The First Extracellular Linker Is Important for Several Aspects of the Gating Mechanism of Human TRPA1 Channel

Abstract

Transient receptor potential ankyrin 1 (TRPA1) is an excitatory ion channel involved in pain, inflammation and itching. This channel gates in response to many irritant and proalgesic agents, and can be modulated by calcium and depolarizing voltage. While the closed-state structure of TRPA1 has been recently resolved, also having its open state is essential for understanding how this channel works. Here we use molecular dynamics simulations combined with electrophysiological measurements and systematic mutagenesis to predict and explore the conformational changes coupled to the expansion of the presumptive channel's lower gate. We show that, upon opening, the upper part of the sensor module approaches the pore domain of an adjacent subunit and the conformational dynamics of the first extracellular flexible loop may govern the voltage-dependence of multimodal gating, thereby serving to stabilize the open state of the channel. These results are generally important in understanding the structure and function of TRPA1 and offer new insights into the gating mechanism of TRPA1 and related channels.

Introduction

The transient receptor potential ankyrin 1 (TRPA1) channel responds to many environmental irritants, electrophilic compounds and proalgesic agents by opening an intrinsic ion channel gate. Specifically, in nociceptor neurons the gating of this cation channel leads to a membrane depolarization and increased intracellular calcium, which in turn modulates its own activity (Story et al., 2003; Nilius et al., 2012; Zygmunt and Hogestatt, 2014). How all the activation stimuli are conveyed through the TRPA1 protein complex to the channel's gate is still unresolved. New high-resolution structures of TRPA1 captured in a closed-pore conformation, compared with the structures of two related channels, TRPV1 and TRPV2, in open, closed and desensitized states, provide exciting clues about the common principles of TRP channel functioning (Cao et al., 2013; Liao et al., 2013; Brewster and Gaudet, 2015; Paulsen et al., 2015; Gao et al., 2016; Zubcevic et al., 2016). All these channels consist of four subunits, each containing six transmembrane segments (S1–S6) flanked by N- and C-terminal cytosolic domains, a sensor module S1–S4, and an ion conduction pore formed between S5 and S6 (Figure 1A). A comparative sequence analysis of the transmembrane regions from almost three thousand different TRP proteins reveals evolutionarily conserved features that indicate a unified model of gating mechanism (Palovcak et al., 2015). This model proposes that dynamic changes in the hydrogen-bonding network around an inherent helical distortion in the middle of S6 are translated into gating by variations in the orientation of the pore-lining residues at the hydrophobic gate. However, the statistically identifed networks of coupled residues only indicate those regions where hundreds of diverse cues that activate the large and functionally versatile family of TRP channels converge. It is known that the energetics of TRP channel gating is regulated from many interdigitated regions outside the membrane domain. Specifically, TRPA1 is unique among other TRP channels in possessing a specific modality of electrophilic activation, an even more remarkable gating promiscuity, and several distinctive structural features: its cytoplasmically located N-terminus is unusually long and its outer pore domain contains two pore helices instead of the typical solitary one present in TRPV1, TRPV2, and in the TRP-related family of voltage-gated potassium (K_V) channels (Paulsen et al., 2015). The structures of TRPV1 in distinct functional states demonstrate a different location of the pore helix in each conformation, suggesting that it is a highly dynamic domain. In contrast, the outer pore region of TRPA1 may not be flexible enough to constitute a regulated gate due to the additional pore helix. Another apparent difference related to the presumptive gating mechanism resides in the existence of a short linker connecting S1 and S2. In TRPV1, the S1–S2 linker is static and lying atop the S1–S4 bundle (Cao et al., 2013). In TRPA1, this region extends into the extracellular space, and a poorly resolved density map in this location suggests that it possesses intrinsic structural flexibility, and thus may be actively involved in channel functioning. An extensive screening of a recombinant membrane-tethered spider toxin library recently identified a 35-residue peptide protoxin-I that antagonizes human TRPA1 through the S1–S2 linker with a high-affinity, and this interaction site is transplantable into the protoxin-I insensitive NaChBac bacterial voltage-gated sodium channel (Gui et al., 2014). Interestingly, the outer pore domain of this channel contains two pore helices like TRPA1 (Zhang et al., 2012).

Figure 1

2004). The height of the particular amino acid at each position indicates its probability of occurance at that position. P742, G743, and N747 are conserved unanimously.

In K_V channels, the upper part of the S1 helix forms a functionally important coevolved contacting interface between the sensor module and the pore helix of an adjacent subunit (Lee et al., 2009). The contribution of the S1–S2 linker is presumably more specific and includes stabilization of the above contacting interface (Füll et al., 2013) or N-linked glycosylation at the NXS/T consensus sequence (Zhu et al., 2003; Schwetz et al., 2010). The question of the rigidity vs. flexibility of the S1–S2 linker has been addressed in structural studies, which demonstrated that the middle section of S1–S2 contains a short helix interacting with S4 in the K_V1.2–K_V2.1 paddle chimera channel (Long et al., 2007), whereas this region is an extended loop in the refined structure of K_V1.2 (Chen et al., 2010).

The amino acid sequence of the S1–S2 linker of TRPA1 is conserved in mammals but different in birds and other non-mammalian organisms (Figures 1B,C), suggesting a role for lineage-specific variations in mediating its presumably regulatory functions. We now know that from this specific region, the channel can be stabilized in its closed conformation by the specific peptide and this mode of inhibition is state-dependent (Gui et al., 2014). From the density map, we also know that this region may be intrinsically flexible. Thus, what could the functional role of the S1–S2 linker be in human TRPA1? What role, if any, does the first extracellular flexible loop play during electrophile-, voltage- and calcium-evoked channel gating? The signals generated by these stimuli likely converge into one allosteric step leading to the opening of the inner gate formed by the two hydrophobic S6 residues I957 and V961 (Paulsen et al., 2015). But how does a channel that opens in this way communicate with the upper part of the protein complex? Does it need an anchor point between the S1–S4 sensor and the pore module similar to that in structurally related voltage-gated potassium channels? Is an N-linked glycosylation at the two putative sites predicted within the S1–S2 linker involved in the gating of TRPA1? In this study, we addressed the issue of whether the opening of the TRPA1 channel may be accompanied by conformational changes near the extracellular surface. We show by modeling, electrophysiology and mutagenesis the importance of the S1–S2 extracellular linker for TRPA1 channel functioning.

Results

Modeling the open state of TRPA1 channel

To gain initial insights into the nature of the mechanical coupling between the lower gate (i.e., amino acids I957 and V961) and the sensor S1–S4 domain (i.e., amino acids S725-L850), we modeled the opening of the TRPA1 channel by using the available structures of TRPA1 in its closed conformation (3J9P; Paulsen et al., 2015) and TRPV1 in its open conformation (3J5Q; Cao et al., 2013). Over the course of Molecular Dynamics Flexible Fitting (MDFF) simulation (for a detailed description of MDFF procedure see Materials and Methods), a marked expansion of the TRPA1 pore occurred in the lower gate region (Video 1, Figures 2A,B). The distances between the alpha-carbon atoms of I957 isoleucines from diagonally positioned subunits of the channel homotetramer increased from 11.6 to 13.1 Å, and the distances between alpha-carbon atoms of V961 valines increased from 10.6 to 13.0 Å (Figure 2C).

Figure 2

The expansion of the lower gate was accompanied by a conformational change and switching of amino acids on the extracellular side, where the S1–S4 sensor approaches the S5–S6 pore domain of an adjacent monomer. Helices S1, S4, and S5 come into close contact—specifically, the amino acids of helices S1 (V737), S4 (W832), and S5 (I890) (Figures 2D–G). Proline at position 742 situated at the beginning of the S1–S2 linker is in close contact with this hydrophobic switch and mediates contacts between the S1–S4 sensor module and the pore-forming domain. Among the hydrophobic amino acids within the S1–S2 linker, phenylalanine F746 (shown in Figure 3) was constantly oriented into the protein bulk and seemed to significantly contribute to stabilization of the sensor module. The results from this MDFF simulation indicate that the extracellular S1–S2 linker may be physically coupled with the conformational changes in the lower gate, and may thus be an important regulator of TRPA1 channel gating.

Figure 3

2015, depicted in salmon color). (B) Electron density map EMD-6267 of TRPA1 superimposed on the S1–S4 sensor domain and two predicted S1–S2 linker conformations I and II (blue and red). (C,D) Details of the S1–S4 sensor domain shown in (A). (C) Conformation I (blue) represents the stable hairpin stabilized by hydrophobic residues. (D) Conformation II (red) interacts with the upper part of the sensor module via hydrophobic residues. Phenylalanine F746 is situated in the middle of the upper part of the S1–S4 sensor domain. (E) The extent of S1–S2 linker flexibility for both conformations represented as a putty representation of the protein colored according to average B-factor was prepared in Chimera using the Render by Attribute menu under Tools and Structure Analysis menu. The diameter and color of the S1–S2 linker are scaled to the maximum B-factor. Orange to red colors and a wider tube indicate a region with a higher average B-factor, whereas shades of yellow and a narrower tube represent a region with a lower average B-factor. Proline P742 at the beginning of the S1–S2 linker is shown.

Structural modeling of S1–S2 linker indicates two distinct conformations stabilized by hydrophobic interactions

To gain further mechanistic insight into the putative role of the S1–S2 linker (amino acids S748-D763), we performed a 500-ns molecular dynamic (MD) simulation for the S1–S4 sensor module (i.e., amino acids S725-L850) in aqueous solution (see Materials and Methods for details). The resulting models were structurally aligned with the published TRPA1 structure 3J9P and compared with the density map of TRPA1 EMD-6267 (Paulsen et al., 2015). The simulation revealed two general conformations of the S1–S2 loop (Figure 3A). A remarkably stable hairpin conformer (conformation I) was found over time from 262 to 412 ns, fitting part of the density well (Figure 3B). In the rest of the MD simulation (from 37.5 to 112.5 ns), the S1–S2 loop was conformationally less pronounced (conformation II) and corresponded to the other part of the electron density map for this region.

The MD simulation indicates that the electron density maps of TRPA1 show a mixture of two conformers. The first one, corresponding to the hairpin structure, appears to be significantly stabilized by hydrophobic interactions between amino acids I751, I752, I761, and L762 (Figure 3C). The second one is a mixture of sub-conformers, wherein the above-mentioned hydrophobic amino acids interact with the upper part of the sensor module S1–S4, including F746 (Figure 3D). Our structural simulations thus substantiate the intrinsic flexibility of the S1–S2 linker (Figure 3E) and also indicate that this region is capable of adopting conformations that can effectively influence the changes carried out by S1–S4 helices during channel gating, presumably through steric interactions of the hydrophobic part of the linker.

Isoleucines I751 and I752 regulate the voltage-dependent activation of TRPA1

The hydrophobic cluster that substantially contributes to the two general conformations of the S1–S2 linker includes a prominent couple of adjacent isoleucines I751 and I752. To explore their presumptive structural significance, we substituted both the residues simultaneously with a smaller and less hydrophobic alanine (I751A/I752A) or an even smaller and flexible glycine (I751G/I752G). The voltage sensitivity was examined by using a voltage step protocol from −80 mV up to +200 mV in extracellular control solution (Figure 4A). For wild-type TRPA1, the Boltzmann fit gave a half-maximal activation voltage (V_1/2) of 127.2 ± 1.9 mV and an apparent number of gating charges (z) of 0.74 ± 0.02 e₀ (n = 121) which agrees with the previously published data (Sura et al., 2012). Compared to wild-type TRPA1, both the double-mutant channels exhibited lower sensitivity to voltage, and their half-activation potential V_1/2 was shifted toward more positive membrane potentials (Figure 4B). To test their chemical sensitivity, we used a protocol in which whole-cell membrane currents were measured in the presence of the full electrophilic agonist allyl isothiocyanate (AITC, 100 μM) in extracellular control solution. The membrane potential was ramped up each second from −80 to +80 mV (Figure 4C). The double mutants exhibited much slower kinetics of AITC-induced responses and a significant reduction in currents at negative membrane potentials (Figures 4D,E and Supplementary Figure 1A). As was previously shown, current-voltage (I-V) relationships of wild-type TRPA1 responses to AITC are linear at the peak, and become less linear as they inactivate (Wang et al., 2008; Kremeyer et al., 2010). In the I751A/I752A mutant, and even more remarkably in I751G/I752G, the I-V relationships exhibited strong outward rectification at the peak of AITC responses (Figures 4D,E, graphs below current traces). The glycine at positions 751 and 752 caused more pronounced effects than alanine, which indicates that the volume of residues and local conformational flexibility may be important for full activation of the channel–specifically at hyperpolarized voltages. Consistent with this finding, the double mutants exhibited lower basal activity and much smaller average responses than wild-type TRPA1 when we measured the AITC-induced activity by a Ca²⁺-imaging technique (F₃₄₀/F₃₈₀; Supplementary Figures 1B,C).

Figure 4

I751G/I752G mutant exhibits different rectification pattern

The above experiments were done in the presence of extracellular control solution which contained 1 mM external Ca²⁺. Calcium ions permeating through the TRPA1 channel exert bimodal effects, both potentiating the current and accelerating its inactivation (Wang et al., 2008). To examine the contribution of Ca²⁺, we used a protocol in which whole-cell membrane currents were measured first in the absence of extracellular Ca²⁺ and in the presence of the full agonist AITC (100 μM, 30 s). Ca²⁺ at a concentration of 1 mM was then added to the bath solution. As before, intracellular Ca²⁺ was buffered to low levels with 5 mM EGTA in the patch pipette to assess the effects of permeating Ca²⁺ ions (Wang et al., 2008). This protocol enabled us to explore not only the AITC sensitivity of the channels to an electrophilic agonist that leads to the persistent activation of TRPA1 (Hinman et al., 2006; Macpherson et al., 2007), but also to permeating Ca²⁺ that modulates the channel through various mechanisms (Doerner et al., 2007; Zurborg et al., 2007). In wild-type TRPA1, AITC in the absence of Ca²⁺ elicited large membrane currents at both membrane potentials which came close to saturation in 30 s (Figure 5A). The subsequent addition of Ca²⁺ evoked a mild potentiation followed by inactivation. In contrast, the I751G/I752G mutant exhibited small responses to AITC in the absence of Ca²⁺ at depolarizing potentials, and no currents were seen at negative potentials (Figure 5B). Calculations of the rectification index (R = −current at −70 mV/current at +70 mV) showed clear differences between wild-type and I751G/I752G channels when exposed to AITC (Figures 5C–F), indicating that the mutation modified the channel's behavior in a voltage-dependent manner.

Figure 5

Impact of the particular physico-chemical properties of residues replacing I751 and I752

To further determine the specificity of the regulation of TRPA1 by the S1–S2 linker, we examined the importance of the physico-chemical properties of various amino acid residues substituted at positions I751 and I752. We explored residues with different hydrophobicities or volumes and measured the sensitivity of the double mutants to depolarizing voltages (Figures 6A,B) and to AITC in the absence or presence of Ca²⁺ (Figures 6C,D and Supplementary Figure 2). There were differences among mutants and wild-type channels in terms of their voltage-dependent gating and the kinetics of their AITC-induced responses. I751P/I752P and I751F/I752F did not differ from the wild type in the peak of AITC responses at both potentials (Figure 6C), however, the time courses of rectification exhibited different kinetics (Figure 6D). Whereas, the former double mutant exhibited a slower onset of the rectification index in AITC and a faster and more complete inactivation at negative potentials after the addition of Ca²⁺, the latter mutations led to opposite effects: a faster onset and a slow and incomplete Ca²⁺-dependent inactivation. When comparing the functional impacts of mutations with the properties of the substituting amino acids (Supplementary Figure 2F), it becomes evident that a specific combination of the volume, hydrophobicity and flexibility imparted by the isoleucines at positions 751 and 752 is critical for the voltage-dependent gating of TRPA1.

Figure 6

I751G/I752G mutant is distinctly blocked by external calcium

To further explore how the S1–S2 region may contribute to chemical activation apart from any allosteric effects of depolarizing voltage, we measured currents from wild-type and I751G/I752G channels in response to AITC applied either in the absence or presence of 1 mM Ca²⁺ at a constant voltage of −70 mV. Two different protocols were used. In the first case, 100 μM AITC in a Ca²⁺-containing bath solution was applied to the wild type or to the I751G/I752G double mutant. As shown in Figure 7, wild-type TRPA1 responded to AITC with robust and fast inward currents followed by rapid inactivation (Figure 7Aa). Removing Ca²⁺ from the extracellular solution immediately (in < 1 s) unblocked monovalent cation flow through the channels (Karashima et al., 2010). If the external Ca²⁺ was removed once again 20 s later, the recovery from the block was not complete and a slowly progressing component of inactivation was apparent. In the second protocol, TRPA1 was first activated by AITC without Ca²⁺, which evoked a slowly rising response and then Ca²⁺ was added 30 s later (Figure 7Ab). The addition of Ca²⁺ caused a rapid increase in currents followed by Ca²⁺-dependent inactivation. The subsequent removal of external Ca²⁺ only partially recovered the channels from inactivation. This finding demonstrates that Ca²⁺, in addition to its well-described potentiating and inactivating effects, reversibly blocks the flow of monovalent cations. The extent of this block is state-dependent, so the less inactivated channels are more blocked. A similar blocking effect of external Ca²⁺ was also apparent for a protocol in which the membrane potential was ramped up each second from −80 to +80 mV (1 V/s) (Figure 7B). Under the same protocols, I751G/I752G exhibited qualitatively different behavior. The channels appeared to be blocked by Ca²⁺ right from the start, and were only transiently unblocked by removing external Ca²⁺ (Figure 7Ac). By using a protocol in which AITC was initially applied without Ca²⁺, we observed very slowly developing responses (Figure 7Ad). The addition of 1 mM Ca²⁺ potentiated and inactivated the channels as in the wild type. However, in contrast to the wild-type, the channels recovered fully from the block after the removal of Ca²⁺ from the bath (compare Figures 7Ab,d), indicating a perturbed gating equilibrium in favor of the closed state. The removal of external Ca²⁺ had immediate and reversible effects on TRPA1, which suggests an external site of Ca²⁺ action. The S1–S2 linker of human TRPA1 has four negatively charged amino acids which could be involved in Ca²⁺ binding. However, when we neutralized all of them (E754Q/D757N/E760Q/D763N), we did not see any significant changes in either the voltage or the chemical responsiveness (Figures 7C,D). These experiments thus more suggest an important contribution of the hydrophobic residues in the S1–S2 linker in Ca²⁺-dependent regulation and further support their active role in regulating the gating equilibrium of the TRPA1 channel.

Figure 7

S1–S2 linker mediates TRPA1 activation at hyperpolarized potentials

The molecular simulations revealed that, during in-silico opening, phenylalanine F746 is constantly oriented toward the sensor module and forms stable interactions with the very beginning of the S1–S2 linker and the upper part of the S1–S4 bundle (Figure 3D). In the TRPA1-related TRPM3 channel, this region forms an alternative ion permeation pathway, distinct from the central pore, which can be uncovered by the specific mutation of an aromatic residue at the top of S4 (Vriens et al., 2014). Although the upper crevice of the S1–S4 sensor is likely to be unsolvated in TRPA1 (Palovcak et al., 2015), F746 is located right in its center and interacts, among others, with W832 from upper S4. This led us to test whether replacing F746 with a small residue would cause changes in the channel's functioning or would enable the formation of an ion permeation pathway analogous to that in TRPM3. Figure 8A shows that the mutation F746A strongly shifted V_1/2 to more positive potentials and significantly increased the gating charge. Compared to the wild type, the AITC-induced currents recorded in the extracellular solution containing 1 mM Ca²⁺ exhibited a delayed time of peak and markedly reduced responses at negative membrane potentials (Figure 8B). This reduction was also observed when AITC was applied in the absence of external Ca²⁺ (Figure 8C). At positive potentials and in the presence of Ca²⁺, the maximum AITC currents were no different from those of wild-type TRPA1. All these observations suggest an active role of F746 in voltage-dependent gating, instead of physically plugging a putative alternative pore. Hydrophobic packing around F746 in the upper part of the S1–S4 module seems to maintain the activity of the channel at negative membrane potentials.

Figure 8

2014) are indicated. Conformation II of the S1–S4 sensor modules with the S1–S2 linkers (S748-D763, yellow) were structurally aligned. Nitrogens from POPC bilayer are shown in blue. (F) Detail from (E) showing binding surface (pink) mapped onto the protoxin-I structure, amino acid residues involved in binding to TRPA1 are indicated. Subunits are distinguished by gray shades.

We next hypothesized that protoxin-I, which inhibits wild-type TRPA1 by binding primarily to the S1–S2 linker but also to other parts of the S1–S4 extracellular surfaces (Gui et al., 2014), might help to explain the functional impact of the F746A mutant. Surprisingly, instead of inhibiting, 1 μM protoxin-I increased the AITC-induced activity of F746A at negative membrane potentials and the rectification ratio (R = −I_−70mV/I_+70mV) became almost immediately constant and close to 1 (Figure 8D), indicating that the binding of protoxin-I may compensate for the intrinsically lower voltage sensitivity of the mutant. This result further supports our predictions that there is a functional relationship between the S1–S4 module and the S1–S2 linker that is substantially involved in the voltage-dependent gating of TRPA1. To predict potential sites of interaction between the transmembrane part of TRPA1 (i.e., amino acids A685-K989) and protoxin-I (Figures 8E,F), we used molecular docking and molecular dynamics simulations (see Materials and Methods for details) taking into account the homology of protoxin-I with the structurally similar vanillotoxin DkTx that binds the pore region of TRPV1 and traps it in its open state (Cao et al., 2013; Bae et al., 2016). The results of this MD simulation showed that, indeed, protoxin-I is capable of binding the S1–S4 sensor of TRPA1. In addition, the simulation confirmed that the S1–S2 linker is an extracellular loop and its conformational preferences predicted above were most likely not affected by the absence of phospholipids. By mapping the known active surface of protoxin-I (Gui et al., 2014) onto the upper part of the S1–S4 sensor and the predicted set of conformations of the S1–S2 linker, we hypothesize that the linker could act as a pivot component in the gating apparatus of TRPA1, which moves against the channel core to affect (or stabilize) its conformations.

To obtain the complete function-based characteristics of the S1–S2 linker, we next measured the responses to voltage and AITC of several additional mutants that were designed to affect their presumable contribution to flexibility, hydrophobicity, charge and N-glycosylation (Figure 9). The relationship between two experimental parameters for voltage-dependent gating were summarized for all the mutants in Figure 9B: the maximum amplitude of currents induced by depolarizing voltage at +200 mV and the relative inward currents induced after 30-s exposure to AITC in the absence of Ca²⁺, measured at −70 mV. This relationship enables us to qualitatively correlate the changes in voltage-dependent gating measured under control conditions with the changes in allosteric coupling between agonist interaction site(s) and the voltage-dependent opening of the channel gate. The graph revealed four different phenotypes of mutants: (1) most of the mutants were negatively correlated with similar values to wild-type TRPA1, from the less voltage-dependent T755A to the gain-of-function mutant G750I, (2) mutations at serines S748 and S756 increased the relative amplitude of the AITC-induced inward currents, (3) the single or double alanine mutants of P742, G743, I751, and I752 had very low voltage-dependent responses at +200 mV, and even lower amplitudes of AITC-induced inward currents and (4), F746A and the double mutant N747T/N753T were clearly set apart from the other constructs, indicative of their specific functional roles.

Figure 9

Mutations at two predicted N-glycosylation sites decrease TRPA1 expression

There are two predicted N-glycosylation sites on the S1–S2 extracellular linker of human TRPA1 according to NetNGlyc 1.0 Server (⁷⁴⁷NSTG and ⁷⁵³NETS), however, there are no consistent reports about the glycosylation state of this channel (Deering-Rice et al., 2015; Egan et al., 2016). In our expression system, the N747T/N753T mutant obviously differed from other less responsive mutants in that its AITC responses were much less rectified at negative membrane potentials (Supplementary Figures 3A–D). To determine whether these changes might stem from changes in N-glycosylation, the cDNA of GFP-tagged TRPA1 or its double mutant N747T/N753T were transiently transfected into HEK293T cells and cell lysates were incubated in the absence or presence of peptide-N-glycosidase F (PNGase F) at 37°C for 3 h (Supplementary Figure 3E). Immunoblot analysis showed that the double mutant achieved in average only 77% of the total wild type protein expression (Supplementary Figure 3F), which roughly corresponds to the reduction in currents. The dominant molecular weight of the double mutant treated with PNGase F was ~140 kDa, which was no different from the wild type. Most likely, in our expression system, TRPA1 does not undergo substantial N-linked glycosylation and the effect of the double mutation N747T/N753T could be attributed to structural changes in the S1–S2 loop that are crucial for proper expression. This is further supported by our two findings: (1) Other mutations disrupting the N-glycosylation consensus sites (S748A and S756A/G) rendered a distinct phenotype of channels with significantly smaller responsiveness to depolarizing voltage but larger and significantly less rectifying AITC currents (Figure 9B). (2) When the transiently transfected HEK293T cells were grown for 24 h in the presence of 5 μg/ml tunicamycin, a drug that prevents the N-glycosylation of proteins, the amplitudes and the time course of rectification of TRPA1-mediated AITC-responses were no different from control measurements (Figure 9A, Supplementary Figures 3C,D).

S1–S2 may distinctly modulate voltage-dependent gating in mammals and birds

The results from molecular modeling support the previous hypothesis that the gating of TRP channels could share a conserved mechanism involving an external contacting interface between the sensor and the pore of an adjacent subunit (Palovcak et al., 2015). The proline P742 is a structurally unique residue that is part of the S1–S2 linker whose sequence from P742 to T755 is highly conserved in mammals but differs from bird species (Figure 1B). If the S1–S2 linker of human TRPA1 is involved in gating, we should expect differences in the voltage-dependent gating properties of chicken TRPA1. Compared to human TRPA1, chicken TRPA1 exhibited a shift in the conductance-to-voltage relationship toward more depolarizing potentials (Figures 10A,B). The less voltage-dependent phenotype of chicken TRPA1 was also detectable as a dramatic reduction in tail currents during repolarization to −70 mV from previous depolarizing steps and by an increase in the voltage-independent component (Figure 10C). The AITC currents through the chicken TRPA1 reached a peak amplitude comparable with that obtained with human TRPA1 at +70 mV, but were much slower and exhibited a lower peak at negative potentials (Figure 10D, Supplementary Figure 4). The maximum rectification R_max for chicken TRPA1 measured at the peak of AITC-induced responses was ~0.5, which means that the channels were far from being fully activated at hyperpolarized potentials (Figure 10D, lower plot). Also, a 30-s application of AITC in the absence of Ca²⁺ and then the addition of Ca²⁺ to the bath solution, did not lead to full activation at negative potentials (R_max = 0.49) (Figure 10E). Apart from the proximal N- and distal C-terminal regions, the lowest sequence similarity between chicken and human TRPA1 is constrained to the S1–S2 linker, which exhibits only a 35% similarity, compared to the rest of the protein (79%). Therefore, it seems plausible that amino acid differences in the extracellular S1–S2 region underlie, at least partially, the observed functional difference between human and chicken channels.

Figure 10

Discussion

The current understanding of how TRP channels gate is based on the recent structural analysis of the TRPA1, TRPV1, and TRPV2 channels (Cao et al., 2013; Liao et al., 2013; Paulsen et al., 2015; Zubcevic et al., 2016). From comparisons among these structures, it appears that the overall gating mechanisms may be partially shared (Palovcak et al., 2015), however, the precise regulatory domains differ. The high sequence variability among species observed in the S1–S2 extracellular linker, surrounded by the largely conserved transmembrane core, is one of the distinctive features of TRPA1, suggesting that this domain may have a unique biological function. This study was ultimately motivated by this enigmatic feature and has six important implications for understanding the gating mechanisms of TRPA1.

First, our results demonstrate that the human TRPA1 channel has evolved an external regulatory domain that must specifically interact with the upper part of the sensor module to stabilize the channel's open state when activated at hyperpolarized potentials. The predicted conformations of the S1–S2 region include structurally important hydrophobic residues that must be properly positioned relative to the sensor as well as to the pore module. The hairpin structure (Figure 3C) likely corresponds to the closed state of the channel, which seems to be more accessible for the specific toxin (Gui et al., 2014). The other conformation (Figure 3D) helps to stabilize the open state, as inferred from our findings that hydrophobic interactions are important for the proper functioning of the channel. This view is also supported by our observation that replacing serines S748 and S756 with hydrophobic residues improved AITC-induced responses (Figure 9).

Second, the impaired voltage-dependent gating in the chicken TRPA1 orthologue, strikingly resembling impaired responses through the F746A mutant of human TRPA1 (compare Figures 8B, 10D), raises the possibility that the S1–S2 linker of chicken TRPA1 lacks the specific contacts required for electromechanical coupling between the sensor and the pore. Given the high degree of sequence variability among species, we hypothesize that this extracellular domain is critical for the species-specific regulation of TRPA1 voltage-dependent gating.

Third, our results may help to interpret the recent cryo-electron microscopic structure of TRPA1. The original density map indicated a poorly resolved α-helical structure in the S1–S2 linker that extends into the extracellular space (Paulsen et al., 2015). We show that this region is intrinsically flexible and represented by two main conformers, one of them dedicated to stabilizing the open channel by interactions with the upper part of the sensor. In this conformer, we indeed observed that structural alterations during MD simulation (from 305 to 356 ns) involved unwinding/rewinding of a helical turn in the region G750–I752.

Fourth, the results from our MDFF simulations indicate that the extracellular S1–S2 linker is physically coupled with the conformational changes in the lower gate, and may thus be an important regulator of TRPA1 channel gating. A number of lines of evidence presented here indicate that there is substantial communication between the activation gate and the upper part of the S1–S4 sensor module. The evidence that mutation at F746 can be rescued by a specific high-affinity peptide binding the S1–S2 linker provides a strong argument in favor of this view.

Fifth, our results demonstrate that the key effect of mutations in the S1–S2 linker (I751G/I752G) is in stabilizing the gate to decrease the voltage-dependent opening of the channel. The transient character of inhibitory effects of calcium that regulates the currents through the channel at hyperpolarized potentials is a further finding that supports the involvement of the S1–S2 region in the direct regulation of TRPA1 gating (Figure 7A). We demonstrated that the four negatively charged amino acids, E754, D757, E760, and D763, located at the extracellular surface of the channel are not important for the voltage-, agonist-, and Ca²⁺-dependent gating and are not involved in ion permeation through the channel. This conclusion is further supported by a recent comprehensive mutagenesis study by Christensen et al. (2016) that calculated electrostatic contributions of acidic residues in the channel vestibule and found that the only acidic residues that serve to attract permeant ions are near or within the pore.

Sixth, we note that the two presumed consensus N-glycosylation motifs on the S1–S2 extracellular linker are most likely not subjected to N-linked glycosylation, although their mutations influence its insertion into the plasma membrane. This is in contrast to the related TRPC3, TRPC6, TRPV5, TRPV6, and some K_V channels, whose activities are tightly regulated from the S1–S2 linker via glycosylation (Dietrich et al., 2003; Chang et al., 2005; Schwetz et al., 2010).

A recent comprehensive analysis of TRP channel transmembrane sequences identifies the conserved and coevolved networks of residues, suggesting a shared general allosteric mechanism underlying the polymodal gating (Palovcak et al., 2015). Our study extends this view and demonstrates that the most divergent regions may dominate, play a fundamental role in tuning the voltage-dependent gating and determine the rectification properties under conditions that resemble those experienced by the TRP channels in their native environment.

Materials and methods

Molecular dynamics simulations of the open state of the TRPA1 channel

The opening of the TRPA1 channel was modeled by using the available structures of TRPA1 in its closed conformation (3J9P; Paulsen et al., 2015) and TRPV1 in its open conformation (3J5Q; Cao et al., 2013). We applied the Molecular Dynamics Flexible Fitting (MDFF) method for combining high-resolution atomic structures with cryo-electron microscopy (cryo-EM) maps, that results in atomic models representing the conformational state captured by cryo-EM (Trabuco et al., 2008, 2009). In fact, we followed standard procedure using software packages, modules and keywords described in detail in “MDFF Tutorial” (http://www.ks.uiuc.edu/Training/Tutorials/). The original electron density map of TRPV1 (Cao et al., 2013) contains structural information about the N- and C-termini, which have a significantly different arrangement in TRPA1. Therefore, we used only the transmembrane portion of TRPV1 (i.e., amino acids K431 to S693 from 3J5Q) and generated its simulated electron map by interpolating the atomic number of each atom onto a 3-D grid and low-pass filtering it to the desired resolution (5 Å) using the software package VMD (Humphrey et al., 1996). The transmembrane part of TRPA1 (i.e., amino acids N722 to A971) was then rigid-docked into the resulting electron density map by means of the software package SITUS (Wriggers et al., 1999) to gain a starting simulated system for subsequent refinement by means of MDFF. SITUS uses for rigid docking the CoLoRes (contour low resolution docking) algorithm emulating in silico the simultaneous registration of surface and volume features (that is at the heart of visual docking as performed by an electron microscopist). A six-dimensional search using fast Fourier transforms is used to rapidly scan the rigid-body degrees of freedom of a probe molecule relative to a fixed target density map as described in detail in Chacón and Wriggers (2002). MDFF takes advantage of the NAMD's grid forces feature, through which an arbitrary external potential defined on a 3-D grid can be added to a molecular dynamics simulation. Here, the simulated TRPV1 electron map served to generate (again using VMD; Humphrey et al., 1996) additional potential U_EM that guided the transmembrane part of TRPA1 (parametrized using the CHARMM force field; Best et al., 2012) to the open conformation in the MDFF simulation produced using the NAMD software package (Phillips et al., 2005). HOLLOW was used for a “casting” of the central pore of TRPA1 by filling the voids, channels and pockets with dummy atoms defined on a grid (Ho and Gruswitz, 2008). Further details of the modeling procedures and coordinates of all structural models presented in this work are available from I.B. upon request.

Molecular dynamics simulations of the S1–S2 linker

Molecular dynamic (MD) simulations of the S1–S2 linker (amino acids S748–D763) were done in the context of the S1–S4 sensor domain (i.e., amino acids S725–L850) which was surrounded with TIP3P water molecules (Jorgensen et al., 1983). This simplified simulated system was chosen to shorten the time for obtaining representative conformers of the S1–S2 linker. Amino acids S725–N747 and T764–L850 were fixed to avoid unfolding of the hydrophobic S1–S4 sensor domain due to the absence of phospholipids in our simulated system. The S1–S2 linker alone forms an extracellular loop. Therefore, its conformational dynamics was not biased by removing of phospholipids. We used the AMBER_ILDN force field (Lindorff-Larsen et al., 2010), which gives reliable results in molecular dynamics simulations of folding for both α-helical as well as β-sheet moieties. The software package GROMACS (Abraham et al., 2015) and Linux computer nodes with powerful NVIDIA GPUs enabled us to accumulate a 500 ns trajectory at 410 K.

Molecular dynamics simulations of interactions between TRPA1 and Protoxin-I

To predict potential sites of interaction between TRPA1 and protoxin-I (Figures 8E,F), we took into account the homology of protoxin-I with the structurally similar vanillotoxin DkTx that binds the pore region of TRPV1 and traps it in its open state (Cao et al., 2013; Bae et al., 2016). We used again rigid docking using the SITUS software package (Wriggers et al., 1999) to place TRPA1 (i.e., amino acids A685-K989) and protoxin-I (2M9L) into a simulated electron map generated based on 3J5Q using VMD (see the “MDFF Tutorial” for details: http://www.ks.uiuc.edu/Training/Tutorials/). Resulting complex was embedded into the POPC phospholipidic bilayer and surrounded by TIP3P water molecules (Jorgensen et al., 1983) leading to a periodic box of size ~120 × ~120 × ~90 Å. 5-ns MD run produced by means of NAMD (Phillips et al., 2005) was used for relaxation. Finally, conformation II of the S1–S4 sensor module (gained within the MD simulation described above) was structurally aligned to produce Figures 8E,F.

Channel constructs, cell culture and transfection

Human embryonic kidney 293T (HEK293T) cells were cultured in Opti-MEM I medium (Invitrogen) supplemented with 5% FBS as described previously (Benedikt et al., 2007). The day before transfection, cells were plated in 24-well plates (2 × 10⁵ cells per well) in 0.5 ml of medium and became confluent on the day of transfection. The cells were transiently co-transfected with 300 ng of cDNA plasmid encoding wild-type or mutant TRPA1 and with 200 ng of GFP plasmid (TaKaRa, Japan) using the magnet-assisted transfection technique (IBA GmbH, Goettingen, Germany) and then plated on poly-L-lysine-coated glass coverslips. The plasmid encoding human TRPA1 was in the pCMV6-XL4 vector (OriGene Technologies, Rockville, MD). The chicken TRPA1 in the pcDNA3.1 vector was kindly provided by Dr. Makoto Tominaga (National Institute of Natural Sciences, Okazaki, Aichi, Japan). At least three independent transfections were used for each experimental group. The wild-type channel was regularly tested in the same batch as the mutants. The cells were used 24–48 h after transfection. For biochemical studies, the plasmid encoding GFP-tagged human TRPA1 in the pCMV6-AC-GFP vector was used (OriGene Technologies, Rockville, MD). The mutants were generated by PCR using a QuikChange II XL Site-Directed Mutagenesis Kit (Agilent Technologies) and confirmed by DNA sequencing (GATC Biotech, Germany).

Patch clamp recording

Whole-cell membrane currents were recorded by employing an Axopatch 200B amplifier and pCLAMP 10 software (Molecular Devices, Sunnyvalle, CA). Patch electrodes were pulled from a glass tube with a 1.5-mm outer diameter. The tip of the pipette was heat-polished, and its resistance was 3–5 MΩ. Series resistance was compensated by at least 70% in all recordings. The experiments were performed at room temperature (23–25°C). Only one recording was performed on any one coverslip of cells to ensure that recordings were made from cells not previously exposed to chemical stimuli. The extracellular control solution contained: 160 mM NaCl, 2.5 mM KCl, 1 mM CaCl₂, 2 mM MgCl₂, 10 mM HEPES, 10 mM glucose, adjusted to pH 7.3 with NaOH, 320 mOsm. The extracellular Ca²⁺-free solution (0 mM Ca²⁺) contained 150 mM NaCl, 10 mM HEPES, 2 mM HEDTA, adjusted to pH 7.3 with NaOH, 290 mOsm. The whole-cell pipette solution contained: 145 mM CsCl, 5 mM EGTA, 3 mM CaCl₂, 10 mM HEPES, 2 mM MgATP, pH 7.3, adjusted with CsOH, 290 mOsm. Allyl isothiocyanate (AITC) solution was prepared prior to use from a 0.1 M stock solution in Me₂SO. All of the chemicals were purchased from Sigma-Aldrich (St. Louis, MO). A system for rapid superfusion of the cultured cells was used for drug application (Dittert et al., 2006).

Whole-cell lysate preparation and PNGase F treatment

HEK293T cells were transfected with 1.5 μg cDNA by Lipofectamine 2000 (Life Technologies) on a 6-well plate. After 48 h, cells were washed with ice-cold phosphate-buffered saline and collected. After brief centrifugation at 15,000 × g for 5 min at 4°C, cells were briefly sonicated in the presence of 10x denaturating buffer (provided by New England Biolabs together with PNGase F) supplemented with a protease inhibitor mixture (Roche Diagnostics). Cell lysate was briefly centrifuged and denaturated at 95°C for 10 min. The protein concentration was measured using BCA protein assay reagent. Peptide: The N-Glycosidase F (PNGase F) digestion of proteins was performed according to the manufacturer's instructions (New England Biolabs). 10x G7 reaction buffer, 10% NP-40 (v/v), and 1000 units of enzyme were added to 20 μg of proteins. The samples were incubated for 3 h at 37°C. After that, 4x sample buffer was added and samples were boiled at 95°C for 10 min. In all cases, a control sample was run in parallel (water was added instead of PNGase F). Samples digested with the enzyme and their controls were loaded side by side on a 7% SDS gel.

Tunicamycin treatment

To study potential effects of N-glycosylation on TRPA1 channel functioning, transiently transfected HEK293T cells were grown in the presence of tunicamycin (5 μg/ml; Sigma Aldrich), an effective inhibitor of N-linked glycosylation, in mammalian cells for 24 h and then measured by the patch clamp technique. Stock of tunicamycin was prepared in Me₂SO at 5 mg/ml. The vehicle was added to the control (i.e., untreated) cells. For each run of experiments an untreated control was done.

SDS/PAGE and immunoblotting

Cell lysates of PNGase F-treated and untreated samples (control) of wild type or N747T/N753T mutant in 4x sample buffer were boiled for 5 min at 95°C. Similar amounts of proteins were applied and separated on 7% SDS/PAGE at 100 V ~3 h. The proteins were then transfered to a polyvinylidene difluoride (PVDF) membrane using a wet blotter at 60 V 3 h. The PVDF membrane was blocked with 5% non-fat milk in TBS-T buffer for 1 h and incubated on a shaker for 1 h with the primary antibody (mouse monoclonal GFP, 1:2000, OriGene). Three 10 min steps in TBS-T were used to wash out nonspecific bindings of the primary antibody. Horseradish-peroxidase-linked anti-mouse secondary IgG (Thermo Scientific; 1:10,000 dilution in 1% non-fat milk) was then added for 1 h at room temperature. After specific binding of secondary IgG, the membrane was washed 5-times for 5 min. The specific proteins were detected after incubating the membrane in an enhanced chemiluminescent HRP substrate kit (Thermo Scientific) on X-ray blue film (FOMA).

Fluorometric Ca²⁺-imaging

Calcium imaging was done using transfected HEK293T cells on coverslips with the same amount (500 ng) of cDNA of wild-type human TRPA1, I751A/I752A, I751G/I752G mutant and GFP (200 ng) for 24–48 h after transfection. Before the experiment, cells were loaded with 4 μM Fura2-AM (Invitrogen) and 0.02% Pluronic acid F-127 (Invitrogen) dissolved in a control extracellular solution (160 mM NaCl, 2.5 mM KCl, 1 mM CaCl₂, 2 mM MgCl₂, 10 mM HEPES, 10 mM glucose, adjusted to pH 7.3 with NaOH, 320 mOsm) for 1 h followed by a 20-min wash in fresh bath solution. The Cell^∧R imaging system was used to capture the fluorescence images obtained with alternating excitation at 340 and 380 nm and emission at >510 nm. Emission ratios (F₃₄₀/F₃₈₀) were calculated for each 0.5 s interval after subtraction of the background. A cellular response to an AITC was defined as an increase in the ratio by more than 0.3 during the 50-s application period. The AITC-induced responses of cells were normalized to their maximum responses evoked by 5 μM ionomycin. All experiments were performed at room temperature.

Data analysis

Electrophysiological data were analyzed in pCLAMP10 (Molecular Devices). Curve fitting and statistical analyses were done in SigmaPlot 10 (Systat Software). Current- and conductance-voltage (I-V and G-V) relationships were obtained from steady-state whole-cell currents measured at the end of voltage steps measured from −80 to +200 mV in increments of +20 mV. Voltage-dependent gating parameters were estimated by fitting the conductance G = I/(V-V_rev) as a function of test potential V to the Boltzmann equation: G = [(G_max-G_min)/(1+exp (- zF(V-V_1/2)/RT))] + G_min, where z is the apparent number of gating charges, V_1/2 is the half-activation voltage, G_min and G_max are the minimum and maximum whole cell conductance, V_rev is the reversal potential, and F, R, and T have their usual thermodynamic meanings. Statistical significance was determined using Student's t-test or one-way analysis of variance, as appropriate. All data are presented as the means ± standard error of mean (SEM). Blots were analyzed with ImageJ software (NIH). The data measured by the Ca²⁺-imaging technique were collected and analyzed with the software Cell^∧R (Olympus).

Author note

During the preparation of this manuscript, a study was published showing that the S1–S2 extracellular loop of the TRPV6 channel contributes to interfaces between the S1–S4 sensor and pore domain (Saotome et al., 2016). Like in TRPA1, the S1–S2 loop in TRPV6 contains two tandems of hydrophobic residues contributing to domain interfaces.

References

• 1

  AbrahamM. J.MurtolaT.SchulzR.PállS.SmithJ. C.HessB.et al. (2015). GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX 1–2, 19–25. 10.1016/j.softx.2015.06.001

  • CrossRef
  • Google Scholar

• 2

  BaeC.AnselmiC.KaliaJ.Jara-OsegueraA.SchwietersC. D.KrepkiyD.et al. (2016). Structural insights into the mechanism of activation of the TRPV1 channel by a membrane-bound tarantula toxin. eLife5:e11273. 10.7554/eLife.11273

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 3

  BenediktJ.TeisingerJ.VyklickyL.VlachovaV. (2007). Ethanol inhibits cold-menthol receptor TRPM8 by modulating its interaction with membrane phosphatidylinositol 4,5-bisphosphate. J. Neurochem.100, 211–224. 10.1111/j.1471-4159.2006.04192.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 4

  BestR. B.ZhuX.ShimJ.LopesP. E. M.MittalJ.FeigM.et al. (2012). Optimization of the Additive CHARMM All-Atom Protein Force Field Targeting Improved Sampling of the Backbone phi, psi and Side-Chain chi(1) and chi(2) Dihedral Angles. J. Chem. Theory Comput.8, 3257–3273. 10.1021/ct300400x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 5

  BrewsterM. S.GaudetR. (2015). How the TRPA1 receptor transmits painful stimuli: inner workings revealed by electron cryomicroscopy. Bioessays37, 1184–1192. 10.1002/bies.201500085

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 6

  CaoE.LiaoM.ChengY.JuliusD. (2013). TRPV1 structures in distinct conformations reveal activation mechanisms. Nature504, 113–118. 10.1038/nature12823

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 7

  ChacónP.WriggersW. (2002). Multi-resolution contour-based fitting of macromolecular structures. J. Mol. Biol.317, 375–384. 10.1006/jmbi.2002.5438

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 8

  ChangQ.HoefsS.van der KempA. W.TopalaC. N.BindelsR. J.HoenderopJ. G. (2005). The β-glucuronidase klotho hydrolyzes and activates the TRPV5 channel. Science310, 490–493. 10.1126/science.1114245

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 9

  ChenX.WangQ.NiF.MaJ. (2010). Structure of the full-length Shaker potassium channel Kv1.2 by normal-mode-based X-ray crystallographic refinement. Proc. Natl. Acad. Sci. U.S.A.107, 11352–11357. 10.1073/pnas.1000142107

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 10

  ChristensenA. P.AkyuzN.CoreyD. P. (2016). The outer pore and selectivity filter of TRPA1. PLoS ONE11:e0166167. 10.1371/journal.pone.0166167

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 11

  CrooksG. E.HonG.ChandoniaJ. M.BrennerS. E. (2004). WebLogo: a sequence logo generator. Genome Res.14, 1188–1190. 10.1101/gr.849004

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 12

  Deering-RiceC. E.ShapiroD.RomeroE. G.StockmannC.BevansT. S.PhanQ. M.et al. (2015). Activation of transient receptor potential ankyrin-1 by insoluble particulate material and association with asthma. Am. J. Respir. Cell Mol. Biol.53, 893–901. 10.1165/rcmb.2015-0086OC

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 13

  DietrichA.Mederos y SchnitzlerM.EmmelJ.KalwaH.HofmannT.GudermannT. (2003). N-linked protein glycosylation is a major determinant for basal TRPC3 and TRPC6 channel activity. J. Biol. Chem.278, 47842–47852. 10.1074/jbc.M302983200

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 14

  DittertI.BenediktJ.VyklickýL.ZimmermannK.ReehP. W.VlachovaV. (2006). Improved superfusion technique for rapid cooling or heating of cultured cells under patch-clamp conditions. J. Neurosci. Methods151, 178–185. 10.1016/j.jneumeth.2005.07.005

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 15

  DoernerJ. F.GisselmannG.HattH.WetzelC. H. (2007). Transient receptor potential channel A1 is directly gated by calcium ions. J. Biol. Chem.282, 13180–13189. 10.1074/jbc.M607849200

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 16

  EganT. J.AcuñaM. A.Zenobi-WongM.ZeilhoferH. U.UrechD. (2016). Effects of N-glycosylation of the human cation channel TRPA1 on agonist-sensitivity. Biosci. Rep.36:e00390. 10.1042/BSR20160149

  • CrossRef
  • Google Scholar

• 17

  FüllY.SeebohmG.LercheH.MaljevicS. (2013). A conserved threonine in the S1-S2 loop of KV7.2 and K V7.3 channels regulates voltage-dependent activation. Pflugers Arch.465, 797–804. 10.1007/s00424-012-1184-x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 18

  GaoY.CaoE.JuliusD.ChengY. (2016). TRPV1 structures in nanodiscs reveal mechanisms of ligand and lipid action. Nature534, 347–351. 10.1038/nature17964

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 19

  GuiJ.LiuB.CaoG.LipchikA. M.PerezM.DekanZ.et al. (2014). A tarantula-venom peptide antagonizes the TRPA1 nociceptor ion channel by binding to the S1-S4 gating domain. Curr. Biol.24, 473–483. 10.1016/j.cub.2014.01.013

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 20

  HinmanA.ChuangH. H.BautistaD. M.JuliusD. (2006). TRP channel activation by reversible covalent modification. Proc. Natl. Acad. Sci. U.S.A.103, 19564–19568. 10.1073/pnas.0609598103

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 21

  HoB. K.GruswitzF. (2008). HOLLOW: generating accurate representations of channel and interior surfaces in molecular structures. BMC Struct. Biol.8:49. 10.1186/1472-6807-8-49

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 22

  HumphreyW.DalkeA.SchultenK. (1996). VMD: visual molecular dynamics. J. Mol. Graph.14, 33–38.

  • Pubmed Abstract
  • Google Scholar

• 23

  JorgensenW. L.ChandrasekharJ.MaduraJ. D.ImpeyR. W.KleinM. L. (1983). Comparison of simple potential functions for simulating liquid water. J. Chem. Phys.79, 926–935. 10.1063/1.445869

  • CrossRef
  • Google Scholar

• 24

  KarashimaY.PrenenJ.TalaveraK.JanssensA.VoetsT.NiliusB. (2010). Agonist-induced changes in Ca²⁺ permeation through the nociceptor cation channel TRPA1. Biophys. J.98, 773–783. 10.1016/j.bpj.2009.11.007

  • CrossRef
  • Google Scholar

• 25

  KremeyerB.LoperaF.CoxJ. J.MominA.RugieroF.MarshS.et al. (2010). A gain-of-function mutation in TRPA1 causes familial episodic pain syndrome. Neuron66, 671–680. 10.1016/j.neuron.2010.04.030

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 26

  LeeS. Y.BanerjeeA.MacKinnonR. (2009). Two separate interfaces between the voltage sensor and pore are required for the function of voltage-dependent K⁺ channels. PLoS Biol.7:e47. 10.1371/journal.pbio.1000047

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 27

  LiaoM.CaoE.JuliusD.ChengY. (2013). Structure of the TRPV1 ion channel determined by electron cryo-microscopy. Nature504, 107–112. 10.1038/nature12822

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 28

  Lindorff-LarsenK.PianaS.PalmoK.MaragakisP.KlepeisJ. L.DrorR. O.et al. (2010). Improved side-chain torsion potentials for the Amber ff99SB protein force field. Proteins78, 1950–1958. 10.1002/prot.22711

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 29

  LongS. B.TaoX.CampbellE. B.MacKinnonR. (2007). Atomic structure of a voltage-dependent K⁺ channel in a lipid membrane-like environment. Nature450, 376–382. 10.1038/nature06265

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 30

  MacphersonL. J.DubinA. E.EvansM. J.MarrF.SchultzP. G.CravattB. F.et al. (2007). Noxious compounds activate TRPA1 ion channels through covalent modification of cysteines. Nature445, 541–545. 10.1038/nature05544

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 31

  NiliusB.AppendinoG.OwsianikG. (2012). The transient receptor potential channel TRPA1: from gene to pathophysiology. Pflugers Archiv.464, 425–458. 10.1007/s00424-012-1158-z

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 32

  PalovcakE.DelemotteL.KleinM. L.CarnevaleV. (2015). Comparative sequence analysis suggests a conserved gating mechanism for TRP channels. J. Gen. Physiol.146, 37–50. 10.1085/jgp.201411329

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 33

  PaulsenC. E.ArmacheJ. P.GaoY.ChengY.JuliusD. (2015). Structure of the TRPA1 ion channel suggests regulatory mechanisms. Nature520, 511–517. 10.1038/nature14367

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 34

  PhillipsJ. C.BraunR.WangW.GumbartJ.TajkhorshidE.VillaE.et al. (2005). Scalable molecular dynamics with NAMD. J. Comput. Chem.26, 1781–1802. 10.1002/jcc.20289

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 35

  SaotomeK.SinghA. K.YelshanskayaM. V.SobolevskyA. I. (2016). Crystal structure of the epithelial calcium channel TRPV6. Nature534, 506–511. 10.1038/nature17975

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 36

  SchwetzT. A.NorringS. A.BennettE. S. (2010). N-glycans modulate K_v1.5 gating but have no effect on K_v1.4 gating. Biochim. Biophys. Acta1798, 367–375. 10.1016/j.bbamem.2009.11.018

  • CrossRef
  • Google Scholar

• 37

  StoryG. M.PeierA. M.ReeveA. J.EidS. R.MosbacherJ.HricikT. R.et al. (2003). ANKTM1, a TRP-like channel expressed in nociceptive neurons, is activated by cold temperatures. Cell112, 819–829. 10.1016/S0092-8674(03)00158-2

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 38

  SuraL.ZimaV.MarsakovaL.HynkovaA.BarvikI.VlachovaV. (2012). C-terminal Acidic Cluster Is Involved in Ca²⁺-induced regulation of human transient receptor potential ankyrin 1 channel. J. Biol. Chem.287, 18067–18077. 10.1074/jbc.M112.341859

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 39

  TrabucoL. G.VillaE.MitraK.FrankJ.SchultenK. (2008). Flexible fitting of atomic structures into electron microscopy maps using molecular dynamics. Structure16, 673–683. 10.1016/j.str.2008.03.005

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 40

  TrabucoL. G.VillaE.SchreinerE.HarrisonC. B.SchultenK. (2009). Molecular dynamics flexible fitting: a practical guide to combine cryo-electron microscopy and X-ray crystallography. Methods49, 174–180. 10.1016/j.ymeth.2009.04.005

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 41

  VriensJ.HeldK.JanssensA.TóthB. I.KerselaersS.NiliusB.et al. (2014). Opening of an alternative ion permeation pathway in a nociceptor TRP channel. Nat. Chem. Biol.10, 188–195. 10.1038/nchembio.1428

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 42

  WangY. Y.ChangR. B.WatersH. N.McKemyD. D.LimanE. R. (2008). The nociceptor ion channel trpa1 is potentiated and inactivated by permeating calcium ions. J. Biol. Chem.283, 32691–32703. 10.1074/jbc.M803568200.

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 43

  WriggersW.MilliganR. A.McCammonJ. A. (1999). Situs: a package for docking crystal structures into low-resolution maps from electron microscopy. J. Struct. Biol.125, 185–195. 10.1006/jsbi.1998.4080

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 44

  ZhangX.RenW.DeCaenP.YanC.TaoX.TangL.et al. (2012). Crystal structure of an orthologue of the NaChBac voltage-gated sodium channel. Nature486, 130–134. 10.1038/nature11054

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 45

  ZhuJ.WatanabeI.PoholekA.KossM.GomezB.YanC.et al. (2003). Allowed N-glycosylation sites on the Kv1.2 potassium channel S1-S2 linker: implications for linker secondary structure and the glycosylation effect on channel function. Biochem. J.375(Pt 3), 769–775. 10.1042/BJ20030517

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 46

  ZubcevicL.HerzikM. A.Jr.ChungB. C.LiuZ.LanderG. C.LeeS. Y. (2016). Cryo-electron microscopy structure of the TRPV2 ion channel. Nat. Struct. Mol. Biol.23, 180–186. 10.1038/nsmb.3159

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 47

  ZurborgS.YurgionasB.JiraJ. A.CaspaniO.HeppenstallP. A. (2007). Direct activation of the ion channel TRPA1 by Ca²⁺. Nat. Neurosci.10, 277–279. 10.1038/nn1843

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 48

  ZygmuntP. M.HogestattE. D. (2014). Trpa1. Handb. Exp. Pharmacol.222, 583–630. 10.1007/978-3-642-54215-2_23

  • Pubmed Abstract
  • CrossRef
  • Google Scholar