ABSTRACT Tityustoxin-Ka (TSTX-Ka) is a 37 amino acid peptide component of the venom of the scorpion, Tityus serrulatus, and blocks voltage-gated K current (1) in highly specific channel types. TsTX-Ka is structurally similar to charybdotoxin, agitoxin-2, and other scorpion toxins that bind to the extracellular surface of the pore-forming region of Kvi-type (Shaker) channels. The action of TsTX-Ka was studied on squid Kt channels using the whole-cell patch clamp technique with dissociated giant fiber lobe (GFL) neurons and with insect Sf9 cells infected with a baculovirus carrying the CDNA for the corresponding channel (Brock et al., 1996). Under standard conditions (12°0, pH 7.6, 20 mM external K*), TsTX-Ka reduces l for a strong voltage step (+40 mV) in GFL neurons with an estimated Kp of 50 nM. Steady state voltage dependence and activation kinetics of residual ly are minimally affected in 300 nM TSTX-Ka, but inactivation kinetics are accelerated. Similar results were obtained in SF9 cells. Block of I, in GFL neurons by TSTX-ka (300 nM) depends on external pH. Less block occurs at low pH, with half-block occurring at pH 6.0, approximately the pK, of histidine (6.02). This amino acid is located in the SgkviA channel (H351) at a position homologous to F425 in Shaker B, a residue critical to binding of charybdotoxin (Goldstein, et al., 1994) and agitoxin-2 (Gross & Mackinnon, 1996). These results suggest that H351 in the external mouth of the squid SqkviA channel is also a determinant of TsTX-ka binding, and that protonation of the side chain of this residue may decrease toxin sensitivity by electrostatic repulsion. INTRODUCTION Kt channels are remarkably diverse. With more channel types than any other ion channel studied, the literature has quickly grown confusing with competing nomenclature and classification schemes (Rudy, 1988). The ability to separate Kt currents on the basis of pharmacology, kinetics genetics, and other methods has indicated the presence of many different K channel types, often within a single cell. The process of identifying and characterizing individual channel types has been aided by a molecular biological approach. A single K“ channel gene can be expressed in a number of model expression systems which lack background Kt channels in order to record currents produced by a single defined channel type. However, it remains difficult, if not impossible, to isolate macroscopic currents of a single K“ channel type from a cell exhibiting many different K“ currents without more specific pharmacological probes than are presently available Certain toxins, however, may hold the key to this need for specificity. Small peptide toxins from the venom of snakes, spiders, and scorpions have been found to produce block at nanomolar concentrations of highly specific K“ channel subtypes (both voltage-gated and Ca2t-activated) (Dreyer, 1990; Strong, 1990). Many scorpion toxins, in particular, have been found to exhibit substantial structural similarity with one another. In each of the toxin sequences shown in Figure 1, there are 6 cysteine residues located at highly conserved positions (Garcia et al., 1994), suggesting that each peptide may share the same arrangement of disulfide bonds and, therefore similar tertiary structures. Based on similarities of structure, these toxins may also share a common mechanism of block. Several scorpion toxins have been shown to bind to the extracellular surface of the ion conduction pore and physically occlude ion flow like a cork in a bottle (Anderson et al., 1988: Gross et al., 1994; Miller, 1988; Mackinnon and Miller, 1988, 1989: Giangiacomo et al., 1992). Specific residues in a-subunits of Kvl (Shaker- like) channels that are important for toxin binding have been localized to the S5-S6 pore-forming region and, in particular, to those sites flanking the outer mouth of the pore (Hartmann et al., 1991; Heginbotham et al., 1994; Gross et al., 1994; Mackinnon and Yellen, 1990; Yellen et al., 1991; Yool and Schwartz, 1991). Thus far, no scorpion toxin studies have been reported on the delayed rectifier K“ channel of the squid giant axon or in the parent cell bodies in the giant fiber lobe (GFL) of the stellate ganglion. One of the goals of this study was to identify which scorpion toxins exhibit potent block of K* currents in squid GFL neurons. Öf those tested, only tityustoxin-Ka (TSTX-Ka) exhibited significant block of GFL Ig. TSTX-Ka is a 37 amino acid peptide from the venom of the Brazilian scorpion, Tityus serrulatus, previously shown to block voltage-gated K“ current in the cloned Kv1.2 channel (Werkman et al., 1993). Its effects were also tested in Sf9 insect cells expressing only the SgkviA gene CDNA, recently identified and proposed to underlie the main component of delayed rectifier K* current in GFL neurons and giant axons (Rosenthal et al, 1996, in press). MATERIALS AND METHODS Giant Fiber Lobe Neurons. Cell bodies were dissociated from the posterior tip of the GFL-stellate ganglion of Loligo opalescens and maintained in primary culture at 16°0 in L¬ 15 medium containing supplemental salts to match the composition of sea water (Gilly et al., 1990). Cells were normally used within one week of isolation. Production of Recombinant Baculovirus. A recombinant baculovirus carrying the SgkV1A CDNA was created by Brock et al. (1996). Briefly, the coding region of SgkviA, spliced to the 5 and 3' untranslated regions of the Xenopus B-globin gene, was subcloned into the baculovirus expression vector pVL1392 (Stratagene, La Jolla, CA). Recombinant baculovirus was isolated and purified by standard procedures (Summers and Smith, 1987). Sf9 cells were cotransfected with the SgkviA expression vector and linearized, wild-type AcMNPV DNA. Occlusion¬ negative plaques were purified and amplified by two rounds of plaque purification. Viral stocks were prepared from final supernatants, titered, and stored at 4° Maintenance and Infection of Sf9 Cells. Sf9 cells were maintained as suspension cultures in complete TNM:FH media (TNM:FH media + 10% fetal bovine serum, 10 mgm streptomycin, 100 units/ml penicillin, and 5 mg/ml fungizone) at 27°0 (Summers and Smith, 1987). For infections, log-phase cells were seeded in culture dishes at 10' cells/cm2 and allowed to settle for 30-60 minutes. The media was then removed, and fresh media was added. Infected cells were maintained at 18°0 Electrophysiology. Recordings from Sf9 cells at 18°0 and GFL neurons at 12°0 were made using standard whole-cell patch clamp techniques. All recordings were made from cells at a holding potential of -80 mV, with linear capacity and ionic currents subtracted online using a standard P/-4 technique. Recordings were sampled at 2 kHz or 20 kHz and filtered at 10 kHz. The recording solutions had the following compositions (in mM): 150K Sf9 Intemal, 20 KCI, 50 KF, 80 K-Glutamate, 10 Lysine-Hepes, 1 EGTA, 1 EDTA, 26.2 Glycine, 85 Sucrose, 4 Mg-ATP, 4.12 TMA-OH, pH 7.0; 20K Sf9 external, 20 KCI, 168 Nacl, 10 Cacl,, 10 MgCl, 10 MgSO,, 5 Hepes, 1.6 NaÖH, pH 7.2; 150K GFL intemal, 20 KCI, 50 KF, 80 K-Glutamate, 10 Lysine-Hepes, 1 EGTA, 1 EDTA, 381 Glycine, 291 Sucrose, 4 Mg-ATP, 5 TMA-OH, pH 7.0; 20K GEL exteral, 20 KCI, 480 Nacl, 10 Cacl,, 10 MgCl», 10 MgSO,, 5 Hepes, pH 7.66. In addition, 200 nM tetrodotoxin was used to block Na’ current, and in some cells 0.8 uM CdCI, was used to block Ca“ current. Scorpion Toxins. Tityustoxin-Ka, a-dendrotoxin, and pandinustoxin were kindly provided by Dr. M. P. Blaustein and D. R. Matteson, Dept. of Physiology, Univ. of Maryland. Stock solutions of toxin (50-250 uM) in H,O were stored in plastic microtubes at 4°0. Toxins were diluted in external solution and slowly perfused (1-2 min for complete solution change) into the recording chamber after stable control currents were recorded. RESULTS Sensitivity of the SgKvi A Channel. The interaction of scorpion toxins with ion channel receptors is highly specific to ion channel subtypes (Adams and Olivera, 1994; Eccles et al, 1994; Rogowski et al., 1994). Several scorpion toxins were examined for their ability to block K“ current in GFL neurons. 250 nM pandinustoxin and 50 nM a-dendrotoxin, known to be potent blockers of voltage-gated K' current in other receptor types, did not significantly affect K+ current («10% change in peak l amplitude at +40 mV) in GFL cells. Tityustoxin-Ka, however, did appreciably block GFL Ig in the nanomolar concentration range. Tityustoxin-Ka causes potent block of lg in GFL neurons. When applied to GFL neurons, 300 nM TSTX-Ka blocked 86% of peak GFL K“ current activated by a strong depolarization to +40mV (Figure 2). A Kp was estimated using the equation: Ko fractional block = 300 nM + Kp Requiring such a relatively high concentration, GFL Ig is considerably less sensitive to TsTX-Ka than is lg in other cell types. For example, in cloned Kv1.2 channels expressed in B82 fibroblast cells, half-block occurs at 0.21 nM, 2.4 orders of magnitude less than in SgkviA (Werkman et al, 1993). The time course of toxin effect appeared to be very slow, with half¬ block occurring after 3-4 minutes and maximum block after 33 minutes. A single exponential (t = 8.9 min) fit these data reasonably well. Reversal was slow (220 min) and incomplete, but these effects were not studied in detail. A similar slowly developing effect of toxin at this concentration also occurs with lg in perfused squid giant axons (Gilly, personal communication). TSTX-Ka did not significantly affect activation properties (Figure 3). Residual lg in 300 nM TSTX-Ka could be scaled such that the time course of activation closely resembled that in control recordings (Figure 3A). Deactivation kinetics, measured by fitting a single exponential to the tail current evoked after a prepulse to +40mV, were not substantially altered by TsTX-Ka (Figure 3B). Similarly, the voltage dependence of l activation was largely unaffected (Figure 30). TsTX-Ka did, however, tend to make inactivation faster and more complete, particularly with increasing toxin exposure time. Ig records taken during the developing block by TsTX-Ka are scaled to the same peak in Figure 4 in order to permit comparison of inactivation kinetics. Single exponential fits of the inactivating phase resulted in the indicated time constant (t) values. In addition, inactivation of GFL K+ channels was substantially more complete in TSTX-Ka. Some of this apparent increase in the degree of completion of inactivation is probably due to small, residual Ca“ current, even though the experiment in Figure 4 was carried out in the presence of 0.8 mM CdCl, in the external solution, which blocks most of the Ca2t current. TsTX-Ka block in Sf9 cells is similar to that in GFL neurons. Although maximum block in Sf9 cells expressing only SgkviA channel CDNA is less than in GFL (45% block vs. 86% in GFL), the kinetics of TsTX-Ka block is similar, with half block occurring after =2-3 minutes (20 K+ ext., pH 7.2, 18°0) (Figure 5A). Similarly, inactivation was found to be faster in TsTX-Ka (t= 45 ms vs. 58 ms in control) (Figure 5B), but this effect was not as strong as in GFL neurons. Degree (i.e. completeness) of inactivation is not much affected by TSTX-Ka. Since Sf9 cells express only the SgkviA channel type, the similarity to TSTX-Ka block of GFL L suggests that its effects are specific to a single channel type in squid GFL neurons. Tityustoxin-Ka block of GFL Ig is pH-dependent. 300 nM TSTX-Ka blocked peak Kt current in GFL neurons in a pH¬ dependent manner with 90% block at pH 9.6, 25% block at pH 5.6, and half block at pH 6.0 (Figure 6). Fractional block values vs. pH were sigmoidal, similar to a 1:1 binding curve centered at pH 6.0 and scaled to 0.86 in order to reflect maximum block achieved at this dose. Control experiments were also performed to assess the effect of pH itself in this range on ly evoked by test pulses to +40 mV (Figure 7). pH effects in the absence of toxin were minor and did not greatly alter the peak current (Figure 7A) or activation and inactivation rates (Figure 7B). DISCUSSION 300 nM Tityustoxin-Ka, a concentration z6 times greater than its estimated Kp of 50 nM in GFL neurons, was found to block Kt current in both GFL neurons and in Sf9 cells expressing the SgkviA channel, presumably the major component of K“ current in GFL cells (Rosenthal et al., 1996, in press). However, this concentration is 2.4 orders of magnitude greater than that needed to block cloned Kv1.2 channels in B82 fibroblast cells (Werkman et al., 1993). This difference in sensitivity is typical of the exquisite specificity of scorpion toxins for their receptors. The pH dependence of TsTX-Ka block of GFL 1g indicates a pH sensitivity of those residues in the toxin or in the Kt channel closely involved in binding. Histidine is the only amino acid capable of buffering in the 5.0 to 8.0 pH range (Lehninger, 1970). The amino acid sequence of TSTX-Ka (Figure 1) contains no histidine residues, therefore the presence of a histidine residue in a position in the Kt channel important for toxin binding might mediate pH dependence of TsTX-Ka block in GFL neurons. Mutagenesis studies have identified specific residues in the Shaker B K“ channel critically important for binding charybdotoxin, a scorpion toxin similar to TSTX-Ka (Figure 1) (Goldstein and Miller, 1992; Mackinnon and Miller, 1989; Mackinnon et al., 1990). One of these residues (F425), when mutated to glycine, led to a 2000-fold increase in toxin affinity, indicating this residue's importance in mediating toxin binding (Goldstein et al., 1994). The amino acid sequence of the Shaker Kt channel is similar to that of SqkviA in the pore-forming region, except at 3 positions (Rosenthal et al., 1996, in press). One of these is histidine (H351 in SgkviA) instead of the F425 of Shaker. It is very plausible that this residue mediates pH dependence of TsTX-Ka block by the degree to which the amino side chains are protonated according to the model: a-subunit - His + Ht a-subunit - His (inaccessible to toxin binding) The sigmoidal curve in Figure 6 is practically identical to the titration curve for histidine with a pK, of 6.02 (Lehninger, 1970). Therefore, as the proportion of protonated histidine residues increases with decreasing pH, the amount of block decreases by the same amount. Interactions of charged residues with toxin binding at position F425 in Shaker have been previously identified. Gross and Mackinnon (1996) have shown that a single molecule of the peptide agitoxin-2, when bound to a single a-subunit of the Shaker K+ channel, is sufficient to completely block the channel. If this is also true in the case in the TsTX-Ka and the SgkviA channel in GFL neurons, protonation of 1, 2, or 3 of the 4 a-subunit H351 residues in a given channel would not necessarily prevent block, since at least 1 subunit would remain available for toxin binding. It would instead reduce the probability of toxin binding by reducing the number of available site Although other scorpion toxins are characterized as having simple mechanisms of block (Goldstein et al., 1994), TSTX-Ka exhibits properties of a more complicated blocking mechanism. TsTX-Ka block in both GFL and Sf9 preparations is accompanied by faster and more complete inactivation of the remaining K“ current. Whether TsTX-Ka is selectively sparing channels with this type of inactivation or whether, allosteric effects act to change inactivation properties is unclear. Inactivation properties are extremely sensitive to the species of amino acid at position 449 in Shaker (threonine) which in SgkviA is a serine. T449 in Shaker, like F425, is critical to charybdotoxin and agitoxin-2 binding. Another interesting puzzle is the slow time course of TsTX-Ka block. Further studies are clearly needed to confirm and further analyze this effect. Also needed is a study using Shaker Kt channels which lack the histidine residue at position 425 to confirm that TsTX-Ka block of this channel, if it occurs at all, is not pH dependent. LITERATURE CITED Adams M. E. and Olivera B. M. (1994) Neurotoxins: overview of an emerging research technology. TINS 17: 151-155. Anderson C. S., Mackinnon R., Smith C., and Miller C. (1988) Charybdotoxin inhibition of Ca2t-activated Kt channels: effects of channel gating, voltage, and ionic strength. J. Gen. Physiol. 91: 317. 333. Brock M. W., Detomaso A. W., Rosenthal J. J. C., Lebaric Z. N., and Gilly W. F. (1996). Temperature-dependent functional expression of squid Kv1.1 channels in an insenct cell line. Biophys. J. 70: A187. Dreyer F Peptide toxins and potassium channels (1990). Rev. Physiol. Biochem. Pharmacol. 115: 93-136. Eccles C. U., Rogowski R. S., Alger B. E., and Blaustein M. P. (1994). Tityustoxin-Ka, from scorpion venom, blocks voltage-gated, non¬ inactivating potassium current in cultured central neurons. Soc. Neurosci. Abstr. 19: 708. Garcia M.L., Garcia-Calvo M., Hidalgo P., Lee A., and Mackinnon R. (1994). Purification and characterization of three inhibitors of voltage¬ dependent K“ channels from Leiurus quinquestriatus var. hebraeus venom. Biochemistry. 33: 6834-6839. Giangiacomo K. M., Garcia M. L., and McManus O. B. (1992). Mechanism of iberiotoxin block of the large-conduction calcium-activated potassium channel from bovine aortic smooth muscle. Biochemistry 31: 6719- 6727. Gilly W. F., Lucero M. T., and Horrigan F. T. (1990). Control of the spatial distribution of sodium channels in giant fiber lobe neurons of the squid. Neuron. 5: 663-74. Goldstein S. A. N., and Miller C. (1992). A point mutation in a Shaker K channel changes its charybdotoxin binding site from low to high affinity. Biophys. J. 62: 5-7. Goldstein S. A. N., Pheasant D. J., and Miller C. (1994). The charybdotoxin receptor of a Shaker Kt channel: peptide and channel residues mediating molecular recognition. Neuron 12: 1377-1388. Gross A., Abramson T., and Mackinnon R. (1994). Transfer of the scorpion toxin receptor to an insensitive potassium channel. Neuron 13: 961¬ 966. Gross A., and Mackinnon R. (1996). Agitoxin footprinting the Shaker potassium channel pore. Neuron 16: 399-406. Hartmann H.A., Kirsch G.E., Drewe J.A., Taglialatela M., Joho R.H., and Brown A.M. (1991). Exchange of conduction pathways between two related K“ channels. Science 251: 942-944. Heginbotham L., Lu Z., Abramson T., and Mackinnon R. (1994). Mutations in the K“ channel signature sequence. Biophys J. 66: 1061-1067 Lehninger, A.L. (1970). Biochemistry. Worth Publishers: New York, New York, p. 75. Mackinnon R., and Miller C. (1988). Mechanism of charybdotoxin block of Ca2t-activated Kt channels. J. Gen. Physiol. 91: 335-349. Mackinnon R., and Miller C. (1989). Mutant potassium channels with altered binding of charybdotoxin, a pore-blocking peptide inhibitor. Science 245: 1382-1385. Mackinnon R., Heginbotham L, and Abramson T. (1990). Mapping the receptor site for charybdotoxin, a pore-blocking potassium channel inhibitor. Neuron 5: 767-771. Mackinnon R., and Yellen G. (1990). Mutations affecting TEA blockade and ion permeation in voltage-activated Kt channels. Science 250: 276-279. Miller C. (1988). Competition for block of a Ca2t-activated K“ channel by charybdotoxin and tetraethylammonium. Neuron 1: 1003-1006. Rogowski R.S., Krueger B.K., Collins J.H., and Blaustein M.P. (1994) Tityustoxin Ka blocks voltage-gated noninactivating K“ channels and unblocks inactivating K“ channels blocked by a-dendrotoxin in synaptosomes. Proc. Natn. Acad. Sci. U.S.A. 91: 1475-1479. Rosenthal, J. C., Vickery R. G., and Gilly W. F. (1996). Molecular identification of Sgkv1A: a candidate for the delayed rectifier K channel in squid giant axon. J. Gen. Physiol., in press. Rudy, B. (1988). Diversity and ubiquity of K channels. Neurosci. 25: 729- 749. Strong, P.N. (1990). Potassium channel toxins. Pharmacol. Ther. 46: 137- 162. Summers M. D. and Smith G. E. (1988). A manual of methods for baculovirus vectors and insect cell culture procedures, 2nd ed. Texas A&M University Press: College Station, TX. Werkman T.R., Gustafson T.A., Rogowski R.S., Blaustein M.P., and Rogawski M.A. (1993). Tityustoxin-Ka, a structurally novel and highly potent K“ channel peptide toxin, interacts with the a-dendrotoxin binding site on the cloned Kv1.2 K“ channel. Mol. Pharm. 44: 430- 436. Yellen G., Jurman M., Abramson T., and Mackinnon R. (1991). Mutations affecting internal TEA blockade identify the probable pore-forming region of a Kt channel. Science. 251: 939-941. Yool A.J., and Schwarz T.L. (1991). Alteration of ionic selectivity of a K channel by mutation of the H5 region. Nature. 349: 700-704. FIGURE LEGENDS Figure 1 Several families of scorpion toxins share substantial structural similarity. All 9 toxins listed contain 37 - 39 amino acids. 6 cysteine residues are located in the same position in each toxin, suggesting a common tertiary structure. X's are spacers for alignment. Figure 2 300 nM TSTX-Ka (pH 7.6) produces potent block of Kt current in GFL neurons. (A) Current activated with depolarizing voltage steps from -80 mV to +40 mV was reduced 86%. A long 250 ms pulse (left) and a short 25 ms pulse (right) are shown. (B) 300 nM TSTX-Ka blocked K- current via a prolonged time course. Following control measurements, TSTX¬ Ka was applied (thick horizontal bar) in external solution perfused into the recording chamber. Washout was slow and incomplete. Points represent measurements of peak current elicited by a voltage step to +40 mV. Figure 3 300 nM TSTX-Ka (pH 7.6) does not markedly affect activation or deactivation kinetics, or voltage dependence of conduction. (A) A 25 ms voltage step from -80 mV to +40 mV elicited K“ current in toxin which could be scaled by a factor of 3.2 to match peak control current (right). The positive-going slope in toxin is largely unaffected and scales well to match the slope in control recordings. (B) Tail currents were elicited by a depolarizing prepulse to +40 mV followed by a test pulse between -130 mV and -50 mV. A single exponential fit to the resulting current gives time constants (t) which are largely unchanged in toxin. (C) Conduction-voltage 16 properties were ascertained by measuring the change in current evoked upon repolarization (to -80 mV) from a series of voltage steps to test potentials between -50 mV and +60 mV and dividing by the voltage change (dI/dV). All measurements were normalized to the maximum conductance value under each condition. Conductances are not markedly changed in toxin, indicating that the voltage dependence of the channels were unaffected by toxin. Figure 4 Inactivation is faster and more complete in 300 nM TSTX-Ka (pH 7.6). Kt current recordings are shown after 0, 11, 20, and 40 minutes in toxin. Current was activated by 250 ms depolarizing voltage steps from -80 mV to +40 mV. Each sweep is scaled to match the peak current level in the control recording. Scaling factors are shown in parenthesis. The time constant (t) decreases from 125 ms in control solution, to 68.9 ms after 40 min in 300 nM TSTX-Ka. In addition to the rate of inactivation, the amount of current at any time point is less than that in control, indicating that more current inactivates in toxin. Figure 5 A delayed time course of block and faster inactivation kinetics are also seen in Sf9 insect cells expressing SgkviA channels, the main component of Kt current in GFL neurons. (A) 300 nM TSTX-Ka applied at time zero blocks K“ current in Sf9 cells to a lesser extent than in GFL neurons. However, the slow rate of toxin effect seen in GFL cells is also exhibited in Sf9 cells. The half-time to block is 2-3 min in Sf9 (3-4 min in GFL). Peak Kt current was elicited using depolarizing voltage steps from -80 mV to +40 mV. (B) Inactivation is faster in Sf9 cells (t-45 ms ys. 58 ms in control) but not more complete as seen in GFL Ig. Actual current traces are shown in control solution and in 300 nM TSTX-Ka (scaled by a factor of 1.86 to match peak control current). Kt current was elicited as in (A). Figure 6 TsTX-Ka block in GFL neurons is pH dependent. The maximum block attained (30 - 40 min after application) with 300 nM TSTX-Ka at pH values between 5.6 and 9.6 are shown. At high pH block is maximal; at low pH block is minimal. A 1:1 binding curve is overlaid, centered at pH 6.0 and scaled by 0.86 to match the maximum block achieved at this dose. The dotted vertical line indicates the pH at which half of the blockable current is blocked (pH 6.0). Peak currents in toxin were measured at each pH following depolarizing 25 ms voltage steps from -80 mV to +40 mV and expressed as a percentage of control current. Figure 7 pH effects alone do not account for differential block at varying pH. (A) Actual current traces from the same cell, elicited by depolarizing voltage steps (from -80 mV to +40 mV) are shown at pH 5.6, 7.6, and 9.6. Peak current increases with increasing pH, however the difference between pH 5.6 and pH 7.6 (control) is less than 10%. (B) When peak Kt currents are scaled to match control, inactivation properties are largely similar at varying pH. FIGURES O OC L NNN 0 0 E 22 O QO a 0 E OOC 0 0 0 88888 GOHINO %) INJaano 10 8 E 5 883 GOAINO %) ZONVIORGNOS b 8 (su) -9 E 2 0 saa- o S eee 8 L kaa 3 - 8 1 8 A 1.0 0.8- 0.6- 0.4- 0.2 0.0- B FIGURE 5 10 TIME (min) CONTROL 300 nM TSTX-Ka (SCALED) 15 5 nA 50 ms 20 c 1.0- 0.6 - 0.4 - 0.2- 0.0 O FIGURE 6 10 . A B (SCALED) FIGURE 7 pH 9.6 S 0 pH 5.6 ktaav- S pH7.6 pH 5.6 pH 9.6 — 10 nA 50 ms