. Multiple channels were present in all three patches but the initial activation showed preferentially openings of one or of few channels. The transmembrane potential was +60 mV, chosen because inactivation of nvTRPM2 was slower when currents were in the outside direction. Similar results were obtained from at least 4 independent experiments.
Fig 5.
Effects of ADPR-2’-phosphate (ADPRP) on channel variants of hTRPM2 and nvTRPM2.
(A-C) Representative patch-clamp experiments either on wild-type hTRPM2, wild-type nvTRPM2 or nvTRPM2- Δ NUD-3xHA as indicated. Stimulation was performed with 0.5 mM (A) or 50 μM (B and C) ADPRP in the pipette solution. The intracellular Ca2+ concentration was adjusted to 1 μM. The respective current characteristics are indistinguishable from those evoked with ADPR as stimulus. (D) Comparison of the effects of different ADPRP concentrations on each channel variant. Note that 50 μM ADPRP failed to stimulate currents in hTRPM2, whereas in both nvTRPM2 variants significant responses were detected already at 5 μM. Asterisks indicate significant differences (* P < 0.05; ** P < 0.01, *** P < 0.001; Student's t-test, n = 5–9) in comparison to the absence of a stimulus. Error bars are s.e.
Fig 6.
Functional effects of different C-terminally attached HA-tags on hTRPM2 and nvTRPM2 channel variants.
(A and B) Characteristic currents of nvTRPM2- Δ NUD-3xHA (A) and nvTRPM2-3xHA (B) developed after the application of H2O2 (10 mM) to the bath (at the time point indicated by an arrow). The intracellular Ca2+ concentration was adjusted to 1 μM. Note that wild-type nvTRPM2 without an HA-tag is not stimulated by H2O2 (see Fig 2B inset). (C and D) Summary of the effects of extracellular H2O2 (10 mM) on several nvTRPM2 (C) and hTRPM2 (D) variants as obtained in calcium-imaging experiments. Note that the 3xHA-tag abolishes H2O2 sensitivity in hTRPM2 but creates it in nvTRPM2. (E) Attenuated current responses in hTRPM2-3xHA stimulated with ADPR (0.6 mM) and Ca2+ (1 μM) in the pipette solution. (F) After shortening of the 3xHA-tag, robust currents were evoked in hTRPM2-1xHA by ADPR (0.2 mM, n = 4). *** indicates a significant difference (P < 0.001) evaluated with a one-way ANOVA and the Bonferroni correction (n = 8–32). Error bars are s.e.
Fig 7.
The sensitivity to H2O2 correlates with the catalytic activity of the NUDT9 domain.
(A), Representative whole cell patch-clamp experiment showing the stimulation of the variant nvTRPM2-NUDenz-AIF with H2O2 (10 mM) applied to the bath at the time point indicated by an arrow. The intracellular Ca2+ concentration was adjusted to 1 μM. Similar results were obtained from 6 independent experiments. (B) Summary of calcium imaging experiments. Maximal increases in (Ca2+)i, as indicated by an increased F340/F380 ratio, were evoked by extracellular application of H2O2 (10 mM). The variants nvTRPM2-NUDenz and nvTRPM2-NUDenz-AIF were compared with mock-transfected cells. (C) Sketch of the NUDT9 enzyme variants (wild-type hNUDT9 enzyme, hNUDT9H domain, nvNUDT9H domain) used for co-expression experiments with nvTRPM2-ΔNUD. (D) Calcium imaging experiments in response to H2O2 (10 mM), performed on cells co-expressing nvTRPM2-ΔNUD and one of the enzyme variants depicted in panel C. Comparison was performed with mock-transfected cells and cells transfected with nvTRPM2-ΔNUD alone. Asterisks indicate significant differences (* P < 0.05; *** P < 0.001) evaluated with a one-way ANOVA and the Bonferroni correction. (n = 8–26). Error bars are s.e.
Fig 8.
Sketch illustrating the different function of the NUDT9H domain in the two TRPM2 channel orthologs.
(A-C) Cartoon interpretation of the putative functional role of the endogenous NUDT9H domain in heterologeously expressed hTRPM2, nvTRPM2 and nvTRPM2-ΔNUD as indicated during cell exposure to H2O2. It is generally accepted that oxidative stress leads to intracellular release of ADPR. In case of hTRPM2, the accumulated ADPR binds to the NUDT9H domain and initiates channel activation without the requirement of ADPRase activity (A). In contrast, there is a second ADPR interaction site in nvTRPM2 responsible for gating, whereas the NUDT9H domain has a strictly enzymatic role, preventing channel activation by low cytosolic concentrations of ADPR as in the presence of H2O2. Hypothetically, H2O2 may additionally enhance the ADPRase activity of the NUDT9H domain (B). When the catalytic activity of the NUDT9H domain of nvTRPM2 is lost due to point mutations or deletion of the entire domain, increasing concentrations of intracellular ADPR activate the channel (C).