Bdelloid rotifers harbor a voltage-gated proton channel with distinct mechanistic features

Abstract

Bdelloid rotifers are renowned for their extraordinary resilience. They can survive desiccation and long-term freezing and revive once favorable conditions return. Such remarkable survival abilities suggest that the evolution of specialized molecular adaptations sustain pH homeostasis under extreme environmental stress. In the present study, we identify and functionally characterize an isoform of the voltage-gated proton channel (H_v) from the bdelloid rotifer Adineta ricciae, which we name ArH_v2. Unlike its paralog ArH_v1 and human H_v1 that activate only at positive membrane potentials and contribute modestly to proton extrusion under normal conditions, the ArH_v2 activates even at negative potentials, enabling rapid and efficient proton efflux or influx when cells face freezing or metabolic arrest. We discover that this behavior is conferred by a positively charged lysine located in the S4 transmembrane segment. This lysine is absent from other known H_v homologs but is highly conserved in bdelloid rotifer H_v2 channels. We further delineated the molecular mechanism by which this lysine fine-tunes ArH_v2 proton channel activation. The discoveries in the present work highlight how the evolution of ion channel architecture at the amino acid level can yield physiological adaptations.

Bdelloid rotifers are renowned for their extraordinary resilience. Here, authors identified a voltage-gated proton channel from bdelloid rotifer, which activates at negative potentials via a conserved lysine, enabling extreme stress pH homeostasis.

Introduction

The voltage-gated proton channel (H_v) is known to play a key role in the regulation of cellular pH^1–5. It mediates the extrusion of protons, preventing intracellular acidification that arises during metabolic activity. H_v channel dysfunction disrupts pH homeostasis and has been linked to diseases^5–12. Structurally, H_v channels are organized as dimers, with each subunit comprising a proton-permeable voltage-sensing domain (VSD) formed by four transmembrane segments (S1–S4), and the two VSDs are held through a C-terminal coiled-coil domain^13–17. The H_v channel S4 helix contains three positively charged residues, typically arginines, that drive voltage-dependent conformational changes essential for channel gating.

Bdelloid rotifers are microscopic organisms renowned for their ability to survive extreme environmental stress^18–21. A study has reported a living bdelloid rotifer from 24,000-year-old Arctic permafrost¹⁸, and the rotifer withstood slow cooling and freezing for a long time and revived when favorable conditions returned. Such extraordinary survival ability suggests that rotifers have evolved unique molecular adaptations that maintain cellular pH homeostasis and sustain metabolism under extreme conditions lethal to most organisms^22,23. In particular, the ability to preserve intracellular pH balance is essential for protecting protein stability, enzymatic activity, and redox homeostasis during prolonged metabolic arrest. These adaptations involve specialized proton transport mechanisms that enable rapid and energy-efficient adjustment of intracellular proton concentrations in bdelloid rotifers and thereby safeguard cellular integrity in the face of desiccation, freezing, and other environmental challenges.

In this work, we identify two H_v isoforms in the bdelloid rotifer Adineta ricciae²⁴, named ArH_v1 and ArH_v2. While the ArH_v1 channel resembles the properties of human H_v1 (hH_v1), the ArH_v2 channel displays different gating properties and activates at more negative potentials, enabling more efficient proton transport near-resting membrane potential. Notably, ArH_v2 contains a positively charged lysine residue in the S4 segment that modulates channel activation. We investigated how this residue and its interactions with conserved acidic residues within the rotifer H_v2 family contribute to ArH_v2 activation and discuss how it regulates this H_v isoform’s function.

Result

Bdelloid rotifers expressed an isoform of the H_v channel

Using the human H_v1 channel (hH_v1) sequence as a query in Protein BLAST, we discovered H_v channel paralogs in bdelloid rotifers including A. ricciae, A. steineri, D. carnosus, R. sordida, and R. magnacalcarata. We performed phylogenetic analysis (Fig. 1a; Table S1), and the results revealed two groups of H_v in bdelloid rotifers; we named them the H_v1 and H_v2 groups, respectively. In this study, the term H_v2 refers to the second H_v isoform identified in rotifers. This nomenclature is consistent with previous studies in H_v proton channels, in which multiple H_v isoforms within the same organism are named sequentially to distinguish coexisting paralogs^25,26. We found that the H_v1 group in rotifers is closely related to H_v channels from other organisms^1,2,25–32, whereas the rotifers H_v2 group shows a distinct lineage, separate from all known H_v homologs identified so far (Fig. 1a).

Fig. 1. H_v channels in bdelloid rotifers.

a Phylogenetic tree places H_v channels of bdelloid rotifers into two groups. The H_v1 group (bond and black) shows a close phylogenetic relationship with homologs from other species, while the H_v2 group (bond and blue) is evolutionarily distinct from other species' H_v homologs. Organisms from bdelloid rotifers include representatives from A. ricciae, D. carnosus, A. steineri, R. sordida, and R. magnacalcarata. Tree scale 0.2 = 20% difference between sequences. The bdelloid rotifer A. ricciae H_v sequences (red arrows, ArH_v1 and ArH_v2) are studied. b View of the bdelloid rotifer A. ricciae. Scale bar represents 50 μm. The image is representative of three images of bdelloid rotifers A. ricciae. c Predicted topology of ArH_v1 and ArH_v2 channels. ArH_v2 shares a similar topology to ArH_v1 and contains four transmembrane domains (S1–S4), an N-terminal S0 domain, and a C-terminal coiled-coil domain (CCD).

We selected one candidate (A. ricciae) from bdelloid rotifers for further investigation (Fig. 1b). H_v paralogs from A. ricciae (ArH_v1 and ArH_v2) have predicted topological features characteristic of other voltage-gated H_v proton channels, including four transmembrane segments forming the voltage-sensing domain (VSD) and a C-terminal coiled-coil domain (CCD) (Fig. 1c; Fig. S1). We investigated the expression of the H_v isoforms in the A. ricciae bdelloid rotifer. Schematic diagrams of A. ricciae H_v1 (ArH_v1) and H_v2 (ArH_v2) genes are shown with primer pair sets designed for each paralog (p1 pair and p2 pair for ArH_v1; p3 pair and p4 pair for ArH_v2) (Supplementary Fig. 2a; Supplementary Fig. 3). PCR amplification was performed using these primer pairs on two independent A. ricciae DNA samples. The results verify that both ArH_v1 and ArH_v2 are present in A. ricciae and can be specifically amplified using paralog-specific primers (Supplementary Fig. 2b).

The ArH_v2 channel exhibited different activities

We next characterized the electrophysiological properties of rotifer H_v channels ArH_v1 and ArH_v2. The ArH_v1 or ArH_v2 genes were fused to pIRES-GFP constructs and expressed in HEK293 cells. Whole-cell configuration recording was performed with intra- and extracellular solutions buffered at pH 7.0. We successfully recorded strong voltage-dependent currents from both ArH_v1 and ArH_v2 under conditions established for other organism H_v channels (Fig. 2). Notably, ArH_v2 displayed markedly different activation ranges and kinetic properties compared to ArH_v1. While ArH_v1 activates slowly and only at positive membrane potentials (Fig. 2a), ArH_v2 activates more rapidly and at both positive and negative membrane potentials (Fig. 2c). ArH_v2 shifted the G–V curve to the negative direction with faster activation kinetics (Fig. 2e, f). The V_1/2 of ArH_v1 is 89.3 ± 3.3 mV, while ArH_v2 is −23.3 ± 2.8 mV.

Fig. 2. Functional expressions of ArH_v1 and ArH_v2 channels.

Representative currents were recorded in HEK293 cells expressing ArH_v1 (a, b) and ArH_v2 (c, d). pH_i = pH_o = 7.0. The corresponding pulse protocols are shown below the current traces. Currents after depolarization in the prestep to +100 mV (for ArH_v1) or +50 mV (for ArH_v2) were used to correct for current rundown. For clarity, only the first and last traces elicited by the depolarization prestep are shown. For ArH_v1, currents were recorded from a holding potential of −40 mV, with test potentials ranging from +180 to 0 mV in 10 mV decrements (a). For ArH_v2, currents were recorded from a holding potential of −80 mV, with test potentials ranging from +70 to −100 mV in 10 mV decrements (c). The red dotted line boxes represent the zoomed area. (b) & (d) show that GFP-co-expressed ArH_v1 (b) or GFP-co-expressed ArH_v2 (d) transfected cells in transmission light and fluorescence were patched; the scale bar is 50 μm. e Conductance (G) versus voltage (V) relationship (G–V) of ArH_v1 and ArH_v2 calculated from current traces like those shown in (a, c). Curves are Boltzmann fits. For ArH_v1, V_1/2 is 89.3 ± 3.3 mV (n = 7); for ArH_v2, V_1/2 is −23.3 ± 2.8 mV (n = 10). n represents the number of independent biological measurements. f The channel opening time constant Ʈ_act in ArH_v1 (n = 4) and ArH_v2 channels (n = 6). Ʈ_act was obtained from exponential fits to rising current like those shown in (a, c). The currents either at small voltages near 0 mV in ArH_v2 or at voltages less than +60 mV in ArH_v1 are too small to reliably determine time constants; therefore, these values are left blank. Data are represented as mean ± SEM.

H_v channels are known for their almost perfect selectivity for protons. Since the rotifer H_v2 isoform is evolutionarily distinct from its H_v1 isoform and other animal H_v homologs, we wonder whether the rotifer H_v2 isoform ArH_v2 keeps this property. We measured its reversal potentials (V_rev) under different transmembrane pH gradients. V_rev values for ArH_v2-mediated currents were obtained using a tail current protocol and compared with the theoretical equilibrium potentials for protons (E_H) predicted by the Nernst equation. Under the range of tested pH gradient conditions, the observed V_rev closely matched the corresponding E_H values, indicating that the ionic flux through ArH_v2 is dominated by protons (Fig. 3).

Fig. 3. ArH_v2 channel is proton selective.

Representative measurements of reversal potentials (V_rev) for ArH_v2-mediated current in the presence of a transmembrane pH gradient (ΔpH) of −1 (a) or 1 (b). Currents were measured at the indicated voltages following a depolarization step to +100 mV (a) or +10 mV (b). Red arrows in insets indicate 0 pA. c V_rev as a function of ΔpH. Data points represent the mean ± SEM of biologically independent replicates (n = 6, 6, and 5 for ΔpH = −1, 0, and 1, respectively). The dashed line indicates the Nernst potential for protons (E_H; slope −59 mV/pH unit). Slope of linear fit of ArH_v2 is -56 ± 5 mV/pH unit. ΔpH values of −1, 0, and 1 correspond to the pH_o/pH_i pairs of 6.0/7.0, 7.0/7.0, and 7.0/6.0, respectively.

Prolonged cold stress is known to enhance anaerobic metabolism, promoting intracellular acidification that is harmful for cellular function. An efficient mechanism enabling H⁺ extrusion near-resting membrane potential would therefore promote survival. We examined ArH_v2 function under defined pH gradients and found that, under intracellular acid load (pH_i 6.0 / pH_o 7.0), ArH_v2 mediates robust outward H⁺ currents at 0 mV (Fig. 4a). Reversing the pH gradient correspondingly reversed the driving force and current direction (Fig. 4b). These results indicate that ArH_v2 can effectively mediate proton transport across the transmembrane gradient at 0 mV. The intracellular acidification combined with an activation range that includes 0 mV would permit sustained H⁺ extrusion via ArH_v2 during cold exposure, helping stabilize intracellular pH and maintain cellular function in extreme environments.

Fig. 4. ArH_v2-mediated proton currents under pH gradients.

a Representative measurements of ArH_v2-mediated current in the presence of pH_i 6.0 / pH_o 7.0. The currents were recorded from a holding potential of −80 mV, with test potentials ranging from +80 to −80 mV in 10 mV decrements. The corresponding pulse protocols are shown below the current traces. Red trace indicates proton current recorded at 0 mV. For clarity, only the trace recorded at 0 mV is shown in the right panel. b Representative measurements of ArH_v2-mediated current in the presence of pH_i 7.0 / pH_o 6.0. The currents were recorded from a holding potential of -80 mV, with test potentials ranging from +80 to −80 mV in 10 mV decrements. The corresponding pulse protocols are shown below the current traces. The red trace indicates proton current recorded at 0 mV. For clarity, only the trace recorded at 0 mV is shown in the right panel.

The lysine (K243) in the S4 segment of ArH_v2 regulates channel voltage-dependent activation

In voltage-gated ion channels, the S4 transmembrane segment of the VSD contains several positively charged residues, mostly arginines, located at every third position³³. The S4 positively charged residues play a central role in channel activation^34–38. They not only control the voltage sensitivity of the channel but also shape the kinetics of activation. Since ArH_v2 exhibited differential activation ranges and kinetic properties (Fig. 2), we analyzed the S4 segment of H_v channels. A lysine (K) in ArH_v2 was shown to take the second position of S4 arginine (R) in ArH_v1 and other H_v channels (Fig. 5a). Notably, this lysine is highly conserved in the rotifer H_v2 family (Supplementary Fig. 4), indicating that it might contribute to a specific role in modifying rotifer H_v2 channel activities.

Fig. 5. The lysine at the second position (K243) in the S4 segment of ArH_v2 regulates channel voltage-dependent activation.

a. Sequence alignment of the S4 segment of H_v channels from human (hH_v1), mouse (mH_v1), sea squirt C. intestinalis (CiH_v1), insect N. phytophila (NpH_v1), sea slug A. californica (AcH_v1, AcH_v2), coccolithophore E. huxleyi (EhH_v1, EhH_v2), and bdelloid rotifer A. ricciae (ArH_v1, ArH_v2). The numbers show three arginines (R1, R2, and R3) in the S4 segment of the H_v channel family. In the ArH_v2 channel, a lysine (K) in the second position of positive S4 residues is shown in bold font and highlighted. b Representative currents from the ArH_v2-K243R mutation. The currents were recorded from a holding potential of -60 mV, with test potentials ranging from +140 to −20 mV in 10 mV decrements. The corresponding pulse protocols are shown below the current traces. pH_i = pH_o = 7.0. c. G–V relationships of ArH_v2 wild-type or ArH_v2-K243R mutation. For ArH_v2, V_1/2 is −23.3 ± 2.8 mV (n = 10), for ArH_v2-K243R, V_1/2 is 55.0 ± 2.6 mV (n = 6). d Representative currents from ArH_v1-R211K mutation. The currents were recorded from a holding potential of −40 mV, with test potentials ranging from +140 to 0 mV in 10 mV decrements. pH_i = pH_o = 7.0. e G–V relationships of ArH_v1 wild-type or ArH_v1-R211K mutation. For ArH_v1, V_1/2 is 89.3 ± 3.3 mV (n = 7); for ArH_v1-R211K, V_1/2 is 58.8 ± 3.7 mV (n = 5). f Representative currents from the hH_v1-R208K mutation. The currents were recorded from a holding potential of -60 mV, with test potentials ranging from +100 to −60 mV in 10 mV decrements. pH_i = pH_o = 7.0. g G–V relationships of human H_v1 wild-type or hH_v1-R208K mutation. For hH_v1, V_1/2 is 60.8 ± 3.3 mV (n = 6); for hH_v1-R208K, V_1/2 is 27.3 ± 3.9 mV (n = 6). Curves in panels (c, e, g) are Boltzmann fits. The lysine (K) at the second position is highlighted in blue, and the arginine (R) at this position is shown in black. n represents the number of independent biological measurements. Data are represented as mean ± SEM.

To determine whether and how this lysine affects the ArH_v2 channel function, we substituted it with an arginine at the second position of S4 in ArH_v2. This charge-conserving replacement (ArH_v2-K243R) dramatically shifted the G–V curve to the right and generated an approximately 80 mV shift in the V_1/2 (Fig. 5b, c). The result suggests that the lysine (K) at the second position of S4 plays a significant role in channel activation and makes the channel much easier to open. Conversely, introducing lysine at the second position of the ArH_v1 channel (ArH_v1-R211K) produced a left-shifted G–V compared to the wild-type ArH_v1 (Fig. 5d, e). The data was consistent with the observation that native ArH_v2 (with K at the second position) favors channel opening. Moreover, we investigated the human H_v1 channel (hH_v1), and similar results were found in hH_v1, in which the lysine substitution at the second position (hH_v1-R208K) notably shifted the G–V curve in the negative direction (Fig. 5f, g). Taken together, these results indicated that the lysine at the second position of S4 regulates the voltage-dependent activation of the H_v channels and plays an essential role in rotifer proton channel ArH_v2 activation and function.

Effects of mutations at position K243 depend on the negative residues in the S1 and S3 transmembrane segments

Since the conserved lysine in rotifer H_v2 controls the channel activity, we set out to determine how this lysine affects the ArH_v2 voltage-dependent activation. There are several negatively charged residues distributed from S1 to S3 segments of the VSD in the voltage-gated ion channels. The function of negatively charged residues was proposed to interact with S4 positively charged residues to mediate the channel activation^39–41. These negatively charged residues are highly conserved in the H_v channel family. However, there are two changes that occurred in ArH_v2 channel: one aspartate (D129 in ArH_v2) takes the position of glutamate (E) in the area near the extracellular side of S1, and conversely, a glutamate (E218 in ArH_v2) takes the position of aspartate (D) in the area near middle of S3 (Fig. 6a). Notably, this “D/E swap” in S1 and S3 is preserved across the rotifers H_v2 family (Supplementary Fig. 4). We propose that the D/E swap of the negatively charged residues in S1/S3 segments of ArH_v2 is associated with the presence of positively charged lysine (K243) the S4, which is also highly conserved in the rotifers H_v2 family (Supplementary Fig. 4).

Fig. 6. Mutant cycle analysis shows that K243 interacts with D129 and E218 during ArH_v2 channel activation.

a Sequence alignment of the transmembrane domains of the H_v channel family. The three identical charged residues in ArH_v2 are highlighted and labeled. Representative currents recorded in HEK293 cells expressing ArH_v2 double mutants K243R-D129E (b), W242A-D129E (c), I244A-D129E (d), K243R-E218D (e), W242A- E218D (f), and I244A- E218D (g). In each double mutant, the currents were recorded from a holding potential indicated, with test potentials ranging from voltages indicated with 10 mV decrements. The corresponding pulse protocol for the individual mutant is shown in the right inset. pH_i = pH_o = 7.0. h–m Effects of ArH_v2 mutations on the G–V relationship of the channels. Voltage-dependent channel activations were shown for D129E with K243R (h), D129E with W242A (i), D129E with I244A (j), E218D with K243R (k), E218D with W242A (l), E218D with I244A (m). In (h–j), the dashed line (black) represented the G–V curve of the WT channel (D129), and the dashed line (red) represented the G–V curve of D129E. In (k–m), the dashed line (black) represented the G–V curve of the WT channel (E218), and the dashed line (red) represented the G–V curve of E218D. Lines in (h–m) indicate fits of the data to a Boltzmann function. Data are represented as mean ± SEM. Details are shown in Table S2. n Summary of |ΔΔG₀| values from double-mutant cycle analysis. Data are represented as mean ± SEM. Residue pairs were considered to have a significant interaction when |ΔΔG₀| > 4.2 kJ/mol (dotted line). The larger values of |ΔΔG₀| indicated interactions between K243 and D219 or E218, see “Methods” section and Table S2.

To investigate whether the S1/S3 negatively charged residues are coupled with the S4 lysine, we applied a mutant cycle approach to probe potential interactions between them. A double-mutant cycle analysis has been used to study residues’ interactions between transmembrane segments in the H_v1 channel⁴². We applied the approaches to investigate the relationship between K243 and D129 or E218 (Fig. 6b–g). The residues near the highly conserved K243 (e.g., W242 and I244) in the S4 segment were investigated as well (Fig. 6b–g, Supplementary Fig. 5).

The G–V curves of the S4 segment mutation (W242, K243, or I244) with or without mutation at D129 or E218 were summarized in Fig. 6h–m. The G–V curve was fitted with the Boltzmann equation, and V_1/2 and k values were determined and used to assess the ΔG₀, which is the free energy difference between the closed and open states of the activation. The thermodynamic coupling energy |ΔΔG₀| between two mutated positions was determined by the thermodynamic coupling analysis⁴³. A significant interaction between two residues was defined as a |ΔΔG₀| > 4.2 kJ/mol (1 kcal/mol)^40,44. The |ΔΔG₀| values obtained were reported in Fig. 6n. It was shown that the combinations K243R-D129E and K243R-E218D produced |ΔΔG₀| values of 6.9 and 5.1 kJ/mol, respectively, supporting the thermodynamic coupling interactions between K243 and D129 or E218 (Fig. 6n, Table S2). In contrast, combinations W242A-D129E and W242A-E218D yielded |ΔΔG₀| values (1.2 and 0.5 kJ/mol, respectively) well below 4.2 kJ/mol, indicating no thermodynamic couplings between W242 and D129 or E218. We further studied another residue, I244, near K243, and the I244A-D129E and I244A-E218D combinations generated much smaller |ΔΔG₀| (1.0 and 0.7 kJ/mol, respectively) when compared with K243R-D129E and K243R-E218D (Fig. 6n, Table S2). Collectively, these data strongly suggest that K243, but not W242 or I244, is coupled to D129 and E218, supporting functional interactions between K243 and D129/E218 during voltage-dependent channel activation.

K243 interacts with E218 to stabilize the closed state of the channel, and K243 interacts with D129 to stabilize the open state of the channel

The activation of VSD involves the outward movement of positively charged amino acids in the S4 segment across the membrane electric field³³. Since the double-mutant cycle analysis indicated that K243 interacts with D129 and E218 during channel activation—a process by which the channel transitions between closed and open states—we propose that interactions between K243 and D129 or E218 play a regulatory role in controlling the conformational transitions of the ArH_v2 VSD.

A charge-conserving mutation at position 129 (D129E) was introduced into the ArH_v2 channel (Fig.7). The D129E mutation shifted the G–V relationship toward more positive direction (Fig. 7g), and exhibited slowed activation (Fig. 7b, h) and accelerated deactivation (Fig. 7e, i), indicating destabilization of the open state of the channel and stabilization of the closed state of the channel. D129E also exhibited less negative ΔG₀ and V_10%max (voltage at 10% of maximum conductance) compared with the WT channel (Supplementary Fig. 6, Table S2). Taken together, these results suggest that the native D129 contributes to stabilizing the open conformation. We further assessed the interactions between D129 and K243 by cysteine pair crosslinking and observed that Cd²⁺ produced sustained tail current and significantly delayed the deactivation process in the double mutant K243C-D129C (Supplementary Fig. 7), indicating that Cd²⁺-mediated coordination locks the channel in the open conformation and impairs channel closing. The data supported that D129 interacts with K243 in the open state of the channel.

Fig. 7. Effects of charge-conserving mutations at D129 or E218 on ArH_v2 channel activation and deactivation kinetics.

Representative rising currents recorded from WT (a), D129E (b), and E218D (c). The currents in (a–c) were measured from a holding potential of −80 mV to test potentials ranging from +80 to −80 mV in 10 mV decrement. pH_i = pH_o = 7.0. Representative tail currents recorded from WT (d), D129E (e), and E218D (f). The tail currents in (d–f) were elicited by a prepulse to +60 mV, in 10 mV decrements from −50 to −80 mV. The dotted line box represents the zoomed area. pH_i = pH_o = 7.0. g The G–V relationships of the WT, D129E, and E218D. Lines indicate fits of the data to a Boltzmann function. For ArH_v2 WT, V_1/2 is −23.3 ± 2.8 mV (n = 10), for D129E, V_1/2 is −9.6 ± 1.4 mV (n = 9), for E218D, V_1/2 is −36.6 ± 3.3 mV (n = 8). n represents the number of independent biological measurements. h The channel opening time constant Ʈ_act in WT, D129E, and E218D. Ʈ_act was obtained from exponential fits to rising current traces like those shown in (a–c). n = 6 for ArH_v2 WT, n = 4 for D129E, n = 5 for E218D. n represents the number of independent biological measurements. i The deactivation (channel closing) time constant Ʈ_deact in WT, D129E, and E218D. Ʈ_deact was obtained from an exponential fit to the tail current traces like those shown in (d–f). n = 6 for ArH_v2 WT, n = 4 for D129E, n = 5 for E218D. n represents the number of independent biological measurements. Examples of exponential fits were shown in Supplementary Fig. 8. Data are represented as mean ± SEM.

In contrast, the E218D mutation shifted the G–V relationship toward more negative potentials (Fig. 7g), accelerated activation (Fig. 7c, h, Supplementary Fig. 8), and markedly delayed deactivation (Fig. 7f, Supplementary Fig. 8), with an increased closing time constant (Ʈ_deact) (Fig. 7i) and more negative ΔG₀ and V_10%max (Supplementary Fig. 6, Table S2). These findings indicate that E218D destabilizes the closed state and stabilizes the open state of the channel. Moreover, the interaction between E218 and K243 was supported by the results of cysteine pair cross-links (Supplementary Fig. 9). In the presence of Cd²⁺, no significant “lock-open” sustained tail currents were observed in the E218C-K243C (Supplementary Fig. 9a), supporting that E218 and K243 are positioned to interact in the resting state. Taken together, these data are consistent with a model where an interaction between E218 and K243 keeps the channel closed and an interaction between D129 and K243 keeps the channel open.

Model of ArH_v2 voltage-sensing domain transitions

To describe the influences of D129 and E218 on the voltage-dependent activation of the ArH_v2 VSD, we proposed a simplified state model describing the conformational transitions of the VSD. This energy-reaction model has been used to describe transitions between functional states of voltage-gated ion channels at a phenomenological level^44,45. The model consists of a closed state and an open state, each represented by an energy well. The transition rates between these states are determined by the height of the energy barrier separating them (Fig. 8).

Fig. 8. Schematic representation of the new isoform of the voltage-gated proton channel ArH_v2 gating.

a The gray lines represent the voltage-dependent activation for the wild-type ArH_v2 channel (WT), and purple lines are changes mediated by the D129E mutation. Destabilization of the open state caused by D129E is achieved by decreasing the relative depth of the energy well for the “Open” state and increasing the energy barrier for the Close -> Open transition to account for the slower activation. The smaller negative ΔG₀ for D129E (−2.2 ± 0.5 Kjmol⁻¹ for ΔG₀D129E vs −5.9 ± 0.8 Kjmol⁻¹ for ΔG₀WT) is satisfied by increased activation kinetics E_actD129E (e.g., at +50 mV, Ʈ_act is 42.3 ± 5.9 ms for D128E vs Ʈ_act is 26.2 ± 3.7 ms for WT) and decreased deactivation kinetics E_deactD129E (e.g., at −50 mV, Ʈ_deact is 15.0 ± 2.7 ms for D128E vs Ʈ_deact is 34.5 ± 4.1 ms for WT). b A similar reaction process can be interpreted for the effects of the E218D mutation on the closed state. Green lines are changes mediated by the E218D mutation. Destabilization of the closed state caused by E218D is consistent with decreasing the relative depth of the energy well for the “Close” state and increasing the energy barrier for the Open -> Close transition to allow for the slower deactivation. And the lager ΔG₀ for E218D (−10.2 ± 0.9 Kjmol^-1 for ΔG₀E218D vs −5.9 ± 0.8 Kjmol-1 for ΔG₀WT) is satisfied by decreased activation kinetics E_actE218D (e.g., at +50 mV, Ʈ_act is 14.2 ± 2.6 ms for E218D vs Ʈ_act is 26.2 ± 3.7 ms for WT) and increased deactivation kinetics E_deactE218D (e.g., at −50 mV, Ʈ_deact is 76.0 ± 7.5 ms for E218D vs Ʈ_deact is 34.5 ± 4.1 ms for WT).

In this model, the effect of the D129E mutation, destabilization of the open state and stabilization of the closed state, was satisfied by decreasing the relative depth of the energy well corresponding to the “Open” state and by increasing the energy barrier for the “Close -> Open” transition to interpret the slower opening (i.e., increased activation kinetics) (purple lines in Fig. 8a). And the smaller negative ΔG₀ value for D129E (Table S2) was consistent with stabilization of closed state and destabilization of open state.

The influence of the E218D mutation on the closed state of the channel was also captured by this model. Destabilization of the closed state and stabilization of the open state by E218D could be satisfied by decreasing the relative depth of the energy well corresponding to the “Close” state and by increasing the energy barrier for the “Open -> Close” transition to account for the slower deactivation observed in Fig. 7f (green lines in Fig. 8b). And the larger negative ΔG₀ value for E218D (Table S2) was consistent with the stabilization of open state and destabilization of closed state of the channel.

Discussion

In the S4 segment of voltage-sensing domains (VSDs), positively charged residues serve as the main charge carriers that drive the stepwise movement of S4 in voltage-gated ion channels³³. Through stable hydrogen bonds and salt bridges with residues in other segments of the VSD, these charges mediate conformational transitions essential for channel activation and serve as the primary molecular mechanism underlying voltage-dependent activation. The positively charged residues, most commonly arginine, play critical roles in channel gating^35–38.

In the bdelloid rotifer A. ricciae, we discovered an isoform of voltage-gated proton channel ArH_v2. The ArH_v2 channel harbors a positively charged lysine (K243) in the second position of the S4 segment. This lysine is highly conserved in the rotifer H_v2 family (Fig. S4) but not present in all other H_v homologs identified so far (Fig. 5a).

Lysine, compared to arginine, has a smaller, more flexible amino group⁴⁶, which better stabilizes interactions with acidic residues and modulates ArH_v2 VSD gating. Substituting lysine at the second position of S4 shifted the G–V curves of ArH_v1, ArH_v2, and hH_v1 toward the negative direction, facilitating channel opening. In rotifers, the native lysine in ArH_v2 likely lowers the energetic barrier for activation, fine-tuning charge-charge interactions during S4 movement and stabilizing specific channel states.

Mutational analyses further clarify these interactions. In the resting state, K243 interacts with E218 (Figs. 9a and 10); the E218D mutation shifted the G–V curve in the negative direction, accelerated activation, and delayed closure, consistent with weakened K243-218 coupling and destabilization of the closed state. In the open state, K243 interacts with D129 (Figs. 9b and 10); the D129E mutation slowed activation and shifted the G–V curve in the positive direction, indicating destabilization of the open conformation. These results highlight a dynamic role of K243, stabilizing distinct channel conformations through context-dependent interactions with conserved acidic residues to fine-tune the channel activity. This function might control ArH_v2 activity to coordinate the regulation of pH homeostasis under various conditions in rotifers.

Fig. 9. Models of AlphaFold-derived voltage-sensing domain (VSD) of the ArH_v2 in the closed state and open state.

a In the closed VSD, K243 interacts with E218. The green lines indicate the interaction between E218 and S4 positively charged residues. The distances between E218 and R240, K243, and R246 are 5.2 Å, 4.5 Å, 10.2 Å, respectively. b In the open VSD, K243 moves upward, interacting with D219. The green lines indicate the interaction between D129 and S4 positively charged residues. The distances between D129 and R240, K243, and R246 are 13.0 Å, 4.4 Å, 9.5 Å, respectively. The models of ArH_v2 were generated by AlphaFold; see “Methods” section.

Fig. 10. Schematic showing the interactions between K243 and negatively charged residues during the ArH_v2 activation.

K243 interacts with E218 to stabilize the closed state of the channel (left panel). When the channel is depolarized, K243 moves upward to interact with D219 to stabilize the open state of the channel (right panel).

A previous study applied mutagenesis to study the role of a lysine/arginine switch in potassium channels⁴⁷. Tao et al. showed that lysine at different positions within the Shaker channel S4 segment can exert opposing effects on channel gating: lysine replacing native arginine at the outer position (R1K) stabilizes the “closed” state; however, lysine at an inner position (K5) stabilizes the “open” state⁴⁷. These results indicate that the functional impact of lysine strongly depends on its position within the S4 segment in the Shaker channel. In the present study, we discovered that the single lysine residue (K243) in the ArH_v2 channel interacts with E218 to keep the channel closed; and upon activation, K243 moves upward to interact with D129 to keep the channel open. These findings reveal a distinct role of lysine in regulating voltage-gated proton channel function.

The presence of lysine in ArH_v2 enables the channel to open at 0 mV. This property can help rotifers survive in harsh environments. During extended low-temperature exposure, cold stress can increase physiological constraints (e.g., promoting anaerobic metabolism) and accumulation of acidic equivalents (e.g., lactate and related metabolites)^48–50. Sustained intracellular acidification is known to disrupt cellular function and induce cell death. ArH_v2 channel is able to generate robust outward H⁺ currents at 0 mV under intracellular acid load, and this function enables rapid and efficient H⁺ extrusion near the resting potential during cold stress, which reduces anaerobic metabolism-induced intracellular acidification and improves survival.

The lysine substitution at the second position of the S4 segment was identified in bdelloid rotifer H_v2 channels. While this appears to represent a molecular adaptation in rotifers, other anhydrobiotic animals may develop diverse or lineage-specific strategies to modify proton channel voltage sensitivity in response to extreme environmental stress.

In contrast to ArH_v2, the rotifer ArH_v1 channel exhibits a more positively shifted G–V relationship. This difference implies that ArH_v2 is the dominant contributor to H⁺ extrusion at resting or near-resting membrane potentials, whereas ArH_v1 is largely inactive under the same conditions. The coexistence of two H_v isoforms in rotifers might allow adaptation to a broader range of voltage or pH conditions over which rotifers can regulate proton flux for fluctuating environments. Future studies are needed to determine how these channels are differentially regulated and whether loss of either proton channel compromises organismal viability.

In summary, our results identify the conserved lysine in the S4 segment of ArH_v2 as a molecular determinant that fine-tunes channel activation through its interactions with negatively charged residues in resting and open states. This distinctive feature sets rotifer H_v2 channels apart from other animal H_v homologs and may underline their ability to regulate proton homeostasis under extreme stress. The discoveries in the present work highlight the connection between molecular specialization at the amino acid level and organism function, illustrating how evolution in ion channel architecture can yield physiological adaptations.

Methods

H_v sequence analysis

The human H_v1 channel (hH_v1) sequence is used as a query in Protein BLAST to identify rotifer H_v homologs⁵¹. For phylogenetic analysis, a multiple H_v sequence alignment and phylogenetic tree were performed using Molecular Evolutionary Genetics Analysis (MEGA12) software⁵². Tree scale 0.2 = 20% difference between sequences.

The following organism H_v protein sequences were used to construct the phylogenetic tree: human (H. sapiens, NP_001035196.1), mouse (M. musculus, NP_001035954.1), zebrafish (D. rerio, NP_001002346.1), coral (A. millepora, XP_029202331.2), snail (H. trivolvis, XP_055871465.1), insect (N. phytophile, AMK01488.1), sea squirt (C. intestinalis, NP_001071937.1), sea slug (A. californica, XP_005100666.1 and XP_005093106.1), coccolithophore (C. braarudii, ADM25825.1 and CAD8615310.1; E. huxleyi, XP_005762299.1 and XP_005794250.1), dinoflagellate (K. veneficum, AEQ59286.1), bdelloid rotifers (A. ricciae, CAF1315436.1 and CAF0979481.1; A. steineri, CAF1145379.1 and CAF1220596.1; D. carnosus, CAF0799904.1 and CAF1276142.1; R. sordida, CAF1201389.1 and CAF1163634.1; R. magnacalcarata, CAF1450926.1 and CAF2039880.1). Detailed information for the protein sequences was summarized in Table S1.

Generation of structural models

Structural models of ArH_v2 were generated by the AlphaFold (https://alphafold.ebi.ac.uk/)⁵³. The input sequence comprises the full-length ArH_v2 protein (i.e., residues 1–325), and the PDB file of ArH_v2 is included in the Supplementary Information. For the model of the closed state, the structure of the isolated VSD was applied (Fig. 9a). For the generation of the model of the open state, the “one-click” model was applied to modify the closed state. In this model, the S4 helix undergoes a single upward step corresponding to one register shift of the charged residues along the helix. Accordingly, the position originally occupied by residue 237 was reassigned to Arg240, Arg240 was reassigned to Lys243, and Lys243 was reassigned to Arg246 (Fig. 9b). Models were visualized, and inter-residue distances were measured using BIOVIA Discovery Studio.

PCR expression analysis

Cultures of Adineta ricciae were maintained in plastic Petri dishes (3–15 cm diameter) with 3–50 ml sterile ddH₂O (depending on the vessel) and fed twice a week with a solution of autoclaved yeast extract and peptone (2.5% w/v each)⁵⁴. Rotifers were harvested by centrifugation at 4000 × g for 6 min at 4 °C, the supernatant removed and the pellet used immediately or frozen at −20 °C until DNA extraction. DNA was extracted with a Qiagen Blood and Tissue Kit (cat. 69504) following the “Tissue” protocol, and the DNA eluted in 20−40 µl of buffer AE. Purification and concentration were measured with a Nanodrop. The DNA concentrations ranged from 20 to 100 ng/µl. The resulting DNA was used as a template for PCR analysis under the following conditions: initial denaturation at 98 °C for 1.5 min, followed by 35 cycles of 98 °C for 20 s, 63 °C for 30 s, and 73 °C for 5 min. A final extension was performed at 72 °C for 15 min, and reactions were cooled to 4 °C for 30 min. The resulting PCR products were analyzed on a 2% agarose gel.

Mutagenesis

DNA constructs encoding wild-type ArH_v1 and ArH_v2 were synthesized by Integrated DNA Technologies (IDT). The coding sequences were subcloned into the pIRES-GFP vector (for HEK293 cells)⁵⁵ or pGEMHE vector (for Xenopus laevis oocytes)⁵⁶ using BamHI/XbaI restriction sites. The construct containing the cDNA sequence from human H_v1 in the pNICE vector was used to express hH_v1. Single-point mutations were introduced with standard PCR techniques. PCR primers were purchased from IDT DNA Technologies. Custom-designed primers for mutagenesis of H_v constructs were listed in Table S3. Mutations were generated by amplifying plasmid templates under the following PCR conditions: initial denaturation at 98 °C for 1.5 min, followed by 25 cycles of 98 °C for 20 s, 60 °C for 30 s, and 72 °C for 5 min. A final extension was performed at 72 °C for 15 min, and reactions were cooled to 4 °C for 30 min. The template plasmid was digested with DpnI (NEB, R0176S), and the resulting PCR products were transformed into Stbl2 competent cells (Invitrogen, 10268019). Plasmids were isolated from bacterial colonies using the QIAprep Spin Miniprep Kit (Qiagen, 27106). All mutations were confirmed by Sanger sequencing before use in subsequent transfection experiments.

Cell culture and H_v transfection

HEK293 cells (Sigma-Aldrich) were cultured on coverslips in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C in 5% CO₂. Upon reaching ~60% confluence, cells were transiently transfected with 2 μg of rotifer H_v cDNA plasmid (pIRES-GFP) using Lipofectamine 3000 reagent (Invitrogen), following the manufacturer’s protocol. For human H_v1, cells were transiently transfected with 2 μg of hH_v1 cDNA plasmid (pNICE) and 0.25 μg of a plasmid encoding GFP. After transfection, cells were incubated for 24 h at 37 °C before electrophysiological studies. For recordings, coverslips containing HEK293 cells were placed in a bath solution within a recording chamber mounted on an Olympus fluorescence microscope. Transfected cells were identified by GFP fluorescence, and whole-cell patch-clamp experiments were performed 1–3 days post-transfection. Xenopus laevis oocytes (Stages V-VI) were obtained from Ecocyte Bioscience (USA) and maintained at 18 °C in ND96 solution containing 96 mM NaCl, 2 mM KCl, 1.8 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, 5 mM pyruvate, and 100 mg/mL gentamicin (pH 7.2). pGEMHE plasmids encoding ArH_v2 constructs were linearized with NheI (New England Biolabs, USA), and RNA synthesis was carried out with a T7 Transcription Kit (Ambion, USA). Oocytes were injected with 50 nL of cRNA (1.5 mg/mL) and incubated for 3–5 days prior to electrophysiological experiments.

Patch-clamp electrophysiology

H_v proton currents in HEK293 cells were recorded in the whole-cell configuration using an Axopatch 200B amplifier (Molecular Devices) controlled by pClamp 11 software via an Axon Digidata 1550B digitizer. The external bath solution at pH 7.0 contained 75 mM N-methyl-D-glucamine (NMDG), 180 mM N-(2-Hydroxyethyl) piperazine-N′-(2-ethanesulfonic acid) (HEPES), 10 mM glucose, 1 mM MgCl₂, and 1 mM CaCl₂, adjusted to pH 7.0 with methanesulfonic acid. The pipette solution at pH 7.0 consisted of 75 mM NMDG, 180 mM HEPES, 10 mM glucose, 1 mM MgCl₂, and 1 mM EGTA, adjusted to pH 7.0 with methanesulfonic acid. The external bath solution at pH 6.0 contained 75 mM NMDG, 110 mM 2-(N-morpholino) ethanesulfonic acid (MES), 80 mM glucose, 1 mM MgCl₂, and 1 mM CaCl₂, adjusted to pH 6.0 with methanesulfonic acid. The pipette solution at pH 6.0 consisted of 75 mM NMDG, 110 mM MES, 80 mM glucose, 1 mM MgCl₂, and 1 mM EGTA, adjusted to pH 6.0 with methanesulfonic acid. H_v proton currents in oocytes were recorded using the inside-out patch-clamp configuration. Both bath and pipette solutions contained 100 mM MES, 30 mM tetraethylammonium (TEA) methanesulfonate, 5 mM TEA chloride, and 5 mM EGTA, with the pH adjusted to 6.0 using TEA hydroxide. Recordings were performed under constant perfusion at 22 ± 3 °C using pipettes with access resistance of 5–10 MΩ. Currents were low-pass filtered at 1 kHz and analyzed with Clampfit 11 (Molecular Devices) and Origin 2019 (OriginLab).

Data analysis

The proton channel conductance is determined by the equation⁵⁷:

$$G\left(Vtest\right)=\left(I_{test}-I_{tail}\right)/\left(V_{test}-V_{tail}\right)$$ 1

where I_test is the current at V_test, measured at the end of the depolarization step, and I_tail is the current at V_tail, measured at the beginning of the repolarization step. Currents after depolarization in the prestep were used to correct for current rundown.

Volage dependent activation G–V curves were fitted by the Boltzmann equation:

$$G/G_{\max }=1/\left(1+\exp \left(V1/2-V\right)/k\right)$$ 2

where G/G_max is the conductance normalized by the maximal conductance, V_1/2 is the half activation potential, V is the test voltage, and k is the slope factor, equal to RT/ɀF,

3

where ɀ is the equivalent charge, R is the gas constant, F is Faraday’s constant, and T is temperature in Kelvin.

Values of V_1/2 and k obtained from the Boltzmann fits of multiple cells were used to calculate the free energy difference at 0 mV (ΔG₀) between the closed and open states of the activation. ΔG₀ was calculated according to the following:

4

ΔG₀ is referred to as the chemical potential difference between the open and closed states in the absence of an electric field. At 0 mV, the voltage-dependent term in the Boltzmann description of gating is zero, so the calculated ΔG reflects the intrinsic energetic preference of the channel (i.e., open vs. closed) instead of the influence of membrane potential. It is a standard parameter for quantifying how mutations shift the baseline stability of gating states⁴³.

Double-mutant cycle analysis was performed to determine whether the effects of two mutations were additive or whether functional interactions existed between the mutation sites. The |ΔΔG₀| (i.e., coupling ΔG₀ value) was calculated according to the following⁴³:

$$\mid \Delta \Delta G_0\mid =\mid \left(\Delta G_0^i-\Delta G_0^{WT}\right)-\left(\Delta G_0^{ij}-\Delta G_0^j\right)\mid$$ 5

where i is the mutation 1, j is the mutation 2. ij is the double mutations at both positions, WT is the wild-type channel.

The standard error of the mean (SEM) value for |ΔΔG₀| was determined according to the following^40,44:

$$\mathrm{SEM}\mathrm{of}\mid \Delta \Delta G_0\mid =\surd \left({\left(\mathrm{SEM}\mathrm{of}\Delta G_0^i\right)}^2+{\left(\mathrm{SEM}\mathrm{of}\Delta G_0^{WT}\right)}^2+{\left(\mathrm{SEM}\mathrm{of}\Delta G_0^{ij}\right)}^2+{\left(\mathrm{SEM}\mathrm{of}\Delta G_0^j\right)}^2\right)$$ 6

The large coupling free energy |ΔΔG₀| values indicate functional interactions between positions i and j, and a coupling energy of greater than 4.2 kJ/mol represents a significant interaction between two positions^40,44.

The channel opening time constant Ʈ_act and closing time constant Ʈ_deact values were calculated by fitting current traces with the single-exponential equation according to the following:

7

where I(t) represents the current at time point t, I₀ is the initial current amplitude, Ʈ is the time constant.

All data were presented as mean ± SEM. Significance between means was determined by Student’s t-test. p < 0.05 was considered to indicate a statistically significant difference.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary Information ^{(1.9MB, pdf)}

Reporting Summary ^{(94.9KB, pdf)}

Transparent Peer Review file ^{(766.6KB, pdf)}

Source data

Source Data ^{(270.1KB, zip)}

Author contributions

L.H. conceived the project and designed the experiments. L.Y. and C.B. performed experiments. L.Y., C.B., and L.H. analyzed data. L.Y. and L.H. wrote the manuscript. All authors reviewed and edited the manuscript.

Peer review

Peer review information

Nature Communications thanks Boris Musset and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This work was supported in part by the NIH Grants R01GM139991 (L.H.) and R01HL174544 (L.H.).

Data availability

All data supporting the discoveries of the study are available within the paper and its Supplementary Information. The source data underlying Figs. 2, 3, 5, 6, and 7, and Supplementary Figs. S2, S6, S7, and S9 are provided as a Source Data file. The PDB file of the AlphaFold-derived ArH_v2 structure is included in the Supplementary Information. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-76314-9.