Phosphocholine – an agonist of metabotropic but not of ionotropic functions of α9-containing nicotinic acetylcholine receptors

We demonstrated previously that phosphocholine and phosphocholine-modified macromolecules efficiently inhibit ATP-dependent release of interleukin-1β from human and murine monocytes by a mechanism involving nicotinic acetylcholine receptors (nAChR). Interleukin-1β is a potent pro-inflammatory cytokine of innate immunity that plays pivotal roles in host defence. Control of interleukin-1β release is vital as excessively high systemic levels cause life threatening inflammatory diseases. In spite of its structural similarity to acetylcholine, there are no other reports on interactions of phosphocholine with nAChR. In this study, we demonstrate that phosphocholine inhibits ion-channel function of ATP receptor P2X7 in monocytic cells via nAChR containing α9 and α10 subunits. In stark contrast to choline, phosphocholine does not evoke ion current responses in Xenopus laevis oocytes, which heterologously express functional homomeric nAChR composed of α9 subunits or heteromeric receptors containing α9 and α10 subunits. Preincubation of these oocytes with phosphocholine, however, attenuated choline-induced ion current changes, suggesting that phosphocholine may act as a silent agonist. We conclude that phophocholine activates immuno-modulatory nAChR expressed by monocytes but does not stimulate canonical ionotropic receptor functions.

In the presence of PC as well as Cho the IL-1β release was inhibited. The inhibitory effect of PC and Cho was dose-dependently antagonized by RgIA4 (*P ≤ 0.05, **P ≤ 0.01, significantly different from cells treated with PC or Cho alone, Mann-Whitney rank-sum test). (b) In LPS-primed U937 cells that were transfected with control siRNA (si) the BzATP-stimulated IL-1β release was inhibited by PC and Cho. In cells transfected with siRNA to Chrna9 or Chrna10, the effects of PC and Cho were blunted (*P ≤ 0.05, different from cells treated with LPS and BzATP; # P ≤ 0.05, ## P ≤ 0.01, different from respective experiments on cells treated with control siRNA; Kruskal-Wallis followed by Mann-Whitney rank-sum test). Data are presented as individual data points, bar represents median, whiskers percentiles 25 and 75.
antagonized by mecamylamine (Mec), α -bungarotoxin and strychnine, suggesting that nicotinic acetylcholine receptors (nAChR) containing α 9 and/or α 10 subunits are involved in signalling 20 . Interestingly, nicotine abolishes ATP-induced ion channel functions of P2X7 receptors in U937 cells, a human monocytic cell line, but does not provoke ion currents itself 20 .
The purpose of this study was to test the hypothesis that binding of PC and Cho to nAChR inhibit P2X7 receptor function similar to nicotine. In addition, we hypothesised that PC is a novel agonist of nAChR containing α 9 and/or α 10 subunits and directly compare the effects of PC to the well-known α 9 * (* indicates the possible presence of additional nAChR subunits) agonist Cho 21 . We provide evidence that PC and Cho induce metabotropic effects via α 9α 10* -containing nAChR in monocytic cells that result in an inhibition of P2X7 receptor function. Canonical ionotropic functions of α 9* nAChR, however, are triggered by Cho but strikingly not by PC.

Results
PC and Cho inhibit BzATP-induced IL-1β release via α9 and α10 nAChR subunits. To test the hypothesis, that the inhibitory effects of PC and Cho are mediated via nAChR containing α 9 and/or α 10 subunits, we used an analogue of α -conotoxin RgIA (RgIA4, synonym CSP-4), a potent and selective antagonist of human α 9* nAChR 22,23 . Human monocytic U937 cells were primed with LPS for 5 h followed by stimulation with the selective P2X7 agonist BzATP (2′ (3′ )-O-(4-benzoyl-benzoyl)ATP trieethylammonium salt) 24 for another 30 min in the presence or absence of PC, Cho and/or RgIA4. Thereafter, IL-1β released into the culture medium was measured by an enzyme linked immunosorbent assay (ELISA; Fig. 1). As expected, PC or Cho (100 μ M each) completely inhibited BzATP-induced IL-1β release (Fig. 1a). RgIA4 fully antagonized the inhibitory effect of PC and Cho in a dose-dependent manner (IC 50 about 10 nM; n = 6; P ≤ 0.05; Fig. 1a).
To corroborate these results, we transfected U937 cells with small interfering RNA (siRNA) to silence the expression of α 9 and α 10 nAChR subunits or with scrambled control siRNA. The efficiency of this treatment in U937 cells was shown before 20 . Transfection of control siRNA neither impaired BzATP-induced release of IL-1β nor altered the inhibitory effects of PC and Cho (n = 4; P ≤ 0.05; Fig. 1b). In contrast, when the expression of α 9 or α 10 subunits was silenced by siRNA the inhibitory effect of PC as well as Cho was blunted (n = 4; P ≤ 0.05; Fig. 1b).

PC and Cho inhibit BzATP-induced ion current stimulation in U937 cells. To investigate if
BzATP-induced ion current stimulation due to P2X7 receptor activation is inhibited by PC and Cho, we performed whole-cell patch-clamp measurements on LPS-primed U937 cells. As shown previously 20 , application of BzATP (100 μ M) consistently induced ion currents (Fig. 3a,d). BzATP-induced ion current stimulation was reversible by washout and repeatable (Fig. 3a,d). No significant changes were detected when comparing the amplitude (∆ I BzATP ) of the first BzATP-induced response with the second (n = 10, P = 0.241; Fig. 3d), indicating that the receptors do not desensitise under these conditions. In the next set of experiments BzATP was first applied alone, which provoked ion currents (Fig. 3b). After washout, the cells were preincubated with Cho (100 μ M) for 30 s, followed by an additional application of BzATP (Fig. 3b). Cho alone did not cause any changes in current (n = 7, Fig. 3b). Moreover, Cho abolished BzATP-induced current stimulation (n = 7, P = 0.018; Fig. 3b,d). In the next experiments, cells were preincubated with the nAChR antagonist Mec (100 μ M), followed by application of Cho (Fig. 3c). Under these conditions (Mec + Cho) a BzATP-induced current stimulation was detectable and the first and second BzATP-induced effects did not differ (n = 5, P = 0.5; Fig. 3c,d).
We performed the same kind of experiments using PC (1 mM) instead of Cho (Fig. 4). PC alone did not induce ion current changes. The BzATP-induced effect was abolished in presence of PC (n = 6, P = 0.028; Fig. 4a,d). Furthermore, the inhibitory effect of PC was antagonized by preincubation of the cells with Mec and the first and second BzATP-induced effect did not differ (n = 6, P = 0.209; Fig. 4b,d). Taken together, we were able to show that PC and Cho inhibit BzATP-induced ion fluxes in U937 cells.
Previous experiments examining the inhibition of BzATP-induced IL-1β release by PC provided evidence that nAChR containing subunits α 9 and/or α 10 are involved (Fig. 1). To confirm these data, we performed whole-cell patch-clamp measurements in the presence of RgIA4 (200 nM) (Fig. 4c). For this purpose, the BzATP-induced current was determined (n = 6; Fig. 4c). After washout of BzATP, cells were preincubated with RgIA4 and followed by PC application (Fig. 4c). Subsequent application of BzATP activated a current that was similar to the In some experiments the expression of nAChR containing the α 9 or α 10 subunit was reduced by using the small interfering RNA (si) transfection. At the end of the experiments, LDH was measured in the cell culture supernatants and is given as % of total release (mean ± standard error of mean, SEM).    PC does not induce ion current stimulation in Xenopus laevis oocytes expressing α9 or α9α10 nAChR. Cho is an agonist of nAChR containing α 9 subunits 21 . To test if PC also evokes ion currents at canonical α 9 containing receptors, human α 9 subunit as well as a combination of human α 9 and α 10 subunits were heterologously expressed in Xenopus laevis oocytes. Two-electrode voltage-clamp (TEVC) measurements were performed to assess ion channel functions of nAChR.
In oocytes expressing homomeric α 9 nAChR, application of Cho (Cho1) resulted in a rapid stimulation of the transmembrane current (I M ; Fig. 5a,c). The effect of Cho was reversible upon washout (Fig. 5a). Subsequent application of Cho (Cho2) activated a current that was significantly smaller than the preceding Cho-induced current (n = 14, P = 0.005; Fig. 5a,c). In contrast, when PC was applied for 2 min, no ion currents were provoked (Fig. 5b). As an internal positive control, Cho was applied at the end of each experiment resulting in a significant current stimulation (n = 12, P = 0.002; Fig. 5c).
The same kinds of experiments were performed on oocytes co-expressing nAChR subunits α 9 and α 10. Application of Cho induced a current stimulation (Fig. 5d,f). In comparison to the Cho-induced effect in oocytes expressing α 9 (Fig. 5a), the current stimulation was faster and shorter (Fig. 5d). The Cho effect was repeatable and significantly blunted compared to the first response (n = 12, P = 0.003; Fig. 5f). Again, application of PC had no impact on the current while Cho induced a current stimulation (n = 10, P = 0.005; Fig. 5e,f). Neither application of Cho (n = 17) nor PC (n = 11) induced changes in currents of water injected control oocytes, which did neither express human α 9 nor α 10 nAChR subunits (Fig. 5g,h).
PC interacts with heterologously expressed α9α10 nAChR. Since PC did not evoke ion currents in oocytes expressing homomeric α 9 and heteromeric α 9α 10 nAChR, we questioned whether PC might function as a silent desensitiser or as an antagonist of α 9* nAChR containing subtypes. Therefore, we performed an additional set of experiments designed to monitor the effects of PC on the Cho-evoked responses over time (Fig. 6a,b). Oocytes expressing heteromeric α 9α 10 nAChR were stimulated with Cho once per min until a  steady-state baseline was achieved. Then, the perfusion solution was switched to one containing 1 mM PC and the Cho-evoked responses were monitored for changes in amplitude. Under these conditions, PC decreased the Cho-evoked responses to 13.1 ± 3.3% (n = 5) of control values after a 20 min perfusion (Fig. 6b). PC was then washed out and Cho-evoked responses were monitored for recovery from inhibition. The responses recovered to 98.2 ± 8.2% (n = 5) of control values after a 15 min perfusion with control solution.

Discussion
In the present study, we identify PC as a novel agonist of monocytic nAChR containing subunits α 9 and α 10. We provide evidence that α 9 and α 10 nAChR subunits are essential for the PC-mediated inhibition of ATP-induced ion channel functions at P2X7 receptor in monocytic cells and hence, for the inhibition of ATP-induced release of IL-1β . Cho, a well-known agonist of nAChR containing α 9 subunits 21 , and PC provoke similar metabotropic but no ionotropic effects in monocytic cells. As expected, Cho induces ion current responses at conventional ionotropic α 9 nAChR homomers or α 9α 10 heteromers. In stark contrast, PC does not provoke any ion current changes at these canonical ionotropic receptors. To the best of our knowledge, we are the first to describe an agonist of nAChR containing α 9 subunits that triggers metabotropic but no ionotropic receptor functions.
In the first part of this study, we demonstrated that PC and Cho are agonists of metabotropic nAChR composed of α 9 and α 10 subunits in monocytic cells. We showed previously that PC and Cho dose-dependently inhibit the ATP-induced release of IL-1β from LPS-primed human monocytic U937 cells via nAChR that are sensitive to Mec, α -bungarotoxin and strychnine 20 . Here, we clarified that PC and Cho act as ligands of monocytic α 9* nAChR by using the potent and selective antagonist RgIA4 which dose-dependently antagonized the inhibition of IL-1β release. These findings were further confirmed by gene-silencing experiments in U937 cells: silencing of Chrna9 expression blunted the inhibitory effects of PC and Cho. The same observation was made, when Chrna10 gene expression was silenced, suggesting that α 9 and α 10 nAChR subunits cooperate in monocytic cells.
The results of the gene-silencing experiments suggested but did not prove that α 9 and α 10 nAChR subunits are essential for the signalling of PC and Cho in monocytes. Therefore to corroborate the role of α 9 nAChR subunit by an independent approach, we investigated freshly isolated PBMC from wild-type and gene-deficient mice. Although primary mouse PBMC were not intentionally primed to induce biosynthesis of IL-1β , they consistently released IL-1β in response to stimulation with BzATP. These observations are in accordance with previous findings on primary human PBMC 20,25 . We assume that freshly isolated PBMC became activated during cell isolation and culture 20 . As expected, PC and Cho efficiently inhibited the BzATP-induced release of IL-1β by PBMC obtained from wild-type mice but did not impair the IL-1β release from PBMC of Chrna9 − /− and Chrna10 − /− mice. From these data, we conclude that α 9 and α 10 subunits are essential for signalling of PC and Cho. Interactions of the α 9 and α 10 nAChR subunits were previously described. Subunits α 10 do not assemble into functional ionotropic homomeric nAChR 26 , while co-expression of α 9 and α 10 nAChR subunits results in formation of functional heteromeric α 9α 10 nAChR [26][27][28] . Transcripts of α 9 and α 10 nAChR subunits have been detected in the auditory system 26,29 , dorsal root ganglion 30 , skin 31 , as well as in mononuclear leukocytes 20,32,33 including human monocytic cells 20 .
Next, we confirmed our hypothesis that PC and Cho inhibit ATP-mediated ion current responses in monocytic cells. In accordance with our previous study 20 , we detected BzATP-induced ion current responses in whole-cell patch-clamp measurements on LPS-primed U937 cells. These currents are most likely due to activation of P2X7 receptors, as BzATP is a specific agonist of this ATP-receptor 34 , and due to a consecutive opening of pannexin hemichannels 35,36 . In line with our hypothesis, BzATP-mediated P2X7 receptor activation was completely abolished in presence of PC and Cho. This inhibitory effect was antagonized by the general nicotinic antagonist Mec and the α 9* -specific conotoxin RgIA4. These results corroborate the involvement of α 9 and/or α 10 nAChR subunits in signalling of PC. A functional interaction of other subtypes of nAChR and P2X receptors was demonstrated previously in neurons 37 and in heterologous expression systems 38,39 . In these studies, co-application of ATP and nicotinic agonists evoked current responses that were smaller than the sum of the individual currents induced by ATP and ACh or nicotine 37,38 . In the present study, PC or Cho alone did not evoke any ion current responses in U937 cells, consistent with functional coupling of a non-canonical metabotropic nAChR to ionotropic P2X7 receptors. While in excitable cells such as neurons nAChR are ligand-gated ion channels, no ionotropic nAChR functions have been observed in leukocytes 20,32,33,40 . At present, we do not know how activation of monocytic nAChR by PC or Cho translates into the observed inhibition of P2X7 receptor function.
In the last part of the study, we investigated the effect of PC at canonical ionotropic nAChR using Xenopus laevis oocytes as a heterologous expression system for human homomeric α 9 nAChR as well as heteromeric α 9α 10 nAChR. Cho was included in these experiments as a positive control and transmembrane ion currents were recorded in TEVC measurements. We detected Cho-induced current responses in oocytes expressing homomeric α 9 and heteromeric α 9α 10 nAChR. Current responses to the first Cho application for 30 s varied from 50 nA to 1200 nA (see Fig. 5a,f). Subsequent application of Cho resulted in smaller current responses, indicating receptor was smaller compared to the first one (Cho1) indicating receptor desensitization. (b,e) Initial application of phosphocholine (PC, 1 mM, white bars) had no impact on IM, whereas application of Cho thereafter induced a current stimulation. Again, oocytes expressing only α9 and those co-expressing α9α10 nAChR subunits led to similar results. (g,h) Representative current traces of water injected control oocytes (no expression of human receptors). Neither repeated application of Cho (n = 17), nor PC (n = 11) induced any changes in I M . Depicted are representative current curves (a,b,d,e,g,h). All changes of the transmembrane current (∆I M ) induced by cholinergic stimulation are shown as individual data points, bars represent median, whiskers percentiles 25 and 75 (c,f). Statistical analyses were performed using the Wilcoxon signed-rank test.
Scientific RepoRts | 6:28660 | DOI: 10.1038/srep28660 desensitisation. This is in contrast to previous studies and to our own observations, where application of ACh or Cho in short 1 s pulses resulted in repeatable current responses without desensitisation 27,41 . We assume that receptor desensitisation is a consequence of extended exposure of nAChR to Cho.
In sharp contrast to Cho, PC did not evoke ion current responses in α 9 and/or α 9α 10 nAChR expressing oocytes. We conclude that at least in our experimental settings, PC is an agonist of metabotropic nAChR containing α 9* subunits, whereas it does not stimulate canonical ligand-gated nAChR functions. This finding suggests that PC acts as a potent regulator of innate immunity but does not activate current responses in neuronal or non-neuronal cells expressing canonical ionotropic nAChR containing α 9 subunits. Hence, PC and other molecules containing a PC group may be promising therapeutics for the prevention and treatment of excessive inflammation involving IL-1β such as life-threatening SIRS, without entailing the risk of adverse effects involving excitable cells.
Although PC did not trigger ion channel functions at α 9 and α 9α 10 subunit containing nAChR expressed by Xenopus laevis oocytes, we obtained evidence that PC interacts with these canonical receptors and hypothesised that PC might act as a silent agonist. Per definition, silent agonists desensitise receptors without activating their function 42,43 . To further asses this, we used α 9α 10 nAChR expressing oocytes in an experimental setup that enabled application of Cho in 1 s pulses every 60 s, where Cho induced repetitive current responses without receptor desensitisation. We found that responses to Cho at heteromeric α 9α 10 nAChR were blunted in the presence of PC. The observed slow kinetics of inhibition and recovery from inhibition by PC was consistent with silent desensitisation. However, we should point out that our results do not rule out that the observed inhibition by PC is due to simple antagonism. Nicotine is an example of a ligand, which acts as an antagonist at ionotropic receptors 44 containing α 9 subunits, but acts as an agonist at monocytic metabotropic α 9 nAChR that inhibit P2X7 function 20 . A silent agonist for homomeric nAChR was recently reported for homomeric α 7 nAChR 43 . These authors showed that ACh-induced current responses were reduced after preincubation of the oocytes with compound NS6740 43 .
In conclusion, we identified PC as a novel agonist of metabotropic nAChR containing α 9 and α 10 subunits. PC and Cho evoke no ion current responses at these receptors expressed by monocytes but efficiently inhibit ATP-mediated P2X7 receptor activation and release of IL-1β . In contrast to Cho, PC does not trigger ionotropic functions at canonical human α 9 nAChR homomers and α 9α 10 nAChR heteromers. These findings suggest that PC may be a valuable active substance for the treatment of inflammatory diseases that targets nAChR of monocytes without disturbing ionotropic functions of excitable cells.  To investigate IL-1β release cells were transferred to 24-well plates (1 × 10 6 cells/ml and per well). Cells were primed with 1 μ g/ml LPS from Escherichia coli (L2654; 1 μ g/ml; Sigma-Aldrich, Deisenhofen, Germany) for 5 h. After priming, the P2X7 receptor agonist BzATP (Sigma-Aldrich; 100 μ M) was added for 30 min in presence or absence of different concentrations of cholinergic agonists and antagonists. Cho chloride (100 μ M), PC chloride calcium salt tetrahydrate (100 μ M), and Mec hydrochloride (100 μ M) were purchased from Sigma-Aldrich. An analogue of α -conotoxin RgIA (RgIA4) 22 was used in concentrations from 0.2 to 200 nM. After cell treatment, cells were spun down (500 g, 8 min) the supernatants were collected and stored at − 20 °C. IL-1β concentrations were measured using a human Quantikine Immunoassays (R&D Systems, Minneapolis, MN) and LDH was determined.

Methods
Silencing of α9 and α10 nAChR subunit expression. In some experiments the expression of nAChR containing the α 9 and/or α 10 subunits in U937 cells was reduced by using siRNA technology. U937 cells were transfected with ON-TARGETplus human Chrna9 or Chrna10 siRNA SMARTpool (Thermo Fisher Scientific, Schwerte, Germany). As a control for unspecific effects of siRNA transfection cells were transfected with negative control ON-TARGETplus Non-targeting Control Pool (Thermo Fisher Scientific). In accordance with the manufacturer's protocol, all cells were transfected with 30 pM siRNA/1 × 10 6 cells using Amaxa Cell Line Nucleofector Kit C and Nucleofector II Device (both from Lonza Cologne, Cologne, Germany). 48 h after siRNA transfection, IL-1β release experiments were performed.
Mononuclear leukocytes from Chrna9 and Chrna10 gene-deficient mice. Male and female gene-deficient Chrna9 (129S-Chrna9 tm1Bedv /J) 45 and Chrna10 (129S4-Chrna10 tm1Bedv /Mmucd) 46 as well as corresponding WT mice (for details see 45,46 ) were used for isolation of PBMC. Experimental animals received humane care according to NIH "Guide for the Care and Use of Laboratory Animals". Animal experiments were approved by the local committee at the Regierungspräsidium Giessen, Hesse, Germany (permit No. Gi 20/23-A10/2011).
Mice were euthanized by neck dislocation, and blood was drawn from the caval vein into heparinized syringes. PBMC were separated by discontinuous Percoll (Ge Healthcare Bio-Sciences AB, Uppsala, Sweden; 1.082 g/ml) density gradient centrifugation and cultured for 2 h in RPMI 1640 (Gibco by Life Technologies GmbH) supplemented with 10% FCS (Biochrome) and 2 mM L-glutamine (Gibco by Life Technologies GmbH), at 5% CO 2 and 37 °C. For investigation of IL-1β release, BzATP (Sigma-Aldrich; 100 μ M) was added for 30 min in the presence or absence of PC (100 μ M) or Cho (100 μ M). Subsequently, cell culture supernatants were collected and stored at − 20 °C. Finally, IL-1β concentrations were measured by using mouse Quantikine IL-1β Immunoassay (R&D Systems).

LDH measurements.
In order to test for cell viability, activity of the cytoplasmic enzyme LDH was assayed by the Non-Radioactive Cytotoxicity Assay (Promega, Madison, WI) according to the supplier's instructions. LDH in the cell culture supernatants of U937 cells and PBMC was measured at the end of the experiments. For calculating the proportion of dead cells, a maximum LDH release control was generated. For this purpose, U937 cells were lysed by freezing them twice (− 80 °C). Subsequently, the samples were analysed according to the supplier's instructions and values determined in cell culture supernatants were compared with the total content of LDH in lysed cells.
Oocytes of stages V and VI (Dumont 1972) were injected with cRNA encoding α 9 nAChR subunits (20 ng per oocyte) or α 9α 10 nAChR subunits (each 20 ng per oocyte) using a microinjector (Nanoject, Drummond Scientific, Broomall, USA). In order to increase expression levels and obtain stable nAChR expression, cRNA encoding RAPSN (5 ng per oocyte) was co-injected in both cases 47,48 . All cRNA was dissolved in nuclease-free water. The injection volume was 50.6 nl. In all TEVC experiments representative controls were performed with oocytes that were injected with 50.6 nl of nuclease-free water.
After an incubation time of 3-5 days, the transmembrane currents (I M ) of water-or RNA-injected oocytes were recorded by the TEVC technique. Oocytes were placed in a perfusion chamber and perfused (gravity driven) with ORi containing (in mM): 90 NaCl, 1 KCl, 2 CaCl 2 , and 5 HEPES (pH 7.4). Intracellular microelectrodes were pulled from borosilicate glass capillaries and filled with 1 M KCl solution. The membrane voltage was clamped to − 60 mV using a TEVC amplifier (Warner Instruments, Hamden, USA), and transmembrane currents were low-pass filtered at 1000 Hertz (Frequency Devices 902, Haverhill, Massachusetts, USA) and recorded with a strip chart recorder (Kipp & Zonen, Delft, The Netherlands). In all experimental groups, measurements were performed on oocytes from at least two different Xenopus laevis individuals.
For experiments examining the effects of continuous exposure to PC on α 9-and α 10-mediated currents, oocytes were injected with a 1:1 ratio of cRNA for human α 9 and α 10 nAChR subunits and incubated at 17 °C for 3 days. To conduct TEVC experiments, the oocytes were placed in a 30 μ l chamber and continuously perfused by gravity with a solution containing 96 mM NaCl, 2.5 mM KCl, 1.8 mM CaCl 2 , and 1 mM MgCl 2. The pH of the solution was adjusted to 7.4 with NaOH. A stock solution of 100 mM PC was prepared in distilled water. A working solution of 1 mM PC was prepared in perfusion solution containing lower CaCl 2 (0.8 mM) such that the final concentration of Ca 2+ in all solutions was 1.8 mM. The oocyte membranes were clamped at a holding potential of − 70 mV and stimulated with 1 s pulses of 1 mM Cho once every 60 s until a steady-state baseline response was observed. The perfusion solution was then switched on one containing 1 mM PC and the oocytes stimulated with 1 s pulses of 1 mM Cho plus 1 mM PC and the Cho-evoked responses monitored for changes in amplitude. The data for inhibition of the Cho-evoked responses were normalized to 3 averaged control pulses and analysed with an exponential decay equation. Data for recovery from inhibition were analysed with an exponential association equation. The data for inhibition by PC were best fit with a double exponential and the data for recovery from inhibition with a single exponential.
Statistical analyses. Data were analysed with the SPSS software (Munich, Germany) or GraphPad Prism 6 software (Ja Lolla, CA, USA). Values derived from different cells were compared, where applicable, by the nonparametric Kruskal-Wallis test, followed by the Mann-Whitney rank-sum test. The Wilcoxon signed-rank test was used for analyses of dependent values. The number (n) of individual experiments is indicated in the Results section and the Figures. In TEVC measurements oocytes from at least two different Xenopus laevis individuals were used.