Article in Press

ARTICLE IN PRESS

Toxicon 50 (2007) 507–517 www.elsevier.com/locate/toxicon

Characterization of the excitatory mechanism induced by Jingzhaotoxin-I inhibiting sodium channel inactivation

Yucheng Xiao, Jiang Li, Meichun Deng, Changliang Dai, Songping LiangÃ

Life Sciences College, Hunan Normal University, Changsha, Hunan 410081, PR China

Received 11 February 2007; received in revised form 15 April 2007; accepted 23 April 2007 Available online 3 May 2007

Abstract

We have recently isolated a peptide neurotoxin, Jingzhaotoxin-I (JZTX-I), from Chinese tarantula Chilobrachys jingzhao venom that preferentially inhibits cardiac sodium channel inactivation and may deﬁne a new subclass of spider sodium channel toxins. In this study, we found that in contrast to other spider sodium channel toxins acting presynaptically rather than postsynaptically, JZTX-I augmented frog end-plate potential amplitudes and caused an increase in both nerve mediated and unmediated muscle twitches. Although JZTX-I does not negatively shift sodium channel activation threshold, an evident increase in muscle fasciculation was detected. In adult rat dorsal root ganglion neurons JZTX-I (1 mM) induced a signiﬁcant sustained tetrodotoxin-sensitive (TTX-S) current that did not decay completely during 500 ms and was inhibited by 0.1 mM TTX or depolarization due to voltage-dependent acceleration of toxin dissociation. Moreover, JZTX-I decreased closed-state inactivation and increased the rate of recovery of sodium channels, which led to an augmentation in TTX-S ramp currents and decreasing the amount of inactivation in a use-dependant manner. Together, these data suggest that JZTX-I acted both presynaptically and postsynaptically and facilitated the neurotransmitter release by biasing the activities of sodium channels towards open state. These actions are similar to those of scorpion a-toxin Lqh II. r 2007 Elsevier Ltd. All rights reserved.

Keywords: Sodium channels; Spider toxin; Dorsal root ganglia; Whole-cell recording

1. Introduction tissues. Their characteristic structures are found to contain a pore-forming functional a subunit and The opening of voltage-gated sodium channels several auxiliary b subunits (b1–b4). The two kinds (VGSCs) plays a critical role in the generation and of subunit interact with each other by covalent or propagation of action potentials on excitable non-covalent bonds (Ogata and Tatebayashi, 1993; Catterall, 2000; Yu et al., 2003). Although up to 10 Abbreviations: dTc, d-Tubocurarine; DRG, dorsal root gang- mammalian VGSC isoforms (Nav1.1–1.9 and Navx) lia; ICK, inhibitor cystein knot; JZTX, Jingzhaotoxin; TTX, have been identiﬁed, cloned and functionally tetrodotoxin; TTX-R, tetrodoxin resistant; TTX-S, tetrodotoxin characterized (Ogata and Ohishi, 2002), they are sensitive; VGSC (and Nav), voltage-gated sodium channel ÃCorresponding author. Tel.: +86 731 8872556; highly conserved during evolution based on the fax: +86 731 8861304. analysis of a subunit protein sequences and main- E-mail address: [email protected] (S. Liang). taining similar pharmacological functions (Catterall

0041-0101/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.toxicon.2007.04.018 ARTICLE IN PRESS 508 Y. Xiao et al. / Toxicon 50 (2007) 507–517 et al., 2003). Some VGSC isoforms may be unrelated to toxin partitioning in the phospholipid expressed selectively in certain tissues and even the bilayer of neuronal membranes (Cohen et al., 2006; same isoform can display quantitative differences in Lee and MacKinnon, 2004). channel kinetics after expression in different tissues. Jingzhaotoxin-I (JZTX-I) is a VGSC toxin newly Cummins et al. (2001) demonstrated that the reported from Chinese tarantula Chilobrachys jingz- repriming kinetics for rat Nav1.3 were twofold hao venom (Zhu et al., 2001). Composed of 33 faster in neurons than in human embryonic kidney residues and stabilized by three disulfide bonds 293 cells. Furthermore, mutants of VGSC a subunit (I–IV, II–V, III–VI), the neurotoxin preferentially proteins can occur naturally, leading to dysfunction inhibits cardiac sodium channel inactivation (Xiao of sodium channels. For example, I848T and L858H et al., 2005). The characteristics of its secondary mutants of human Nav1.7 cause a significant structure are similar to that of huwentoxin-I in hyperpolarizing shift in the V1/2 of channel activa- solution, which mainly contains a short triple- tion (Cummins et al., 2004). To date, over 150 stranded antiparallel b-sheet (Liang et al., 2000; naturally occurring VGSC mutants have been Zeng et al., 2005). More recently, we reported that characterized in several human hereditary diseases, the toxin also inhibited Kv2.1 and Kv4.1, but the such as skeletal muscle myotonias, episodic ataxia affinities were over fivefold lower than to neuronal and epilepsy (Meisler and Kearney, 2005). There- TTX-S VGSCs (Yuan et al., 2007). In this study, we fore, novel potent VGSC ligands will be of great characterized the excitatory responses of JZTX-I on significance as drugs for alleviating diseases and as several in vitro preparations and the underlying tools for investigating the structure and function mechanism on TTX-S VGSCs expressed in adult rat relationships of sodium channels. dorsal root ganglion (DRG) neurons. For defense or predation, many animals evolve a venom gland to secrete toxins targeting sodium 2. Methods channel. Most of them are of protein structures stabilized by three or four intramolecular disulfide 2.1. Toxin purification bridges. They can be folded into inhibitor cystein knot (ICK) or a/b motifs, the former being frequent JZTX-I was fractionated from Chinese tarantula in spider toxins, sea anemone toxins and conotoxins C. jingzhao venom, using ion-exchange chromato- and the latter in scorpion toxins (Xiao et al., 2004; graphy followed by reverse-phase high pressure Karbat et al., 2004). These VGSC modulating liquid chromatography as previously described toxins have been classified into two groups accord- (Xiao et al., 2005). After the purity of the sample ing to the properties: (1), pore-blocking toxins, such was up to 99% calculated by the analysis of matrix- as m-conotoxins, which occlude the channel pore to assisted laser desorption/ionization time-of-flight block inward sodium currents by binding to site 1 in mass spectrometry and eluted peak area, it was a manner similar to tetrodotoxin (TTX) and stored at 20 1C until required. saxitoxin (Shon et al., 1998). Unaffecting sodium channel activation and inactivation kinetics, several tarantula toxins, such as huwentoxin-IV, hainan- 2.2. Tissue preparation toxin-III and hainantoxin-IV, are also assumed to belong to the family (Peng et al., 2002; Xiao and Phrenic nerve-diaphragm preparation was acutely Liang, 2003); (2), gate-modifying toxins, such as isolated from the 20 g Kunming mouse of either sex a/b-scorpion toxins and sea anemone toxins, which as described by Zhou et al. (1997). After isolation, modulate the movement of the voltage sensor to the fresh preparations were immediately immersed affect the activities of activation gate and inactiva- into Tyrode’s solution containing (mM): NaCl tion ball by binding to neuronal site 2–6 (Rogers 143.0, KCl 5.4, NaH2PO4 0.3, MgCl2 0.5, glucose et al., 1996; Richard Benzinger et al., 1999). 10.0, HEPES 5.0, CaCl2 1.8 at pH7.2. Tyrode’s b-Scorpion toxins binding to neuronal site 4 toxins solution was kept to be bubbled at 30–32 1C by 95% trap the voltage sensor in the activated state in a O2 and 5% CO2. Electrical stimulation was applied voltage dependent but concentration-independent indirectly to the phrenic nerve with a suction manner (Cestele et al., 2006). Different from electrode, or directly to the muscle using a same potassium channel modifiers, the underpinning protocol at a frequency of 0.06 Hz (supramaximal, mechanism of this is clearly disclosed to be 0.2 s, square wave). ARTICLE IN PRESS Y. Xiao et al. / Toxicon 50 (2007) 507–517 509

Vas deferens preparation was acutely dissociated The output electrical signals after amplification from male Sprague–Dawley rats with about 200 g were displayed and simultaneously recorded using and bisected to be prostatic and epididymal a BL-420 biology function laboratory system segments as described by Liang et al. (2000). Before (Chengdu instrument, China). All sciatic nerve– both segments of the vas deferens with length of satorius muscle preparations were immersed in 2 cm were immerged into 5 ml physiological saline Ringer solution containing 5 mM d-tubocurarine solutions bubbled by 95% O2 and 5% CO2 and (dTc) for 10 min, and then the innervated muscle maintained at 33 1C, the top of each tissue was fiber at end-plate area was impaled with a micro- attached to an electromechanical transducer and the electrode. Suprathreshold electrical stimulation bottom attached to a movable support. The tissue (3 V, 1 ms) was continuously applied on the sciatic was straddled with platinum stimulating electrodes. nerve at a constant frequency of 1 Hz for 1 min. For For prostatic segments, the solution contained control and toxin tested groups, the preparations (mM): NaCl 119.0, KCl 4.7, CaCl2 2.5, NaHCO3 were treated under the same processes. 2.5, MgSO4?7H2O 1.2, KH2PO4 1.2, Glucose 11.0, EDTA 0.026 at pH 7.3, in which EDTA was used to 2.4. Patch-clamp recordings in rat DRG neurons eliminate the fasciculation, while for epididymal segments the solution contained (mM): NaCl 118.4; Rat DRG neurons were acutely dissociated and KCl 4.7; MgSO4 1.2; KH2PO4 1.2; CaCl2 2.5; maintained in a short-term primary culture accord- NaHCO3 25; D-glucose 11.1, as provided by Rash ing to the procedures adapted from (Xiao et al. et al. (2000). The prostatic segment of the vas deferens (2005). Sodium currents were recorded from ex- was stimulated with single electrical field pulses perimental cells using an EPC9 patch clamp (100 V, 0.1 ms wave width) for every 5 s while no amplifier (HEKA, Germany) at room temperature electrical stimulation was applied on the epididymal (22–25 1C). Recording pipettes (2–3 mm, diameter) segment to detect the fasciculation. Before application were made from borosilicate glass capillary tubes of the testing toxin, the isolated vas deferens was (VWR, USA) and its resistances were around equilibrated in Krebs solution for 30 min. 2.0 MO when filled with internal solution contained Heart preparation without vagosympathetic (mM): CsF 135, NaCl 10, HEPES 5 at pH 7.0. The nerve was acutely isolated from the 50-g toads as external bathing solution contained (mM): NaCl 30, described by Liang et al. (2000). During dissection, CsCl 5, D-glucose 25, MgCl2 1, CaCl2 1.8, HEPES 5, the blood in heart was replaced completely with tetraethylammonium chloride 20, tetramethylam- Ringer solution through a frog-heart tube contain- monium chloride 70 at pH 7.4. Series resistance was ing (in mM) NaCl 109.9, KCl 4.0, CaCl2 1.5, compensated to be 80–85%. After establishing the NaHCO3 4.3, Glucose 4.0, and the auto-heartbeat whole-cell recording configuration, the resting for fresh preparations were detected. Assays were potential was held at 80 mV for at least 4 min to conducted at room temperature. allow adequate equilibration between the micropip- The twitch responses induced in all tissues were ette solution and the cell interior. The P/4 protocol transformed into an electric signal by an electro- was used to subtract linear capacitive and leakage mechanical transducer. The output signals were currents. Experimental data were acquired and collected by RM6240B biological signal recording analyzed by the program pulse+pulsefit8.0 system (Biological electronic instrument, Chengdu, (HEKA, Germany). The needed concentrations of China). toxin dissolved in external solution were applied onto the experimental cell surface by low-pressure 2.3. End-plate potential recording injection with a microinjector (IM-5B, Narishige).

End-plate potentials (EPPs) were recorded from 2.5. Analysis the endplate regions in toad sciatic nerve–satorius muscle preparations using a conventional technique Data analysis was performed using the Pulsefit as described by Zhou et al. (1997). The resistances (HEKA, Germany) and Sigmaplot 8.0 (Sigma, of intracellular microelectrodes ranged from 10 to USA) software programs. All data were presented 15 MO when filled with 3 M KCl. The Ag–AgCl as mean7S.E.M, n represents the number of wire from the microelectrode was connected to a independent experiments, and I5ms is the current ME-100 amplifier (Chengdu instrument, China). measured at the depolarization of 5 ms. Statistical ARTICLE IN PRESS 510 Y. Xiao et al. / Toxicon 50 (2007) 507–517 significance of toxin effect was determined using paired Student’s t-test with Po0.05 considered significant.

3. Results

3.1. Biological assays

Animal venoms contain many components, which are toxic to preys or enemies by disrupting the activities of organs. Huwentoxin-I from Chinese tarantula Ornithoctomus huwena blocks synaptic transmission, inhibiting muscle contraction induced by an indirect stimulation (Liang et al., 2000). a-Pompilidotoxin from wasp venom facilitates excitatory and inhibitory synaptic transmission (Sahara et al., 2000). JZTX-I is lethal to American cockroaches (LD50 26.42 mg) after abdominal injection and Kunming mice (1.48 mg/kg) by intraper- itoneal injection. However, in contrast to Fig. 1. Typical effects of JZTX-I on rat vas deferens preparation. Huwentoxin-I, JZTX-I did not block synaptic (A) JZTX-I at 2.3 mM significantly increased the twitch tension transmission but significantly increased muscle on the prostatic segment of the rat vas deferens induced by twitches induced by indirect or direct electrical electrical stimulation (above). Stimulation condition: single rectangular current pulse of width (0.1 ms, strength 100 V at stimulations. On exposure to 2.3 mM toxin, the 33 1C for every 5 s). JZTX-I at 2.3 mM twofold increased the augmentation was over twofold more in rat vas muscle contraction. JZTX-I at 2.3 mM also significantly poten- deferens (Fig. 1A and B), mouse phrenic nerve tiated the fasciculation of the epididymal segment of the rat vas diaphragm (Fig. 2A) and toad heart preparations deferens on which no additional stimulation was applied (below). (Fig. 2B and C), followed by a slight decrease after (B), JZTX-I affected the muscle tension in a time-dependent manner on the stimulated prostatic segment and on the 10 min toxin treatment (seen particularly in Fig. 2A, fasciculation of the epididymal segment of the rat vas deferens. n ¼ 4). Muscle contraction induced by indirect The data points were derived from the data in (A) (n ¼ 3). stimulation could be inhibited completely by the nicotinic receptor antagonist dTc at 15 mM in mouse phrenic nerve-diaphragm preparation (Fig. 2A, parations from adult toad. Only the preparations n ¼ 3). Addition of 2.3 mM toxin to dTc-pretreated giving stable EPPs to nerve stimulation within an tissues also evoked twofold augmentation of muscle initial 10 min were used in our experiments. As seen contraction induced by direct stimulation (Fig. 2A, in Fig. 3, the averaged control EPP amplitude was n ¼ 3). In addition, the toxin at the same concen- quickly increased from 1.370.2 to 2.570.4 mV tration significantly potentiated the fasciculation of after 10 mM toxin treatment for 20 s (n ¼ 5) and the epididymal segment of rat vas deferens within no significant change was detected after the addition 40 min (Fig. 1B, n ¼ 3). JZTX-I prolonged the of Ringer solution containing 5 mM dTc (n ¼ 5). No duration of action potentials on frog sartorial repetitive EPPs were detected following a single muscles and sciatic nerves (data not shown, presynaptic stimulation after JZTX-I treatment n ¼ 4). These results supported that JZTX-I could (data not shown, n ¼ 8). modulate the activities of skeletal, smooth and cardiac muscles and neurons, and also suggested 3.3. Effects of JZTX-I on rat DRG TTX-S sodium that JZTX-I might target the VGSC isoforms currents expressed in these tissues. We have recently demonstrated that different 3.2. Effects of JZTX-I on end-plate potentials from other spider toxins but similar to scorpion a-like toxins, JZTX-I neither evidently shifts the The effects of JZTX-I on the EPPs were measured steady-state inactivation nor changes the current– on dTc-treated sciatic nerve-satorius muscle pre- voltage relationship curve of neuronal TTX-S ARTICLE IN PRESS Y. Xiao et al. / Toxicon 50 (2007) 507–517 511

Fig. 2. JZTX-I increased the twitch response of mouse phrenic nerve diaphragm and toad heart preparation. (A) Mouse phrenic nerve diaphragm preparation. The twitch tensions were induced by indirectly (Above) or directly (Below) electrical stimulation at a frequency of 0.06 Hz (supramaximal, 0.2 s, square wave). D-tubocurarine (dTc) at 15 mM eliminated the indirectly induced muscle contraction completely within 5 min, and then the twitch tension was induced by directly electrical stimulation (Below). In the presence of 2.3 mM JZTX-I, an increasing response was observed within 30 min. (B) Toad heart preparation. Cardiac muscle contraction was elicited spontaneously and no additional stimulation was applied. The twitch response was increased around threefold after the treatment of 2.3 mM JZTX-I. (C) Graph illustrating the ratio of the peak contraction on several in vitro preparations, such as mouse phrenic nerve diaphragm, rat vas deferens and toad heart preparations, induced by 2.3 mM JZTX-I.

VGSCs including peak current amplitudes (Fig. 4A, from 21.072.0% to 12.871.5% at the depolariza- n ¼ 3) (Nicholson et al., 1998; Xiao et al, 2003, tion from 0.1 to 0.5 s (n ¼ 6). The time constants of 2005). The midpoint of activation was 17.575.0 channel inactivation were significantly prolonged and 16.575.7 mV before and after application of before (5.571.5 ms, n ¼ 3) and after (51.8713.0 ms, 1 mM toxin, respectively (P40.05). Here, we chose n ¼ 3) toxin treatment. The induced sustained the IC50 concentration (1 mM) to further characterize current could be blocked completely by 100 nM the modulation of JZTX-I on TTX-S VGSCs from TTX in Fig. 4B (n ¼ 5), implying that JZTX-I did rat DRG neurons. Fig. 4B showed that JZTX-I not induce the opening of other ion channels, but induced a significant sustained current on sensory VGSCs. We also checked the effect of the toxin on the neurons by inhibiting VGSC inactivation, not currents induced by slow ramp (0.2 mV/ms) depolar- decaying completely even after application of a izations from 80 to +40 mV. The ramp current 500 ms depolarizing pulse of 10 mV. Percent of initially activated around 55 mV was augmented sustained currents to peak currents was decreased significantly by 1 mMtoxin(Fig. 4C, n ¼ 3). ARTICLE IN PRESS 512 Y. Xiao et al. / Toxicon 50 (2007) 507–517

Fig. 3. Effect of JZTX-I on EPPs recorded from frog sciatic nerve-satorius muscle preparations. (A) Left: control; right:20s after addition of 10 mM toxin. (B), Left: control; right: 20 s after addition of Ringer solution containing 5 mM dTc. Note that the stimulus artifact was deleted.

Rogers et al. (1996) demonstrates that inhibiting VGSC fast inactivation, the binding of scorpion a-toxin LqTx is substantially dissociated during stronger depolarization. A similar dissociation of JZTX-I-VGSC complex was detected in our study. Experimental cells were clamped at 80 mV and then strongly depolarized to +80 or +120 mV for an increasing duration from 0 to 512 ms to induce toxin dissociation. After restoring to the holding potential for 20 ms to reverse channel inactivation, the cells were depolarized by a 10 ms test potential of 0 mV to measure available sodium currents in Fig. 5A (inset). The toxin dissociation rate was determined by the ratio of I5ms to Ipeak elicited during the test pulse and plotted as a function of conditioning pulse duration. In Fig. 5B, JZTX-I binding could be substantially dissociated from TTX-S VGSCs in a voltage-dependent manner. At +80 mV, the ratio of I5ms to Ipeak was markedly decreased from 0.5570.04 to 0.1470.05 whereas the ratio was from 0.5070.03 to 0.0170.01 at +120 mV (n ¼ 5–6). Their time constants of dissociation were 37.7710.6 and 50.874.3 ms, respec- Fig. 4. Effects of JZTX-I on TTX-S sodium currents from rat tively. The dissociation of JZTX-I was not complete DRG neurons. (A) Current–voltage (I–V) relationship for at +80 mV even with duration of 512 ms. However, VGSCs before (open circles) and after JZTX-I treatment (ﬁlled at the same strong depolarization the dissociation of circles) (n ¼ 3). I–V curves before and after JZTX-I treatment were induced by 50 ms depolarizing steps to various potentials LqTx-channel is complete in about 400 ms, indicat- from a holding potential of –80 mV. Test pulses ranged from ing that JZTX-I-channel complexes might be more –80 mV to +60 mV. I5ms was shown as the current inactivated at stable than that of scorpion a-toxin and channel. 5 ms (ﬁlled squares). (B) Sustained currents induced by JZTX-I did not decay evidently within over 500 ms (n ¼ 6). Single current 3.4. Effects of JZTX-I on recovery rate of rat DRG trace was elicited by a 500 ms depolarizing potential of 10 mV. Five data points coming from six experimental cells were shown VGSC inactivation as mean7S.E. (C) Augmentation of ramp currents induced by JZTX-I. Currents were elicited by 600 ms voltage ramps from The effect of JZTX-I on recovery rate from 80 to +40 mV and the cells were clamped at 80 mV (inset). inactivation was determined using a standard two- The peaks of ramp currents were indicated before (K) and after J pulse protocol, in which the cell clamped at 80 mV toxin application ( ). ARTICLE IN PRESS Y. Xiao et al. / Toxicon 50 (2007) 507–517 513

Fig. 5. Voltage-dependent dissociation of JZTX-I from neuronal TTX-S VGSCs. (A) After 1 mM JZTX-I treatment for 4 min, family of sodium current traces was induced on rat DRG neurons by a 20 ms test pulse of 0 mV. Before applying the test pulse, experimental cells held at 80 mV were depolarized by a conditioning pulse of +120 mV for 0–512 ms and then backed to holding potential for 20 ms. The current traces showed that sustained current amplitudes were decreased more when the conditioning pulse lasted for longer, indicating that the dissociation of JZTX-I from VGSCs could be accelerated with longer conditioning pulse. (B) The rate of dissociation of 1 mM JZTX-I from VGSCs at conditioning pulse of +80 (filled circles) and +120 mV (open circles) (n ¼ 5–6). The ratio of current at the depolarization of 5 ms (I5ms) to peak current (Ipeak) was plotted as a function of time during conditioning pulse. was applied with a 50 ms conditioning pulse of critical role in determining the generation of 10 mV, then repolarized to the holding potential subthreshold currents and further affects the firing of 80 mV for an interval between 0.5 ms and 1 s frequency of action potentials on excitable tissues followed by a 20 ms test pulse of 10 mV to activate (Cummins et al., 1998). In order to measure the VGSCs available (Fig. 6A, inset). In control effects of JZTX-I on development of TTX-S conditions, almost no currents were elicited by test inactivation on rat DRG neurons, another two- pulse after the interpulse duration of 0.5 ms, and pulse protocol was introduced. Clamped at almost all inactivated VGSCs had recovered after 100 mV, experimental cells were stepped to a 50 ms. Recovery rate from inactivation was assayed depolarizing potential of 60 or 50 mV for by calculating the ratio of Itest to Icond. In presence increasing duration from 0 to 512 ms, followed by a of 1 mM JZTX-I for 1 min, fast inactivation inhibi- test pulse of 10 mV for measurement of activating tion was detected markedly in all induced current TTX-S sodium currents (Fig. 6C, inset). The ratio of traces by both test and conditioning pulse (Fig. 6A, Itest to Icond was plotted against the increasing n ¼ 4). Fig. 6B showed that exposure to 1 mM duration (Fig. 6D). The closed-state inactivation JZTX-I could result in 39.574.0% VGSCs activat- was slower for TTX-S VGSCs expressed in DRG ing after the interpulse duration of 0.5 ms when neurons at 60 than at 50 mV (n ¼ 5). JZTX-I at TTX-S VGSCs returned to resting (closed) state. 1 mM evidently decreased the fraction of channels Although the time constants were changed before inactivated at 50 but not at 60 mV. In absence (2.370.3 ms) and after (4.070.7 ms, n ¼ 4) applica- and presence of toxin, the time constants were tion of toxin, the interpulse duration (T1/2) at which 43.8714.3 and 26.473.6 ms at 50 mV (Po0.05) half the maximal VGSCs had recovered was and 58.7716.6 and 54.8712.0 ms at 60 mV decreased from 2.3 to 1.1 ms. The results supported (P40.05), respectively (Fig. 6D, n ¼ 5). It indicates that JZTX-I, unlike tarantula toxin Hainantoxin- that JZTX-I reduced the development of TTX-S III, increased the repriming kinetics of TTX-S VGSC inactivation from rat DRG neurons. VGSCs on rat DRG neurons (Xiao et al., 2003). 3.6. Use-dependent effect of JZTX-I on rat DRG 3.5. Effects of JZTX-I on development of rat DRG TTX-S VGSCs VGSC inactivation In Fig. 7A, a series of inward current traces was Closed-state inactivation is another important elicited by a train of pulses at firing frequencies of kinetics property for VGSCs, which might play a 30 Hz (inset). The data obtained showed that the ARTICLE IN PRESS 514 Y. Xiao et al. / Toxicon 50 (2007) 507–517

Fig. 6. Effects of JZTX-I on the rate of recovery from inactivation and on the development of inactivation on rat DRG neurons. (A) The cell was held at 80 mV and a dual-pulse voltage clamp protocol was detailed in the inset using two 50 ms pulses (conditioning and test pulses) from 80 to 10 mV with a varying interval from 0.5 ms to 1 s. Typical current traces represent the recovery from inactivation before (above) and after (bellow) of 1 mM JZTX-I treatment. (B) Data points represent the ratio of the peak currents induced by test pulse

(Itest) to the peak currents by conditioning pulse (Icond)(n ¼ 5). Plotted against the duration of the interval between two pulses, they have been fitted using a single exponential equation. The filled and open circle symbols present for control and 1 mM JZTX-I, respectively. (C) Family of current traces represent the development of inactivation before (above) and after (below) of 1 mM JZTX-I treatment. Current traces were elicited by a dual-pulse protocol shown in inset, in which the cell held at 100 mV was applied with a conditioning pulse of 50 mV for an increasing duration from 0 to 512 ms and then stepped to a test pulse of 10 mV to measure the fraction of current available. (D) Time course of development of inactivation for peak currents (n ¼ 6). Data points show the ratio of peak currents to that measured at conditioning pulse for 0 ms and were plotted as a function of time during the conditioning pulse. The data are fitted to a single exponential function. The circle and square symbols present for control (filled) and 1 mM JZTX-I (open) at 50 and 60 mV, respectively. inhibition of channel inactivation by 1 mMJZTX-Iwas C. jingzhao venom, is a 33-residue peptide stabilized detected at every pulse and the peak current amplitude by three disulfide bonds (I–IV, II–V, III–VI) (Xiao for the first pulse remained unchanged before and after et al., 2005). Different from other tarantula toxins toxin treatment. However, percent of the peak current such as ProTx I-II, Huwentoxin-IV and Hainantox- for the second pulse was increased from 82.5% to in-IV (Middleton et al., 2002; Peng et al., 2002; Xiao 95.0% and that for the 20th pulse was from 37.6% to et al., 2003), JZTX-I shows no evident effect on the 85.1% at 30 Hz (Fig. 7B). A similar increasing response voltage–current relationship or steady-state inacti- was also observed at 10 Hz (Fig. 7B). vation, although they may share a common structure with ICK motif (Zeng et al., 2005). In 4. Discussion this study, we further characterized its excitatory responses on several in vitro preparations and We have recently reported that JZTX-I, a VGSC investigated the underlying related mechanism on inactivation inhibitor from Chinese tarantula TTX-S VGSC inactivation kinetics. ARTICLE IN PRESS Y. Xiao et al. / Toxicon 50 (2007) 507–517 515

Sodium channel toxins are abundant in scorpion, spider, sea anemone and sea cone venoms. Altering the activation or inactivation kinetics of channel isoforms, they exert different responses on muscular tension from nerve–muscle preparations. Spider toxins are found to selectively target neuronal isoforms and few to target muscular and cardiac isoforms. Gregio et al. (1999) reported that Ancy- lometes spider venom could induce an excitatory response of muscle tension on phrenic nerve–diaphragm preparation only when not pretreated with dTc. Similarly, JZTX-I induced an increasing response to the muscle contractions on mouse phrenic nerve–diaphragm, toad heart, and rat vas deferens preparations. However, in our study, addition of toxin to dTc-pretreated mouse phrenic nerve-diaphragm preparation can still potentiate muscle twitches significantly. The duration of action potentials was also prolonged on both toad sartorial muscle and sciatic nerve. We therefore conclude that JZTX-I is an interesting spider toxin, which may act both presynaptically and postsynaptically. The toxin is also predicted to target skeletal muscular sodium channel subtype as well as neuronal and cardiac subtypes. d-Missulenatoxin-Mb1a, from the venom of the male Australian eastern mouse spider Missulena bradleyi, can be considered as a typical representative of known spider excitatory toxins (Gunning et al., 2003). They are found to induce both the inhibition of channel inactivation and the hyperpolarizing shift of channel activation. In contrast to d-missulenatoxin-Mb1a, JZTX-I exhi- bits no evident effect on the activation threshold of sodium channels, although both of them cause an increase in smooth muscle fasciculation. JZTX-I is Fig. 7. JZTX-I caused a use-dependent decrease in current an excitatory toxin facilitating the release of amplitude decline on rat DRG neurons. (A) Family of current neurotransmitter Ach to increase muscle twitching. traces before (above) and 4 min after (below) adding toxin was Rapid activation and fast inactivation kinetics are elicited by a 30 Hz pulse train. The cell, held at 80 mV, was depolarized to 10 mV for 22 ms for every 11 ms during the two important processes for VGSCs and directly 20-pulse train (inset). (B) The decline is determined by norma- determine the time course of action potentials lizing peak currents to the peak induced by the first pulse for (Cummins et al., 1998). The transition between control (filled symbols) and 1 mM JZTX-I (open symbols) at 10 them is considered as in the context of a simple (triangle) and 30 Hz (circle). (C) Gating transition scheme for channel gating scheme as shown in Fig. 7C, in which VGSC inactivation and activation multistate. The model was cited from the Ref. (Cummins et al., 1998) and thought that there are multiple closed and closed-inactivated VGSCs had multistate for inactivation and activation. Closed states as well as open and inactivated states. While state, open state, open-inactivated state and closed-inactivated site 1 toxins (TTX and m-conotoxins) simply occlude state are indicated by Cs, O, IO and ICs, respectively. These channel pore (Shon et al., 1998), other site toxins states can be transited from each other at certain depolarizing (site 2–6) often profoundly modulate the processes conditions. JZTX-I biased the transition of VGSC multistate as indicated by arrows. of VGSC kinetics including the transition among these states. Site 3 toxins, such as d-atracotoxins and scorpion a-toxins, are found to increase the recovery rate from channel inactivation (Nicholson ARTICLE IN PRESS 516 Y. Xiao et al. / Toxicon 50 (2007) 507–517 et al., 1998). Our study revealed that a similar also in JZTX-I binding to VGSCs (Xiao et al., increasing response was observed after JZTX-I 2005). Therefore, the mechanism for JZTX-I- treatment. JZTX-I markedly inhibits the conversion channel dissociation might be similar to that for of open state to open-inactivated state, allowing a LqTx-channel dissociation. However, the difference fraction of current not inactivating completely between spider and scorpion toxins is also obvious. during a 500 ms depolarization. Consequently, the Exhibiting limited sequence identity with each transition should be reduced indirectly from open- other, they represent different three-dimensional inactivated state to closed-inactivated states accord- structural motifs. The secondary structure elements ing to the channel gating scheme in Fig. 7C. For the from 1H NMR spectra and disulfide bridge pattern repriming kinetics of channel recovery was in- indicate that like Huwentoxin-IV and Hainantoxin- creased with the T1/2 decreased from 2.3 to 1.1 ms, IV, JZTX-I should be folded into an ICK motif in it clearly indicates that VGSCs treated by toxin solution (Zeng et al., 2005). So JZTX-I perhaps can undergo the transition to closed states at a much contribute to further disclosing the structure–func- faster rate. Thus, similar to d-atracotoxins, JZTX-I tion relationship of VGSCs. might cause a more rapid transition from open state In conclusion, in this study we profoundly directly to closed states of VGSCs. In our previous characterized the excitatory responses of JZTX-I work, JZTX-I exhibited no evident effect on steady- from Chinese tarantula C. jingzhao and the under- state inactivation of TTX-S VGSCs based on the lying mechanism on VGSCs. The toxin significantly voltage midpoint value (Xiao et al., 2005), but a increased muscle tension in nerve–muscle prepara- more elaborated work here demonstrated that a tions, especially on nerve-unmediated cardiac and slight reduce with statistical significance was de- skeletal muscles, where no spider toxins are reported tected on the transition from closed states to closed- to induce similar responses so far. Muscle contrac- inactivated states at higher conditioning pulse tion might be potentiated by JZTX-I facilitating the (50 mV). The result is consistent with that in Fig. neurotransmitter release into synaptic gap. Also, 6C of Xiao et al. (2005), where larger current JZTX-I treatment can bias the activities of sodium amplitudes compared to control was also observed channels towards open state to enhance the excit- at conditioning pulse from 50 to 30 mV. It has ability of neuronal, muscular and cardiac mem- been proposed that ramp currents, generated by branes. slow closed-state inactivation, can modulate the excitability of tissues (Cummins et al., 1998). Acknowledgments Indeed, JZTX-I treatment augments the ramp currents significantly and decreases the amount of The work was supported by grants from National inactivation in a use-dependant manner in our Natural Science Foundation of China under Con- study. The result may partially explain the increas- tract no. 30500146 and National 973 Project of ing response to muscle twitches on nerve–muscle China under Contract no. 2006CB708508. preparations. Similar to scorpion a-toxin LqTx, JZTX-I binding to VGSCs is also voltage dependent, but the References JZTX-I-channel complex is more stable than the LqTx-channel complex. Rogers et al. (1996) sug- Catterall, W.A., 2000. From ionic currents to molecular gested that voltage-dependent dissociation for mechanisms: the structure and function of voltage-gated scorpion a-toxins is likely to be driven by the sodium channels. Neuron 26 (1), 13–25. voltage-dependent conformational change to the Catterall, W.A., Striessnig, J., Snutch, T.P., Perez-Reyes, E., low affinity state. After Glu1613 was mutated to be 2003. International Union of Pharmacology. International Union of Pharmacology. XL. Compendium of voltage-gated Arg or His for Nav1.2, the dissociation of LqTx- ion channels: calcium channels. Pharmacol. Rev. 55 (4), channel became faster and the channel mutant 579–581. blocked toxin binding. Our previous work demon- Cohen, L., Gilles, N., Karbat, I., Ilan, N., Gordon, D., Gurevitz, strated that JZTX-I inhibited channel inactivation M., 2006. Direct evidence that receptor site-4 of sodium of both neuronal TTX-S and cardiac TTX-R channel gating modifiers is not dipped in the phospholipid bilayer of neuronal membranes. J. Biol. Chem. 281 (30), subtypes, but its affinity for the latter is about 20673–20679. 30-fold higher than for the former. The residue Cummins, T.R., Howe, J.R., Waxman, S.G., 1998. Slow closed- Glu1613 was assumed to play an important role state inactivation: a novel mechanism underlying ramp ARTICLE IN PRESS Y. Xiao et al. / Toxicon 50 (2007) 507–517 517

currents in cells expressing the hNE/PN1 sodium channel. Rash, L.D., Birinyi-Strachan, L.C., Nicholson, G.M., Hodgson, J. Neurosci. 18 (23), 9607–9619. W.C., 2000. Neurotoxic activity of venom from the Austra- Cummins, T.R., Aglieco, F., Renganathan, M., Herzog, R.I., lian eastern mouse spider (Missulena bradleyi) involves Dib-Hajj, S.D., Waxman, S.G., 2001. Nav1.3 sodium modulation of sodium channel gating. Br. J. Pharmacol. 130 channels: rapid repriming and slow closed-state inactivation (8), 1817–1824. display quantitative differences after expression in a mamma- Richard Benzinger, G., Tonkovich, G.S., Hanck, D.A., 1999. lian cell line and in spinal sensory neurons. J. Neurosci. 21 Augmentation of recovery from inactivation by site-3 Na (16), 5952–5961. channel toxins. A single-channel and whole-cell study of Cummins, T.R., Dib-Hajj, S.D., Waxman, S.G., 2004. Electrophy- sustained currents. J. Gen. Physiol. 113 (2), 333–346. siological properties of mutant Nav1.7 sodium channels in a Rogers, J.C., Qu, Y., Tanada, T.N., Scheuer, T., Catterall, W.A., painful inherited neuropathy. J. Neurosci. 24 (38), 8232–8236. 1996. Molecular determinants of high affinity binding of Gregio, A.M., Heleno, M.G., Von Eicksted, V.R., Fontana, alpha-scorpion toxin and sea anemone toxin in the S3–S4 M.D., 1999. The neuromuscular action of Ancylometes sp. extracellular loop in domain IV of the Na+ channel alpha spider venom in the rat phrenic nerve-diaphragm preparation. subunit. J. Biol. Chem. 271 (27), 15950–15962. Toxicon 37 (3), 545–550. Sahara, Y., Gotoh, M., Konno, K., Miwa, A., Tsubokawa, H., Gunning, S.J., Chong, Y., Khalife, A.A., Hains, P.G., Broady, Robinson, H.P., Kawai, N., 2000. A new class of neurotoxin K.W., Nicholson, G.M., 2003. Isolation of delta-missulena- from wasp venom slows inactivation of sodium current. Eur. toxin-Mb1a, the major vertebrate-active spider delta-toxin J. Neurosci. 12 (6), 1961–1970. from the venom of Missulena bradleyi (Actinopodidae). FEBS Shon, K.J., Olivera, B.M., Watkins, M., Jacobsen, R.B., Gray, Lett. 554 (1–2), 211–218. W.R., Floresca, C.Z., Cruz, L.J., Hillyard, D.R., Brink, A., Karbat, I., Frolow, F., Froy, O., Gilles, N., Cohen, L., Turkov, Terlau, H., Yoshikami, D., 1998. m-Conotoxin PIIIA, a new M., Gordon, D., Gurevitz, M., 2004. Molecular basis of the peptide for discriminating among tetrodotoxin-sensitive Na high insecticidal potency of scorpion alpha-toxins. J. Biol. channel subtypes. J. Neurosci. 18, 4473–4481. Chem. 279 (30), 31679–31686. Xiao, Y., Liang, S., 2003. Inhibition of neuronal tetrodotoxin- Lee, S.Y., MacKinnon, R., 2004. A membrane-access mechanism sensitive Na+ channels by two spider toxins: Hainantoxin-III of ion channel inhibition by voltage sensor toxins from spider and Hainantoxin-IV. Eur. J. Pharmacol. 477 (1), 1–7. venom. Nature 430 (6996), 232–235. Xiao, Y., Tang, J., Yang, Y., Wang, M., Hu, W., Xie, J., Zeng, Liang, S.P., Chen, X.D., Shu, Q., Zhang, Y., Peng, K., 2000. The X., Liang, S., 2004. Jingzhaotoxin-III, a novel spider toxin presynaptic activity of Huwentoxin-I, a neurotoxin from the inhibiting activation of voltage-gated sodium channel in rat venom of the Chinese bird spider Selenocosmia huwena. cardiac myocytes. J. Biol. Chem. 279 (25), 22220–22626. Toxicon 38 (9), 1237–1246. Xiao, Y., Tang, J., Hu, W., Xie, J., Maertens, C., Tytgat, J., Meisler, M.H., Kearney, J.A., 2005. Sodium channel mutations Liang, S., 2005. Jingzhaotoxin-I, a novel spider neurotoxin in epilepsy and other neurological disorders. J. Clin. Invest. preferentially inhibiting cardiac sodium channel inactivation. 115 (8), 2010–2017. J. Biol. Chem. 280 (13), 12069–12076. Middleton, R.E., Warren, V.A., Kraus, R.L., Hwang, J.C., Liu, Yu, F.H., Westenbroek, R.E., Silos-Santiago, I., McCormick, C.J., Dai, G., Brochu, R.M., Kohler, M.G., Gao, Y.D., K.A., Lawson, D., Ge, P., Ferriera, H., Lilly, J., DiStefano, Garsky, V.M., Bogusky, M.J., Mehl, J.T., Cohen, C.J., P.S., Catterall, W.A., Scheuer, T., Curtis, R., 2003. Sodium Smith, M.M., 2002. Two tarantula peptides inhibit activation channel b4, a new disulfide-linked auxiliary subunit with of multiple sodium channels. Biochemistry 41 (50), similarity to b2. J. Neurosci. 23 (20), 7577–7585. 14734–14747. Yuan, C., Yang, S., Liao., Z., Liang, S., 2007. Effects and Nicholson, G.M., Walsh, R., Little, M.J., Tyler, M.I., 1998. mechanism of Chinese tarantula toxins on the Kv2.1 Characterization of the effects of robustoxin, the lethal potassium channels. Biochem. Biophys. Res. Commun. 352 neurotoxin from the Sydney funnel-web spider Atrax (3), 799–804. robustus, on sodium channel activation and inactivation. Zeng, X.Z., Zhu, Q., Liang, S.P., 2005. Sequence-specific Pflu¨gers Arch. 436, 117–126. assignment of 1H-NMR resonance and determination of the Ogata, N., Ohishi, Y., 2002. Molecular diversity of structure secondary structure of Jingzhaotoxin-I. Acta. Biochim. and function of the voltage-gated Na+ Channels. Jpn. Biophys. Sin. 37 (8), 567–572. J. Pharmacol. 88, 365–377. Zhou, P.A., Xie, X.J., Li, M., Yang, D.M., Xie, Z.P., Zong, X., Ogata, N., Tatebayashi, H., 1993. Kinetic analysis of two types of Liang, S.P., 1997. Blockade of neuromuscular transmission by Na+ channels in rat dorsal root ganglia. J. Physiol. 466, 9–37. huwentoxin-I, purified from the venom of the Chinese bird Peng, K., Shu, Q., Liu, Z., Liang, S., 2002. Function and solution spider Selenocosmia huwena. Toxicon 35 (1), 39–45. structure of Huwentoxin-IV, a potent neuronal tetrodotoxin Zhu, M.S., Song, D.X., Li, T.H., 2001. A new species of (TTX)-sensitive sodium channel antagonist from Chinese bird the family Theraphosidae, with taxonomic study on the spider Selenocosmia huwena. J. Biol. Chem. 277 (49), species Selenocosmia hainana. J. Baoding Teachers College 47564–47571. 14, 1–6.