Pflügers Arch – Eur J Physiol (2000) 439:423–432 Digital Object Identifier (DOI) 10.1007/s004249900181
© Springer-Verlag 2000
O R I G I N A L A RT I C L E
Haijun Chen · Dalia Gordon · Stefan H. Heinemann
Modulation of cloned skeletal muscle sodium channels by the scorpion toxins Lqh II, Lqh III, and LqhαIT
Received: 29 June 1999 / Received after revision: 15 September 1999 / Accepted: 28 September 1999 / Published online: 20 November 1999
Abstract The scorpion α-toxins Lqh II, Lqh III, and LqhαIT from Leiurus quinquestriatus hebraeus are representatives of typical α-toxins, specific for either mammals (Lqh II) or insects (LqhαIT), and α-like toxins (Lqh III) which act on both mammals and insects. For a comparative study of the effects of these toxins on mammalian sodium channels we stably expressed rat skeletal muscle sodium channel α subunits (µI) in HEK 293 cells and measured Na+ currents in the whole-cell patchclamp mode. The α- and α-like toxins strongly slowed down channel inactivation with a half-maximal effect at 1.4 nM (Lqh II), 5.4 nM (Lqh III), and 0.5 nM (LqhαIT). The recovery from fast inactivation was accelerated by all toxins with the potency sequence: Lqh II>LqhαIT> Lqh III. The voltage dependence of inactivation and recovery from inactivation were reduced while the threshold for activation was only slightly shifted by ≅10 mV without altering the slope factors, suggesting uncoupling of the impaired inactivation from the activation. The toxins induced an increase in peak inward current, which was accounted for by an increased maximal open-channel probability. Although all three toxins induced similar modifications of the channel properties, their kinetics of association and dissociation were very different. Between –140 and –80 mV toxin association was not voltage dependent. In 100 nM toxin the association time constants were: 1.3 s (Lqh II), 20 s (Lqh III), and 3.8 s (LqhαIT). At positive voltages the toxin dissociated from the channel; at +100 mV the dissociation time constants were 30, 321, and 135 ms, respectively. In contrast to the association, dissociation was voltage dependent with a similar slope of about 12 mV per e-fold change H. Chen · S.H. Heinemann (✉) Research Unit Molecular and Cellular Biophysics, Medical Faculty of the Friedrich Schiller University Jena, Drackendorfer Str. 1, D-07747 Jena, Germany e-mail:
[email protected] Tel.: +49-3641-304540, Fax: +49-3641-304542 D. Gordon CEA, Département d’Ingéniere et d’Edtudes des Protéines, C.E. Saclay, F–91191 Gif-sur-Yvette, France
for all three toxins. The strong differences in the association and dissociation kinetics of these toxins may identify them as members of different scorpion α-toxin subgroups. Key words Inactivation · Patch clamp · Receptor site 3 · Scorpion toxin · Skeletal muscle · Sodium channel Abbreviations ApA, ApB: Antopleurin A/B, a-toxins from the sea anemone Anthopleura xanthogrammica · ATX II: toxin II from the sea anemone Anemonia sulcata · AaH II: a-toxin from the venom of the scorpion Androctonus australis Hector · LqTx: a-toxin from the venom of the scorpion Leiurus quinquestriatus quinquestriatus, also called Lqq V · Lqh II, Lqh III, and LqhaIT: a-toxins from the venom of the scorpion Leiurus quinquestriatus hebraeus
Introduction Voltage-gated sodium channels play a major role in rapid electrical signaling in excitable cells. These channels are modulated by a variety of toxins, which bind to various sites at the channel polypeptide [5, 12]. The α-toxins from scorpions are thought to bind to receptor site 3 [5], which is formed by amino acid residues in the extracellular linker between segments S3 and S4 in the fourth homologous domain (D4) of the channel polypeptide [21]. The extracellular regions between segments S5 and S6 in domains D1 and D4 are also thought to contribute to receptor site 3 [26]. The major functional effect of scorpion α-toxins on voltage-gated sodium channels is a marked slowing of fast inactivation. The binding of scorpion α-toxins to mammalian sodium channels has been shown to be voltage dependent in a sense that the affinity decreases with membrane depolarization. Hence, receptor site 3 was suggested to undergo a conformational change during depolarization, leading to decreased affinity to the toxin [4, 5]. Besides the functional importance of scorpion α-toxins for toxicology and membrane phys-
424
iology, they are therefore interesting tools for studying structure–function relationships of voltage-gated sodium channels. One of the intriguing aspects is the molecular mechanism underlying the α-toxin effect, namely the removal of fast inactivation. While the toxins bind to site 3, which is located at the extracellular side, fast inactivation is mainly determined by the hinged-lid mechanism involving the movement of the intracellular linker between domains D3 and D4 to block the channel from the inside [28]. Thus, one of the open questions is how the extracellular binding of α-toxins affects rapid channel inactivation, mediated by intracellular protein domains. A putative coupling element between these two mechanisms might be the voltage-sensing S4 segment of domain D4, which may also provide indirect voltage dependence to the process of channel inactivation [9]. However, the details of this complex interaction between scorpion α-toxins and channel structures being responsible for activation and inactivation remains elusive. Different scorpion α-toxins have been shown to be active on sodium channels in various excitable cells (for reviews see [14, 19]). Recently, the class of scorpion αtoxins has been subdivided into three major groups according to their activities and properties of binding to voltage-gated sodium channels in mammals and insects [14]. The classic α-toxins are highly active in mammals (so-called α-mammal toxins, such as AaH II and Lqh II [19, 23]), α-toxins, highly active in insects (e.g., LqhαIT [10]), form the second group, and the third group comprises the so-called α-like toxins (e.g., Lqh III, Bom III, and Bom IV [7, 13, 23]). The latter are highly active in both insects and mammals and do not compete with αmammal toxin binding in rat brain synaptosomes. Lqh II, LqhαIT, and Lqh III are toxins from the venom of the scorpion Leiurus quinquestriatus hebraeus and they are representatives of these three α-toxin groups. Despite the overall similarity in structure, several turn regions involved in the functional binding surface of LqhαIT, the α-toxin highly active on insects [29], reveal significant differences in the α-like Lqh III structure [17]. The electrostatic charge distribution in the three αtoxins is also surprisingly different [17]. No molecular information is presently available to explain the selectivity of these α-toxins for different sodium channels in mammals and insects. These toxins have been shown to bind to sodium channels in rat brain (Lqh II [18]), rat skeletal muscle (LqhαIT) and cockroach axon (LqhαIT, Lqh III [10]). However, there is no comparative study of the effects of Lqh II, Lqh III, and LqhαIT on the function of sodium channels. Thus, it is not clear whether these toxins, which belong to different α-toxin groups and are expected to bind to receptor site 3, induce similar or different functional modulations of a single cloned sodium channel subtype. The effect of some of the toxins that bind to receptor site 3, like the sea anemone toxins ATX II, ApA, and ApB, have been assayed for cloned sodium channels such as rat brain II (RbII) [21], rat skeletal muscle I (µI),
human cardiac I (hH1) [1, 2, 9], and Drosophila para sodium channels [27] using electrophysiological methods. Interestingly, all toxins that compete for scorpion α-toxin binding to receptor site 3 on sodium channels reveal similar effects of inhibition of inactivation [5, 6, 12, 13, 17, 21, 27], despite the marked differences in their length and structure [14, 20]. All scorpion α-toxins are highly toxic to mammals and, surprisingly, the α-toxin highly active on insects, LqhαIT, is only about four times less toxic to mice than the α-mammal toxin Lqh II [13, 23]. Previous reports showed that high concentrations of LqhαIT prolong the action potential duration of rat skeletal muscle fibers by slowing the inactivation of sodium channels [10]. Since one of the main targets of scorpion toxins in the body may be the skeletal muscle, we attempted to examine the interaction and mode of action of these three different scorpion α-toxins with the cloned rat skeletal muscle sodium channel µI. It is important to know how toxins slow inactivation if scorpion α-toxins are to be applied in future as biopesticides and pharmacological agents in medicine. The first step is to further characterize details of their action, binding sites and interaction kinetics with sodium channels. In this study, we expressed µI sodium channels in HEK 293 cells, and then examined and compared the functional impacts of Lqh II, Lqh III, and LqhαIT on these channels using patch-clamp methods.
Materials and methods Cell culture and transfection HEK 293 cells (CAMR, Porton Down, Salisbury, UK) and HEK 293 µI cells, stably expressing µI sodium channel α subunits, were maintained in Dulbecco’s Minimal Eagles Medium (DMEM), supplemented with 10% fetal calf serum in a 5% CO2 incubator. For stable transfection of HEK 293 cells with µI sodium channel α subunits, the plasmid DNA NaµI-pcDNA3 was linearized with PvuI. The cells were grown to 30–50% confluence on 35-mm plates in culture medium. Transfection with linearized NaµIpcDNA3 DNA was then performed with the standard calcium phosphate procedure [22]. The expression level was assayed electrophysiologically 48–72 h after transfection. About 24 h after transfection cells were cultured in a selection medium [DMEM with 10% FCS and 400 µg/ml of the aminoglycoside antibiotic Geneticin (G418, Invitrogen, Leek, The Netherlands)]. The stable cell line (HEK 293_µI) was isolated after keeping the cells for 3 weeks in selection culture. Electrophysiological measurements Whole-cell voltage-clamp experiments were performed as described previously [15]. Patch pipettes were fabricated from Kimax-51 glass (WPI, Sarasota, Fla., USA) and were coated with RTV615 (Paul Hellermann, Pinneberg, Germany) to reduce their capacitance. After fire polishing pipette resistances of 0.9–2.0 MΩ were obtained. The series resistance was compensated for by more than 80% in order to minimize voltage errors. As patch-clamp amplifier we used an EPC9, operated by Pulse+PulseFit software (both HEKA Elektronik, Lambrecht, Germany). Leak and capacitive currents were corrected with a p/n method. In most cases currents were low-pass filtered at 5 kHz and sampled at a rate of
425 25 kHz. For nonstationary noise analysis data were sampled at a rate of 50 kHz. All experiments were performed at room temperature, 19–23°C. The patch pipettes contained (mM): 35 NaCl; 105 CsF; 10 EGTA; 10 HEPES (pH 7.4 with CsOH). The bath solution contained (mM): 150 NaCl; 2 KCl; 1.5 CaCl2; 1 MgCl2; 5 glucose; 10 HEPES (pH 7.4 with NaOH). Toxins were purified according to Sautière et al. [23] and were a generous gift from Dr. Pierre Sautière, Laboratoire de Chimie des Biomolecules, Institut Pasteur de Lille, Lille, France. Toxins were dissolved in the bath solution without glucose, supplemented with 1 mg/ml bovine serum albumin (BSA) in order to prevent adherence of toxins to the vials and the perfusion apparatus. Aliquots of the toxin samples were frozen at the final concentration in order to minimize freezing and thawing cycles. Toxin was applied using an application pipette of about 10 µm opening diameter. The collection of control data was initiated 15 min after establishing the whole-cell voltage-clamp configuration. The time course with which the toxin effect was established was rather slow. Therefore, measurements of toxin effects were started between 5 and 30 min after toxin application, depending on the toxin concentration. Application of 1 mg/ml BSA alone did not alter sodium channel function. Data analysis Data analysis was performed using PulseFit, PulseTools (HEKA) and IgorPro (WaveMetrics, Lake Oswego, Ore., USA) running on Macintosh computers. All data are presented as mean ±standard error of the mean (n=number of independent experiments). Current–voltage relationships From a holding potential of –120 mV test depolarizations in the range of from –80 to +70 mV were applied at 6-s intervals. The peak currents were measured and plotted as a function of test voltage. The control data were then fitted with an activation curve assuming three independent gating units and a single-channel conductance obeying the Goldman–Hodgkin–Katz equation: −(V − Erev ) / 25mV 1 (1) I (V ) = ΓV 1 − e −V / 25mV 1− e (1 + e −(V −Vm ) / km )3 where Γ is the conductance and Erev the reversal potential. The last term describes the open probability of the channels, characterized by the voltage of half-maximal activation per subunit, Vm, and the slope factor, km. The data points, both for control and the corresponding data after toxin application, were then normalized to the peak inward current predicted by the data fit (see Fig. 2, left).
Steady-state inactivation Conditioning pulses of 500 ms duration ranging between –120 and –20 mV were applied before available channels were assayed with a short depolarization to –20 mV. The resulting inward currents were normalized to the maximal current (see Fig. 2, right). The voltage dependence was described with a Boltzmann function of first order, characterized by the half-inactivation voltage Vh, the corresponding slope factor kh, and a non-inactivating component I ∞: 1 − I∞ (2) Inorm (V ) = I∞ + 1 + e(V −Vh ) / kh Inactivation The degree of inactivation was assayed by measuring the peak current as well as the mean current level between 4.5 and 5 ms after the start of the depolarization. The ratio I5ms/Ipeak gives an estimate of how many channels inactivate after opening; a value of
zero represents complete inactivation in 4.5 ms; a value of one, no inactivation (see Fig. 3). Alternatively, inactivation was quantified by fitting single- or double-exponential functions to the decaying section of the Na+ currents (see Fig. 4A). Dose–response curves The dose dependence for toxin-induced removal of inactivation was measured by plotting the relative degree of inactivation after 4.5–5 ms at 0 mV (I5ms/Ipeak) as a function of toxin concentration. The concentration dependence was described with the Hill equation (see Fig. 3D): I5ms 1 (0mV ) = h I peak 1 + ( EC50 /[toxin])
(3)
where h is the Hill coefficient, [toxin] the toxin concentration, and EC50 provides a measure for the concentration of half-maximal inactivation inhibition. Noise analysis Fluctuations in the macroscopic current signal provide information on the unitary current size (i) and the number of channels (NC) [24]. We performed nonstationary noise analysis of each 500 consecutive current recordings (depolarization to 0 mV) compiling the mean current (I) and the current variance (σ2), based on the differences of consecutive current traces, using an algorithm described by Heinemann and Conti [16]. The binned variance was then plotted as a function of the mean current and a nonlinear mean–variance fit [25] was performed according to the following equation:
σ 2 = σ b2 + iI − I 2 / NC
(4)
where σb2 is the background variance. The maximal open-channel probability (popen) was estimated by analyzing the peak macroscopic current using the parameters of Eq. 4.
Results Lqh II, Lqh III, and LqhαIT modify activation and inactivation of µI sodium channels Families of Na+ currents from whole-cell recordings of HEK 293_µI cells in the absence and presence of 5 nM Lqh II, Lqh III, and LqhαIT are shown in Fig. 1. While under control conditions µI channels inactivate rapidly, all three toxins strongly slowed down the inactivation indicating that these toxins are very active on µI sodium channels. In addition, Lqh II, Lqh III, and LqhαIT induced persistent currents during long depolarizing potentials (data not shown). Both effects are similar to the effects exerted by the sea anemone toxin ATX II on µI sodium channels [8]. Figure 2 shows the current–voltage (I–V) relationships in the absence and presence of indicated concentrations of Lqh II, Lqh III, and LqhαIT. I–V curves were fit with a Hodgkin–Huxley formalism (Eq. 1), yielding the reversal potential (Erev), the half-activation voltage (Vm) and the slope factors for the voltage dependence of activation (km). As can be seen in the left panels of Fig. 2, the reversal potential did not change upon toxin application. The other parameters are listed in Table 1, indicat-
426 Table 1 Parameters for the voltage dependence of activation and inactivation. The voltage dependence of activation and inactivation was analyzed from experiments as shown in Fig. 2 in the absence (control) and presence of 100 nM of the indicated toxins. Vm
and km are the parameters for activation (Eq. 1) and Vh and kh characterize inactivation (Eq. 2). The statistical significance of a deviation from the control parameters, obtained from the corresponding experiments, was determined with a two-tailed t-test
Activation
Control Lqh II Lqh III LqhαIT *PLqhαIT> Lqh III. In addition, compared with the control conditions, the toxins reduced the voltage dependence of the recovery rates by about a factor of two. Lqh II, Lqh III, and LqhαIT increased the peak currents through µI sodium channels (Figs. 1, 2, 3); the maximal peak inward currents obtained in I–V families were increased up to about 1.5 times (Fig. 2). This effect can result either from an increased channel open probability or from an increased single-channel conductance. Both parameters were measured at 0 mV with nonstationary noise analysis. Estimates of the mean current (INa) and variance (σ2) from 500 consecutive traces are shown in Fig. 6A, B for a representative cell. The variance-current analysis (using Eq. 4, Fig. 6C) indicates that the single-channel conductance did not change, while the channel open-probability increased. Similar experiments were performed for 100 nM Lqh III and LqhαIT. Single-channel currents before and after toxin application were – Lqh II: 1.36 and 1.23 pA (n=1); Lqh III: 1.32±0.05 and 1.35±0.10 pA (n=5); LqhαIT: 1.19±0.02 and 1.09±0.03 pA (n=5). Channel open-prob-
429
Fig. 6A–C Nonstationary noise analysis. A Mean currents recorded in whole-cell voltage-clamp mode in response to test pulses to 0 mV (average of 500 traces) in the absence and presence of 100 nM Lqh II. B Mean ensemble variance based on the differences of successive current traces. C Variance as a function of mean current before (open circles) and after application of 100 nM Lqh II (filled circles). The continuous curves represent data fits according to Eq. 4 yielding a single-channel current of 1.36 and 1.23 pA and a maximal open-channel probability of 0.51 and 0.76 for the control and in the presence of 100 nM Lqh II, respectively. The estimated number of channels was 2800. For mean values and results for the other toxins see text
abilities were – Lqh II: 50.9 and 76.4% (n=1); Lqh III: 50.2±2.7 and 67.2±3.2% (n=5); LqhαIT: 54.9±3.7 and 69.3±1.6% (n=5). Kinetics of toxin association and dissociation Application and removal of Lqh II, Lqh III, and LqhαIT during current recording from µI channels revealed that these toxins have different time courses of binding to and dissociation from these channels. Therefore, we attempted to investigate the association and dissociation kinetics of these toxins. Previous work has shown that scorpion α-toxin binding is inhibited by depolarization due to the acceleration of toxin dissociation [21]. Therefore, both association and dissociation kinetics of toxin binding were measured in the presence of 100 nM toxin. Dissociation rates were determined in double-pulse protocols (Fig. 7A) at positive potentials. Conditioning pulses (20–140 mV) of increasing duration were applied, returning to –120 mV for 20 ms to recover channels from fast inactivation. Na+ current was then elicited with a 30-ms test pulse to –20 mV. The extent of removal of fast inactivation was assayed with the ratio I5ms/Ipeak, plotted as a function of the duration of the conditioning pulse (Fig. 7B). Single-exponential fits revealed the time constants, τoff, for various conditioning voltages. At the same conditioning voltage the three toxins exhibited quite different dissociation time constants with the order τoff(Lqh II)
