Tityustoxin effect on nerve compound action potentials requires extracellular sodium

Neuroscience Letters 282 (2000) 25±28 www.elsevier.com/locate/neulet

Tityustoxin effect on nerve compound action potentials requires extracellular sodium J.S. Cruz a,*, A.C.S. Matavel a,1, H.M. LeaÄo-Filho a,1, T. Moraes-Santos b, P.S.L. BeiraÄo a a

Departamento de BioquõÂmica e Imunologia, Instituto de CieÃncias BioloÂgicas, Universidade Federal de Minas Gerais, Caixa Postal 486, CEP 30161±970, Belo Horizonte, Minas Gerais, Brazil b Departamento de Alimentos, Faculdade de FarmaÂcia, Universidade Federal de Minas Gerais, Caixa Postal 486, CEP 30161±970, Belo Horizonte, Minas Gerais, Brazil Received 30 December 1999; accepted 15 January 2000

Abstract Previous studies have demonstrated that Li 1 ions can substitute for Na 1 in a variety of functional systems. Using the single sucrose-gap recording technique, we measured the nerve compound action potential to study the effects of tityustoxin (an a-scorpion toxin that selectively inhibits fast Na 1 channel inactivation) upon removal of extracellular Na 1. Our results suggest that tityustoxin requires the presence of extracellular Na 1 to produce its typical pharmacological effect on Na 1 channel inactivation kinetics, but not to bind to its site. q 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Single sucrose-gap; Sciatic nerve; Tityustoxin; a-scorpion toxin; Inactivation; Na 1 channel; Compound action potentials; Li 1 ions

The voltage-dependent sodium channels underlie the electrical excitability of nerve and muscle cells. With the use of molecular biology techniques the identi®cation of different Na 1 channel isoforms was possible. Our understanding of where in the Na 1 channel molecule are localized the major structural motifs that are responsible, at least in part, for normal gating has increased due to a combination of structural and functional studies. The voltage-dependent Na 1 channels are targets for a number of toxins. One group, well studied, is the family of site-3 toxins which comprise toxins isolated from several species of scorpions and sea anemones. These toxins, also called a-toxins, bind to the extracellular surface of voltagegated Na 1 channels and inhibit fast inactivation kinetics that causes prolonged membrane depolarization. Evidences are being obtained that support the notion that the gating behavior of some types of ion channels could be altered by changing the permeant ion [11,16,20]. Lithium (Li 1) has been widely used as a substitute for * Corresponding author. 725 West Lombard Street, Department of Molecular Biology and Biophysics, University of Maryland Biotechnology Institute, Baltimore, MD 21201, USA. Tel.: 11-410-706-2663. E-mail address: [email protected] (J.S. Cruz) 1 These authors contributed equally to this work.

Na 1 ions in different preparations. There is a large body of electrophysiological evidence showing that Li 1 permeates voltage-dependent Na 1 channels, with no apparent modi®cation in the gating properties [8±10,13,21]. At present, little has been reported as to the effect of Na 1 substitution on the pharmacological pro®le of Na 1 channels [1,12]. To address this issue, we have investigated tityustoxin (Tstx), a toxin isolated from the venom of the Brazilian scorpion, Tityus serrulatus, [7] which binds to site-3 of Na 1 channels slowing the kinetics of inactivation, thereby leading to a dramatic lengthening of compound action potential (CAP) duration [18]. Toxins that affect the inactivation kinetics can be precisely and quantitatively assayed using the single sucrose-gap method [5,18]. In this study, we ®nd that Tstx required extracellular Na 1 in order to produce its action on prolongation of CAP duration in the isolated sciatic frog nerve preparation. Tstx was puri®ed using the method described by Sampaio et al. [17], which uses a combination of gel ®ltration on Sephadex G-25 (®ne grade) followed by ion exchange chromatography on CM-Cellulose-52, eluted with a slightly modi®ed ammonium bicarbonate gradient (0.01±0.7 M). The toxic peptide T2IIII thus puri®ed was found to be identical with Tstx, puri®ed and partially characterized by Coutinho-Neto and Diniz [4]. Other authors have also puri-

0304-3940/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 0 0) 00 86 2- 4

26

J.S. Cruz et al. / Neuroscience Letters 282 (2000) 25±28

®ed the Tityus serrulatus venom using different methods and have given other names to toxins that are now recognized as identical to Tstx (TsIV-5 and Ts IV [6]). Tstx was then dissolved in water and stored in aliquots at 2208C until used. Procedures for the experiments were fundamentally the same as described in previous papers [5,18]. Brie¯y, the sciatic nerves from bullfrogs (Rana catesbeiana) were carefully removed and desheathed. One nerve bundle was positioned across the ®ve compartments of the experimental chamber, which contained vaseline at the partitions to electrically isolate them. Compartments 1 and 2, at one end of the nerve bundle, were used to apply supramaximal stimulation, which consisted of 100±140 ms isolated rectangular voltage pulses (4±6 V), delivered by a SD5 stimulator (Grass Instruments, USA), triggered by a computer. These parameters were chosen to selectively stimulate fast-conducting myelinated ®bers (Aa). All compartments were ®lled with normal Ringer solution (composition in mM: NaCl 115; KCl 1.8; CaCl2 1.6; [N±(2±hydroxyethyl)piperazine±N 0 ±2±ethanesulfonic acid] (HEPES) 5; pH 7.2), except for the fourth compartment, which was ®lled with isotonic (180 mM) sucrose solution that was continuously renewed, to electrically isolate the neighboring recording compartments. Tstx and modi®ed Ringer solutions were introduced into the test (central) compartment, with a total volume of 0.2 ml. Li 1containing Ringer solutions, were prepared by stoichiometrically replacing part or all of the NaCl with LiCl. The potential difference between the test and the 5th compartment was recorded every 5 or 10 min. Data were converted to digital form by a microcomputer-based 12-bit A/D converter at a rate of 13 kHz and later analyzed using a suite of programs (written by P.S.L. BeiraÄo) in Pascal programming language. To quantify the effects of Tstx in normal and Li 1substituted Ringer we used the relative post-potential (RPP) that is de®ned as the ratio of a parameter V (which is the potential difference between the baseline and the voltage of the CAP at either 5 or 20 ms after its onset) to the amplitude (which is the potential difference between the baseline and the maximal voltage of the CAP). RPP was shown to quantitatively describe the effect of Tstx [18]. All experiments were conducted at room temperature (25±288C). Data are presented as mean ^ SEM of the indicated number of independent experiments. To assess signi®cance level, two tailed Student's t-test was used, and differences between experimental and control groups were considered signi®cant when P , 0:05. Fig. 1A illustrates a typical example of the effects of Tstx (225 nM) on the CAP. The toxin concentration used throughout this paper was chosen based on previous work [14], where the authors reported that 225 nM was a concentration that produced a signi®cant increase in internal Na 1 concentration, as measured by micro¯uorimetry. The CAP developed a plateau after 5 min of exposure to Tstx (225 nM) in normal Ringer solution, which leads to a marked increase of CAP duration. The plateau reached its maximum height after approximately 20 min of exposure to the toxin

Fig. 1. Effect of tityustoxin (Tstx) on nerve compound action potential (CAP). Panel (A) shows representative superimposed CAP's, in the absence (control, trace a) and 20 min after the addition of Tstx (225 nM, trace b) to the normal Ringer solution. Panel (B) shows superimposed CAP's recorded in normal Ringer solution in the absence (control, trace a) and in the presence of Tstx 225 nM (20 min exposure, trace b), after wash-off the toxin with Li 1-Ringer solution 20 min, trace c) and after toxin-free Na 1 -Ringer (10 min, trace d). Stimulation parameters: 4 V/120 ms.

(Fig. 1A, trace b). The relative post-potential (RPP) at 5 ms averaged 0.059 ^ 0.007 (n ˆ 8) and 0.398 ^ 0.053 (n ˆ 4) for control (trace a) and in the presence of Tstx (225 nM, P , 0:05, trace b), respectively. The described effect of Tstx was not reversible even with prolonged washing with normal Ringer solution (data not shown, see Ref. [18]). In the experiment of Fig. 1B, Tstx (225 nM) was applied for a period of 20 min in normal Ringer solution (Fig. 1B, trace b). At the end of this period we replaced the solution inside the test compartment with a solution containing 115 mM Li 1 without Na 1 (trace c). One can see that the typical effect of Tstx disappeared almost completely. In the presence of 115 mM Li 1, and absence of Na 1, the CAP recorded had the same shape as the one recorded in the control conditions (trace a). RPP at 5 ms averaged 0.059 ^ 0.007 (n ˆ 8) and 0.066 ^ 0.036 (n ˆ 4, P . 0:05) for normal Ringer and Li 1-Ringer ([Li 1]e ˆ 115 mM), respectively. The major difference observed was the CAP amplitude when Li 1 replaced Na 1. It was always lower than that measured in normal Ringer (73.8 ^ 4.2 mV for control, n ˆ 8 and 65.6 ^ 2.9 mV for Li 1-Ringer, n ˆ 4, P , 0:05). This is not surprising as Li 1 ions reportedly conduct 30% less current than Na 1 ions through myelinated nerve Na 1 channels [9]. Surprisingly, the effect on the CAP recovered completely in about 10 min when Li 1-Ringer was replaced with Na 1-Ringer (Fig. 1B, trace d). Although the toxin was not present during the washout period in the Na 1-Ringer solution, the recovered CAP was prolonged, thus suggesting that the toxin was still bound to its receptor site in the Na 1 channel. These data were representative of the results from four different nerve preparations. In order to determine whether there was a relationship between Li 1 concentration and the inhibition of Tstx action, as evaluated by the RPP at either 5 or 20 ms, we performed a series of experiments in which we varied the extracellular Na 1 and Li 1 concentrations, but keeping their sum equal to

J.S. Cruz et al. / Neuroscience Letters 282 (2000) 25±28

115 mM. Fig. 2 shows that as we decrease external [Na 1] with the Tstx present from 115 to 65 mM the RPP at 5 ms signi®cantly increased from 0.40 ^ 0.05 (n ˆ 4) to 0.61 ^ 0.08 (n ˆ 4, P , 0:05). Interestingly, if we look at the RPP values at 20 ms there is a clear difference in the amount of change as the [Na 1]e decreases. When all extracellular Na 1 was replaced by Li 1, the mean RPP taken at 5 and 20 ms, after 20 min incubation, measured 0.21 ^ 0.06 (n ˆ 4) and 0.12 ^ 0.04 (n ˆ 4), respectively. These values are closer to control RPP's suggesting that Tstx had signi®cantly less effect on Na 1 channel inactivation when Li 1 fully replaced Na 1 as the main charge carrier. The next question is whether the observed lack of effect of Tstx is a consequence of the absence of Na 1 or of the presence of Li 1 in the extracellular solution. The results shown in Fig. 2 favor the ®rst possibility, because of the marked difference of the RPP values in the presence of similar concentrations of Li 1 (95 and 115 mM, Fig. 2, last two groups of bars). The only difference between these groups is the absence of Na 1. To test this hypothesis, in the experiment shown in Fig. 3 the Li 1 concentration was kept constant at 95 mM, in the presence of 20 mM of either Na 1 or choline 1 (i.e. zero Na 1), which is a non-permeant cation through voltagedependent Na 1 channels. The effect obtained in the presence of 20 mM Na 1, during 30 min incubation with Tstx (225 nM), was not signi®cantly different from that observed in Normal Ringer solution. However, there was a marked difference when 20 mM Na 1 was replaced with 20 mM choline 1 (Fig. 3). These data support the idea that to fully exert its effect on Na 1 channels fast inactivation, Tstx requires the presence of extracellular Na 1. Where is Na 1 required to promote Tstx effect? It has been shown that a reduction of the cationic concentration in the external millieu led to a decrease in the KD value of the Leiurus-toxin, an a-scorpion toxin that shows similarity

Fig. 2. Effect of Na 1 removal on Tstx action. Relative post-potential (RPP) changes under different external Na 1 concentrations with Li 1 replacing Na 1 equimolarly keeping the sum equals to 115 mM. Open columns represents the mean RPP values measured at 5 ms. Filled columns indicate mean RPP values at 20 ms. *P , 0:05 compared with 225 nM Tstx (Normal Ringer). The averaged results are means ^ SEM of at least four nerve bundles.

27

Fig. 3. Tstx action requires extracellular Na 1. Tstx was incubated at 225 nM ®nal concentration and we measured CAP's every 5 min. The ®gure shows the mean values obtained after 30-min incubation in the different conditions as described in the text. Open columns represent RPP values measured at 5 ms and ®lled columns the RPP values at 20 ms. The averaged results are mean ^ SEM for three different nerve bundles. *Statistically signi®cant compared with control at P , 0:05.

with Tstx [15]. The authors suggested a competitive inhibition of the toxin binding by Na 1 ions. So far our results suggest that Tstx is not removed from its binding site in the Na 1 channel upon substitution of Li 1 for Na 1. If this had occurred, rinsing the nerve with normal Na 1 Ringer (without Tstx) would not have restored its full effect, as shown in Fig. 1, panel B, trace d. This conclusion was further supported by the experiment depicted in Fig. 4, showing that the Tstx (225 nM) can bind to its target in the total absence of Na 1, although exerting a modest effect (Fig. 4). The presence of bound toxin was revealed by the addition of Normal Ringer solution (without Tstx), condition where the full effect of Tstx was then observed. It is assumed that the opening and closing (gating process) of an ion channel is normally a process independent from the permeation of ions through the open pore. However, several examples have been reported in the literature where the concentration or nature of the permeant ion appears to have effects on gating [3,13,19,20]. Changes in extracellular Na 1 concentration have been shown to produce small effects on either the kinetics or voltage-dependence of Na 1 currents of squid axons [2]. Yet, an anomalous effect was reported in which there was a decrease in Na 1 current despite an increase of driving force when extracellular Na 1 concentration was raised [19]. Further, with low extracellular concentrations of permeant cation (Li 1, Na 1 or hydrazinium), a fraction of Na 1 channels do not open at large depolarizations. In addition, both permeant and impermeant alkali metal cations in the extracellular solution affected the kinetics and steadystate levels of slow inactivation in human heart Na 1 channels expressed transiently [20]. On the other hand, changes of intracellular Na 1 concentration barely affects the rates and steady-state inactivation of Na 1 channels [2].

28

J.S. Cruz et al. / Neuroscience Letters 282 (2000) 25±28

Fig. 4. Tstx can bind to its site in the presence of Li 1. The nerve trunks were pre-incubated ®rst with Ringer solution containing neither Na 1 nor Tstx. Afterwards the toxin (225 nM) was added to the Na 1-free Ringer solution. After 30 min there was a slightly increase in the RPP at 5 ms which was further increased when Na 1 was added to the test chamber. *Statistically signi®cant compared with Tstx (225 nM, zero Na 1, 30 min, n ˆ 5).

Lilley and Robbins [12] studying the effects of local anesthetics on isolated frog sciatic nerve concluded that procaine, lignocaine and benzocaine were more potent when Li 1 rather than Na 1 was used as permeant ion and suggested that a simple interaction between the permeant ion and the drug binding sites may not fully explain their results but rather a complex allosteric interaction could at least in part be involved. Although the mechanism underlying the phenomenon observed here have to be further explored, we concluded, on the basis of present results, that Tstx requires the presence of low concentrations (20 mM) of Na 1 to produce its typical effect on the inactivation kinetics of Na 1 channels. Furthermore, we suggested that the binding of Tstx to its site does not require Na 1, and indicate that the Tstxinduced modi®cation of the fast inactivation kinetics is the Na 1-dependent step. The authors wished to thank Mauricio Mudado for help in some of the experiments. We would also like to thank Dr. Laura Martin for helpful comments on the manuscript. This study was supported by FAPEMIG, CNPq, Fundo-FUNDEP. P.S.L. BeiraÄo is recipient of a research fellowship from CNPq. [1] Adam, K.R., Schmidt, H., StaÈmp¯i, R. and Weiss, C., The effect of scorpion venom on single myelinated nerve ®bres of the frog. Br. J. Pharmacol., 26(3) (1966) 666±677. [2] Chandler, W.K. and Meves, H., Rate constants associated with changes in sodium conductance in axons perfused with sodium ¯uoride. J. Physiol. (Lond.), 211(3) (1970) 679±705. [3] Correa, A.M. and Bezanilla, F., Gating of the squid sodium channel at positive potentials I. Macroscopic ionic and gating currents. Biophys. J., 66(6) (1994) 1853±1863.

[4] Coutinho-Neto, J. and Diniz, C.R., Properties of Tityustoxin obtained by a new method of puri®cation. Ann. Acad. Bras. Cienc, 46(3/4) (1974) 698. [5] Cruz, J.S., Cotta, G., Diniz, C.R. and BeiraÄo, P.S.L., Partial puri®cation and pharmacological characterization of a neurotoxic fraction isolated from the venom of the spider Lycosa erythrognatha. Braz. J. Med. Biol. Res., 27(11) (1994) 2653±2659. [6] de Lima, M.E. and Martin-Eauclaire, M.-F., The toxins puri®ed from Tityus serrulatus (Lutz & Mello) venom. J. Toxicol. Toxin Rev., 14(3) (1995) 457±481. [7] Gomez, M.V. and Diniz, C.R., Separation of toxic components from Brazilian scorpion Tityus serrulatus venom. Mem. Inst. Butantan, 33(3) (1966) 899±902. [8] Haynes, L.W., Permeation of internal and external monovalent cations through the cat®sh cone photoreceptor cGMPgated channel. J. Gen. Physiol., 106(3) (1995) 485±505. [9] Hille, B., The Permeability of the sodium channel to metal cations in myelinated nerve. J. Gen. Physiol., 59(6) (1972) 637±658. [10] Ismailov, I.I., Shlyonsky, V.G., Alvarez, O. and Benos, D.J., Cation permeability of a cloned rat epithelial amiloridesensitive Na 1 channel. J. Physiol. (Lond.), 504(2) (1997) 287±300. [11] Kurata, Y., Sato, R., Hisatome, I. and Imanishi, S., Mechanisms of cation permeation in cardiac sodium channel: description by dynamic pore model. Biophys. J., 77(4) (1999) 1885±1904. [12] Lilley, S.J. and Robbins, J., The action of local anaesthetics on the compound action potential is altered by the nature of the permeant ion in frog nerve. Neurosci. Lett., 252(1) (1998) 41±44. [13] Lonnendonker, U., Neumcke, B. and Stamp¯i, R., Interaction of monovalent cations with tetrodotoxin and saxitoxin binding at sodium channels of frog myelinated nerve. P¯uÈgers Arch., 416(6) (1990) 750±757. [14] Massensini, A.R., Moraes-Santos, T., Gomez, M.V. and Romano-Silva, M.A., Alpha- and beta-scorpion toxins evoke glutamate release from rat cortical synaptosomes with different effects on [Na 1]i and [Ca 21]i. Neuropharmacology, 37(3) (1998) 289±297. [15] Okamoto, H., Takahashi, K. and Yamashita, N., One-to-one binding of a puri®ed scorpion toxin to Na channels. Nature, 266(5601) (1977) 465±468. [16] Ogielska, E.M. and Aldrich, R.W., A mutation in S6 of Shaker potassium channels decreases the K 1 af®nity of an ion binding site revealing ion-ion interactions in the pore. J. Gen. Physiol., 112(2) (1998) 243±257. [17] Sampaio, S.V., Laure, C.J. and Giglio, J.R., Isolation and characterization of toxic proteins from the venom of the Brazilian scorpion Tityus serrulatus. Toxicon, 21(2) (1983) 265±277. [18] Silveira, J.N., Heneine, I.F. and BeiraÄo, P.S., Reversion by polyclonal antibodies of alpha effects of Tityus serrulatus venom on frog sciatic nerve. Toxicol. Lett., 76(3) (1995) 187± 193. [19] Townsend, C., Hartmann, H.A. and Horn, R., Anomalous effect of permeant ion concentration on peak open probability of cardiac Na 1 channels. J. Gen. Physiol., 110(1) (1997) 11±21. [20] Townsend, C. and Horn, R., Effect of alkali metal cations on slow inactivation of cardiac Na 1 channels. J. Gen. Physiol., 110(1) (1997) 23±33. [21] Zacharova, D., Zachar, J., Hencek, M. and Nasledov, G.A., Selectivity of sodium channels in denervated tonic muscle ®bre membrane of the frog. Gen. Physiol. Biophys., 2(6) (1983) 425±435.