Voltage‐gated proton channels in polyneopteran insects

Voltage‐gated proton channels (HV1) are expressed in eukaryotes, including basal hexapods and polyneopteran insects. However, currently, there is little known about HV1 channels in insects. A characteristic aspartate (Asp) that functions as the proton selectivity filter (SF) and the RxWRxxR voltage‐sensor motif are conserved structural elements in HV1 channels. By analysing Transcriptome Shotgun Assembly (TSA) databases, we found 33 polyneopteran species meeting these structural requirements. Unexpectedly, an unusual natural variation Asp to glutamate (Glu) at SF was found in Phasmatodea and Mantophasmatodea. Additionally, we analysed the expression and function of HV1 in the phasmatodean stick insect Extatosoma tiaratum (Et). EtHV1 is strongly expressed in nervous tissue and shows pronounced inward proton conduction. This is the first study of a natural occurring Glu within the SF of a functional HV1 and might be instrumental in uncovering the physiological function of HV1 in insects.

Voltage-gated proton channels (H V 1) are expressed in eukaryotes, including basal hexapods and polyneopteran insects. However, currently, there is little known about H V 1 channels in insects. A characteristic aspartate (Asp) that functions as the proton selectivity filter (SF) and the RxWRxxR voltage-sensor motif are conserved structural elements in H V 1 channels. By analysing Transcriptome Shotgun Assembly (TSA) databases, we found 33 polyneopteran species meeting these structural requirements. Unexpectedly, an unusual natural variation Asp to glutamate (Glu) at SF was found in Phasmatodea and Mantophasmatodea. Additionally, we analysed the expression and function of H V 1 in the phasmatodean stick insect Extatosoma tiaratum (Et). EtH V 1 is strongly expressed in nervous tissue and shows pronounced inward proton conduction. This is the first study of a natural occurring Glu within the SF of a functional H V 1 and might be instrumental in uncovering the physiological function of H V 1 in insects.
Voltage-gated proton channels are found in most eukaryote kingdoms, from coccolithiophores [1] to dinoflagellates [2], chordata, fungus, plants and mammals [3,4]. In dinoflagellates (Lingulodinium polyedrum), H V 1 channels trigger light emission [5]. In mammals, H V 1 channels play a pivotal role in many physiological processes including pH homeostasis, respiratory burst of phagocytes and maturation of sperm [6][7][8]. In several breast and colorectal cancers, H V 1 is significantly upregulated [9,10]. Much less is known about H V 1 channels in hexapoda, especially in insects. Recently, a H V 1 channel of the Zygentoma Nicoletia phytophila was characterized and other H V 1 sequences have been found in phylogenetically more basal hexapodes [11].
As a member of the voltage-gated superfamily of ion channels, H V 1 possesses four transmembrane regions (S1-S4) with a typical voltage-sensor element in the fourth transmembrane segment (S4) [3,4]. Compared to other voltage-gated ion channels, such as potassium, sodium and calcium channels, H V 1 does not have the last two transmembrane regions (S5-S6) of the typical six transmembrane alpha helices, usually composing the pore region. Instead, a different ion conduction pathway is established for protons, including all four transmembrane regions with a typical negatively charged aspartate in S1 as proton selectivity filter (SF) [17][18][19]. Additional site-directed mutagenesis of this aspartate residue (position 112 in human H V 1) revealed new aspects of structure-function relationships within H V 1 channels [2,11,17,20,21]. To date, H V 1 channels are considered to be dimeric with the intracellular C-terminal domain connecting both subunits [22][23][24].
The aim of this study was to characterize the H V 1 channel of the stick insect Extatosoma tiaratum. Besides the H V 1 channel of the Zygentoma Nicoletia, this is the second H V 1 channel from Hexapoda and the first of the class Insecta. Our database analysis provided a detailed picture of the presence and absence of H V 1 genes in different hexapodan and insect orders. With the analysis of the tissue-specific expression in Extatosoma and the electrophysiological characterization, we hope to blaze a trail for uncovering the physiological function of H V 1 channels in insects. The presence of an unusual glutamate residue as SF within the S1 domain of Extatosoma H V 1 was analysed and compared to the other hexapodan H V 1 channel from Nicoletia.

Database analysis
The BLAST algorithm was used to analyse insect transcriptome shotgun assembly (TSA) and genomic databases at NCBI. The recently described sequence of the Zygentoma Nicoletia phytophila (KT780722 [11,25]) was used as query sequence. Only sequences harbouring the typical RxWRxxR motif in the S4 segment were used for further analysis.

Heterologous expression
Extatosoma tiaratum EtH V 1 gene was synthesized commercially (Eurofins/Genomics, Ebersberg, Germany). The synthesized DNA including a 5 0 BamHI and 3 0 EcoRI restriction site was cloned into a pEX-A2 plasmid. The gene was later subcloned into a pQBI25-fC3 or pcDNA3.1, using 5 0 BamHI and 3 0 EcoRI restriction sites and GFP fused to N-terminal as previously described [2,11,17,25]. tSA201 cells (human kidney cell line) were grown to 85% confluency in 35 mm culture dishes. Cells were transfected with 1.0 μg plasmid DNA using polyethylenimine (Sigma, St. Louis, MO, USA). After 12 h at 37°C in 5% CO 2 , cells were trypsinized and replated onto glass coverslips at low density for patch clamp recording the same day and the next day. Green cells were selected under fluorescence for recording. As in [11,25], whole cell patch clamp showed no other voltage-or time-dependent conductance under our recording conditions. The level of expression of EtH V 1 was sufficiently high so that potential contamination by native H V 1 currents was negligible.

Electrophysiology
Patch-clamp recordings were done as described in [11,25]: A patch-clamp amplifier EPC 10 (HEKA, Lambrecht, Germany) was used. Recordings were stored on hard discs and analysed with Origin (Origin 2017, Northampton, MA, USA). Patch pipettes were made from borosilicate capillaries GC 150TF-10 (Harvard Apparatus, Holliston, MA, USA) and pulled using Flaming Brown automatic pipette puller P-1000 (Sutter Instruments, Novato, CA, USA). Pipettes were heat polished to a tip resistance ranging typically from 5 to 9 MΩ with pipette solutions used. Electrical contact with the pipette solution was achieved by a chlorinated silver wire and connected to the bath with an agar bridge made with Ringer's solution. Seals were formed with Ringer's solution (in mM 160 NaCl, 4.5 KCl, 2 CaCl 2 , 1 MgCl 2 , 5 Hepes, pH 7.4) in the bath, and the potential zeroed after the pipette was placed above the cell. Whole-cell and inside-out solutions (pipette and bath) included 100 mM buffer close to its pK a with tetramethylammonium (TMA + ) and methanesulfonate (CH 3 SO 3 − ) as the main ions, 1 mM EGTA, and 1-2 mM Mg 2+ with an osmolarity of 300 mOsmÁkg −1 . Buffers were 2-(N-morpholino) ethanesulfonic acid (MES) at pH 5.5 and pH 6.0, Bis-(2hydroxyethyl)imino-tris-(hydroxymethyl)-methane (BIS-TRIS) at pH 6.5 and PIPES at pH 7.0. Resistance of the seals was usually > 3 GΩ. Currents are shown without correction for leak or liquid junction potentials. Data were collected between 19°C and 23°C. Currents were fitted to a rising exponential to obtain the activation time constant (τ act ). The maximal proton conductance (g H, max ) was calculated from the steady-state current (the fitted current extrapolated to infinite time) using reversal potentials (V rev ) measured in each solution in each cell. In these fits, the initial delay was ignored and the remaining current usually fitted a single exponential well. The threshold potential, V thres , was determined from families of pulses as the potential where the first tail current was observed once the membrane was repolarized. The reversal potential was measured by two methods. When V thres was negative to V rev , it could be readily determined by the zero current. When V thres was positive to V rev , then V rev was determined with the tail current method. Voltage dependence of activation was obtained by linear fittings of the activation kinetics plots at the region of the curve where τ act becomes faster with depolarization. Selectivity for protons was determined by comparison of measured reversal potentials to the Nernst potential for protons (E H ) at the experimental ΔpH (pH o − pH i ). The pH dependence of gating was evaluated in a pH range from 5.5 to 7.0 by linear regression of data from V thres against V rev graphs, in a potential range from −70 mV to +70 mV. Overexpression of the channels in small cells resulted in large proton currents which removed enough protons from the cell to change pH i considerably. Proton channel gating kinetics depend strongly on pH; therefore, proton depletion is a significant source of error. To minimize this problem, families with different pulse lengths were applied. Longer pulses were used to determine pulses close to V thres where τ act is slow, while shorter pulses were used at more positive voltages. Zinc inhibition assays were tested extracellular and performed at 0, 10 and 100 μM ZnCl 2 . EGTA was omitted from zinc-containing solutions. Families of pulses of different lengths were collected in each zinc condition and exchanges of external solutions recorded during test-pulse protocols. The data are shown without corrections for buffer binding.

Structural model
A structural model of the transmembrane domain was constructed via homology modelling using Modeller [26,27] and the crystallographic structure of Ci-VSD in an open state as template (PDB: 4G7V [28]). The amino acid sequences (residues G23 to S158 for Ci-VSD and residues G31 to V169 for Extatosoma) were aligned with MUSCLE [29]. A few inaccuracies in the alignment were corrected manually. The two sequences have 18% identity and 40% similarity. 100 models were generated. The best model according to the Modeller objective function was refined with 3DRefine [30]. Five solutions were generated and ranked according to the 3Drefine and RWPlus scores. The solution with the best ranking was conserved.

Polyneopteran TSA-database analysis
Using the typical proton channel signature motif of the voltage sensor in S4 (RxWRxxR), we identified 33 putative polyneopteran H V 1 channels: nineteen in stick insects (Phasmatodea), eight in locusts/crickets (Orthoptera), three in webspinners (Embioptera), two in gladiators (Mantophasmatodea) and one in stoneflies (Plecoptera). A complete list of all identified channels with the respective GenBank Acc. No. can be found in Table S1. All amino acid sequences compiled from TSA files are shown in Fig. S1. No H V 1 sequence homolog was found in cockroach TSA databases and in the species-poor polyneopteran groups of ice crawlers and ground lice. The only sequences found initially in mantis (Metallyticus splendidus, GATB01324360, see Table S1) and earwigs (Forficula auricularia, GAYQ01077212) were subsequently removed from 1KITE datasets as they most likely represent a fungal contamination of the animal sample ( [31], B. Misof personal communication, Table S1). Further database analysis identified seven more partial sequences (six stick insects, one locust) with significant homology to H V 1 channels, however not or only partially covering the signature motif RxWRxxR (Table S1B).
Within the 18 full-length clones, sequence identity between species from different polyneopteran orders were 42%-55% within the core region of the channels (S1-S4, 63%-75% homology). Within a polyneopteran order, sequence identity was > 80% (Table S2). All sequences have the typical four-transmembrane structure comparable to other known H V 1 channels: short loops between the transmembrane region (8-16 amino acids) and rather short C-and N-terminal domains (~50 amino acids). Sequence length varied between 211 and 273 amino acids in total.
A striking sequence variation is found within the SF in the S1 segment. In most known H V 1 channels, a negatively charged aspartate residue (D112 in human H V 1) is important for proton selectivity. A similar aspartate is found in all orthopteran, mantodean and embiopteran sequences. In phasmatodean and mantophasmatodean sequences however, this aspartate is replaced by an also negatively charged glutamate residue (Figs 1 and S2). A similar exchange (D to E) within the S1 selectivity filter has been artificially generated by different mutagenesis  [11,17,20,21,32], resulting in more negative activation, speeding up of activation kinetics but maintaining proton selectivity. Here, for the first time, we show that a glutamate residue occurs naturally at the SF position.

Do any hemipteran or holometabolan insects harbour a H V 1 channel homolog?
Despite being overrepresented in protein and nucleotide databases, analysis of all TSA and genomic databases of hemipteran and holometabolan insects revealed only seven TSA sequences encoding the S4 signature motif. The genomic sequence data (mainly from Diptera) showed absolutely no evidence of the presence of an H V 1 homolog in Hemiptera and Holometabola. A closer look at the respective TSA sequences showed that none of them has high homology to known hexapodan H V 1 homologs, but show strong homology to fungal H V 1 sequences (in six cases) and to Chelicerata (one case). Therefore, all identified putative H V 1 channels within these insect orders are likely due to parasitic contamination of the animal sample investigated (Table S1C). Indeed, especially fungal contaminations are easily uncovered by sequence analysis of H V 1 homologs, as the third The three voltage-sensor arginines in S4 are shown in blue, the glutamate SF (in S1) in green, and the conserved tryptophan in red sticks. Other arginine and lysine residues are depicted in blue, aspartate and glutamate residues in red lines. A hydrophobic gasket, depicted as yellow surface, is formed by the side chains of V59, F101, V128 and I129 that separates the inner and outer aqueous vestibules. The network of stabilizing interactions shown as gold dots (distances inÅ) is similar to that in human H V 1 (hH V 1). R1, here R156, interacts with a glutamate residue, E69 (E119 in hH V 1), R2 (R159) with the SF E62 (D112 in hH V 1) and E136 (D185 in hH V 1), and R3 with E104 (E153 in hH V 1) and D125 (D174 in hH V 1). (B and C) Sequence logo representation of H V 1 S1-(B) and S4-domains (C) showing the amino acid frequencies at respective positions of four different orders of polyneopteran insects. The first green residues shown (position 10 in B) belong to the SF; the voltage-sensor arginine residues are highlighted in blue and the conserved S4 tryptophan residue in red. Number of sequences used for analysis in S1, S4; respectively: Orthoptera (n = 3, 4), Embioptera (n = 3, 3), Phasmatodea (n = 15, 18) and Mantophasmatodea (n = 2, 2). arginine of the RxWRxxR motif is usually mutated to a lysine residue in fungus. We conclude that there is no evidence for the presence of H V 1 channel homologs in Hemiptera or Holometabola.

526
Structure of the Extatosoma tiaratum H V 1 channel (EtH V 1) For further characterization, we selected the H V 1 channel of the stick insect Extatosoma tiaratum (EtH V 1, GenBank Acc. No. GAWG01024136). EtH V 1 is 236 amino acids (aa) in length, possesses the usual four transmembrane regions and 52 aa N-terminal and 65 aa C-terminal intracellular domains. This sequence harbours the phasmatodean-specific glutamate (E62) as SF in S1 and a typical S4 voltage sensor. Within the core segment, S1-S4 EtH V 1 is 33% identical and 63% homologous to human H V 1.

Tissue expression of EtH V 1
An RT-PCR analysis of five different tissues isolated from two animals showed strongest expression in the nervous system as a conglomerate of all Extatosoma ganglia. Moderate expression was found in the digestive system and weak expression was detected in eyes, whereas no clear expression could be detected in muscle and antenna. In Fig. 2, an agarose gel of the RT-PCR is shown; the 363 bp EtH V 1-PCR product is indicated. As a positive control, Extatosoma histone H3 expression was detected in all five tissue samples. PCR products from ganglia and digestive system were verified by DNA sequencing.
Electrophysiological characterization of EtH V 1 Extatosoma EtH V 1 was expressed as a GFP fusion protein in tSA cells and was distinguishably localized in the cell membrane when detected under fluorescence. Transfected cells had a capacitance of 9.62 AE 2.03 pF (mean AE SD, n = 9 cells) and presented a mean conductance density of 1.07 AE 0.44 nSÁpF −1 (mean AE SD, n = 9 cells), demonstrating reliable expression levels.
Typical proton selective currents were detected during patch-clamp experiments. Consistent with reports of other species [1][2][3][4]11,33,34], robust H + currents presented threshold potential, V thres , and time-dependent behaviour in the order of seconds. The time course of currents has sigmoidal shape which has previously been attributed to the dimeric nature of H V 1 [22,35]. After a short delay, currents rise exponentially during membrane depolarization and large tail currents appear at repolarization steps. Figure 3A depicts an example of a whole-cell patchclamp measurement of EtH V 1 in two different pH conditions. The amplitude of currents is time-dependent and increases with every depolarizing step, clearly indicating voltage-dependent activation. Large and relatively slow tail currents are also seen once the channel deactivates as consequence of repolarization of the cell membrane (e.g. V hold = −80 mV).
In common with other H V 1 channels, g H of EtH V 1 is also regulated by the pH gradient across the membrane, ΔpH (pH o − pH i ). When ΔpH increases (ΔpH > 0) or decreases (ΔpH < 0), EtH V 1 adjusts its g H to more positive or to more negative potentials accordingly. Figure 3B shows a clear rightward shift of the conductancevoltage relationship, g H -V, of 25 mV once external pH (pH o ) was diminished from 7.0 to 6.5 (black arrow). The same behaviour was detected in inside-out patches where pH i was exchanged to generate the same ΔpH = −0.5 (Fig. S3).
The effect of g H -V change can also be seen on EtH V 1 activation kinetics, τ act (Fig. 3C). The voltage dependence of τ act of EtH V 1 is represented by a slope of −0.10 AE 0.02 e-fold sÁmV −1 (mean AE SD, n = 6) which renders into 10 mV/e-fold change.
We further tested the classical proton channel inhibitor, zinc, on EtH V 1. Figure 4 shows how Zn 2+ affects EtH V 1 H + currents in the same cell. The divalent cation drastically reduces the amplitude and kinetics of proton currents activation. In our experiments, we increased [Zn 2+ ] from 0 (Control) to 10 and 100 μM. During test pulse protocols, H + activation and tail currents reduce their amplitude once zinc is added to the bath solution (Fig. 4A). Figure 4B depicts families of pulses under the three different zinc conditions in a whole-cell configuration. Two main effects are shown by the families of pulses: a reduction of activation currents at the same depolarization and a slowing of activation kinetics. In agreement with other H V 1 studies [5,25,[34][35][36], g H -V curves shift rightwards along the voltage axe (Fig. 4C), and τ act becomes slower once [Zn 2+ ] increases (Fig. 4D). Both are the two main effects of zinc inhibition on proton channels. The data demonstrate that EtH V 1 is sensitive to the classical proton channel inhibitor, zinc, in a micromolar range.

Proton selectivity and pH-dependent gating of EtH V 1
The reversal potential of EtH V 1 was analysed in a pH o range between 5.5 and 7.0, and pH i 6.5-7.0. Because activation of EtH V 1 was negative to V rev for most of the cases, V rev was determined directly as the zero current in a family of depolarizing pulses.
The recorded values follow accurately the predicted Nernst potential for proton conduction, E H , indicating that EtH V 1 is highly proton selective (Fig. 5A). Deviations of V rev values from E H are a consequence of incomplete pH i control even though high pH buffer concentrations were used, for example strong depolarization causes H + depletion that increases pH i . Increase in internal proton concentration, [H + ] i , and the consequent drop of internal pH (pH i ) causes divergences between measured V rev and calculated E H . A rise in [H + ] i is provoked by consistent inward H + currents during channel's activation, when V thres < V rev , or by large tail currents observed during membrane repolarization.
In H V 1, voltage and pH modulate the channel's gating. To evaluate the pH dependence of gating of EtH V 1, we applied the 'threshold versus reversal' strategy previously used in other studies [2,11,37]. The approach consists of determining the reversal and threshold potential at a wide range of pH to obtain an equation of the form V thres = slopeÁV rev + offset. We measured V rev and V thres at pH ranges from 5.5 to 7.0 applying different ΔpH (Fig. 4B). In a total of 16 determinations, data permit to define the voltage dependence of EtH V 1 as: Interestingly, EtH V 1 and NpH V 1 present the same voltage dependence of gating translated to a slope of 0.77 V thres /V rev . Nevertheless, a major difference in the offsets of both channels can be seen. Extatosoma is more negatively activated (−23 mV) than Nicoletia (−2.4 mV). The dotted line in Fig. 5B represents equality between V thres and V rev . Data located under the dotted line stand for inward H + conduction, while data points above the dotted line represent outward H + currents. By definition, if V thres is positive to V rev , H V 1 conducts protons outwards, alkalinizing the cytosol. In opposition to this, threshold values negative to V rev show inwardly directed proton currents. Thus, Fig. 5B enables the fast determination of the proton currents direction. In the whole investigated pH range, EtH V 1 activation is negative to V rev . EtH V 1 permits proton influx. In contrast, NpH V 1 activation is 20 mV more positive and permits proton extrusion, while mostly preventing inward H + flux.

Distribution among hexapodans
Proton channels are unique members of the voltagegated ion channel superfamily, as they are represented in most species by a single gene or by no gene at all. This indicates that H V 1 offers an evolutionary advantage over some species, whereas other species may dispense an H V 1 homolog. The common ancestor of Hexapoda, Crustacea, Myriapoda and Chelicerata clearly possesses a single H V 1 gene. Figure 6 shows a Considering Polyneoptera as monophyletic group, it is obvious that the common ancestor of the sister groups Mantodea and Blattodea lost an H V 1 homolog. The absence of H V 1 within these orders is very likely since sequence coverage of these orders is high. Furthermore, it is extremely unlikely that in all species analysed the putative H V 1 homolog has simply been missed by sequencing, instead of been lost during evolution. During this study, nine TSA sequences were identified with significant homology to known fungal sequences, with up to 99% identity. These clones undoubtedly represent sample contaminations. The 1KITE project and other related TSA studies provide us with a huge amount of transcriptomal sequence data [31]. Despite the overall data being of very good quality and sequence coverage also being high (depending somewhat on the species analysed), a major drawback is sample contaminations from insect parasites, mainly fungus. As whole insects were analysed by TSA studies, such contaminations cannot be excluded within the first sequence drafts. Indeed, algorithms were used to eliminate such noninsect sequences from the dataset; however, some contaminations are still found. Actually, from ten insect TSA database entries representing clear contaminations, only three were subsequently removed by the 1KITE staff (~30% of all contaminated TSA entries,~50% of contaminated 1KITE TSA data).

Glutamate as SF of EtH V 1
The unusual glutamate residue in the S1 selectivity filter (E62 in Extatosoma) is found only in the two closely related polyneopteran orders of Phasmatodea and Mantophasmatodea. Indeed, homologous positions at the SF have been characterized in detail by site-directed mutagenesis in human [17], dinoflagellate [2] and also in the Zygentoma Nicoletia phytophila H V 1 [11]. Asp to Glu substitutions at the SF were Comparison of the pH dependence of gating between EtH V 1 and NpH V 1. Threshold potential, V thres , are plotted versus the reversal potential for both, EtH V 1 (blue circles) and NpH V 1 (red triangles), in a voltage range from −70 mV to + 70 mV. Dotted line represents equality between V rev and V thres . Blue and red solid lines show the linear regression of data from EtH V 1 and NpH V 1, respectively. For the same voltage range, NpH V 1 presents a pHdependent gating equal to V thres = 0.77 V rev -2.4 mV (n = 41); meanwhile, EtH V 1 shows a more negative activation defined by V thres = 0.77 V rev -23 mV. n = 16 (9 cells), pH i was 6.5 or 7.0, and measurements made in a pH o range from 5.5 to 7.0. (C) Upper recording: activation of proton currents (the conductance activated negative to V rev ) in a whole-cell patch-clamp configuration show a V rev between +50 and +60 mV when pH i = 6.5 and pH o 5.5, accordingly with a predicted E H of +58 mV in the same pH conditions. Pulses were applied in 10 mV increments from the holding potential (−40 mV) to +60 mV. Lower recording: Tail current records of a patch at symmetrical pH i // pH o 7.0 indicating a V rev close to 0 mV. Test pulses were applied in 10 mV increments after a depolarizing pulse (+55 mV) from the holding potential (−40 mV) to +10 mV. done considering that both residues are negatively charged at physiological conditions. The investigations proved that H V 1 is still proton selective once a glutamate is present at the SF position in S1 [2,11,17]. On the other hand, the extreme proton selectivity is lost once Asp is mutated to a neutral amino acid, for example alanine (Ala), making the channel also permeable to anions [2,11,17]. A potential mechanism explaining the necessity of a negatively charged amino acid at this position has been reported by Dudev et al. [18]. They analysed the selectivity mechanism of H V 1 applying a quantum-based model. In the open state of the channel, the SF is composed of a salt bridge interaction between the Asp and the second or third Arg of the voltage-sensor motif RxWRxxR, in a constricted part of the channel. When placed in-between the Asp-Arg SF, H 3 O + enables protonation of the Asp, breaking the electrostatic Asp-Arg interaction in an energetically favourable process. Later, the gained H + is transferred from the protonated aspartate (AspH) to a neighbouring nucleophile and the Asp-Arg interaction restored, allowing other protons to initiate the process again. In this way, protons can travel through the SF of a H V 1. Other competing ions as Cl − and Na + are repelled by the residue bearing the same charge (e.g. Asp for Cland Arg for Na + ) or trapped by the residue of opposite charge and cannot cross though the Asp-Arg SF [18]. In voltage-gated proton channels (H V 1), the existence of a negatively charged residue at the SF is mandatory for proton selectivity and E62 in Extatosoma meets this requirement. The pk a of the glutamate side chain is~4.25, which enables the residue to be deprotonated at pH > 5.0 and to remain negatively charged as consequence. This would permit EtH V 1 to have a SF composed of a Glu-Arg interaction working in a similar manner as the Asp-Arg selectivity mechanism.
In our experiments, protons were in a concentration from 0.1 μM (pH 7.0) to 3.12 μM (pH 5.5), between four to five orders of magnitude lower than the concentrations of the main ions TMA + and CH 3 SO 3 − (90-125 mM). Despite this great disproportion for protons, all measured V rev follow nernstian behaviour of protons (Eqn 3). Small variations are mainly consequence of an imperfect control of the pH i . Depletion and accumulation of H + due to high depolarization and robust inward H + currents, respectively, are a common source of error while measuring H V 1 in the whole-cell patch-clamp configuration. In this configuration, the accuracy of the control of cytosolic pH is limited by the diffusion rate of the buffer between the pipette and the cell. This diffusion exchange lasts from seconds to even minutes [38]. In our experiments, we try to circumvent this problem by shortening pulses at very positive voltages and/or increasing resting times between pulses. Our data specify EtH V 1 as a proton selective channel.

EtH V 1 is inhibited by Zn 2+
Inhibition of proton currents by external addition of zinc is considered one of the main characteristics of H V 1. We tested the response of EtH V 1 to external Zn 2+ in micromolar range concentrations. Our experiments recorded inhibition of H + currents at 10 μM Zn 2+ which was augmented once [Zn 2+ ] further increased (Fig. 4A,B). The inhibitory effect is better seen as a shift of the g H -V relationship and as slowing of the kinetics of activation (Fig. 4C,D). The tendency is similar to other tested H V 1 [5,25,[34][35][36]. The mechanism of inhibition of proton channels by zinc is still an ongoing discussion; nevertheless, several studies have identified some of the amino acids involved.
Mammalian proton channels possess two identified Zn 2+ binding sites composed exclusively of external His residues [3,36]. The first one is located at the top of S2 alpha helix and the second is placed in the S3-S4 loop (H140 and H193 in the human H V 1). Substitution of these two His residues to Ala renders the channel zinc insensitive [3,35]. In contrast, proton channels of further species show more diversity. For example, the other characterized insect proton channel, NpH V 1, conserves a His residue (H92) at the same relative position of H140 of the human channel but presents a variation to Asp (D145) in the second biding site, H193. A detailed zinc inhibition analysis demonstrated that in the case of NpH V 1, the main inhibitory effect is caused by zinc binding to the first position, H92, with minimal participation of the second binding site, D145. The inhibition of NpH V 1 by Zn 2+ is smaller than in human and rat channels [25]. Nevertheless, zinc sensitivity of Nicoletia is greatly increased by mutation of D145 to His and completely abolished once both amino acids are mutated to Ala, similar to mammalian H V 1 channels [25]. It seems that histidine residues at these two precise locations of H V 1 are important for high zinc sensitivity.
Interestingly, the putative Zn 2+ binding sites of EtH V 1 consist of a lysine residue (Lys91) and of an aspartate residue (Asp144) at the first and second positions, respectively (see alignment of Fig. S2). A Lys residue at the first zinc binding site is a peculiarity of EtH V 1 but remarkably common to all identified polyneopteran H V 1. However, this noteworthy difference is lost in other orders as Zygentoma, Diplura, Protura and Archeognatha, whose H V 1 have the regular His residue also present in mammal channels. On the other hand, the second putative binding position (Asp144) on the S3-S4 linker of Extatosoma is preceded by two consecutive histidine residues, His142 and His143, which are potentially coordinating Zn 2+ . The same -His-His-Asppattern is shared with other phasmatodean H V 1 homologs: SsH V 1, RaH V 1 and MeH V 1 (Fig. S2). Yet, other phasmatodean species (Aretaon asperrimus and Peruphasma schultei), the Embioptera Aposthonia japonica and the Archeognatha Pedetontus okajimae, present in contrast only one His residue next to the Asp of the S3-S4 linker.
Investigations in Nicoletia confirmed the dimeric nature of an insect proton channel and the possibility of zinc binding at the interface of both monomers [25]. Coordination of zinc in-between H V 1 subunits has also been suggested in inhibition studies of the human channel [35,39]; hence, EtH V 1 stoichiometry might be a factor to be considered.
Structural differences between Zn 2+ binding sites among species also generate different zinc sensitivities. Therefore, potency of Zn 2+ on H V 1 can be related to the surrounding [Zn 2+ ] and the function of proton channels in the organisms. Thus, low zinc concentration in human and mouse serum ranges from 13 to 20 μM [40] and the respective H V 1 shows consequently higher zinc sensitivity than in other species, for example Nicoletia phytophila (insect), Ciona intestinalis (sea squirt) and Helisoma trivolvis (snail) [34]. The sensitivity to Zn 2+ revealed by the animal model Danio rerio (zebrafish) is even lower, which associates with the considerably higher zinc concentrations in the serum of the animal (~150 μM) [40]. Unfortunately, there are no data available determining the concentration of zinc in the haemolymph of Extatosoma.
Further studies, including the pH dependence of Zn 2+ inhibition, site-directed mutagenesis of putative binding sites and the analysis of the channel oligomerization are still necessary to address the nature of zinc inhibition of polyneopteran H V 1 channels.
EtH V 1 has conventional pH dependence of gating with strong voltage-dependent kinetics of activation The pH dependence of gating in EtH V 1 is described by a slope of 0.77 V thres /V rev which translates into a shift  Table 1) with exception of Helisoma trivolvis which reports an anomalous pH dependence of gating [34]. Moreover, the two insects Extatosoma and Nicoletia have identical pH dependence of gating for the same voltage range (Fig. 5B). The conformity of the pH-dependent gating of H V 1 of different species indicates a common pH sensing mechanism which to date is still unknown. However, differences in the offsets among different species are evident. Our analysis shows an offset of −23 mV for EtH V 1. A negative offset of the V thres -V rev relationship reflects an early activation which permit protons to flow from the external solution into the cell. Along the whole pH range tested, EtH V 1 conducts H + inwards consistently. The results contradict the more positive activation of NpH V 1 and mammalian channels, whose physiological roles relate to elimination of excessive cytosolic acidification [37] and compensation of electrical charges during the respiratory burst of phagocytes [41]. The negative activation of EtH V 1 is in contrast more similar to kH V 1 from the dinoflagellate Karlodinium veneficum [2]. In dinoflagellates, inward H + currents acidify the interior of membrane specialized compartments (scintillons) which triggers bioluminescence [5,34]. Hypothetically, EtH V 1 in Extatosoma plays a role in an acidification process or in the generation of action potentials. To analyse the activation of EtH V 1 in more detail, we measured the voltage dependence of EtH V 1 kinetics. We applied linear regressions to τ actvoltage plots at different pH. Results show that EtH V 1 has a stark steepness of the τ act -V relationship of 10.0 mV/ e-fold change, similar to the snail channel HtH V 1, and much stronger than mammalian channels which values vary between 40 and 72 mV/e-fold change [37] (  [34]. The other hexapod proton channel, NpH V 1, activates also in the range of seconds [11]. Interestingly, in the human channel, data suggest that a glutamate at the position of the SF speeds up the channel activation kinetics. The hH V 1-D112E mutant is~5 times faster than the wild-type [32] and also shifts V thres to more negative potentials [20,21], indicating a shift of free energy to open the channel.

Possible physiological role
For a functional analysis of insect proton channels, a detailed cellular expression pattern would be of great importance. So far, only tissue distributions of the H V 1 expression are available. Compared to the Zygentoma Nicoletia phytophila, EtH V 1 showed a more restricted expression pattern in the different tissues tested. In both, Nicoletia and Extatosoma, H V 1 is strongly expressed in the nervous system. Interestingly, no expression in leg muscle was found for Extatosoma, despite it being present in leg and body muscle in Nicoletia. Harrison [42] describes several patterns of acid-base regulation in insects. The passive transport of protons through H V 1 could be related to the pH-homeostasis maintenance in some of these processes in polyneopteran species.
There are pH differences across the digestive system of some insects. In crickets and grasshoppers (Orthoptera), passive distribution of protons across the midgut epithelium is associated with low pH in the lumen [42]. Coincidently, we found a mild expression of EtH V 1 in digestive system (Fig. 2).
Discontinuous ventilation of insects generating variations of partial CO 2 pressure (P CO2 ) is also mentioned. The fluctuations on P CO2 during discontinued ventilation change the pH of the haemolymph, for example in grasshoppers, where haemolymph pH correlates with fluctuations of P CO2 and the nonbicarbonate buffer values [42]. Nevertheless, these pH variations due to discontinuous ventilation are considered small [42].
Other pH-homeostasis changes in insects are associated with periods of activity. In general, the increase of activity, for example during flight, is accompanied by the use of anaerobic metabolism that generates acid production. In locust, for example, tracheal and fluid P CO2 during flight increases two-to threefold in comparison to the resting state [42]. Accumulation of CO 2 due to insect's activity translates to a drop of haemolymph pH of~0.2 units for grasshoppers and even to 0.9 for cockroaches (do not express H V 1) during flight [42]. Despite EtH V 1 was not found in leg muscle, the haemolymph circulates through the whole body of the animal. Hence, we cannot discard the channel involved in pH regulation of the haemolymph during activity periods.
The pH of the haemolymph of some invertebrates decreases linearly with temperature [43]. Similarly, in orthopteran insects (which do have H V 1), the haemolymph pH appears to be also dependent on temperature although the temperature-pH relationship loses linearity. Thus, the orthopterans M. bivittatus and S. nitens are able to keep a constant haemolymph pH at temperatures of 10°C-25°C but the value drops with a rate of 0.017 unitsÁ°C −1 at temperatures higher than 25°C [42]. Transmembrane acid-base transport controlled by the renal system has been suggested to explain this behaviour [44]. EtH V 1 could also play an important role in relation to acid regulation.
Another possible physiological role of EtH V 1 could be related to the sensitivity of chemoreceptors to haemolymph pH. For example, cockroaches (which lack of H V 1) abdominal pumping rates are regulated by the pH of solutions in contact with the nerve cord [45].
Grasshoppers on the other hand possess H V 1 and their ventilation rates are unaltered once the haemolymph pH is changed [46].
The EtH V 1 channel is highly expressed in the nervous system (Fig. 2). Remarkably, H V 1 was first discovered in snail neurons by Thomas and Meech [47]. Subsequent studies in neurons of other snail species [48,49] confirmed the existence of H V 1 presenting τ act of few milliseconds [50]. The activation of EtH V 1 is negative to V rev in the whole pH range tested. It implies that EtH V 1 conducts H + inwardly and therefore could depolarize the cell membrane. In the case that H + conductance is dominant in the membrane of neurons under ionic conditions of the animal at certain membrane potentials, small inward currents could effectively depolarize the neuron to action potential threshold. Proton channels of mammals, activating in the order of seconds, restore pH i of small cells after an acid load in the order of tens of seconds because of their surface/volume ratio [50]. However, a role of EtH V 1 in the generation of action potentials is presumably limited due to its relatively slow activation. In neurons of Locusta migratoria (Orthoptera) for example, the times-to-peak range from~2 to 10 ms [51]. Our data do not confirm or discard the participation of EtH V 1 in the generation of action potentials in Extatosoma. Further in vivo electrophysiological studies in Extatosoma neurons are required to evaluate involved conductances and the effects of pH variations on triggering of action potentials.
A striking difference we found between the hexapodans proton channels NpH V 1 and EtH V 1 is the more negative opening of the later. Consistently, EtH V 1 activates approximately 20 mV more negative than NpH V 1. This means that in comparison with NpH V 1, activation of EtH V 1 presents a shift of free energy that favours the close→open transition due to the influence of the membrane potential. EtH V 1 requires less membrane depolarization to activate. Hypothetically, the natural occurring variation to Glu in the SF of EtH V 1 might be responsible for it. Site-directed mutations of Asp112 to Glu in the SF of the human H V 1 have revealed negative shifts of threshold of activation [20,21]. Mutations of other amino acids in other parts of the channel also provoke ΔV thres to more negative potentials. However, in accordance with a metaanalysis of mutation studies [32], of all Asp mutants in the SF, only the Asp to Glu mutation shifts V thres negatively.
The activation of EtH V 1 is also negative to V rev , which translates to an inward H + current that acidifies the cytosol. Teleologically, EtH V 1 task is related to the Similarly, marine dinoflagellates, whose H V 1 channels activate also negative to V rev , use H V 1 channels to acidify scintillons and trigger bioluminescence [2,5]. The chemistry of the physiological environment must always be considered. Thus, if [Zn 2+ ] is elevated in Extatosoma, then the V thres is shifted to positive potentials. Assuming the voltage shift is sufficient to set V thres positive to V rev , in this case, EtH V 1 functions similar to most known H V 1 and extrude protons out of the cell. Embioptera (Asp in SF) and Phasmatodea (Glu in SF) belong to sister branches with a common ancestor [12]. Interestingly, their cousin branch Mantophasmatodea also has a Glu in the SF [12]. Perhaps the answer to the function of Glu as SF and its relationship with the physiology of the insect lies on the physiological differences between those polyneopteran orders.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Table S1. List of all identified H V 1 homologs compiled from TSA files and correspondent GenBank accession number. Table S2. Sequence identity percentage between species from different polyneopteran H V 1 proteins. Fig. S1. Amino acid sequences of polyneopteran insects proteins possessing a typical S4 RxWRxxR motif. Fig. S2. Alignment of putative polyneopteran H V 1 channels. Fig. S3. Inside-out patch-clamp measurement of EtH V 1.