Novel Effects of Conus californicus Venom on Voltage-gated Potassium Channels Abstract Cone snails are predatory marine gastropods of the genus Conus which proliferated to over 500 species worldwide largely through the use of novel peptide toxins (conotoxins) that are injected into prey items through a harpooning apparatus. Conotoxins have evolved for maximal effect on prey species hunted by each cone snail species (Duda and Palumbi 1999). All cone snail venoms examined thus far contain peptides that act with great specificity on a variety of channels including K', Na“, Ca", and ACh receptors (Olivera et al. 1999). Conus californicus is indigenous to the California coastline. Venom samples were collected by extruding the contents of one venom duct into 0.5 ml external recording solution and were further diluted during experiments. Voltage-dependent K' currents were recorded from Shaker B K' channels, in which both N- and C-type inactivation were eliminated, using the whole-cell patch-clamp technique. When applied to the bathing solution, the venom was found to induce rapid inactivation of Ig. K current was reduced in à time- and concentration-dependent manner that fits satisfactorily with a simple model of open-channel block that applies to tertiary amines and quarternary ammonium (OA) that block K' channels at the inner mouth of the channel, but only after the channel has opened following a voltage step. Several lines of evidence are consistent with the venom acting on the external side of the membrane. First, the venom interacts with externally applied tetraethylammonium (TEA), an impermeant QA. An internal blocker would be expected to be unaffected by external TEA. Second, recovery from venom-induced Iy block is similar to that from QA-block in its dependence on voltage, but different from QA in that recovery shows no acceleration in solutions with elevated external K' concentration. Lack of this effect suggests that the venom component in question is not acting to block internally, because it cannot be dislodged by inward Ig at negative voltages. Third, channel-blocking action was preserved following dialysis of the venom at a 3500-molecular weight cut off. This suggests a large molecule such as a peptide is the active component. Introduction Cone snail venoms Peptidic components of the venoms of predatory Conus species have been used in previous studies to characterize external regions of ion channels because of the great affinity and specificity they show for binding sites. Many of these conopeptides target voltage-gated calcium and sodium channels, as well as the acetylcholine receptor. In this study, voltage gated Shaker K channels were shown to be affected by some component of the raw venom of Conus californicus in a way that has not been previously described. An active component in this venom, suspected to be a peptide, acts to induce fast inactivation of Shaker K' channels in a way that is similar to open-channel block that is demonstrated by many quaternary ammonium (QA) and tertiary amine compounds. However, unlike known open-channel blockers, the Conus venom component shows no accelerated recovery from block in solutions with high external K' concentration. Another inconsistency of the venom with known open channel blockers is the fact that venom effects are inhibited by the presence of external TEA, suggesting an interaction in the external region similar to that seen in k-conotoxins (Olivera et al. 1998), which target voltage-gated K“ channels as well. All other open channel blockers acting to cause rapid inactivation in K' channels act on the internal region of the channel. Methods Preparation of venom Venom was collected by removing the entire venom gland and its attached muscular bulb from the snail. The gland was then put on a glass plate and the contents exuded by applying pressure with a pair of bent forceps. Gland contents normally had a final volume of about 4 ul. The contents of the gland were then suspended in 293 external solution and placed in an Eppendorf tube to a final volume of 0.5 ml at a dilution of approximately 1:125. The solution containing venom was then centrifuged two times for approximately 25 seconds in an Eppendorf centrifuge model 5415 C at 13,000 rmp. The supernate was decanted and heated for 5 minutes at 100 °C to denature any active enzymes, and then placed in a 4 °C cooler for storage. Cells Human embryonic kidney (HEK) 293 cells were purchased from ATTC, Rockville, MD, and grown in Dulbecco's Modified Eagle Medium with 10% fetal bovine serum, 100 U/ml penicillin, and 100 ug/ml streptomycin. Cells for transfection were grown to 75% confluency and split 1:50 one day prior to transfection. Cells were transfected with plasmid vectors containing various K channel cDNAs using the calcium phosphate transfection methods of Chen and Okayama (1987). Electrophysiological Recording Conventional whole-cell patch clamp experiments were performed with the HEK 293 cells. External 293 solution contained 20 mM KCl, 180 mM NaCl, 4 mM CaCl2, 5 mM MgCI2, 10 HEPES, 1 mM EDTA with a pH of 7.2 (adjusted with NaOH). Currents were recorded at room temperature using a whole-cell voltage clamp amplifier and a lowpass filter set at 15 KHz. Electrode resistance was 1.0 to 1.8 MQ when filled with an internal solution containing 60 mM KF, 140 mM KCl, 1 mM MgCI2, 8 mM EGTA, and 10 mM HEPES for a pH of 7.3 (390 mOsm). All currents were compensated for linear capacitive and ionic currents using a standard P/4 subtraction protocol. Voltage commands were generated using VCLAMP/WCDATA software and a direct memory access (DMA) interface. Data acquisition and analysis were performed with the same program. Holding potential was -80 mV throughout all experiments. Recordings were made at room temperature (20-22°C). Application of venom The setup for HEK 293 recording included an Air Cadet vacuum/pressure station (Cole-Parmer Instrument) to allow exchange of the external solution. A volume of 0.5 to 1.5 ml of the experimental solution was perfused into the chamber with this system while aspirating the overflow. Results Conus californicus venom component acts as an open channel blocker of Iy Control Ig (Fig. IA) was evoked in response to a 740-ms depolarization to +20 mV and represents maximal activation of K' conductance. At each venom concentration. Iy begins to activate normally in response to depolarization but then becomes blocked in a time-dependent manner, declining exponentially to a steady state at full concentration that shows inactivation of 85% of channels. In Figure 1B the concentration dependence of the venom effect is seen in which greater venom concentrations produce a faster time constant of inactivation and the amount of steady-state block is increased. Figure 2 shows that steady-state block is independent of voltage at all concentrations tested. The blocking behavior produced appears to be similar to others described by a simple model of open-channel block, such as that seen in the QA tetrapentylammonium (TPeA) as developed by Armstrong (1971), that assumes no block at rest. Steady state I in Conus venom represents an equilibrium between open and blocked channels with an equilibrium constant Kd = (L/K) where k and L represent the rates of drug block and dissociation from the open channel and k is proportional to the venom concentration. 1o determine whether the acting venom compound is binding singularly at a specific region of the K+ channel, a model must be used that assumes 1:1 blocking of agonist at the binding site. To make this conclusion it is first assumed that by varving relative concentrations of venom you are altering the blocking rate (k) of the venom. This model is tested when the fraction of channels blocked at the three different concentrations is plotted on an isotherm that assumes 1:1 binding of receptor and agonist (Fig. 3). Another model was also used in which voltage-dependent toxin block was analyzed using time constant of block and on and off rates of toxin binding. Time constant rates were analyzed from-20 mV to +20 mV at each of the three concentrations of venom tested and compared to the time constants that would be predicted by the equation T.... 17Kolvenom + K.g. Actual measured values for time constant at each of the concentrations were found to fit reasonably well with the values predicted by the equation: ta - lyenom,1(1no ontol-Lyeno t venom, (1+(yenom/Teontro-Ivenom)2 Conus venom component interacts with TEA at an external site A conventional conopeptide of 25-40 amino acids in length (m.w. 3500-5600) would be large in size and unable to cross the cell membrane and act internally in the way modeled by the QA, TPeA. In order to determine at which side of the membrane the relevant component in Conus venom is acting, a competition experiment was used with the small QA tetraethylammonium (TEA). TEA is unable to cross the cell membrane but can act at either outer or inner mouth of the channel to produce a time-independent decrease in L This decrease in Ig is caused by TEA's rapid movement in and out of the pore region that causes à transient block that also prevents the channel from inactivating. Competition between TEA and an intrinsic inactivation mechanism was used to localize N- and C-type inactivation in Shaker B channels to the inner and outer mouth of the channel, respectively (Choi et al. 1991). 5 In the present study, it was assumed that if Conus venom and TEA were applied concurrently in the external solution, then an inhibitory effect of TEA on the action of venom would be indicative of competitive interaction between the two agents at the outer mouth of the channel. 90 mM TEA added to the external solution (in the absence of venom) has the effect of blocking approximately 60% of steady state Ig (Fig 4A). When venom is added and TEA has been washed out, there is a decrease in the time constant of inactivation to 44 ms. and an increase the fraction of blocked channels to about 85% of control (Fig 4B), as would be expected when the competing agonist has been removed. Now it is necessary to compare channels exposed to venom only to those exposed to venom and TEA simultaneously. When channels are exposed to a solution containing both venom and TEA, the results show both a increase in time constant of inactivation from the amount that is seen in venom without TEA to around 100 ms (Fig. 4B, 40), as well as à 25% decrease in the fraction of channels blocked from maximum Ig to steady-state (Fig. 5). This interaction between external TEA and the venom suggest an external binding site for the venom component. In a model that assumes 1:1 competition between two blocking molecules (Choi et al. 1991), two separate experiments reasonable agreement with the 1:1 competition hypothesis (Fig. 6). Recovery from venom component block is slow and independent of external K concentration Recovery of 1g from block in the presence of Conus venom was tested using a double pulse procedure as described by Horrigan and Gilly (1996). Recovery from venom component block was found to be faster at more negative potentials (Fig. 7). This follows the expected model for recovery in which larger inward Ig acts to remove open channel blockers from their internally bound position (Armstrong, 1971). Using the same recovery procedure it was found that at high external Kt concentrations, the rate of recovery from block was not substantially affected (Fig, 8). Larger inward lg does not relieve block, which does not agree with the expected model for recovery from block in K' channels. Discussion This study demonstrates that a component of the venom of Conus californicus blocks Ig in Shaker B channels. The mechanism of this block was shown to involve a direct interaction of the venom component with K' channels in a time-dependent manner that is well described by the model of open channel block developed by Armstrong (1971) to describe the effects of quaternary ammonium ions (QA). However, unlike the QA open¬ channel blocker model in which the blocker crosses the membrane and binds to the internal region, the Conus venom component appears to bind externally. Consistent with the theory that the Conus venom is acting externally are the apparent interaction of external tetraethylammonium (TEA) and the venom component, and the independence of recovery from block from changes in external K. Based on a competition model established for TEA and an intrinsic open channel c- type inactivator of Shaker channels, it seems competition exists between the Conus venom component and external TEA. If the full effect of the toxin were seen in the presence of external TEA, it would be assumed that the venom component was acting on the internal region of the channel. However, in the presence of external TEA, the venom was unable to exert its full effect of both the rate of inactivation and the degree of steady-state block of channels. Although recovery was found to be voltage dependent in the way described for öther open channel blockers, experiments performed in which recovery was tested at O. 20. 100 and 200 K showed no significant acceleration of recovery at higher K' concentrations. The results from the high external K' experiment are inconsistent with models shown for internally bound open-channel blockers in which the blocking compound is destabilized and kicked out of its position of blocking the pore region by increased inward Iy. A test was needed to determine if a externally-bound protein was indeed the active component in Conus venom. For this experiment, dialysis tubing with a 3500-molecular weight cut off was used. Here, a sample of 0.5 ml (toxin + 293 external) was dialyzed for 26 hours with solution changes of 1 liter approximately every two hours, 13 in distilled water and a final two times in 293 external solution. If the active component was a traditional open channel blocker such as a QA or tertiary amine of low m.w. then these compounds should have been dialyzed out of the sample. However, large compounds such as conopeptides would be retained. Following the dialysis, the sample showed the full effect of a non-dialyzed toxin sample on the Shaker B channels. A large conopeptide would be expected to bind externally to the channel, as it would be unable to cross the cell membrane. Together, these results support a hypothesis in which a conopeptide is the active component in Conus californicus venom. More tests must be run in order to firmly establish the conclusion that a peptide is working in the external region of the channel, but if this holds true, then perhaps a new class of conotoxins could be identified. A conotoxin agonist working from the external region of the pore to produce an effect like open-channel block would indeed be novel among the conotoxins documented thus far. — Fraction Inactivated 0 Figure 3: Concentration-dependence of steady-state block 1.( 0.5 O- 0.01 0.1 1.0 10.0 Relative Venom Concentration C Figure 4: External TEA (90 mM) reduces the effects of cone snail venom Control TEA wash-out +venom TE 2 nA TEA+venom 200 ms 200 ms pan n nn n nen nenenenen Control ien nennntene +venom 1nA TEA +venom 200 ms 8 O Fraction Inactivated Channels — N A ed t Fold-slowing in TEA N 1 - Fraction Recovered 3N o . 1 - Fraction Recovered — D O D Do -O X X O O E 0 Figure Legends Figure 1: Basic venom effect. Externally applied Conus venom causes time-dependent block of Shaker B K' channels. A) Ig evoked by 250-ms depolarizations to -10 mV in the absence (control) and presence of venom. B) Ig at +20 mV in the presence of the indicated relative concentrations of Conus venom. Control record (0.0 venom concentration) was obtained immediately before application of the venom at 0.1 dilution. Traces were recorded after envenomation in sequential order of increasing relative concentration. Figure 2: Voltage-dependence of steady-state block of Ig by Conus venom. Calculated by taking 1 minus the steady state current at the end of a 740-ms pulse over the peak current (1-L) versus voltage. As relative venom concentration rises from 0.1 to 1.0, the fraction of blocked channels increases. A mild, U-shaped voltage¬ dependence exists for all venom concentrations and is also evident in the control condition. Figure 3: Concentration-dependence of steady state Iy block. Mean fraction of steady state block was calculated from data between -20 mV and +20 mV at each of the three relative concentrations of venom. These three values are compared with a 1:1 binding isotherm fitted to the data points by eye. Figure 4: External TEA (90 mM) reduces the effects of Conus venom. A) Depolarizations to +20 mV in control solution and following application of 90 mM TEA in which Ig is reduced by 55% and inactivation of channels is blocked. B) Ig evoked from depolarizations in the same cell. First, TEA is applied concurrently with toxin and a slight decrease in time constant of inactivation and depression of Ig is seen in comparison to the TEA-only sweep. Next, only venom is perfused in (no TEA). After TEA is washed out the effects of the venom are established immediately, with a large decrease in time constant of inactivation and large increase in fraction of channels blocked from peak Iy to steady-state Iy C Another TEA competition experiment at 740-ms depolarizations to +20 mV. Here, venom is first added to the solution to produce the fast inactivation and increased block effect as in B. Next, TEA is added with toxin and the competition effect is again seen. TEA acts to slow down (decrease) the time constant of inactivation, and shows a decrease in the fraction of channels blocked. Figure 5: Inhibition of venom-block by external TEA. The fraction of steady-state Iy was decreased at all voltages by external TEA. Figure 6: Competition model for TEA and venom. The solid line predicts a 1:1 competition between two agonists for a blocking site at the channel. Two separate experiments in which TEA and venom were applied simultaneously show some deviation from the 1:1 blocking competition. This is calculated by the equation 1-(T/Tyan) ys. 1-(LEAyenon) from-20 mV to +20 mV in each experiment (one of the voltage steps could not be accounted for in the solid circle experiment). Figure 7: Recovery of Ig from Conus venom block is slow and dependent upon holding potential. Recovery of peak current is shown using a double pulse protocol as described in Horrigan and Gilly (1996). Currents evoked by test pulses after 250 ms recovery intervals at -120 and -80 mV demonstrate that recovery of Iy from venom block is accelerated at more hyperpolarized holding potentials. Figure 8: Recovery of Ig from Conus venom is independent of external K' concentrations. Recovery from cone snail venom block was tested at 0 and 100 K' external concentrations at a holding potential of -80 mV. No appreciable difference in recovery of ly can be seen between the two. References Armstrong, C.M. 1971. J. Gen. Physiol. 58: 413-437 Chen, C., Okayama. 1987. High-efficiency transformation of mammalian cells by plasmid DNA. Molecular Cellular Biology. 7: 2745-2752 Choi, L.L., R.W. Aldrich, and G. Yellen. 1991. Tetraethylammonium blockade distinguishes two inactivation mechanisms in voltage-gated K' channels. Proc. Natl. Acad. Sci. USA. 88: 5092-5095 Duda, T.F., and S.R. Palumbi. 1999. Molecular genetics of ecological diversification: Duplication and rapid evolution of toxin genes of the venomous gastropod Conus. Proc. Natl. Acad. Sci. USA. 96: 6820-6823 Horrigan, F.T. and W.F. Gilly. 1996. Methadone Block of K' Current in Squid Giant Fiber Lobe Neurons. J. Gen. Physiol. 107: 243-260 Mclntosh, J.M., B.M. Olivera, and L.J. Cruz. 1999. Conus Peptides as Probes for Ion Channels. Methods in Enzymology. 294: 605-624 Shon, K., M. Stocker, H. Terlau, W. Stuhmer, R. Jacobsen, C. Walker, M. Grilley, M. Watkins, D.R.Hillyard, W.R. Gray, and B.M. Olivera. 1998. k-Conotoxin PVIIA Is a Peptide Inhibiting the Shaker K“ Channel. Journal of Biological Chemistry. 273:33-38 Acknowledgements 1 would like to thank Professor Gilly and Mat Brock for helping make this project a successful and enjoyable endeavor. Jon Sack, Joseph Schultz and Henry Jerng deserve much credit for their tolerance of my questions and occupation of their time and space in the lab. Thanks also to Jim Watanabe and Joe Tyburzcy for collecting the snails and making the project possible. Spring quarter ’99 was without a doubt the most enjoyable experience Thave had at Stanford and this is entirely due to the awesome personalities and abilities of everybody in the Gilly lab.