The function of the two-pore channel TPC1 depends on dimerization of its carboxy-terminal helix
Two-pore channels (TPCs) constitute a family of intracellular cation channels with diverse permeation properties and functions in animals and plants. In the model plant Arabidopsis, the vacuolar cation channel TPC1 is involved in propagation of calcium waves and in cation homeostasis. Here, we discovered that the dimerization of a predicted helix within the carboxyl-terminus (CTH) is essential for the activity of TPC1. Bimolecular fluorescence complementation and co-immunoprecipitation demonstrated the interaction of the two C-termini and pointed towards the involvement of the CTH in this process. Synthetic CTH peptides dimerized with a dissociation constant of 3.9 µM. Disruption of this domain in TPC1 either by deletion or point mutations impeded the dimerization and cation transport. The homo-dimerization of the CTH was analyzed in silico using coarse-grained molecular dynamics (MD) simulations for the study of aggregation, followed by atomistic MD simulations. The simulations revealed that the helical region of the wild type, but not a mutated CTH forms a highly stable, antiparallel dimer with characteristics of a coiled-coil. We propose that the voltage- and Ca2+-sensitive conformation of TPC1 depends on C-terminal dimerization, adding an additional layer to the complex regulation of two-pore cation channels.
KeywordsVacuole Calcium signaling MD simulation Patch-clamp Microscale thermophoresis Mutation
Two-pore cation channels (TPCs) are intracellular ion channels residing in the vacuolar membrane of plant cells, and in lysosomes and endosomes of mammalian cells [1, 2]. Depending on their biophysical properties and host species, they play diverse roles, ranging from cation and pH homeostasis, sensing of the metabolic state, cell-to-cell signaling, control of the membrane potential and membrane trafficking to pigmentation, and even to the control of Ebola virus invasion in animal cells [1, 2, 3, 4, 5, 6, 7, 8]. The underlying molecular mechanisms are in most cases fairly unknown.
TPCs belong to the superfamily of voltage-gated ion channels that are built from Shaker-like domains of 6 transmembrane segments (S1-S6) and a pore-forming helix between S5 and S6. One TPC monomer contains two Shaker-like domains connected by a cytosolic linker, and the channel functions as a dimer [1, 9, 10]. The cytosolic N-terminus contains a dileucine motif, which is responsible for targeting of the channel to the tonoplast in plant cells or lysosomes in animal cells [11, 12]. Two EF-hand motifs in the linker domain of plant, but not animal TPCs mediate binding and channel activation by cytosolic calcium ions [13, 14]. In contrast, animal TPC1 and TPC2 are directly activated by phosphatidyl-inositol-3,5-bisphosphate PI(3,5)P2 [6, 7, 15] and have been identified as mediators of nicotinic acid adenine dinucleotide phosphate (NAADP)-induced Ca2+-release from endolysosomes [9, 12, 16, 17, 18].
While animal cells express two or more different channels, in most plant species there is only one TPC isoform, TPC1 [1, 2]. TPC1 encodes the slow vacuolar (SV) channel  and mediates the passage of K+ and Na+, and other monovalent as well as divalent cations, although a direct involvement in calcium release from the vacuole in vivo is strongly debated [20, 21, 22, 23, 24, 25].
Recently, a role for TPC1 in a rapid, long-distance signaling system based on Ca2+ waves has been uncovered . In response to a localized salt stimulus at the Arabidopsis root, the calcium wave mainly travels through the cortex and endodermis at speeds of up to 420 μm/s, and is required for activation of stress responsive genes in the shoot. Lack of TPC1 largely reduces the travel speed of this trigger wave, while TPC1 overexpression accelerates it. TPC1 thus plays an important role in long-distance signaling in response to salt stress . How the density or distribution of TPC1 in the vacuolar membrane affects the speed of the long-distance calcium wave remains largely unknown.
Deregulation of the voltage-dependent activity of TPC1 leads to imbalanced cation homeostasis in the vacuole of fou2, which has a threefold higher Ca/K ratio as compared to the wild type . The expression profile of fou2 resembles that of wild type plants under K starvation, and fou2 plants produce more oxylipins in response to wounding [27, 28]. The fou2/TPC1D454 N mutation introduces an amino acid exchange in the binding site for luminal Ca2+, which abolishes the inhibition of TPC1 by luminal calcium ions and shifts the activity range towards more negative potentials [26, 29]. These results demonstrate the important role of TPC1 for cation homeostasis and vacuolar storage function.
A tight regulation of SV channels prevents loss of potassium and other cations from the vacuole, and many factors down-regulating or blocking TPC1 have been identified, including luminal calcium ions , protons , sodium ions , and polyamines . A further negative regulation is mediated by polyunsaturated fatty acids, which like luminal Ca2+ ions shift the voltage dependence towards more negative potentials . In comparison to these ionic and metabolic factors, less is known about interactions of TPC1 with regulatory proteins and their sites of interaction. 14-3-3 proteins rapidly reduce SV currents [35, 36], and regulation by kinases and phosphatases is most likely [37, 38].
Except for a few cases like calcium binding [13, 29] or block by polyamines , little is known about the structure function relation in TPC1. Here, we report that the predicted carboxy-terminal helical regions of two TPC1 monomers are prone to dimerization. These dimers are shown to be essential for the function of the (dimerized) TPC1 channel. We employed wild type coarse-grained and atomistic simulations, as well as coarse-grained mutant simulations, which revealed that the wild type C-terminal domain forms a stable antiparallel coiled-coil. In contrast, the mutant showed a highly promiscuous dimerization pattern, pointing to a significantly decreased affinity. We suggest that the dimerization of the wild type TPC1 carboxyl-termini stabilizes the channel in a conformation, which is sensitive to Ca2+-binding and depolarization, adding an additional layer to the complex regulation of two-pore cation channels.
Materials and methods
Plant material and growth conditions
Arabidopsis thaliana Col-0 wild type and tpc1-2 mutant plants  were used. After seed stratification at 4 °C for 3 days, plants were grown on soil in a growth chamber under 8 h light/16 h dark conditions at 22 °C. Nicotiana benthamiana wild type plants were cultivated on soil in the greenhouse at 22 °C under a 16 h light/8 h dark cycle and used after 6 weeks for infiltration.
Cloning procedures for C-terminal GFP-fusions are described elsewhere , and eGFP was used in all cases. PCR-primers were designed accordingly with the respective cloning sites and with or without stop-codon and purchased from Sigma. For construct details and primer sequences see Supplemental Table 1 and 2.
For BiFC analyses the gateway entry vector pENTR-D-TOPO and the destination vectors pDEST–GWVYCE and pDEST–GWVYNE were used for C-terminal fusion of the Venus C- and N-terminus, respectively, and pDEST–VYCE(R)GW and pDEST–VYNE(R)GW for N-terminal tagging .
For the Co-IP tests the C-terminally tagged GFP-fusions were used as templates to amplify the soluble C-terminus (amino acids 673–733) of TPC1 or TPC1-3LP fused to GFP with PCR. The constructs were subcloned into the pENTR-D-TOPO vector and afterwards brought into the expression vectors with EcoRI and XhoI (New England Biolabs), so that they were N-terminally tagged either with 6x Myc (pCS2+MT) or with the FLAG-tag (pCS2 + Flag) .
Protoplast isolation and transformation
Expression of GFP-fusion proteins for electrophysiological measurements or confocal microscopy was performed in A. thaliana tpc1-2 mesophyll protoplasts as described [11, 13]. Briefly, leaves were enzymatically digested to release mesophyll protoplasts, which were then transformed by the polyethylene glycol method. Protoplasts were used for electrophysiological measurements and subcellular localization of GFP fluorescence 2–4 days after transformation.
Fluorescence signals were detected and documented with a TCS SP2 confocal laser scanning microscope and the Leica Confocal Software (Leica Microsystems). A 488 nm Argon laser was used for excitation of GFP and of the auto-fluorescence of chlorophyll. GFP-signals were detected from 500 to 556 nm and chlorophyll signals from 675 to 767 nm. Venus was excited with a 543 nm helium–neon laser and detected in the range from 520 to 556 nm, the corresponding chlorophyll signals were detected from 637 to 736 nm. Images were processed with Photoshop (Adobe Systems).
BiFC, tobacco infiltration, and fluorescence quantification
BiFC studies were performed as described  using a Split Venus system . The Agrobacterium strain C58C1  harboring the plasmid of interest and the helper strain p19  were grown in LB medium supplemented with 100 µg/ml Kanamycin at 29 °C overnight. The cultures were brought to an optical density (OD600) of 1.0 in infiltration buffer (10 mM MES, 10 mM MgCl2, 100 µM acetosyringone, pH 5.7/KOH) and the combinations to test for co-expression mixed at equal amounts. The p19 helper strain (OD600 = 1.0) was added to these mixtures in a 1:1 ratio. Suspensions were incubated for 2 h before infiltration into the abaxial side of 6-week-old N. benthamiana leaves. Each leaf was infiltrated with several plasmid combinations and a positive control (TPC1/pDEST-GWVYCE and TPC1/pDEST-GWVYNE) as well as a negative control (p19 alone) at separated areas to monitor comparability of the transient expression and to define background fluorescence.
For fluorescence quantification leaf disks (Ø 6 mm) were cut out of the infiltrated areas with a cork borer 2.5 day after infiltration. Leaf disks were placed upside-down in a black 96-well plate prefilled with 70 µl H2O per well to minimize dehydration effects. Fluorescence was measured with a plate-reader (Infinite F200, Tecan) with an excitation wavelength of 485 ± 20 nm and an emission wavelength of 525 ± 25 nm .
Cell culture, transfection, Co-IP and western blot analysis
HEK293T cells were propagated in DMEM GIBCO Glutamax (Life Technologies) supplemented with 10 % fetal calf serum. Cells were seeded in 6–well plates (20–30 % confluency) and transfected the next day using 10 µl Roti-Fect (Carl Roth) per 2 µg total DNA (1 µg DNA per plasmid). Two days after transfection cells were washed with 2 ml PBS buffer (Life Technologies) and harvested for immunoprecipitation in 150 µl cold lysis buffer (10 mM HEPES, 150 mM NaCl, 1 % Nonidet P40, 5 % glycerol, pH 7,4/KOH) supplemented with protease inhibitors and phosphatase inhibitors (Complete mini EDTA free and PhosStop, Roche). Cells were disrupted mechanically with a syringe (cannula diameter 0.55 mm). The lysate was cleared by centrifugation at 13.500 rpm and 4 °C for 10 min. Total protein concentrations were measured with a bicinchoninic acid protein assay (Applichem), typically the total protein amount was 500–600 µg. Loading controls of 50 µg of the lysate were taken accordingly and filled up to 20 µl with lysis buffer before adding 5 µl of 5x sample buffer (250 mM Tris, 5 % SDS, 25 % glycerol, 0.25 % bromophenol blue, 250 mM β–mercaptoethanol, pH 6.8/HCl) and boiling. For immunoprecipitation, 3 µg of mouse anti-c-Myc Epitope antibody 9E10 (sc-40, Santa Cruz) was used per sample. Samples were incubated with the antibody for 45 min at 4 °C with a rotator, then 10 µl of protein G Dynabeads (Life Technologies) equilibrated in the lysis buffer was added to each sample and incubation continued for another 45 min. Samples were washed three times with cold lysis buffer before eluting three times with 20 µl of 100 mM glycine (pH 2.5/HCl). After neutralizing the pH 15 µl of 5x sample buffer was added and samples were boiled.
Proteins of the loading controls and samples were separated with SDS-PAGE (12 % gel) and transferred to a PVDF membrane in duplicate. Membranes were blocked for 1 h with blocking reagent (Roche) and incubated over night at 4 °C with rabbit anti-Myc or anti-Flag antibodies. After washing three times with TBST buffer, the membranes were incubated for 1 h at RT with anti-rabbit antibody coupled to alkaline phosphatase. Membranes were washed three times with TBST, followed by two times with alkaline phosphatase buffer (100 mM NaCl, 100 mM Tris, pH 9.5/HCl). Colorimetric detection of the proteins was performed with NBT/BCIP substrate (Roche) in alkaline phosphatase buffer until a sufficient staining was achieved. All antibodies used for the western blots were purchased from Cell Signaling Technology.
Molecular dynamics simulations
Secondary structure prediction of the TPC1 carboxyl-terminus
Five atomistic [force field Amber 14SB, 52] starting configurations were generated from selected coarse-grained frames of the wild type simulations using the backward method . The K+-ion concentration of 100 mM was chosen similar to the cytoplasmic salt concentrations in plants . Additionally, the unit cells were altered to smaller ones, containing approximately 5700 TIP3P  water molecules as well as counterions (Cl−).
The modeled sequence is a small part of a larger protein sequence. To describe the helices accurately, the termini were kept neutral in all coarse-grained simulations or acetylated and amidated, respectively, in the atomistic simulations.
After automatically generating each 100 coarse-grained (CG) starting structures using DAFT, the wild type and mutant systems were subjected to a relaxation and equilibration process using martinate . The energy was minimized using steepest-descent (500 steps) and a 10 ps position-restrained simulation with a time step of 2 fs. Subsequently, the systems were relaxed in a 100 ps NpT simulation with an integration time step of 20 fs under constant pressure p and temperature T. In all simulations, the isotropic pressure of 1 bar was kept constant using a weak coupling scheme  with a 3 ps time constant. The temperature was maintained at 310 K with the v-rescale thermostat and a time constant of 1 ps . The relative dielectric constant was globally set to 2.5 and long-range electrostatic interactions were treated using particle-mesh Ewald  summation with a real-space cutoff of 1.2 nm. Dispersion interactions were described by a Lennard-Jones 12–6 potential that was shifted to zero between 0.9 nm and 1.2 nm. The production simulations were run for 250 ns each, using a time step of 20 fs.
Selected backmapped coarse-grained (CG) structures of the wild type were studied at atomistic resolution at 300 K and 1 bar using the Amber14SB force field. The simulation length was 200 ns (4 simulations) and 270 ns (1 simulation), respectively, with an integration time step of 2 fs. The temperature was kept constant using the Nosé-Hoover [60, 61] thermostat with a time constant of 0.5 ps. The pressure was modulated isotropically with the Parinello-Rahman  barostat and a time constant of 10 ps. The relative permittivity was set to 1. Long-range Coulomb interactions were calculated with the particle-mesh Ewald summation and a real-space cutoff of 1 nm. Van der Waals interactions were treated with a cutoff of 1 nm as typical for Amber force fields . Furthermore, long-range dispersion corrections were applied for the energy and pressure.
Relative orientations of the C-terminal helices of AtTPC1 and mutated AtTPC1-3LA were described using Euler angles (Fig. 7b), which characterize the relative orientation of two peptides [47, 64]. The matrix with Euler angles was obtained by least-square fitting a structure on a reference structure. To investigate the stability of formed dimers, the tilt angle between the dimers was determined over the last 50 ns of each simulation and its distribution plotted in a histogram. The tilt, ranging from 0° to 180°, was then divided into three subintervals (0°–50°, 50°–130°, and 130°–180°) in the case of wild type dimers and four subintervals (0°–35°, 35°–100°, 100°–150°, and 150°–180°) in the case of mutant dimers. These intervals were chosen according to the minima in the histogram (Fig. S6a). Subsequently, the number of transitions between these intervals was counted for each simulation and further evaluated.
MicroScale thermophoresis (MST) binding assay
For MST experiments, peptides corresponding to the consensus helical sequence (RSQRVDTLLHHMLGDEL) and the 3LA mutant (RSQRVDTAAHHMAGDEL) were synthesized (Peptide Speciality Laboratories GmbH, Heidelberg, Germany). To allow for label-free MST binding experiments, additional wild type and mutant peptides were synthesized, which were C-terminally extended by addition of two tryptophanes and a short linker (AAWW).
Labelfree MicroScale Thermophoresis binding experiments were performed in cooperation with the 2bind GmbH (Regensburg, Germany), using 750 nM tryptophane containing target peptide in PBS pH 7.5, 1 % DMSO, 0.1 % Pluronic with varied concentrations of the ligand peptide at 80 % MST power, 100 % LED power in hydrophilic zero background capillaries on a Monolith NT.labelfree device at 25 °C (NanoTemper Technologies, Munich, Germany). Normalized fluorescence data sets (WT peptide and MT peptide) were analyzed in the thermophoresis and temperature jump. For determination of the binding affinity (KD) of the wild type peptide, the recorded fluorescence was normalized to the fraction bound (0 = unbound, 1 = bound), and fitted using the KD fit formula derived from the law of mass action. Technical duplicates were performed for each experimental setup.
Sequence data from this article can be found in the EMBL/GenBank data libraries under the following accession numbers: Aly-Arabidopsis lyrata (D7M2M4); Ath-Arabidopsis thaliana (B9DFD5); Bna-Brassica napus (A0A078G686); Bol-Brassica oleracea (A0A0D3E0B2); Cru-Capsella rubella (R0GT49); Csa-Cucumis sativus (A0A0A0K5Q7); Egr-Eucalyptus grandis (A0A059A094); Gma-Glycine max (1M3S8); Gso-Glycine soja(A0A0B2R1M3); Hvu-Hordeum vulgare (Q6S5H8); Jcu-Jatropha curcas (A0A067JJP4); Mtr-Medicago truncatula (A0A072VBZ7); Nta-Nicotiana tabacum (Q75VR1); Osa-Oryza sativa (Q5QM84); Ptr-Populus trichocarpa (U5FYB3); Sit-Setaria italica (K3XEV7); Sly-Solanum lycopersicum (K4CFU2); Tae-Triticum aestivum (Q6YLX9); Tca-Theobroma cacao (A0A061E309); Zma-Zea mays (B6SP34).
A C-terminal region is essential for TPC1 function
We previously reported that deletion of the last 55 amino acids of AtTPC1, corresponding to the cytosolic carboxyl-terminus, resulted in a mutant (TPCΔC) which is correctly targeted to the tonoplast, but lacks channel function . This indicates an essential role of this region for the activity of TPC1, but not for the targeting or trafficking process.
To investigate whether a specific sub-region of the carboxyl-terminus is involved in this regulation, two additional truncated channel versions were created: TPC1ΔC8 contained the residues 1–725, and TPC1ΔC29 the residues 1–704. Both mutants were localized in the tonoplast, indicating their expression, efficient ER export, and correct targeting (Fig. 1b, c). The lack of the last 8 amino acids did not interfere with the activity of TPC1ΔC8 (Fig. 1b), and the amplitudes and current–voltage behavior of the mutant were wild type like (Fig. 1b, e). In contrast, although TPC1ΔC29 was expressed normally (Fig. 1c, Fig. S1), this mutant stayed silent like the TPC1ΔC mutant lacking the whole carboxyl-terminus (Fig. 1c–e). These results identified amino acids 705–725 to include a region indispensable for channel function.
Mutation in a predicted helical domain abolishes TPC1 activity
So far, no structural information is available about TPCs of plant or animal origin. To address the secondary structure of the C-terminus, different prediction tools were applied, which revealed a consensus helical sequence to span amino acids 707–723 (Table 1).
A current reduction was also obtained, when a point mutation was introduced at Ser706, which is located at the N-terminal end of the helix (Table 1). Since Ser706 represents a putative phosphorylation site, it was replaced either by the phosphorylation-mimicking Asp, introducing a negative charge, or by Ala. In both cases, the whole-vacuolar current amplitudes were largely reduced, by 58 % and 53 % for TPC1-S706D and TPC1-S706A, respectively (Fig. 2d, Fig. S3). The current reductions were also observed at non-saturating Ca2+-concentrations of 0.2 mM and 0.05 mM (Fig. 2d). Similar open probabilities of wild type and mutants at saturating Ca2+ concentrations (1 mM) and 0.2 mM, corresponding to the half-maximal activation concentration , showed that the current reductions did not result from a largely impaired Ca2+-dependent shift of the voltage dependence (Fig. S3). Besides shifting the voltage dependence to less negative potentials, an elevation of the cytosolic Ca2+ concentration also increases the maximum conductance . While TPC1-S706 wild type currents were reduced to 53 ± 10 % at 107 mV following the reduction of Ca2+ from 1 mM to 200 µM, this value was 28 ± 3 % for TPC1-S706A and 24 ± 4 % for TPC1-S706D (Fig. 2d). The S706 mutation therefore appears to modestly affect the Ca2+-sensitivity of channels via the link between the Ca2+-binding and change in the number of voltage-sensitive channels.
Together, these results show that Ser706 appears to have a structural role instead of being involved in channel regulation by phosphorylation. Furthermore we conclude that mutations in or near the C-terminal helix reduce the number of open channels rather than affecting the voltage dependence itself.
An alignment of the TPC1 carboxyl-termini of different species shows that the C-terminal helix (CTH), including S706 as well as L714, L715, and L719, is highly conserved among plants (Fig. 2e), implicating an important function of this domain. Our results strongly suggest that vacuoles harboring mutations in the C-terminal helix formed less functional two-pore channels. As helical structures are often involved in mediating protein–protein interactions, the CTH of TPC1 may be involved in protein–protein interactions required as a prerequisite for channel gating or stabilization of the open state.
TPC1 dimerizes via its C-termini
One possibility for a CTH-mediated interaction would be a dimerization of the cytosolic C-termini of two TPC1 subunits. To test this possibility, interaction studies were performed using bimolecular fluorescence complementation (BiFC) and co-immunoprecipitation (Co-IP).
TPC1-3LP C-terminally tagged with the Venus halves produced BiFC signals that were reduced compared to the wild type (Fig. 3d). Confocal fluorescence microscopy revealed that compared to the wild type, TPC1-3LP tagged with the Venus halves was less efficiently transported to the vacuole (Fig. 3c), since the fluorescence emerged not only from the vacuolar membrane, but also to significant amounts from other endomembranes, mostly the ER. This localization pattern of TPC1-3LP with BiFC was not seen with GFP-tagged TPC1-3LP in Arabidopsistpc1-2 cells (Fig. 2b, Fig. S1) and might hint to a slowed trafficking associated with the tobacco expression system compared to the Arabidopsis mesophyll protoplasts. Nevertheless, the reduced BiFC signals of the helix-breaking point mutant TPC1-3LP (Fig. 3c) in comparison to the wild type may indicate that the dimer formation via the C-terminus was hampered. However, the flexibility of the C-termini and Venus halves would still allow the formation of the fluorophore.
In contrast, combinations with N-terminally tagged TPC1 subunits did not result in Venus-fluorescence making intramolecular interactions of the N-terminus seem unlikely (Fig. 3d). Functional expression and tonoplast localization of an N-terminally tagged TPC1 were verified using a GFP tag (Fig. S4).
The CTH mediates formation of an antiparallel coiled-coil dimer
The CTH-mediated dimerization was further analyzed in silico, both for the wild type and for the CTH-3LA mutant. To this end, the dimerization was first analyzed from 100 aggregation simulations (250 ns each) for each system. Each dimerization simulation started from two C-terminal TPC α-helices at an initial peptide center of mass (COM) distance between 5.5 and 6.5 nm solvated in a box of water at coarse-grained resolution. Dimers were formed for both systems in all simulations within tens of nanoseconds, reflecting the observed enhanced stickiness of the coarse-grained MARTINI force field .
The relative instability and promiscuity of the CTH-3LA dimers observed in MD simulations corresponds to the lack of dimer formation in MST experiments. In the following, we focused on the analysis of dimer formation for the CTH wild type peptide.
Within approximately 50 ns, the wild type peptides approached each other in all simulations to a distance below 1 nm, reflecting dimerization (81 % of the monomers dimerized within 25 ns). Figure 7a shows the center of mass distances for each simulation. With increasing simulation time, cluster formation is seen, with one cluster at a COM distance of ≈1 nm (c1) and a second cluster at ≈1.45 nm (c2). Peptides in dimers of cluster c2 adopted a shifted configuration (see below).
The orientation of the C-terminal dimers of TPC1 was analyzed using Euler angles [47, 64]. A tilt angle below 50° indicates that the helices are in a parallel conformation, while an angle above 130° reflects an antiparallel orientation (Fig. 7b). The latter configuration was adopted in 86 % of the coarse-grained simulations, characterized by an average tilt angle of 160°. In 7 % of the simulations, the dimers ended up in a parallel orientation (average tilt of 18°), the remaining simulations ended up in diverse intermediate configurations. To reach the antiparallel conformation most of the dimers followed a specific dimerization pathway (Fig. 7c): At larger intermolecular distances (3–7 nm), i.e., at the beginning of the simulation, the distribution of tilt angles was random (Fig. 7c). Upon approach, the tilt was constrained to values of 10–50°, the peptides thus oriented in a parallel head-to-tail configuration. Subsequently, the peptides reoriented and ended up in the antiparallel configuration (130°–180°, Fig. 7c).
To evaluate the frequency of the different dimer conformations at the end of the simulations, combinations of the binding position β and rotation angle ϕ (compare Fig. 7b) obtained from the trajectories were plotted as a two-dimensional kernel density map (Fig. 7d). While the position of one monomer to a reference monomer is defined by the position β, the exact binding site that faces the reference structure is given by the phase ϕ, the rotation around its helical axis (Fig. 7b).
In the bound state, i.e., at the end of the simulations, positively and negatively charged amino acids come into close proximity, in particular in the antiparallel configuration (Fig. 7d). These antiparallel dimers (cluster A in Fig. 7d) are electrostatically favored by proximity of the N-terminal arginines of one monomer and the C-terminal glutamic and aspartic acids of the second monomer. Additionally, the hydrophobic residues leucine and valine are positioned at the dimer interface.
Besides the main cluster A, four additional clusters (B–E) were identified. The second-largest cluster B consists of antiparallel dimers that interact as well through electrostatic interactions. Due to a decreased number of hydrophobic residues at the dimer interface, this conformation is expected to be metastable. Parallel dimers were found in clusters C and D. In these arrangements, the helices are shifted with respect to each other, allowing for a better packing of hydrophobic residues (parallel dimer, Fig. 7d). Therefore, the center of mass distances between the corresponding monomers is increased as compared to the antiparallel dimers (compare c2 in Fig. 7a). For cluster E (tilt from 50°–130°) the inner leucines are in close proximity to each other. Furthermore, in this configuration the central aspartic acids are able to interact with the arginines.
To evaluate the binding strength, the non-bonded interaction energy was calculated for the different conformations. In agreement with the high frequency of antiparallel dimers (cluster A and B), their mean interaction energy was lowest among all dimers with a value of ≈−600 kJ/mol, while the parallel dimers interact with ≈−420 kJ/mol, and the rest found in cluster E with ≈−500 kJ/mol.
Less stable conformations were observed for the other dimers (clusters A, B, C). Here, leucines and valines were partially exposed to the solvent (Fig. 8a, ii and iii) or hydrogen bonds within the helix were broken and consequently the helix disrupted (Fig. 8a, i and ii). Additionally, hydrophilic histidines were partially found at the interface.
The combined coarse-grained and atomistic study thus strongly suggests that the wild type, but not the mutant CTHs of two TPC1 monomers form a stable helical dimer in an antiparallel coiled-coil conformation (Fig. 8b). Together with the results of the mutagenesis study, we propose the formation of an antiparallel CTH dimer as a prerequisite for TPC1 activity. This conclusion is further supported by the ability of the synthetic CTH peptide to rapidly inhibit TPC1 currents, when applied to the active channels in electrophysiological recordings (Fig. S7).
Here, we identified the presence of a carboxy-terminal helix (CTH) in TPC1 and documented its role for channel dimerization and function. Loss of the CTH rendered the channel inactive while point mutations within the CTH resulted in severely reduced channel activity. Correspondingly, BiFC analysis and Co-IP experiments reported on the interaction of the C-terminus, which was reduced by mutations within the CTH. A residual current of 10 % in the TPC1-3LA and -3LP mutant may indicate that in contrast to the CTH peptide the complete C-terminus may dimerize to a small extent even with three leucines replaced. Although we cannot exclude the possibility that in addition to the CTH other parts of the C-terminus are involved in the dimerization process, the lack of the CTH was sufficient to render the channel silent.
Dimerization of the CTH was directly shown in MST assays and MD simulations. The sequential multiscale simulations revealed that the wild type CTH dimer preferably adopts an antiparallel coiled-coil conformation. This antiparallel orientation is in accordance with an assumed head-to-tail configuration of the channels . The fact that leucine to alanine mutations in the inner helix drastically impaired channel function is in accordance with the observation that although alanine is a residue of high helix propensity, it can decrease the stability of the helix (dimer) [66, 68]. Indeed, coarse-grained MD simulations discovered mutated dimers that exhibit a significant increase of flexibility, pointing to less stable dimers. These results were corroborated by MST analyses, which showed that the wild type CTH dimerized with a Kd of 3.9 µM while the mutants did not interact. Apparently, the CTH is not required for correct assembly of the channel in the membrane. However, our results suggest that the CTH is crucial for stabilization of the TPC1 dimer and coupled with dimerization also for channel gating.
Gating of ion channels by their soluble termini is a common feature of ion channels [69, 70, 71, 72, 73], and often modification of the C-terminus induces alterations of the voltage dependence [74, 75, 76, 77]. In comparison, mutations within and near the CTH did not feedback onto the voltage dependence of TPC1. In this respect, TPC1 shares structural and functional properties with voltage-dependent sodium (NaV) channels. Deletion of the complete or distal part of the C-terminus from a prokaryotic NaV resulted in a complete or almost complete loss of sodium currents, respectively, while the trafficking of the channel and the voltage dependence of activation were unaffected . Combined MD simulations and electron paramagnetic resonance spectroscopy showed that a C-terminal helix of this channel forms a coiled-coil bundle involving four subunits, and that this tetramerization is essential for coupling to channel opening via a proximal C-terminal linker following S6. Most notably, the linker contains a negatively charged cluster, and several glutamate residues are also conserved in the plant TPC C-termini following S6 (Fig. 2). A role of the coiled-coil in stabilizing the sodium channel tetramer or dimer in case of TPC1, and in enabling the opening and closing of the pore during gating without disrupting the quaternary structure may thus represent a mechanism also valid for two-pore channels, which has to be further evaluated in future studies.
Besides voltage changes, TPC1 is activated by binding of Ca2+ ions to the EF-hands in the central linker domain between transmembrane S6 and S7, which apparently stabilizes the open state [13, 78, 79]. The rabbit skeletal muscle type 1 ryanodine receptor RyR1 is an intracellular Ca2+ release channel, which also belongs to the six-transmembrane superfamily. Similar to TPC1, activation of RyR1 by Ca2+ involves cytosolic EF-hands, here present in the N-terminus, while the C-terminus homodimerizes. An essential function of the RyR1 C-terminus requires the last 15 amino acids, deletion of which abolishes channel function . Recently, the crystal structure of the RyR1 revealed a putative mechanism for Ca2+-mediated gating involving the C-terminus [73, 81]. In RyR1 Ca2+-dependent changes in the conformation of the N-terminal EF-hand containing domain are transmitted to the pore via contacts with the C-terminal domain, inducing a change of the cytosolic aperture of the channel and stabilizing the open state [73, 81]. A RyR1-like allosteric mechanism may therefore also account for the role of the C-terminus in the Ca2+- and voltage-dependent activity of TPC1.
TPC1 is a large-conductance channel and is therefore tightly regulated to prevent ion leakage from the vacuole [reviewed in 2]. The many ionic cytosolic regulators, such as H+, Ca2+, Mg2+, K+, and Na+, exert their effects via alteration of the voltage dependence of the channel. In contrast, 14-3-3 proteins inhibit TPC1 activity by about 90 % within 10 s without any changes in the voltage dependence . A similar reduction in channel activity without affecting the voltage dependence was observed for the CTH mutants described here. Likewise, application of the synthetic CTH peptide to the vacuolar membrane rapidly inhibited TPC1 currents (Fig. S7). It is therefore tempting to speculate that the 14-3-3 protein GRF6, which regulates TPC1 in Arabidopsis , interferes with the dimerization of the carboxyl-termini. A region partly overlapping with the CTH contains two serines (S706, S708) and constitutes a putative 14-3-3 binding site [705-KSRSQR, 82]. Accumulation of GRF6 dimers may thus push the CTHs apart, which will destabilize the TPC1 dimer and induce a rapid closure of the channel pore. Alternatively, 14-3-3 proteins may act more indirectly, because a second putative 14-3-3 binding site is predicted to be located in the linker domain between the two EF-hands at T359 . Given a RyR1-like allosteric mechanism, 14-3-3 binding in the linker may interfere with the coordinated mechanism, which links Ca2+ binding to the activation of the channel via the C-terminal α-helical region. In any case, the coiled-coil mediated dimerization of wild type CTH’s as a prerequisite for channel opening adds a very rapid regulatory mechanism to shut down this large vacuolar conductance, either by destabilizing the TPC1 dimer, or by preventing the allosteric coupling to the channel pore.
This work was supported by the German Science Foundation (DFG): Research Training Group 1962-Dynamic Interactions at Biological Membranes. Computer time was provided by the Computing Center of the University Erlangen-Nürnberg (RRZE). We would like to thank Joanna Bogdanska-Urbaniak and Gudrun Steingräber for technical assistance, and Norbert Sauer (FAU Erlangen-Nürnberg) for sharing the confocal microscope.
- 4.Lin-Moshier Y, Keebler MV, Hooper R, Boulware MJ, Liu X, Churamani D, Abood ME, Walseth TF, Brailoiu E, Patel S, Marchant JS (2014) The two-pore channel (TPC) interactome unmasks isoform-specific roles for TPCs in endolysosomal morphology and cell pigmentation. Proc Natl Acad Sci USA 111(36):13087–13092. doi:10.1073/pnas.1407004111 CrossRefPubMedPubMedCentralGoogle Scholar
- 5.Sakurai Y, Kolokoltsov AA, Chen C-C, Tidwell MW, Bauta WE, Klugbauer N, Grimm C, Wahl-Schott C, Biel M, Davey RA (2015) Two-pore channels control Ebola virus host cell entry and are drug targets for disease treatment. Science 347(6225):995–998. doi:10.1126/science.1258758 CrossRefPubMedPubMedCentralGoogle Scholar
- 6.Wang X, Zhang X, Dong XP, Samie M, Li X, Cheng X, Goschka A, Shen D, Zhou Y, Harlow J, Zhu MX, Clapham DE, Ren D, Xu H (2012) TPC proteins are phosphoinositide- activated sodium-selective ion channels in endosomes and lysosomes. Cell 151(2):372–383. doi:10.1016/j.cell.2012.08.036 CrossRefPubMedPubMedCentralGoogle Scholar
- 8.Galione A (2011) NAADP Receptors. In: Perspectives in Biology, vol 3, 1. Cold Spring Harbor. doi:10.1101/cshperspect.a004036
- 12.Brailoiu E, Rahman T, Churamani D, Prole DL, Brailoiu GC, Hooper R, Taylor CW, Patel S (2010) An NAADP-gated two-pore channel targeted to the plasma membrane uncouples triggering from amplifying Ca2+ signals. J Biol Chem 285(49):38511–38516. doi:10.1074/jbc.M110.162073 CrossRefPubMedPubMedCentralGoogle Scholar
- 15.Boccaccio A, Scholz-Starke J, Hamamoto S, Larisch N, Festa M, Gutla P, Costa A, Dietrich P, Uozumi N, Carpaneto A (2014) The phosphoinositide PI(3,5)P2 mediates activation of mammalian but not plant TPC proteins: functional expression of endolysosomal channels in yeast and plant cells. Cell Mol Life Sci. doi:10.1007/s00018-014-1623-2 Google Scholar
- 16.Pitt SJ, Funnell TM, Sitsapesan M, Venturi E, Rietdorf K, Ruas M, Ganesan A, Gosain R, Churchill GC, Zhu MX, Parrington J, Galione A, Sitsapesan R (2010) TPC2 is a novel NAADP-sensitive Ca2+ release channel, operating as a dual sensor of luminal pH and Ca2+. J Biol Chem 285(45):35039–35046. doi:10.1074/jbc.M110.156927 CrossRefPubMedPubMedCentralGoogle Scholar
- 18.Calcraft PJ, Ruas M, Pan Z, Cheng X, Arredouani A, Hao X, Tang J, Rietdorf K, Teboul L, Chuang KT, Lin P, Xiao R, Wang C, Zhu Y, Lin Y, Wyatt CN, Parrington J, Ma J, Evans AM, Galione A, Zhu MX (2009) NAADP mobilizes calcium from acidic organelles through two-pore channels. Nature 459(7246):596–600. doi:10.1038/nature08030 CrossRefPubMedPubMedCentralGoogle Scholar
- 28.Bonaventure G, Gfeller A, Rodríguez VM, Armand F, Farmer EE (2007) The fou2 gain-of-function allele and the wild-type allele of Two Pore Channel 1 contribute to different extents or by different mechanisms to defense gene expression in Arabidopsis. Plant Cell Physiol 48(12):1775–1789. doi:10.1093/pcp/pcm151 CrossRefPubMedGoogle Scholar
- 29.Dadacz-Narloch B, Beyhl D, Larisch C, Lopez-Sanjurjo EJ, Reski R, Kuchitsu K, Müller TD, Becker D, Schönknecht G, Hedrich R (2011) A novel calcium binding site in the slow vacuolar cation channel TPC1 senses luminal calcium levels. Plant Cell 23(7):2696–2707. doi:10.1105/tpc.111.086751 CrossRefPubMedPubMedCentralGoogle Scholar
- 35.Latz A, Becker D, Hekman M, Müller T, Beyhl D, Marten I, Eing C, Fischer A, Dunkel M, Bertl A, Rapp UR, Hedrich R (2007) TPK1, a Ca2+-regulated Arabidopsis vacuole two-pore K+ channel is activated by 14-3-3 proteins. Plant J 52(3):449–459. doi:10.1111/j.1365-313X.2007.03255.x CrossRefPubMedGoogle Scholar
- 45.Schrödinger, LLC (2010) The PyMOL Molecular Graphics System, Version 1.3r1Google Scholar
- 49.Pluhackova K, Wassenaar TA, Böckmann RA (2013) Molecular Dynamics Simulations of Membrane Proteins. In: Rapaport D, Herrmann JM (eds) Membr Biogen, vol 1033, 1st edn. Humana Press, New york, pp 85–101. doi:10.1007/978-1-62703-487-6Google Scholar
- 52.Case DA, Babin V, Berryman JT, Betz RM, Cai Q, Cerutti DS, Cheatham TE, Darden TA, Duke RE, Gohlke H, Goetz AW, Gusarov S, Homeyer N, Janowski P, Kaus J, Kolossváry I, Kovalenko A, Lee TS, LeGrand S, Luchko T, Luo R, Madej B, Merz KM, Paesani F, Roe DR, Roitberg A, Sagui C, Salomon-Ferrer R, Seabra G, Simmerling CL, Smith W, Swails J, Walker RC, Wang J, Wolf RM, Wu X, Kollman PA (2014) AMBER 14. University of California, oaklandGoogle Scholar
- 64.Wassenaar TA (2006) Molecular dynamics of sense and sensibility in processing and analysis of data. University Library Groningen, GroningenGoogle Scholar
- 65.Kotelchuk D, Scheraga HA (1968) The influence of short-range interactions on protein conformation, I. Side chain-backbone interactions within a single peptide unit. Proc Natl Acad Sci USA 61:4Google Scholar
- 76.Hatano N, Ohya S, Muraki K, Clark RB, Giles WR, Imaizumi Y (2004) Two arginines in the cytoplasmic C-terminal domain are essential for voltage-dependent regulation of A-type K+ current in the Kv4 channel subfamily. J Biol Chem 279(7):5450–5459. doi:10.1074/jbc.M302034200 CrossRefPubMedGoogle Scholar
- 82.Dinkel H, Van Roey K, Michael S, Davey NE, Weatheritt RJ, Born D, Speck T, Krüger D, Grebnev G, Kuban M, Strumillo M, Uyar B, Budd A, Altenberg B, Seiler M, Chemes LB, Glavina J, Sanchez IE, Diella F, Gibson TJ (2014) The eukaryotic linear motif resource ELM: 10 years and counting. Nucl Acids Res 42:D259–D266. doi:10.1093/nar/gkt1047 (Database issue) CrossRefPubMedGoogle Scholar
- 89.Remmert M, Biegert A, Hauser A, Soding J (2012) HHblits: lightning-fast iterative protein sequence searching by HMM-HMM alignment. Nat Meth 9 (2):173–175. doi:http://www.nature.com/nmeth/journal/v9/n2/abs/nmeth.1818.html—supplementary-information
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.