TRPV1 activation power can switch an action mode for its polypeptide ligands

TRPV1 (vanilloid) receptors are activated by different types of stimuli including capsaicin, acidification and heat. Various ligands demonstrate stimulus-dependent action on TRPV1. In the present work we studied the action of polypeptides isolated from sea anemone Heteractis crispa (APHC1, APHC2 and APHC3) on rat TRPV1 receptors stably expressed in CHO cells using electrophysiological recordings, fluorescent Ca2+ measurements and molecular modeling. The APHCs potentiated TRPV1 responses to low (3–300 nM) concentrations of capsaicin but inhibited responses to high (>3.0 μM) concentrations. The activity-dependent action was also found for TRPV1 responses to 2APB and acidification. Thus the action mode of APHCs is bimodal and depended on the activation stimuli strength—potentiation of low-amplitude responses and no effect/inhibition of high-amplitude responses. The double-gate model of TRPV1 activation suggests that APHC-polypeptides may stabilize an intermediate state during the receptor activation. Molecular modeling revealed putative binding site at the outer loops of TRPV1. Binding to this site can directly affect activation by protons and can be allosterically coupled with capsaicin site. The results are important for further investigations of both TRPV1 and its ligands for potential therapeutic use.


Introduction
TRPV1 is the founding member of a TRP channel sub-family, which has six members divided into two groups: thermo-sensitive, non-selective ion channels (TRPV1-TRPV4) and channels highly selective for Ca 2+ ions (TRPV5-TRPV6) (see [1] for review). TRPV1 is known as the capsaicin receptor or vanilloid receptor 1. It is the most thoroughly studied channel among TRPVs. TRPV1 mediate pronounced cation influx and have weak selectivity to Ca 2+ . Various activation stimuli including capsaicin, resiniferatoxin, heat, H + , endocannabinoid lipids such as anandamide, eicosanoids, and 2APB were identified as current agonists of TRPV1 [2][3][4][5][6]. PLOS  coli BL21 (DE3) cells expressing thioredoxin fusions of polypeptides were cultured overnight at 25˚C, harvested, ultrasonicated in a buffer for metal-affinity chromatography, and centrifuged to remove all insoluble particles. Fusion proteins were purified using a TALON Superflow Metal Affinity Resin (Clontech, USA) and subjected to CNBr cleavage, as described [30]. The recombinant polypeptides were purified on a reverse-phase column Jupiter C 5 (Phenomenex, USA) 250×10 mm. The target polypeptides' purity was verified by MALDI-TOF massspectrometry.
Fluo-4 based intracellular calcium assay CHO (Chinese hamster ovary) cells were obtained from Evrogen company (http://evrogen. ru). The CHO cell lines stably expressing rat TRPV1 was generated using T-Rex System (Invitrogen) according to the manufacturer's instructions and fluorescent assays were performed using NOVOstar (BMG LABTECH, Germany) as described [31]. CHO cells stably expressing rat TRPV1 were seeded into black-walled, clear-bottomed 96-well plates at a density of 75,000 cells per well (complete media without antibiotics and containing 1 μg/ml tetracycline to induce channel expression) and were cultured overnight at 37˚C. The cells were then loaded with the cytoplasmic calcium indicator Fluo-4AM using Fluo-4 Direct™ Calcium Assay Kits (Invitrogen) and incubated in the dark at 37˚C for 60 minutes and then at 25˚C for 60 minutes. The buffer alone (control) or APHC-polypeptides were added to the cells. Changes in cell fluorescence (λex = 485 nM, λem = 520 nM) were monitored before and after the addition of TRPV1 agonist (capsaicin or 2APB). The measurements were performed at pH 7.4 and 25˚C.

Electrophysiology
Recordings were made by a patch-clamp technique in whole-cell configuration using CHO cells stably expressing rat TRPV1. Whole-cell currents were recorded and digitized by Instru TECH LIH 8+8 data acquisition system (HEKA, Germany) at -80 mV holding voltage using EPC-8 (HEKA, Germany) amplifier, filtered at 5 kHz, sampled at 10kHz and stored.

Data analysis and statistics
All data were expressed as the mean ± standard deviation (SD). Significance of the effect was tested by ANOVA followed by Tukey's post hoc analysis with the p-value of 0.05.

Results
The cell line stably expressing rat TRPV1 was used for electrophysiological and internal Ca 2+ fluorescent recordings. Application of TRPV1 agonists (capsaicin, 2APB or low pH) evoked currents with the reversal voltage close to zero mV. Capsaicin-evoked currents were reversibly blocked by 10 μM of antagonist of TRPV1 channels capsazepine (data shown in S1A Fig).
Application of the APHC1, APHC2, APHC3 (100 nM) without activating stimuli did not cause any detectable response. In Fluo-4-based intracellular calcium assay experiments, capsaicin or 2APB were used as channel agonists. Cell responses to capsaicin were fully blocked by 10 μM capsazepine (S1B Fig

Modulation of capsaicin-evoked currents
Recombinant analogs of sea anemone polypeptides APHC1, APHC2, APHC3 were previously described as selective inhibitors of TRPV1 in several expression systems [23,24]. In our present experiments, 100 nM of APHC1, APHC2 and APHC3 caused 22±8, 41±14 and 61±15% inhibition of the response caused by 3 μM of capsaicin, respectively (Fig 1A-1C). Increase of APHC concentrations up to 3 μM did not induce complete inhibition of capsaicin evoked currents. The saturation level of APHC3 effect estimated from concentration-dependence was 71 ±6% and IC 50 value was 18±4 nM. Maximal inhibition for APHC2 was 42±12% and IC 50 value was 23±9 nM ( Fig 1H). The same analysis for APHC1 estimated the maximal inhibitory effect 31±9% and IC 50 value 60 ±20 nM that were similar to a result obtained earlier for APHC1 in an electrophysiology study in oocyte expression system [24]. Thus, all polypeptides were found to be potent and incomplete inhibitors, but APHC3 induced the maximal inhibition and had the lowest IC 50 value. Next, we studied the action of the polypeptides on TRPV1 responses evoked by low (0.3 μM) capsaicin concentration (Fig 1D-1F). Surprisingly, responses to 0.3 μM capsaicin were markedly potentiated by all three polypeptides: 250±80%, 100±40% and 150±50% potentiation was found for 100 nM APHC1, APHC2 and APHC3 correspondingly (Fig 1D-1F). Thus, direction of APHCs action was dependent on capsaicin concentration ( Fig 1I). Concentration-dependence of an inhibition of TRPV1 currents activated by 3 μM capsaicin is shown in Fig 1H. Application of 1 nM APHCs' concentration was not effective. Effect at APHCs' concentrations above 1000 nM reached saturation level and was not statistically differ from the one at 100 nM peptide's concentration. Rundown of the responses together with large cell-tocell variations did not allow to made precise estimation of IC 50 parameters.
Analysis of the voltage-dependence of action of APHC1 was performed by 5 sec ramp protocol ( Fig 1G). Percentage of current inhibition by 100 nM APHC1 decreased from 27±7% at -70 mV to 21±5% (n = 5) at +50 mV holding voltage. Potentiation of responses evoked by 0.3 μM capsaicin demonstrated the same trend. The effect decreased from 250±80% at -80 mV to 180±60% at +5 mV. However, in both cases the effect reduction at positive voltages was not statistically significant (P>0.1). The reversal potential of capsaicin evoked responses was close to zero mV for all conditions.

Calcium assay experiments
The total calcium responses of cell population to different capsaicin concentrations were measured in control and in the presence of 500 nM of polypeptides. APHC1-3 alone did not  produce any response (data not shown). All three polypeptides demonstrated potentiating activity for TRPV1 activation by 3 nM capsaicin (Fig 2A and 2C); no significant effect was found for 30 nM-1 μM capsaicin ( Fig 2C) and partial inhibition was found in the case of TRPV1 activation by 3 and 10 μM capsaicin (Fig 2B and 2C). The summary plot of capsaicindependence of the APHC action is presented in Fig 2C. For APHC3 peptide the dose dependent potentiation was measured for 3 nM stimulus of capsaicin (S3 Fig). Small fluorescence response was detected both for control and APHCs' treated cells, however 13-30% average potentiation by APHC3 in 65-2000 nM range was observed. ] responses were measured as changes in fluorescence intensity before (FI base ) and after agonist addition (FI). The data shown are representative average plots (n = 4) of the fluorescence signals against time during assays. C, Chart values of inhibiting/activating effects of APHC 1-3 (500 nM) relative to control in dependence on capsaicin concentration applied. In all graphs the values are given as mean ± SD of at least 5 independent experiments. Reliability (*) was checked by ANOVA followed by Tukey's post hoc analysis, p = 0.05. D, "Dose-response" of TRPV1 activation by capsaicin alone (black squares) and in the presence of 500 nM APHC3 (red squares) determined in Fluo-4 based intracellular calcium assay tests. Relative responses were measured as (FI-FI base )/FI base , where FI is the measured peak fluorescence intensity after agonist addition, FI base -fluorescence intensity of the cells before agonist addition. Data are expressed as mean ± SD (n = 4-8).
https://doi.org/10.1371/journal.pone.0177077.g002 Thus, the results obtained in electrophysiological experiments on individual cells correlated with the results of Fluo-4-based intracellular calcium assay experiments, although certain quantitative differences were observed. Fig 2D presents the experimentally obtained dependence of responses in control and in the presence of 500 nM APHC3 on capsaicin concentration. Potentiation at low concentrations and inhibition at high concentrations can be interpreted as increase of capsaicin affinity and decrease of its efficacy in the presence of the toxin.

Modulation of the response evoked by 2APB or low pH
Taking into account the dependence of APHCs mode of action (potentiation or inhibition) on capsaicin concentration, we used different concentrations of protons and 2APB to further characterize action of APHC1, APHC2 and APHC3. The currents evoked by drops of extracellular pH from 7.3 to 6.2 were significantly potentiated by application of 100 nM APHC2 (200±80%) and APHC3 (180±40%). Representative currents for APHC3 are presented in Fig 3A. In contrast, APHC1 did not cause significant effects on the responses evoked by the pH drop. This potentiating effect of APHC2 and APHC3 polypeptides decreased when the channel was activated by stronger acidifications. The effect of APHC3 (100 nM) was reduced to 80±20% at pH 5.5 and became non-significant at pH 4.5 (Fig 3B and 3C). APHC2 showed the same tendency.  Effects of APHCs on TRPV1 activated by different concentration of 2APB were qualitatively similar to their effect on responses evoked by pH drops. Polypeptides at 100 nM concentration potentiated the currents activated by low concentrations of 2APB (30 μM)-160±40%, 170±30% and 170±50% for APHC1, APHC2 and APHC3 correspondingly. However, currents evoked by 300 μM 2APB were not affected significantly, although certain inhibitory trend was observed (see Fig 3D and 3E for APHC3 action).
Since APHC3 polypeptide was the most active, we studied its action on the TRPV1 activated by co-application of capsaicin with 2APB or with low pH. The results shown in Fig 4 demonstrate that action of 100 nM APHC3 on response evoked by 3 μM capsaicin together with 0.3 mM 2APB is similar to that response evoked by 2APB alone. In both cases the effect of APHC3 was not statistically significant. The inhibitory action of APHC3 on channel activation by 3 μM capsaicin with low pH was approximately equal to the polypeptide effect at channel activation by capsaicin alone, contrary to APHC3 potentiating action at application of pH 6.2 alone. Thus, effects of the APHCs on the response activated by combinations of different stimuli are non-additive.

Kinetics of APHC action
Existence of two opposite effects (potentiation and inhibition) raises the question about site(s) of action. Indeed, the effects could be induced by complex allosteric interaction with a single binding site or by independent binding to distinct sites. To check a possibility of APHC binding to distinct sites we studied kinetics of action in the presence of capsaicin. Fig 5 demonstrates the representative results. APHC3 effects on the currents evoked by low ( Fig 5A) and high (Fig 5B) capsaicin concentrations as well as potentiation of the proton evoked current (Fig 5C) developed monotonously and were well-fitted by single-exponential function. Development of the effects was much slower than the overall speed of the application system (see S4  Fig). Recovery from the effects was also monotonous and fast in all cases suggesting that APHC3 does not significantly affect TRPV1 desensitization.  Moreover, kinetics of potentiation and inhibition of the capsaicin-evoked responses did not differ significantly (Fig 5D). It is unlikely that two independent effects mediated by APHC binding to distinct sites can demonstrate matching kinetics. Thus, our data suggest that inhibition and potentiation of capsaicin-evoked responses are caused by APHC binding to the single site. The kinetics of APHC3 action on proton-evoked currents was significantly slower (P<0.05). The time constants were about 2 sec in the case of capsaicin-evoked responses and were more than 4 sec for proton-evoked currents (Fig 5D).

Molecular modeling
To reveal possible binding site of the APHC-polypeptides and predominant structural determinants of their action we employed the molecular modeling approach. This became possible due to a recent resolution of the atomic-scale structure of rat TRPV1 in different functional states [28]. Since 3D structures of the APHC-polypeptides are unavailable we used a model based on the structure of highly homologous trypsin inhibitor from sea anemone S. helianthus SHPI-1 (PDB code 1SHP) (Fig 6A). We docked APHC1 to the closed TRPV1 structure (PDB code 3J5P) using a two-stage protocol. At the first stage 30,000 random positions and orientations of the APHC1 in the vicinity of the TRPV1 outer pore were generated. Each structure  was shortly optimized to remove steric clashes. For the next stage of docking we selected 100 best-energy structures. These structures were optimized until 1000 consecutive energy minimization did not improve the complex energy. For the minimal-energy complex ( Fig 6B) the key interacting pairs of residues were determined (Fig 6C).
In this optimal binding mode, the C-terminal helix of the APHC1 is incorporated between P-loops of two adjacent subunits. APHC1 interacts significantly with the pore helix and two external loops of TRPV1, which are involved in the extracellular gating and undergo significant conformational rearrangements. The strongest toxin-channel interactions in the model are provided by positively charged Arg residues in the C-terminal helix (positions 48, 51 and 55). Arg48 interacts with Glu648, Glu651 and Tyr653. Arg51 interacts with Glu636 and  Tyr627 of another subunit. The terminal Arg55 interacts with Asp646 of this subunit. There are also significant interactions between Lys28 of the toxin and Asp601 of the channel and between Glu45 and Lys656. These interacting residues are shown in Fig 6C. Thus, strong interactions were found in the model between the toxin molecule and adjacent subunits of TRPV1.
To reveal possible determinants of the activity-dependent action of APHCs we performed the docking procedure for the TRPV1 open state model (PDB code 3J5Q). The superimposition of two minimal-energy binding models for the channel open and close states (Fig 6D) showed similar interactions in both cases. However, since the loop between pore helix and S6 segment has different conformations in the open and close state, the slot, which accommodates the C-helix of APHC1, is narrower in the open state than in the closed state. As a result, in the open state APHC1 is pushed about 2.5 Å in the external direction ( Fig 6D). Thus, conformation rearrangements associated with external gating change the binding position of the APHCs and vice versa; APHCs binding can affect equilibrium between close, intermediate and open states.

Discussion
The unique properties of TRPV1, as a sensor and an integrator of different stimuli, make it a very exciting target for the study and treatment of pain as well as other pathological states [36,37]. The usual directions of drug development searching for potent agonists and antagonists of TRPV1 were significantly discredited because of the side effects of new compounds [36]. On the other hand, the multimodality of TRPV1 activation provides additional opportunities for drug design [38]. Therefore, understanding the mechanisms underlying the effects of partial antagonists is an important step to developing effective drugs affecting TRPV1 function.
Usually, considering the ligand-receptor interaction implies that the ligand can induce stable and invariable effects on the receptor-activation, potentiation or inhibition. In the present work we aimed to characterize action of recently discovered sea anemone peptides APHC1-3 on the TRPV1 channels. The main finding of our work is that action of these polypeptides depends not only on the nature of activation stimuli, which is known for many TRPV ligands, but also on the strength of the stimuli. At low activation-strength stimuli, all APHC-polypeptides mainly potentiated TRPV1 response while at increasing activation strength the potentiating effects disappeared or switched to inhibition in the case of activation by capsaicin.
Although both electrophysiological and Ca 2+ assay approaches have revealed activitydependence of the APHCs action, some quantitative details do not coincide. For example, the range of agonist concentrations resulted to potentiation or inhibition effect mismatched for electrophysiological recordings and fluorescent Ca 2+ measurements. Activation of TRPV1 by 0.3 μM capsaicin in the presence of APHCs has significant potentiation effect in electrophysiological experiments (Fig 1D-1F) while at the same capsaicin concentration in calcium assay experiments ( Fig 2D) the effect was close to zero, and potentiating was observed at lower agonist concentrations. Also it should be emphasized that APHC3 produce unequal effects on pH activation in different experimental conditions; it could either potentiate pH response in electrophysiological experiments (this paper) or inhibit it in single cell Ca 2+ imaging [23]. Such dissimilarities could occur due to differences in the experiment conditions. For example, in patch clamp experiments the membrane voltage is kept constant, whereas in calcium measurements activation of TRPV1 and their modulation by polypeptides obviously affect the voltage and thus the calcium influx. As a result of this negative feedback the results of calcium measurements are not proportional to the conductance. From the other side, voltage clamping is artificial approach and results of calcium measurements may be more relevant to physiological conditions. Important question is to discriminate direct effect of a drug or toxin and indirect influence mediated by secondary messengers. For instance, it has been shown that Gigantoxin I isolated form sea anemone potentiates TRPV1 by modulation of the PLA2 pathway [39]. From the other side, a bivalent tarantula toxin activates TRPV1 by direct action [21]. We cannot rule out a possibility of indirect action but our data favor direct mechanism. First, kinetics of APHCs action is rather fast. Potentiation and inhibition develop approximately as fast as response to capsaicin or the pH drop. Recovery from the APHC action also takes place in timescale of seconds (see Fig 5). Second, in the whole-cell patch clamp experiments the intracellular content is largely dissociated and involvement of intracellular messengers in the modulation is unlikely. Therefore we concentrated on the putative mechanisms of direct action.
Mechanism of APHCs action is complex. Our data suggest that in the presence of APHCs the agonist apparent affinity is increased but the efficacy is decreased. It means stabilization of some intermediate non-conducting state of the channel. Rationalization of these data is possible in view of recently proposed double-gate scheme of TRPV1 activation, which is based on atomic-scale structures of TRPV1 [28,29]. In this scheme, two distinct gate regions recognized by cryo electron microscopy [28] suggest four principal states of the receptor. In the absence of activating stimuli both gates are closed. Binding of an activating ligand causes sequential opening of both gates, which are coupled allosterically. According to the results of our modeling calculations, APHCs bind to the external region involved in the proton activation and can directly affect opening of the external gate of TRPV1. This can promote allosteric opening of external gate after capsaicin binding and can cause potentiation of proton binding. The mechanism of decreasing of apparent capsaicin efficacy revealed in our experiments is unclear. There are several possibilities, which cannot be discriminated at present. First, positively charged toxin residues in our model occurs rather close to the permeation pathway and can cause a reduction of channel conductance. Second possibility is that open-time or fast desensitization characteristics of the external gate are affected by the toxin binding. The third possibility is that allosteric coupling between external and internal gates is affected by APHCs that results in stabilization of an intermediate non-conductive state in the double-gate scheme of activation. Further studies are required to resolve these uncertainties.
Modulators of TRPV1 activation can produce positive therapeutic effects without common side effects such as hyperthermia. Response modulation to low pH stimuli is considered an important factor determining the hyperthermia effect of the TRPV1 antagonists in vivo.
Antagonists that are able to inhibit pH-induced TRPV1 currents elicit a hyperthermic effect, while potentiation of the pH-induced activation of TRPV1 either decrease or do not change body temperature [8,40]. APHC1 did not produce significant effects on pH activation of TRPV1 and caused significant decrease of body temperature in vivo tests [23]. APHC3 is able to produce dissimilar effect on pH activation from potentiation (this study) to inhibition and this result in moderate decrease of core body temperature [23]. It should be emphasized that APHC3 produce dissimilar effects on pH activation in the different experimental conditions; it could either potentiate pH response in electrophysiological experiments (this paper) or inhibit it in single cell Ca 2+ imaging [23].
Thus reliable prediction of in vivo effects of APHC polypeptides cannot be made at present time taking into account the complexity of physiological processes associated with TRPV1 and the complexity of polypeptides action demonstrated above. In particular, we showed that action of APHC3 on the TRPV1 currents evoked by a mixture of capsaicin and 2APB or capsaicin and acidic pH is not a simple sum of its effects on the currents evoked by individual stimuli. Nevertheless, modulators with dualistic effects may have a distinct advantage in terms of their practical medical application. Such drug compounds don't inhibit normal receptor function. Only in the case of pathologically strong activation the desired therapeutic effect will occur.