Revealing the glutamate switch: Insights into proton translocation in plant H⁺-pyrophosphatase

Abstract

Membrane-bound pyrophosphatases (M-PPases) are responsible for the hydrolysis of pyrophosphate (PPi), coupled with the pumping of H⁺ and/or Na⁺ across the membrane. In Vigna radiata H⁺-translocating pyrophosphatase (VrH⁺-PPase), proton translocation involves both a “binding change” mechanism, where PPi binding induces proton translocation, and the “Grotthuss-chain” mechanism, which describes proton translocation along the ion gate, hydrophobic gate, and exit channel. However, the dynamic coupling between protonation states and conformational changes in VrH⁺-PPase remains unclear, partly due to the challenges of experimentally capturing transient states during transport. To address this, we employed constant pH molecular dynamics and classical molecular dynamics simulations to elucidate the proton transport mechanism from the ion gate (R242/D294/K742/E301) to the hydrophobic gate (L232/A305/L555/V746). Our simulations reveal that K742 becomes deprotonated upon PPi binding, suggesting its potential role as an internal proton donor. When K742 is deprotonated, E301 penetrates the hydrophobic gate and creates a hydrophilic environment for proton transport. Following PPi hydrolysis, D294 accepts a hydrolysis-generated proton to become protonated, inducing R242 to act as a positive plug that prevents the reprotonation of D294. Meanwhile, the protonation of D294 causes E301 to rebound and close the hydrophobic gate. We propose that E301 acts as a molecular switch, regulating proton transport through the hydrophobic gate. Furthermore, we suggest that the penetration of glutamate is a conserved feature among plant H⁺-PPases, maintaining a consistent hydrophilic environment at the hydrophobic gate in H⁺-PPases. In conclusion, proton translocation in plant H⁺-PPases involves lysine deprotonation in the PPi-bound state and aspartate protonation in the PPi-hydrolyzed states, with the glutamate switch dynamically regulating the opening and closing of the hydrophobic gate.

Significance

By integrating constant pH molecular dynamics (CpHMD) simulation and classical molecular dynamics (MD) simulation, this study connects experimental findings with computational models to understand the proton translocation in plant H⁺-PPases. We reveal a detailed mechanism involving R242, D294, K742, and E301 that defines the dynamics of the ion gate and directs proton flow. Specifically, our findings identify E301 as a molecular switch that regulates the opening and closing of the proton translocation pathway and functions as a conditional proton carrier. Meanwhile, R242 acts as a positive plug that blocks extracellular protons, while K742 and D294 directly mediate proton transfer as the proton donor and acceptor, respectively. These molecular-level insights provide a comprehensive view of the “pumping-before-hydrolysis” mechanism and highlight a conserved strategy across plant H⁺-PPases.

Introduction

Membrane-bound pyrophosphatases (M-PPases) (E.C 3.6.1.1) play a crucial role in pyrophosphate (PPi) hydrolysis or synthesis, coupled with the translocation of protons (H⁺) and/or sodium (Na⁺) ions (1). These proteins are essential for survival under diverse abiotic stress conditions (2). M-PPases have been identified in plants and various unicellular organisms, with the notable exceptions of fungi and animals (3,4,5). The absence of M-PPase activity in protists is associated with abnormal biological functions, and their growth is inhibited by PPi analogs (4). M-PPases are typically dimeric, with a molecular weight of 66–90 kDa, and composed of 16 transmembrane helices (TMs) (1,6). Concerning the residue-determining function (alanine or lysine), M-PPases are divided into two main types: K⁺ dependent and K⁺ independent (7). K⁺-dependent M-PPases contain a conserved alanine and require K⁺ ion for their function, while K⁺-independent M-PPases have a conserved lysine residue that substitutes for K⁺ ion and activates the M-PPases (2). M-PPases utilize Mg₂PPi as their substrate and require Mg²⁺ ions for activity (2), binding two activating Mg²⁺ ions (K_d = 20–42 μM and 0.25–0.46 mM) (8) with an additional inhibitory Mg²⁺ ion binding site (K_d ≈ 100 mM) (9) present in the active site.

The three-dimensional structures of the plant H⁺-PPase from Vigna radiata (VrH⁺-PPase) (10) and the bacterial Na⁺-PPase from Thermotoga maritima (TmNa⁺-PPase) (11) have been determined, revealing that M-PPases exhibit similar inner/outer rosette folds composed of 16 TMs (10,11,12). The six core TMs (TM5-6/TM11-12/TM15-16) collectively constitute the PPi-binding pocket, the ion gate, the hydrophobic gate, and the exit channel leading to the lumen (in plants and protozoan parasites) or periplasm (in archaea and bacteria) (10,11,12) (Fig. 1). The ion gate of VrH⁺-PPase is primarily composed of three highly conserved residues, R242, D294, and K742, and one semiconserved residue, E301 (Fig. 1 B). Additionally, D294 and K742 form an ion pair, while E301 forms a hydrogen bond with the ion selectivity residue, S235 (13). While R-D-K-E propels protons via the “Grotthuss-chain” mechanism in VrH⁺-PPase (10), TmNa⁺-PPase utilizes D-K-E as an ion triplet for sodium transfer (11). The hydrophobic gate of VrH⁺-PPase consists of L232, A305, L555, and V746, preventing the proton gradient counterflow (10) (Fig. 1 B). Recently, Tsai et al. (14) proposed that E225 and R562 may act as the latch to regulate proton release in the exit channel of VrH⁺-PPase (Fig. 1 B), and identified the 10 ordered waters (WAT1-WAT10), outlining the pathway of proton transfer along the ion gate, the hydrophobic gate, and the exit channel.

Figure 1.

Overall structure of VrH⁺-PPase. (A) The 3D structure of the VrH⁺-PPase-IDP complex monomer (PDB: 4A01). The six core TMs and outer TMs are colored green and white, respectively. IDP and five Mg²⁺ ions are colored in CPK. (B) The substructures are in four functional regions: PPi binding site, ion gate, hydrophobic gate, and exit channel. D287 and D731 coordinated a nucleophilic water. R242, D294, K742, and E301 are involved in the proton translocation. The ion selectivity residue S235. L232, A305, L555, and V746 constitute the hydrophobic gate. Crucial exit channel residues include E225, T228, and R562.

Previously, Baykov et al. (15) proposed a “direct-coupling” mechanism for M-PPases, suggesting that PPi hydrolysis is the initial step for H⁺ and/or Na⁺ pumping. However, the binding of nonhydrolyzable PPi analogs, imidodiphosphate (IDP) and methylenediphosphonate (MDP), could induce a single turnover ion-pumping event (16,17), proving that substrate binding drives pumping and revealing a “binding change” mechanism, which unifies the order of both H⁺ and Na⁺ pumping and PPi hydrolysis during the catalytic cycle. Furthermore, Vidilaseris et al. (18) proposed that M-PPases could be an asymmetric dimer for their biological functions, based on the crystal structure of TmNa⁺-PPase in complex with the IDP and ATC (PDB: 6QXA), where the ATC dimer was bound only to one subunit of TmNa⁺-PPase. Anashkin et al. (19) suggested that one subunit would hydrolyze PPi and initiate proton translocation, while the other subunits might store the energy to induce the dimeric M-PPase to change the conformation. Additionally, recent findings by Strauss et al. (13), based on time-resolved data sets from TmNa⁺-PPase, indicate that M-PPases exhibit anticooperative behavior, suggesting that PPi binding cannot occur simultaneously in both subunits. In the unified model of M-PPase catalysis (13), which includes three possible dimers during the asymmetric cycle, PPi/Apo, 2Pi/Apo, and 2Pi/PPi dimers, it is proposed that each subunit independently undergoes PPi binding and PPi hydrolysis. However, the detailed mechanism of proton translocation and the associated conformational dynamics in VrH⁺-PPase remain unclear, particularly regarding the molecular basis linking PPi hydrolysis to proton pumping.

In this study, we employed constant pH molecular dynamics (CpHMD) and classical molecular dynamics (MD) simulations to explore proton translocation along the ion gate and the hydrophobic gate, focusing on the protonation and deprotonation in PPi-bound and PPi-hydrolyzed states. Our findings integrate a direct-coupling mechanism (15), the Grotthuss-chain mechanism (10), and a binding change mechanism (16) to elucidate the potential role of R242, D294, K742, and E301 in proton transport. Specially, we found that R242 prevents the reprotonation of D294. D294 acts as a hydrolysis-generated proton acceptor. K742 serves as an internal proton donor initiating proton translocation. E301 is proposed to function as a molecular switch at the hydrophobic gate, regulating proton translocation. Furthermore, we suggest that the penetration of E301 in VrH⁺-PPase represents a conserved feature among plant H⁺-PPases, supporting a consistently hydrophilic environment within the hydrophobic gate in H⁺-PPase.

Materials and methods

Initial structures

To generate the initial structure for MD simulation of VrH⁺-PPase complexes with PPi and 2Pi, we established the force field for ligands using the antechamber of AmberTools20 (20). Following the proposed asymmetric dimer of M-PPase (13,21), we constructed three catalytic states of the VrH⁺-PPase, including the ligand-free state as the Apo/Apo dimer, the PPi-bound state as the PPi/Apo dimer, and the PPi-hydrolyzed state as the 2Pi/Apo dimer. The Apo/Apo dimer was modeled based on the VrH⁺-PPase-Pi complex (PDB: 5GPJ) (16) with the Pi removed from both subunits. The PPi/Apo dimer was modeled based on the VrH⁺-PPase-IDP complex (PDB: 4A01) (10), where one subunit had IDP replaced with PPi, while the other subunit had its three IDP-related Mg²⁺ ions and IDP removed (17,19). Similarly, the 2Pi/Apo dimer was modeled based on the VrH⁺-PPase-2Pi complex (PDB: 6AFS) (14), where one subunit retained 2Pi, while the other subunit had its three 2Pi-related Mg²⁺ ions and 2Pi removed (17,19). The missing loops in all VrH⁺-PPase structures were built using MODELLER 9v14 with default settings in Discovery Studio 2019 (22).

Furthermore, the Apo/Apo, PPi/Apo, and 2Pi/Apo dimer complexes were embedded, using the Membrane Builder of CHARMM-GUI (17,19,23,24), into the same membrane system of a combination with a lipid mixture of phosphatidylethanolamine and phosphatidylcholine at a ratio of 1:1 and a water layer with mimic 250 mM MgCl₂. The size of the MD simulation system is 150.375 × 150.375 × 137.707 ų. The crystallographic water molecules in the core TMs from the VrH⁺-PPase-IDP complex and the VrH⁺-PPase-2Pi complex were preserved. The VrH⁺-PPase-Pi complex added the same water molecules from the VrH⁺-PPase-2Pi complex. The tLEaP of AmberTools20 (20) was employed, utilizing the force field parameters, ff14SB for protein and ions (25), GAFF2 for PPi and 2Pi, Lipid14 for lipids (26), and the water model TIP3P.

Residue modification

According to the hypothesis proposed in this study, residues D294, K742, and E301 were modified in protonated and deprotonated states. In the classical MD simulations, the deprotonated forms were represented as ASP, LYN, and GLU, while the protonated forms were represented as LYS, ASH, and GLH. Additionally, in CpHMD simulations, the titratable residues were defined using AS4, LYS, and GL4.

Classical MD simulation

Each MD simulation for VrH⁺-PPase complexes was done using AMBER20 (20) with three independent repeats. Firstly, preminimization was performed by Sander to relieve the strain and steric clashes generated by newly modeled VrH⁺-PPase complexes. The minimization is calculated as 2500 steps of steepest descent and then 2500 steps of a conjugate gradient in all three sections. The first section constrains all atoms of protein. In the second section, the backbone of the protein was constrained, and the side chains of the protein, water, and lipid were minimized. In the third section, the entire system was optimized without any constraints. Subsequently, the system was heated from 0 to 300 K over 100 ps under the NVT ensemble with restraints applied to the protein backbone, then equilibrated under the NPT ensemble for 100 ps with the same restraints. Next, the unrestrained equilibration was conducted under the NPT ensemble for 100 ps. Finally, the production of MD simulation for 200 ns was performed from the NPT ensemble at 300 K and 1 atm. Different parameters were taken for the trajectory analysis, that is, root mean-square deviation and atom-atom distance, which were performed using the analysis tools by Cpptraj of AmberTools20 (20). All structures are displayed using Discovery Studio 2019 (22).

CpHMD simulation

CpHMD simulation was done using AMBER20 (20), applying discrete protonation states in an explicit solvent for VrH⁺-PPase complexes. To determine the fraction of different protonation states at the ion gate, the titratable residues D294, K742, and E301 were analyzed at different pH values. The method involves standard MD simulations being propagated in explicit solvent and protonation state changes being attempted in implicit solvent at fixed intervals. Three 100 ns CpHMD simulations were performed with the pH range of 5.5, 6.5, and 7.2. Protonation state changes were attempted every 200 fs for all CpHMD simulations. The solvent relaxation time was set to 200 fs as suggested by Swails et al. (27).

Results

The protonation states of D294, K742, and E301 at the ion gate

To accurately identify H⁺ donors using MD simulations and pKa calculations, it is essential to incorporate protonation dynamics (28). We applied CpHMD simulations to three catalytic states of the VrH⁺-PPase, including the ligand-free state, the PPi-bound state, and the PPi-hydrolyzed state. These states are modeled as the Apo/Apo dimer, PPi/Apo dimer, and 2Pi/Apo dimer, respectively. Our aim was to elucidate how PPi binding induces proton translocation, following the binding change mechanism (16). The ligand-free state was modeled as the Apo/Apo dimer, based on the VrH⁺-PPase-Pi complex (16) with Pi removed from both subunits. The PPi-bound state was modeled as an asymmetric PPi/Apo dimer (13,21) based on the VrH⁺-PPase-IDP complex (10), in which one subunit was modified by replacing IDP with PPi, and the other was generated by removing IDP along with its associated Mg²⁺ ions. Similarly, the PPi-hydrolyzed state was modeled as an asymmetric 2Pi/Apo dimer (13,21) based on the VrH⁺-PPase-2Pi complex (14), with one subunit generated by removing 2Pi along with its associated Mg²⁺ ions.

Subsequently, we performed three independent CpHMD simulations on each dimer for 100 ns at pH 5.5, 6.5, and 7.2, representing the physiological pH range between the vacuole and cytoplasm in the plant cells. The proton translocation pathway in the VrH⁺-PPase involved three titratable residues, D294, K742, and E301, generating eight possible protonation states (e.g., D⁰/K⁰/E⁰, D⁰/K⁰/E¹, D⁰/K¹/E⁰, D⁰/K¹/E¹, D¹/K⁰/E⁰, D¹/K⁰/E¹, D¹/K¹/E⁰, and D¹/K¹/E¹ states, where “0” denotes the deprotonated state and “1” denotes the protonated state).

We examined the probability of each protonation state in the ligand-free subunit of the Apo/Apo dimer, the PPi-binding subunit of the PPi/Apo dimer, and the 2Pi-binding subunit of the 2Pi/Apo dimer under pH 5.5, 6.5, and 7.2 (Fig. 2). In the Apo/Apo dimer, the D⁰/K¹/E⁰ state was predominant at pH 5.5. At pH 6.5, the D⁰/K⁰/E⁰ and D⁰/K¹/E⁰ states occurred with nearly equal probabilities, whereas at pH 7.2 the D⁰/K⁰/E⁰ state became dominant (Fig. 2 A). These results suggest that increasing pH may affect the protonation state of K742, shifting it from a protonated to a deprotonated state. In the PPi/Apo dimer, the D⁰/K⁰/E⁰ state was consistently the most probable across all three pH conditions (Fig. 2 B), implying that PPi binding might stabilize K742 in a deprotonated state. In the 2Pi/Apo dimer, the D⁰/K⁰/E¹ state was dominant at pH 5.5, whereas the D⁰/K⁰/E⁰ state became more probable at pH 6.5 and 7.2 (Fig. 2 C). These findings indicate that, while K742 remains deprotonated after PPi hydrolysis, E301 might become protonated at pH 5.5.

Figure 2.

The probability of eight protonation states of D294, K742, and E301 under different solvent pH in three catalytic states. The probability of eight protonation states in (A) Apo/Apo dimer, (B) PPi/Apo dimer, and (C) 2Pi/Apo dimer. The “0” denotes the deprotonated state and “1” denotes the protonated state. Three different pH values of 5.5, 6.5, and 7.2 are colored in black, light gray, and gray, respectively. Bars represent the mean ± SD from three independent simulations for each condition.

Given that the vacuolar pH typically ranges from 5.5 to 6.0 (29,30), our results suggest that K742, initially protonated in the ligand-free state, becomes deprotonated upon PPi binding and remains deprotonated after PPi hydrolysis. At higher pH, K742 deprotonation persists across all catalytic states. Taken together, these observations support the hypothesis that PPi binding induces K742 deprotonation at the ion gate, which suggests that K742 might serve as a potential proton donor in VrH⁺-PPase.

E301 penetrates the hydrophobic gate

As we mentioned above, K742 potentially undergoes deprotonation at the ion gate upon PPi binding, where the ion gate mainly adopts the D⁰/K⁰/E⁰ state. To explore the detailed conformational change at the ion gate, we performed three independent equilibrium classical MD simulations for 200 ns on the PPi/Apo dimer (Fig. S1 A), where D294, K742, and E301 were deprotonated in the PPi-binding subunit. To investigate the conformational change of the ion gate, we measured the distances from R242, D294, K742, and E301 to T228 (Fig. S2 A). R242 and D294 consistently maintained distances of 22.7 ± 0.7 and 18.6 ± 0.7 Å from T228, respectively. Interestingly, K742 exhibited a fluctuating distance from 8.5 to 15.1 Å, suggesting that the deprotonated K742 undergoes a swinging motion between D294 and E301 (Fig. S3 A). Notably, E301 showed a descending conformational change from 7.8 to 3.6 Å.

To further understand the dynamics of E301, we captured three snapshots at 2.49, 2.6, and 3.0 ns, revealing a rapid multistep transition (Fig. 3). At 2.49 ns, E301 was positioned above the hydrophobic gate and was 7.6 Å away from T228. At 2.6 ns, E301 turned toward L555 and approached T228, reaching a distance of 5.9 Å. Ultimately, at 3.0 ns, E301 penetrated the hydrophobic gate and stabilized its conformation at 4.1 Å from T228. Initially, E301 was stabilized a hydrogen bond with S235. Upon E301 penetrating the hydrophobic gate and entering into the exit channel, the hydrogen bond shifted from S235 to T228, with the Cα of E301 moving downward by 1.5 Å (Fig. 3 B). We then measured the pore diameter along the proton translocation pathway using PoreWalker (31), and observed a significant widening at the hydrophobic gate. When E301 penetrated, the pore diameter reached 3.4 Å (Fig. S3 B). The potential water wire in VrH⁺-PPase was analyzed before and after the penetration of E301 using MOLEonline (32) (Fig. S4). Before penetration, the water wire was interrupted at the ion gate and hydrophobic gate (Fig. S4 A). After penetration, a continuous water wire was observed, extending from K742 to E301 and through the hydrophobic gate to the exit channel (Fig. S4 B).

Figure 3.

The conformational change of penetrating E301 in the PPi-bound state. (A) Distance between E301 (CD) and T228 (CA) in the initial 6 ns. Three independent MD simulations for 200 ns, indicated in black, gray, and light gray lines. (B) The three representative structures from Run1 are colored white, cyan, and orange, indicating the process of penetrating E301 and forming a hydrogen bond with T228. The deprotonation (dp) of D294, K742, and E301 are labeled. The structure shows the region of the ion gate (R242/D294/K742/E301), hydrophobic gate (L555), exit channel (T228), and related TMs (TM5-6,12,16). The hydrogen bond was displayed as a dashed line and colored in black.

These conformational changes suggest that deprotonation of K742 increases its flexibility at the ion gate, and enables E301 to penetrate and open the hydrophobic gate, thereby creating a hydrophilic environment that may facilitate proton translocation.

Essential features for E301 penetration

Since we identified the D⁰/K⁰/E⁰ state in the PPi-bound state that features both deprotonated K742 and E301, we sought to determine whether similar conformational changes of the ion gate could occur when K742 or E301 is protonated. Specifically, we were interested in whether the swinging motion of K742 and the penetration of E301 would still be observed under these alternative protonation states.

To address this, we performed three independent equilibrium classical MD simulations for 200 ns under three other protonation states, including D⁰/K⁰/E¹, D⁰/K¹/E⁰, and D⁰/K¹/E¹ (Fig. S1, B–D). In these simulations, conformational measurements at the ion gate revealed stable profiles, with no detectable swinging motion of K742 or downward movement of E301 (Fig. S2, B–D). Representative structures from these protonation states revealed distinct hydrogen bond networks at the ion gate, which likely restrict the dynamics of K742 and E301 (Fig. 4). In the D⁰/K⁰/E¹ state (Fig. 4 A), protonated E301 formed hydrogen bonds with S235 and the backbone of K742. In the D⁰/K¹/E⁰ state (Fig. 4 B), protonated K742 formed the hydrogen bond with D294 and E301, while E301 maintained a hydrogen bond with S235. In the D⁰/K¹/E¹ state (Fig. 4 C), protonated K742 formed the hydrogen bond with D294, and protonated E301 formed hydrogen bonds with S235 and the backbone of K742. None of these protonation states exhibited the characteristic conformational features observed in the D⁰/K⁰/E⁰ state, such as the swinging of K742 (Fig. S3 A) or penetration of E301 into the hydrophobic gate (Fig. 3 B). These results suggest that deprotonation of both K742 and E301 is a prerequisite for initiating the conformational changes necessary to open the hydrophobic gate. Therefore, the deprotonated states of K742 and E301 appear essential for inducing conformational changes at the ion gate.

Figure 4.

The conformation of the ion gate in other protonation states. The representative structures at 200 ns in PPi/Apo dimer under (A) D⁰/K⁰/E¹, (B) D⁰/K¹/E⁰, and (C) D⁰/K¹/E¹ states. The protonation (p) and deprotonation (dp) of D294, K742, and E301 are labeled. The structure shows the region of the ion gate (R242/D294/K742/E301), S235, hydrophobic gate (L555), exit channel (T228), and related TMs (TM5-6,12,16). The hydrogen bond was displayed as a dashed line and colored in black.

E301 rebounds above the hydrophobic gate

Previous research has proposed a direct-coupling mechanism (15), which posits that the hydrolysis-generated proton is transferred after PPi hydrolysis. Since we identified the D⁰/K⁰/E⁰ state as the predominant state upon PPi binding, the ion gate is expected to accept the hydrolysis-generated proton after PPi hydrolysis, leading us to assume that the ion gate could transition to the D¹/K⁰/E⁰ state. However, CpHMD simulations revealed that the D¹/K⁰/E⁰ state appeared only at pH 5.5 with a probability of less than 1% (Fig. 2 C). This suggests that the D¹/K⁰/E⁰ state may represent a transient intermediate that is not stable enough to be maintained. Nevertheless, based on our assumption, we performed three independent equilibrium MD simulations for 200 ns on the 2Pi/Apo dimer (Fig. S1 E) to explore the detailed conformational change at the ion gate. In this simulation, we considered both the deprotonated state of K742, as the first proton had been pumped, and the protonated state of D294, as it accepted the hydrolysis-generated proton in the 2Pi-binding subunit. From the measurements of conformational changes at the ion gate (Fig. S2 E), R242 exhibits the ascending conformational change from 20.7 to 25.3 Å, different from the PPi/Apo dimer. Besides, D294 and K742 maintained distances of 17.1 ± 0.8 and 13.9 ± 0.7 Å from T228, respectively. K742 did not exhibit a swinging motion. Particularly, E301 displayed an obvious ascending conformational change from 3.8 to 9.8 Å.

To further analyze the ascending dynamics of E301, we captured snapshots at 1.0, 1.5, and 2.7 ns, exploring the upward shift of E301 (Fig. 5). At 1 ns, E301 initially penetrated the hydrophobic gate and located 4.1 Å away from T228, consistent with the effect observed when K742 is deprotonated. At 1.5 ns, E301 appeared to raise its sidechain, similar to the observation in K742A, positioned 7.4 Å away from T228. Eventually, at 2.7 ns, E301 rebounded above the hydrophobic gate, maintaining this conformation, and located 10.0 Å away from T228 (Fig. 5 B). Remarkably, E301 did not form a hydrogen bond with T228 during the MD simulation. Additionally, when E301 rebounded above the hydrophobic gate, K742 would form a hydrogen bond with D294 and E301.

Figure 5.

The conformational change of rebounding E301 in the PPi-hydrolyzed state. (A) Distance between E301 (CD) and T228 (CA) in the initial 15 ns. Three independent MD simulations for 200 ns, indicated in black, gray, and light gray lines. (B) The three representative structures from Run1 are colored white, cyan, and orange, indicating the process of rebounding E301 and forming a hydrogen bond with K742. The protonation (p) and deprotonation (dp) of D294, K742, and E301 are labeled. The structure shows the region of the ion gate (R242/D294/K742/E301), hydrophobic gate (L555), exit channel (T228), and related TMs (TM5-6,12,16). The hydrogen bond was displayed as a dashed line and colored in black.

Based on these findings, we proposed that PPi hydrolysis provides the second proton to protonate D294, leading to the rebound of E301. This structural transition from penetration to rebound provides valuable insights into the proton translocation process in VrH⁺-PPase.

Two distinct conformations of the ion gate

In both key protonation states, the D⁰/K⁰/E⁰ state in the PPi/Apo dimer and the D¹/K⁰/E⁰ state in the 2Pi/Apo dimer, we identified the distinct conformational changes of E301, including E301 penetration (Fig. 3) and E301 rebound (Fig. 5). Beyond these findings, structural comparisons between these two states revealed a notable difference in the conformation of R242, which appears to be regulated by the protonation state of D294 (Fig. 6). In the PPi/Apo dimer, deprotonated D294 formed a hydrogen bond with R242, causing R242 to adopt a conformation oriented toward D294 (Fig. 6 A). Conversely, in the 2Pi/Apo dimer, protonated D294 did not form a hydrogen bond with R242. Instead, protonated D294 interacted with K742, restricting its swinging motion. As a result, R242 shifted 3.8 Å toward the cytosolic side and established a hydrogen bond with D287 (Fig. 6 B).

Figure 6.

Two distinct conformations of R242 depending on the protonation state of D294. The conformations of R242 are distinguished by the protonation of D294. (A) Upon PPi binding, D294 is deprotonated, and R242 forms a hydrogen bond with D294. (B) Upon PPi hydrolysis, D294 becomes protonated, promoting R242 to shift upward by 3.8 Å and forming a hydrogen bond with D287, while D294 forms a hydrogen bond with K742. The protonation (p) and deprotonation (dp) of D294 and K742 are labeled. The structures show the residue, R242, D294, K742, D287, D731, related TMs (TM5-6,12,16), and PPi or 2Pi. The hydrogen bond was displayed as a dashed line and colored in black.

These structural comparisons suggest that, upon PPi hydrolysis, D294 accepts the hydrolysis-generated proton, triggering the conformational change of R242 toward the cytosolic side. This movement causes D294 to reorient toward K742, potentially facilitating proton transfer from D294 to K742. Our findings indicate that R242 retains its positive charge at the entry of the ion gate and participates in proton translocation through the interaction with D294 or D287, which depends on the protonation state of D294.

Discussion

K742 acts as an initial proton donor

Previously, Lin et al. (10) proposed that the proton transfer mechanism in VrH⁺-PPase is similar to the Grotthuss-chain mechanism observed in bacteriorhodopsin, suggesting that K742 plays a role in regulating the protonation and deprotonation of D294 and E301 in VrH⁺-PPase. Typically, rhodopsins use Asp (33,34,35,36) or Glu (37,38) residues as proton donors to reprotonate the retinal Schiff base. However, in the case of Exiguobacterium sibiricum rhodopsin, the retinal Schiff base accepted the proton from K96, which acts as an internal proton donor, different from the typical involvement of Asp or Glu residues (39,40,41). Based on these results, we suggest that K742 functions as an internal proton donor in VrH⁺-PPase, similar to K96 in Exiguobacterium sibiricum rhodopsin. Despite differences in their structural environments, K742 and K96 illustrate how lysine side chains within proton transport in membrane proteins can deprotonate to supply protons. The observed depression of K742’s pKa below the standard 10.4 for lysine in water further supports this role, aligning with prior evidence that buried lysine residues can adopt low pKa values when buried in proteins without conformational reorganization (42).

Furthermore, K742 acting as an internal proton donor explains the proton-transmitted signal observed from nonhydrolyzable inhibitors such as IDP and MDP (16,17). As Li et al. (16) reported, a positive signal is observed with PPi and absent with KH₂PO₄ in Nanion SURFE²R N1, indicating that VrH⁺-PPase exhibits substrate specificity that is essential for proton pumping. Consequently, our data showing the probability of K742 being deprotonated in the PPi-binding subunit (Fig. 2) supports a binding change mechanism that PPi binding induced K742 supply proton then proton pumping in VrH⁺-PPase. Furthermore, mutagenesis studies of K742M and K742R, which result in the loss of PPi hydrolysis and proton translocation in VrH⁺-PPase (10), provide additional support for the idea that K742 may need to alternate between deprotonated and protonated states in VrH⁺-PPase.

The comparison between K742 in VrH⁺-PPase and K707 in TmNa⁺-PPase highlights the different roles of these highly conserved lysine residues. In TmNa⁺-PPase, K707 forms the ion triplet with D243 and E246 to stabilize the Na⁺ binding site, which is crucial for Na⁺ transmission (11,12). In contrast, VrH⁺-PPase, following the Grotthuss-chain mechanism (10,12), does not necessitate a specific binding site for the pumped proton. Our findings suggest that K742 in VrH⁺-PPase undergoes deprotonation upon PPi binding, serving as the initial proton donor for proton translocation. Additionally, K742 undergoes reprotonation after PPi hydrolysis, storing the hydrolysis-generated proton for the next proton translocation.

E301 acts as a molecular switch at the hydrophobic gate

Proton translocation forming a water wire through coordinated water molecules is observed in various biological systems, including bacteriorhodopsin (35,43), cytochrome c oxidase (44), and VrH⁺-PPase (10,14). In VrH⁺-PPase, Tsai et al. (14) identified the 10 functional waters through the ion gate (WAT1-WAT2) to the hydrophobic gate (WAT3-WAT4) and exit channel (WAT5-WAT10). The presence of WAT3 and WAT4 within the hydrophobic gate implies that solvation may occur to facilitate proton translocation. E301 is located at the end of the ion gate, positioned right next to the hydrophobic gate (Fig. 1 B). Mutagenesis studies demonstrated that mutations at E301, such as E301A (10) and E301Q (14), result in the loss of proton translocation activity. Notably, the crystal structure of E301Q (PDB: 6AFT) revealed that a hydrogen bond between Q301 and the main chain of K742 fixes the flexibility of Q301 and thus affects its proton pumping activity, suggesting that E301 requires flexibility during proton translocation (14).

Our study contributes to understanding the initiation of conformational changes in E301. We propose that, upon PPi binding, induces deprotonation of K742 and activates E301 to penetrate the hydrophobic gate. Remarkably, the penetrated E301 is located at the corresponding position of WAT3 (Fig. S5) and creates a suitable pore at the hydrophobic gate that facilitates proton transfer (Fig. S3 B). Our findings are consistent with the suggestion by Tsai et al. (14) that the hydrophobic gate can be solvated and its diameter adjusted to facilitate proton transfer. Furthermore, during E301 penetration, a continuous water wire was observed, which extended from the ion gate through the hydrophobic gate to the exit channel (Fig. S4), similar to the water wire from WAT1-WAT10 proposed by Tsai et al. (14).

Moreover, the conformational changes of E301 in VrH⁺-PPase correspond to similar actions observed in other proteins involved in proton transfer. For instance, in respiratory complex I from both Thermus thermophilus and humans, the transient opening of the proton pathway is facilitated by the downward flip of E32 and E34, functionally analogous to the penetration of E301 in VrH⁺-PPase. This highlights a common structural strategy for modulating proton pathways across diverse biological systems (45,46). Similarly, in anion channelrhodopsin from Guillardia theta (GtACR1) the reorientation of E68 leads to the closing of the channel bottleneck (47), akin to the rebound of E301 in VrH⁺-PPase, further emphasizing the versatile role of glutamate residues in proton channel regulation.

While E301 is part of the proton translocation pathway, it has been suggested that E301 directly participates in proton transfer and undergoes protonation and deprotonation (10,16). The classical MD simulation of protonated E301 (Fig. 4, A and C) and the crystal structure of E301Q (14) indicate that E301/Q301 forms a hydrogen bond with the backbone of K742, which restricts the flexibility required for E301 to penetrate. However, the CpHMD simulation of the 2Pi/Apo dimer reveals that the D⁰/K⁰/E¹ state becomes the major state during 2Pi binding (Fig. 2 C). Taken together, we propose that the penetration of E301 plays a key role in opening the hydrophobic gate and creating a hydrophilic environment for protons to move during the proton translocation. After the first proton has been transported into the lumen side, VrH⁺-PPase requires another proton into the ion gate for the next proton translocation. Specifically, upon PPi hydrolysis, D294 accepts a hydrolysis-generated proton, triggering E301 rebound back to the ion gate. Following the E301 rebound, D294 might transfer the proton to E301. Consequently, E301 can act as a molecular switch at the hydrophobic gate, controlling the "on" or "off" state of proton translocation upon PPi binding and PPi hydrolysis.

R242 prevents reprotonation of D294

Previous mutagenesis studies of VrH⁺-PPase demonstrated the significance of R242 in PPi hydrolysis and proton translocation (10). Hsiao et al. (48), Lin et al. (10), and Huang et al. (49) have proposed that R242 participates in the Grotthuss-chain mechanism, serving as the initial proton acceptor in VrH⁺-PPase. However, when arginine is partially embedded within protein structures or membranes, it remains protonated at the guanidino group (50,51,52). Some channel proteins have the mechanism of assisting proton translocation by arginine. For example, in P-type ATPase, there exists a positive-negative amino acid pair of R655 and D684. R655 points upward toward D684, creating a positively charged plug that prevents extracellular protons from reprotonation D684 (53). Additionally, in the MD simulation of light-driven proton pump bacteriorhodopsin, R82 is suggested to act as a barrier, separating two water clusters and stabilizing protonated waters and change of the arginine hydrogen bonding by a deprotonated water cluster (54).

Considering our observation of two conformations of R242 upon PPi-bound and PPi-hydrolyzed states (Fig. 6), we propose that the role of R242 may involve acting as a positive plug specifically during the PPi-hydrolyzed state, after the proton has been transferred to D294. In this state, R242 likely functions similarly to R655 in P-type ATPase, preventing proton reflux and blocking extracellular protons from reprotonating D294.

The consistent hydrophilic environment at the hydrophobic gate in H⁺-PPases

Among H⁺-PPases, the conserved functional region includes the ion gate, which has the uniform composition of “R-D-K-E,” and the hydrophobic gate that exhibits sequence diversity (10,12). Based on the phylogenetic tree of M-PPases (15), which identified four types of H⁺-PPases, we selected six representative species of H⁺-PPases for comparison. These include plant type H⁺-PPases such as VrH⁺-PPase, Arabidopsis thaliana H⁺-PPase (AtH⁺-PPase), Oryza sativa japonica group H⁺-PPase (OsH⁺-PPase), Pyrobaculum aerophilum H⁺-PPase (PaH⁺-PPase) from K⁺-independent H⁺-PPase, Carboxydothermus hydrogenoformans H⁺-PPase (ChH⁺-PPase), and Flavobacterium johnsoniae H⁺-PPase (FjH⁺-PPase). To elucidate the structural variability and conservation among different H⁺-PPase types, particularly focusing on the ion gate and hydrophobic gate regions, we compared the structural differences between the crystal structure of VrH⁺-PPase (PDB: 4A01) (10) and PaH⁺-PPase (PDB: 8B37) (13) and the AlphaFold models of AtH⁺-PPase, OsH⁺-PPase, ChH⁺-PPase, and FjH⁺-PPase (Fig. 7).

Figure 7.

The structural comparison of the semiconserved glutamate and composition of the hydrophobic gate in H⁺-PPases. H⁺-PPases explore the three different positions of glutamate in the ion gate and indicate the hydrophilic residue present at the hydrophobic gate in (A) VrH⁺-PPase in light green (PDB: 4A01), (B) AtH⁺-PPase in green (UniProt: P31414, AlphaFold), (C) OsH⁺-PPase in dark green (UniProt: Q75U53, AlphaFold), (D) PaH⁺-PPase in brown (PDB: 8B37), (E) ChH⁺-PPase in yellow (UniProt: Q3AFC6, AlphaFold), and (F) FjH⁺-PPase in purple (UniProt: A5FJW6, AlphaFold). The structure shows the region of ion gate, hydrophobic gate, and related TMs (TM5-6,12,16).

In VrH⁺-PPase, AtH⁺-PPase, and OsH⁺-PPase, the hydrophobic gate is highly conserved with the sequence "L-A-L-V," and their glutamates within the ion gate are positioned near the hydrophobic gate (Fig. 7, A–C). In contrast, in PaH⁺-PPase, ChH⁺-PPase, and FjH⁺-PPase, the glutamates within the ion gate are positioned one helical turn above the hydrophobic gate (Fig. 7, D–F). Interestingly, these H⁺-PPases exhibit a hydrophilic residue around the hydrophobic gate. Specifically, PaH⁺-PPase and ChH⁺-PPase have the sequence "F-A-L-T" in the hydrophobic gate, where T695 and T675 are hydrophilic residues, respectively (Fig. 7, D and E). FjH⁺-PPase has the sequence "L-A-L-L" in the hydrophobic gate, with a hydrophilic residue, S181, located nearby (Fig. 7 F).

Based on our finding that E301 is able to penetrate the hydrophobic gate in VrH⁺-PPase (Fig. 3 B), we propose that plant H⁺-PPases utilize their glutamate to penetrate the hydrophobic gate, creating a hydrophilic environment equivalent to T695 in PaH⁺-PPase, T675 in Ch H⁺-PPase, and S181 in FjH⁺-PPase. This comparison suggests that H⁺-PPases have a consistent hydrophilic environment at the hydrophobic gate but highlights the different roles of glutamate among plant type H⁺-PPases. The plant type H⁺-PPases lack an intrinsic hydrophilic residue at the hydrophobic gate, relying on glutamate penetration to assist in proton translocation. Conversely, PaH⁺-PPase, ChH⁺-PPase, and FjH⁺-PPase have inherent hydrophilic residues at the hydrophobic gate, obviating the need for additional hydrophilic contributions from glutamate. This may explain the evolutionary differences in the positioning of glutamate residues. Therefore, we strongly support the necessity for the penetration of glutamate to temporarily create an accessible pathway for protons during the proton translocation in plant H⁺-PPases.

The detailed mechanism of proton translocation in plant H⁺-PPases

Three mechanisms have been proposed for M-PPase, including the Grotthuss-chain mechanism, which outlines the proton translocation pathway (10,12,16); a binding change mechanism, which explains the proton/sodium transfer before PPi hydrolysis (11,12,16); and a direct-coupling mechanism, which posits that the hydrolysis-generated proton is directly transported (12,15). Our study sought to integrate features from these hypotheses for M-PPases (10,12,15,16) to propose a more comprehensive mechanism for the proton translocation in plant H⁺-PPases, including VrH⁺-PPase, AtH⁺-PPase, and OsH⁺-PPase. Our findings provide structural insights into the conformational changes at the ion gate during PPi binding and hydrolysis in VrH⁺-PPase (Figs. 3 and 5). According to the working model of VrH⁺-PPase proposed by Lin et al. (10), we focus on the protonated or deprotonated state at D294 and K742, as well as the conformational dynamics of E301, particularly its penetration and rebound at the hydrophobic gate (Fig. 8).

Figure 8.

The E301 switch mechanism during proton translocation in VrH⁺-PPase. (A) In the ligand-free state, the first proton is retained at K742. (B) Upon PPi binding (PPi-bound state), K742 releases the first proton, inducing TM6 to move downward and allowing E301 to penetrate the hydrophobic gate, thereby opening the proton translocation pathway. (C) Following PPi hydrolyzed to 2Pi (PPi-hydrolyzed state), the second proton is transferred to D294. (D) Protonated D294 triggers E301 to rebound, while R242 turns toward the cytosolic side. (E) After the E301 rebound, the second proton transfer to E301, which subsequently reprotonates K742 after 2Pi release. Finally, VrH⁺-PPase returns to the ligand-free state. The residues involved in the proton translocation are shown and labeled. PPi and 2Pi are shown as spheres and colored orange. The proton is labeled and shown as spheres and colored red.

In a ligand-free state (Fig. 8 A), VrH⁺-PPase retains two Mg²⁺ ions for enzyme function, and D294 remains deprotonated, while K742 remains protonated, which stores the first proton involved in proton translocation. Upon PPi binding, VrH⁺-PPase transitions to the PPi-bound state (Fig. 8 B), leading to the deprotonation of K742 and release of the first proton. The disruption of the ion pair between D294 and K742 induces a 1.5 Å downward movement of E301 in TM6, allowing E301 to penetrate the hydrophobic gate. The penetration of E301 opens the hydrophobic gate and creates a hydrophilic environment within the hydrophobic gate, enabling the first proton to be pumped through the hydrophobic gate and exit channel toward the lumen side. After PPi hydrolysis, VrH⁺-PPase proceeds to the PPi-hydrolyzed state. The hydrolysis-generated proton, serving as the second proton, is transferred to D294 (Fig. 8 C). Protonation of D294 induces R242 to turn toward the cytosolic side and reestablish a hydrogen bond with K742, while E301 rebounds back to the ion gate and closes the hydrophobic gate (Fig. 8 D). The second proton is then transferred from D294 to E301, during which K742 remains in a deprotonated state (Fig. 8 E). Following product release, the conformational change in VrH⁺-PPase promotes the formation of a hydrogen bond between E301 and K742, enabling E301 to reprotonate K742, thereby completing the proton translocation cycle in VrH⁺-PPase (Fig. 8 A).

In summary, we propose that the first proton is released from K742 during PPi binding, allowing E301 to penetrate the hydrophobic gate and facilitate proton transfer. The second proton generated by PPi hydrolysis is transferred to D294, which subsequently induces the rebound of E301. Finally, D294 transfers the second proton to E301 and, upon product release, the resulting conformational change enables E301 to reprotonate K742.

Conclusions

In this study, we employed CpHMD simulation and classical MD simulation to investigate the mechanism of proton translocation within the ion gate of VrH⁺-PPase during PPi binding and PPi hydrolysis. We identified the potential protonation state of the ion gate in three catalytic states, including the ligand-free state, the PPi-bound state, and the PPi-hydrolyzed state. However, the hybrid-solvent CpHMD implementation in AMBER is not rigorous for membrane proteins, since a generalized Born implicit solvent model does not consider the membrane environment. Despite this limitation, the CpHMD simulations provided a valuable starting point to infer plausible protonation states, which we further validated and expanded upon using classical MD simulations. Based on assumptions regarding the proton transfer process, we elucidated the roles and dynamics of R-D-K-E under conditions that proposed proton transfer.

Our findings support a sequential mechanism in which the first proton is released from K742 upon PPi binding, allowing E301 to penetrate the hydrophobic gate and enable proton translocation. The second proton, generated by PPi hydrolysis, is accepted by D294, which subsequently triggers the rebound of E301. R242 likely acts as a positive plug, preventing D294 from being reprotonated prematurely. Finally, D294 rapidly donates the proton to E301 and, upon product release, protonated E301 forms a hydrogen bond with K742, transferring a proton to reprotonate K742. These findings suggest that D294 and K742 are directly involved in proton transfer, R242 serves a regulatory role, and E301 functions both as a dynamic molecular switch and a conditional proton carrier. The proposed mechanism for VrH⁺-PPase is consistent with previous functional data and explains the dynamics of ion gates with “pumping-before-hydrolysis” in the binding change mechanism. Importantly, our proposed mechanism is applicable to plant H⁺-PPases, including AtH⁺-PPase and OsH⁺-PPase. Furthermore, the observed penetration of E301 highlights a consistently hydrophilic environment in H⁺-PPases, offering insights into structural differences across species.

To more accurately capture proton dynamics, future studies may employ advanced methods such as QM/MM and multiscale reactive MD simulations (55,56,57) to elucidate the explicit proton translocation in H⁺-PPase. Nevertheless, insights gained from MD simulations reveal the detailed dynamics of the ion gate, providing a comprehensive understanding of proton translocation in plant H⁺-PPase.

Data and code availability

All data supporting the findings of this study are available from the corresponding author upon reasonable request.

Acknowledgments

This work was supported by the National Science and Technology Council, Taiwan (113-2311-B-007-002 and 113-2311-B-007-007-MY3 to Y.-J.S., 111-2311-B-001-001 and 111-2311-B-001-002 to C.-D.H.), and National Tsing Hua University, Taiwan (113QF009E1 to Y.-J.S.). We thank the bioinformatics core at the Institute of Molecular Biology, Academia Sinica, Taiwan.

Author contributions

Conceptualization, Y.-S.H., J.-Y.T., and Y.-J.S.; data curation, Y.-S.H.; formal analysis, Y.-S.H. and J.-Y.T.; investigation, Y.-S.H.; methodology, Y.-S.H. and J.-Y.T.; software, Y.-S.H.; validation, Y.-S.H.; visualization, Y.-S.H., J.-Y.T., and Y.-J.S.; writing – original draft, Y.-S.H.; writing – review & editing, Y.-S.H., J.-Y.T., S.-M.L., S.-C.W., C.-D.H., and Y.-J.S.; funding acquisition, C.-D.H. and Y.-J.S.; resources, C.-D.H. and Y.-J.S.; supervision, C.-D.H. and Y.-J.S.; project administration, Y.-J.S.

Declaration of interests

The authors declare no competing interests.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used ChatGPT to improve the language and readability of the manuscript. After using this tool/service, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

Editor: Ana Nicoleta Bondar.