Voltage-gated sodium channel isoform-specific effects of pompilidotoxins Emanuele Schiavon 1, *, Marijke Stevens 2, *, Andre ´ J. Zaharenko 3 , Katsuhiro Konno 4 , Jan Tytgat 2 and Enzo Wanke 1 1 Dipartimento di Biotecnologie e Bioscienze, Universita ` di Milano-Bicocca, Milan, Italy 2 Laboratory of Toxicology, University of Leuven, Belgium 3 Center of Biotechnology, Instituto de Pesquisas Energe ´ ticas e Nucleares & Depto. de Fisiologia, Instituto de Biocie ˆ ncias, Universidade de Sa˜o Paulo, Brazil 4 Institute of Natural Medicine, University of Toyama, Japan Introduction Voltage-gated sodium channels (VGSCs) play a major role in the generation and propagation of action poten- tials in all excitable tissues. They are composed of a pore-forming a-subunit associated with up to four known different b-subunits. To date, nine mammalian Na v channel isoforms (Na v 1.1–Na v 1.9) have been func- tionally characterized. They are differentially distributed in the central and peripheral nervous system, in skeletal Keywords channel isoforms; ion channels; sodium channel inactivation; toxin binding; wasp toxins Correspondence E. Wanke, Dipartimento di Biotecnologie e Bioscienze Universita ` di Milano-Bicocca, Piazza della Scienza 2, 20126 Milano, Italy Fax: +39 02 64483565 Tel: +39 02 64483303 E-mail: enzo.wanke@unimib.it *These authors contributed equally to this work (Received 13 October 2009, revised 13 November 2009, accepted 4 December 2009) doi:10.1111/j.1742-4658.2009.07533.x Pompilidotoxins (PMTXs, a and b) are small peptides consisting of 13 amino acids purified from the venom of the solitary wasps Anoplius samari- ensis (a-PMTX) and Batozonellus maculifrons (b-PMTX). They are known to facilitate synaptic transmission in the lobster neuromuscular junction, and to slow sodium channel inactivation. By using b-PMTX, a-PMTX and four synthetic analogs with amino acid changes, we conducted a thorough study of the effects of PMTXs on sodium current inactivation in seven mammalian voltage-gated sodium channel (VGSC) isoforms and one insect VGSC (DmNa v 1). By evaluating three components of which the inactivat- ing current is composed (fast, slow and steady-state components), we could distinguish three distinct groups of PMTX effects. The first group concerned the insect and Na v 1.6 channels, which showed a large increase in the steady- state current component without any increase in the slow component. Moreover, the dose-dependent increase in this steady-state component was correlated with the dose-dependent decrease in the fast component. A sec- ond group of effects concerned the Na v 1.1, Na v 1.2, Na v 1.3 and Na v 1.7 iso- forms, which responded with a large increase in the slow component, and showed only a small steady-state component. As with the first group of effects, the slow component was dose-dependent and correlated with the decrease in the fast component. Finally, a third group of effects concerned Na v 1.4 and Na v 1.5, which did not show any change in the slow or steady- state component. These data shed light on the complex and intriguing behavior of VGSCs in response to PMTXs, helping us to better understand the molecular determinants explaining isoform-specific effects. Abbreviations AFT-II, Anthopleura fuscoviridis toxin; ATX-II, Anemonia sulcata toxin; Bc-III, Bunodosoma caissarum toxin; Cn2, Centruroides noxius toxin; PMTX, pompilidotoxin; a-PMTX, a-pompilidotoxin; b -PMTX, b-pompilidotoxin; TFA, trifluoroacetic acid; TTX, tetrodotoxin; VGSC, voltage- gated sodium channel. 918 FEBS Journal 277 (2010) 918–930 ª 2010 The Authors Journal compilation ª 2010 FEBS muscle, and in cardiac muscle [1]. In insects, only three orthologous sodium channel isoforms have been suc- cessfully expressed in Xenopus oocytes so far, namely DmNa v 1, BgNa v 1, and MdNa v 1. DmNa v 1 (isolated from Drosophila melanogaster) was isolated first. This channel is encoded by the para gene, and is coexpressed with TipE for fully functional expression [2,3]. VGSCs are associated with a broad range of chan- nelopathies and channel-related diseases. Channelopa- thies result from mutations in channel genes and cause a channel dysfunction. They are, for instance, mani- fested as cardiac [4], neuronal [5] and pain disorders [6], as well as various forms of epilepsy and febrile sei- zures [7]. Not only can mutations in genes cause a channel dysfunction, but aberrant expression of VGSC is also a contributory factor in a growing series of channel-related diseases. For example, upregulation of Na v 1.6 was recently shown to play an important role in multiple sclerosis [8]. Furthermore, elevated levels of VGSCs are observed in Alzheimer patients [9] and various types of cancer [10]. To date, little is understood about the components of the venom of solitary wasps. However, solitary wasps such as the spider wasps (Pompilidae) are known to paralyze their prey instead of killing it, as they make a nest for their larvae so that they can feed on living prey. Hence, their venom promises to contain interesting neurotoxins acting on VGSCs. In the search for such new neurotoxins, Konno et al. [11,12] identi- fied two new compounds, named a-pompilidotoxin (a-PMTX) and b-pompilidotoxin (b-PMTX), from the spider wasps Anoplius samariensis and Batozonel- lus maculifrons, respectively. Although they originate from two different species of the Pompilidae family, they differ in only one amino acid at position 12. During further characterization, this single amino acid difference (Arg12 for b-PMTX versus Lys12 for a-PMTX) appeared to be responsible for a difference in potency, as b-PMTX appeared to be five times as potent as a-PMTX in lobster neuromuscular junctions [12,13]. In structure–activity studies, the location of three other basic residues at positions 1, 3 and 12 was found to be crucial for toxin action [13], but none of the synthesized analogs was more potent than b-PMTX. Simultaneously, Sahara et al. [14] reported that a-PMTX slows down the inactivation in tetrodo- toxin (TTX)-sensitive VGSCs of rat trigeminal neu- rons. In chimeric channel studies with Na v 1.2 and Na v 1.5 from rats, b-PMTX selectively targeted the Na v 1.2 channel, and its binding site was localized in the S3–S4 extracellular loop of domain 4 at Glu1616 [15]. b-PMTX effects were compared with those of Anemonia sulcata toxin (ATX-II), which is also known to slow inactivation. Differential actions were observed, although their tertiary structures showed similarities [16,17]. More recently, Grieco and Raman [18] described the ability of b-PMTX to increase or induce resurgent currents in Purkinje cells expressing or lacking Na v 1.6, respectively. By using a series of six pompilidotoxin (PMTX) ana- logs (Table 1) [16], we studied the effects of PMTXs on sodium current inactivation in seven mammalian VGSC isoforms and one insect VGSC (DmNa v 1) (Table 2). By evaluating three components of which the recorded cur- rent is composed (fast, slow and steady-state compo- nents), we could distinguish three distinct groups of PMTX effects. Additionally, this work is unique in including extensive characterization of b-PMTX and analogs over the tested VGSC isoforms. Results Screening of peptides for robustness of the slowing of the inactivation process both in mammalian isoforms and in one insect isoform As only a limited amount of each compound was available, we initially decided to use a simple and fast method to detect the robustness of the effects. To investigate whether a peptide is able to bind to the ion Table 1. Primary sequence alignment of b-PMTX and its analogs, tested in this study. Normal type and bold type indicate identical and homologous amino acids, respectively. b-PMTX RIKIGLFDQLSRL-NH 2 3Rb-PMTX RIRIGLFDQLSRL-NH 2 1K3Rb-PMTX KIRIGLFDQLSRL-NH 2 1Kb-PMTX KIKIGLFDQLSRL-NH 2 a-PMTX RIKIGLFDQLSKL-NH 2 1Ka-PMTX KIKIGLFDQLSKL-NH 2 Table 2. Position and amino acid sequences of hNav1.x and para ⁄ - TipE (DmNav1) channels. Segments 3 and 4 of domain IV and its corresponding extracellular linker are shown. Identical amino acids are in normal type, homologous amino acids are in bold type, and different amino acids are in italic. Position IV S3 Linker IV S4 Navl.1 1616 IVGMFLAELIEKYFVSPTLFRVI Navl.2 1606 IVGMFLAELIEKYFVSPTLFRVI Navl.3 1601 IVGMFLAEMIEKYFVSPTLFRVI Navl.4 1628 IVGLALSDLIQKYFVSPTLFRVI Navl.5 1603 IVGTVLSDIIQKYFFSPTLFRVI Navl.6 1597 IVGMFLADIIEKYFVSPTLFRVI Navl.7 1579 IVGMFLADLIETYFVSPTLFRVI para 1694 ILGLVLSDIIEKYFVSPTLLRVV E. Schiavon et al. Pompilidotoxins and voltage-gated Na v isoforms FEBS Journal 277 (2010) 918–930 ª 2010 The Authors Journal compilation ª 2010 FEBS 919 channel protein and cause slowing of the physiological fast inactivation process, we routinely measured the sodium current traces at 20 ms (30 s for the insect isoform) after eliciting the current, i.e. about 20 times after the end of the fast inactivation (test voltage of 0 mV). Determining the ratio between this current and the peak current (I 20 ⁄ I peak ) is the standard procedure for quantification of these effects independently of the mode of action, either for a slow or a steady-state component increase [19]. In Fig. 1, a 3D plot illustrates the average I 20 ⁄ I peak ratio for both the eight isoforms and the PMTX ana- logs. DmNa v 1 ⁄ TipE was only tested with b-PMTX. Almost no effects were seen for Na v 1.4 and Na v 1.5, and an increase in potency from Na v 1.7 up to Na v 1.1 through Na v 1.3 and Na v 1.2 was observed. Of all synthe- sized PMTX analogs, b-PMTX generally showed the largest effects on the tested isoforms. Consequently, we decided to use b-PMTX for subsequent experiments, as it was also previously demonstrated to be the most potent PMTX toxin [13], and a sufficient amount of this compound was available. Although the I 20 ⁄ I peak proce- dure has been frequently used in many laboratories, it cannot easily show whether the effects produced by the toxin peptide are more related to a change in the inacti- vation time constant of the toxin-bound channels, or to an incomplete process of the inactivation machinery caused by the toxin itself [20,21]. An in-depth investigation of the three components of the sodium current was performed both on seven mammalian and one insect VGSC with b-PMTX. As illustrated in Fig. 2, under control conditions, the fast component, A f , was generally large. In contrast, the slow (A s ) and steady-state (A ss ) components were very low or negligible (see Experimental procedures). Dur- ing toxin action, the fast and slow time constants (s f and s s ) did not change significantly, but, the ampli- tudes of the three components did. Moreover, these components interchanged their relative amplitudes, so the ratio between each component and the total ampli- tude (T) is illustrated in Figs 3–6 in a dose-dependent manner. Collectively, with these data, we studied in detail how the PMTXs affected the sodium currents in each of the tested VGSC isoforms. Na v 1.1, Na v 1.2, Na v 1.3 and Na v 1.7 dose–response relationships for fast, slow and steady-state components and their voltage-dependent activation The results are divided into Figs 3 and 4 for Na v 1.1, Na v 1.2 and Na v 1.7, and for Na v 1.3, respectively, Fig. 1. Effects of different analogs of PMTX on the eight VGSC iso- forms. The ratio I 20 ms ⁄ I peak is plotted, on the z-axis, for the six dif- ferent analogs of PMTX and the eight VGSC isoforms. A maximal concentration of 46 l M was used (n = 4 for each data point). See [17] and Experimental procedures for details of the sequences of the other peptides. Fig. 2. Dissection analysis of the sodium currents to separate the fast, slow and steady-state components. A representative response obtained with b-PMTX and its quantitative analysis are shown. The macroscopic current traces in controls and in the presence of b-PMTX are shown as a line plus open squares and open circles, respectively. The best-fitted decomposed components of the mac- roscopic current (fast, slow, and steady-state) for the b-PMTX trace are shown as lines. Inset: the same, but zoomed-in to reveal only the early currents and the beginning of the fitting. The fast, slow and steady-state components are indicated as lines. Pompilidotoxins and voltage-gated Na v isoforms E. Schiavon et al. 920 FEBS Journal 277 (2010) 918–930 ª 2010 The Authors Journal compilation ª 2010 FEBS according to their similarity with regard to the observed effects. Figures 3 and 4 show, for each repre- sentative isoform: (a) the typical time course of the currents; (b) the fast and slow fractional component change versus toxin concentration; (c) the A ss ⁄ T dose– response relationship; and (d) the voltage-dependent conductances of each component. In Fig. 3, these data are presented for Na v 1.2 together with the exemplary recordings from cells expressing Na v 1.1 and Na v 1.7, as they exhibit similar effects. For these three isoforms, s f , the time constant of the fast component, lay in the range from 0.4 ± 0.02 to 0.7 ± 0.03 ms (n = 4 for each type). In contrast, s s , the time constant of the slow component, which was about 4.3 ± 0.3 ms in controls (but its amplitude was well below 0.5%), increased up to 8.5 ± 0.02 ms (n = 4 for each type) and 15 ± 0.02 ms (n = 4 for each type) when tested at low or high concentrations, i.e. from about 10 up to 140 lm. These values were not consistently increasing in the membrane voltage range between )20 and +30 mV (not shown). In Fig. 3B, the dose–response relationship for Na v 1.2 is presented, describing the exchange between the amount of channels that are successively bound by the increased number of peptide molecules. The A f and A s data crossover at about 50% suggests a concentration value of about 30 lm. In Fig. 3C, the concentration- dependent increase of the A ss component is described, and it was found that that this effect was relatively small, amounting to a maximum of < 20% of the total current. To investigate whether the emergence of the slow component was voltage-dependent, conductances were determined for a range of membrane voltages, at a fixed concentration of 45 lm. Theoretically, we would expect the A f ⁄ T plot obtained in the presence of toxin to follow the behavior seen in controls, as this ampli- tude originates from the unbound channels. The nor- malized data do reveal such a type of curve for the A f ⁄ T plot. The A s ⁄ T component, in contrast, remains constant after having reached its maximum at )20 mV. The results for Na v 1.3 are presented in Fig. 4. The A s time constants were not significantly different from those mentioned above for the other isoforms. It can be Fig. 3. Effects on the Na v 1.1, Na v 1.2 and Na v 1.7 isoforms. (A) Data from three exemplary cells expressing Na v 1.1, Na v 1.2 and Na v 1.7 chan- nels. Superimposed traces elicited at )5 mV from a holding potential of )110 mV in controls and during perfusion of b-PMTX at 46 and 140 l M. (B) Fractional change of the fast component (triangles) and slow component (circles) versus [b-PMTX] for Na v 1.2 currents (n = 4). (C) Dose–response curve of the fractional steady-state component (n = 4) for Na v 1.2 currents. The line is the best fit with a Hill curve with EC 50 , power and maximum value equal to 21 ± 2.4 lM, 1.1 ± 0.13, and 0.16, respectively. (D) Conductance–voltage plots obtained in con- trols (solid squares) and in the presence of toxin (solid circles and triangles) with 46 l M b-PMTX, for Na v 1.2 currents. The fractional conduc- tances of the slow components (A s ⁄ T, closed circles) and the fast components (A f ⁄ T, solid inverted triangles) are plotted against the applied membrane voltage. The same data are also shown resized to 1 (open circles, open triangles). E. Schiavon et al. Pompilidotoxins and voltage-gated Na v isoforms FEBS Journal 277 (2010) 918–930 ª 2010 The Authors Journal compilation ª 2010 FEBS 921 noticed that, in Fig. 4B, the crossover between the amplitudes of A s and A f occurs at about 100 instead of 20 lm, as shown in Fig. 3 for Na v 1.2. Moreover, the predicted EC 50 of the A ss component is also around 100 lm, instead of the value of 21 lm seen for Na v 1.2 in Fig. 3. The voltage-dependent conductance of the slow component started to decrease in the region from )10 to +30 mV, after displaying a peak at )20 mV. This effect was somewhat unexpected, as it could suggest that the binding of the toxin could be voltage-dependent by itself. All together, these data suggest that the toxin action shows complex behavior and depends on the par- ticular amino acid sequence of each isoform. Effects on the sodium currents of Na v 1.6 are characterized by a robust voltage-dependent steady-state component and the absence of the slow component The effects of the toxin on Na v 1.6 currents are remark- ably different from those described above for other isoforms. As illustrated in Fig. 5A, the slow compo- nent was not detected at all and the fast component completely disappeared at 45 and 140 lm. In contrast, a steady-state component appeared early after activa- tion, with a maximal amplitude of about one-third of the fast peak observed in controls. This is illustrated in Fig. 5B, which confirms that the population of toxin- bound channels is in equilibrium with the population of unbound channels – here represented by the steady- state component instead of the slow component, as compared with Figs 3 and 4. To evaluate the A ss com- ponent, its dose–response relationship was defined, as outlined in Fig. 5C. At a fixed concentration of 45 lm, the A ss component also appeared to demonstrate volt- age-dependent behavior, as well as the A f (Fig. 5D). This surprising result underlines the peculiarities of this isoform, which are already known: (a) the possibil- ity of generating a resurgent current in control condi- tions in cells expressing this isoform (cerebellar Purkinje cells, [22]); and (b) the ability of Centruro- ides noxius toxin (Cn2), a scorpion b-toxin, to induce a resurgent current, both in cells expressing Na v 1.6 chan- nels and in cerebellar neurons [23]. Moreover, Cn2 has been reported to cause a strong hyperpolarizing shift in the voltage dependency of the activation. This coincidence prompted us to design experiments to examine whether b-PMTX and Cn2 could eventu- ally compete for the same site on the channel. Specifi- cally, we investigated whether the Na v 1.6 channels were still targeted by Cn2 if b-PMTX is already bound by the channel, i.e. whether the two toxins act via dif- ferent binding sites. The results of these experiments, presented in Fig. 6, recapitulate the Cn2 effects and the b-PMTX effects as if they were produced by two independent pathways. In these experiments, peptide concentrations in the range of their EC 50 values were used. In Fig. 6A,B, the superimposed traces of the cur- rents are shown at )30 and +20 mV, in the absence (Fig. 6A) and presence (Fig. 6B) of b-PMTX. Only at +20 mV was there clear evidence that b-PMTX induced the A ss component, as would be expected from the data shown in Fig. 5. When Cn2 was added after addition of b-PMTX (Fig. 6C), a large peak current did arise at )30 mV, owing to a hyperpolarizing shift of the activation, which was produced by the scorpion b-toxin (a blocking effect at +20 mV also occurred; Fig. 4. Effects on Na v 1.3. (A) Superimposed traces elicited at )5 mV from a holding potential of )110 mV in controls and during perfusion of b-PMTX at 46 and 140 l M. (B) Fractional change of the fast and steady-state components versus [b-PMTX] (n = 4). (C) Dose–response curve of the fractional steady-state component (n = 4). The line is the best fit with a Hill curve with EC 50 , Hill coef- ficient and maximum value equal to 99 ± 1 l M, 1.11 ± 0.06, and 0.21, respectively. (D) Conductance–voltage plots obtained in con- trols (closed squares) and in the presence of toxin with 46 l M b-PMTX (solid circles and triangles). The fractional conductances of the fast components (A f ⁄ T, solid inverted triangles) and the slow components (A s ⁄ T, solid circles) are plotted against the applied membrane voltage. The same data are also shown resized to 1 (open symbols). Pompilidotoxins and voltage-gated Na v isoforms E. Schiavon et al. 922 FEBS Journal 277 (2010) 918–930 ª 2010 The Authors Journal compilation ª 2010 FEBS Fig. 5. Effects on Na v 1.6. (A) Superimposed traces elicited at )5 mV from a holding potential of )110 mV in controls and during perfusion of b-PMTX at 46 and 140 l M. (B) Fractional change of the fast and steady-state components versus [b-PMTX] (n = 4). (C) Dose–response curve of the fractional slow component (n = 4) obtained at three different membrane potentials of )20, )5 and +25 mV. The three lines are the best fit with Hill curve parameters (EC 50 , Hill coefficient), as follows: 43.2 ± 1.7 l M, 1.2 ± 0.05 ()20 mV); 30 ± 1 l M, 1.5 ± 0.06 ()5 mV); 23.6 ± 1.7 l M, 1.6 ± 0.13 (+25 mV). (D) Conductance–voltage plots obtained in controls (solid squares) and in the presence of toxin with 46 l M b -PMTX (triangles). The fractional conductances of the fast compo- nents (A f ⁄ T, solid inverted triangles) and the steady-state components (A ss ⁄ T, solid triangles) are plotted against the applied membrane voltage. The same data are also shown resized to 1 (open triangles). Fig. 6. The competition between b-PMTX and Cn2. Upper panel. Superimposed traces of currents elicited at )90, )30 and +20 mV from a holding potential of )100 mV. The four panels show the three traces, for the same cell, from left to right, in controls and in the presence of b-PMTX [46 l M (note the steady-state component)], after subsequent addition of 36 nM Cn2 and 1 min after starting the washout of b-PMTX. The washout of Cn2 has a time course of more than 20 min. Lower left panel. The normalized conductance–voltage plots obtained in controls (open squares) and in the presence of Cn2 (open circles). Note the left-shift induced increase in conductance in the region below )30 mV (n = 3). Lower right panel. The ratio of the steady-state conductance with respect to peak control conductance (measured at +10 mV) is plotted against the applied membrane potential in the four conditions as shown in the upper panels. Note the large increase in the steady-state component produced also at membrane potentials below the threshold of activation of )30 mV (n = 3). E. Schiavon et al. Pompilidotoxins and voltage-gated Na v isoforms FEBS Journal 277 (2010) 918–930 ª 2010 The Authors Journal compilation ª 2010 FEBS 923 see [23]). In both traces of Fig. 6C, a steady-state com- ponent was also present. Consecutively, a rapid wash- out of b-PMTX was performed, with Cn2 still present. As a result, the steady-state component immediately disappeared (Fig. 6D). Nevertheless, we did not see additional effects when both toxins were present. To validate the recordings, an analysis of the voltage dependence of normalized peak conductances in con- trols and in the presence of Cn2 is shown in Fig. 6E, and the conductance ratios (G ss ⁄ G control ) derived from the previous four experiments (Fig. 6F) clearly con- firmed that the two toxins should work by an indepen- dent mechanism. Effects on the sodium currents of the insect isoform show a higher affinity with respect to those observed in Na v 1.6 In order to test the effects of the peptide not only on mammalian isoforms, but also on insect sodium chan- nels, we were forced to use another expression system, namely the oocyte expression system. The effects of the toxin on the DmNa v 1 ⁄ TipE isoform are shown in Fig. 7. These data qualitatively recapitulated the pep- tide action shown in Fig. 5 for Na v 1.6. In Fig. 7A, the superimposed traces suggest com- plete removal of the fast component at the expense of the steady-state one. The data shown in Fig. 7B,C con- firm quantitatively that the action follows a dose– response relationship, with an EC 50 about 10-fold lower than that presented in Fig. 5 for Na v 1.6. The interesting difference was the fact that, contrary to any expectation, the final amplitude of the currents pro- duced by the toxin could be greater than the original values present in controls. Na v 1.4 and Na v 1.5 effects and voltage-dependent activation ⁄ inactivation at maximal concentration Although the data presented in Fig. 2 for Na v 1.4 and Na v 1.5 excluded any slow and steady-state components in the action of b-PMTX, we nevertheless wanted to Fig. 7. Effects on the insect isoform DmNa v 1 ⁄ TipE (expression in X. laevis oocytes). (A) Superimposed whole cell sodium current traces in the presence of 0.5, 3 and 50 l M b-PMTX and in the control situation in oocytes expressing the DmNa v 1 ⁄ TipE insect channel. An arrow indi- cates the zero-current level. Currents were elicited at 0 mV from a holding potential of )90 mV. (B) Fractional change of the fast and steady- state components of the current versus the applied b-PMTX concentrations (n = 2–7). (C) Dose–response curve of the fractional steady-state components obtained at membrane potentials of )20, 0, and +20 mV (n = 2–7). Curves were fitted with Hill curve parameters (EC 50 , Hill coefficient) as follows: 18.5 ± 15.1 l M, 1.9 ± 1.6 (+20 mV); 4.6 ± 0.5 lM, 2.7 ± 0.5 (0 mV); 5.8 ± 0.6 lM, 2.03 ± 0.25 ()20 mV). (D) Con- ductance–voltage plots obtained in controls (solid squares) and in the presence of toxin with 10 l M b-PMTX (triangles). The fractional con- ductances of the fast (A f ⁄ T, solid inverted triangles) and steady-state (A ss ⁄ T, solid triangles) components of the current are plotted against the applied voltage. The same data are also shown resized to 1 (open triangles). Pompilidotoxins and voltage-gated Na v isoforms E. Schiavon et al. 924 FEBS Journal 277 (2010) 918–930 ª 2010 The Authors Journal compilation ª 2010 FEBS investigate whether other effects could be observed by examining in detail the voltage-dependent activation and inactivation. These results are given Fig. S1. Exem- plary traces for both types of channels did not reveal any effect on the slow or steady-state component, although some very small inhibitory effects were seen for the cardiac (Na v 1.5) isoform. In Fig. S1B, a small hyperpolarizing shift of the steady-state fast inactiva- tion of Na v 1.5 was observed, and could probably be responsible for the small inhibition seen in Fig. S1A. Summary of the effects on the slow and steady-state components in Na v 1.1, Na v 1.2, Na v 1.3, Na v 1.6, Na v 1.7 and DmNa v 1 ⁄ TipE Collectively, the previous data suggest that b-PMTX is able to recognize the insect as well as the mammalian Na v 1.1, Na v 1.2, Na v 1.3, Na v 1.6 and Na v 1.7 isoforms, but that the exerted actions are consistently different. To summarize, Fig. 8 provides an overview of the dif- ferent effects on the tested VGSC isoforms, for the increase in the slow component as well as in the steady-state component. Excluding Na v 1.4 and Na v 1.5, for which we did not obtain significant results, and based on the obtained results, two divergent groups of effects can be distinguished. The first group concerns the insect and Na v 1.6 isoforms. Both showed a large increase in the steady-state component without any increase in the slow component. The second group concerns Na v 1.1, Na v 1.2, Na v 1.3, and Na v 1.7, which showed a large increase in the slow component, with only a small steady-state component increase. These data could suggest that the two isoform groups inter- act in different ways with the toxin by offering differ- ent binding sites. Discussion It is very common for a number of peptides purified from various animal venoms to potentially alter the inactivation process that occurs in VGSCs during the physiological upstroke of the action potential. Undeni- ably, the inactivation of VGSCs is their most vulnera- ble kinetic feature, as it is influenced, being mostly slowed or abolished, by many kinds of chemical sub- stances such as drugs and toxins, and by mutations. These mutations often encompass only a single amino acid in the channel’s sequence. The sequences of the VGSCs, as well as their transmembrane topology, have been understood for a while. However, relating struc- tural channel components to a profound function still remains a difficult task, and has not been accom- plished to date. Nevertheless, site-directed mutagenesis is increasingly being applied to resolve this issue. The molecular exploration of hereditary diseases has also contributed significantly to the identification of relevant regions. According to the well-known definition derived through the extensive work conducted by Ce ` stele and Catterall [24], ‘site 3¢ is considered to be responsible for a large part of the toxin-induced effects on the inactivation of VGSCs known to date. This site is located in domain 4 in the extracellular linker between segments S3 and S4. It is the target site for sea anem- one, scorpion, spider and wasp toxins. These toxins are peptides with different structures and lengths. In the current study, eight different VGSC isoforms were tested for their selective response to PMTXs, which are known to cause slowing of the inactivation of VGSCs. This study resembles previous work con- ducted by Oliveira et al. [25], in which it was demon- strated that the 47–48 amino acid peptides from sea anemones (type 1 toxins) very selectively targeted some isoforms more than others. Additionally, one natural single amino acid substitution [K36A, between ATX-II and Anthopleura fuscoviridis toxin (AFT-II)] is suffi- cient to cause large diversities in potency and selectiv- ity [25]. The results of the current study demonstrate that the PMTXs were also able to discriminate between the different VGSCs, classifying the tested VGSCs into three groups according to their PMTX- induced kinetics. Fig. 8. Three-dimensional plot of the normalized increase of the slow or steady-state components seen in the different isoforms. Data were obtained at a concentration of 46 l M (n = 5). E. Schiavon et al. Pompilidotoxins and voltage-gated Na v isoforms FEBS Journal 277 (2010) 918–930 ª 2010 The Authors Journal compilation ª 2010 FEBS 925 For Na v 1.1, Na v 1.2, Na v 1.3, and Na v 1.7, the addi- tion of b-PMTX mainly resulted in a large increase in the slow component of current inactivation and, to a minor extent, a small increase in a steady-state compo- nent. These data quantitatively demonstrate an almost perfect dose–response relationship between two popu- lations of channels, those bound by the peptide and those unbound. This produces a net decrease in the fast component and a corresponding increase in the slow component of the inactivation. The maximal effects seen at the level of the steady-state component, represented by the ratio A ss ⁄ T, did not reach a value of 0.2, and the voltage-dependent inactivation curves also did not show any significant effect (data not shown). A completely different biophysical effect was revealed by experiments performed on Na v 1.6 and the DmNa v 1 ⁄ TipE. No slow component increase was noticed in these experiments. In contrast, the steady- state component largely increased, reaching a value of 1(A ss ⁄ T). Moreover, its growth was comparable to the decrease in the fast component. This suggests that, in this isoform, the action of the peptide was accompa- nied by a mechanism somewhat different from that observed in the other isoforms. In addition, previous studies demonstrated that Na v 1.6 and Na v 1.3 are also prone to produce a large steady-state component under the action of AFT-II and Bunodosoma caissarum toxin (Bc-III) [25], although in this case these toxins also produced a slow inactivation component. The data for DmNa v 1 shown in Fig. 7A are very striking, because they bear a strong resemblance to data obtained in earlier experiments. In those experi- ments, enzymes capable of irreversibly removing the short peptide responsible for the inactivation mecha- nism were injected intracellularly [26,27]. In more recent, analogous, experiments, a similar effect was shown with the use of a scorpion a-toxin [28] and the sea anemone toxins BgII and BgIII (isolated from Bunodosoma granulifera) [29]. Similarly to what is seen in Fig. 7A, the currents of the DmNa v 1 ⁄ TipE channels also did not show any A s component upon BgII ⁄ BgIII application. Na v 1.6 has distinctive properties, because it is excep- tionally sensitive to the actions of other neurotoxins, such as scorpion b -toxins [23]. This led us to investi- gate competition between b-PMTX and Cn2. As shown in Fig. 6, it became clear that the two peptides do not significantly interact, as the increase in the steady-state component was also present in those channels bound by Cn2, which opened at voltages lower than the threshold values typical of the normal conditions. In another study, a chimera channel was constructed by inserting domain 2 of DmNa v 1 channels in the rat brain rBIIA (rNa v 1.2) sodium channel [30]. It is important to mention that this chimera exhibited insect channel properties in the activation of the rBIIA chan- nel, and also determined the selectivity of the excit- atory AahIT scorpion toxin from Androctonus australis hector over insect channels. The same study demon- strated the unexpected involvement of domain 2 in the inactivation process, even though scorpion a-toxins had not changed their binding ability because of the presence of the insect domain 2 in the rat channel. Considering our present results on the coapplication of Cn2 and b-PMTX, this is one more indication that b-PMTX would not bind to domain 2 (the target of Cn2). As expected from previous results reported by Kinoshita et al. [15], Na v 1.4 and Na v 1.5 did not show any effect on the slow component or the steady-state component of inactivation. In general, the dose– response EC 50 values of the slow or steady-state com- ponent illustrated in Figs 3, 4, 5 and 7 correlated well with the data showing the crossover of the populations of toxin-bound and toxin-unbound channels. Indeed, Na v 1.2 and Na v 1.6 were the isoforms with the best affinity in the range from 20 to 30 lm;Na v 1.7 had an intermediate value of about 55–60 lm (data not shown), and Na v 1.3 had the lowest affinity, at about 100 lm. The described specificity of the b-PMTX-induced effects on the tested isoforms can be explained by the model previously obtained from a study by Kinoshita et al. [15]. They suggested that specific site-directed mutagenesis in the rBII (rNa v 1.2) isoform was able to eliminate an increase in the so-called modification ratio (in our case: I 20 ⁄ I peak ). This amino acid modification involved a Glu fi Gln mutation at position 1616, which is exactly reproduced in natural sequences of the Na v 1.4 and Na v 1.5 isoforms, and is a fingerprint of these channels (presented in Table 2; human isoforms and invertebrate channels are aligned). Consequently, our results confirmed that Na v 1.4 and Na v 1.5 did not show any effect in the slow or steady-state component, as shown in Fig. 2 and Fig. S1. In contrast, the differ- ence in amino acid sequence of the Na v 1.6 and DmNa v 1 isoforms with respect to the Na v 1.1, Na v 1.2, Na v 1.3 and Na v 1.7 group can be restricted to the DI sequence versus EL, EM or DL for the Na v 1.1, Na v 1.2, Na v 1.3 and Na v 1.7 group. Moreover, DmNa v 1 insect channels present the sequence DIIE, similar to the sequence starting at Asp1604 in Na v 1.6. Considering the preferential affinity of b-PMTX, the individual or combined role of these residues still remains unclear. Pompilidotoxins and voltage-gated Na v isoforms E. Schiavon et al. 926 FEBS Journal 277 (2010) 918–930 ª 2010 The Authors Journal compilation ª 2010 FEBS Within the last 15 years, much research has been conducted on ‘site 3 mutagenesis’, as well as chimera channels, in order to analyze the alleged binding sites of scorpion a -toxin and b-toxin and type 1 sea anem- one toxins in VGSCs [30–32]. Researchers discovered that the sea anemone toxin ATX-II and scorpion a-toxins bind to a common extracellular site. In addi- tion, they found that the negatively charged Glu1613 in Na v 1.2 was crucial for toxin binding [31]. More recently, Heinemann and co-workers [32] contributed greatly to the field by scrutinizing the role of the E ⁄ DE, E ⁄ Q and KYFV equivalent in positions 1623, 1626 and 1627, 1628, 1629 and 1630 of Na v 1.1 (Table 2), respectively. The assayed toxins consisted of a quantity of scorpion a-toxins (Lqh-2, Lqh-3, and LqhaIT) from Leiurus quinquestriatus hebraeus as well as d-SVIE conotoxin from Conus striatus. Each of these peptides binds to site 3 of VGSCs and impairs the inactivation process in a similar way as b-PMTX. Taking into consideration the current results, com- bined with previously discussed site-directed mutagene- sis findings [30–32] and studies on chimera channels [15,30], it has become clear that certain ‘microhetero- geneous residues’ in S3–S4 linker segments have a very specific affect on toxin interactions. However, even with the current knowledge of the S3 and S4 segments of domain 4, it still remains unclear whether other topological components of the VGSCs may be involved in the binding of toxins that impair the inacti- vation process and increase steady-state components. Further study on chimeras and site-directed mutagene- sis will undoubtedly prove beneficial in addressing this issue. Experimental procedures Peptide synthesis Peptides were synthesized by a stepwise solid-phase method using Fmoc chemistry with TGS-RAM resin (Rapp Poly- mere GmbH, Tu ¨ bingen, Germany) on a Shimadzu PSSM-8 peptide synthesizer (Shimadzu Corp., Kyoto, Japan). All Fmoc-l-amino acids were purchased from Nova Biochem. The side chain protective groups were t-butyloxycarbonyl for Lys, 2,2,5,7,8-pentamethylchroman-6-sulfonyl for Arg, and t-butyl for Thr. Cleavage of the peptide from the resin was achieved by treatment with a mixture of trifluoroacetic acid (TFA) ⁄ phenol ⁄ thioanisole ⁄ 1,2-ethanedithiol ⁄ ethyl- methyldisulfide ⁄ H 2 O (80 : 5 : 5 : 3 : 2 : 5, by volume), using 10 mLÆg )1 resin at room temperature for 8 h. After removal of the resin by filtration and washing twice with TFA, the combined filtrate was added dropwise to diethyl ether at 0 °C and then centrifuged at 3000 r.p.m. for 10 min. The crude synthetic peptide obtained was purified by semipreparative RP-HPLC using YMC-PAK ODS (20 · 150 mm; Yamamura Kagaku, Kyoto, Japan) with isocratic elution of 28–30% CH 3 CN ⁄ H 2 O ⁄ 0.1% TFA at a flow rate of 7 mLÆmin )1 . The homogeneity and the sequence were confirmed by HPLC and MALDI-TOF MS [17]. Besides b-PMTX, the synthetic peptides used were: 3Rb (same as b-PMTX except for a Lys fi Arg change at position 3), 1Kb (same as b-PMTX except for an Arg fi Lys change at position 1), a-PMTX, 1K3Rb (same as b-PMTX except for Arg fi Lys and Lys fi Arg changes at positions 1 and 3, respectively), and 1Ka (same as a-PMTX except for an Arg fi Lys change at position 1). Human Na v 1.x isoforms Cell culture HEK293 cell lines stably expressing human Na v 1.1, Na v 1.2, Na v 1.3, Na v 1.5 and Na v 1.6 (generously donated by Glaxo- SmithKline, Medicines Research Centre, Stevenage, UK) were cultured in modified DMEM supplemented with 10% fetal bovine serum as previously described [25]. Na v 1.4- expressing cells were obtained by stable transfection of a plasmid containing the hNa v 1.4 construct (a kind gift from D. Conti-Camerino, University of Bari, Italy). Na v 1.7- expressing cells were obtained by transient transfection of a plasmid containing the hNa v 1.7 construct (a kind gift from F. Hofmann through A. Wada, University of Miyazaki, Japan). Approximately 2 · 10 4 cells were transfected with 2 lg of the hNa v 1.7 vector along with 0.2 lg of green fluorescent protein in a pEGFP-C1 vector (Clontech, Euro- clone, Milan, Italy), using the Lipofectamine reagent kit (Invitrogen, Milan, Italy), following the instructions of the manufacturer. Currents were recorded 24–72 h following transfection. Solutions and drugs The standard extracellular solution contained: 70 mm NaCl, 67 mm N-methyl-d-glucamine, 1 mm CaCl 2 , 1.5 mm MgCl 2 , 5mm Hepes, and 10 mmd-glucose (pH 7.40). The standard pipette solution contained: 105 mm CsF, 27 mm CsCl, 5mm NaCl, 2 mm MgCl 2 ,10mm EGTA, and 10 mm Hepes (pH 7.30). About 6–8% of the cells expressing the Na v 1.6 channel clone had a persistent sodium current, as reported by Burbidge et al. [33]. We systematically tested these cells, and discarded those showing incomplete inacti- vation (a residual current after 250 ms of < 0.1% of the peak sodium current). Known quantities of the toxins were dissolved in the extracellular solution just before the start of the experiments. TTX (Sigma, Milan, Italy) was used at 300 nm on the Na v 1.1, Na v 1.2, Na v 1.3, Na v 1.4, Na v 1.6 and Na v 1.7 currents, and the resulting traces were subtracted from the control traces to obtain the TTX-sensitive currents; the Na v 1.5 clone (which has a much higher TTX E. Schiavon et al. Pompilidotoxins and voltage-gated Na v isoforms FEBS Journal 277 (2010) 918–930 ª 2010 The Authors Journal compilation ª 2010 FEBS 927 [...]... vzw) E Schiavon was a PhD student of physiology at the Department of Biotechnologies and Biosciences of the University of Milano-Bicocca References 1 Catterall WA, Goldin AL & Waxman SG (2005) International Union of Pharmacology XLVII Nomenclature and structure–function relationships of voltagegated sodium channels Pharmacol Rev 57, 397–409 2 Dong K (2007) Insect sodium channels and insecticide resistance... interaction of delta-conotoxins with voltage-gated sodium channels FEBS Lett 579, 3881–3884 Burbidge SA, Dale TJ, Powell AJ, Whitaker WR, Xie XM, Romanos MA & Clare JJ (2002) Molecular cloning, distribution and functional analysis of the NA(V)1.6 voltage-gated sodium channel from human brain Brain Res Mol Brain Res 103, 80–90 Liman ER, Tytgat J & Hess P (1992) Subunit stoichiometry of a mammalian K+ channel. .. Gordon D & Lotan I (2002) Domain 2 of Drosophila para voltage-gated sodium channel confers insect properties to a rat brain channel J Neurosci 22, 4364–4371 Rogers JC, Qu Y, Tanada TN, Scheuer T & Catterall WA (1996) Molecular determinants of high affinity binding of alpha-scorpion toxin and sea anemone toxin in the S3–S4 extracellular loop in domain IV of the Na+ channel alpha subunit J Biol Chem 271,... Pompilidotoxins and voltage-gated Nav isoforms 7 Mantegazza M, Gambardella A, Rusconi R, Schiavon E, Annesi F, Cassulini RR, Labate A, Carrideo S, Chifari R, Canevini MP et al (2005) Identification of an Nav1.1 sodium channel (SCN1A) loss -of- function mutation associated with familial simple febrile seizures Proc Natl Acad Sci USA 102, 18177–18182 8 Waxman SG (2008) Mechanisms of disease: sodium channels and... in Nav1.6 channel Significance in mice Purkinje neurons J Biol Chem 281, 20326–20337 ` Cestele S & Catterall WA (2000) Molecular mechanisms of neurotoxin action on voltage-gated sodium channels Biochimie 82, 883–892 Oliveira JS, Redaelli E, Zaharenko AJ, Cassulini RR, Konno K, Pimenta DC, Freitas JC, Clare JJ & Wanke E (2004) Binding specificity of sea anemone toxins to Nav 1.1–1.6 sodium channels: unexpected... Ca(V) channels Channels (Austin) 2, 100–117 4 Ackerman MJ (2004) Cardiac channelopathies: it’s in the genes Nat Med 10, 463–464 5 Kullmann DM (2002) The neuronal channelopathies Brain 125, 1177–1195 6 Dib-Hajj SD, Binshtok AM, Cummins TR, Jarvis MF, Samad T & Zimmermann K (2009) Voltage-gated sodium channels in pain states: role in pathophysiology and targets for treatment Brain Res Rev 60, 65–83 Pompilidotoxins. .. Heinemann SH (2004) Combinatorial interaction of scorpion toxins FEBS Journal 277 (2010) 918–930 ª 2010 The Authors Journal compilation ª 2010 FEBS 929 Pompilidotoxins and voltage-gated Nav isoforms 20 21 22 23 24 25 26 27 28 29 930 E Schiavon et al Lqh-2, Lqh-3, and LqhalphaIT with sodium channel receptor sites-3 Mol Pharmacol 65, 685–691 Ulbricht W (2005) Sodium channel inactivation: molecular determinants... from a holding potential of )90 mV The fractional components of the currents were plotted against the logarithm of the applied concentrations To assess the concentration dependency of the b-PMTX-induced effects, dose–response curves of the steady-state component were constructed and fitted with the Hill equation to obtain EC50 values (i.e the toxin concentration that produces 50% of the maximum effect)... class of neurotoxin from wasp venom slows inactivation of sodium current Eur J Neurosci 12, 1961–1970 15 Kinoshita E, Maejima H, Yamaoka K, Konno K, Kawai N, Shimizu E, Yokote S, Nakayama H & Seyama I (2001) Novel wasp toxin discriminates between neuronal and cardiac sodium channels Mol Pharmacol 59, 1457–1463 16 Kawai N & Konno K (2004) Molecular determinants of two neurotoxins that regulate sodium. .. used a ‘free’ or fixed (control) value of sf An example of how we analyzed the observed effects and obtained these theoretical components and their corresponding macroscopic current traces is fully shown in Fig 2 This strongly suggests that the currents recorded in the presence of toxin were always the sum of two types of current: those deriving from toxin-bound channels (modified), FEBS Journal 277 (2010) . Voltage-gated sodium channel isoform-specific effects of pompilidotoxins Emanuele Schiavon 1, *, Marijke. the presence of toxin were always the sum of two types of cur- rent: those deriving from toxin-bound channels (modified), Pompilidotoxins and voltage-gated Na v isoforms