Zinc activation of OTOP proton channels identifies structural elements of the gating apparatus
Abstract
Otopetrin proteins (OTOPs) form proton-selective ion channels that are expressed in diverse cell types where they mediate detection of acids or regulation of pH. In vertebrates there are three family members: OTOP1 is required for formation of otoconia in the vestibular system and it forms the receptor for sour taste, while the functions of OTOP2 and OTOP3 are not yet known. Importantly, the gating mechanisms of any of the OTOP channels are not well understood. Here, we show that zinc (Zn2+), as well as other transition metals including copper (Cu2+), potently activates murine OTOP3 (mOTOP3). Zn2+ pre-exposure increases the magnitude of mOTOP3 currents to a subsequent acid stimulus by as much as 10-fold. In contrast, mOTOP2 currents are insensitive to activation by Zn2+. Swapping the extracellular tm 11–12 linker between mOTOP3 and mOTOP2 was sufficient to eliminate Zn2+ activation of mOTOP3 and confer Zn2+ activation on mOTOP2. Mutation to alanine of H531 and E535 within the tm 11–12 linker and H234 and E238 within the 5–6 linker reduced or eliminated activation of mOTOP3 by Zn2+, indicating that these residues likely contribute to the Zn2+ activating site. Kinetic modeling of the data is consistent with Zn2+ stabilizing the opn2+en state of the channel, competing with H+ for activation of the channels. These results establish the tm 11–12 and tm 5–6 linkers as part of the gating apparatus of OTOP channels and a target for drug discovery. Zn2+ is an essential micronutrient and its activation of OTOP channels will undoubtedly have important physiological sequelae.
Editor's evaluation
This valuable study discovers that zinc ions can activate some OTOP proton channels, identifying a pharmacological tool for research, and further establishing that OTOP channels gate. The data presented provides convincing support for conclusions. This work is expected to be of great interest to physiologists studying OTOP channels and other proton-permeation pathways.
https://doi.org/10.7554/eLife.85317.sa0Introduction
Pharmacological agents that can activate or inhibit ion channels have long been used as probes to describe the fundamental processes of channel gating and ion permeation (Hille, 2001). For example, the discovery of the charged molecule TEA and the scorpion toxin charybdotoxin as a specific blocker of K+ channels allowed for the early identification of residues lining the channel pore well before the channel structures were determined (MacKinnon et al., 1990; Yellen et al., 1991; Banerjee et al., 2013). Similarly, gating modifiers have been used to probe structural rearrangements that accompany the opening of voltage-gated ion channels (Swartz and MacKinnon, 1997; Sack and Aldrich, 2006; Catterall et al., 2007; Goldschen-Ohm and Chanda, 2014). More recently, toxins that target pain-sensing ASIC and TRPV1 channels have been used to probe the conformational states of these channels (Bohlen et al., 2010; Baconguis et al., 2014). One of the most common modulators of channel activity is the trace metal zinc (Zn2+), which can affect gating, permeation, or both (Gilly and Armstrong, 1982; Chu et al., 2004; Noh et al., 2015; Peralta and Huidobro-Toro, 2016). Zn2+ binds to proteins with high affinity and specificity and regulates a wide range of cellular processes, including metabolism and gene expression (Vallee and Falchuk, 1993). Zn2+ is a potent inhibitor of proton transport molecules including the voltage-gated proton channel Hv1 and the proton-selective ion channel OTOP1 (Decoursey, 2003; Ramsey et al., 2006; Bushman et al., 2015; Tu et al., 2018; Teng et al., 2019).
OTOP1 is a member of a family of proteins (Hughes et al., 2008; Hurle et al., 2011), which includes, within vertebrates, two other members, OTOP2 and OTOP3, that also function as proton channels (Tu et al., 2018). OTOP proton channels are expressed throughout the body, where they play diverse and still poorly understood roles in pH sensing and homeostasis. In vertebrates and invertebrates, OTOP channels expressed in the gustatory system sense acids and function as sour taste receptors (OTOP1 for vertebrates; OTOPL1 for Drosophila) (Teng et al., 2019; Zhang et al., 2019; Ganguly et al., 2021; Mi et al., 2021). In mice and zebrafish, OTOP1 plays an essential role in the formation of force-sensing calcium carbonate-based otoconia in the ear (Hurle et al., 2003; Hughes et al., 2004), likely by regulating pH in the endolymph. OTOP2 and OTOP3 are both found throughout the digestive system, and their expression has been shown to correlate with disease progression in some forms of colon cancer (Tu et al., 2018; Parikh et al., 2019; Yang et al., 2019). Most recently, an OTOP channel was shown to be critically involved in calcification and the formation of a skeleton in sea urchin embryos (Chang et al., 2021).
Given the recent discovery of OTOP proteins as forming ion channels (Tu et al., 2018), much remains to be discovered about how they function. For example, it was not known if the channels occupy open and closed states or if those terms even apply to these proteins, which bear no structural similarity to other ion channels, and that could conduct protons through a non-aqueous pathway (Decoursey, 2003). Recently, we showed that OTOP channels are gated by extracellular protons, acting mostly likely on multiple titratable residues on the extracellular domain of the protein (Teng et al., 2022). Here, we report the first evidence that OTOP channels can be activated by Zn2+ and Cu2+. This confirms that OTOP channels are gated, like nearly all other ion channels. Using a chimeric channel approach and point mutations, we identify residues in the linkers between tm 11-12 and tm 5-6 that represent key determinants for and likely form the Zn2+ activating site. We further propose that the elements of the gating apparatus of the channels that we identify are potential targets for pharmacological manipulation.
Results
Zn2+ both blocks and potentiates OTOP3 currents
While measuring the sensitivity of the three murine OTOP channels to inhibition by Zn2+, we noticed that mOTOP3 currents were larger following removal of Zn2+ than before its introduction. We refer to this activating effect of Zn2+ as potentiation. For these and other experiments, murine OTOP channels were expressed in HEK-293 cells and studied by patch-clamp recording. As shown in Figure 1, all three murine OTOP channels carried inward currents in response to a pH 5.5 stimulus and were subsequently inhibited by 1 mM Zn2+ at pH 5.5. However, only mOTOP3 currents showed recovery during the Zn2+ exposure and a large rebound, of nearly threefold, following its removal (Figure 1A and B). The potentiating effect of Zn2+ was dose-dependent over a concentration range of 0.1–10 mM and did not show evidence of saturation (Figure 1C and D). This contrasts with the inhibitory effect of Zn2+ on mOTOP3 currents, which for a stimulus of pH 5.5 showed clear saturation at 3 mM and could be fit with an IC50=0.31 mM (Figure 1D). The difference in the dose dependence of inhibition and potentiation suggests that Zn2+ interacts with distinct binding sites on the channels to produce the two effects (see below).

Zn2+ blocks and potentiates mOTOP3 currents.
(A) Representative traces show Zn2+ both inhibits and potentiates mOTOP3 currents. Proton currents were elicited in HEK293 cells expressing each of the three mOTOP channels in response to lowering the extracellular pH to 5.5 in the absence of extracellular Na+ as indicated. Zn2+ (pink bar, 1 mM) inhibits currents through all three channels, but only mOTOP3 currents are potentiated following Zn2+ removal. Vm was held at –80 mV. (B) Average and all points data for experiments as in (A) showing fold potentiation as measured by comparing the current magnitude before and after Zn2+ application (arrows shown in A) .(C) Zn2+ applied at varying concentrations (pink bar, concentration indicated in mM) produces a dose-dependent inhibition and potentiation of mOTOP3 currents. (D) Average data from experiments in (C) show the dose dependence of potentiation and inhibition (n=4 for Zn2+ potentiation, n=6 for Zn2+ inhibition). The dose dependence of Zn2+ inhibition was fit with a Hill equation, with an IC50=0.31 mM, and Hill coefficient = 0.94. (E) Average potentiation of mOTOP3 currents in response to 1 mM extracellular Zn2+ with (gray) or without (pink) 1 mM Zn2+ loaded in the pipette. There was no difference between the two conditions (Student’s t-test, p=0.88).
-
Figure 1—source data 1
Source data for Figure 1.
- https://cdn.elifesciences.org/articles/85317/elife-85317-fig1-data1-v2.xlsx
We next asked whether the effect of Zn2+ to potentiate the mOTOP3 currents was through actions on the extracellular or the intracellular side of the channel, possibly following entry into the cytosol as is the case for TRPA1 (Hu et al., 2009). Introducing 1 mM Zn2+ into the patch pipette did not change the degree of potentiation in response to 1 mM extracellular Zn2+, indicating that Zn2+ likely acts on extracellular domains of the channel (Figure 1E).
Pre-exposure to Zn2+ potentiates OTOP3 currents
To study the activating effects of Zn2+ on mOTOP3 and avoid confounds due to its inhibitory effect, we devised a recording protocol in which the cells were pre-exposed to Zn2+ at pH 7.4, prior to evoking currents with an acidic stimulus (pH 5.5). As shown in Figure 2A, this produced a robust potentiation of mOTOP3 currents evoked in response to the pH 5.5 stimulus. Notably, the currents were both faster and larger after exposure to Zn2+.

Pre-exposure to Zn2+ potentiates mOTOP3 currents in a dose- and time-dependent manner.
(A) Solution exchange protocol designed to measure effects of Zn2+ on gating of OTOP currents without confounds due to its inhibitory effects. Vm was held at –80 mV. In this example, currents were elicited to a pH 5.5 stimulus without pre-exposure to Zn2+ and then following a 16 s exposure to 1 mM Zn2+. (B) Representative traces show mOTOP3 currents elicited in response to pH 5.5 stimulus with pre-exposure to 0.3 mM (blue), 1 mM (green), and 3 mM (orange/red) Zn2+ for durations from 1 to 64 s as indicated. (C, D) The fold potentiation (C) and time to Ipeak (D) as a function of Zn2+ pre-exposure time from experiments as in (B) (n=5–7). Fold potentiation was measured as the ratio of the current evoked to the pH 5.5 stimulus after Zn2+ to the control response in the absence of Zn2+. Data are plotted as mean ± s.e.m.
-
Figure 2—source data 1
Source data for Figure 2.
- https://cdn.elifesciences.org/articles/85317/elife-85317-fig2-data1-v2.xlsx
We measured the dose and time dependence of potentiation by Zn2+, using three concentrations of Zn2+: 0.3, 1, and 3 mM and by varying the duration of the Zn2+ pre-exposure from 1 to 64 s. The response to Zn2+ was compared to the response in the absence of Zn2+ from the same cell. As shown in Figure 2B and C, 3 mM Zn2+ caused a more than 10-fold increase in the peak current, the lowest concentration of Zn2+, 0.3 mM, applied for up to 64 s produced only a negligible increase in the peak current and 1 mM Zn2+ had an intermediate effect. Thus, the potentiating effect of Zn2+, as measured by the peak current, was dose-dependent, with an apparent threshold of >0.3 mM Zn2+. Examination of the time dependence of the response showed increasing potentiation with exposure times up to ~16 s for concentrations of 1 and 3 mM Zn2+, at which point the effect tended to saturate, although there was some variability from cell to cell (Figure 2C).
In addition to increasing the peak current, Zn2+ pre-exposure also increased the apparent rate of activation to a pH 5.5 stimulus that otherwise slowly activates mOTOP3 currents (Teng et al., 2022). This change in apparent activation kinetics showed a dose and time dependence (Figure 2D). At the higher concentrations (1 and 3 mM), an exposure of 1 s was sufficient to observe a maximal decrease in the time to peak current from ~4 s to <1 s. At 0.3 mM Zn2+, this effect required longer exposure times, and activation rates never reached the speed obtained with 1 s exposure to 1 mM Zn2+. To assess the stability of the Zn2+ bound conformation, we measured the rate of recovery from potentiation by Zn2+ (Figure 3). For these experiments, the cells were exposed to 1 mM Zn2+ (pH 7.4) for 16 s. This was followed by a ‘wash-off’ period of 0–64 s in a Zn2+-free solution (pH 7.4) before currents were activated with a pH 5.5 solution (Figure 3A). A wash-off period of 1 s was sufficient to reduce potentiation by 50% while a period of >16 s allowed for a complete recovery of currents to baseline.

Time dependence of the recovery of mOTOP3 currents from Zn2+ pre-potentiation.
(A) Solution exchange protocol designed to measure the recovery of mOTOP3 currents following exposure to Zn2+. In this example, the cell expressing mOTOP3 was first exposed to 1 mM Zn2+ for 16 s which was followed by an 8 s wash-off phase in pH 7.4 solution before currents were elicited in response to the pH 5.5 solution. (B, C) The fold potentiation (B) and time to Ipeak (C) as a function of Zn2+ wash-off time from experiments as in (A) (n=4–6). Data are plotted as mean ± s.e.m.
-
Figure 3—source data 1
Source data for Figure 3.
- https://cdn.elifesciences.org/articles/85317/elife-85317-fig3-data1-v2.xlsx
Thus, application of millimolar concentrations of Zn2+ at pH 7.4 elicits upon its removal a robust concentration- and time-dependent potentiation of mOTOP3 currents.
OTOP1 is mildly potentiated by Zn2+
Using this new protocol, we went back and assessed the effect of Zn2+ on mOTOP1 and mOTOP2. mOTOP1 and mOTOP2 currents evoked in response to a solution at pH 5.5 showed little to no evidence of potentiation by exposure to 1 mM Zn2+ applied for 16 s at pH 7.4 (Figure 4A and B). As mOTOP1 and mOTOP2 differ from mOTOP3 channels in that they display a greater degree of activation at pH 5.5 (Teng et al., 2022), we considered whether this might preclude further potentiation by Zn2+. Indeed, we found that mOTOP1 currents evoked in response to a pH 6.0 stimulus could be potentiated by as much as threefold after a 64 s exposure to 1 mM Zn2+ (Figure 4C), with a time dependence similar to what we observed for mOTOP3 (Figure 4D). Activation rates of the more rapidly activating OTOP1 channels were not measurably enhanced by pre-exposure to Zn2+ exposure (Figure 4E). Thus, potentiation by Zn2+ is a feature shared by mOTOP1 and mOTOP3.

mOTOP1 is potentiated by Zn2+ when activated by a mild acid stimulus.
(A) Proton currents recorded from HEK293 cells expressing each of the three mOTOP channels as indicated, in response to pH 5.5 with (red) or without (black) Zn2+ pre-exposure. Vm was held at –80 mV. The cells were exposed to 1 mM Zn2+ for 16 s prior to pH 5.5 solutions. (B) Average and all points data from experiments as in (A) showing the fold potentiation in response to 1 mM Zn2+ for 16 s. (C) Representative traces showing mOTOP1 currents evoked in response to a pH 6.0 stimulus before and after exposure to Zn2+ for varying times as indicated. (D, E) The fold potentiation (C) and time to Ipeak (D) as a function of Zn2+ pre-exposure time from experiments as in (C) (n=5). Data are plotted as mean ± s.e.m.
-
Figure 4—source data 1
Source data for Figure 4.
- https://cdn.elifesciences.org/articles/85317/elife-85317-fig4-data1-v2.xlsx
Divalent transition metal ions potentiate and block OTOP3
Zn2+ modulates gating of a wide range of ion channels and neurotransmitter receptors, some of which are sensitive to other divalent transition metals interacting with the same residues as Zn2+ (Mathie et al., 2006; Shcheglovitov et al., 2012). For example, the zinc-activated ion channel, a member of the family of Cys-loop receptors, is also activated by copper (Cu2+) (Trattnig et al., 2016) while the cyclic nucleotide-gated ion channel from rods is potentiated by nickel (Ni2+), cadmium (Cd2+) and cobalt (Co2+), as well as by Zn2+, acting through a histidine residue in the mouth of the channel (Karpen et al., 1993; Gordon and Zagotta, 1995a). Divalent transition metals such as Co2+, Ni2+, Cu2+, and Cd2+ are predicted to have distinct preferred coordination geometries in metalloproteins and other proteins to which they bind but are often coordinated by the same acidic and/or polar residues including histidine, glutamic acid, aspartic acid, and cysteine (Rulísek and Vondrásek, 1998). To gain insights into the nature of the Zn2+ binding site, we tested whether other transition metals could potentiate mOTOP3 currents. Each metal ion was presented at a concentration of 1 mM for 16 s prior to activation of currents with a pH 5.5 stimulus. Pre-exposure to Cu2+ caused a dramatic increase in the magnitudes of the currents, potentiating them by 11.6±0.1-fold (n=10). Pre-exposure to Cd2+ and Ni2+ had more modest effects, potentiating mOTOP3 currents by 1.7±0.1 (n=10) and 1.6-fold±0.0 (n=8). In contrast, Co2+ and iron (Fe2+) had little to no effect on the magnitude or kinetics of the currents (Figure 5). These results suggest that the binding site occupied by Zn2+ may also be shared by other d-block transition metals.

Divalent transition metal ions also potentiate mOTOP3.
(A) Proton currents in response to a pH 5.5 stimulus following exposure (1 mM, 16 s) to various d-block transition metals recorded from HEK293 cells expressing wildtype mOTOP3. Vm was held at –80 mV. Black trace is the control from the same experiment (cell). (B) Average (mean ± s.e.m.) and all points data showing the fold potentiation measured from experiments as in (A). (C) Average (mean ± s.e.m.) and all points data for latency to Ipeak, measured from experiments as in (A). The latency to peak in (C) was scored as 8 s when peak magnitudes were not reached before 8 s.
-
Figure 5—source data 1
Source data for Figure 5.
- https://cdn.elifesciences.org/articles/85317/elife-85317-fig5-data1-v2.xlsx
We also tested if the metals that activate also inhibit mOTOP3, like Zn2+. Following activation of mOTOP3 by a pH 5.5 stimulus, and in the continued presence of the stimulus, each metal was applied at a concentration of 1 mM. Cu2+, Cd2+, and Ni2+ all inhibited mOTOP3 currents to varying degrees, with Cu2+ serving as the most potent inhibitor (Figure 5—figure supplement 1). Importantly, Ni2+ inhibited mOTOP3 currents to similar degree as Zn2+ but was a less potent activator. We conclude that the inhibitory and activating metal binding sites have different ligand specificity, consistent with the interpretation that they are distinct (see below).
The tm 11-12 linker is necessary and sufficient to confer sensitivity to Zn2+ potentiation
Given the marked difference between mOTOP3 and mOTOP2 in potentiating effects of Zn2+, we reasoned that chimeras between the two channels might allow us to identify its structural basis. As Zn2+ is likely to bind to an extracellular domain, we tested chimeras in which each of the six external linkers between transmembrane domains were exchanged (Teng et al., 2022). A total of twelve chimeras were tested, six in which the backbone was the mOTOP2 channel and six in which the backbone was the mOTOP3 channel. Each chimera was tested for potentiation following pre-exposure to 1 mM Zn2+ for 16 s (Figure 6).

The tm 11–12 linker is both necessary and sufficient for Zn2+ potentiation.
(A, B) Proton currents in response to a pH 5.5 stimulus with (red) or without (black) Zn2+ pre-exposure (1 mM, 16 s) recorded from HEK293 cells expressing either wildtype (WT) OTOP channels or chimeric channels as indicated. Vm was held at –80 mV. The WT traces are the same set as shown in Figure 3A. (C) Average data showing the fold potentiation after Zn2+ pre-exposure (1 mM, 16 s) measured from experiments as in (A) and (B). Bars are mean ± s.e.m. mOTOP2 and its chimeras are shown in blue, mOTOP3 and its chimeras are shown in red. Statistical significance determined with an ANOVA using Kruskal-Wallis (non-parametric) statistics. P values are shown where less than 0.05. (D, E) Same data as in (C) plotted to show current magnitudes before and after Zn2+ for WT channels and each of the chimeras. (F) Representative traces of O2/O3(L11-12) currents in response to pH 5.5 after pre-exposure to Zn2+ for varying times as indicated. (G) Average data for experiments as in (F) showing the time dependence of the potentiation by Zn2+ for the O2/O3(L11-12) chimera as compared with WT (n=4–5). Data of the WT mOTOP3 are the same set as shown in Figure 2C and D. (H) Left panel: response of mOTOP2 to alkaline stimulus (pH 9.0) with (red) and without (black) Zn2+ pre-exposure (1 mM, 16 s). Right panel: magnitude of currents at pH 9 for WT and mutant channels with and without Zn2+ pre-exposure. Currents were smaller after Zn2+ pre-exposure for both.
-
Figure 6—source data 1
Source data for Figure 6.
- https://cdn.elifesciences.org/articles/85317/elife-85317-fig6-data1-v2.xlsx
Strikingly, we found that Zn2+ potentiation was eliminated in a chimera containing the mOTOP3 backbone with the tm 11–12 linker from mOTOP2 (O3/O2(L11-12); Figure 6B, C and E). Indeed, even a 64 s exposure to 1 mM Zn2+ had no effect on the magnitude or activation kinetics of the currents (Figure 6—figure supplement 1). Interestingly, the O3/O2(L11-12) chimera is activated at a higher pH and currents are more rapidly activating than currents carried by mOTOP3 (Figure 6B and Teng et al., 2022). This suggests that the O3/O2(L11-12) channels may be partly locked in a potentiated state. None of the other chimeras with a mOTOP3 backbone showed a loss of potentiation.
Remarkably, simply transplanting the 11–12 linker from mOTOP3 onto mOTOP2 (O2/O3(L11–12)) conferred sensitivity to potentiation by Zn2+. Notably, pre-exposure of O2/O3(L11–12) channels to 1 mM Zn2+ for 16 s caused a nearly fourfold potentiation of the subsequent currents evoked in response to a pH 5.5. stimulus (Figure 6A, C and D). The other five chimeras containing an mOTOP2 backbone remained resistant to Zn2+ potentiation. Zn2+ potentiation of O2/O3(L11–12) showed a time dependence similar to that of wildtype mOTOP3 channels (Figure 6F and G). mOTOP2 and O2/O3(L11–12), but not mOTOP3, carry outward currents in response to alkaline stimuli (Teng et al., 2022). Thus, we wondered if the sensitivity to potentiation by Zn2+ would be observed for outward currents. Following exposure to 1 mM Zn2+ for 16 s, currents elicited in response to a pH 9.0 stimulus were smaller than in the absence of Zn2+ for both WT and O2/O3(L11–12) channels (Figure 6H). We conclude that the 11–12 linker contributes to the potentiation of OTOP channels by Zn2+ in response to acidic but not alkaline stimuli.
While none of the other chimeras containing an mOTOP3 backbone showed a reduction in Zn2+ potentiation, several showed an increase (Figure 6C and E). Interestingly, a chimera with a swap of the tm 1–2 linker, O3/O2(L1–2), which was previously observed to be non-conductive in response to changes in extracellular pH (Teng et al., 2022), was strongly activated by Zn2+. Thus, for this chimeric channel, the increase in fold potentiation mostly reflects the large decrease in the acid-induced currents, rather than an increase in the magnitude of the currents after Zn2+ exposure and suggests that the 1–2 linker plays a role in acid activation of mOTOP3.
To determine if potentiating and inhibiting effects of Zn2+ were mediated by the same binding site, we next tested whether currents carried by O3/O2(L11–12) and other mOTOP3 chimeras retained sensitivity to inhibition by Zn2+ following activation at pH 5.5 (Figure 7). All chimeras were inhibited by 1 mM Zn2+, including the O3/O2(L11–12) chimera, although the extent of the inhibition, which ranged from ~40% to 80%, varied significantly between some of the chimeras and WT channels (Figure 7). Interestingly, potentiation following Zn2+ inhibition was absent not just in the O3/O2(L11–12) chimera, as expected, but also in the O3/O2(L3–4) chimera which showed potentiation with the pre-exposure protocol. This suggests that while differences in L3–4 do not account for differences in Zn2+ potentiation between mOTOP2 and mOTOP3, residues in L3–4 may nonetheless contribute to activation of the channels.

Inhibition of mOTOP3 by Zn2+ is retained in chimeric channels.
(A) Zn2+ sensitivity of chimeric mOTOP3-mOTOP2 channels as measured with a pre-exposure protocol (left panel in each) or by adding 1 mM Zn2+ to the pH 5.5 stimulus (blocking protocol; right panel in each). Chimeras containing mOTOP2 N-domain and C-domain linkers are shown in the left column and right columns, respectively. Data from pre-exposure experiments is also presented in Figure 6, and here is shown for comparison to results with the blocking protocol. (B) Representative traces of wildtype mOTOP3 currents in response to the same protocols as in (A). The arrow indicates the time point in this trace where inhibition by Zn2+ was measured, by comparison with the current magnitude before adding Zn2+. (C) Average (mean ± s.e.m.) and all points data showing fractional inhibition of currents by 1 mM Zn2+ measured from wildtype mOTOP3 and its chimeras. All channels were similarly inhibited by 1 mM Zn2+. Significance determined by ANOVA with Dunnett’s test corrected for multiple comparisons.
-
Figure 7—source data 1
Source data for Figure 7.
- https://cdn.elifesciences.org/articles/85317/elife-85317-fig7-data1-v2.xlsx
Together we conclude that the mOTOP3 11–12 linker is both necessary and sufficient to confer sensitivity to potentiation by Zn2+. Other parts of the channel may contribute to gating by Zn2+ either directly or through allosteric effects.
Contribution of H531 and other residues to Zn2+ potentiation
The tm 11–12 linker is relatively short, consisting of sixteen amino acids, of which five are conserved between the three murine OTOP channels (Figure 8A). In this region, the residues that could coordinate Zn2+ and that vary between mOTOP3 and mOTOP2 are H531, E533, and E535 (mOTOP3 numbering, Figure 8A). Within mOTOP3, we mutated each residue to that found in mOTOP2 or to alanine. Strikingly, mutation of H531 to either arginine (found in mOTOP2) or alanine eliminated the ability of Zn2+ to potentiate mOTOP3 currents, assessed by measuring either the magnitude of the currents or their activation kinetics (Figure 8B-D). However, the converse was not true: introducing histidine at the same position in mOTOP2 (R517H) was not sufficient to produce potentiation by Zn2+ under the conditions tested (Figure 8—figure supplement 1). mOTOP1 has an arginine at the equivalent residue to H531, but is still potentiated by Zn2+, albeit more weakly. Mutation to histidine at this position in OTOP1 (R554H) was sufficient to significantly increase Zn2+ potentiation (Figure 8—figure supplement 1).

H531 in mOTOP3 L11–12 is essential for Zn2+ potentiation.
(A) Sequence alignment of three mOTOP channels. The residues that were exchanged between mOTOP2 and mOTOP3 in the L11–12 chimeras is indicated with a blue box. Residues that differed between the two channels and that were tested are indicated by red arrows. (B) Representative traces of mOTOP3_H531A and H531R currents in response to pH 5.5 after pre-exposure to Zn2+ for varying times as indicated. (C, D) Average data for fold potentiation (C) and latency to Ipeak (D) measured from experiments as in (B), plotted as a function of pre-exposure time to Zn2+ (n=3–7). Data from wildtype (WT) mOTOP3 are the same set as shown in Figure 2C and D. (E, F) Responses of mOTOP3 mutants as indicated in response to pH 5.5 with (red) or without (black) Zn2+ pre-exposure (1 mM, 16 s). Vm was held at –80 mV. (G) Average data for fold potentiation measured from experiments as in (E and F). Statistical significance compared with WT determined using the Kruskal-Wallis (non-parametric) test corrected for multiple comparison. (H) Latency to peak currents of WT mOTOP3 or mutant currents, measured from experiments as in (E and F). Each set of points represents a separate cell.
-
Figure 8—source data 1
Source data for Figure 8.
- https://cdn.elifesciences.org/articles/85317/elife-85317-fig8-data1-v2.xlsx
Taken together, we conclude that H531 is a critical element of the Zn2+ activating site in mOTOP3. However, the complete binding site is undoubtedly formed by multiple residues that together coordinate Zn2+. We, therefore, set out to identify these residues. Mutation of the acidic residues (E533 and E535) within the L11–12 linker had more subtle effects: mutation to residues found in mOTOP2 (H and S, respectively) had no effect on potentiation while mutation of E535 to alanine significantly reduced, but did not eliminate, potentiation (Figure 8E-H). Thus, the potentiation of mOTOP3 by Zn2+ may involve contributions, either direct or indirect, from both H531 and E535 in the tm 11–12 linker, and the relative contribution of each residue may vary between different channels and under different conditions.
Inspection of the structure of mOTOP3 predicted by AlphaFold (Jumper et al., 2021) reveals a possible Zn2+ binding site formed by H531, E535, and residues in the linker between tm 5–6, H234, and E238 (Figure 9B). Note that a histidine at the position equivalent to 234 (mOTOP3 numbering) is conserved among all murine OTOP channels while glutamic acid is present at a position equivalent to 238 in mOTOP1 (Figure 9A). To test the contributions of these residues to Zn2+ potentiation, we mutated each alone or in combination to alanine. The single mutations H234A and E238A each showed significantly reduced potentiation (3.5±0.64 and 2.4±0.12-fold potentiation, respectively) as compared with wildtype (10.0±1.0-fold) and the double mutation, H234A/E238A showed a further reduction in potentiation (1.3±0.21) as compared with either of the single mutants (Figure 9F and G). Interestingly, all mutants retained some degree of potentiation, and current activation was faster after Zn2+ exposure, which was not observed for the mOTOP3 H531A/R single mutation or the mO3_mO2_L11-12 chimera (Figure 8D and Figure 6—figure supplement 1C).

H234 and E238 in transmembrane domain 5 contribute to Zn2+ potentiation of mOTOP3.
(A) Sequence alignment of the three mOTOP channels. Alpha helices shown above the sequence are based on the AlphaFold prediction of the structure of mOTOP3. Red arrows indicate residues neutralized to alanine in subsequent experiments. (B) Images generated using the AlphaFold predicted structure of mOTOP3. Left panel shows a sideview and right panel shows a topview of mOTOP3. Each zoom-in highlights the predicted Zn2+ potentiation binding site. (C,D,E) Representative traces of mOTOP3 H234A, E238A, and a double mutation of H234A/E238A in response to lowering the extracellular pH to 5.5 after (red) and without (black) pre-exposure to Zn2+ (1 mM, 16 s). Vm was held at –80 mV. (F) Average (mean ± s.e.m.) and all points data showing the fold potentiation measured from experiments as in (C,D,E). Statistical significance determined using an ANOVA with the Kruskal-Wallis (non-parametric) test. (G,H) Same data as in (F) plotted to show current magnitudes before and after Zn2+ (G) and the time to Ipeak with and without pre-exposure to Zn2+ (H) from experiments in (C,D,E).
-
Figure 9—source data 1
Source data for Figure 9.
- https://cdn.elifesciences.org/articles/85317/elife-85317-fig9-data1-v2.xlsx
The data collectively demonstrate that gating of OTOP channels is regulated by Zn2+ acting through residues within the tm 11–12 and tm 5–6 linkers.
Kinetic model for Zn2+ potentiation
These data suggest that Zn2+ may act to lock the channel in an open state, possibly by binding more strongly to the same site that is titrated by H+ ions to activate the channel (Teng et al., 2022). The data also suggest that a separate site with a faster off-rate mediates inhibition by Zn2+. Thus, in the presence of Zn2+, the channels may enter a state that is simultaneously activated (open gate) and inhibited. Upon removal of Zn2+, the channels may then transit through a fully open state, as the inhibition is relieved faster than the activation. To formally test these predictions, we generated a kinetic model reflecting these properties and asked if it could recapitulate our experimental observations.
We postulated a model comprised of interacting elements that can each transition between two configurations: a pore gate that is either closed or open, a binding site for protons, and two types of binding sites for Zn2+ (one activating and one blocking), that are either unoccupied or occupied (Figure 10A). For a detailed description of this type of model representation, see Goldschen-Ohm et al., 2014. The proton and activating Zn2+ sites are energetically coupled to the pore such that binding speeds pore opening and/or slows pore closing. To model the competition of protons and Zn2+ for the same activating site(s), we destabilized all states where both the proton and activating Zn2+ sites are simultaneously occupied. We modeled the blocking Zn2+ site as independent of all other model elements, except that when occupied all channel current is blocked. Simulated currents in response to the same pH and Zn2+ protocols used in our experiments qualitatively describe our observed mOTOP3 current responses (Figure 10B and C), supporting the plausibility of our proposed mechanism. For example, the model recapitulates the rebound following addition and removal of Zn2+ during an acid stimulus. It also recapitulates the speeding up of current activation for a pre-exposure to Zn2+ at 0.3 mM, and the increase in current magnitudes with pre-exposure to Zn2+ at 1 and 3 mM. It also qualitatively recapitulates the decay of the potentiated currents, which in the model is due to unbinding of the Zn2+ (see Figure 10—figure supplements 1–3). Conceptually, upon occupancy of the activating site by Zn2+ the channel will enter a potentiated state, and thereafter upon removal of Zn2+ the channel will slowly return to its baseline activity in the absence of Zn2+ with a time course that is largely described by the time course of Zn2+ unbinding from the activating site (Figure 10—figure supplement 2). We note that such a decay could potentially also reflect a desensitization or ion accumulation process which we did not attempt to model.

mOTOP3 kinetic model for Zn2+ potentiation and block.
(A) Kinetic model for activation of mOTOP3 by H+ and Zn2+. The channel moves from a closed state (C) upon binding H+ or Zn2+ to an open state (Zn-O, H–O) in which it permeates protons. The doubly bound Zn-H-O and Zn-H-C states are disfavored energetically. A separate site binds Zn2+ and inhibits channel permeation, independently of the gating state. See methods for a more detailed description of the model. (B) Example mOTOP3 current recordings from Figure 1C and Figure 2B. Each set of recordings was from a different cell. (C) Simulated currents for the model under the same protocols as in (B). Model parameters were adjusted manually and were the same for all traces. The model replicates the rebound seen upon addition of Zn2+ at pH 5.5, and the potentiation seen with pre-exposure to Zn2+ at pH 7.4.
However, this model also undoubtedly represents an oversimplification of the true number and properties of the states the channel adopts. For example, the model does not recapitulate the observed decay of the currents below baseline when strongly potentiated, which may reflect contributions from an inactivation process or ion accumulation not modeled. Making things more complicated, and interesting, it is also entirely possible that there is more than one permeation pathway and gate (Chen et al., 2019; Saotome et al., 2019), and that Zn2+ opens one of the permeation pathways and not the other. The 11–12 loop sits close to the intrasubunit interface, so that at present it is not possible to predict if it would open a permeation pathway in the N domain, the C domain, or the intrasubunit interface.
Discussion
The OTOP proton channels were discovered in 2018 when mOTOP1 was cloned from taste tissue as a putative sour receptor (Tu et al., 2018). At that time, mOTOP1 was shown to have near-perfect selectivity for protons over other ions, and to generate inward currents as the pH was lowered. Vertebrate OTOP2 and OTOP3, as well as invertebrate channels also carry inward proton currents in response to extracellular acidification, and where tested have been shown to function as proton-selective ion channels (Tu et al., 2018; Chen et al., 2019) (see also Teng et al., 2022). Nearly all descriptions regarding the functional properties of OTOP channels come from the heterologous expression of OTOP channels and save for the description of proton currents now attributed to mOTOP1 channels in taste cells (Chang et al., 2010; Bushman et al., 2015), all descriptions of OTOP channels postdate the discovery of the genes encoding the proton channels. That is, even by scouring the literature, it is difficult to find a description of a proton current that could be attributed to an OTOP channel in a native cell type. Thus, in contrast to K+ channels where there was a vast literature regarding their functional properties before their cloning, for OTOP channels, this kind of information was not available.
Critically, before this work and that described in Teng et al., 2022, it was not known if OTOP channels were gated, and if so by what. Here, we provide evidence that OTOP channels are gated by Zn2+ and other transition metals. In the apo, Zn2+-free condition, mOTOP3 currents are small and slowly activating. With pre-exposure to 1–3 mM Zn2+ for several seconds, the currents increase in magnitude by up to 10-fold. This can only be explained if Zn2+ either increases the probability that the channels open or produces a change in their conductance – gating them. We also find that pre-exposure to copper (Cu2+) strongly potentiates OTOP3 currents. Together with other recent evidence for pH-dependent gating (Teng et al., 2022), we can now say with certainty that some OTOP channels are gated.
Structural considerations
The structures of two of the three OTOP channels have been reported at near-atomic resolution ([zebrafish OTOP1, chicken OTOP3, and Xenopus OTOP3; Chen et al., 2019; Saotome et al., 2019)]. In addition, predictions are available for structures of mouse and human OTOP channels that appear to be reliable (Jumper et al., 2021; Teng et al., 2022; Varadi et al., 2022). All the structures to date share common features: The protomers assemble as dimers, and each protomer shows a twofold symmetry, leading to an overall pseudo-tetrameric stoichiometry. The four pseudo-subunits come together to form a central cavity that is filled with lipids and cannot support ion permeation. Instead, ions may permeate through the barrel-like structures formed from transmembrane domains 2–6 (N domain) and 8–12 (C domain) or at the intrasubunit interface (formed mostly by tm 6 and 12). From these static structures, it is not possible to tell if the channels are gated, and if so, what state they are in (open or closed). Based on the pH sensitivity of the channels (Teng et al., 2022), we presume they are closed.
We have focused on mOTOP2 and mOTOP3, which show the most divergent functional properties. In addition to differences in the effects of Zn2+ described here, we have also described differences in pH sensitivity of the two channels (Teng et al., 2022). mOTOP3 is gated by protons, and thus conducts currents only in response to extracellular acidification (<pH 5.5), while OTOP2 is constitutively open and conducts currents over a large pH range, including outward currents when the extracellular solution is alkalinized. The strikingly different functional properties of the two channels, but otherwise overall similar architecture, provided us the opportunity to identify motifs involved in gating using a chimeric approach.
Remarkably, we found that a short stretch of amino acids linking transmembrane domains 11–12 was necessary for Zn2+ potentiation of mOTOP3 and sufficient to confer Zn2+-sensitive gating on mOTOP2. Within that stretch, we identified one amino acid, H531, mutations of which (to R or A) completely abolished Zn2+-sensitive gating in mOTOP3. As histidine is well documented to form part of the Zn2+ binding site in other Zn2+-sensitive proteins (Vallee and Auld, 1990), including the voltage-gated proton channel HV1 (Ramsey et al., 2006), H531 likely contributes to coordinating Zn2+ in mOTOP3. However, Zn2+ is typically coordinated by sidechains of four or more residues, which in addition to the imidazole rings of histidine, includes sulfhydryl groups of cysteine, carboxyl group of acidic residues (aspartic and glutamic acid). Moreover, water can also participate in coordinating Zn2+ (Vallee and Auld, 1990). In addition to H531, we have identified three other residues, H234, E238, and E535, that when mutated to alanine lead to reduced efficacy of Zn2+ to potentiate mOTOP3. Based on the functional data provided in this paper and an AlphaFold structure of mOTOP3, we tentatively predict these four residues form the Zn2+ activating binding site. Interestingly, alkali activation of mOTOP1 was recently shown to require a basic residue at a position equivalent to H531 in mOTOP3 and in the tm 11–12 linker at a position close to H234 (Tian et al., 2023). It is tempting to speculate that Zn2+ binding could bring together residues in the tm 5–6 linker and tm 11–12 linker that are at a distance in the closed state (8 Å; Figure 9B) to stabilize an open state, as shown for other ion channels (Gordon and Zagotta, 1995b; Zhang et al., 2009). A similar mechanism could open mOTOP1 channels in response to alkaline pH.
Role of zinc in health and disease
Zinc is required for the functioning of a wide range of enzymes and proteins in the body and supplementary zinc has both beneficial and detrimental health effects. Given that it is mostly not known which cells express OTOP channels, or what functions they play in those cells, it is hard to predict the functional role of Zn2+-sensitive gating in physiology. The dual blocking and potentiating effect of Zn2+ make it likely that the enhanced activity would be mostly evident under conditions where Zn2+ concentrations dropped rapidly or that favored binding of Zn2+ to its activating site over its blocking site. These studies also raise the interesting prospect that other molecules could be found that function as positive allosteric modulators of OTOP channels, with selectivity for the activating site over the blocking site, and that could ameliorate effects of reductions in OTOP channel function that lead to vestibular or other disorders.
Materials and methods
Reagent type (species) or resource | Designation | Source or reference | Identifiers | Additional information |
---|---|---|---|---|
Gene (Mus musculus) | Otop1, Otop2, and Otop3 | Tu et al., 2018. PMID:29371428 | ||
Cell line (Homo sapiens) | HEK293 | ATCC | CRL-1573 | |
Cell line (Homo sapiens) | PAC-KO HEK293 cells | Yang et al., 2019. PMID:31023925 | ||
Recombinant DNA reagent | Otop1, Otop2 and Otop3 in pcDNA3.1 | Tu et al., 2018. PMID:29371428 | ||
Recombinant DNA reagent | Otop1, Otop2 and Otop3 – GFP | Saotome et al., 2019. PMID:31160780 | ||
Recombinant DNA reagent | mO2_O3 loop swap mutations | This paper | cDNAs encode chimeric channels (see Materials and methods and Figure 6—figure supplement 1). Available upon request | |
Recombinant DNA reagent | mO3_O2 loop swap mutations | This paper | cDNAs encode chimeric channels (see Materials and methods and Figure 6—figure supplement 1). Available upon request | |
Recombinant DNA reagent | pHluorin in pcDNA3 | Miesenbock, et al., 1998. PMID:9671304 | ||
Chemical compound, drug | CHES | Sigma | C2885 | |
Chemical compound, drug | PIPES | Sigma | P6757 | |
Chemical compound, drug | Homopiperazine-1,4-bis(2-ethanesulfonic acid) | Sigma | 53588 | |
Software, algorithm | GraphPad Prism 8 and 9 | GraphPad | RRID:SCR_002798 | |
Software, algorithm | pClamp and clampfit | Molecular Devices | RRID:SCR_011323 | |
Software, algorithm | Origin | OriginLab corporation | RRID:SCR_002815 | |
Software, algorithm | CorelDraw | Corel | RRID:SCR_014235 | |
Software, algorithm | SimplePCI | HCImage | https://hcimage.com/simple-pci-legacy/ |
Clones, cell lines, and transfection
Request a detailed protocolMouse OTOP1, OTOP2, and OTOP3 cDNAs were in a pcDNA3.1 vector with an N-terminal fusion tag consisting of an octahistidine tag followed by eGFP, a Gly-Thr-Gly-Thr linker, and 3C protease cleavage site (LEVLFQGP) were as previously described (Saotome et al., 2019). All chimeras and mutations were generated using In-Fusion Cloning (Takara Bio) and were confirmed by Sanger sequencing (Genewiz) (see Teng et al., 2022).
HEK293 cells were purchased from ATCC (CRL-1573). The cells were cultured in a humidified incubator at 37°C in 5% CO2 and 95% O2. The high glucose DMEM media (Thermo Fisher) is implemented with 10% fetal bovine serum (Life Technology) and 1% penicillin-streptomycin antibiotics. Cells were passaged every 3–4 days. Cells were tested and found free of mycoplasma using a PCR detection kit (Sigma-Aldrich, USA).
Cells used for patch-clamp recordings were transfected in 35 mm Petri dishes, with 600–1000 ng DNA and 2 µL TransIT-LT1 transfection reagents (Mirus Bio Corporation) following the manufacturer’s protocol. After 24 hr, the cells were lifted using trypsin-EDTA and plated into a recording chamber.
Patch-clamp electrophysiology
Request a detailed protocolWhole-cell patch-clamp recording was performed as previously described (Teng et al., 2019). Briefly, recordings were obtained with an Axonpatch 200B amplifier, digitized with a Digidata 1322a 16-bit data acquisition system, acquired with pClamp 8.2, and analyzed with Clampfit 9 or 10 (Molecular Devices). Records were sampled at 5 kHz and filtered at 1 kHz. Patch pipettes with a resistance of 2–4 MΩ were fabricated from borosilicate glass (Sutter Instrument). Solution exchange was achieved with a fast-step perfusion system (Warner Instrument, SF-77B) custom modified to hold seven microcapillary tubes in a linear array. Cells were treated with trypsin-EGTA and plated into the recording chamber immediately before each experiment. After a gigaohm seal was formed and whole-cell recording was achieved, the cell was lifted and moved in front of the microcapillary tubes. The membrane potentials were held at –80 mV unless otherwise indicated.
Patch-clamp electrophysiology solutions
Request a detailed protocolTyrode’s solution contained 145 mM NaCl, 5 mM KCl, 1 mM MgCl2, 2 mM CaCl2, 20 mM dextrose, 10 mM HEPES (pH adjusted to 7.4 with NaOH). Standard pipette solution contained 120 mM Cs-aspartate, 15 mM CsCl, 2 mM Mg-ATP, 5 mM EGTA, 2.4 mM CaCl2 (100 nM free Ca2+), and 10 mM HEPES (pH adjusted to 7.3 with CsOH; 290 mOsm). Standard Na+-free extracellular solutions contained 160 mM NMDG, 2 mM CaCl2, and 10 mM buffer (HEPES for pH 7.4, MES for pH 6.0 and 5.5), and pH was adjusted with HCl to 7.4.
ZnCl2 was directly introduced into the Na+-free solutions where the final concentration was less than 3 mM, which caused a change in osmolarity of <10 mOsm. The pH was re-adjusted with NMDG-OH or HCl as needed. For the experiment in Figure 1C, the concentration of NMDG in the 10 mM Zn2+ solution was reduced to maintain an ~300 mOSM osmolarity.
For experiments in Figure 5 and Figure 5—figure supplement 1, FeCl2, CoCl2, NiCl2, CuCl2, and CdCl2 were directly introduced into the Na+-free solutions at a concentration of 1 mM immediately before the experiment was performed. The pH was re-adjusted with NMDG-OH or HCl as needed.
Quantification and statistical analysis
Request a detailed protocolAll data are presented as mean ± SEM if not otherwise noted. Statistical analysis was performed using GraphPad Prism 9 (GraphPad Software Inc). The sample sizes of 3–10 independent recordings from individual cells per data point are similar to those in the literature for similar studies. For comparison of fold potentiation, an ANOVA was used with the Kruskal-Wallis (non-parametric) test (GraphPad Prism 9). All replicates are biological replicates. Number of replicates are indicated in the figure or figure legend. They represent recordings from different cells transfected with the same plasmid DNA. Data were excluded if a gigaohm seal was not established or maintained, as indicated by an inward current of >80 pA (–80 mV) in the presence of the OTOP channel blocker Zn2+ applied at pH 7.4. Data sets that represent time series were excluded if four or fewer time points were collected out of a possible eight time points, due to seal instability.
Representative electrophysiology traces shown in the figures were acquired with pClamp, and in some cases, the data was decimated by 10-fold before exporting into graphic programs, Origin (Microcal) and Coreldraw (Corel).
Kinetic modeling
Request a detailed protocolA kinetic model was generated based on interacting binary elements as described in Goldschen-Ohm et al., 2014. Briefly, each element can adopt one of two possible configurations (closed or open for the pore, unbound or bound for each binding site) with intrinsic rate constants for transitioning between them. The elements are energetically coupled such that binding at a particular site either promotes or inhibits pore opening. To simulate competition between H+ and Zn2+ for the activating site(s), occupation of one site energetically destabilizes the other.
Model parameters were manually adjusted to qualitatively recapitulate the experimental observations. Rate constants for pore opening/closing (ko/kc), proton binding/unbinding (konH/koffH), Zn2+ binding/unbinding to activating sites (konZnA/koffZnA) and blocking site (konZnB/koffZnB) are (s–1 or M–1s–1 for binding rates): ko = 0.02, kc = 25, konH = 5 × 104, koffH = 1, konZnA = 35, koffZnA = 1, konZnB = 5 × 104, and koffZnB = 20. In Figure 10 these rates reflect the transition pairs from the unbound closed state C to either the unbound open state O (ko, kc) or each closed state having only a single binding site occupied (H∙C, Zn∙C, Blocked∙C) (konH, koffH, konZnA, koffZnA, konZnB, koffZnB). To describe how binding to a site influences pore opening and/or binding to other sites, and vice versa, we define state-dependent interaction energies between the pore gate and binding site domains. These energies modulate the rate constants as defined in Goldschen-Ohm et al., 2014. The interaction between the proton site and pore gate is such that proton binding reduces the energy barrier for pore opening by –4 kcal/mol without affecting pore closure. For example, this means that the transition rate from H∙C to H∙O will be exp(–(–4 kcal/mol)/RT)~850 times faster than that from C to O, where RT is the product of the molar gas constant R and temperature T and has value 0.593 kcal/mol at a room temperature of 298 K. Reciprocally, pore opening reduces the energy barrier for proton binding by –1 kcal/mol. These energetic effects imply that pore opening must also increase the barrier for proton unbinding by 3 kcal/mol (assuming transition rates are described by a single transition state energy barrier). The interaction between the activating Zn2+ site and pore gate is such that Zn2+ binding both reduces the energy barrier for pore opening by –3 kcal/mol and increases the barrier to pore closure by 3 kcal/mol, thereby increasing opening frequency and stabilizing the open pore. Reciprocally, pore opening reduces the barrier for Zn2+ binding to the activating site by –5 kcal/mol. These energetic effects imply that pore opening must also increase the barrier for Zn2+ unbinding from the activating site by 1 kcal/mol, thereby stabilizing bound Zn2+ at the activating site. To simulate competition between protons and Zn2+ for the activating site, states with both proton and activating Zn2+ sites occupied were destabilized by 6 kcal/mol, the value for which was simply chosen to be a somewhat large destabilizing energy. The Zn2+ blocking site was assumed to be independent of all other elements, but when occupied blocks all channel current. This model includes 14 free parameters: eight rate constants describing the opening/closing of the pore gate and binding unbinding at the proton and Zn2+ activating and blocking sites, and six state-dependent interaction energies defining how activation/occupation of the above elements influences the rates of the other elements. However, the model is relatively insensitive to some of the parameters (within reason). For example, binding/unbinding of Zn2+ from the blocking site is very fast in comparison to the macroscopic current kinetics of proton activation and decay of Zn2+ potentiation. Thus, any set of blocking kinetics that give a relatively rapid onset and offset of the block will likely perform similarly in the model.
For comparison with model simulations, the entire set of currents across all cells was uniformly scaled under the assumption that the maximal response following preapplication of 3 mM Zn2+ is reflective of channels with an open probability of ~0.9, a measure that has not been experimentally verified. It is likely that the same model structure will similarly describe responses with lower maximal open probabilities, albeit with slightly different parameter values. Given these uncertainties, as well as the relatively large number of free parameters (11 parameters defining pore gating, binding to activating sites, and their interactions are most relevant), we did not attempt to optimize the model parameters, but instead manually identified parameters that qualitatively recapitulate the data. The time-dependent probability in each state was simulated as described by Colquhoun and Hawkes, 1995, after first generating the matrix of transition rates between states from the model’s binary elements and interactions (Goldschen-Ohm et al., 2014). Currents were computed from the simulated open probability assuming a conductance of 1 pS and a driving force of –80 mV in pH 5.5 and 0 in pH 7.4 as estimated from ramp experiments (Teng et al., 2022). The choice of conductance is arbitrary given that we do not know how many channels are in each recording, and thus only contributes to the overall scaling of the current.
Material availability
Request a detailed protocolAll materials generated during this study including mutant channels are available upon request.
Data availability
Request a detailed protocolAll data generated or analyzed during this study are included in the manuscript and supporting files. The source code for simulations is provided in supplementary information.
Data availability
All data generated or analysed during this study are included in the manuscript and supporting files.
References
-
Subunit-Dependent high-affinity zinc inhibition of acid-sensing ion channelsThe Journal of Neuroscience 24:8678–8689.https://doi.org/10.1523/JNEUROSCI.2844-04.2004
-
Voltage-Gated proton channels and other proton transfer pathwaysPhysiological Reviews 83:475–579.https://doi.org/10.1152/physrev.00028.2002
-
Divalent cations and the activation kinetics of potassium channels in squid giant axonsThe Journal of General Physiology 79:965–996.https://doi.org/10.1085/jgp.79.6.965
-
BookProbing gating mechanisms of sodium channels using pore blockersIn: Goldschen-Ohm MP, editors. Handbook of Experimental Pharmacology. Springer. pp. 183–201.https://doi.org/10.1007/978-3-642-41588-3_9
-
A nonequilibrium binary elements-based kinetic model for benzodiazepine regulation of GABAA receptorsThe Journal of General Physiology 144:27–39.https://doi.org/10.1085/jgp.201411183
-
Zinc activates damage-sensing TRPA1 ion channelsNature Chemical Biology 5:183–190.https://doi.org/10.1038/nchembio.146
-
Otopetrin 1 is required for otolith formation in the zebrafish Danio rerioDevelopmental Biology 276:391–402.https://doi.org/10.1016/j.ydbio.2004.09.001
-
Interactions between divalent cations and the gating machinery of cyclic GMP-activated channels in salamander retinal rodsThe Journal of General Physiology 101:1–25.https://doi.org/10.1085/jgp.101.1.1
-
Zinc and copper: pharmacological probes and endogenous modulators of neuronal excitabilityPharmacology & Therapeutics 111:567–583.https://doi.org/10.1016/j.pharmthera.2005.11.004
-
Molecular and cellular basis of acid taste sensation in DrosophilaNature Communications 12:3730.https://doi.org/10.1038/s41467-021-23490-5
-
The direct modulatory activity of zinc toward ion channelsIntegrative Medicine Research 4:142–146.https://doi.org/10.1016/j.imr.2015.07.004
-
Zinc as allosteric ion channel modulator: ionotropic receptors as metalloproteinsInternational Journal of Molecular Sciences 17:1059.https://doi.org/10.3390/ijms17071059
-
Coordination geometries of selected transition metal ions (co2+, ni2+, cu2+, zn2+, cd2+, and hg2+) in metalloproteinsJournal of Inorganic Biochemistry 71:115–127.https://doi.org/10.1016/s0162-0134(98)10042-9
-
Binding of a gating modifier toxin induces intersubunit cooperativity early in the shaker K channel’s activation pathwayThe Journal of General Physiology 128:119–132.https://doi.org/10.1085/jgp.200609492
-
Structures of the otopetrin proton channels otop1 and otop3Nature Structural & Molecular Biology 26:518–525.https://doi.org/10.1038/s41594-019-0235-9
-
Molecular and biophysical basis of glutamate and trace metal modulation of voltage-gated Ca (V) 2.3 calcium channelsThe Journal of General Physiology 139:219–234.https://doi.org/10.1085/jgp.201110699
-
Vertebrate OTOP1 is also an alkali-activated channelNature Communications 14:26.https://doi.org/10.1038/s41467-022-35754-9
-
Copper and protons directly activate the zinc-activated channelBiochemical Pharmacology 103:109–117.https://doi.org/10.1016/j.bcp.2016.02.004
-
The biochemical basis of zinc physiologyPhysiological Reviews 73:79–118.https://doi.org/10.1152/physrev.1993.73.1.79
Decision letter
-
Leon D IslasReviewing Editor; Universidad Nacional Autónoma de México, Mexico
-
Kenton J SwartzSenior Editor; National Institute of Neurological Disorders and Stroke, National Institutes of Health, United States
-
Leon D IslasReviewer; Universidad Nacional Autónoma de México, Mexico
-
Jon T SackReviewer; University of California, Davis, United States
Our editorial process produces two outputs: (i) public reviews designed to be posted alongside the preprint for the benefit of readers; (ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.
Decision letter after peer review:
Thank you for submitting your article "Zn2+ potentiation of OTOP proton channels identifies structural elements of the gating apparatus" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, including Leon D Islas as Reviewing Editor and Reviewer #1, and the evaluation has been overseen by Kenton Swartz as the Senior Editor. The following individual involved in the review of your submission has agreed to reveal their identity: Jon T Sack (Reviewer #2).
The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.
Essential revisions:
After consultation, the reviewers concur on these essential revisions to the manuscript.
1) In the explanation of the model predictions, it is not clear how the allosteric model can predict the desensitization of currents. This part needs a more detailed explanation. Also, it was difficult to tell if the description in methods of the kinetic model fully describes the model. There appear to be 24 different rate constants. Perhaps labeling a table listing all the free parameters using a nomenclature clearly addressable to Figure 9A would help readers understand how the model works.
2) Given the effect of H531 in OTOP3 it is reasonable to wonder if the histidine is sufficient to recapitulate the O3/O2tm11-12 phenotype or at least part of it. The same position should be replaced with histidine in OTOP2 and OTOP1, which have arginines. Related to this point: can the authors provide an explanation for the OTOP1 potentiation given that this isoform does not have the corresponding H531 residue?
3) Please comment on to what extent the desensitization of the proton currents could be a reflection of proton depletion given the large amplitude of the proton currents.
Also please attend to the specific comments of reviewers below. Only the experiments in section 2 above are viewed as essential.
Reviewer #1 (Recommendations for the authors):
The authors of this manuscript have discovered that the recently characterized proton (H+) gated, proton permeable channel OTOP3 is strongly potentiated by the extracellular application of zinc (Zn2+) ions, while the related isoform OTOP2 is not. After characterizing the main features of the potentiation effects, the authors show, through the construction of chimeric channels, that one of the extracellular linkers, the one between transmembrane domains 11-12 is capable of encoding the potentiation phenotype. Further experiments demonstrate that a histidine residue at position 531 in the 11-12 linker is mainly responsible for Zn2+ potentiation of OTOP3 channels.
In an attempt to provide a quantitative explanation for their findings, the authors produce a transition-state gating model that is capable of explaining most of the experimental findings by postulating that gating is influenced by the interaction of two separate binding sites for H+ and Zn2+.
The experiments and analyses performed in this manuscript are of very high quality and the interpretation of the data and conclusions are supported by the data.
This is a contribution that should help understand the emerging physiological roles of this recent addition to the family of proton permeable channels.
Specific comments:
1) In several places in the manuscript, the authors mention that Zn potentiation indicates that the channels are gated, however, isn't activation by H an indication that the channels are are gated? Perhaps this statement should be explained in a more nuanced manner.
2) The solutions employed have a relatively low concentration of buffer, thus the proton currents will incur in proton depletion artifacts, especially for large amplitude currents. Please comment on to what extent the desensitization of the proton currents could be a reflection of proton depletion.
3) The data in figure 3A shows the time course of the wash-off of the Zn effect. What process do the authors think is reflected here? A slow koff for Zn unbinding or a slow rate constant for recovery from the potentiated state?
4) In the methods section, please explain in more detail the procedure to perfuse cells in whole-cell recordings. Were the cells lifted and brought to the perfusion tubes? Or were the cells attached and the surrounding solution changed? If the latter, how complete is the solution exchange to be expected?
5) In the explanation of the model predictions, it is not clear to me how the allosteric model can predict the desensitization of currents. This part needs a more detailed explanation.
Reviewer #2 (Recommendations for the authors):
It was curious that it didn't seem like potentiation could occur in the presence of Zn. Despite the identification of an important His residue and the different concentration response of inhibition vs potentiation I do wonder whether the Zn potentiation is due to a second Zn site or is a special slow allosteric consequence of Zn inhibition dependent on the loop. Zn inhibition and potentiation both begin to occur at 0.3 mM. Perhaps the concentration response of inhibition vs potentiation by Cu could address the question from another angle. Two separate divalent sites might not have the same relative affinities for Zn and Cu.
It was difficult to tell if the description in methods of the kinetic model fully describes the model. There appear to be 24 different rate constants in the model. Perhaps labeling a table listing all the free parameters using a nomenclature clearly addressable to Figure 9A would help readers understand how the model works.
Reviewer #3 (Recommendations for the authors):
– Line 4 typo "vesibular".
– Figure 1B: typo in figure legend "Inhibtion".
– Figure 1B: mOTOP 3 currents: is there a reason for removing the Zn2+ before reaching a steady-state current (the current during Zn2+ application shows the sum of the two different actions but does it reach a steady-state)?
– Figure 5. I suppose that all the divalent ions tested show a block of OTOPs, although data are not shown. Is there a correlation between the potency of inhibition and the potency of activation for each of the ions? For instance, Cu2+ is the most potent activator among the tested ones, does it also block with the highest potency? It would be interesting if, among these transition metals, the authors could isolate one that has different affinities of inhibition/activation to find a concentration suitable to study the two modulation mechanisms independently.
– Figure 6 Tm11-12 linker: from the chicken OTOP3 structure this linker seems to be the longest on the extracellular side. To characterize the properties of this linker to the fullest the authors should test two important aspects: does the length of the linker affect Zn2+ potentiation? This would also partly answer the extent of the interaction of Zn2+ coordination with other parts of the channel. Obviously, a long linker has the flexibility to reach out to residues not in the closest proximity. The second aspect that should be worth testing is if the position of this linker, located between tm 11 and 12, is also important for Zn2+ potentiation. What would happen if the linker is moved around OTOP3 extracellular side? Does the Zn2+ effect correlate only with the linker itself (aminoacidic sequence) or even with its position in the protein? This could also provide insights into the gating mechanism itself.
– Figure 8. Given the position of the loop in the chicken OTOP3 structure (PDB 6NF6), supported by the quality of the electron density map, H531 could potentially flip towards H234 and E238 (mOTOP3 numbering). Those residues are both conserved in chicken and mouse isoforms, so it should be worth testing if they can be part of the zinc-binding site through site-directed mutagenesis, as their position is reasonably close to H531 and not too buried into the membrane.
– Figure 8 Given the position of E535A in the chicken OTOP3 structure I would suggest testing the E535H mutation, to measure possible enhancement of the Zn2+ potentiation.
– Figure 8 Given the effect of H531 in OTOP3 it would be reasonable to wonder what happens when the same position is replaced with a histidine in OTOP2 and OTOP1, both bearing an arginine. Is the histidine sufficient to recapitulate the O3/O2tm11-12 phenotype? Or at least part of it?
– Figure 8 Regarding the previous point: do the authors have an explanation for the OTOP1 potentiation given that this isoform does not have the corresponding H531 (OTOP3 numbering) residue?
– Line 301 "It is tempting to speculate that the binding site spans parts of the channel that are at a distance in the closed state and that come together in the open state as shown for other ion channels" (Gordon and Zagotta, 1995b; Zhang et al., 2009).
The Zn2+ potentiation occurs even during pre-treatment at pH 7.4, a condition in which the channel is closed (Teng et al. 2022). Doesn't this suggest that the binding site is available also during the closed state? The authors suggested direct competition with H+, so that at pH 7.4 Zn2+ can bind to the extracellular side (interacting with the unprotonated H531) transitioning the channel into a state that facilitates pore opening at pH 5.5. If this is the case what is the mechanism of Zn2+ binding when H531 is protonated (experiment at pH5.5 Figure 1B)? I think that this aspect of the mechanism is still a bit elusive and needs clarification.
https://doi.org/10.7554/eLife.85317.sa1Author response
Essential revisions:
After consultation, the reviewers concur on these essential revisions to the manuscript.
1) In the explanation of the model predictions, it is not clear how the allosteric model can predict the desensitization of currents. This part needs a more detailed explanation. Also, it was difficult to tell if the description in methods of the kinetic model fully describes the model. There appear to be 24 different rate constants. Perhaps labeling a table listing all the free parameters using a nomenclature clearly addressable to Figure 9A would help readers understand how the model works.
We apologize for not making the model parameters sufficiently clear. We have updated the methods section on the model to better describe the parameters and how they map onto the scheme in Figure 9A. We note that there are 14 free parameters in the model, 11 of which are likely to be of the most importance. We also note that we did not attempt to optimize the model parameters, in part because there are so many, but instead manually found a parameter set that qualitatively describes our observations. To further help illustrate how the model works, we have included two new Supplementary Figures. Supplementary Figure 2 shows the model’s simulated time course for each individual gating or binding domain, and Supplementary Figure 3 contrasts the simulated time course for the decay of the zinc-potentiated current with that of zinc unbinding from the activating site to show that they are similar (i.e., the current decay can be largely accounted for by zinc unbinding in the model simulation). We also added a few lines describing this conceptually in the model section of the main text.
2) Given the effect of H531 in OTOP3 it is reasonable to wonder if the histidine is sufficient to recapitulate the O3/O2tm11-12 phenotype or at least part of it. The same position should be replaced with histidine in OTOP2 and OTOP1, which have arginines. Related to this point: can the authors provide an explanation for the OTOP1 potentiation given that this isoform does not have the corresponding H531 residue?
The reviewers raise an interesting point. We have now tested the effect of introducing the equivalent residue to H531 in OTOP2 and OTOP1 (new Figure 8, supplementary figure 1). In the case of OTOP2, the channels remain insensitive to potentiation from Zn2+. Thus, H531 is not sufficient in the context of OTOP2 to recapitulate the effect of swapping the entire 11-12 linker. We assume that other residues in the 11-12 linker, or changes in the overall conformation of the linker, are required to confer sensitivity to Zn2+. For OTOP1, we did observe an enhancement in the degree of potentiation for channels with H replacing R, although the channels are not as strongly potentiated as OTOP3 channels. This data is consistent with the notion that the Zn2+ binding site consists of several residues and that their relative contributions to energetically stabilizing Zn2+ in the open/potentiated state may vary among different channels. It also is consistent with our model, which predicts that for a channel that is already activated by H+ (as OTOP1 is at the pH tested), Zn2+ will have a relatively smaller effect. There is certainly more to be done in determining the nature of the Zn2+ activating site, in closed and open conformations of the channels, and within the context of different OTOP isoforms.
3) Please comment on to what extent the desensitization of the proton currents could be a reflection of proton depletion given the large amplitude of the proton currents.
The decay of the currents was not a focus of this manuscript. It likely includes contributions from three factors: Zn2+ unbinding (de-potentiation), channel desensitization and ion accumulation (collapse of the H+ gradient). In a separate manuscript, Teng et al., 2022a, we discuss this question in the context of OTOP1 and find that the decay of the currents is likely due to collapse of the H+ gradients. For the experiments shown with OTOP3, our modeling shows that the decay of the currents can be be explained by unbinding of Zn2+. However, we cannot rule out that there is a component of either desensitization or gradient collapse. We now discuss this point when we describe the model.
Reviewer #1 (Recommendations for the authors):
Specific comments:
1) In several places in the manuscript, the authors mention that Zn potentiation indicates that the channels are gated, however, isn't activation by H an indication that the channels are are gated? Perhaps this statement should be explained in a more nuanced manner.
Wording changed throughout. Line 58; wording changed to indicate that evidence that channels are gated is confirmatory.
2) The solutions employed have a relatively low concentration of buffer, thus the proton currents will incur in proton depletion artifacts, especially for large amplitude currents. Please comment on to what extent the desensitization of the proton currents could be a reflection of proton depletion.
The buffer concentration was 10 mM, which is standard for this type of experiment. Higher concentrations did not change the kinetics of decay of other OTOP channels, in our hands, so we did not try higher concentrations for these experiments. The decay of the currents in these experiments is likely due to unbinding of Zn2+ and recovery of channels from the potentiated state (see essential revision #1 above).
3) The data in figure 3A shows the time course of the wash-off of the Zn effect. What process do the authors think is reflected here? A slow koff for Zn unbinding or a slow rate constant for recovery from the potentiated state?
For the model in Figure 9A these are one and the same as the potentiated state is defined as that having Zn bound to the activating/potentiating site. This model, and thus Zn unbinding, can qualitatively explain the observed wash-off time course. We did not attempt to experimentally distinguish between them so we cannot provide a test of this hypothesis with available data.
4) In the methods section, please explain in more detail the procedure to perfuse cells in whole-cell recordings. Were the cells lifted and brought to the perfusion tubes? Or were the cells attached and the surrounding solution changed? If the latter, how complete is the solution exchange to be expected?
We apologize for leaving out this important detail. The cells were lifted. In similar experiments (Teng et al., 2022), we showed that complete solution exchange takes around 30 ms. This is now explained in the methods, Line 382.
5) In the explanation of the model predictions, it is not clear to me how the allosteric model can predict the desensitization of currents. This part needs a more detailed explanation.
Please see our response to essential revision #1.
Reviewer #2 (Recommendations for the authors):
It was curious that it didn't seem like potentiation could occur in the presence of Zn. Despite the identification of an important His residue and the different concentration response of inhibition vs potentiation I do wonder whether the Zn potentiation is due to a second Zn site or is a special slow allosteric consequence of Zn inhibition dependent on the loop. Zn inhibition and potentiation both begin to occur at 0.3 mM. Perhaps the concentration response of inhibition vs potentiation by Cu could address the question from another angle. Two separate divalent sites might not have the same relative affinities for Zn and Cu.
Actually, we do observe potentiation in the presence of Zn, see Figure 1 C, where one can see a recovery of the currents following an initial rapid block. In the manuscript, we presented several pieces of evidence showing that the Zn blocking site is distinct from the Zn potentiating site. This includes: (1) a difference in dose-dependence shown in Figure 1D, where it can be seen that inhibiton saturates at 3 mM Zn, while activation does not saturate even at 10 mM, implying that the inhibitory site is higher affinity, (2) the observation that all OTOP channels are blocked by Zn but only OTOP3, and to a lesser extent OTOP1, is potentiated and (3) the observation that potentiation can be eliminated in the OTOP3/OTOP2 11-12 chimera, which retains inhibition.
We followed the reviewer’s suggestion to look a little more carefully at inhibition by other transition metals that we found could activate mOTOP3 channels. Thus, we now include new data addressing this point. This data, now shown in Figure 5, shows that Ni, Cd, and Cu all block mOTOP3 channels. Interestingly, Ni produces nearly identical block of the channels as Zn, but much less potentiation, consistent with different metal binding affinities for the two sites.
It was difficult to tell if the description in methods of the kinetic model fully describes the model. There appear to be 24 different rate constants in the model. Perhaps labeling a table listing all the free parameters using a nomenclature clearly addressable to Figure 9A would help readers understand how the model works.
Please see our response to essential revision #1.
Reviewer #3 (Recommendations for the authors):
– Line 4 typo "vesibular".
Fixed
– Figure 1B: typo in figure legend "Inhibtion".
Fixed
– Figure 1B: mOTOP 3 currents: is there a reason for removing the Zn2+ before reaching a steady-state current (the current during Zn2+ application shows the sum of the two different actions but does it reach a steady-state)?
The duration of the Zn2+ stimulus in these experiments was based on the time needed to observe its inhibitory effect and we did not attempt to measure the steady-state activating effects under these conditions. This was done more carefully when we switched to a protocol that allowed us to study activation in the absence of inhibition.
– Figure 5. I suppose that all the divalent ions tested show a block of OTOPs, although data are not shown. Is there a correlation between the potency of inhibition and the potency of activation for each of the ions? For instance, Cu2+ is the most potent activator among the tested ones, does it also block with the highest potency? It would be interesting if, among these transition metals, the authors could isolate one that has different affinities of inhibition/activation to find a concentration suitable to study the two modulation mechanisms independently.
This is an excellent point. As described in the response to Reviewer 2, we now show that Ni, Cd and Cu (1 mM) all block mOTOP3 currents (pH 5.5). Thus, we could not find a metal ion that activates the currents without blocking them. There was no direct relationship between potency of inhibition and activation, and Ni which produced very little activation was as good a blocker as Zn.
– Figure 6 Tm11-12 linker: from the chicken OTOP3 structure this linker seems to be the longest on the extracellular side. To characterize the properties of this linker to the fullest the authors should test two important aspects: does the length of the linker affect Zn2+ potentiation? This would also partly answer the extent of the interaction of Zn2+ coordination with other parts of the channel. Obviously, a long linker has the flexibility to reach out to residues not in the closest proximity. The second aspect that should be worth testing is if the position of this linker, located between tm 11 and 12, is also important for Zn2+ potentiation. What would happen if the linker is moved around OTOP3 extracellular side? Does the Zn2+ effect correlate only with the linker itself (aminoacidic sequence) or even with its position in the protein? This could also provide insights into the gating mechanism itself.
The tm11-12 linkers are exactly the same length for mOTOP2 and mOTOP3, so length of the linker cannot explain the difference in sensitivity to Zn2+ potentiation. We have no doubt that changing the length of the linker could change function, but it is not clear what one could conclude from this experiment. Similarly, we do not propose that the linker on its own, irrespective of protein context, would potentiate the channels, and thus, moving the linker or swapping linkers would be extremely unlikely to open (potentiate) the channels.
– Figure 8. Given the position of the loop in the chicken OTOP3 structure (PDB 6NF6), supported by the quality of the electron density map, H531 could potentially flip towards H234 and E238 (mOTOP3 numbering). Those residues are both conserved in chicken and mouse isoforms, so it should be worth testing if they can be part of the zinc-binding site through site-directed mutagenesis, as their position is reasonably close to H531 and not too buried into the membrane.
Excellent point. We tested mutations of H234 and E238 to alanine and found they reduced potentiation. A double mutation had a more severe effect (although it did not completely eliminate potentiation like H531A) which suggests that the four residues may constitute the Zn2+ (metal ion) potentiating site. We now include this data in a new figure 8. Future structural studies will be necessary to confirm that metal ions bind to this site and to understand and how metal binding stabilizes an open state.
– Figure 8 Given the position of E535A in the chicken OTOP3 structure I would suggest testing the E535H mutation, to measure possible enhancement of the Zn2+ potentiation.
Indeed, this might be an interesting mutant to test, but we do not believe the results, however it would turn out, would impact the conclusion of our study, and for this reason we did not prioritize is among the many mutants we could generate and test.
– Figure 8 Given the effect of H531 in OTOP3 it would be reasonable to wonder what happens when the same position is replaced with a histidine in OTOP2 and OTOP1, both bearing an arginine. Is the histidine sufficient to recapitulate the O3/O2tm11-12 phenotype? Or at least part of it?
See responses to essential reviews.
– Figure 8 Regarding the previous point: do the authors have an explanation for the OTOP1 potentiation given that this isoform does not have the corresponding H531 (OTOP3 numbering) residue?
This is an interesting question. We did find that mutation to H increased the degree of potentiation, but it is still curious that the positive charge on the arginine would not disrupt potentiation in the context of mOTOP1 as it does in the context of mOTOP3.
– Line 301 "It is tempting to speculate that the binding site spans parts of the channel that are at a distance in the closed state and that come together in the open state as shown for other ion channels" (Gordon and Zagotta, 1995b; Zhang et al., 2009).
The Zn2+ potentiation occurs even during pre-treatment at pH 7.4, a condition in which the channel is closed (Teng et al. 2022). Doesn't this suggest that the binding site is available also during the closed state?
There are two mechanisms by which Zn2+ could bind to the channel at pH 7.4. It could either bind to the closed state (which is something that our model predicts can happen at a low rate) or the channel can briefly transit through an open state (at low probability) from which Zn2+ will more rapidly bind. In either case, unbinding is predicted to be slow. This is consistent with the slow apparent on rate of Zn2+ at pH 7.4 (many seconds) and slow off rate (>1 s). Essentially, Zn2+ traps the channel in an open state.
The authors suggested direct competition with H+, so that at pH 7.4 Zn2+ can bind to the extracellular side (interacting with the unprotonated H531) transitioning the channel into a state that facilitates pore opening at pH 5.5. If this is the case what is the mechanism of Zn2+ binding when H531 is protonated (experiment at pH5.5 Figure 1B)? I think that this aspect of the mechanism is still a bit elusive and needs clarification.
For the experiment where the pH is 5.5, and the His residues is protonated, we need to presume that Zn2+ can bind, perhaps weakly, due to interactions with other residues, and that binding of Zn2+ causes the deprotonation of HIs531, which then may stabilize the Zn2+ ion. This is speculation, and we agree with the reviewer that there is still some question regarding the exact physical mechanism by which Zn binds to closed vs. open states of the channels and stabilizes the open state. We did not presume to have defined the precise role of H531 and other residues in this process, which may require structural information.
https://doi.org/10.7554/eLife.85317.sa2Article and author information
Author details
Funding
National Institutes of Health (R01GM131234)
- Emily R Liman
National Institutes of Health (R01GM148591)
- Marcel P Goldschen-Ohm
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.
Acknowledgements
We thank Jackson Walker and Anne Tran for expert technical support, Carly Hamel and Roei Zakut, for help in generating mutant channels, all members of the Liman lab for helpful discussions, and all the reviewers of the manuscript for their constructive comments, including the suggestion to include an assessment of contributions of residues in the tm 5–6 linker. This research was supported by NIH grants R01GM131234 to ERL and R01GM148591 to MG-O.
Senior Editor
- Kenton J Swartz, National Institute of Neurological Disorders and Stroke, National Institutes of Health, United States
Reviewing Editor
- Leon D Islas, Universidad Nacional Autónoma de México, Mexico
Reviewers
- Leon D Islas, Universidad Nacional Autónoma de México, Mexico
- Jon T Sack, University of California, Davis, United States
Version history
- Received: December 2, 2022
- Preprint posted: December 18, 2022 (view preprint)
- Accepted: March 27, 2023
- Version of Record published: April 13, 2023 (version 1)
- Version of Record updated: April 14, 2023 (version 2)
Copyright
© 2023, Teng, Kaplan et al.
This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.
Metrics
-
- 408
- Page views
-
- 49
- Downloads
-
- 0
- Citations
Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.
Download links
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Further reading
-
- Neuroscience
- Structural Biology and Molecular Biophysics
Ca2+/calmodulin-dependent protein kinase II (CaMKII) is essential for long-term potentiation (LTP) of excitatory synapses that is linked to learning and memory. In this study, we focused on understanding how interactions between CaMKIIα and the actin-crosslinking protein α-actinin-2 underlie long-lasting changes in dendritic spine architecture. We found that association of the two proteins was unexpectedly elevated within 2 minutes of NMDA receptor stimulation that triggers structural LTP in primary hippocampal neurons. Furthermore, disruption of interactions between the two proteins prevented the accumulation of enlarged mushroom-type dendritic spines following NMDA receptor activation. α-Actinin-2 binds to the regulatory segment of CaMKII. Calorimetry experiments, and a crystal structure of α-actinin-2 EF hands 3 and 4 in complex with the CaMKII regulatory segment, indicate that the regulatory segment of autoinhibited CaMKII is not fully accessible to α-actinin-2. Pull-down experiments show that occupation of the CaMKII substrate-binding groove by GluN2B markedly increases α-actinin-2 access to the CaMKII regulatory segment. Furthermore, in situ labelling experiments are consistent with the notion that recruitment of CaMKII to NMDA receptors contributes to elevated interactions between the kinase and α-actinin-2 during structural LTP. Overall, our study provides new mechanistic insight into the molecular basis of structural LTP and reveals an added layer of sophistication to the function of CaMKII.
-
- Neuroscience
- Structural Biology and Molecular Biophysics
Interactions between an enzyme kinase, an ion channel and cytoskeletal proteins maintain the structure of synapses involved in memory formation.